1. “Importance of Reactive Power Management, Voltage Stability and FACTS Applications in today’s Operating Environment”
Sharma Kolluri
Manager of Transmission Planning
Entergy Services Inc
Engineering Seminar
Organized by IEEE Mississippi Section
Jackson State University
August 20, 2010

2. Outline Introduction VAR Basics Voltage Stability FACTS
Applications at Entergy
Summary .

3. Voltage Profile during Aug 14th Blackout
Voltages decay to almost 60% of normal voltage. This is probably the point that load started dropping off.
However, the recovery is too slow and generators are not able to maintain frequency during this condition.
Many generators trip, load shedding goes into effect, and then things just shut down due to a lack of generation.

4. A “Near” Fast Voltage Collapse in Phoenix in 1995
North American Electric Reliability Council, System Disturbances, Review of Selected 1995 Electric System Disturbances in North America, March 1996.

5. Recommendation#23
Strengthen Reactive Power and Control Practices in all NERC Regions
“Reactive power problem was a significant factor in the August 14 outage, and they were also important elements in the several of the earlier outages”
-Quote form the outage report

6. Reactive Power

7. Laws of Reactive Physics
System load is comprised of resistive current (such as lights, space heaters) and reactive current (induction motor reactance, etc.).
Total current IT has two components.
IR resistive current
IQ reactive current
IT is the vector sum of IR & IQ
IT = IR + jIQ IT IQ IR
North American Electric Reliability Corporation

8. Laws of Reactive Physics
Complex Power called Volt Amperes (“VA”) is comprised of resistive current IR and reactive current IQ times the voltage.
“VA” = VIT* = V (IR – jIQ) = P + jQ
Power Factor (“PF”) = Cosine of angle between P and “VA”
P = “VA” times “PF”
System Losses
Ploss = IT2 R (Watts)
Qloss = IT2 X (VARs) VA Q P
North American Electric Reliability Corporation

9. Reactive Physics – VAR loss
Every component with reactance, X: VAR loss = IT2 X
Z is comprised of resistance R and reactance X
On 138kV lines, X = 2 to 5 times larger than R.
One 230kV lines, X = 5 to 10 times larger than R.
On 500kV lines, X = 25 times larger than R.
R decreases when conductor diameter increases. X increases as the required geometry of phase to phase spacing increases.
VAR loss
Increases in proportion to the square of the total current.
Is approximately 2 to 25 times larger than Watt loss.
North American Electric Reliability Corporation

11. Reactive Power for Voltage Support
Reactive Loads
VARs flow from High voltage to Low voltage; import of VARs indicate reactive power deficit

12. Reactive Power Management/Compensation
What is Reactive Power Compensation?
Effectively balancing of capacitive and inductive components of a power system to provide sufficient voltage support.
Static and dynamic reactive power
Essential for reliable operation of power system
prevention of voltage collapse/blackout
Benefits of Reactive Power Compensation:
Improves efficiency of power delivery/reduction of losses.
Improves utilization of transmission assets/transmission capacity.
Reduces congestion and increases power transfer capability.
Enhances grid reliability/security.

13. Transmission Line Real and Reactive Power Losses vs. Line Loading
Source: B. Kirby and E. Hirst 1997, Ancillary-Service Details: Voltage Control,
ORNL/CON-453, Oak Ridge National Laboratory, Oak Ridge, Tenn., December 1997.

14. Static and Dynamic VAR Support
Static Reactive Power Devices
Cannot quickly change the reactive power level as long as the voltage level remains constant.
Reactive power production level drops when the voltage level drops.
Examples include capacitors and inductors.
Dynamic Reactive Power Devices
Can quickly change the MVAR level independent of the voltage level.
Reactive power production level increases when the voltage level drops.
Examples include static VAR compensators (SVC), synchronous condensers, and generators.

15. Voltage Stability

16. Common Definitions
Voltage stability - ability of a power system to maintain steady voltages at all the buses in the system after disturbance.
Voltage collapse - A condition of a blackout or abnormally low voltages in significant part of the power system.
Short term voltage stability - involves the dynamics of fast acting load components such as induction motors, electronically controlled loads, and HVDC converters.
Long term voltage stability - involves slower acting equipments such as tap-changing transformer, thermostatically controlled loads, and generator limiters.

