Control Loop Performance

Monitoring

© 2020 General Electric Company

Contents

Chapter 1: Control Loop Performance Monitoring Overview 1

What is Process Control? 2
What is a Control Loop? 2
What is a PID Control Loop? 2
What is Control Loop Performance Monitoring? 3
The Asset Model in CLPM 4
CLPM Components 5
User Workflow in CLPM 6
CLPM Terminology 7

Chapter 2: Get Started with CLPM 8

Getting Started with CLPM 9
Prerequisites for CLPM 9
Define the Asset Model 10
Deploy Data Collection 10
Deploy CLPM Analytics 15
Access Loop and Fleet Dashboards 45

Chapter 3: The Fleet Report 46

The Fleet Report 47
Access the Fleet Report 47
Modify the Date Range for a Fleet Report 47
Average Control Loop Performance Chart 48
The Diagnostic Alerts Filtering Smart Filter 48
The Control Loop Performance Table 49

Chapter 4: The Loop Report 52

About the Loop Report 53
Access the Loop Report 53
Modify the Date for a Loop Report 54

ii Control Loop Performance Monitoring

 The Process Variable Performance Chart 54
The Manipulated Variable Distribution Chart 55
The Control Overview Chart 57
The Error Distribution Chart 59
The Control Mode Summary Table 60
The Controller Performance Table 60
The PV Performance Table 63
The Error Statistics Table 63
The Controller Configuration Table 64

Chapter 5: The Loop Analysis Template 65

About the Loop Analysis Template 66
Access the Loop Analysis Template 66
Modify the Date for a Loop Analysis 66

Chapter 6: Manage Loops 68

Add a Loop 69
Modify a Loop 69
Delete a Loop 70

Chapter 7: Reference 71
CLPM Terminology 72
Control Loop Asset Definition 72
KPI Reference for CLPM 107

Chapter 8: Release Notes 126

Control Loop Performance Monitoring Release Notes Q2 2019 127
Control Loop Performance Monitoring Release Notes Q1 2019 127

iii
Copyright GE Digital
© 2020 General Electric Company.

GE, the GE Monogram, and Predix are either registered trademarks or trademarks of General Electric
Company. All other trademarks are the property of their respective owners.
This document may contain Confidential/Proprietary information of General Electric Company and/or its
suppliers or vendors. Distribution or reproduction is prohibited without permission.
THIS DOCUMENT AND ITS CONTENTS ARE PROVIDED "AS IS," WITH NO REPRESENTATION OR
WARRANTIES OF ANY KIND, WHETHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO
WARRANTIES OF DESIGN, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. ALL OTHER
LIABILITY ARISING FROM RELIANCE UPON ANY INFORMATION CONTAINED HEREIN IS EXPRESSLY
DISCLAIMED.
Access to and use of the software described in this document is conditioned on acceptance of the End
User License Agreement and compliance with its terms.

iv © 2020 General Electric Company

 Chapter

1
Control Loop Performance Monitoring Overview
Topics:

• What is Process Control?

• What is a Control Loop?
• What is a PID Control Loop?
• What is Control Loop
Performance Monitoring?
• The Asset Model in CLPM
• CLPM Components
• User Workflow in CLPM
• CLPM Terminology

© 2020 General Electric Company 1

What is Process Control?
Process control is an engineering discipline for maintaining the output of a specific process within a
desired range.
A control system is a device or set of devices used to manage the behavior of other devices or systems in
an industrial process facility.

What is a Control Loop?

A control loop is the fundamental building block of industrial control systems. It consists of all the physical
components and control functions necessary to automatically adjust the value of a measured Process
Variable (PV) to equal the value of a desired Setpoint (SP).
A control loop includes the process sensor, the controller function, and the Final Control Element (FCE), all
of which are required for automatic control.
The difference between the SP value and the current value of the PV is called the Error, and this is used by
the control algorithm to calculate adjustments to the Manipulated Variable (MV). The MV is sent to the
FCE (also referred to as the actuator) to bring the PV back to the SP value, using a closed loop feedback
mechanism. This mechanism allows control systems to adapt to varying process operating circumstances
and disturbances. Disturbances can sometimes be measured and compensated for by means of a
Disturbance Variable (DV).

What is a PID Control Loop?

A PID control loop is a generic control loop feedback mechanism that uses a well-known PID(F) algorithm.
It is very common in industrial control systems.
The PIDF (Proportional–Integral–Derivative-Filter) control algorithm involves four separate tuning
constant parameters:
• Proportional value (P)
◦ The variable control required in proportion to the current needs of the system.

2 © 2020 General Electric Company

 • Integral value (I)
◦ The integral control that applies ever-increasing control actions, until the error is reduced to zero.
• Derivative value (D)
◦ The derivative value of the algorithm is determined by how quickly the error changes over time.
• Filter value (F)
◦ Adding a filter (F) to the PID controller helps to remove noise signals from the system and enables
better control under some conditions.
Many different implementations (variations) of the PID(F) algorithm are used in industry today.

What is PID Control Loop Tuning?

When a disturbance to the process occurs, the task of the controller is to reduce the error signal to zero
and restore the PV back to the SP value, as effectively, efficiently, and quickly as possible. Choosing the
right PIDF values for the parameters in the PIDF controller algorithm allows the controller to achieve this
goal in the most effective way possible. This is referred to as loop tuning: selecting the best PIDF values to
get the most accurate and effective control loop performance behavior. If the loop is tuned properly, the
response time to disturbances is reduced. If the loop is tuned too aggressively, cycling and overshoot of
the setpoint value may occur.

What is Control Loop Performance Monitoring?

Control Loop Performance Monitoring (CLPM) provides a solution to monitor the performance of an
enterprise fleet of industrial process control loops, to detect and identify underperforming control loops,
and to help to diagnose control loop problems.
Industrial process plants typically have hundreds or thousands of control loops configured inside PLC
(Programmable Logic Controller), DCS (Distributed Control System), and SCADA (Supervisory Control and
Data Acquisition) systems to automate and control plant operations.
Control Loops are located at the interface between equipment and process. They also operate industrial
equipment to form an industrial process.
The stability, reliability, and optimal performance of an industrial process is dependent on the stability
and performance of the fleet of control loops used to operate and control that process. Industrial process

© 2020 General Electric Company 3

 plants are also capital-intensive assets. Suboptimal control loops can directly impact equipment
reliability, equipment lifetime, and operational performance.
It is therefore essential to manage the performance of control loops as valuable assets in an industrial
enterprise facility that should be continuously monitored and optimized.

Evaluating Control Loop Performance

• Control loops are performing well when the monitored process variable (PV) varies within the defined
upper and lower control limits from the desired SP value.
• Control loops are performing poorly where the monitored PV functions outside of the desired process
control limits.

The Asset Model in CLPM

The asset model is central to CLPM in a number of ways:
• Loops are defined as assets, according to a predefined template/definition for a control loop.
• Deployment of data collections in Edge Analytics is driven by the asset model.
• CLPM configuration is driven from the asset.
• Loop thresholds and control limits are defined on the asset.
• Edge data is extracted from the asset.
• Actions on a control mode are defined on the asset.
Consult the Control Loop Asset Definition for details.

4 © 2020 General Electric Company

CLPM Components
CLPM is made up of the following components working together.

Figure 1: CLPM components

These components enable the following functionality.

1. Loops are defined as assets, according to a predefined template for a control loop.
2. Selected tags are collected into TimeSeries for each loop.
3. Loop analytics are deployed.
4. Alerts produced by these analytics are triggered under conditions defined in analytic configuration.
5. Data produced by these analytics is viewed via the Fleet Dashboard and Loop Dashboard.
6. Issues identified with loops (cases) are resolved.

© 2020 General Electric Company 5

User Workflow in CLPM
Following the CLPM workflow / process described below helps users monitor and optimize the
performance of an enterprise fleet of control loops. Underperforming control loops can be identified,
prioritized, and optimized.

Figure 2: User workflow / process

1. The Fleet Dashboard

Use the Fleet Dashboard to rank the performance of all control loops contained in any level of an
enterprise asset model (Enterprise, Segment, Site, etc.) over a specified time period. This helps in
identifying underperforming control loops and prioritizing them for optimization. Consult the Fleet
Dashboard section of this documentation for details.

OR

1. Alerts
Alerts notify you about potential problems with control loops. Consult the section of this documentation
on deploying CLPM analytics for details on analytics that trigger alerts and how to configure them.

2. The Loop Dashboard

Use the Loop Dashboard to visualize the performance of any individual control loop in detail over a
specified time period. This helps in assessing control loop performance and diagnosing potential causes of
suboptimal performance. Consult the Loop Dashboard section of this documentation for details.

3. Loop KPI Analysis

Analyze the historical performance of loop KPIs over any historical time period to gain further insight.
Loop KPIs can be assessed in relation to other loop KPIs or compared with the same loop KPIs calculated
using other data and other tags available in Predix TimeSeries. Consult the KPI Reference for CLPM for
more on loop KPIs.

6 © 2020 General Electric Company

 4. Case Management
Action, collaborate on, and resolve any issues identified with control loops. Consult the Application
Analytics documentation for details.

CLPM Terminology
The following terms are frequently used in control loop performance monitoring (CLPM).

Term Description

CLPM Control Loop Performance Monitoring

PV (Process Variable) This is the measured process variable, reflecting the actual state
of the process that is being controlled.

SP (Setpoint) This is the desired value (target value) to which the PV is

controlled.

Error This is the difference between the PV and SP. It is calculated as

Error = PV-SP. Note that this is not an absolute value.

Controller The subsystem containing the control algorithm that

manipulates the MV to control the PV to the SP.

MV (Manipulated Variable) The variable manipulated (adjusted) by the controller to control

the PV to the SP.

DV (Disturbance Variable) A measured disturbance variable that can be used by the

controller to compensate for the impact of the disturbance on
the PV.

P (Proportional Value) Normally the Proportional Gain value of the PID controller.

I (Integral Value) Normally the Integral Gain value of the PID controller.

D (Derivative Value) Normally the Derivative Gain value of the PID controller.

F (Filter Value) Normally the Filter Time Constant value of the PID controller.

Final Control Element (FCE)/ Actuator The device that is physically affected by a change in the MV so
as to have an impact on the process.

© 2020 General Electric Company 7

 Chapter

2
Get Started with CLPM
Topics:

• Getting Started with CLPM

• Prerequisites for CLPM
• Define the Asset Model
• Deploy Data Collection
• Deploy CLPM Analytics
• Access Loop and Fleet
Dashboards

8 © 2020 General Electric Company

Getting Started with CLPM
This section provides instructions for getting started with CLPM.
At the high level, this involves the following steps, to be performed by the CLPM solution administrator.

1. Set up the prerequisites for CLPM.

2. Define the asset model.
3. Deploy data collection. This includes the following:
a. Set up control mode logic.
4. Deploy CLPM analytics.
5. Access loop and fleet dashboards.

Prerequisites for CLPM

To get started with CLPM, first ensure that you have met these prerequisites.

Tenancy Requirements
Ensure that your OPM application instance has the following apps and services set up:
• Application Analytics
• The Asset Model Service
• The Time Series Service
• The Dashboard Service
• Predix Event Hub
• Predix Insights
• KPI Management
• Analysis Service
• Alert Service

Analytics Requirements
Ensure that you have access to the CLPM analytics in the Analytics Catalog.

Browser Requirements
Only the Chrome 43 or higher browser is supported for accessing the Loop and Fleet Dashboards.

© 2020 General Electric Company 9

Define the Asset Model
Construct the asset ingestion file required by Application Analytics and perform asset ingestion.
The asset ingestion files (there can be several files or they can be combined into one) are JSON files that
specify customer assets and their hierarchical structure or asset model. In the context of CLPM, the
assets in this file are control loops that together define an industrial control system.
The basic structure of the files is as outlined in the APM Assets documentation for asset ingestion.
In addition to the required attributes that are standard for all such files, files describing control loop
assets require custom attributes defined in the Control Loop Asset Definition for CLPM that you need to
specify.
Note: It is important to ensure that your asset ingestion file conforms to the control loop asset definition
for CLPM. If your file does not conform to this definition, CLPM will be unable to function correctly.
To define the asset model, carefully follow the instructions given with the Control Loop Asset Definition
for CLPM.

Deploy Data Collection

Collect data into TimeSeries for all the analytic input tags as defined in your asset model. The
recommended rate for data collection is every 5 seconds or slower.
Once you have set up data collection, you are ready to set up control mode logic.

Set Up Control Mode Logic

Set up control mode logic for CLPM by creating and deploying your Control Mode Analytic. Use the Python
example provided here to create a Predix Insights-based analytic.

What Is the Control Mode Analytic?

CLPM analytics require an input tag that describes the current control mode for each loop. The control
mode tag must contain values of type double and must be one of the following:

Tag Value Description

1 A tag value of 1 indicates a control loop that is currently

operating manually. The controller is not controlling the process.

2 A tag value of 2 indicates a control loop where the process is

currently being controlled by the controller.

3 A tag value of 3 indicates a controller that is currently set to

CASCADE.

4 A tag value of 4 indicates a process that is currently shut down.

Note: Most KPIs will not be calculated for this control mode.

CLPM provides an example Predix Insights analytic to get you started towards developing your own
control mode analytic. The resulting analytic will vary depending on your available data, but each control
mode analytic must produce a result that matches one of the tag values in the table above.

10 © 2020 General Electric Company

 Example Control Mode Analytic
The Python script that follows can be used as the basis for you Control Mode Analytic.

'''
Created on Aug 30, 2018

@author:
'''
from pyspark.sql import SparkSession
from pyspark.sql import *
from pyspark.sql import functions
from pyspark.sql.types import *
from pyspark.sql.dataframe import *
from pyspark.sql.functions import *
import sys
import time
from datetime import datetime

class ControlModeScript():
# ################ DO NOT EDIT ########################
MANUAL = "1.0"
AUTO = "2.0"
CASCADE = "3.0"
SHUTDOWN = "4.0"
QUALITY_GOOD = "3"
QUALITY_BAD = "0"
# ################ END ########################

# ################ MODIFY THIS FUNCTION ########################

# When modifying this function, TQVs for the custom tags (Tag_VV,
Tag_XX, Tag_YY, Tag_ZZ) can be obtained as follows:
# * timestamp is common amongst all tags
# * quality_XX is quality of Tag_XX, quality_YY is quality of
Tag_YY and so on
# * value_XX is value of Tag_XX, value_YY is value of Tag_YY
and so on
#
def control_mode_query(self):

# The example SQL query below does the following:

# select timestamp as the timestamp of the tag
# always select quality as good
# if quality is bad
# select value as shutdown
# else if value of tag_XX is 1
# select value as MANUAL
# else if value of tag_YY is 1
# select value as AUTO
# else if value of tag_ZZ is 1
# select value as CASCADE
# else
# select value as SHUTDOWN
string = "timestamp as timestamp, CASE WHEN quality_XX != 3 OR
quality_YY != 3 OR quality_ZZ != 3 OR quality_VV != 3 THEN " +
self.SHUTDOWN
string += " WHEN value_XX = 1 THEN "+self.MANUAL+" WHEN
value_YY = 1 THEN "+self.AUTO+" WHEN value_ZZ = 1 THEN "+self.CASCADE
+" WHEN value_VV = 1 THEN "+self.SHUTDOWN+" ELSE "+self.SHUTDOWN+" END
as value," + self.QUALITY_GOOD +" as quality"
return string

© 2020 General Electric Company 11

 # ################ END ########################

# ################ DO NOT EDIT ########################

def run_job(self, spark_session, runtime_config, job_json,
context_dict, logger):
try:
spark = spark_session
logger.info("Starting analytic...")
configContext = context_dict["configDS"]
tsContext = context_dict["timeseriesReadDS"]
configDF = configContext.sql("select * from " +
context_dict["configDS"].table_name)
configDF.createOrReplaceTempView("configDF")
tsDF = tsContext.sql("select * from " +
context_dict["timeseriesReadDS"].table_name)
tsDF.createOrReplaceTempView("timeseriesReadDF")

Tag_XXDDF = tsContext.sql("SELECT c.AssetSourceKey as

asset, t.timestamp as timestamp, t.value as value_XX, t.quality as
quality_XX FROM `timeseriesReadDF`"
+ " t JOIN `configDF` c on t.tag = c.MappingValue
WHERE c.MappingKey = 'Tag_XX' DISTRIBUTE BY asset")
Tag_XXDDF.createOrReplaceTempView("tag_XXDDF")

Tag_YYDDF = tsContext.sql("SELECT c.AssetSourceKey as

asset, t.timestamp as timestamp, t.value as value_YY, t.quality as
quality_YY FROM `timeseriesReadDF"
+ "` t JOIN `configDF` c
on t.tag = c.MappingValue WHERE c.MappingKey = 'Tag_YY' DISTRIBUTE BY
asset")
Tag_YYDDF.createOrReplaceTempView("tag_YYDDF")

Tag_ZZDDF = tsContext.sql("SELECT c.AssetSourceKey as

asset, t.timestamp as timestamp, t.value as value_ZZ, t.quality as
quality_ZZ FROM `timeseriesReadDF"
+ "` t join `configDF` c
on t.tag = c.MappingValue WHERE c.MappingKey = 'Tag_ZZ' DISTRIBUTE BY
asset")
Tag_ZZDDF.createOrReplaceTempView("tag_ZZDDF")

Tag_VVDDF = tsContext.sql("SELECT c.AssetSourceKey as

asset, t.timestamp as timestamp, t.value as value_VV, t.quality as
quality_VV FROM `timeseriesReadDF"
+ "` t JOIN `configDF` c
on t.tag = c.MappingValue WHERE c.MappingKey = 'Tag_VV' DISTRIBUTE BY
asset")
Tag_VVDDF.createOrReplaceTempView("tag_VVDDF")

