system impact study final report · system impact study – final report electranix ii may 31 st,...

i

TI_04-1214 WIND PROJECT

SYSTEM IMPACT STUDY FINAL REPORT

Prepared for

Tri-State Generation & Transmission Association, Inc.

By

Dennis Woodford, Garth Irwin, and Andrew Isaacs. Electranix Corporation 107 – 865 Waverley St., Winnipeg, MB R3T 5P4

May 31st, 2004

Phone: +1 204 953 1832

Fax: +1 204 953 1839

e-mail: [email protected]

mailto:[email protected]

TI-04-1214 WIND PROJECT SYSTEM IMPACT STUDY – FINAL REPORT

i

May 31st, 2004

SUMMARY A system impact study for the TI_04-1214 Wind Ranch has been commission by Tri-State Generation and Transmission Association (Tri-State). A previous report [1] examined power flow, dynamic and short circuit studies for the TI_04-1214 Wind Ranch rated at 40 MW, without the proposed Tri-State’s Gladstone to Walsenburg 230 kV transmission line. This second study includes the Gladstone to Walsenburg 230 kV line1, and investigates the TI_04-1214 Wind Ranch at the same interconnection point, but with 80 and 120 MW injections in addition to the original 40 MW output level. If power from the wind ranch supplies its designated load by backing off an equivalent amount of power at San Juan generating station, acceptable steady state performance during contingencies for the three acceptance criteria of line/transformer overload, undervoltage deviations and maximum/minimum bus voltage limits was generally achieved in the North East Area of New Mexico (also known as the NEA) for the operation of the TI_04-1214 Wind Ranch up to 120 MW. Marginal voltage deviations exceeding 5% were observed on some of Tri-State’s 115 kV buses for the following contingencies: Ojo-Taos 345kV line (or 345/115kV Taos transformer), Colinas-Rowe Tap 115kV line, and the Black Lake-Taos 115kV line. The list of contingencies studied is in Appendix I. If designated load at or around Hernandez substation is additional to the normal projected loads for the area, and is to be supplied from the wind ranch without backing off scheduled generation, then major additional facilities are required in the Taos – Ojo – Hernandez triangle for a rating of the TI_04-1214 Wind Ranch exceeding approximately 40 MW. An increase in wind ranch rating above 40 MW would be possible with increased investment in facilities that have yet to be defined in order to comply with acceptable steady state performance during contingencies. Acceptable performance during stability analysis was generally achieved. A pre-existing problem in this part of the system was identified for a three-phase fault on the Gladstone to Walsenburg 230 kV line2. This fault causes the Bravo Dome compressor motors to trip on under frequency. When wind generation at the TI-04-1214 project is added at levels of 80 MW or higher, the load at Bravo Dome is able to remain in service3.

1 At this time, the proposed Gladstone – Walsenburg 230 kV line is projected to be in service in December 2006. 2 Although the Gladstone – Walsenburg 230 kV line does not currently exist, in terms of the TI-04-1214 Wind Ranch, the potential to trip Bravo Dome load for underfrequency conditions exists today and was confirmed to be possible after the addition of the Gladstone – Walsenburg 230 kV line. 3 The Tri-State transmission system serving the Bravo Dome load is intentionally operated up to a steady state stability limit with all transmission lines in service.


ELECTRANIX ii

May 31st, 2004

Studies using PSCAD (a 3-phase instantaneous solution simulation program [3]) were performed to determine short circuit current levels, determine the level of negative sequence imbalance, study wind turbine and SVC interactions, determine the effect of the pulsations from compressor motors on the wind turbines, look for amplification of wind turbine harmonics and to study the transient impact of closing the breaker on the Gladstone to Walsenburg 230 kV line. From steady state, short circuit, electromagnetic transients and also transient stability analysis considerations, the operation of the TI_04-1214 Wind Ranch up to 120 MW still appears a realistic objective for the operating condition where San Juan generation can be backed off against wind power generated by the wind ranch. The facilities required to achieve operation of the wind ranch at 120 MW are minimal providing generation can be backed off at San Juan. These include replacing the current transformers on the 115 kV transmission lines that are presently limiting and to identify if conductors on the 115 kV transmission lines have to be raised. The 115 kV transmission line section between the TI_04-1214 interconnect point and the proposed Gladstone substation is of concern at this stage. If load is added in the area of the Hernandez substation that is supplied with power from the TI_04-1214 Wind Ranch so that no scheduled generation can be backed off, then major facilities will be required in the area to accommodate load and wind ranch capacity above about 40 MW. The Facilities Study will address these issues provided the load sink can be defined. The TI_04-1214 project is modeled as an interconnecting bus and radial facilities to the wind ranch. Although the System Impact Study does not detail the interconnection facilities required, full sectionalizing will be required at the point of interconnection via a 115 kV substation.