17. What is Voltage Instability/Collapse?
A power system undergoes voltage collapse if post-disturbance voltages are below “acceptable limits”
voltage collapse may be due to voltage or angular instability
Main factor causing voltage instability is the inability of the power systems to “maintain a proper balance of reactive power and voltage control”

18. Voltage Instability/Collapse
The driving force for voltage instability is usually the load
The possible outcome of voltage instability:
loss of loads
loss of integrity of the power system
Voltage stability timeframe:
transient voltage instability: 0 to 10 secs
long-term voltage stability: 1 – 10 mins
Randy Graves of Distribution Asset Planning is compiling the list

19. Voltage stability causes and analysis
Causes of voltage instability
Increase in loading
Generators, synchronous condensers, or SVCs reaching reactive power limits
Tap-changing transformer action
Load recovery dynamics
Tripping of heavily loaded lines, generators
Methods of voltage stability analysis
Static analysis methods
Algebraic equations, bulk system studies, power flow or continuation power flow methods
Dynamic analysis methods
Differential as well as algebraic equations, dynamic modeling of power system components required

20. Generator Capability Curve
Over-excitation Limit
Lagging (Over-excited)
0.8 pf line
Stator Winding Heating Limit
- Per unit MVAR (Q) +
Normal Excitation (Q = 0, pF = 1) MW
Turbine Limit
Leading (Under-excited)
Under-excitation Limit
Stability Limit

21. P-V Curve

22. Q-V Curve 200 Q-V Curve with Detailed Load Model -80 -60 -40 -20 20 40
Peak Load with Fixed Taps
-80
-60
-40
-20 20 40 60 80
100
120
0.5
0.6
0.7
0.8
0.9 1
1.1
1.2
1.3
1.4
1.5
Voltage (p.u.)
Mvars
Base Case
Contingency

23. Key Concerns Limit UVLS activation Minimize motor tripping
Voltage (pu)

24. Possible Solutions for Voltage Instability
Install/Operate Shunt Capacitor Banks
Add dynamic Shunt Compensation in the form of SVC/STATCOM to mitigate transient voltage dips
Add Series Compensation on transmission lines in the problem area
Implement UVLS Scheme
Construct transmission facilities

25. Voltage Collapse

26. Fault Induced Delayed Voltage Recovery (FIDVR)
FIDVR Definition
Load Models

27. Fault Induced Delayed Voltage Recovery (FIDVR)
What is it?
After a fault has cleared, the voltage stays at low levels (below 80%) for several seconds
Results in dropping load / generation or fast voltage collapse
4 key factors drive FIDVR:
Fault Duration
Fault Location
High load level with high Induction motor load penetration
Unfavorable Generation Pattern

28. Southern Company was really unaware that the system was evolving to the point where concern about exposure to FIDVR was warranted. However, as a result of the Union City event in 1999, we became acutely aware that FIDVR exposure existed.
The Union City event was actually a multiple contingency event – an event well beyond even NERC Category D, and thus you would never plan and build for such an event.
This is a DFR voltage versus time recording at a plant about 50 miles NW of the fault location. Note that the voltage did not fully recover for about 15 seconds. The depth of the voltage depression was more severe at load serving substations more removed from generation (source of dynamic Mvars)

30. Load characteristics
The accuracy of analytical results depends on modeling of power system components, devices, and controls.
Power system components - Generators, excitation systems, over/under excitation limiters, static VAr systems, mechanically switched capacitors, under load tap changing transformers, and loads among others.
Loads are most difficult to model.
Complex in behavior varying with time and location
Consist of a large number of continuous and discrete controls and protection systems
Dynamics of loads, especially, induction motors at low voltage levels should be properly modeled.