Tag_XXYYDF = tsContext.sql("SELECT
COALESCE(tag_XXDDF.asset, tag_YYDDF.asset) as
asset,COALESCE(tag_XXDDF.timestamp,tag_YYDDF.timestamp) as timestamp,
tag_XXDDF.value_XX, tag_XXDDF.quality_XX,
tag_YYDDF.quality_YY,tag_YYDDF.value_YY FROM tag_XXDDF FULL OUTER JOIN
tag_YYDDF ON tag_XXDDF.asset = tag_YYDDF.asset and tag_XXDDF.timestamp
= tag_YYDDF.timestamp DISTRIBUTE BY asset")
Tag_ZZVVDF = tsContext.sql("SELECT
COALESCE(tag_ZZDDF.asset, tag_VVDDF.asset) as
asset,COALESCE(tag_ZZDDF.timestamp,tag_VVDDF.timestamp) as
timestamp,tag_ZZDDF.quality_ZZ, tag_ZZDDF.value_ZZ,
tag_VVDDF.quality_VV,tag_VVDDF.value_VV FROM tag_ZZDDF FULL OUTER JOIN

12 © 2020 General Electric Company

 tag_VVDDF ON tag_ZZDDF.asset = tag_VVDDF.asset and tag_ZZDDF.timestamp
= tag_VVDDF.timestamp DISTRIBUTE BY asset")
Tag_XXYYDF.createOrReplaceTempView("tag_XXYYDF")
Tag_ZZVVDF.createOrReplaceTempView("tag_ZZVVDF")
timeseriesDF = tsContext.sql("SELECT
COALESCE(tag_XXYYDF.asset,tag_ZZVVDF.asset) as asset,
COALESCE(tag_XXYYDF.timestamp,tag_ZZVVDF.timestamp) as timestamp,
tag_XXYYDF.quality_XX, tag_XXYYDF.value_XX,
tag_XXYYDF.quality_YY,tag_XXYYDF.value_YY,
tag_ZZVVDF.quality_ZZ,tag_ZZVVDF.value_ZZ,
tag_ZZVVDF.quality_VV,tag_ZZVVDF.value_VV FROM tag_XXYYDF FULL OUTER
JOIN tag_ZZVVDF ON tag_XXYYDF.asset = tag_ZZVVDF.asset and
tag_XXYYDF.timestamp = tag_ZZVVDF.timestamp")
timeseriesDF.createOrReplaceTempView("collectedDS")

queryString = "SELECT configDF.MappingValue as tag, "

queryString += self.control_mode_query()
queryString += " FROM `collectedDS` JOIN `configDF` on
collectedDS.asset = configDF.assetSourceKey WHERE configDF.mappingType
= 'OutputMappings'"
resultDF = tsContext.sql(queryString)
resultDF.createOrReplaceTempView("resultDF")
timeseriesWriteDF = tsContext.sql("SELECT tag,
timestamp, CAST((value) as double) as value, quality FROM `resultDF`")
logger.info("Returning result...")
result = {"timeseriesWriteDS" : timeseriesWriteDF}
return result

except Exception as e:
print("ERROR RETURNED")
logger.info("Error: " + str(e))
exc_tb = sys.exc_info()[2]
logger.info("Line number: " + str(exc_tb.tb_lineno))
# ################ END ########################

Control Mode Constants

The script includes the following constants to represent the various control modes:
• MANUAL
• AUTO
• CASCADE
• SHUTDOWN

Available Tags
The following configurable tags can be queried in the script. These tags will be mapped to tags in Predix
TimeSeries.
• Tag_VV
• Tag_XX
• Tag_YY
• Tag_ZZ
For these tags, the timestamp, quality, and value (T,Q,V) are represented in a table with the following
columns:

© 2020 General Electric Company 13

 Timestamp value_VV quality_VV value_XX quality_XX value_YY quality_YY value_ZZ quality_ZZ

Note: The timestamp is common among all these tags.

Example Query
The script includes the following example query:

def control_mode_query(self):
string = "timestamp as timestamp, CASE WHEN quality_XX != 3 OR
quality_YY != 3 OR quality_ZZ != 3 OR quality_VV != 3 THEN " +
self.SHUTDOWN
string += " WHEN value_XX = 1 THEN "+self.MANUAL+" WHEN
value_YY = 1 THEN "+self.AUTO+" WHEN value_ZZ = 1 THEN "+self.CASCADE
+" WHEN value_VV = 1 THEN "+self.SHUTDOWN+" ELSE "+self.SHUTDOWN+" END
as value," + self.QUALITY_GOOD +" as quality"
return string

The preceding query does the following:

1. Sets the control mode timestamp to the timestamp used by these tags.
2. Sets the control mode quality to good.
3. Sets the control mode value as follows:
a. If the quality of any of the tags is bad, sets the control mode value to SHUTDOWN.
b. Else if the value of tag_XX is 1, sets the control mode value to MANUAL.
c. Else if the value of tag_YY is 1, sets the control mode value to AUTO.
d. Else if the value of tag_ZZ is 1, sets the control mode value to CASCADE.
e. Else, sets the control mode value to SHUTDOWN.
You should create your own query, based on your own business needs.

Modify the Control Mode Analytic

Modify the control_mode_query() function in the Python script, building up a SQL string to query
one or more of the available tags and (based on the T,Q,V of these tags) to return a T,Q,V for the control
mode.
Note: You should modify ONLY the control_mode_query() function in this template. Do not modify
anything else in the template.
Note: You MUST use Tag_VV in your query. Using the other available tags in your query is optional.

Create Your Analytic Template

Once you have modified the Python script to include the control mode logic you require, do the following.

Procedure
1. For the Control Mode Template Analytic, create a Spark analytic called OPM-CLPM-Control_Mode,
following the guidance given in the Spark and Application Analytics documentation.
2. Publish the analytic to the Analytics Catalog.
3. Set up the input and output definitions as follows:

14 © 2020 General Electric Company

 Table 1: INPUT DEFINITION: TAGS

Name Type Required

Tag_VV Double Yes

Tag_XX Double No

Tag_YY Double No

Tag_ZZ Double No

Note: For each of the tags used in your Control Mode Analytic, ensure that you specify the
corresponding tag as Required in the input definition.

Table 2: OUTPUT DEFINITION: TAGS

Name Type Required

Control Mode Double Yes

Next Steps
Now that you have set up the control mode logic, you are ready to deploy the CLPM analytics.

Deploy CLPM Analytics

Deploy each of the CLPM analytics as described in this section.

Overview of CLPM Analytics

The following specialized analytics are provided with CLPM, and these are used for data collection and
transformation.

Table 3: CLPM Analytics

Analytic Name Use Location

OPM-CLPM-Control_Mode Operates on scheduled data and contains Will be in your Analytics Catalog if you
the control mode logic. have correctly performed the step to Set
Up Control Mode Logic.

OPM-CLPM-PV_Statistics Scheduled and produces the following Located in your Analytics Catalog. If you
KPIs based on the incoming data: do not find it there, contact your tenant
administrator.
• PV Error
• Lower SP Threshold
• Upper SP Threshold

Consult the KPI Reference for details on

the KPIs produced by this analytic.

OPM-CLPM-Performance Scheduled and produces a set of Located in your Analytics Catalog. If you
performance KPIs based on the incoming do not find it there, contact your tenant
data. administrator.

Consult the KPI Reference for details on

the KPIs produced by this analytic.

© 2020 General Electric Company 15

 Analytic Name Use Location

OPM-CLPM-Performance_Ext Scheduled and produces a set of Located in your Analytics Catalog. If you
performance KPIs based on the incoming do not find it there, contact your tenant
data. administrator.

Consult the KPI Reference for details on

the KPIs produced by this analytic.

OPM-CLPM-Config_Change Scheduled and produces the Total PIDF Located in your Analytics Catalog. If you
Changes KPI based on the incoming data. do not find it there, contact your tenant
administrator.
Consult the KPI Reference for details on
this KPI.

OPM-CLPM-Alert_Perf Scheduled and produces an alert for poor Located in your Analytics Catalog. If you
overall performance. do not find it there, contact your tenant
administrator.
The Overall Performance KPI is compared
to a threshold configured on the analytic.
For the last 3 hourly calculated samples, if
the Overall Performance KPI is greater
than the threshold, an alert is triggered.

Consult the KPI Reference for details on

the Overall Performance KPI.

OPM-CLPM-Alert_Limits Scheduled and produces an alert for Located in your Analytics Catalog. If you
limits being exceeded. do not find it there, contact your tenant
administrator.
The Percentage Limits Exceeded KPI is
compared to a threshold configured on
the analytic. For the last 3 hourly
calculated samples, if the Percentage
Control On KPI is 100% and the
Percentage Limits Exceeded KPI is greater
than the threshold, an alert is triggered.

Consult the KPI Reference for details on

the Percentage Limits Exceeded KPI and
Percentage Control On KPI.

16 © 2020 General Electric Company

 Analytic Name Use Location

OPM-CLPM-Alert_Manual Scheduled and produces an alert for Located in your Analytics Catalog. If you
control mode changing to Manual Mode. do not find it there, contact your tenant
administrator.
Control mode changes are examined over
a data window of one hour. If the last
control mode change detected is a
change from Auto, Cascade, or Shutdown
Mode to Manual Mode, an alert is
triggered.

For example, the following control modes

recorded over the window have the
following corresponding results. (Auto =
A, Cascade = C, Shutdown = S, Manual =
M)

• |SMASAM|: A to M is last change.

Alert!
• |AAAMMM|: A to M is last change.
Alert!
• |MMCMCM|: C to M is last change.
Alert!
• |AAAAMS|: M to S is last change. No
alert.
• |MMMMMS|: M to S is last change. No
alert.
• |SMMMMM|: S to M is last change.
Alert!

OPM-CLPM-Alert_PV_Quality Scheduled and produces an alert for poor Located in your Analytics Catalog. If you
PV sensor data, indicating that a possible do not find it there, contact your tenant
PV sensor health problem has been administrator.
detected.

This alert is triggered when the control

mode is not Shutdown and either of the
following is true:

• Percentage of PV sensor data of good

quality is lower than a threshold
configured on the analytic.
• The PV Variance KPI is zero (that is,
the PV is flatlining).

Consult the KPI Reference for details on

the PV Variance KPI.

Recommendation: If running this analytic

triggers an alert, check the PV sensor
data collection and sensor health.

© 2020 General Electric Company 17

 Analytic Name Use Location

OPM-CLPM-Alert_Tuning Scheduled and produces an alert Located in your Analytics Catalog. If you
indicating possible loop tuning or design do not find it there, contact your tenant
problems. administrator.

This alert is triggered when all of the

following are true:

• The Reversal Count KPI is greater

than a threshold configured on the
analytic.
• The Reversal Amplitude KPI is greater
than a threshold configured on the
analytic.
• The Percentage MV Saturation KPI is
less than a threshold configured on
the analytic.
• The Percentage Control On KPI is
100%.
• The Percentage Limits Exceeded KPI
is greater than a threshold configured
on the analytic.

Consult the KPI Reference for details on

each of the KPIs involved in triggering this
alert.

Recommendation: If running this analytic

triggers an alert, review the loop tuning
and design.

OPM-CLPM-Alert_MV_Quality Scheduled and produces an alert for poor Located in your Analytics Catalog. If you
MV sensor data, indicating that a possible do not find it there, contact your tenant
actuator health problem has been administrator.
detected.

This alert is triggered when the controller

is on and any of the following is true:

• Percentage of MV sensor data of

good quality is lower than a threshold
configured on the analytic.
• The Movement Index KPI is zero (that
is, the MV is flatlining).
• The Percentage MV Saturation KPI is
greater than a threshold configured
on the analytic.

Consult the KPI Reference for details on

each of the KPIs involved in triggering this
alert.

Recommendation: If running this analytic

triggers an alert, check the MV sensor
data collection and actuator health.

18 © 2020 General Electric Company

 Deploy analytics

Before You Begin

Make sure that you have set up control mode logic.

About This Task

This task describes how to deploy a single CLPM analytic.

Note: Using the steps described here, all CLPM analytics must be deployed, starting with the OPM-CLPM-
Control_Mode analytic.
Note: Variations in configuration between the different kinds of CLPM analytics are explained in the
tables that follow the procedure.

Procedure
1. Log into your tenant.
2. In the module navigation menu, select Analytics > Deployments.

3. Select the Add icon ( ) next to Deployments to add a deployment.

4. In the New Deployment window, select Analytic Template.
5. Enter a value in the Deployment Name box.
Make sure you enter a unique name. The system checks for duplicates.

© 2020 General Electric Company 19

 The Author box will be pre-populated with your login name and cannot be modified.
6. Enter a value in the Template Name box for the particular CLPM Analytic. All CLPM analytics start
with OPM-CLPM-...
This field is pre-populated to autocomplete with all matching options for the template name you
specify. You cannot modify any of these names; you must select one of them.
7. Step through the deployment wizard, in the same way you would for other kinds of analytics. Refer to
the APM Analytics documentation for details on how to map analytic inputs and outputs to tags. For
CLPM analytic configuration, it is important to adhere to the CLPM-specific guidance in the tables that
follow.
8. Select Deploy.

CLPM Analytic Configuration

These tables give a configuration reference for the different types of CLPM analytics.
• OPM-CLPM-Control_Mode
• KPI analytics (share common configuration)
◦ OPM-CLPM-Performance
◦ OPM-CLPM-Performance_Ext
◦ OPM-CLPM-PV_Statistics
• KPI analytic: OPM-CLPM-Config_Change
• Alert analytic: OPM-CLPM-Alert_Perf
• Alert analytic: OPM-CLPM-Alert_Limits
• Alert analytic: OPM-CLPM-Alert_Manual
• Alert analytic: OPM-CLPM-Alert_PV_Quality
• Alert analytic: OPM-CLPM-Alert_Tuning
• Alert analytic: OPM-CLPM-Alert_MV_Quality

20 © 2020 General Electric Company

OPM-CLPM-Control_Mode Analytic
Configure the OPM-CLPM-Control_Mode analytic as described in the table that follows.

Wizard Page Option Values Comments

1. Asset Selection: Select the Asset Filters All Assets Filter on All Assets. Use the
target control loop assets for search to select the target
the analytic deployment. control loop assets for the
analytic deployment.

Note: All analytics are

selected by default, so you will
need to deselect all those you
do not want and use the
search to find the ones you do
want.

2. I/O Mapping: Map inputs SELECTED STEP: OPM- Map for All Assets
and outputs in your analytic CLPM-Control_Mode > Tag
template definitions to tags on Mapping
the control loop asset.
INPUT DEFINITION: TAGS & It is mandatory to map the Map the analytic inputs to tags
ATTRIBUTES Tag_VV input. The Tag_XX, on the selected control loop
Tag_YY, and Tag_ZZ inputs are asset. These tags are used to
optional. determine control mode logic.

OUTPUT DEFINITION: TAGS It is mandatory to map the Map the analytic outputs to
& ATTRIBUTES Control Mode output. This tag tags on the selected control
specifies the control mode loop asset. Ensure that the
logic. analytic output name matches
the corresponding mapped tag
name exactly.

3. Schedule SCHEDULE Required value: Recurrent

SCHEDULE > Repeats Every Required value: 10 minutes

DATA REQUEST > Offset Required value: 10 minutes

Before Schedule

DATA REQUEST > Sample Required value: 10 minutes

Duration

DATA REQUEST > Sample Required value: Interpolated

Data

DATA REQUEST > Sampling Required value: 1 second

Interval

4. Review Review the deployment to

ensure that all configuration is
correct.

© 2020 General Electric Company 21

KPI Analytics
Configure the following KPI analytics as described in the table that follows.
• OPM-CLPM-Performance
• OPM-CLPM-Performance_Ext
• OPM-CLPM-PV_Statistics

Wizard Page Option Values Comments

1. Asset Selection: Select the Asset Filters All Assets Filter on All Assets. Use the
target control loop assets for search to select the target
the analytic deployment. control loop assets for the
analytic deployment.

Note: All analytics are

selected by default, so you will
need to deselect all those you
do not want and use the
search to find the ones you do
want.

2. I/O Mapping: Map inputs SELECTED STEP: Map for All Assets
and outputs in your analytic <analytic_name> > Tag
template definitions to tags on Mapping
the control loop asset.
INPUT DEFINITION: TAGS & Map the analytic inputs to tags
ATTRIBUTES on the selected control loop
asset. Ensure that each
analytic input name matches
the corresponding mapped tag
name exactly.

OUTPUT DEFINITION: TAGS Map the analytic outputs to

& ATTRIBUTES tags on the selected control
loop asset. Ensure that the
analytic output name matches
the corresponding mapped tag
name exactly.

22 © 2020 General Electric Company

 Wizard Page Option Values Comments

3. Schedule SCHEDULE Required value: Recurrent

SCHEDULE > Repeats Every Required value: 60 minutes

DATA REQUEST > Offset Required value: 60 minutes

Before Schedule

DATA REQUEST > Sample Required value: 60 minutes

Duration

DATA REQUEST > Sample Required value: Interpolated

Data

DATA REQUEST > Sampling Required value: 1 second

Interval

4. Review Review the deployment to

ensure that all configuration is
correct.

OPM-CLPM-Config_Change Analytic
Configure the OPM-CLPM-Config_Change analytic as described in the table that follows.

Wizard Page Option Values Comments

1. Asset Selection: Select the Asset Filters All Assets Filter on All Assets. Use the
target control loop assets for search to select the target
the analytic deployment. control loop assets for the
analytic deployment.

Note: All analytics are

selected by default, so you will
need to deselect all those you
do not want and use the
search to find the ones you do
want.

2. I/O Mapping: Map inputs SELECTED STEP: OPM- Map for All Assets
and outputs in your analytic CLPM-Config_Change > Tag
template definitions to tags on Mapping
the control loop asset.
INPUT DEFINITION: TAGS & Not all the P, I, D, and F inputs Map the analytic inputs to tags
ATTRIBUTES need to be mapped, as these on the selected control loop
inputs are optional. However, asset. Ensure that each
some will have to be mapped. analytic input name matches
the corresponding mapped tag
name exactly.

OUTPUT DEFINITION: TAGS Map the analytic outputs to

& ATTRIBUTES tags on the selected control
loop asset. Ensure that the
analytic output name matches
the corresponding mapped tag
name exactly.

© 2020 General Electric Company 23

 Wizard Page Option Values Comments

3. Schedule SCHEDULE Required value: Recurrent

SCHEDULE > Repeats Every Required value: 60 minutes

DATA REQUEST > Offset Required value: 60 minutes

Before Schedule

DATA REQUEST > Sample Required value: 60 minutes

Duration

DATA REQUEST > Sample Required value: Interpolated

Data

DATA REQUEST > Sampling Required value: 1 second

Interval

4. Review Review the deployment to

ensure that all configuration is
correct.

24 © 2020 General Electric Company

OPM-CLPM-Alert_Perf Analytic
Configure the OPM-CLPM-Alert_Perf analytic as described in the table that follows.

Wizard Page Option Values Comments

1. Asset Selection: Select the Asset Filters All Assets Filter on All Assets. Use the
target control loop assets for search to select the target
the analytic deployment. control loop assets for the
analytic deployment.

Note: All assets are selected

by default, so you will need to
deselect all those you do not
want and use the search to
find the ones you do want.

2. I/O Mapping: Map inputs SELECTED STEP: OPM- Map for All Assets
and outputs in your analytic CLPM-Alert_Perf > Tag
template definitions to tags on Mapping
the control loop asset.
INPUT DEFINITION: TAGS & Required input tag: Overall Map the analytic input to the
ATTRIBUTES Performance, Type: Double. tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly. The Overall
Performance tag is a KPI tag
produced by a previously run
CLPM analytic.

CONSTANTS: Name: Threshold, Type: Define a threshold constant

Double, Value: <a value for the analytic. The Overall
between 0 and 100>, Data Performance KPI is compared
Format: Constant to this threshold maximum
value. For the last 3 hourly
calculated samples, if the
Overall Performance KPI is
greater than the threshold, an
alert is triggered. Consult the
KPI Reference for details on
the Overall Performance KPI.

[Prepopulated Constant] Note: This constant is already

Name: alertTemplateName, defined on the analytic. Do not
Value: CLPM- modify this constant.
Loop_Performance, Type:
String, Data Format: Constant

OUTPUT DEFINITION: [Prepopulated Alert] Name: Note: This alert mapping is

ALERTS CLPM-Loop_Performance, already defined on the
Active: On analytic. Do not modify this
alert mapping.

© 2020 General Electric Company 25

 Wizard Page Option Values Comments

3. Schedule SCHEDULE Required value: Recurrent

SCHEDULE > Repeats Every Required value: 60 minutes

DATA REQUEST > Offset Required value: 60 minutes

Before Schedule

DATA REQUEST > Sample Required value: 180 minutes

Duration

DATA REQUEST > Sample Required value: Interpolated

Data

DATA REQUEST > Sampling Required value: 1 hour

Interval

4. Review Review the deployment to

ensure that all configuration is
correct.

26 © 2020 General Electric Company

OPM-CLPM-Alert_Limits Analytic
Configure the OPM-CLPM-Alert_Limits analytic as described in the table that follows.

© 2020 General Electric Company 27

 Wizard Page Option Values Comments

1. Asset Selection: Select the Asset Filters All Assets Filter on All Assets. Use the
target control loop assets for search to select the target
the analytic deployment. control loop assets for the
analytic deployment.

Note: All assets are selected

by default, so you will need to
deselect all those you do not
want and use the search to
find the ones you do want.

28 © 2020 General Electric Company

 Wizard Page Option Values Comments

2. I/O Mapping: Map inputs SELECTED STEP: OPM- Map for All Assets
and outputs in your analytic CLPM-Alert_Limits > Tag
template definitions to tags on Mapping
the control loop asset.
INPUT DEFINITION: TAGS & Required input tag: Percentage Map the analytic input to the
ATTRIBUTES Limits Exceeded, Type: Double. tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly. The Percentage
Limits Exceeded tag is a KPI
tag produced by a previously
run CLPM analytic.

Required input tag: Percentage Map the analytic input to the

Control On, Type: Double. tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly. The Percentage
Control On tag is a KPI tag
produced by a previously run
CLPM analytic.

CONSTANTS: Name: Threshold, Type: Define a threshold constant

Double, Value: <a value for the analytic. The
between 0 and 100>, Data Percentage Limits Exceeded
Format: Constant KPI is compared to this
threshold maximum value. For
the last 3 hourly calculated
samples, if the Percentage
Control On KPI is 100% and
the Percentage Limits
Exceeded KPI is greater than
this threshold, an alert is
triggered. Consult the KPI
Reference for details on the
Percentage Limits Exceeded
KPI.

[Prepopulated Constant] Note: This constant is already

Name: alertTemplateName, defined on the analytic. Do not
Value: CLPM-Limits_Exceeded, modify this constant.
Type: String, Data Format:
Constant

OUTPUT DEFINITION: [Prepopulated Alert] Name: Note: This alert mapping is

ALERTS CLPM-Limits_Exceeded, already defined on the
Active: On analytic. Do not modify this
alert mapping.

© 2020 General Electric Company 29

 Wizard Page Option Values Comments

3. Schedule SCHEDULE Required value: Recurrent

SCHEDULE > Repeats Every Required value: 60 minutes

DATA REQUEST > Offset Required value: 60 minutes

Before Schedule

DATA REQUEST > Sample Required value: 180 minutes

Duration

DATA REQUEST > Sample Required value: Interpolated

Data

DATA REQUEST > Sampling Required value: 1 hour

Interval

4. Review Review the deployment to

ensure that all configuration is
correct.

30 © 2020 General Electric Company

OPM-CLPM-Alert_Manual Analytic
Configure the OPM-CLPM-Alert_Manual analytic as described in the table that follows.

Wizard Page Option Values Comments

1. Asset Selection: Select the Asset Filters All Assets Filter on All Assets. Use the
target control loop assets for search to select the target
the analytic deployment. control loop assets for the
analytic deployment.

Note: All assets are selected

by default, so you will need to
deselect all those you do not
want and use the search to
find the ones you do want.

2. I/O Mapping: Map inputs SELECTED STEP: OPM- Map for All Assets
and outputs in your analytic CLPM-Alert_Manual > Tag
template definitions to tags on Mapping
the control loop asset.
INPUT DEFINITION: TAGS & Required input tag: Control Map the analytic input to the
ATTRIBUTES Mode, Type: Double. tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly.

CONSTANTS: [Prepopulated Constant] Note: This constant is already

Name: alertTemplateName, defined on the analytic. Do not
Value: CLPM-Manual_Control, modify this constant.
Type: String, Data Format:
Constant

OUTPUT DEFINITION: [Prepopulated Alert] Name: Note: This alert mapping is

ALERTS CLPM-Manual_Control, Active: already defined on the
On analytic. Do not modify this
alert mapping.

© 2020 General Electric Company 31

 Wizard Page Option Values Comments

3. Schedule SCHEDULE Required value: Recurrent

SCHEDULE > Repeats Every Required value: 60 minutes

DATA REQUEST > Offset Required value: 60 minutes

Before Schedule

DATA REQUEST > Sample Any window size acceptable.

Duration Recommended value: 60
minutes

DATA REQUEST > Sample Required value: Interpolated

Data

DATA REQUEST > Sampling Recommended value: 1 second

Interval or the rate at which control
mode values are written to
Predix Timeseries

4. Review Review the deployment to

ensure that all configuration is
correct.

32 © 2020 General Electric Company

OPM-CLPM-Alert_PV_Quality Analytic
Configure the OPM-CLPM-Alert_PV_Quality analytic as described in the table that follows.

Wizard Page Option Values Comments

1. Asset Selection: Select the Asset Filters All Assets Filter on All Assets. Use the
target control loop assets for search to select the target
the analytic deployment. control loop assets for the
analytic deployment.

Note: All assets are selected

by default, so you will need to
deselect all those you do not
want and use the search to
find the ones you do want.

2. I/O Mapping: Map inputs SELECTED STEP: OPM- Map for All Assets
and outputs in your analytic CLPM-Alert_PV_Quality >
template definitions to tags on Tag Mapping
the control loop asset.
INPUT DEFINITION: TAGS & Required input tag: PV, Type: Map the analytic input to the
ATTRIBUTES Double tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly. This tag is
needed to calculate
percentage good quality. For
the calculated samples over
the last data window, the
percentage of PV values that
are of good quality is
compared to the
PV_Quality_Threshold constant
you will configure on this page.

Required input tag: PV Map the analytic input to the

Variance, Type: Double tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly. The PV Variance
tag is a KPI tag produced by a
previously run CLPM analytic.
If the PV Variance KPI is zero
(i.e., the PV is flatlining) for the
calculated samples over the
last data window, and the
control mode is not Shutdown,
an alert is triggered.

© 2020 General Electric Company 33

 Wizard Page Option Values Comments

CONSTANTS: Required input tag: Control Map the analytic input to the
Mode, Type: Double tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly.

Name: PV_Quality_Threshold, Define a threshold percentage

Type: Double, Value: <a value constant for the analytic. For
between 0 and 100>, Data the calculated samples over
Format: Constant the last data window, the
percentage of PV values that
are of good quality is
compared to this threshold
percentage. If that quality is
lower than the threshold, and
the control mode is not
Shutdown, an alert is
triggered.

[Prepopulated Constant] Note: This constant is already

Name: alertTemplateName, defined on the analytic. Do not
Value: CLPM-Sensor_Health, modify this constant.
Type: String, Data Format:
Constant

OUTPUT DEFINITION: [Prepopulated Alert] Name: Note: This alert mapping is

ALERTS CLPM-Sensor_Health, Active: already defined on the
On analytic. Do not modify this
alert mapping.

3. Schedule SCHEDULE Required value: Recurrent

SCHEDULE > Repeats Every Required value: 60 minutes

DATA REQUEST > Offset Required value: 60 minutes

Before Schedule

DATA REQUEST > Sample Minimum value: 60 minutes,

Duration Recommended value: 60
minutes

DATA REQUEST > Sample Required value: Interpolated

Data

DATA REQUEST > Sampling Recommended value: 1 second

Interval or the rate at which control
mode and PV values are
written to Predix Timeseries

4. Review Review the deployment to

ensure that all configuration is
correct.

34 © 2020 General Electric Company

OPM-CLPM-Alert_Tuning Analytic
Configure the OPM-CLPM-Alert_Tuning analytic as described in the table that follows.

© 2020 General Electric Company 35

 Wizard Page Option Values Comments

1. Asset Selection: Select the Asset Filters All Assets Filter on All Assets. Use the
target control loop assets for search to select the target
the analytic deployment. control loop assets for the
analytic deployment.

Note: All assets are selected

by default, so you will need to
deselect all those you do not
want and use the search to
find the ones you do want.

36 © 2020 General Electric Company

 Wizard Page Option Values Comments

2. I/O Mapping: Map inputs SELECTED STEP: OPM- Map for All Assets
and outputs in your analytic CLPM-Alert_Tuning > Tag
template definitions to tags on Mapping
the control loop asset.
INPUT DEFINITION: TAGS & Required input tag: Reversal Map the analytic input to the
ATTRIBUTES Count, Type: Double. tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly. The Reversal
Count tag is a KPI tag
produced by a previously run
CLPM analytic. This KPI is
compared to the
Reversal_Count_Threshold
constant you will configure on
this page.

Required input tag: Reversal Map the analytic input to the

Amplitude, Type: Double. tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly. The Reversal
Amplitude tag is a KPI tag
produced by a previously run
CLPM analytic. This KPI is
compared to the
Reversal_Amplitude_Threshold
constant you will configure on
this page.

Required input tag: Percentage Map the analytic input to the

MV Saturation, Type: Double. tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly. The Percentage
MV Saturation tag is a KPI tag
produced by a previously run
CLPM analytic. This KPI is
compared to the
Percentage_MV_Saturation_Th
reshold constant you will
configure on this page.

© 2020 General Electric Company 37

 Wizard Page Option Values Comments

Required input tag: Percentage Map the analytic input to the

Control On, Type: Double, tag on the selected control
Value: <a value between 0 and loop asset. Ensure that the
100> analytic input name matches
the corresponding mapped tag
name exactly. The Percentage
Control On tag is a KPI tag
produced by a previously run
CLPM analytic. If the
Percentage Control On KPI is
100%, this satisfies one of the
requirements for an alert to be
triggered.

Required input tag: Percentage Map the analytic input to the

Limits Exceeded, Type: Double. tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly. The Percentage
Limits Exceeded tag is a KPI
tag produced by a previously
run CLPM analytic. This KPI is
compared to the
Percentage_Limits_Exceeded_
Threshold constant you will
configure on this page.

38 © 2020 General Electric Company

 Wizard Page Option Values Comments

CONSTANTS: Name: Define a reversal count

Reversal_Count_Threshold, threshold constant for the
Type: Double, Value: <any analytic. For the calculated
whole number>, Data Format: samples over the last data
Constant window, the Reversal Count
KPI is compared to this
threshold. If the KPI is greater
than the threshold, this
satisfies one of the
requirements for an alert to be
triggered.

Name: Define a reversal amplitude

Reversal_Amplitude_Threshold threshold constant for the
, Type: Double, Value: <any analytic. For the calculated
whole number>, Data Format: samples over the last data
Constant window, the Reversal
Amplitude KPI is compared to
this threshold. If the KPI is
greater than the threshold,
this satisfies one of the
requirements for an alert to be
triggered.

Name: Define a percentage MV

Percentage_MV_Saturation_Th saturation threshold constant
reshold, Type: Double, Value: for the analytic. For the
<any whole number>, Data calculated samples over the
Format: Constant last data window, the
Percentage MV Saturation KPI
is compared to this threshold.
If the KPI is less than the
threshold, this satisfies one of
the requirements for an alert
to be triggered.

Name: Define a percentage limits

Percentage_Limits_Exceeded_ exceeded threshold constant
Threshold, Type: Double, Value: for the analytic. For the
<any whole number>, Data calculated samples over the
Format: Constant last data window, the
Percentage Limits Exceeded
KPI is compared to this
threshold. If the KPI is greater
than the threshold, this
satisfies one of the
requirements for an alert to be
triggered.

[Prepopulated Constant] Note: This constant is already

Name: alertTemplateName, defined on the analytic. Do not
Value: CLPM- modify this constant.
Loop_TuningOrDesign, Type:
String, Data Format: Constant

© 2020 General Electric Company 39

 Wizard Page Option Values Comments

OUTPUT DEFINITION: [Prepopulated Alert] Name: Note: This alert mapping is

ALERTS CLPM-Loop_TuningOrDesign, already defined on the
Active: On analytic. Do not modify this
alert mapping.

3. Schedule SCHEDULE Required value: Recurrent

SCHEDULE > Repeats Every Required value: 60 minutes

DATA REQUEST > Offset Required value: 60 minutes

Before Schedule

DATA REQUEST > Sample Minimum value: 60 minutes,

Duration Recommended value: 60
minutes

DATA REQUEST > Sample Required value: Interpolated

Data

DATA REQUEST > Sampling Required value: 1 hour

Interval

4. Review Review the deployment to

ensure that all configuration is
correct.

40 © 2020 General Electric Company

OPM-CLPM-Alert_MV_Quality Analytic
Configure the OPM-CLPM-Alert_MV_Quality analytic as described in the table that follows.

© 2020 General Electric Company 41

 Wizard Page Option Values Comments

1. Asset Selection: Select the Asset Filters All Assets Filter on All Assets. Use the
target control loop assets for search to select the target
the analytic deployment. control loop assets for the
analytic deployment.

Note: All assets are selected

by default, so you will need to
deselect all those you do not
want and use the search to
find the ones you do want.

42 © 2020 General Electric Company

 Wizard Page Option Values Comments

2. I/O Mapping: Map inputs SELECTED STEP: OPM- Map for All Assets
and outputs in your analytic CLPM-Alert_MV_Quality >
template definitions to tags on Tag Mapping
the control loop asset.
INPUT DEFINITION: TAGS & Required input tag: MV, Type: Map the analytic input to the
ATTRIBUTES Double tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly. This tag is
needed to calculate
percentage good quality. For
the calculated samples over
the last data window, the
percentage of MV values that
are of good quality is
compared to the
MV_Quality_Threshold
constant you will configure on
this page.

Required input tag: Movement Map the analytic input to the

Index, Type: Double tag on the selected control
loop asset. Ensure that the
analytic input name matches
the corresponding mapped tag
name exactly. The Movement
Index tag is a KPI tag produced
by a previously run CLPM
analytic. If the Movement
Index KPI is zero (i.e., the MV is
flatlining) for the calculated
samples over the last data
window, this satisfies one of
the requirements for an alert
to be triggered.

Required input tag: Percentage Map the analytic input to the

MV Saturation, Type: Double, tag on the selected control
Value: <a value between 0 and loop asset. Ensure that the
100> analytic input name matches
the corresponding mapped tag
name exactly. The Percentage
MV Saturation is a KPI tag
produced by a previously run
CLPM analytic. For the
calculated samples over the
last data window, the
Percentage MV Saturation KPI
is compared to the
Percentage_MV_Saturation_Th
reshold constant you will
configure on this page.

© 2020 General Electric Company 43

 Wizard Page Option Values Comments

Required input tag: Percentage Map the analytic input to the

Control On, Type: Double, tag on the selected control
Value: <a value between 0 and loop asset. Ensure that the
100> analytic input name matches
the corresponding mapped tag
name exactly. The Percentage
Control On tag is a KPI tag
produced by a previously run
CLPM analytic. If the
Percentage Control On KPI is
100%, this satisfies one of the
requirements for an alert to be
triggered.

CONSTANTS: Name: MV_Quality_Threshold, Define a threshold percentage

Type: Double, Value: <a value constant for the analytic. For
between 0 and 100>, Data the calculated samples over
Format: Constant the last data window, the
percentage of MV values that
are of good quality is
compared to this threshold
percentage. If that quality is
lower than the threshold, this
satisfies one of the
requirements for an alert to be
triggered.

Name: Define a percentage MV

Percentage_MV_Saturation_Th saturation threshold constant
reshold, Type: Double, Value: for the analytic. For the
<a value between 0 and 100>, calculated samples over the
Data Format: Constant last data window, the
Percentage MV Saturation KPI
is compared to this threshold.
If the KPI is less than the
threshold, this satisfies one of
the requirements for an alert
to be triggered.

[Prepopulated Constant] Note: This constant is already

Name: alertTemplateName, defined on the analytic. Do not
Value: CLPM-Actuator_Health, modify this constant.
Type: String, Data Format:
Constant

OUTPUT DEFINITION: [Prepopulated Alert] Name: Note: This alert mapping is

ALERTS CLPM-Actuator_Health, Active: already defined on the
On analytic. Do not modify this
alert mapping.

44 © 2020 General Electric Company

 Wizard Page Option Values Comments

3. Schedule SCHEDULE Required value: Recurrent

SCHEDULE > Repeats Every Required value: 60 minutes

DATA REQUEST > Offset Required value: 60 minutes

Before Schedule

DATA REQUEST > Sample Minimum value: 60 minutes,

Duration Recommended value: 60
minutes

DATA REQUEST > Sample Required value: Interpolated

Data

DATA REQUEST > Sampling Recommended value: 1 second

Interval or the rate at which MV values
are written to Predix
Timeseries

4. Review Review the deployment to

ensure that all configuration is
correct.

Access Loop and Fleet Dashboards

About This Task
Loop and Fleet dashboards are deployed automatically when you deploy CLPM as part of tenant set up.
No additional steps are required. You can access the Loop and Fleet dashboards.

© 2020 General Electric Company 45

 Chapter

3
The Fleet Report
Topics:

• The Fleet Report

• Access the Fleet Report
• Modify the Date Range for a
Fleet Report
• Average Control Loop
Performance Chart
• The Diagnostic Alerts Filtering
Smart Filter
• The Control Loop Performance
Table

46 © 2020 General Electric Company

The Fleet Report
The Fleet Report provides a summary of all loops in the selected asset context for the selected time
period. The report contains the following information:
• Average Control Loop Performance chart
• The Diagnostic Alerts Filtering smart filter
• Control Loop Performance table
Note:
• CLPM analytics process raw tag data and produce results that are interpolated. These interpolated
results are then used to produce the Fleet Report chart and table.
• The measures presented in the Fleet Report are derived from KPIs that are produced by CLPM
analytics. Consult the KPI Reference for details on these KPIs.

Access the Fleet Report

Procedure
In the module navigation menu, select Intelligence and then go to Control Loop Performance.
The Fleet Report appears, displaying the Average Control Loop Performance graph, diagnostics alerts
filtering, and the Control Loop Performance table on selecting the required asset context, in the Select
Context window. By default, the data reported in the Fleet Report is for seven days prior to the current
date.
Note: Control Loop Performance reports are also accessible from the Dashboard menu.
Tip: If needed, you can perform the following steps:
• Modify the asset context by selecting SELECT ASSET CONTEXT, navigating to the asset, and then
selecting OPEN.
• Modify the date range for which the data is reported.

Modify the Date Range for a Fleet Report

Procedure
1. Access the Fleet Report.
2. In the Fleet Report workspace, specify the new date range by selecting the start date and end date.
The report is updated with the new data.
Note: The date and time range for the reported data also appears in the report.
Important: The Fleet Report supports only a date range of up to 30 days.
Note: When you navigate from a Fleet Report to a Loop Report, the Loop Report is plotted with the
end date selected in the Fleet Report. For example, if the date range selected in the Fleet Report is
12th November, 2018 00:00:00 to 18th November, 2018 23:59:59 (the last data point is from 19th
November, 2018 00:00:00 to include any last micro second data point), after you navigate to the Loop
Report, the Loop Report is plotted with data reported for 18th November, 2018 (that is, 18th
November, 2018 00:00:00 to 19th November, 2018 00:00:00).

© 2020 General Electric Company 47

Average Control Loop Performance Chart
About the Average Control Loop Performance Chart
The Average Control Loop Performance chart contains a stacked-bar graph and a trend line. Each item in a
stack represents the percentage of time the control element operated at each of the following control
modes during the reporting period:
• Manual
• Auto
• Cascade
• Shutdown
The trend line represents the percentage of time the control element operated out of the specified limits
during the reporting period.

The chart contains the following axes:

• x-axis: Represents the date for which the data is plotted.
• y-axis: Represents the percentage of time the control element operated at each control mode. It is
calculated based on hourly samples for each date.

The Diagnostic Alerts Filtering Smart Filter

The Diagnostic Alerts Filtering smart filter is a visual tool to identify faulty control loops. Using this filter,
you can perform the following tasks:
• Identify control loops based on diagnostic alerts.
• Classify the loops in three fault categories, namely Sensor, Actuator, and Tuning.
• Drill down into the status of the alerts to find the list of control loops where action is needed.
• Identify bad actors (poorly performing loops) among a large set of control loop assets.
Note: The filter is limited to the control loop diagnostic alert generated based on pre-defined probabilistic
analytics.

48 © 2020 General Electric Company

 The filter displays the number of alerts for each fault mode and the total number of alerts using
interactive pie charts.

The pie charts illustrate the different fault modes of your control loops and the corresponding number of
alerts. Each pie chart is further classified based on the alert status. The pie charts display the total the
number of open and closed alerts. Each pie chart has a list displaying the top three control loops
contributing the most number of alerts, in descending order.
On selecting the Open area of the pie chart, the details about the loops with the open alerts are updated
in the list below the pie charts. Also, the Control Loop Performance table is updated to list only the
control loops with open alerts.
Tip: Select Reset to reset the Diagnostic Alerts Filtering smart filter and the Control Loop
Performance table.

The Control Loop Performance Table

Note: By default, the Control Loop Performance table is not populated with details about the control
loops. Select Show all to access the list of all the control loops in the table.

In the Fleet Report, you can modify the reporting period by selecting a date prior to the current date in the
Date Selector box. You can sort data based on any column in this table. By default, the table is sorted
based on the value in the Performance Index AVG (100-0) column.
You can filter data in the table by:
• Applying a global filter on the table
• Using multiple-level filters
• Using text in the filter field
• Using conditional parameters
You can search for a specific control loop using different search options.
For each loop in the selected asset context, the table contains the following information.

© 2020 General Electric Company 49

 Loop Statistic Interpretation

Loop The name of the control loop asset. If you select the link in this
column, the Loop Report appears.

Limits Exceeded AVG (%) The percentage of time for which the loop was in Auto, Cascade,
or Manual mode, where the limits for these modes were
exceeded. The time when the loop was in shutdown mode is not
included in this statistic.

Performance Index AVG (100-0) The average of the performance index during the specified
reporting period. The value ranges from 100 to 0, where 100
indicates poor performance and 0 indicates good performance.
By default, the table is sorted based on the value in this column.

The performance index is an aggregate of the following values in

the following proportions:

• One-third of the performance index value is composed of

the percentage of time for which the loop limits were
exceeded. The time when the loop was in the shutdown
mode is not included in this statistic.
• One-third of the performance index value is composed of
the percentage of time for which the loop was in the
manual control mode.
• One-third of the performance index value is composed of
the percentage of time for which the controller output was
saturated. The controller output is considered saturated
when the MV has a value of 0 or 100, which are the lower
and upper limits for controller output saturation,
respectively.

A higher average performance index indicates that the loop is

not performing well.

MV Movement The average number of oscillations of the manipulated variable.

You can use this statistic to investigate the ability of the control
element to maintain the PV operating within process limits.

MV Saturation AVG (%) The percentage of time that the MV for each loop was
saturated.

Manual AVG (%) The percentage of time that the loop was in the manual control
mode.

Auto AVG (%) The percentage of time that the loop was in the auto control
mode.

Cascade AVG (%) The percentage of time that the loop was in the cascade control
mode.

Shutdown AVG (%) The percentage of time that the loop was in the shutdown
control mode.

Good Quality Data (%) The percentage of time that the data was of good quality.

50 © 2020 General Electric Company

 Loop Statistic Interpretation

PV Variability AVG The average PV variability. This value is a measure of the data
spread in the PV data set, expressed as a percentage.

This is the average value of the PV Variability KPI. Refer to the

KPI Reference for the definition of this KPI.

Total Configuration Changes The number of times the controller configuration was changed
during the reporting period.

Absolute Error (Aggregate) AVG The average of the absolute values of all errors during the
reporting period. Error is calculated as the difference between
PV and SP.

Tip: Sorting the table based on this field is useful only if all loops
use the same unit of measure.

© 2020 General Electric Company 51

 Chapter

4
The Loop Report
Topics:

• About the Loop Report

• Access the Loop Report
• Modify the Date for a Loop
Report
• The Process Variable
Performance Chart
• The Manipulated Variable
Distribution Chart
• The Control Overview Chart
• The Error Distribution Chart
• The Control Mode Summary
Table
• The Controller Performance
Table
• The PV Performance Table
• The Error Statistics Table
• The Controller Configuration
Table

52 © 2020 General Electric Company

About the Loop Report
The Loop Report provides a systematic approach to optimizing the performance of a control loop by
monitoring the loop performance and providing diagnostic data about the loop. The report contains
various charts and tables that help in assessing the performance of a control loop. The data used for
assessing the performance of the loop is derived from the tags associated with the control loop assets.
Note: CLPM analytics process raw tag data and produce results that are interpolated. These interpolated
results are then used to produce the Loop Report charts and tables.
Note: The measures presented in the Loop Report are derived from KPIs that are produced by CLPM
analytics. Consult the KPI Reference for details on these KPIs.
Note: Unless otherwise stated, only data of good quality is considered when calculating the measures
represented in these tables and charts. Bad quality data is ignored.

The Loop Report: Charts

The Loop Report contains the following charts:
• Process Variable Performance
• Manipulated Variable Distribution (%)
• Control Overview
• Error Distribution

The Loop Report: Tables

The Loop Report contains the following tables:
• Control Mode Summary
• Controller Performance
• PV Performance
• Error Statistics
• Controller Configuration

Access the Loop Report

About This Task
This topic describes how to access the Loop Report from the module navigation menu.

Procedure
1. In the OPM navigation menu, go to Intelligence > Control Loop Performance.
The fleet dashboard appears, displaying the Average Control Loop Performance graph, diagnostic
alerts filtering, and the Control Loop Performance Table on selecting the required asset in the Select
Context window.
2. In the Control Loop Performance Table of an associated fleet report, select the link in the loop
column.
The Loop Report appears, displaying the preconfigured widgets for graphs and tables. The report
contains data for the selected asset context from the Fleet Report.

© 2020 General Electric Company 53

 Note: If there is no context selected, select the required asset by selecting the Select Context tab in
the header.
Note: Control Loop Performance reports are also accessible from the Dashboard menu.

Modify the Date for a Loop Report

About This Task

This topic describes how to modify the date for a Loop Report.

Procedure
Note: By default, the Loop Report is plotted with data reported for the current date. For example, when
you directly access the loop report on 12th November, 2018 20:00:00, the dashboard is plotted with data
reported from 12th November, 2018 00:00:00 to 12th November, 2018 20:00:00 (till the latest available
data point). The loop report can be displayed for a time period of one day only.
1. Access the Loop Report whose date you want to modify.
The Date Selector box appears, displaying the default date. You can access data for a single day. Data
is reported till the current time of the day. For example, if the current date and time is 20th November,
2018 20:00:00, and you select 20th November, 2018, then data is reported from 20th November, 2018
00:00:00 to 20th November, 2018 20:00:00. However, if the current date and time is 20th November,
2018 20:00:00, and you select 19th November, 2018, then data is reported from 19th November, 2018
00:00:00 to 20th November, 2018 00:00:00.
2. Select the Date Selector box.
The date selector window appears.
3. Select the date you want to set, and then select Apply.
The date is modified, and the report is updated. The date for which the data is reported appears in the
upper-right corner of the report.
Tip: You can reset the date to the default value by selecting reset.

The Process Variable Performance Chart

The process variable performance (PVP) chart helps in analyzing the trend of the process variable (PV)
against the setpoint (SP).
The chart contains the following axes:
• x-axis: Represents the time period selected in the control loop report.
• y-axis: Represents the PV and SP values.
The trend of each value over the selected time period appears as a step chart, which helps you
understand the magnitude of change at a specific timestamp. The following trend lines are plotted on the
chart:

54 © 2020 General Electric Company

 Trend Line Description

PV A step chart representing the trend of PV over the time period

selected in the report. Consult the CLPM Terminology table for
the definition of the PV.

SP A step chart representing the trend of SP over the time period

selected in the report. Consult the CLPM Terminology table for
the definition of the SP.

Upper Limit A step chart representing the trend of the acceptable upper
limit for the PV value. The Upper Limit on this chart represents
the Upper SP Threshold KPI. Consult the KPI Reference for the
definition of the Upper SP Threshold.

Lower Limit A step chart representing the trend of the acceptable lower
limit for the PV value. The Lower Limit on this chart represents
the Lower SP Threshold KPI. Consult the KPI Reference for the
definition of the Lower SP Threshold.

Interpreting the PVP Chart

Based on the values plotted on the PVP chart, you can identify performance issues in the process control
loop. The following table lists a few scenarios that may appear in the chart and what each scenario
indicates about the loop performance:
Note: If data for a tag is missing, a message appears at the top of the page, specifying the tag for which
the data is missing.

What You See Interpretation Suggested Action

The trend of the PV values is centered Process is under control. None. Maintain the parameter settings.
around the trend of the SP values.

The trend of the PV values exceeds the Process is not under control. Tune the control loop to reduce process
trend of the upper or lower limit. variation.

Gaps in the trend line. The plotted data contains a mix of good Investigate why the quality of some data
and bad quality data. is bad.

A flat trend line appears. • There has been no change in data for • Investigate why there is no change in
the time period selected. The last data.
recorded value is maintained. • Fix the connection to the real-time
-or- server.
• The connection to a real-time server
has been lost.

An empty chart. The quality of all the data for the selected Investigate why the quality of data is bad.
time period is bad.

The Manipulated Variable Distribution Chart

The manipulated variable distribution (%) chart is a histogram that helps you analyze the efficiency of a
control element and evaluate the suitability of the element for the process.
A control element receives the output signal from the controller and controls the operating conditions
such as flow, pressure, temperature, and liquid level in response to the signal. The manipulated variable

© 2020 General Electric Company 55

 (MV) is a percentage measure of the value at which the control element is functioning. For example, in the
temperature control loop in an air conditioner, based on how much cooling is required (that is, setpoint),
the thermostat (that is, the controller) sends a signal to manipulate the flow of cooling agent such that
the room temperature (that is, the process variable) is adjusted to match the desired temperature. In this
example, MV is the percentage of flow of cooling agent allowed through the valve.
The chart contains the following axes:
• x-axis: Represents the value of the MV measured in percentage
• y-axis: Represents the count of samples of the MV
The histogram represents the count of samples that the control element was set to at each percentage
value, and not the length of time the element remained at each value. Therefore, you must always
consider the distribution pattern of MV when interpreting the statistical significance of the samples
displayed in this histogram.

Interpreting the Manipulated Variable Distribution Chart

Based on the values plotted on the chart, you can identify performance issues in the controller. The
following table provides a few scenarios that may appear in the chart and what each scenario indicates
regarding the controller performance.
Note: If data for a tag is missing, a message appears at the top of the page, specifying the tag for which
the data is missing.

What you see Interpretation Required action

Red bars appear. The samples can be interpreted as Install a different type of controller, either
follows: a larger one that is more suitable for the
purpose, or a smaller one that is more
• < 6%: Indicates that the control
energy-efficient and cost-effective.
element is operating at a very low
saturation level and that the element
is oversized.
• > 95%: Indicates that the control
element is operating at a very high
saturation level and that the element
is undersized.

A controller operating at the extreme high

or low values of its capacity will result in
asset fatigue over time.

Yellow bars appear. The samples can be interpreted as Monitor the control element for signs of
follows: asset fatigue.

• 6% - 30%: Indicates that the control

element is operating close to its
lower saturation level.
• 71% - 95%: Indicates that the control
element is operating close to its
higher saturation level.

A controller operating at high or low

values of its capacity will result in asset
fatigue over time.

56 © 2020 General Electric Company

 What you see Interpretation Required action

Blue bars appear. The controller is operating at the desired None. This is the ideal value.
capacity, that is, in the range of 31%
through 70%. More samples in this range
indicate that the controller is well-
designed.

Image displays a message stating that Data is not available for the selected time Investigate if there is a loss of connection
there is no data. period. with the database.

An empty histogram appears. The quality of all the data for the selected Investigate why the quality of data is bad.
time period is bad.

One trend interval (bin) appears with This indicates that there has been no data Verify and fix the connection to the real-
majority of data counts. change, and that the last recorded value time server.
has been maintained.

This can possibly occur as a result of loss

of connection to a real-time server.

Flat histogram, with no discernible peaks, The control element is operating over its Optimize the size of the control element.
appears. whole range to control the PV. This
typically indicates that the control
element is too small for the system and
will cause asset fatigue over time.

The Control Overview Chart

The control overview chart displays the trend of the manipulated variable (MV). The chart contains the
following axes:
• x-axis: Represents the time period selected in the report.
• y-axes (two in number): Represent the MV value and the control mode in which the control loop was
operating over the time period. The following table provides the number that represents each control
mode:

Number Control Mode

1 Manual

2 Auto

3 Cascade

4 Shutdown

Trend lines appear in the following colors in the chart:

• Blue: Indicates the control mode in which the system is operating. The control mode can change over
the recording period.
• Red: Indicates the trend of the MV value over the total calibrated range.

Interpreting the Control Overview Chart

Based on the values plotted on the chart, you can identify performance issues in the controller. The
following table provides a few scenarios that may appear in the chart and what each scenario indicates
regarding the controller performance.

© 2020 General Electric Company 57

 Note: If data for a tag is missing, a message appears at the top of the page, specifying the tag for which
the data is missing.

What you see Interpretation Required action

The blue trend line is set to Shutdown. Indicates that the control element is in Ensure that the system is operating in the
the Shutdown mode. required control mode.

Investigate causes of a change to

Shutdown mode.

