Chapter 2

Real Time Control of Power Systems

 Contents
• EMS Functions
• Security Analysis
• Preventive Control/Restorative Control
Constraints in unit Commitment
• Solution Methods of Unit Commitment
• Dynamic Programming and
• Lagrange Relaxation
 1. EMS functions
• What is EMS?
• Need for EMS
• Basic EMS functions
• EMS configurations
 What is EMS?

• EMS stands for Energy management system

• Centralized control of power system needs
– Collection of data to central location
– Processing the data
– Making decision
• Requirements
– Power system encompass large area
– Interconnected system with large number of nodes
– Operate synchronously
– Any disturbance in one part affects the others
– Disturbances propagate rapidly
– Need control requirements with complex hierarchies
– Need coordination of actions between various entities
• Schematic outline of two level centralized
control
– SCADA and EMS
Anatomy of SCADA system
• Supervisory needs- Control actions need confirmation
• Data acquisition +supervisory control=SCADA
• Basic elements
– Interface in the field
– Scan interfaces (real and reactive power, current, voltage,
switch and CB position)
– Transmission to central station
– Processing and display
– Determine control actions
– Transmitting request for control to the field equipment
– Monitoring the completion of control request
– Building data base
Need for EMS
• To determine control action by simulating the
operation of the power system close to real
time
• Basic EMS functions
• AGC – to determine real power output of all
generating units
• ECD or EDC – to determine optimal real power
output of generators
• Reserve monitoring – to compute real power
capability
Additional functions
• State estimator and static security analysis
• Alarm processing
• Display generation
• Report formatting and printing
• Storage of real time data
Infrastructure requirements for the combined
SCADA/EMS system
•Large telecommunication network
•Powerful computer at central location
•Subordinate computers at other locations
•Operator interface (display and request entering )
•Large data storage and historical data archival
 AGC has two main control loops
• Supplementary control action
– Regulate frequency sensing load changes
 AGC control …
• Tie line control

– assume equal generation and load characteristics and interchange

agreement is 100MW from 1 to 2
– System 2 has sudden increase of 30MW
– System 1 will experience a net increase of 15MW
Tie line control

• Uses two inputs

– System frequency
– Net power flowing over the interchange line
• Algorithm
Conceptual model
Conceptual overview of SCDA/EMS system
 Thailand’s SCADA/EMS system

HP-Alpha platform
Tru-64 OS
Oracle relation DB
Meidensha and Wisp+
protocols, DNP3 serial protocol
 2. Security analysis
• Power system security
• Factors affecting power system security
• Contingency analysis detection of network
problems
– Overview of security analysis
– Linear sensitivity factors
– AC power flow methods
– Contingency selection
– Concentric relaxation
 2.1 power system security
• Security
– A major function in power system operation
– It refers to practices designed to keep the system
operating when a component fails
• Example: how to handle when
• equipment fail in a generating unit
• a transmission line is damaged
2.2 Factors affecting power system security

• Operate system so that power is delivered

reliably
• With in the constraints placed on system
operation for reliability, system has to be
operated economically
 2.3 Contingency analysis: detection of
network problems
• Used to predict the impact of outages
• Contingency analysis
– Model single failures or outages (one line or one
generator )
– Multiple equipment failures
– For each outage, checks all lines and voltages and
check against their limits
– Computation time is very critical – less than 1min
• operator has to know if current operation is safe
What would
happen if TL from
3 to 5 open?
Voltage and power in TL’s
when TL 3 to 5 open
Methods used for contingency analysis

• Linear sensitivity factors

– Approximate change in line flows for a change in
generation on the network flows
• Generation shift factors , ali
• Line outage distribution factors dl,k
• If a generator having capacity Po fail, then

• Raise alarm if new flow is greater than rating

• Similar procedure is used for line outage
• Used to calculate DC power flows, MW
• Has only 5% accuracy
• No MVAR analysis is done
Contingency analysis flow chart
Example: In the 6 bus system shown above, outage occurred on the
generator on bus 3 which was having base generation of 60MW. Calculate
the flow on line 1-4 after the outage. Base case flow in line 1-4 is 43.6MW

• Solution
– In sensitivity analysis, assumptions are
• Pick up of lost generation come from reference bus
• Sensitivity factors are pre calculated
– Generation shift distribution factor a =-0.29 1-4,3

– Flow after outage of generator 3 = base case flow 1-4 + a1-

4,3*Pgen3=43.6MW+(-0.29)*(-60MW)=61MW
• AC power flow method
– Considers ac power flows
– Used when voltage magnitudes are critical in
assessing contingencies
– Procedure
• Run AC power flow analysis for each possible
generator, transmission line and transformer outage
• Output – overloads and voltage limit violations
– Newton Raphson method is used
– More accurate than sensitivity method but takes
long time
• Which one to choose?
 Contingency selection