31. Induction motor characteristics
Impact of fault on transmission grid
Depressed voltages at distribution feeders and motor terminals
Reduction of electrical torque by the square of the voltage resulting in slow down of motors
The slow down depends on the mechanical torque characteristics and motor inertias
With fault clearing
Speed – per unit
Torque -per unit
Electric torque
1.0
Constant load torque
Square-law load torque
Fig. 1 Induction motor characteristics
Partial voltage recovery
Slowed motors draw high reactive currents, depressing voltage magnitudes
Motor will reaccelerate to normal speed if, electrical torque>mechanical torque
else, the motors will rundown, stall, and trip
The problem is severe in the summer time with large proportion of air conditioner
motors

32. Air conditioner motor characteristics
Main portion (80-87%) consumed by compressor motor
Electromagnetic contactor drop out between (43-56%) of the nominal voltage and reclose above drop out voltage
Stalling at (50-73%) of the nominal voltage
Thermal overload protection act if motors stall for 5-20 seconds
The operation time of thermal over load (TOL) protection relay is inversely proportional to the applied voltage at the terminal
Air conditioner should be modeled to analyze the short term voltage stability problem
Quite important for utilities in the Western interconnection

33. Fig. 2 Traditional load model
Load modeling
Old models – Loads are represented as lumped load at distribution feeder
Does not consider the electrical distance between the transmission bus and the end load components
The diversity in composition and dynamic behavior of various electrical loads is not modeled
Modeling
WECC interim model
20% of the load as generic induction motor load
80% constant current P and constant impedance Q
Distribution Capacitor
Distribution Bus
OLTC
Transmission Bus
Lumped Load (ZIP load)
Fig. 2 Traditional load model 33

34. Composite load modeling
Representation of distribution equivalent
Feeder reactance
Substation transformer reactance
Parameters of various load components
Discharge lighting
Electronic Loads
Constant Impedance loads
Motor loads
Distribution Capacitor
Bus 3
Distribution Capacitor
Transmission Bus
Bus 1
Bus 2
Distribution Bus
Substation Capacitor
OLTC
Distribution Feeder
Feeder Equivalent
Static Loads
(Constant impedance, constant current, constant impedance loads)
Dynamic Loads
(Small motor, Large motor, trip motor loads)
Fig. 3 Composite load model structure

35. FACTS

36. What is FACTS?
Alternating Current Transmission Systems Incorporating Power Electronic Based and Other Static Controllers to Enhance Controllability and Increase Power Transfer Capability.
power semi-conductor based inverters
information and control technologies

37. Major FACTS Controllers
Static VAR Compensator (SVC)
Static Reactive Compensator (STATCOM)
Static Series Synchr. Compensator (SSSC)
Unified Power Flow Controller (UPFC)
Back-To-Back DC Link (BTB)

38. FACTS Applications UPFC STATCOM BTB SSSC Voltage Control
Power System Stability
SSSC
S/S
UPFC
Power Generation
Load
Increased
Transmission Capacity
Inter-area Control
Inter-tie Reliability
Power Flow Control
System Reliability
Improved
Power Quality
Enhanced
Import Capability
Inter-connected
RTO System
Power System
BTB
STATCOM

39. Static VAr compensator (SVC)
Variable reactive power source
Can generate as well as absorb reactive power
Maximum and minimum limits on reactive power output depends on limiting values of capacitive and inductive susceptances.
Droop characteristic XC
TCR XL I V
Firing angle control
Fig. 4 Schematic diagram of an SVC

40. Static compensator (STATCOM)
Voltage source converter device
Alternating voltage source behind a coupling reactance
Can be operated at its full output current even at very low voltages
Depending upon manufacturer's design, STATCOMs may have increased transient rating both in inductive as well as capacitive mode of operation
Transformer
DC-AC switching converter I X
System bus Cs
Vdc V E
Fig. 5 Schematic diagram of STATCOM

41. Technology Applications at Entergy

42. Technology Applications at Entergy to Address Reactive Power Issues
Large Shunt Capacitor Banks
UVLS
Series Compensation
SVC
Coordinated Capacitor Bank Control
DVAR
AVR

43. 2006 Power Systems Conference and Exposition
Determining Reactive Power Requirements in the Southern Part of the Entergy System for Improving Voltage Security – A Case Study
Sharma Kolluri
Sujit Mandal
Entergy Services Inc
New Orleans, LA
Panel on Optimal Allocation of Static and Dynamic VARS for Secure Voltage Control
2006 Power Systems Conference and Exposition
Atlanta, Georgia
October 31, 2006

44. Areas of Voltage Stability Concern
North Arkansas
Mississippi
West of the Atchafalaya Basin
(WOTAB)
Southeast Louisiana
Western Region
Amite South/DSG

45. Study Objective Identify Voltage Stability Problems in the DSG area
Determine the proper mix of reactive power support to address voltage stability problem
Determine size and location of static and dynamic devices.