The red trend line indicates the MV • < 5%: Indicates that the control Optimize the size of the controller.
percentage functioning value over the element is operating at a very low
total calibrated range. saturation level and that the element
is oversized.
• 5 - 30%: Indicates that the control
element is operating close to its
lower saturation level.
• 30 - 70%: Indicates that the control
element is operating at its optimal
value. This is the ideal value for the
variable.
• 70 - 95%: Indicates that the control
element is operating close to its
higher saturation level.
• > 95%: Indicates that the control
element is operating at a very high
saturation level and that the element
is undersized.

The blue trend line is set to Bad. The data was of bad quality and loop Identify the reason for bad quality data.
statistics could not be calculated.
Check the connection to the server or
database.

A report with no value for MV in the y-axis Either the MV has no data or the data is of Investigate which of these is true and
appears. bad quality. why.

Gaps in the trend line appear. Data was a mixture of good and bad Investigate why the quality of some data
quality. is bad.

A flat trend line appears. • There has been no change in data for • Investigate why there is no change in
the time period selected. The last data.
recorded value is maintained. • Fix the connection to the real-time
-or- server.
• The connection to a real-time server
has been lost.

The Control Mode and MV values are The control element has an appropriate None. The controller is well-tuned.
responding quickly relative to the PV and reaction to changes in the PV.
SP values in the PV trend.

The Control Mode and MV values are The control element has a sluggish Tune the control loop PIDF parameters.
responding very slowly relative to the PV reaction to changes in the PV.
and SP values in the PV trend.

58 © 2020 General Electric Company

The Error Distribution Chart
The error distribution chart displays a histogram that represents the distribution of the amount of time
the controller operated at each error value. Controller error is calculated as the difference between the
values, PV and SP. The chart contains the following axes:
• x-axis: Represents the controller error.
• y-axis: Represents the time for which the controller was operating at the PV error value.

Interpreting the Error Distribution Chart

Based on the values plotted on the chart, you can identify performance issues in the controller. The
following table provides a few scenarios that may appear in the chart and what each scenario indicates
regarding the controller performance.
Note: If data for a tag is missing, a message appears at the top of the page, specifying the tag for which
the data is missing.

What you see Interpretation Required action

The histogram bars appear in red. The error is out of acceptable limits. Control the PV so that it functions within
the defined process limits.

The histogram bars appear in blue. The error is in acceptable limits. None.

Bimodal or oscillating distribution of Possible friction exists in control element. Check the functioning of the control
histogram bars appears. element.

Histogram bars appear primarily in the Possible errors in loop tuning exist, or the Check the functioning of the control
higher error range. control element is not working effectively. element. Tune the PIDF controller
parameters.

© 2020 General Electric Company 59

 What you see Interpretation Required action

One trend interval (bin) appears with There has been no change in PV error for Investigate why there is no change in
majority of data counts. the time period selected. The last data. Fix the connection to the real-time
recorded value is maintained. Or, the server.
connection to a real-time server has been
lost.

An empty histogram appears. The quality of all the data for the selected Investigate why the quality of data is bad.
time period is bad.

The Control Mode Summary Table

The control mode summary table helps you assess the performance of the control loop based on the
percentage of time the loop operated in each control mode. The table contains a row for each of the
following control modes:
• Manual
• Auto
• Cascade
• Shutdown
The table contains the following information for each control mode:

Column name Description

Duration (%) Indicates the percentage of time the control loop operated in
each controller mode. This is the average value of the
Percentage <control mode> KPI.

Lower Limit Exceeded (%) Indicates the percentage of time the loop operated below the
lower limit. This is the average value of the Percentage <control
mode> LL Exceeded KPI.

Upper Limit Exceeded (%) Indicates the percentage of time the loop operated above the
upper limit. This is the average value of the Percentage <control
mode> UL Exceeded KPI.

Total Limits Exceeded (%) Indicates the total percentage of time that the loop operated
out of limit. This value is calculated as the sum of the Lower
Limit Exceeded (%) and Upper Limit Exceeded (%) values.

Note: Consult the KPI Reference for the definition of each KPI referenced in the preceding table.
The last row in the table, Total, contains the sum of values in each column for all control modes, except
the Shutdown mode.
In the Total row, if the value in the Duration (%) column is not 100%, it indicates missing data for the
selected time period, possibly due to bad quality data.

The Controller Performance Table

The controller performance table helps you analyze the performance of the control element using a range
of KPIs. It contains the following information.

60 © 2020 General Electric Company

 KPI Interpretation Required action

Manipulated Variable [%] The average value of the control variable, If the variable is not in the ideal range,
expressed as a percentage. install a different type of controller.
Depending on the variable value, install
• < 5%: Indicates that the control
either a larger controller that is more
element is operating at a very low
suitable for the process, or a smaller one
saturation level and that the element
that is more energy-efficient and cost-
is oversized.
effective.
• 5 - 30%: Indicates that the control
element is operating close to its
lower saturation level.
• 30 - 70%: Indicates that the control
element is operating at its optimal
value. This is the ideal value for the
variable.
• 70 - 95%: Indicates that the control
element is operating close to its
higher saturation level.
• > 95%: Indicates that the control
element is operating at a very high
saturation level and that the element
is undersized.

This is the average value of the

Percentage Controller Output KPI.

Total MV Movement An indication of the absolute sum of all Investigate and optimize the size of the
changes made in the MV. The closer the control system.
value for this variable is to zero, the closer
the MV is to staying at a constant value.

You can use this statistic to help you keep

the PV operating within process limits.

This is the average value of the Movement

Index KPI.

Average MV Change [%] The average percentage change in the MV Investigate and optimize the size of the
between samples. The closer the value for control system.
this variable is to zero, the closer the MV
is to staying at a constant value.

This is the average value of the Control

Amplitude KPI.

MV Oscillation Count [#] The number of times that the MV changes Add a better filter to the control system.
direction. This indicates the level of noise
in the system. The closer this value is to
zero, the less noise there is in the system.

All data is taken into account for this

statistic, regardless of the quality.

This is the average value of the MV

Oscillation Count KPI.

© 2020 General Electric Company 61

 KPI Interpretation Required action

Average oscillation amplitude [%] The average percentage change in the MV Investigate and optimize the control
before changing direction. element.
• 0: The closer this value is to zero, the
closer the MV is to staying at a
constant value.
• <5% Indicates a good control system.
• > 5%: Indicates a possibility of
damage or failure, but must be
viewed in conjunction with the
number of controller oscillations.

This is the average value of the Reversal

Amplitude KPI.

MV Saturation [%] The percentage of the reporting window Investigate for a possibility of an
for which the controller output is undersized or oversized control element,
saturated. a malfunctioning control element, or
other causes for MV saturation, and
The controller output is considered
optimize the control element.
saturated when the MV has a value of 0
(the lower limit for controller output
saturation) or 100 (the upper limit for
controller output saturation). The closer
this is to zero, the more likely it is that the
control element is appropriately sized and
functioning correctly.

This is the average value of the

Percentage MV Saturation KPI.

Duration Not Utilized [%] The percentage of the reporting window If the percentage of time not utilized is
for which the loop is not in use. too high this can be for either or both of
the following reasons:
The loop is considered to be not in use if it
is in manual mode or if the controller • The loop spends a large proportion of
output is saturated. The controller output time in manual mode. Investigate the
is considered saturated when the MV has reasons for that behavior.
a value of 0 (the lower limit for controller • Investigate for a possibility of an
output saturation) or 100 (the upper limit undersized or oversized control
for controller output saturation). The element, a malfunctioning control
closer this is to zero, the more likely it is element, or other causes for MV
that the loop is in a mode other than saturation, and optimize the control
manual and the control element is element.
appropriately sized and functioning
correctly.

This is the average value of the

Percentage Not Utilized KPI.

Note: Consult the KPI Reference for the definition of each KPI referenced in the Interpretation column of
the preceding table.

62 © 2020 General Electric Company

The PV Performance Table
The process variable performance table helps you analyze the performance of the process variable based
on values as specified in the following table.

Column name Interpretation

Average PV Limits exceeded This value is an average of the percentage of time the PV
exceeded the upper or lower limits.

This is the average value of the Percentage Limits Exceeded KPI.

PV Variance [%] This value is a measure of the average deviation of the PV from
the average PV, expressed as a percentage of the reporting
window span.

This is the average value of the PV Variance KPI.

PV Variability This value is an averaged measure of how spread or closely

clustered the PV data set is, expressed as a percentage

This is the average value of the PV Variability KPI.

Note: Consult the KPI Reference for the definition of each KPI referenced in the preceding table.

The Error Statistics Table

The error statistics table contains the statistics that indicate the extent at which the process variable
deviates from the set point. The table contains the following values:

Measured variable Interpretation

Integrated Error This value indicates the percentage time the process has spent
out of limits.

If there is a substantial integrated error, investigate the

functioning of the PV, and tune the system to maintain the PV
within the defined limits.

This is the sum of the Integrated Sum KPI values for the window.

Average Error This value indicates the average tag error. Tag error is the
difference between process variable and set point, expressed as
a percentage.

This is the average value of the PV Error Average KPI.

Average Absolute Error This value indicates the average absolute value of the tag error.

This is the average value of the PV Error Absolute Average KPI.

PV Error Standard Deviation This value indicates the standard deviation of the tag error,
where the tag error is the difference between process variable
and set point, expressed as a percentage.

This is the average value of the PV Error Standard Deviation KPI.

© 2020 General Electric Company 63

 Note: Consult the KPI Reference for the definition of each KPI referenced in the preceding table.

The Controller Configuration Table

The controller configuration table contains the Proportional (P), Integral (I), Derivative (D), or Filter (F)
values at the beginning and end of the selected time period.

64 © 2020 General Electric Company

 Chapter

5
The Loop Analysis Template
Topics:

• About the Loop Analysis

Template
• Access the Loop Analysis
Template
• Modify the Date for a Loop
Analysis

© 2020 General Electric Company 65

About the Loop Analysis Template
The Loop Analysis Template allows you to visualize and analyze loop data with grouped KPIs in various
pre-built charts. Using the template, you can investigate potential issues with a loop, perform analysis to
manually troubleshoot issues, and optimize the performance of the loop.

Access the Loop Analysis Template

Procedure
1. Access the loop dashboard whose analysis template you want to access.

2. In the workspace, select .

The loop analysis template appears, displaying the graphs, tags, tag expressions, and other
information about the template.
For more information, refer to the Analysis section of the documentation.

Modify the Date for a Loop Analysis

About This Task
This topic describes how to modify the date range for a loop analysis.

Procedure
1. Access the loop analysis whose date you want to modify.
The Date Selector box appears, displaying the default date as current date. For example, if today's
date is 21st November, 2018, then the default time range in the loop analysis is 20th November, 2018
00:00:00 to 21st November, 2018 00:00:00.
2. In the drop-down list box, select an option as specified in the following table.
Note: For each day that is included in the date range, data is reported for the complete day (that is, 24
hours). Therefore, the time that appears in the drop-down list box is not considered in the report. For
example, even if 1st November, 2018 3:00:00 to 2nd November, 2018 1:00:00 is the specified date
range, data is reported for the following date range: 1st November, 2018 00:00:00 to 3rd November,
2018 00:00:00.
Option Description
Date Select this option if you want to access data for a single day. If, however, you select the current date, data is
Selector reported for the previous day. For example, if you select 19th November, 2018, then data is reported from
19th November, 2018 00:00:00 to 20th November, 2018 00:00:00. If, however, the current date is 19th
November, 2018, then data is reported from 18th November, 2018 00:00:00 to 19th November, 2018
00:00:00.
Time Span Select this option if you want to access data for the past day, week, month, three months, six months, or year.
For example, if you select 1W, and if the current date is 21st November, 2018, then data is reported from 14th
November, 2018 00:00:00 to 21st November, 2018 00:00:00.
Date & Time Select this option if you want to access data for a day, week, month, three months, six months, or year up to a
Span selected date. If, however, you select the current date, data is reported till the previous day. For example, if you
select 13th November, 2018, and then select 1D, then data is reported from 13th November, 2018 00:00:00 to

66 © 2020 General Electric Company

 Option Description
14th November, 2018 00:00:00. If, however, the current date is 13th November, 2018, then data is reported
from 12th November, 2018 00:00:00 to 13th November, 2018 00:00:00.
Custom Select this option if you want to access data for a specific date range. If, however, you select current date, data
Time Span is reported till the previous day. For example, if you select 10th November, 2018 to 15th November, 2018, data
is reported from 10th November, 2018 00:00:00 to 16th November, 2018 00:00:00. If, however, the current
date is 15th November, 2018, then data is reported from 10th November, 2018 00:00:00 to 15th November
2018 00:00:00.

The date range is modified, and the analysis is updated. The date and time range for which the data is
reported appears in the upper-right corner of the analysis.
Tip: You can reset the date range to the default value by selecting reset.

© 2020 General Electric Company 67

 Chapter

6
Manage Loops
Topics:

• Add a Loop
• Modify a Loop
• Delete a Loop

68 © 2020 General Electric Company

Add a Loop
To add a loop:

About This Task

Note: It is important to follow these instructions carefully and to ensure that your asset ingestion file
conforms to the Control Loop Asset Definition for CLPM. If your file does not conform to this definition,
CLPM will be unable to function correctly.

Procedure
1. Define the asset classification.
2. Create asset instances.
(Follow the instructions in the APM Asset documentation for creating asset instances.)
3. Set up tag classifications for outputs.
4. Associate inputs and outputs for control loop assets.
5. Modify all existing analytic deployments to add the asset to the deployments. Consult the section of
this documentation on deploying CLPM analytics for details.
IMPORTANT: If this step is not performed, analytics will be unable to produce KPIs and alerts for the
loop.

Modify a Loop
To modify a loop:

Procedure
1. Log in to your tenant.
2. In the module navigation menu, select Assets.
3. Use the Context Browser to select the control loop asset instance you want to modify.
4. In the control loop asset instance, hover over, then select the field on the asset you want to modify. It
will only be possible to modify those properties that you should be able to modify.
For Custom Attributes on the control loop asset and on the tags associated with the control loop
asset, consult the Control Loop Asset Definition for CLPM to ensure that the changes you make comply
with the definition/template. The following custom attributes are important for CLPM configuration:
Custom Attribute Function
Attributes > Custom Attributes > Lower SP Offset Used to calculate the Lower SP Threshold.
Attributes > Custom Attributes > Upper SP Offset Used to calculate the Upper SP Threshold.
Tags > PV or MV > Custom Attributes > Minimum Operating Value Used for data clipping.
Tags > PV or MV > Custom Attributes > Maximum Operating Value Used for data clipping.

Note: It is important to ensure that your modifications to custom attributes conform to the control
loop asset definition for CLPM. If your change does not conform to this definition, CLPM will be unable
to function correctly.
For other properties common to both control loop assets and other kinds of assets, consult the APM
Asset documentation for asset instances.

© 2020 General Electric Company 69

 5. Make the change you want to make, then press Enter or select outside the field.

Results
The newly modified value for the field you have changed appears on the control loop asset instance page.

Delete a Loop
To delete a loop:

About This Task

Note: This action cannot be reversed.

Procedure
1. Log in to your tenant.
2. In the module navigation menu, select Assets.
3. Use the Context Browser to select the asset instance you want to delete.
4. Select the trash icon and confirm that you want to delete the loop.
5. Modify all existing CLPM analytic deployments to remove the asset from the deployments.

70 © 2020 General Electric Company

 Chapter

7
Reference
Topics:

• CLPM Terminology
• Control Loop Asset Definition
• KPI Reference for CLPM

© 2020 General Electric Company 71

CLPM Terminology
The following terms are frequently used in control loop performance monitoring (CLPM).

Term Description

CLPM Control Loop Performance Monitoring

PV (Process Variable) This is the measured process variable, reflecting the actual state
of the process that is being controlled.

SP (Setpoint) This is the desired value (target value) to which the PV is

controlled.

Error This is the difference between the PV and SP. It is calculated as

Error = PV-SP. Note that this is not an absolute value.

Controller The subsystem containing the control algorithm that

manipulates the MV to control the PV to the SP.

MV (Manipulated Variable) The variable manipulated (adjusted) by the controller to control

the PV to the SP.

DV (Disturbance Variable) A measured disturbance variable that can be used by the

controller to compensate for the impact of the disturbance on
the PV.

P (Proportional Value) Normally the Proportional Gain value of the PID controller.

I (Integral Value) Normally the Integral Gain value of the PID controller.

D (Derivative Value) Normally the Derivative Gain value of the PID controller.

F (Filter Value) Normally the Filter Time Constant value of the PID controller.

Final Control Element (FCE)/ Actuator The device that is physically affected by a change in the MV so
as to have an impact on the process.

Control Loop Asset Definition

The Control Loop Asset Definition specifies all the necessary attributes of a control loop asset. The various
sections of the definition are detailed here. This is the definition to be used for the JSON asset ingestion
file(s).

Define the Control Loop Asset

To define the control loop asset, perform the following steps:
1. Define the asset classification.
2. Create asset instances.
• Follow the instructions in the APM Asset documentation for creating asset instances.
3. Set up tag classifications for outputs.
4. Associate inputs and outputs for control loop assets.
Note: It is important to follow these instructions carefully and to ensure that your asset ingestion file
conforms to the control loop asset definition for CLPM. If your file does not conform to this definition,
CLPM will be unable to function correctly.

72 © 2020 General Electric Company

Define the Asset Classification
Define the asset classification for the control loop asset.

Asset Classification: Reserved Attributes

The attributes in the following table are inherited as standard asset classification required attributes. No
action is required on your part to modify them.

Property/Attribute Data Type

Family Type String

Equipment Type String

Make String

Model String

Series String

Serial Number String

Maintenance Criticality Risk Score Integer

Fault Mode String

Asset Classification: Custom Attributes

The attributes in the table that follows are custom asset classification attributes specifically for control
loop assets. You need to specify values for them.

Property/Attribute Data Type Range Property Notes

Required?

Name String N/A Required Required as an analytic input in

order to be able to log the loop
name with any individual logging
entries.

Description String N/A Optional

Loop Priority Enumeration Enumeration (High, Required Used for filtering during
Medium, Low) visualization.

Type String N/A Required. Must

be set to PID
Control Loop

Lower SP Offset Numeric >0 Required Used to calculate the lower SP

threshold.

Upper SP Offset Numeric >0 Required Used to calculate the upper SP

threshold.

© 2020 General Electric Company 73

Set Up Tag Classifications for Outputs
Define each analytic output using the Control Loop KPI tag classification described in the table that
follows.

Tag Classification

Required Property / Attribute Data Type

Source Unit of Measure String

Data Type String

Category String

Maximum Operating Value Double

Minimum Operating Value Double

Highest Threshold Value Double

Lowest Threshold Value Double

Resolution String

Sensor Health Index Low Threshold Double

Sensor Health Index High Threshold Double

Sensor Health Index Flat Line Number Integer

Sensor Health Index Flat Line Epsilon Double

Bad Observation Persistent / Penalty Window Integer

False Alarm Probability Double

Missed Alarm Probability Double

Sample Failure Magnitudes Mean Integer

Sample Failure Magnitudes Variance Integer

Null Hypothesis Variance Double

NaN Test Weight Double

Outlier Test Weight Double

Flat Test Weight Double

Associate Analytic Inputs for Control Loop Assets

Specify analytic inputs by mapping timeseries tags onto asset instances defined by the control loop asset
classification.
Each analytic input is uniquely identified by the Tag Alias, and the data type must match the
corresponding type specified in the table that follows.
However, no specific tag classification is required for analytic inputs. You may use whatever tag
classification you want to use.

74 © 2020 General Electric Company

 Note: As you can see in the table that follows, the Maximum Operating Value and Minimum
Operating Value tag attributes for the PV are significant because they are used for data clipping.
Note: It is important to carefully follow this specification. If your file does not conform to this definition,
CLPM will be unable to function correctly.

Tag Tag Tag Attributes (Required) Tag User Require

Alias Descrip Value Input d?
UOM Data Categor Maximu Minimu Highest Lowest
tion Ranges Require
Type y m m Thresho Thresho
d?
Operati Operati ld Value ld Value
ng ng
Value Value

Setpoint The Specifies Numeric Any Any Any Any Any Any No Yes
setpoint the
for the appropri
process ate unit
variable of
(PV) measure
represen .
ts the
desired
value for
PV.

PV Process Specifies Numeric Any Used as Used as Any Any Any No Yes
variable the a a
appropri custom custom
ate unit attribute attribute
of to to
measure specify specify
. the the
upper lower
limit of limit of
the PV the PV
tag. tag.

This is This is
the the
upper lower
Engineer Engineer
ing ing
(clipping (clipping
) limit ) limit
for PV. for PV.
Data Data
outside outside
of this of this
limit is limit is
not used not used
for for
calculati calculati
ons. ons.

© 2020 General Electric Company 75

 Tag Tag Tag Attributes (Required) Tag User Require
Alias Descrip Value Input d?
UOM Data Categor Maximu Minimu Highest Lowest
tion Ranges Require
Type y m m Thresho Thresho
d?
Operati Operati ld Value ld Value
ng ng
Value Value

MV The Specifies Numeric Any Any Any Any Any Percent No Yes
controlle the age
r output appropri [0%-100
or ate unit %]
"Manipul of
ated measure
Variable .
"

P Proporti Specifies Numeric Any Any Any Any Any >0 Optional No
onal the
value appropri
ate unit
of
measure
.

I Integral Specifies Numeric Any Any Any Any Any >0 Optional No
value the
appropri
ate unit
of
measure
.

D Derivati Specifies Numeric Any Any Any Any Any >0 Optional No
ve value the
appropri
ate unit
of
measure
.

76 © 2020 General Electric Company

 Tag Tag Tag Attributes (Required) Tag User Require
Alias Descrip Value Input d?
UOM Data Categor Maximu Minimu Highest Lowest
tion Ranges Require
Type y m m Thresho Thresho
d?
Operati Operati ld Value ld Value
ng ng
Value Value

F Filter Specifies Numeric Any Any Any Any Any >0 Optional No
value the
appropri
ate unit
of
measure
.

Control Control Specifies Numeric Any Any Any Any Any Enumer No Yes
Mode Mode the ation
appropri
(Auto,
ate unit
Manual,
of
Cascade
measure
,
.
Shutdo
wn)

Associate Analytic Outputs for Control Loop Assets

Specify analytic outputs by mapping timeseries tags onto asset instances defined by the control loop
asset definition.

Analytic Outputs
Each output is identified with the Control Loop KPI tag classification.
Consult the KPI Reference topic for descriptions of the KPIs corresponding to each output tag, along with
formulas and notes on interpreting each KPI.
Note: The following applies to the Tag Source ID.
• The Tag Source ID must be of the form <AssetID>.<Output Tag Name> for all analytic outputs.
• The <AssetID> in the table that follows refers to the ID of the asset to which the tag belongs.
• For example, observe that the Tag Association in the Example Loop Asset for Percentage
Controller Output, for the Loop1 asset with ID Loop1, the Tag Source ID must be
Loop1.Percentage Controller Output.
• When manually creating the tags on the asset, this property is referred to as the Source ID.
• When constructing Asset Ingestion files, this property is referenced using the ID in the tags
section.
Note:
The Master Quality referenced in the table that follows is good if all the following tags are of good quality:
• Control Type
• MV
• PV
• SP

© 2020 General Electric Company 77

 Otherwise, it is bad.

Output Tag Tag Attributes (Required) Tag Quality

Tag Descrip Value
UOM Data Categor Maximu Minimu Highest Lowest Tag
Name tion Ranges
Type y m m Thresho Thresho Source
Operati Operati ld Value ld Value ID
ng ng
Value Value

Percent The Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age average the D>.Perc age Quality
Controll MV appropri entage [0%-100
er value ate unit Controll %]
Output (express of er
ed as a measure Output
%). .

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age Not age Not the D>.Perc age Quality
Utilized Utilized appropri entage [0%-100
ate unit Not %]
of Utilized
measure
.

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age MV age MV the D>.Perc age Quality
Saturati Saturati appropri entage [0%-100
on on ate unit MV %]
of Saturati
measure on
.

Overall Overall Specifies Numeric Any Any Any Any Any <AssetI Percent Master
Perform Perform the D>.Over age Quality
ance ance appropri all [0%-100
ate unit Perform %]
of ance
measure
.

Moveme Gives Specifies Numeric Any Any Any Any Any <AssetI Master
nt Index the sum the D>.Move Quality
of all appropri ment
changes ate unit Index
of the of
MV, measure
divided .
by 100

78 © 2020 General Electric Company

 Output Tag Tag Attributes (Required) Tag Quality
Tag Descrip Value
UOM Data Categor Maximu Minimu Highest Lowest Tag
Name tion Ranges
Type y m m Thresho Thresho Source
Operati Operati ld Value ld Value ID
ng ng
Value Value

Control Calculat Specifies Numeric Any Any Any Any Any <AssetI Master
Amplitu es the the D>.Cont Quality
de sum of appropri rol
all MV ate unit Amplitu
rate of of de
change measure
values .
divided
by the
total
number
of
samples

Reversal The Specifies Numeric Any Any Any Any Any <AssetI MV
Count number the D>.Reve Quality
of times appropri rsal
MV ate unit Count
changes of
direction measure
.

Reversal Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
Amplitu age the D>.Reve age Quality
de Reversal appropri rsal [0%-100
Amplitu ate unit Amplitu %]
de of de
measure
.

Integrat Integrat Specifies Numeric Any Any Any Any Any <AssetI Master
ed Sum ed Sum the D>.Integ Quality
appropri rated
ate unit Sum
of
measure
.

PV Error PV Error Specifies Numeric Any Any Any Any Any <AssetI Master
Average Average the D>.PV Quality
appropri Error
ate unit Average
of
measure
.

© 2020 General Electric Company 79

 Output Tag Tag Attributes (Required) Tag Quality
Tag Descrip Value
UOM Data Categor Maximu Minimu Highest Lowest Tag
Name tion Ranges
Type y m m Thresho Thresho Source
Operati Operati ld Value ld Value ID
ng ng
Value Value

PV Error PV Error Specifies Numeric Any Any Any Any Any <AssetI Master
Absolut Absolut the D>.PV Quality
e e appropri Error
Average Average ate unit Absolut
of e
measure Average
.

PV Error PV Error Specifies Numeric Any Any Any Any Any <AssetI Master
Standar Standar the D>.PV Quality
d d appropri Error
Deviatio Deviatio ate unit Standar
n n of d
measure Deviatio
. n

PV PV Specifies Numeric Any 100 0 100 0 <AssetI Master

Variance Variance the D>.PV Quality
appropri Variance
ate unit
of
measure
.

PV PV Specifies Numeric Any Any Any Any Any <AssetI Master

Variabili Variabili the D>.PV Quality
ty ty appropri Variabili
ate unit ty
of
measure
.

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age age the D>.Perc age Quality
Good Good appropri entage [0%-100
Quality Quality ate unit Good %]
of Quality
measure
.

Total Total Specifies Numeric Any Any Any Any Any <AssetI >=0 Worst
PIDF PIDF the D>.Total Quality
Changes Changes appropri PIDF of P, I, D,
ate unit Changes F
of
measure
.

80 © 2020 General Electric Company

 Output Tag Tag Attributes (Required) Tag Quality
Tag Descrip Value
UOM Data Categor Maximu Minimu Highest Lowest Tag
Name tion Ranges
Type y m m Thresho Thresho Source
Operati Operati ld Value ld Value ID
ng ng
Value Value

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age age the D>.Perc age Quality
Manual Manual appropri entage [0%-100
ate unit Manual %]
of
measure
.

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age age the D>.Perc age Quality
Manual Manual appropri entage [0%-100
LL LL ate unit Manual %]
Exceede Exceede of LL
d d measure Exceede
. d

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age age the D>.Perc age Quality
Manual Manual appropri entage [0%-100
UL UL ate unit Manual %]
Exceede Exceede of UL
d d measure Exceede
. d

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age Auto age Auto the D>.Perc age Quality
appropri entage [0%-100
ate unit Auto %]
of
measure
.

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age Auto age Auto the D>.Perc age Quality
LL LL appropri entage [0%-100
Exceede Exceede ate unit Auto LL %]
d d of Exceede
measure d
.

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age Auto age Auto the D>.Perc age Quality
UL UL appropri entage [0%-100
Exceede Exceede ate unit Auto UL %]
d d of Exceede
measure d
.

© 2020 General Electric Company 81

 Output Tag Tag Attributes (Required) Tag Quality
Tag Descrip Value
UOM Data Categor Maximu Minimu Highest Lowest Tag
Name tion Ranges
Type y m m Thresho Thresho Source
Operati Operati ld Value ld Value ID
ng ng
Value Value

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age age the D>.Perc age Quality
Cascade Cascade appropri entage [0%-100
ate unit Cascade %]
of
measure
.

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age age the D>.Perc age Quality
Cascade Cascade appropri entage [0%-100
LL LL ate unit Cascade %]
Exceede Exceede of LL
d d measure Exceede
. d

Percent Percent Specifies Numeric Any 100 0 100 0 AssetID Percent Master
age age the >.Percen age Quality
Cascade Cascade appropri tage [0%-100
UL UL ate unit Cascade %]
Exceede Exceede of UL
d d measure Exceede
. d

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age age the D>.Perc age Quality
Shutdo Shutdo appropri entage [0%-100
wn wn ate unit Shutdo %]
of wn
measure
.

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age age the D>.Perc age Quality
Shutdo Shutdo appropri entage [0%-100
wn LL wn LL ate unit Shutdo %]
Exceede Exceede of wn LL
d d measure Exceede
. d

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age age the D>.Perc age Quality
Shutdo Shutdo appropri entage [0%-100
wn UL wn UL ate unit Shutdo %]
Exceede Exceede of wn UL
d d measure Exceede
. d

82 © 2020 General Electric Company

 Output Tag Tag Attributes (Required) Tag Quality
Tag Descrip Value
UOM Data Categor Maximu Minimu Highest Lowest Tag
Name tion Ranges
Type y m m Thresho Thresho Source
Operati Operati ld Value ld Value ID
ng ng
Value Value

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age age the D>.Perc age Quality
Control Control appropri entage [0%-100
On On ate unit Control %]
of On
measure
.

Statistic The time Specifies Numeric Any Any 0 Any 0 <AssetI >=0 None
s Period period in the D>.Stati
seconds appropri stics
over ate unit Period
which of
the KPIs measure
have .
been
calculat
ed.

Percent Percent Specifies Numeric Any 100 0 100 0 <AssetI Percent Master
age age the D>.Perc age Quality
Limits Limits appropri entage [0%-100
Exceede Exceede ate unit Limits %]
d d of Exceede
measure d
.

PV Error PV Error Specifies Numeric Any Any Any Any Any <AssetI Master
the D>.PV Quality
appropri Error
ate unit
of
measure
.

Lower Lower Specifies Numeric Any Any Any Any Any <AssetI Setpoint
SP SP the D>.Lowe Quality
Threshol Threshol appropri r SP
d d ate unit Threshol
of d
measure
.

© 2020 General Electric Company 83

 Output Tag Tag Attributes (Required) Tag Quality
Tag Descrip Value
UOM Data Categor Maximu Minimu Highest Lowest Tag
Name tion Ranges
Type y m m Thresho Thresho Source
Operati Operati ld Value ld Value ID
ng ng
Value Value

Upper Upper Specifies Numeric Any Any Any Any Any <AssetI Setpoint
SP SP the D>.Uppe Quality
Threshol Threshol appropri r SP
d d ate unit Threshol
of d
measure
.

MV MV Specifies Numeric Any Any 0 Any 0 <AssetI

Oscillati Oscillati the D>.MV
on on appropri Oscillati
Count Count ate unit on
of Count
measure
.

Example Loop Asset for APM Asset Ingestion

The following sample JSON gives an example of the definition of a control loop asset for ingestion into
APM.

The sample JSON below is split up into asset classification, asset instances, tag classifications, and tag
associations.
This example produces two loops (Loop1 and Loop2).

Asset Classification

{
"classifications": [
{
"id": "CLPM-Enterprise",
"name": "Control Loop Enterprise Classification",
"description": "",
"ccomClass": "ENTERPRISE_TYPE",
"parent": null
},
{
"id": "CLPM-Site",
"name": "Control Loop Site Classification",
"description": "",
"ccomClass": "SITE_TYPE",
"parent": null
},
{
"id": "CLPM-Segment",
"name": "Control Loop Segment Classification",
"description": "",

84 © 2020 General Electric Company

 "ccomClass": "SEGMENT_TYPE",
"parent": null
},
{
"id": "ControlLoop",
"name": "Control Loop",
"description": "",
"ccomClass": "ASSET_TYPE",
"parent": null,
"properties": [
{
"id": "Name",
"value": [],
"type": "String"
},
{
"id": "Description",
"value": [],
"type": "String"
},
{
"id": "Loop Priority",
"value": [
"Medium"
],
"type": "String"
},
{
"id": "Type",
"value": [],
"type": "String"
},
{
"id": "Lower SP Offset",
"value": [

],
"type": "Double"
},
{
"id": "Upper SP Offset",
"value": [
],
"type": "Double"
}
]
}
]
}

Instances

{
"instances": [
{
"id": "SavoryFoodsInc",
"name": "Savory Foods Inc",
"description": "This is a digital manufacturing
enterprise",

© 2020 General Electric Company 85

 "classification": "CLPM-Enterprise",
"ccomClass": "ENTERPRISE"
},
{
"id": "CapeTownSouthAfricaFactory",
"name": "Cape Town, South Africa Factory",
"description": "This is the Cape Town, South African
factory",
"properties": [
{
"id": "Address",
"value": [
"Cape Town"
],
"type": "string"
}
],
"classification": "CLPM-Site",
"ccomClass": "SITE"
},
{
"id": "TortillaPlant",
"name": "Tortilla Plant",
"description": "This is the tortilla plant",
"properties": [
{
"id": "PlantNumber",
"value": [
"TortillaPlant001"
],
"type": "string"
}
],
"classification": "CLPM-Segment",
"ccomClass": "SEGMENT"
},
{
"id": "Loop1",
"name": "Loop1",
"description": "The tortilla fryer",
"properties": [
{
"id": "Lower SP Offset",
"value": [
20
],
"type": "Double"
},
{
"id": "Upper SP Offset",
"value": [
120
],
"type": "Double"
}
],
"classification": "ControlLoop",
"ccomClass": "ASSET"
},
{
"id": "Loop2",

86 © 2020 General Electric Company

 "name": "Loop2",
"description": "The tortilla fryer",
"properties": [
{
"id": "alias",
"value": [
"Fryer"
],
"type": "string"
},
{
"id": "Lower SP Offset",
"value": [
30
],
"type": "Double"
},
{
"id": "Upper SP Offset",
"value": [
150
],
"type": "Double"
}
],
"classification": "ControlLoop",
"ccomClass": "ASSET"
}
],
"connections": [
{
"from": {
"id": "CapeTownSouthAfricaFactory",
"ccomClass": "SITE"
},
"to": [
{
"type": "parent",
"id": "SavoryFoodsInc",
"ccomClass": "ENTERPRISE"
}
]
},
{
"from": {
"id": "TortillaPlant",
"ccomClass": "SEGMENT"
},
"to": [
{
"type": "parent",
"id": "CapeTownSouthAfricaFactory",
"ccomClass": "SITE"
}
]
},
{
"from": {
"id": "Loop1",
"ccomClass": "ASSET"
},

© 2020 General Electric Company 87

 "to": [
{
"type": "parent",
"id": "TortillaPlant",
"ccomClass": "SEGMENT"
}
]
},
{
"from": {
"id": "Loop2",
"ccomClass": "ASSET"
},
"to": [
{
"type": "parent",
"id": "TortillaPlant",
"ccomClass": "SEGMENT"
}
]
}
]
}

Tag Classifications

{
"tagClassifications": [
{
"id": "ControlLoopKPI",
"name": "Control Loop KPI",
"description": "",
"unitGroup": "",
"reservedProperties": {
"uom": "",
"dataType": "Double",
"category": ""
}
},
{
"id": "AnalyticInputs",
"name": "AnalyticInputs",
"description": "Tag Classification for analytic input
tags",
"unitGroup": ""
}
]
}

Tag Associations

{
"tagAssociations": [
{
"monitoredEntity": {
"id": "Loop1",
"ccomClass": "ASSET"
},

88 © 2020 General Electric Company

 "tags": [
{
"name": "Control Mode",
"id": "Loop1.ControlMode",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"timeseriesLink": "Loop1.ControlMode",
"dataType": "String"
},
"aliases": [
"Control Mode"
]
},
{
"name": "D",
"id": "Loop1.D",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"D"
]
},
{
"name": "F",
"id": "Loop1.F",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"F"
]
},
{
"name": "I",
"id": "Loop1.I",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"I"
]
},
{
"name": "MV",
"id": "Loop1.CV",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"MV"
],

© 2020 General Electric Company 89

 "properties": [
{
"id": "Maximum Operating Value",
"value": [
120
],
"type": "Double"
},
{
"id": "Minimum Operating Value",
"value": [
0
],
"type": "Double"
}
]
},
{
"name": "P",
"id": "Loop1.P",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"P"
]
},
{
"name": "PV",
"id": "Loop1.PV",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"PV"
],
"properties": [
{
"id": "Maximum Operating Value",
"value": [
120
],
"type": "Double"
},
{
"id": "Minimum Operating Value",
"value": [
0
],
"type": "Double"
}
]
},
{
"name": "Setpoint",
"id": "Loop1.Setpoint",
"classification": "AnalyticInputs",

90 © 2020 General Electric Company

 "unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"Setpoint"
]
},
{
"name": "Control Amplitude",
"id": "Loop1.Control Amplitude",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Integrated Sum",
"id": "Loop1.Integrated Sum",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Lower SP Threshold",
"id": "Loop1.Lower SP Threshold",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Movement Index",
"id": "Loop1.Movement Index",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "MV Oscillation Count",
"id": "Loop1.MV Oscillation Count",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"minOperatingValue": 0,
"lowestThresholdValue": 0
}
},
{
"name": "Overall Performance",
"id": "Loop1.Overall Performance",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {

© 2020 General Electric Company 91

 "dataType": "Double"
}
},
{
"name": "Percentage Auto LL Exceeded",
"id": "Loop1.Percentage Auto LL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Auto UL Exceeded",
"id": "Loop1.Percentage Auto UL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Auto",
"id": "Loop1.Percentage Auto",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Cascade LL Exceeded",
"id": "Loop1.Percentage Cascade LL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Cascade UL Exceeded",
"id": "Loop1.Percentage Cascade UL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {

92 © 2020 General Electric Company

 "dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Cascade",
"id": "Loop1.Percentage Cascade",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Controller Output",
"id": "Loop1.Percentage Controller Output",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Control On",
"id": "Loop1.Percentage Control On",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Good Quality",
"id": "Loop1.Percentage Good Quality",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Limits Exceeded",

© 2020 General Electric Company 93

 "id": "Loop1.Percentage Limits Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Manual LL Exceeded",
"id": "Loop1.Percentage Manual LL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Manual UL Exceeded",
"id": "Loop1.Percentage Manual UL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Manual",
"id": "Loop1.Percentage Manual",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage MV Saturation",
"id": "Loop1.Percentage MV Saturation",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0

94 © 2020 General Electric Company

 }
},
{
"name": "Percentage Not Utilized",
"id": "Loop1.Percentage Not Utilized",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Shutdown LL Exceeded",
"id": "Loop1.Percentage Shutdown LL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Shutdown UL Exceeded",
"id": "Loop1.Percentage Shutdown UL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Shutdown",
"id": "Loop1.Percentage Shutdown",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "PV Error Absolute Average",
"id": "Loop1.PV Error Absolute Average",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"

© 2020 General Electric Company 95

 }
},
{
"name": "PV Error Average",
"id": "Loop1.PV Error Average",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "PV Error Standard Deviation",
"id": "Loop1.PV Error Standard Deviation",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "PV Error",
"id": "Loop1.PV Error",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "PV Variability",
"id": "Loop1.PV Variability",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "PV Variance",
"id": "Loop1.PV Variance",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Reversal Amplitude",
"id": "Loop1.Reversal Amplitude",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,

96 © 2020 General Electric Company

 "lowestThresholdValue": 0
}
},
{
"name": "Reversal Count",
"id": "Loop1.Reversal Count",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"minOperatingValue": 0,
"lowestThresholdValue": 0
}
},
{
"name": "Total PIDF Changes",
"id": "Loop1.Total PIDF Changes",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Upper SP Threshold",
"id": "Loop1.Upper SP Threshold",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Statistics Period",
"id": "Loop1.Statistics Period",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
}
]
},
{
"monitoredEntity": {
"id": "Loop2",
"ccomClass": "ASSET"
},
"tags": [
{
"name": "Lower PV Limit",
"id": "Loop2.LowerPVLimit",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"Lower PV Limit"
]
},

© 2020 General Electric Company 97

 {
"name": "Upper PV Limit",
"id": "Loop2.UpperPVLimit",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"Upper PV Limit"
]
},
{
"name": "Setpoint",
"id": "Loop2.Setpoint",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"Setpoint"
]
},
{
"name": "PV",
"id": "Loop2.PV",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"PV"
],
"properties": [
{
"id": "Maximum Operating Value",
"value": [
120
],
"type": "Double"
},
{
"id": "Minimum Operating Value",
"value": [
0
],
"type": "Double"
}
]
},
{
"name": "MV",
"id": "Loop2.MV",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [

98 © 2020 General Electric Company

 "MV"
],
"properties": [
{
"id": "Maximum Operating Value",
"value": [
100
],
"type": "Double"
},
{
"id": "Minimum Operating Value",
"value": [
0
],
"type": "Double"
}
]
},
{
"name": "P",
"id": "Loop2.P",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"P"
]
},
{
"name": "I",
"id": "Loop2.I",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"I"
]
},
{
"name": "D",
"id": "Loop2.D",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "Double"
},
"aliases": [
"D"
]
},
{
"name": "F",
"id": "Loop2.F",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {

© 2020 General Electric Company 99

 "dataType": "Double"
},
"aliases": [
"F"
]
},
{
"name": "Control Mode",
"id": "Loop2.ControlMode",
"classification": "AnalyticInputs",
"unit": "",
"reservedProperties": {
"dataType": "String"
},
"aliases": [
"Control Mode"
]
},
{
"name": "Total PIDF Changes",
"id": "Loop2.Total PIDF Changes",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Percentage Manual",
"id": "Loop2.Percentage Manual",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Manual LL Exceeded",
"id": "Loop2.Percentage Manual LL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Manual UL Exceeded",
"id": "Loop2.Percentage Manual UL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,

100 © 2020 General Electric Company

 "minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Auto",
"id": "Loop2.Percentage Auto",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Auto LL Exceeded",
"id": "Loop2.PercentageAuto LL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Auto UL Exceeded",
"id": "Loop2.Percentage Auto UL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Cascade",
"id": "Loop2.Percentage Cascade",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Cascade LL Exceeded",
"id": "Loop2.Percentage Cascade LL Exceeded",
"classification": "ControlLoopKPI",

© 2020 General Electric Company 101

 "unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Cascade UL Exceeded",
"id": "Loop2.Percentage Cascade UL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Shutdown",
"id": "Loop2.Percentage Shutdown",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Shutdown LL Exceeded",
"id": "Loop2.Percentage Shutdown LL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Shutdown UL Exceeded",
"id": "Loop2.Percentage Shutdown UL Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},

102 © 2020 General Electric Company

 {
"name": "Percentage Control On",
"id": "Loop2.Percentage Control On",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Controller Output",
"id": "Loop2.Percentage Controller Output",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Good Quality",
"id": "Loop2.Percentage Good Quality",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage Not Utilized",
"id": "Loop2.Percentage Not Utilized",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Percentage MV Saturation",
"id": "Loop2.Percentage MV Saturation",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,

© 2020 General Electric Company 103

 "highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Overall Performance",
"id": "Loop2.Overall Performance",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Movement Index",
"id": "Loop2.Movement Index",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Control Amplitude",
"id": "Loop2.Control Amplitude",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Reversal Count",
"id": "Loop2.Reversal Count",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"minOperatingValue": 0,
"lowestThresholdValue": 0
}
},
{
"name": "Reversal Amplitude",
"id": "Loop2.Reversal Amplitude",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "Integrated Sum",
"id": "Loop2.Integrated Sum",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {

104 © 2020 General Electric Company

 "dataType": "Double"
}
},
{
"name": "PV Error Average",
"id": "Loop2.PV Error Average",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "PV Error Absolute Average",
"id": "Loop2.PV Error Absolute Average",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "PV Error Standard Deviation",
"id": "Loop2.PV Error Standard Deviation",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "PV Variance",
"id": "Loop2.PV Variance",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,
"highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "PV Variability",
"id": "Loop2.PV Variability",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Percentage Limits Exceeded",
"id": "Loop2.Percentage Limits Exceeded",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"maxOperatingValue": 100,
"minOperatingValue": 0,

© 2020 General Electric Company 105

 "highestThresholdValue": 100,
"lowestThresholdValue": 0
}
},
{
"name": "PV Error",
"id": "Loop2.PV Error",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Lower SP Threshold",
"id": "Loop2.Lower SP Threshold",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "Upper SP Threshold",
"id": "Loop2.Upper SP Threshold",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
},
{
"name": "MV Oscillation Count",
"id": "Loop2.MV Oscillation Count",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double",
"minOperatingValue": 0,
"lowestThresholdValue": 0
}
},
{
"name": "Statistics Period",
"id": "Loop2.Statistics Period",
"classification": "ControlLoopKPI",
"unit": "",
"reservedProperties": {
"dataType": "Double"
}
}
]
}
]
}

106 © 2020 General Electric Company

KPI Reference for CLPM
This reference details all the KPIs calculated over the data window.
These KPIs are produced by deployed CLPM analytics and are associated with assets in Predix TimeSeries.
The KPIs are also used to produce the tables and charts on the Loop Dashboard and Fleet Dashboard,
which help you monitor and diagnose loop performance.
The KPIs are categorized in the tables that follow as:
• Aggregate KPIs
• Non-Aggregate KPIs
Note: The samples over which these KPIs are calculated are resampled to be at regular intervals. The
resulting samples are therefore equally weighted in the calculations. All KPIs are calculated off
interpolated data.

© 2020 General Electric Company 107

Aggregate KPIs
These aggregate KPIs are calculated over the data window.

KPI Description Formula Interpretation Source KPI

Analytic

Control The Control Amplitude Control Amplitude = SUM(| The closer this KPI is to zero OPM-CLPM-
Amplitude gives the sum of all MV_ROC|) / Number of Samples the better. An amplitude Performance_E
changes of the MV, divided close to zero indicates that xt
where:
by the number of samples. the manipulated variable
Only data of good quality is • MV_ROC = MV current - MV does not need to be greatly
considered for the previous adjusted in order to control
calculation of this value. • MV_ROC Quality is the worst the PV to the SP.
Bad quality data is ignored. quality of MV current and
If this KPI is large, it indicates
MV previous.
a large adjustment to the
MV. One of the following is
likely the cause of this:

• The sensor or FCE/

Actuator is broken.
• The FCE/Actuator is
incorrectly sized.
• An unsuitable MV with
little or no leverage is
currently selected to
control the PV.
• The loop settings (for
tuning) are too
aggressive.

Integrated Sum This gives the sum of the Integrated Sum = SUM(|PV-SP|) This KPI gives an indication OPM-CLPM-
absolute values of all PV for samples of good quality in of the total size of the Performance_E
Error (PV - SP) samples window absolute PV Error over the xt
within the data window whole reporting period. The
that are of good quality. closer this KPI is to zero the
better.

It is not normalized, so it
cannot be directly compared
across different loops, but is
best analyzed over time for
the same loop to see how
well a specific loop is
controlling over time.

108 © 2020 General Electric Company

 KPI Description Formula Interpretation Source KPI
Analytic

Movement The Movement Index gives Movement Index = SUM(| The closer this KPI is to zero OPM-CLPM-
Index the sum of all changes of MV_ROC|) / 100 the better. An index close to Performance_E
the MV, divided by 100. zero indicates that the xt
where:
Only data of good quality is manipulated variable does
considered for the • MV_ROC = MV current - MV not need to be adjusted a lot
calculation of this value. previous in order to control the PV to
Bad quality data is ignored. • MV_ROC Quality is the worst the SP.
quality of MV current and
If this KPI is large, it indicates
MV previous.
a lot of adjustment to the
MV. One of the following is
likely the cause of this:

• The sensor or the FCE/

Actuator is broken.
• The FCE/Actuator is
incorrectly sized.
• An unsuitable MV with
little or no leverage is
currently selected to
control the PV.
• The loop settings (for
tuning) are too
aggressive.

MV Oscillation This calculates the number Reversal Count The closer this KPI is to zero OPM-CLPM-
Count of times that the MV the better. A large value for Performance_E
= Reversal Count / 2
oscillates (changes this KPI indicates that MV xt
= COUNT(MV direction change) /
direction twice) over the changes tend to overshoot
2
reporting period. the mark and require MV
changes in the opposite
direction to compensate.

© 2020 General Electric Company 109

 KPI Description Formula Interpretation Source KPI
Analytic

Overall This indicates the overall OverallPerformance = This gives a broad picture of OPM-CLPM-
Performance performance of the loop by the performance of the Performance
Percentage Manual / 3 +
combining the (equally system. This should be low.
Percentage MV Saturation / 3 +
weighted) Percentage
Percentage Limits Exceeded / 3 If this KPI is large, one of the
Manual, Percentage MV
following is likely the cause
Saturation, and Percentage where Percentage Manual,
of this:
Limits Exceeded KPIs into Percentage MV Saturation, and
a single percentage value. Percentage Limits Exceeded are • The sensor, or FCE/
KPIs described elsewhere in this Actuator is broken.
table. • The SP is set at an
unrealistic/unreachable
value.
• The FCE/Actuator is
incorrectly sized.
• The loop is set to
Manual mode for most
of the time.
• The loop control limits
(specifying the target
control range) are set to
an unrealistically narrow
range.
• The current control loop
design, settings, and
algorithm are unable to
effectively control the
target process or
equipment.

Examine the individual KPIs

that comprise this KPI for a
more precise indicator of
what is contributing to the
behavior of the system.

110 © 2020 General Electric Company

 KPI Description Formula Interpretation Source KPI
Analytic

Percentage This indicates the % Auto = Number of samples For Auto control mode, a OPM-CLPM-
Auto percentage of the samples WHERE (Auto control mode AND controller within a control Performance
of good quality within the samples of good quality)/ system implements
data window for which the Number of samples of good automated changes to a
control mode used was quality variable when a sensor
Auto. measures the process
operating outside the
defined process limits. The
output is calculated by the
controller using the error
signal, which is the
difference between the set
point and the process
variable. Ideally, the control
loop would always be in
either Auto or Cascade
mode, implying that the loop
is switched ON to actively,
automatically, and
continuously control the PV
to SP.

© 2020 General Electric Company 111

 KPI Description Formula Interpretation Source KPI
Analytic

Percentage This indicates the % Auto LL Exceeded = Number of If PV violates the lower OPM-CLPM-
Auto LL percentage of the samples samples WHERE (Auto control control limits while the loop Performance
Exceeded of good quality within the mode AND PV < Lower SP Offset is in Auto mode, it means
data window for which the AND samples of good quality) / one of the following:
control mode was set to Number of samples of good • The loop tuning is
Auto and the PV was quality incorrect.
below the Lower SP
• The control loop is not
Offset..
able to do its job or is
not suitable for its job.
Possible reasons can
include:
◦ The loop is complex
loop and you need
an Advanced
Controller like MPC
to control this type
of loop.
◦ The control loop is
not designed
correctly.
◦ The sensor or FCE/
actuator is
unsuitable or is
faulty.
◦ The equipment
being controlled is
malfunctioning.
• The control loop and
process/equipment are
OK, but the control
limits are too tight/strict
and should be reviewed/
updated.

112 © 2020 General Electric Company

 KPI Description Formula Interpretation Source KPI
Analytic

Percentage This indicates the % Auto UL Exceeded = Number If PV violates the upper OPM-CLPM-
Auto UL percentage of the samples of samples WHERE (Auto control control limits while the loop Performance
Exceeded of good quality within the mode AND PV > Upper SP Offset is in Auto mode, it means
data window for which the AND samples of good quality) / one of the following:
control mode was set to Number of samples of good • The loop tuning is
Auto and the PV was quality incorrect.
above the Upper SP Offset.
• The control loop is not
able to do its job or is
not suitable for its job.
Possible reasons can
include:
◦ The loop is complex
loop and you need
an Advanced
Controller like MPC
to control this type
of loop.
◦ The control loop is
not designed
correctly.
◦ The sensor or FCE/
actuator is
unsuitable or is
faulty.
◦ The equipment
being controlled is
malfunctioning.
• The control loop and
process/equipment are
OK, but the control
limits are too tight/strict
and should be reviewed/
updated.

© 2020 General Electric Company 113

 KPI Description Formula Interpretation Source KPI
Analytic

Percentage This indicates the % Cascade = Number of samples In Cascade control mode the OPM-CLPM-
Cascade percentage of the samples WHERE (Cascade control mode loop output/MV is not Performance
of good quality within the AND samples of good quality)/ directly driving an actuator,
data window for which the Number of samples of good but rather it is linked to or
control mode used was quality setting or calculating the SP
Cascade. for another controller loop.

For example, a flow control

loop is trying to control the
flow, measured by a flow
meter, to a desired SP by
changing a valve position.
The flow control loop
calculates a desired valve
position, which in turn
becomes the requested
position SP for a valve
positioning control loop that
changes a pressure to move
the valve to a desired
position. The two loops are
linked together in cascade.
The flow control loop
depends on the valve
positioning control loop to
do its job; if the latter does
not work well, the former
will also not work well.

114 © 2020 General Electric Company

 KPI Description Formula Interpretation Source KPI
Analytic

Percentage This indicates the % Cascade LL Exceeded = If PV violates the lower OPM-CLPM-
Cascade LL percentage of the samples Number of samples WHERE control limits while the loop Performance
Exceeded of good quality within the (Cascade control mode AND PV < is in Cascade mode, it means
data window for which the Lower SP Offset AND samples of one of the following:
control mode was set to good quality) / Number of • The loop tuning is
Cascade and the PV was samples of good quality incorrect.
below the Lower SP Offset.
• The control loop is not
able to do its job or is
not suitable for its job.
Possible reasons can
include:
◦ The loop is complex
loop and you need
an Advanced
Controller like MPC
to control this type
of loop.
◦ The control loop is
not designed
correctly.
◦ The sensor or FCE/
actuator is
unsuitable or is
faulty.
◦ The equipment
being controlled is
malfunctioning.
◦ The downstream,
dependent
cascaded loop has a
problem.
• The control loop and
process/equipment are
OK, but the control
limits are too tight/strict
and should be reviewed/
updated.

© 2020 General Electric Company 115

 KPI Description Formula Interpretation Source KPI
Analytic

Percentage This indicates the % Auto UL Exceeded = Number If PV violates the upper OPM-CLPM-
Cascade UL percentage of the samples of samples WHERE (Auto control control limits while the loop Performance
Exceeded of good quality within the mode AND PV > Upper SP Offset is in Cascade mode, it means
data window for which the AND samples of good quality) / one of the following:
control mode was set to Number of samples of good • The loop tuning is
Cascade and the PV was quality incorrect.
above the Upper SP Offset.
• The control loop is not
able to do its job or is
not suitable for its job.
Possible reasons can
include:
◦ The loop is complex
loop and you need
an Advanced
Controller like MPC
to control this type
of loop.
◦ The control loop is
not designed
correctly.
◦ The sensor or FCE/
actuator is
unsuitable or is
faulty.
◦ The equipment
being controlled is
malfunctioning.
◦ The downstream,
dependent
cascaded loop has a
problem.
• The control loop and
process/equipment are
OK, but the control
limits are too tight/strict
and should be reviewed/
updated.

Percentage This indicates the % Control On = Number of This KPI corresponds to OPM-CLPM-
Control On percentage of the samples samples WHERE ((Auto or either Auto or Cascade Performance
of good quality within the Cascade control mode) AND control modes. The control
data window for which the samples of good quality)/ loops are meant to be
control mode is enabled. Number of samples of good actively, automatically,
quality continuously controlling.

116 © 2020 General Electric Company

 KPI Description Formula Interpretation Source KPI
Analytic

Percentage This indicates the average Percentage Controller Output = The MV Average should be OPM-CLPM-
Controller MV value (expressed as a SUM(MV for samples of good around 50% most of the Performance
Output %) for data within the data quality in window) / Number of time.
window that is of good samples of good quality
If the MV is permanently
quality.
stuck at a minimum of 0%,
or a maximum of 100%, it
shows that the MV cannot
control the PV to the SP. One
of the following is likely the
cause of this:
• The sensor, or FCE/
Actuator is broken.
• The SP is set at an
unrealistic/unreachable
value.
• The FCE/Actuator is
incorrectly sized.
• An unsuitable MV with
little or no leverage is
currently selected to
control the PV.

For example, if the valve/

pump acting as the FCE is
too small, the MV Average
will be closer to 80/90%.
Conversely, if the valve/
pump is too big, then the MV
Average will be closer to
10/20% most of time.

Percentage This indicates the % Good Quality = Number of This KPI gives you an idea of OPM-CLPM-
Good Quality percentage of data within samples of good quality / how trustworthy the data Performance
the data window that is of Number of samples within the data window is.
good quality. The higher the percentage
good quality, the more
trustworthy the data is.
Other KPIs should be
understood with this
measure in mind, bearing in
mind that control loop KPI
calculations exclude bad
quality data samples.

© 2020 General Electric Company 117

 KPI Description Formula Interpretation Source KPI
Analytic

Percentage This indicates the total PercentageLimitsExceeded = This gives a broad picture of OPM-CLPM-
Limits percentage of time that (Total number of samples the performance of the Performance
Exceeded the system is in Auto, WHERE upper or lower limits system. This should be low.
Cascade, or Manual modes exceeded in Auto, Cascade, or
If this KPI is large, then the
where the upper or lower Manual modes / Total duration in
upper or lower control limits
limits for those modes Auto, Cascade, or Manual
are being exceeded more
have been exceeded. modes) * 100
than is healthy.

Examine the individual KPIs

for each mode indicating
that limits have been
exceeded for a more precise
indicator of what is
contributing to the behavior
of the system.

Percentage This indicates the % Manual = Number of samples A loop is typically set to OPM-CLPM-
Manual percentage of the samples WHERE(Manual control mode Manual control mode if the Performance
of good quality within the AND samples of good quality) / Operator wants the ability to
data window for which the Number of samples of good manually set MV values. For
control mode used was quality example, the plant may be in
Manual. abnormal operating
conditions, but not
completely shutdown. For
Manual control mode, the
MV is set manually by the
operator to make control
changes. When a measured
variable operates out of the
limits, an operator
personally implements a
control change to restore
the system to within
operating limits. Ideally, the
control loop should be in
Manual mode as little as
possible.

118 © 2020 General Electric Company

 KPI Description Formula Interpretation Source KPI
Analytic

Percentage This indicates the % Manual LL Exceeded = Number If the PV violates the lower OPM-CLPM-
Manual LL percentage of the samples of samples WHERE (Manual control limits while the loop Performance
Exceeded of good quality within the control mode AND PV < Lower is in Manual mode, this is
data window for which the SP Offset AND samples of good undesirable and it means
control mode was set to quality) / Number of samples of that the operator is not
Manual and the PV was good quality successfully controlling the
below the Lower SP Offset. loop to remain within
acceptable control limits. In
such a case, the loop should
rather be set to Auto control
mode so it can be
automatically controlled to
remain within acceptable
control limits around a
chosen SP. If the loop is
working correctly, Auto
control will generally be
better than Manual control,
unless the loop has tuning,
design, or equipment
problems.

Percentage This indicates the % Manual UL Exceeded = If the PV violates the upper OPM-CLPM-
Manual UL percentage of the samples Number of samples WHERE control limits while the loop Performance
Exceeded of good quality within the (Manual control mode AND PV > is in Manual mode, this is
data window for which the Upper SP Offset AND samples of undesirable and it means
control mode was set to good quality) / Number of that the operator is not
Manual and the PV was samples of good quality successfully controlling the
above the Upper SP Offset. loop to remain within
acceptable control limits. In
such a case, the loop should
rather be set to Auto control
mode so it can be
automatically controlled to
remain within acceptable
control limits around a
chosen SP. If the loop is
working correctly, Auto
control will generally be
better than Manual control,
unless the loop has tuning,
design, or equipment
problems.

© 2020 General Electric Company 119

 KPI Description Formula Interpretation Source KPI
Analytic

Percentage MV This indicates the Percentage MV Saturation = This gives an indication of OPM-CLPM-
Saturation percentage of samples of (Number of samples of good how much MV is at one Performance
good quality where the MV quality WHERE (MV is 0% or extreme or the other. This
is saturated. The MV is 100%) / Number of samples of should be low.
considered saturated good quality) * 100
If this KPI is large, one of the
when it is at 0% or at
following is likely the cause
100%.
of this:

• The sensor, or FCE/

Actuator is broken.
• The SP is set at an
unrealistic/unreachable
value.
• The FCE/Actuator is
incorrectly sized.

Percentage Not This indicates the Percentage Not Utilized = This gives an indication of OPM-CLPM-
Utilized percentage of samples of (Number of samples of good how much MV is not being Performance
good quality where the quality WHERE (MV is 0% or MV used effectively. This should
loop is not being utilized. is 100% or control mode is be low.
The loop is considered not Manual) / Number of samples of
If this KPI is large, one of the
utilized when the MV is at good quality) * 100
following is likely the cause
0% or at 100% (saturated),
of this:
or when the control mode
is Manual. • The sensor, or FCE/
Actuator is broken.
• The SP is set at an
unrealistic/unreachable
value.
• The FCE/Actuator is
incorrectly sized.
• The loop is in Manual
mode for other reasons.

120 © 2020 General Electric Company

 KPI Description Formula Interpretation Source KPI
Analytic

Percentage This indicates the % Shutdown = Number of In Shutdown control mode, OPM-CLPM-
Shutdown percentage of the samples samples WHERE (Shutdown the loop is switched off Performance
of good quality within the control mode AND samples of because the plant is in an
data window for which the good quality)/ Number of abnormal operating mode.
control mode used was samples of good quality For example, the plant may
Shutdown. be shut down for
maintenance.

Shutdown mode means that

the loop is off, and the plant
is off or not in a condition to
be automatically controlled.
The CLPM solution excludes
all data from loop KPI
calculations during this time.

During CLPM deployment, an

analytic is created that
decides based on
operational rules, when each
control loop should be
considered in Shutdown
mode.

Percentage This indicates the % Shutdown LL Exceeded = It is OK if a loop violates OPM-CLPM-

Shutdown LL percentage of the samples Number of samples WHERE limits during Shutdown Performance
Exceeded of good quality within the (Shutdown control mode AND PV mode, since the loop is off
data window for which the < Lower SP Offset AND samples and not meant to be
control mode was set to of good quality) / Number of controlling, and data is not
Shutdown and the PV was samples of good quality used in KPI calculations.
below the Lower SP Offset.

Percentage This indicates the % Shutdown UL Exceeded = It is OK if a loop violates OPM-CLPM-

Shutdown UL percentage of the samples Number of samples WHERE limits during Shutdown Performance
Exceeded of good quality within the (Shutdown control mode AND PV mode, since the loop is off
data window for which the > Upper SP Offset AND samples and not meant to be
control mode was set to of good quality) / Number of controlling, and data is not
Shutdown and the PV was samples of good quality used in KPI calculations.
above the Upper SP Offset.

PV Error This gives the average of PV Error Absolute Average = The closer this KPI is to zero OPM-CLPM-
Absolute the absolute values of all SUM(|PV Error Samples of good the better. An absolute Performance_E
Average PV Error (PV - SP) samples quality in window|) / Number of average of zero indicates xt
within the data window PV Error samples of good quality that the loop is controlling
that are of good quality. exactly on the SP. The KPI
gives an indication of how
far away the PV is on
average from the SP over the
reporting period.

© 2020 General Electric Company 121

 KPI Description Formula Interpretation Source KPI
Analytic

PV Error This gives the average of PV Error Average = SUM (PV The closer this KPI is to zero OPM-CLPM-
Average all PV Error (PV - SP) Error Samples of good quality in the better. An average of Performance_E
samples within the data window) / Number of PV Error zero indicates that the loop xt
window that are of good samples of good quality is on average controlling
quality. centered around the SP.

PV Error This gives the standard PV Error Standard Deviation = This KPI gives an indication OPM-CLPM-
Standard deviation of all PV Error SQUARE ROOT (SUM(PV Error - of the amount of variation in Performance_E
Deviation (PV - SP) samples of good PV Error Average) ^2 for each PV PV Error sample values of xt
quality. Error sample of good quality) / good quality over the
Number of PV Error samples of reporting period.
good quality)

PV Variability This gives a ratio which is a PV Variability = (PV Standard This KPI gives an indication OPM-CLPM-
measure of the standard Deviation * 200) / |PV Average| of how spread out or closely Performance_E
deviation relative to the PV clustered the PV data xt
sample average, for PV samples are, expressed as a
samples of good quality percentage. It is a measure
within the reporting of relative PV variability. This
period. is also sometimes referred
to as the coefficient of
variation.

PV Variance This gives the variance of PV Variance = SUM((PV Sample - This KPI gives an indication OPM-CLPM-
all PV samples of good PV Average) ^2 for each PV of the amount of variation in Performance_E
quality over the reporting sample of good quality) / PV sample values of good xt
period. Number of PV samples of good quality over the reporting
quality period.

Reversal This calculates the sum of Reversal Amplitude The closer this KPI is to zero OPM-CLPM-
Amplitude all the changes in the MV the better. A low value likely Performance_E
= SUM(|MV_ROC|) / Reversal
relative to the number of means that the system is xt
Count
times the MV changes responsive and smaller
direction. Only data of = SUM(|MV_ROC|) / COUNT(MV changes are made to keep
good quality is considered direction change) the system on track. A large
for the calculation of this reversal amplitude likely
value. Bad quality data is indicates that the system is
ignored. unresponsive and hence
requires larger changes in
the MV to keep it on track.

Reversal Count This calculates the number Reversal Count = COUNT(MV The closer this KPI is to zero OPM-CLPM-
of times that the MV direction change) the better. A large value for Performance_E
changes direction over the this KPI indicates that MV xt
reporting period. changes tend to overshoot
the mark and require MV
changes in the opposite
direction to compensate.

122 © 2020 General Electric Company

 KPI Description Formula Interpretation Source KPI
Analytic

Statistics This is the time period in Statistics Period = End time - The time period in seconds OPM-CLPM-
Period seconds over which the Start time, expressed in seconds over which the KPIs have Performance_E
KPIs have been calculated. been calculated. xt

Total PIDF This indicates the total Total PIDF Changes = COUNT (All Different settings will OPM-CLPM-
Changes number of times that one samples where P, I, D, or F values change this KPI and can Config_Change
of the P, I, D, or F have changed) explain sudden changes in
(Proportional, Integral, control loop performance.
Derivative, or Filter) values For example, there may have
has changed inside the been some change in the
window of data. loop tuning settings and that
has affected loop
performance.

© 2020 General Electric Company 123

Non-Aggregate KPIs
These non-aggregate KPIs are given for the data window.

124 © 2020 General Electric Company

 KPI Description Formula Interpretation Source KPI
Analytic

Lower SP This returns a dataset that For each sample in window, add This returns a dataset that is OPM-CLPM-
Threshold is the set of all differences a datapoint, IFF it differs from the set of all differences PV_Statistics
between the SP and the the previous datapoint: between the SP and the
Lower SP Offset for each Lower SP Offset for each
SP-LowerSPOffset
point in the data window point in the data window
where that difference has where quality is where that difference has
changed from the previous WORST(SP,LowerSPOffset). changed from the previous
difference. difference.

The quality of each This represents the

datapoint in the dataset is acceptable lower control
the quality of the SP at limit within which the PV
that point in time. should be controlled.

This is used to produce a

step graph indicating
changes in the Lower SP
Threshold.

PV Error This returns a dataset that For each sample in window, add This returns a dataset that is OPM-CLPM-
is the set of all differences a datapoint: PV-SP, where quality the set of all differences PV_Statistics
between the PV and SP for is WORST(PV,SP). between the PV and SP for
each point in the data each point in the data
window. window.

The quality of each Ideally, the PV Error should

datapoint in the dataset is be zero.
the worst quality of the PV
A large positive or negative
and SP at that point in
PV Error means that the
time.
control loop is not
This is used to produce the controlling the PV to the SP
error histogram. effectively.

Upper SP This returns a dataset that For each sample in window, add This returns a dataset that is OPM-CLPM-
Threshold is the set of all sums of the a datapoint, IFF it differs from the set of all sums of the SP PV_Statistics
SP and the Upper SP Offset the previous datapoint: and the Upper SP Offset for
for each point in the data each point in the data
SP+UpperSPOffset
window where that sum window where that
has changed from the where quality is difference has changed from
previous sum. WORST(SP,UpperSPOffset). the previous difference.

The quality of each This represents the

datapoint in the dataset is acceptable upper control
the quality of the SP at limit within which the PV
that point in time. should be controlled.

This is used to produce a

step graph indicating
changes in the Upper SP
Threshold.

© 2020 General Electric Company 125

 Chapter

8
Release Notes
Topics:

• Control Loop Performance

Monitoring Release Notes Q2
2019
• Control Loop Performance
Monitoring Release Notes Q1
2019

126 © 2020 General Electric Company

Control Loop Performance Monitoring Release Notes Q2 2019
This topic provides a list of product changes released for this module on the dates listed below.

Release Date: June 3, 2019

Table 4: Enhancements and New Features

The following enhancements and new features have been added.

Description Tracking ID

The Loop Dashboard now has the capability to display information till the latest available data US340762
point for the specified dates.

In the Fleet Dashboard, you can now access the historical data by specifying a date range of up to US325581
30 days.

Control Loop Performance Monitoring Release Notes Q1 2019

This topic provides a list of product changes released for this module on the dates listed below.

Release Date: March 29, 2019

Table 5: Enhancements and New Features

The following enhancements and new features have been added.

Note Tracking ID

Additional alerting analytics are now available for scheduling. Once configured, these alerts can F41291
notify you of the following situations:

• Poor PV sensor data.

• Potential problems with loop tuning or design.
• Poor MV sensor data.

Consult the Deploy CLPM Analytics section of this documentation for more information.

Enhancements have been made to the Control Loop Performance Table in the Fleet Dashboard. F41287

• Advanced filters are now available, including a global filter, multiple-level filters, text filters,
and conditional parameters to filter data in the table.
• Updates have also been made to improve responsiveness when loading the table.

The Diagnostic Alerts Filtering smart filter has been added to the Fleet Dashboard in this release. F36574
This smart filter is a visual tool to identify faulty control loops.

Table 6: Known Issues

The following issues are unresolved in this release.

© 2020 General Electric Company 127

 Note Tracking ID

Not all PIDF configuration changes are currently detected by the Configuration Change Detection DE91981
analytic.

For interpolated data, APM Timeseries returns bad quality data until the first sample of good DE95724
quality in the data window is encountered.

Table 7: Limitations

The following limitations apply to this release.

Note Tracking ID

Only the Chrome 43 or higher browser is supported for accessing the Loop and Fleet dashboards. NA

CLPM has been benchmarked for deployment over a maximum of 200 assets. NA

All analytic inputs are read as interpolated data. Ensure that your raw data in Predix Timeseries is NA
sampled at a lower frequency than the data interval of your deployed analytics. Results for data
recorded at higher frequencies may not be statistically accurate representations of the
performance of the control loop in question.

All CLPM analytics must currently be configured to use recurrent scheduling and interpolated NA
sample data. Streaming scheduling and raw sample data options are not currently available.

128 © 2020 General Electric Company