ELECTRANIX iii

May 31st, 2004

CONTENTS 1 INTRODUCTION ..................................................................................... 1 2 BASIC ASSUMPTIONS............................................................................ 2

2.1 Originating Case for Study ....................................................................2 2.2 TI-04-1214 Wind Ranch Representation................................................2 2.3 SVC at Clapham....................................................................................7 2.4 Thermal Capacity of the Springer to Rosebud 115 kV Line....................7 2.5 Power Flow Acceptance Criteria............................................................8 2.6 Transient Stability Acceptance Criteria..................................................8 2.7 Simulation Software ..............................................................................8 2.8 Load Sink..............................................................................................9

3 LOAD FLOW STUDY METHODOLOGY.............................................. 9 4 RESULTS (STEADY STATE) .................................................................. 9

4.1 Summary of Power Flow Results...........................................................9 4.2 Discussion of Power Flow Results ....................................................... 11

5 STABILITY STUDY METHODOLOGY ........................ .......................11 6 RESULTS (TRANSIENT STABILITY)..................................................12

6.1 Summary of Transient Stability Results ............................................... 12 6.2 Discussion of Bravo Dome Compressor Motor Tripping...................... 12 6.3 Discussion of Tripping of TI-04-1214 Wind Ranch ............................. 13 6.4 Sympathetic Tripping .......................................................................... 13

7 SHORT CIRCUIT AND TRANSIENTS STUDY ...................................14 7.1 Model Development ............................................................................ 14 7.2 Short Circuit Current Calculations ....................................................... 15 7.3 Measurement of Negative (-ve) Sequence Voltage Levels.................... 18 7.4 Interactions between the Clapham SVC and the Wind Turbines........... 19 7.5 Impact of Pulsating Bravo Dome Compressor Motors on the Wind

Turbines .............................................................................................. 19 7.6 Harmonic Studies ................................................................................ 20 7.7 Impact of Closing the Gladstone to Walsenburg 230 kV line................ 22

8 CONCLUSIONS .......................................................................................24 9 REFERENCES .........................................................................................25 10 ACKNOWLEDGEMENTS......................................................................25

APPENDICES

Appendix I…. Power flow case list

Appendix II….Power flow analysis with generation backed off at San

Juan to accommodate the power from the TI_04-1214 Wind Ranch

Appendix III…Transient stability analysis results with generation backed

off at San Juan to accommodate the power from the TI_04-1214 Wind Ranch


ELECTRANIX iv

May 31st, 2004

Appendix IV…NERC/WECC Performance Standard

Appendix V….Transient stability analysis results with load added at Hernandez to accommodate the power from the TI_04-1214 Wind Ranch

Appendix VI.. Power flow analysis with load added at Hernandez to

accommodate the power from the TI_04-1214 Wind Ranch Appendix VII..GE Wind turbine generator model parameters Appendix VIII..Sympathetic tripping report Appendix IX..Power flow single line diagrams Appendix X…Single line diagrams for the TI_04-1214 Wind Ranch.


ELECTRANIX 1

May 31st, 2004

1 INTRODUCTION A system impact study for the TI_04-1214 Wind Ranch has been commissioned by Tri-State for the requestor; “Customer”. The wind ranch will be located in the North East area of New Mexico (also known as the NEA) and will interconnect onto the existing 115 kV transmission line that extends from Springer substation to Rosebud substation. The interconnection point will be approximately half way between Clapham and Rosebud substations or 11.5 miles each way (see Figure 1). A previous report [1] examined power flow, dynamic and short circuit studies for the TI-04-1214 Wind Ranch rated at 40 MW, and without the proposed Gladstone to Walsenburg 230 kV transmission line. This second study investigates the TI-04-1214 Wind Ranch at the same interconnection point with the Gladstone to Walsenburg line in service, with 80 and 120 MW injections in addition to the original 40 MW output.

230 kV

Gladstone

115 kV

Storrie

Springer Black Lake

York Canyon

Clapham – location of existing and replacement SVC

Rosebud (Bravo Dome)

115 kV

Walsenburg

Taos

Figure 1: Single line diagram of the area under study showing the location of the proposed TI -04-1214 Wind Ranch.

New 230 kV transmission line

Colorado

New Mexico

345 kV

34.5 kV

TI-04-1214 Wind Ranch

Customer (new)

Hernandez

To Norton

To San Juan

To Arriba_T

Ojo


ELECTRANIX 2

May 31st, 2004

This is a final report of the power flow, transient stability and PSCAD electromagnetic transients study results. Two operating conditions were examined, which were based on where the power generated by the TI-04-1214 Wind Ranch was delivered. The first operating condition accommodated the power from the TI-04-1214 Wind Ranch by backing off an equivalent amount of generation at San Juan. The second operating condition was to lump a load at Hernandez that was equivalent to the power being generated from the wind ranch at TI-04-1214. The cases run are summarized in Appendix I. The steady state power flow results are summarized in Appendix II (with power backed off at San Juan) and Appendix VI (with load added at Hernandez). The transient stability results are shown in Appendix III with generation backed off at San Juan to accommodate the wind power, and in Appendix V with load added at Hernandez to accommodate the wind power generated. Due to the GE Wind turbine/generator proposed for TI-04-1214 that uses doubly fed generators, it was necessary to undertake the following studies using PSCAD, which are discussed in Section 7 below:

- