QIP course on

‘Smart Grid Technology’

Distribution Management System

Presented by:
Dr. Abheejeet Mohapatra
Assistant Professor
Department of Electrical Engineering
IIT Kanpur
[email protected]
Typical Active Distribution Network
Components of Distribution Network Typical configuration of distribution network
 AC and DC Loads
Radial
 Renewable Energy Sources
Looped
 Energy Storage Devices
Networked Primary and secondary
 Protective Devices

May 9, 2019 2
Need of Distribution Management System (DMS)
Demands of the customers should be met economically and reliably.
 Economy : Low operational cost, low losses, best customer services
 Reliability: Power quality and availability
In other words, a balance between supply and demand is maintained.

Requirement of Distribution System

To have a Decision Support System to assist the control

room and field operating personnel with the
monitoring and control of the electric distribution
system in an optimal manner while improving safety
and asset protection

Scope of DMS
May 9, 2019 3
Basic DMS Functions
DSSE Distribution System State Estimation
NCA Network Connectivity Analysis

LFA Load Flow Analysis

VVC Volt-Var Control

ILSR Intelligent Load Shedding and Restoration

FMSR Fault Management & System Restoration

LBFR Load Balancing via Feeder Reconfiguration

FDN Forecasting in Distribution Networks

May 9, 2019 4
Advanced DMS Functions
• Real-time Simulations • Medium and Long-term
• Off-line Simulations Forecasting
• What-if Analysis • Network Automation
• Historical information • Network Reinforcement
Training • Optimal Device sizing and
Placement
Planning
• DSSE
• LFA Optimize • Volt/Var Optimization
• Fault Calculation • Network Reconfiguration
• Reliability Analysis • Short-term Forecasting
• Contingency Analysis Operations • Demand Response
• Device Capability, etc. • Distributed Energy
Analyze management

Monitoring
• Telemetry • Fault Management
• Alarming • Switch Management
• Reporting • Load Shedding
• Trending • Restoration, etc.

May 9, 2019 5
Distribution System State Estimation (DSSE)
Core driver of other DMS functions

Calculates system states (voltages and phase angle) based on real time measurements

DSSE faces challenges such as

• unobservable system due to lack of sufficient measurements
• unbalanced/ incorrect network topology
• exposure to cyber attacks and bad data
• delayed or missing measurements

Challenges in hybrid measurement devices - different standards and data integration

Need for real-time system to simulate field measurements

May 9, 2019 6
DSSE Contd.

Differences in DSSE, as compared to transmission system state estimation are

• High R/X ratio due to shorter line lengths
• Unobservable system due to lack of measurement sources
• Complex measurement functions due to unbalanced phases
• Missing lines and zero injections
• Scalability
• Poor convergence due to high condition number of gain matrix

May 9, 2019 7
DSSE Contd.
 Typical techniques – least square estimation, Kalman filter
 Obtained states significantly depend on
• System topology obtained from NCA
• Bad data detection, identification and correction process
• Availability and time stamping of obtained measurements from various meters
 Large number of bad data affects the computational time of DSSE
 Cyber attacks pose additional challenge in DSSE
 Measurements are hybrid in nature – micro PMUs, RTUs, IEDs
 Measurements arrive at different times and hence, measurement utilization and
synchronization in DSSE is a challenge
 Need for statistical techniques for robust DSSE

May 9, 2019 8
Network Connectivity Analysis
 Evaluates connectivity between various network elements

 Is done in coherency/ parallel to DSSE

 Operates with the help of metered data in real-time

• Bus connectivity (Live/ dead status)

• Feeder connectivity
• Abnormally energised sections
• Grounded network sections
• Loop detection
• Status of Circuit Breakers/ Isolators
May 9, 2019 9
Load Flow Analysis (LFA)
 LFA provides bus voltage magnitudes and associated phase angles for specified bus
injections and network topology and parameters
 DSSE is a real time system monitoring exercise
 LFA is an offline exercise used for analysing the impact of possible future scenarios on
the present system
 Compared to transmission system, distribution system LFA has the following differences
• High R/X ratio
• Matrix ill-conditioning – Newton’s based approach fails
• Mostly three phase unbalanced analysis
 Several direct/ indirect, recursive/ non-recursive non-Newton based techniques for
distribution system LFA

May 9, 2019 10
LFA Contd.

Distribution
System Load
Flow Analysis

Forward and Modified Miscellaneous

Compensation Implicit ZBus
Backward sweep Newton/Newton Power Flow
methods Gauss Method
Algorithm like methods Methods

Current Power Admittance Direct method

Loop impedence
summation summation Summation (BIBC/BCBV
matrix method
methods method method matrix method)

May 9, 2019 11
Volt - VAr Control (VVC)
• Active loss minimization • Maintain acceptable node voltages
• Load demand reduction
OBJECTIVE • Savings under all loading conditions in offline/
• A weighted sum approach real time across a day

• Requires coordinated control of

• Capacitor banks voltage regulation and reactive power
CONTROLS • OLTCs, Voltage regulators
• Smart Inverters control systems

• Determines the best set of control

• No. of tap and capacitor bank
operations
actions for all voltage regulating
CONSTRAINTS • Voltage magnitude across the feeder devices and VAR control devices to
• Inverter operational limits
• Storage state of the charge
achieve the objectives without
violating the fundamental operating
constraints
May 9, 2019 12
VVC Contd.
Conventional
• Capacitor Banks
Distribution Loss • OLTC
Reduction Methods

• Load Profile
Smart Meters • Real-time V/I/PF Data
Adaptive
Real-Time
• Multi Level: Offline + Real time
Volt/VAr Optimization CVR and
• Smart Inverters
Volt/VAr
Optimization
Conservation Voltage • Dynamic consumer Voltage Reduction
• Adaptive voltage regulators and OLTCs
Reduction (CVR) • Smart Inverters, peak load reduction

Volt/WATT • To squeeze maximum real power during

availability
Optimization • To reduce the burden on energy storage devices

May 9, 2019 13
VVC – Optimization: Then and Now!
Off-line &
Off-line Real time
control
Real time
More
No real time Less data
data intelligence intelligence
availability

Sub optimal Optimal

Localized Un-
Wide Centrally
visibility Coordinated Visibility coordinated

Independent
functions FAST
and
constraints

May 9, 2019 14
VVC in active distribution networks: How?
Both day-ahead and real time control of devices which can manage reactive power
and instruments which can control voltage magnitudes

For day-ahead control, optimization engine is run to meet the objective(s) by finding
the optimal settings of slow acting devices such as OLTCs, VRs and capacitor banks
based on forecasted loads and renewable resources generation and desirable storage
requirement

For real time control, quick control actions are implemented over and above the
day-ahead controls by exploiting the speed of action of inverters and storage based
on real time change in loads as well as generation of renewable resources (or any
other system uncertainties!)

May 9, 2019 15
Intelligent Load Shedding & Restoration (ILSR)
Intentional Load • This is an operator initiated load shedding sequence, which will trip selective feeders
Shedding based on the date and time of day.

• This is an automatic load shedding sequence, where if the grid frequency drops
Under Frequency
below a certain threshold, then a group of feeders need to be tripped
Load Shedding immediately.
Transformer
• This is also an automatic load shedding sequence, which is initiated when the total
Overload Load
power flow through a transformer exceeds a certain threshold.
Shedding

Distributed Load • This is required when the overall drawl from the grid for the utility is above the
Reduction scheduled/granted drawl.

Function Performed
by RTU and/or • This is required for telemetry and control.
SCADA
May 9, 2019 16
Demand Side Management (DSM)
World Bank definition:
Systematic utility and government activities designed to change the amount and/or timing of the customer’s use of
electricity for the collective benefit of the society, the utility and its consumers.

Energy supplied by
Energy is stored storage to meet excess Peak
loads
Power

Base Load

High demand
Low demand

time
May 9, 2019 17
 DSM Contd.

Reduce/shift Demand Price

demand reduction reduction

Reduce/shift demands
Demand

during peaks
Generation

Electricity Price
capacity limit
Price
Reduction

Demand
Reduction

time Quantity
May 9, 2019 18
Benefits of DSM
DSM focuses on efficient usage and allocation of available resources rather than going for
additional capacity installation
Customer
• Benefits Utility Benefits Societal Benefits
Satisfy electricity Reduce environmental
demands Lower cost of service degradation

Reduce / stabilize Improve operating Conserve resources

costs or electricity bill efficiency,
Flexibility
Maintain/improve Improve customer Protect global
lifestyle and service environment
productivity

May 9, 2019 19
DSM strategies
Strategy Aim Impact on Energy Impact on Peak
Demand Demand
Energy Reduce the overall energy
Conservation demand (energy
conservation)
Peak ‘Clip’ demand at peak load
Clipping periods (load levelling)

Load Shifting to off peak hours

Shifting (load levelling) No change

Flexible Induce change in load as

Reliability per supply (load May reduce
controlling)
Strategic Promotion of applications
Load Growth requiring electricity – May increase
electric vehicles
Valley Increasing load during off
Filling peak hours (load levelling) No change
May 9, 2019 20
Protection in Distribution Networks
 Typically, Overcurrent (OC) relays have been conventionally used for protection of
passive distribution networks
 Load – way and loop – way level – directional OC relay
 In current microgrid framework
• Use of differential relays for unit protection schemes
• Adaptive fault current limiters
• Adaptability to varying system conditions

May 9, 2019 21
Protection in active distribution networks: Challenges
1. Malfunction of relays due to downstream
faults:
• In a downstream fault, utility grid and DER unit currents (Ig
and DER, respectively) contribute to the total fault current.
• Interfaced DER is large, Ig is low because of a higher voltage
contributed by DER at PCC.
• PD1 may not trip because of a lower fault current even
though feeder 1 experiences a higher fault current.

2. Sympathetic tripping:
• PD3 should trip to clear the fault. Figure : The impact of DER units on relay operation.
(a) The malfunction of PDs due to downstream faults
• If DER unit contribution to the fault current is large, PD2 and
may trip in response to high current DER, which would (b) The sympathetic tripping of PD.
disconnect feeder 2 from the utility grid.

May 9, 2019
22
Fault Location, Isolation and Service Restoration (FLISR)
Objective of FLISR

• Distribution automation application.

• Reduce outage time to the end customers.

Problem due to the fault:

• Substation feeder protection normally shuts-down power on the entire

feeder.
• Disruption in the service to many end-user customers, including industrial,
hospitals, commercials and residential loads.

May 9, 2019 23
How FLISR can reduce outage time?

May 9, 2019 24
Steps in FLISR
Fault location (followed by fault detection):
• Triggered by substation protection devices (Intelligent Electronic Device, IED or recloser controller).
• After faulty feeder tripping, the faulty section on the tripped feeder needs to be located
Fault isolation:
• After identifying faulty feeder section, both sides of fault need to be isolated using switches/ reclosers.

Capability Estimation:
• After isolation and before restoration, a capability estimation need to be carried out to determine if
service restoration from a healthy feeder is possible
Service Restoration:
• From capability estimation, it is determined whether complete or partial load of the faulty feeder can
be transfer to healthy feeder
• Accordingly, the service restoration process closes tie-switch and corresponding feeder switches
(which can feed healthy portion of the faulty feeder)
May 9, 2019 25
FLISR Architectures
1. Centralized FLISR (C-FLISR)
• Each relay communicate to control center directly
• high bandwidth communication network

2. De-Centralized FLISR (DC-FLISR)

• DC-FLISR system deployed at substation level using a single or a redundant automation device installed in each
substation.
• Remote I/O modules installed at each switch/recloser locations connected to the distribution substation
automation device over communication network
• faster with lower bandwidth requirements.

3. Distributed FLISR (D-FLISR)

• D-FLISR uses controlled devices at each switch/recloser locations
• These communicate among each other to determine where the fault has occurred and to determine the
appropriate switching actions necessary for the restoration.

 IEC 61850 Generic Object Oriented Substation Events (GOOSE) based peer-to-peer communication technology is a
good fit for such applications.
May 9, 2019 26
Contingency in Microgrid
Upon occurrence of a contingency, microgrids will
experience one of the following situations:

1. Fault:
Disconnection of all the load or partial load

2. Grid-Connected:
After fault isolation, some sections/ microgrids might
be able to import the required power from upstream
network or inject their excess power to it
Fig. Sample system for illustrating different operating modes.
3. Islanded:
Fault isolation might disconnect some microgrids/
sections from the upstream network. In this case,
microgrids may or may not continue supplying loads
in islanded mode

May 9, 2019 27
Data/Methods for fault location
Apparent impedance measurement

 Direct three-phase circuit analysis

Superimposed components

Traveling waves

 Power quality monitoring data

Artificial intelligence/ machine learning approach

May 9, 2019 28
Network Reconfiguration
Network reconfiguration can be defined as altering the topological structure of
feeders by changing the open/closed status of the sectionalizing (normally
closed) and tie (normally open) switches

Normal Condition

Constrains
• Minimization of losses
• Reliability improvement
• Transformer capacity • Load balancing

• Feeder thermal capacity

Faulty Condition
• Voltage drop
• Isolation of faulty line
• Maximum service restoration

May 9, 2019 29
Need for reconfiguration
Need for Depending on the current loading conditions, reconfiguration may become
network necessary in order to eliminate overloads on specific system components
reconfiguration such as transformers or line sections. In this case, it is called load balancing.

For reducing real power losses in the network, reconfiguration can be done.
This is usually referred to as network reconfiguration for loss reduction

To maximize reliability of the system.

During line outage or schedule maintenance of line.

May 9, 2019 30
Process for Offline Reconfiguration

Calculate power
losses, Reliability
Run load flow
for each
• Now run load flow configuration
Permutation for all above
possible • Calculated total power
combination configuration loss and reliability for
• Find all possible above all possible
Perform load configuration such configuration
flow that system is radial
and all load are
• Find total power connected
losses for above
initial state
Read input data
• Read initial status of
switches before
reconfiguration
May 9, 2019 31
Unbalance Minimization In Distribution System
Unbalances in the system increases mostly due to
the following reasons

• Increase in number of single phase loads

• Connection of sources in single phase modes (Solar PV)

and due to the uncertainty associated with them

• Uncertainty due to loads such as Electric Vehicles

• Due to increased harmonics (Power Electronics interfaces)

Voltage drops can be very high due to the unbalances

which is an unwanted scenario

May 9, 2019 32
 Methods to Unbalance Minimization
LOAD BALANCING
• Measuring power consumption at a bus across
the feeders and switch the loads

• Use transfer switches (Eg.: three phase input-

single phase output static transfer switches)

• Individual monitoring and centralized control is

involved.

• Use voltage unbalance indices to control the

settings of centralized controllers. (Eg.: Phase
Voltage Unbalanced Rate (as per IEEE 936-1987,
IEEE 112-1991), VU (ratio of negative sequence
component to positive sequence component) Load Current

• Requires smart meters, central controllers, good

communication network, hardware switch
Load Voltage
May 9, 2019 33
Methods to Unbalance Minimization Contd.
CONTROL OF INVERTER BASED SCHEMES

• Requirement of centralized control, communication

and switching in load balancing can be reduced if
smart inverters can be involved in unbalance
minimization

• Smart inverters which are participating in VVC can

be extended to reduce unbalances too, which can
increase the performance in a dynamic
environment.

• Smart inverters are quick and flexible. This can

reduce the burden on network reconfiguration
scheme which can even be slower in time frame.
An inverter can quickly respond to a
reference change as shown in the figure.
May 9, 2019 34
Forecasting in Distribution Networks
What? Predict the future value of a time varying quantity

Using data from past and present

How? Analyze the trends in data, judgement based on experience, knowledge
and judgement

Why? Crucial for estimation, planning, control

Any parameter of importance which is time varying in nature

Where?
Touching almost every spectrum of human interest

May 9, 2019 35
Distribution Load Forecasting
Estimation for future load by an industry or utility company

Vital for electric industry in the deregulated economy

Essential to establish procurement policies; future fuel requirement; types of fuel
requirements;
Accurate model for electric load forecasting are essential to the operation and
planning of a utility company

Important for energy suppliers, ISOs, financial institutions, and other participants in
energy distribution level

A difficult task; load series is a complex and exhibits several levels of seasonality

Other important exogenous variables such as weather must be considered

May 9, 2019 36
Forecasting periods and accuracy levels
Short term Medium term Long term
Influencing factor Influencing factor Influencing factor
Weather Weather Weather
Events, Holidays, festivals, Growth Rate Growth Rate
TV programs New Customers New Customers
Lifestyle Change
Accuracy

Hours/days Months Years

Benefits Benefits Benefits
Network planning Network planning Capacity/ Investment
Supply/Demand Matching Supply/Demand Matching planning/ Infrastructure
Spot Power Procurement Power Procurement development
Load Shedding Strategy Fuel Mix Decision
May 9, 2019 37
Typical Forecasting Methodologies
Simple Time Series Model

Regression Model

Simpler Day Approach/ Seasonal

Consumption = function(season, weather) + residue
Seasonality can be modelled through ARIMA
Major seasonal component
Time of the day
Day of the week
Month of the year

Machine Learning based Model (ANN, SVM, Fuzzy)

May 9, 2019 38
Load Forecasting Model Development

Collect Historical Load and Load Shedding Data

Prepare Unconstrained Collect Historical Collect Historical Event

Load Data Weather Data Data

Prepare Model Input Select a Model /

Analyse the Data
and Test Data Methodology

Run and Refine the Select the Best Model Fit the Data and Tune
Model and Implement the Model
May 9, 2019 39
Forecasting Challenges

Data Related Challenges

Lack of “good’ data
Selection of optimal data size
Process and Methodology Challenges
Selection of appropriate model
Technology Challenges
Only empirical relation exist between load and known variables
Correlation of load variation to other variables still unknown
“Act of God” Challenge
Cannot be modelled

May 9, 2019 40
Flaws of forecasting models
Statistical process; individual forecast of customer do not add to total forecast

Forecasting done of unconstrained load

A combination of good data, good process, and good model; a good model alone can not give good
accuracy

Accuracy improvement is a gradual process and involve significant human intervention. An

automated process does not provide good accuracy on regular basis

More data does not always lead too good forecast; an optimal size of data set is important

One model does not fit all

May 9, 2019 41
Distribution Automation – Typical Scope

• Centralized Monitoring and Control of Distribution Networks

• Enables implementation of various DMS functions

• Load Assessment and Energy Monitoring

• Integration with Outage Management Systems (OMS)

May 9, 2019 42
Elements of Distribution Automation

May 9, 2019 43
Elements of Distribution Automation
• SCADA System Infrastructure at Operations Control Centre (OCC)

• Remote Terminal Units (RTU) at Grid Substations for Remote Monitoring and Control

• Metering at Grid and Distribution Transformer (DT) Levels

• RMUs and Reclosers

• DT Monitoring Units (DTMU)

• Metering Data Acquisition Software (MDAS) and Data Concentrator Units (MDCU)

• Fault Passage Indicators (FPI) for Fault Reporting

May 9, 2019 44
Communication Protocols in DMS
• At Substation/MG level
• IEC61850, IEC 60870-5-103 (for protection devices)
• DLMS, Modbus (for metering devices)
• IEEE C37.118 (for phasor measurements)

• At control center level

• IEC 60870-5-104 (between RTUs, DCUs and SCADA)
• File-based exchanges (XML, CSV, etc.)

• Inter control center

• IEC 60870-6 aka. TASE.2 (between control centers)
• IEC 60870-5-104 (in peer-to-peer mode)

May 9, 2019 45
Communication Mechanisms in DMS

• Between Field Equipment and substation gateways

• 3G/4G
• LoRA / eq.
• Meshed Radio
• Power-line Communications (PRIME, G3-PLC, etc.)

• Within substation / MG
• Typically wired networks like RS232, RS485, Ethernet
• Legacy systems work on hardwired I/O

• Between substation and control centers

• Ethernet (MPLS, VSAT, leased-line, etc.)
• 3G / 4G public networks

May 9, 2019 46
THANK YOU