North America Custody Transfer Measurement Conference

April 14 -17 2015

ULTRASONIC FLOW METER COMPLIANCE WITH API 14.10

MEASUREMENT OF FLOW TO FLARES

Wil Camara, SICK Maihak USA

Ekkehard Riedel and Sven Holzbächer, SICK Germany

1 ABSTRACT

Ultrasonic flow meters have been used for over 30 years to measure flow to flare stacks. Due to changes
in plant construction philosophies to more compact building, environmental regulations, and an increased
offshore activity the demands to these measurements have changed during recent years. New
regulations implemented in Europe and North America requires the measurement of fuel and flare gas to
calculate CO2 and VOC emissions.

Compact and more optimized production facilities have higher capacities and therefore result in higher
maximum flow conditions over the flare. Flow rates of more than 100 m/s have been recorded during
shut down conditions. Waste or flare gas can vary greatly in its composition and abrupt flow condition
changes should be expected. Furthermore the low flow conditions became more importance for process
control and optimization and for identification of valve leakage in normal operational conditions. Minimum
flow rates demands high efficient transducer technology and advanced signal processing algorithms to
ensure exact and high resolution time measurement. High flow velocities will introduce soaring noise
within the flowing gas and at the ultrasonic probe itself. Even drift of the “ultrasonic beam” has to be
considered at high gas velocities.

With all of these considerations, the main guidance for flare flow installation is API 14.10. This paper
reviews API 14.10 and compliance with the design requirements established relating to flare flow and the
SICK ultrasonic mass flow meter FLOWSIC100 Flare.

2 INTRODUCTION

API 14.10 is a standard which addresses flare flow meters and includes discussion of the following
criteria.

 Application Considerations.

 Selection Criteria for Instrumentation.

 Installation Considerations

 Calibration.

 Operations

 Uncertainty

For the purposes of this paper we examine application and installation considerations, calibration, and
uncertainty. The other topics are either not applicable or are well covered by the standard.
1
 North America Custody Transfer Measurement Conference
April 14 -17 2015

3 APPLICATION CONSIDERATIONS (Section 4 of API 14.10)

API 14.10 section 4.1 recommends that Error curve, 6" CrossDuct
Manufacturers be able to provide test 3
reports quantifying the effects of various
2
influences parameters on their devices to
substantiate their performance specification 1

Dev. [%]
on request. Reports should be based on 0
flow loop data from a recognized flow meter
test facility. This data can be provided by -1

SICK and this is a representative sample. -2

-3
0 10 20 30 40 50 60 70 80 90 100 110 120 130
VOG [m/s]
Volumetric Flow Rate. 4.2.2.2.2

Ultrasonic flow meters measure velocity and calculate volumetric flow by multiplying the velocity by the
cross sectional area of the pipe, I.E. ACFM. Temperature and Pressure are provided as analog inputs to
calculate Standardized or Normalized flow, I.E. SCFM, Molecular Weight, and Mass Flow.

Meter Sizing 4.4

Section 4.6.1 suggests that meters which have 1000:1 or better turndown should have the capability of
multiple analog outputs of at least one low flow and one high flow. This is possible with the Sick flow
meter as you may add outputs in groups of two via Output modules in the MCU.

For Sick ultrasonic flow meters, meter sizing consists of a combination of pipe size and stream
composition considerations. High levels of attenuating gasses such as CO2 or H2S may reduce the
upper velocity and measuring distance between sensors but in general piping should be size to keep the
velocity between 0.98 and 394 feet per second for cross duct applications and 0.1 and 295 feet per
second for single port probe versions of the flow meter. Providing a full accounting of possible stream
compositions, pressure and temperature conditions and flow rates will ensure a properly sized meter.

SICK flare flowmeter enables a cross- duct measurement also in very large pipes. A fully diametric
measurement path without sensor insertion in the flow can be realized up to 72” (depending on gas
attenuation)

Figure: Easy installation of I/O modules.

2
 North America Custody Transfer Measurement Conference
April 14 -17 2015

Below is an example of a typical meter repeatability for an

Ultrasonic flow meter. Calculations can be provided for
any given application. Consideration of the repeatability
should be given before deciding to utilize two or more
outputs over a large range. The maximum turndown of a
Sick Ultrasonic flow meter is 4000:1.

Entrained Liquids and Fouling 4.6.3

API 14.10 suggests that Ultrasonic flow meters have a low sensitivity to entrained liquids and to fouling
compared to other technologies, ultrasonic flow meters have a relatively HIGH tolerance to fouling.
Pollution/Contamination of the ultrasonic sensors can be monitored by diagnostic functions unless the
process will have sticky tar like liquids in the line. The perceived limitation is overcome with the adaptation
of a diagnostic circuit for automatic gain control on the sensors. If the transducers begin to foul drastically,
the gain on the transducers will increase to overcome the effects of the fouling. Once the gain value has
reached 70% of max, the MCUP will alarm with a maintenance fault. The meter will continue to operate
and read accurately, and the fault will alert the operator that the sensors need cleaning.

Diagnostic Software 4.6.4

SICK’s software provides with the

meters the documentation
parameters to detect fouling and
other conditions affecting the
meter accuracy.

Figure: Diagnostic report.

3
 North America Custody Transfer Measurement Conference
April 14 -17 2015

Commissioning and Initial Field Calibration. 4.6.5

A detailed commissioning procedure should be available in

the instruction manual provided with the instrument. Initial
field testing is part of this commissioning and printouts are
provided from the diagnostic software that will document the
initial parameters and values relative to the calibration and
proper operation of the meter. A meter check cycle run can
be initiated to verify proper calibration in the field.

Figure: Zero Flow Test Box

4 INSTALLATION CONSIDERATIONS

Safety 4.2.1

Appropriate access should be provided for sensor maintenance. Remote flow computer or electronics
may be located up to 3000 feet from the sensors for easy and safe access. Since the remote electronics
communicate to the sensors via digital communications, all sensor parameter and diagnostics are
available from the Multi Control Unit Processing MCUP.

Sick transducers are certified for Class 1, Division 1 and Class 1, Division 2, Groups A-D hazardous
locations and the remote electronics (Multi Control Unit Processing MCUP) are also certified for
hazardous locations.

When mounted near the flare flame, such as on a flare boom on an offshore platform, a radiant heat
shield may be required to block radiant heat from the flame to the sensor electronics package.

The performance of the MCUP allows connecting up to three independent flow meters. The MCUP
supports a complete different measuring value settings, parameter settings and I/O for each
measurement point. Further customer requirements like redundant metering, pipe incremental size or two
path measurements can be realized to save costs.

Figure: Use of MCUP with three independent measurements in one spoolpiece for three customers.

4
 North America Custody Transfer Measurement Conference
April 14 -17 2015

Piping Runs 4.2.2.1

Section 4.2.2.1 of API 14-10 recommends 20

upstream and 10 downstream piping runs for flare
flow meters. The sections also states that meter
runs minimums can vary depending on the
technology and piping configurations. Sick
provides expertise supports for installations 10
upstream inlet pipe and 5 downstream outlet pipe.

More complex isometric configurations can require

20 upstream and 10 downstream pipe runs [1].The
diagram below shows the minimal piping
requirements for Sick Flare Flow Meters:

Ultrasonic Flow Meters 4.7.4

Spool and Hot tap versions are available; units are offered as
extractable under pressure; fouling is identified through use of
monitoring automatic gain on the sensors and fouling is identified
before loss of signal; a proprietary algorithm is incorporated to infer
molecular weight at 2% accuracy from 2 to 120. SICK ultrasonic flow
meters offer all of the recommendations in this section.

The sections makes note that care must be taken to install insertion
units at the proper depth and angle to ensure the proper path length.
SICK supplies the installation tools to make this a reality, including a
nozzle installation tool and electronic spread sheet to calculate the
precise nozzle angle and path length after installation. This takes
installation error out of the uncertainty for the meter installation.

Figure: EXCEL tool for easy calculation of path angle and path length
[1] (SICK scope of supply).

Location of secondary Devices 4.8.1

The temperature transmitter is suggested to be 3 pipe

diameters downstream of the center line of the meter. The
pressure transmitter should be 1 pipe diameter downstream
or inline with the downstream ultrasonic sensor.

Figure: Required installation setup

5
 North America Custody Transfer Measurement Conference
April 14 -17 2015

Maintenance Considerations 4.8.1.

While all of the considerations are valid, provide 3 ft of clearance on all sides for removal of the sensors.

Record Keeping. 4.11

See diagnostics software above. A print out of the parameterization and diagnostic values as well as a
downloadable file can be saved for record keeping purposes.

Factory Calibrations/Verifications (Section 5 of API 14.10)

Flowmeter 5.1

For type testing, the primary element performance has been established using flow loop testing, and all
primary elements are shown to have been manufactured to within the manufacturer’s tolerance for the
flow meter that was type tested. Sick flare flow meters are both individually tested for proper zero and
speed of sound calculation and type tested over the capable velocity range of the flow meter.

The SICK flow test facilities are approved and traceable to national standards.

Figure: SICK flow test facilities in Dresden Germany

SICK has the capability for calibration in their test

benches near Dresden, Germany.

Below it’s a typical meter repeatability for an Ultrasonic

flare meter. The figure show an error curve “as left” of a
FLOWSIC100 EX-S. The repeatability is lower than
0.25% [1].

Figure: Meter repeatability of a 16”meter of normalized

flow [1]

6
 North America Custody Transfer Measurement Conference
April 14 -17 2015

Zero Flow Check

The shown measuring results as regards accuracy and

repeatability are total results of the measurement set-up,
which includes uncertainties caused in the test bench, device
under test, temperature, pressure and, more than any other,
flow rate fluctuations during the measurements.

This is why the analysis of a zero flow measurement of the

gas meter is particularly revealing, because that test almost
exclusively shows the uncertainty of the device itself. SICK
finalize each meter fabrication with a zero flow check.