46. Downstream of Gypsy Area - Critical Facilities
Ninemile Units MW MW MW MW MW
115 kV
230 kV
Michoud Units MW MW MW
115 kV kV
Little Gypsy-South Norco 230kV line
Waterford-Ninemile 230kV line

47. DSG Issues Area load growth 1.6% projected for 2003 - 2013
Weather normalized to 100º F
Projected peak load – 3800 MW
Area power factor - Low
94% at peak load
Worst double contingency
Loss of the Waterford to Ninemile 230 kV transmission line and one of the 230 kV generating units at Ninemile or Michoud
Michoud
Ninemile
New Orleans area voltage profile on June 2, 2003
(with 2 major generators offline)
Area Problems
Thermal overloads of underlying 115 kV and 230 kV transmission system
Depressed voltages throughout New Orleans metro area potentially leading to voltage collapse and load shedding

48. Various Steps Used for Determining Reactive Power Requirements
Step 1 – Problem identification
Step 2 – Determining total reactive power requirements
Step 3 – Sizing and locating dynamic devices
Step 4 – Sizing and locating static shunt devices
Step 5 – Verification of reactive power requirements

49. Tools & Techniques Used
Various tools and techniques used for analysis purposes
PV analysis using PowerWorld
Transient stability using PSS/E Dynamics
Mid-term stability using PSS/E Dynamics
PSS/E Optimal Power Flow
Detailed Models used
Motor models and appropriate ZIP model for dynamic analysis
Tap-changing distribution transformers, overexcitation limiters, self-restoring loads modeled in mid-term stability study

50. Criteria/Requirements
Minimize motor tripping
Improve post-fault voltage
Voltage (pu)

51. Steady State Analysis Results

52. PV Curve Ninemile Unit 4 out-of-service Trip Ninemile Unit 5 and Waterford – Ninemile 230 kV linecc

53. Dynamic Analysis

54. Stability Simulation Ninemile Unit 4 out-of-service Trip Ninemile Unit 5 and Waterford – Ninemile 230 kV line

55. Process for Determining Reactive Power Requirements
Approx 700 MVAr of reactive power shortage identified in the DSG
How much static and how much dynamic?
Criteria for determining static and dynamic requirements
Voltage at critical buses should recover to 1 pu in several seconds
Voltage at critical buses should recover to 0.9 pu within seconds
Voltage should not dip below 0.7 pu for more than 20 cycles
Generator reactive power output should be below Qmax
Factors considered in sizing static/dynamic devices
Short circuit levels, size & location of the stations, number and existing size of cap banks, back-to-back switching, etc

56. SVC Size and Location Sites considered Size Ninemile 230 kV
Gretna 115 kV
Paterson 115 kV
Size
300 MVAR
500 MVAR
Optimal size and location

57. Steps to locate Static Shunt Devices
Static shunt requirements – 400 MVAR approximately
Options available to locate the static shunt devices on the transmission or distribution systems
OPF Program used to come up with size and location of shunt devices

58. OPF Application PSS/E OPF Program used
Objective Function – Minimize adjustable shunts
OPF simulated for critical contingencies

59. List of Shunt Capacitor Banks Banks Recommended

60. Simulation Results with the Capacitors and SVC Ninemile Unit 4 out-of-service Trip Ninemile Unit 5 and Waterford – Ninemile 230 kV line

61. SVC Performance Ninemile Unit 4 out-of-service Trip Ninemile Unit 5 and Waterford – Ninemile 230 kV line

62. Summary
Process for determining static and dynamic reactive power requirements discussed
OPF program utilized for sizing/locating static shunt capacitor banks
Results verified using mid-term stability simulations
Study recommendation – 400 MVAR of static shunt devices and 300 MVAR of dynamic shunt compensation

63. Ninemile SVC Configuration
= 75 MVAr
= 150 MVAr

64. External Device Control Single line diagram of SVC and MSC

65. SVC Ninemile

66. SVC Ninemile

67. SVC Topology: 2 x 75MVAr TSC & 1 x 150MVAr TSC
Porter 0/+300Mvar SVC
SVC Topology: 2 x 75MVAr TSC & 1 x 150MVAr TSC

68. Porter Static Var Compensator (SVC)
Maintains system voltage by continuously varying VAR output to meet system demands Controls capacitor banks on the transmission system to match reactive output to the load requirements.