• To solve the dilemma in selection of method

• Select those contingencies which are likely to
result in overload
• Calculate the overload performance index

– Where n is a large number

• Order the PI index in descending order and then
select those on top of the list
3. Preventive Control/Restorative Control constraints in
unit commitment

• Preventive control
– Control to avoid faults
– Applied before a fault occurs
• Restorative control
– Control after fault has occurred
– Used to restore system to its normal operation
Need for Unit commitment

• Load variation- daily and weekly variations

Which one to drop as load varies?

If load is greater than 1000, run all three units
If load is less than 1000 but greater than 600, run unit 1 and unit 2
If load is less than 600 run unit 1 only
Unit commitment constraints

• Unit commitment
– Unit generation schedule
• Minimize operating cost
• Satisfy prevailing constraints
– Load demand
– System reserve requirements over a time period

• Classic UC problem
– Minimize operational cost subject to minimum up and down time
constraints, crew constraints, unit capability limits, generation
constraints and reserve constraints
– Objective function consists of generation cost function and start up
cost functions
 Example: Three generating units were to supply a load of
550MW and each of them having the following generating
capacity and input-output characteristics
Generator Min(MW) Max(MW) Input-output Cost
G1 150 600 1.1
G2 100 400 1.0
G3 50 200 1.2

What will be the optimal combination of the generators to supply the demanded power?
 4. Solution Methods of Unit Commitment

• Priority method
– Generating cost function
• For Hydroelectric plants

– Calculate average generating cost and arrange

them in increasing order
Priority method cont…
• Arrangement based on Full load average production cost

• The unit commitment order will be unit 2-1-3

• Compare with the demand ordering 1-2-3
• Which one has a better spinning reserve?
Spinning reserve=Total generation available-(present
load + losses being supplied)
Priority based algorithm
 5. Dynamic Programming
• Is a systematic way of deciding the unit to start or
stop
• Suppose there are nn1units, total possible
combination 2
• DP reduces the dimensionality of the problem
– Forward DP
• Start with present time and step by step decide until desired
time K
– Backward DP
• Start from the last time needed K and calculate to present time
 Dynamic programming
• Consider the following directed graph, the
goal is to find the minimum cost path from a
to i
 Dynamic programming

• The problem has four decision steps

• At each stage, we apply Bellman’s principle of
optimality
– For a system having N decision stages, what ever the
decision taken initially, the choice of N-1 stages must
constitute the optimal path at each stage viewed as
starting point
• Using the backward procedure
– First stage is hi=10 and gi=12
– Stage 2, we have three paths
• From d, e or f di=dh+hi= 16, fi=fg+gi=19 but e has two
possible paths
• ei=min(eh+hi, eg+gi)=min(18,20)=18
 Dynamic programming
• The stage 3
– Two possible choices
• bi=min(bd+di,be+di)=min(13+16,5+18)=23
• Ci=min(ce+ei, cf+fi)=min(12+18,9+19)=28
• The last stage is
– ai=min(ab+bi, ac+ci)=min(10+23,13+28)=33
• Note: the same result can be obtained by
following the forward procedure
 Dynamic programming exercise
• For the 5x5 grid shown, verify that the
minimum path is the one given in bold arrows
while the maximum path is given by the
narrow arrows
 Recursive DP algorithm

• Minimum cost at time t and state I is given by

• The constraints
Example: use priority method and dynamic programming to solve
the UC problem of the four units given below. The load variation
in 8 hours is also given.

• Given
Complete enumeration
 1. Using priority list method

• Priority order for the units is

– 3-2-1-4
– Load at t=1 is 450
– States {1-11} cannot be used, generation <450
– State 13 violates priority order
 DP method of UC

• Select states
– For t=1-4, states ={12,14,15}
– For t=5-8, states ={5,12,14,15}
– Calculate minimum cost according to the recursive
algorithm
Starting state is {12}
At t=1, we can go to either state {12, 14,15}
• At t=2, we can remain in same state or jump
to new state
 6. Lagrange Relaxation
• Disadvantage of DP
– Computational burden
• In the Lagrangian relaxation method, the UC
problem is formulated as a constrained
optimization problem
• Formulation

Subject to
 Constraint relaxation
• Relax coupled constraints
– 1 and 3 are coupled
– Include them into the objective function using
Lagrangian multipliers
– Modified problem is minimize

Subject to