Additional the speed of sound is recorded as well. Due to the

dependency of speed of sound and temperature it gives a
good expression whether thermal influences are present or
not. The very impressive stability of the recorded data shows
that there is no thermal convection detectable.

Figure 1: Measured velocity of gas inside the zero flow box [1]

All measuring data and results are documented in a report

which is part of the delivery documentation for FLOWSIC100
Flare. The report contains meter identification data like serial
number and sensor type. Based on the ambient medium
conditions humidity, pressure and temperature a theoretical
speed of sound is calculated. For that a algorithm suggested
by Owen Cramer is used [3].

Deviation and limits which was measured and calculated for

the device are shown in tables and diagrams as easy
overview about the meter performance is possible.

Figure 2: Zero Flow Report [1]

Periodic Verification (Section 7 of API 14.10)

Periodic Verification Method – Flowmeter 7.2

A detailed periodic verification procedure is available and can

be performed without removal of the primary sensors.
Alternatively, the sensors can be removed for Zero and speed
of sound verification with a Zero Flow Box.

Documentation printouts are provided from the supplied

software for both as found and as left values. For periodic

7
 North America Custody Transfer Measurement Conference
April 14 -17 2015

testing inline, the recommended procedure, see the “Check Cycle” routine described in the Instruction
Manual.

Zero Flow box

The meter check cycle satisfies all manufacturers’ requirements for annual calibration without actual
removal of the sensor from the line. Recognizing the industry trend to offer a secondary calibration check
for ultrasonic flow meters, a zero flow box is offered to assist in testing the zero and speed of sound.

In a specially designed measuring chamber, the velocity of gas and the velocity of sound are determined
at a flow rate of 0 m/s. The special chamber is designed such that flow rate fluctuations and effects of
temperature have been almost fully eliminated. This is why the test must yield a gas velocity of 0 m/s and
the theoretically calculated velocity of sound in the chamber for each properly working measuring device.

Figure: Using the zero flow box

In addition, based on the recorded measured velocity values, it is possible to check resolution and
general measuring signal noise of the device. The test is successfully finalized if the recorded data within
the specified limits. All the data, diagnostics and conclusions are stored, presented and automatically
generated in an accurate EXCEL report.

Additionally, SICK offers training for end users as suggested in section 4.6.6.

8
 North America Custody Transfer Measurement Conference
April 14 -17 2015

Performance Testing Protocol (Section 9 of API 14.10)

Performance Testing Protocol Requirements – 9.2

While performance testing is not required, a spool piece type for flow meter should be desired, a spool
piece can be sent to a third party lab for performance verification. Insertion sensors can be placed in a

flow loop with air or other gasses and verified as well; however the user should understand that an
insertion meter without a spool may be subject to installation uncertainties that cannot be accounted for in
a factory flow test. Those uncertainties can be accounted for on installation by taking proper
measurements and calculating

Uncertainty (Section 9 of API 14.10)

Uncertainty Analysis – 10.3

The analysis assumes correct installation of the meter in

accordance with the recommendations of the manufacturer.
The analysis differentiates between three calibration conditions:

 Devices installed in the field without meter body or

calibration
 Devices with meter body and geometric measurement
(“dry calibration”)
 Devices with meter body and flow calibration

Figure: Uncertainty analyse of FLOWSIC100

The analyses are based on the method "Guide to the

Expression of Uncertainty in Measurement" (GUM) according
to ISO/IEC Guide 98-3:2008, for each of the FLOWSIC100
Flare meter types.

The uncertainty analysis has been calculated for the following gas flow velocity values, as specified in the
API Standards: 0.3 m/s; 1.0 m/s; 10 m/s; 90 m/s

The example cited is based on a meter in a pipeline with a nominal width of 16”, and can be transferred
for other widths. The analysis is valid for a device with one cross-duct path. The values measured are: the
volume flow rate during operation, the normalized flow rate and the mass flow rate.
The uncertainty analysis thereby fulfils the requirements for the following standards and regulations:

 Manual of Petroleum Measurement Standards, Chapter 14 - Natural Gas Fluids Measurement,

Section 10, July 2007
 NORSOK Standard I-104

9
 North America Custody Transfer Measurement Conference
April 14 -17 2015

The diagram below shows one result for the measuring value “Normalized Flowrate”. The calculation
always consider the given velocity points and the different type of installation. It will be clear that the best
uncertainty will be achievable by using of a calibrated meter which includes a meter body certainly.

Figure: Uncertainty result for measurement

value “Normalized Flowrate” [1].

5 REFERENCES

[1] SICK Engineering GmbH – Bergener Ring 27 D-01458 Ottendorf-Okrilla, Germany

[2] Manual of Petroleum Measurement Standards Chapter 14 – Natural Gas Fluids Measurement,
Section 10 Measurement of Flow to Flares. First edition, July 2007

[3] The variation of the specific heat ratio and the speed of sound in air with temperature, pressure
and humidity, and CO2 concentration. Owen Cream, National Metrology Program

10