69. Porter SVC

70. Series Capacitor – Dayton Bulk 230kV Station
The Capacitor offsets reactance in the line, making it appear to the system to be half of its actual length. Power flows are redirected over this larger line, unloading parallel lines and increasing transfer capability.

71. DSMES Unit Stores Energy in a superconducting coil
Automatically releases energy to the system when needed to ride through voltage dips caused by faults. This unit improves power quality and reduces customer loss of production.

72. Industry Issues Coordination of reactive power between regions
No clearly defined requirements for reactive power reserves
Proper tools for optimizing reactive power requirements
Incentive to reduce losses

73. Summary
The increasing need to operate the transmission system at its maximum safe transfer limit has become a primary concern at most utilities
Reactive power supply or VAR management is an important ingredient in maintaining healthy power system voltages and facilitating power transfers
Inadequate reactive power supply was a major factor in most of the recent blackouts

74. Questions?

75. Under Voltage Load Shed Logic - Western Region
T&D Planning
April 2010

76. Western Region – Overview
≤ 230 kV Tie Lines
Generation
Load Center

77. Load Projection
2010 peak: 1770 MW
2012 peak: 1852 MW

78. Sample PV Curve Result Lewis Creek Unit 1 & China-Porter 230kV Out - 2010

79. 2010 Summer PV Curve Analysis
Scenarios
P Limit (MW)
With 3% Margin (MW)
Voltage (4/8 Buses) (pu)
Lewis Creek U1 out
2385
2313
0.84 – 0.89
Lewis Creek U1 + China-Jacinto out
2260
2192
0.83 – 0.89
Lewis Creek U1 + Grimes-Crockett out
2230
2163
0.86 – 0.91
Lewis Creek U1 + China-Porter out
2065
2003
0.85 – 0.93
Approved Construction Plan Projects included:
*Relocate Caney Creek 138kV

80. Dynamic Analysis Results

81. Results: 2010 case without load shed
Case 3 Voltages (pu): Goslin: 0.810; Conroe: 0.855; Cleveland: 0.909; Jacinto: 0.924; Dayton: 0.944; Huntsville: 0.944
Case 4 Voltages (pu): Goslin: 0.757; Conroe: 0.800; Dayton: 0.913; Huntsville: 0.928; Cleveland: 0.928; Rivtrin: 0.941

82. 2010 Summer Conditions - Dynamics Analysis
Lewis Creek Unit 1 outaged in the base case
50% induction motor load is modeled
Result: Shed Load Block 1 (183 MW)

83. Observations for 2010 Summer Peak Conditions
Existing load shed logic in Western Region OK for 2010 Summer conditions
Voltage at some critical buses drop below 0.7 pu for more than 20 cycles – Potential of motor load tripping
Conclusions for 2010 Summer
Reducing load shed blocks to MW in Western Region has no negative impact

84. Results: 2010 case with load shed (Block 1)
Case 3 Voltages (pu): Goslin: 0.872; Conroe: 0.902; Cleveland: 0.934; Jacinto: 0.948; Dayton: 0.966; Huntsville: 0.968
Case 4 Voltages (pu): Goslin: 0.827; Conroe: 0.855; Dayton: 0.939; Cleveland: 0.951; Huntsville: 0.954; Jacinto: 0.964

85. Conclusions and Recommendations
Retain the exiting UVLS logic
Change the load blocks
Block one: 180 MW
Block two: 70 MW (existing size 111 MW)

86. Proposed Load Shed Logic
4/8 buses <0.90 pu
Armed all time
Drop load
One or more Lewis Creek units in-service?
OEL at Lewis Creek units
4/8 buses
< 0.92 pu
Time Delay 3 seconds
Load Blocks:
Block 1: 175 MW
Alden: 50 MW
Metro: 35 MW
Oakridge:30 MW
Goslin: 60 MW
Block 2: 75 MW
In the vicinity of Block 1
Monitored Buses:
Metro 138kV
Goslin 138kV
Alden 138kV
Oakridge 138kV
Huntsville 138kV
Rivtrin 138 kV
Poco 138 kV
Conroe 138 kV
Reset the Process for next LVSH block
Load Blocks:
Block 1: 175 MW
Block 2: 75 MW
The above conditions need to be met for 3 scans to trigger load shedding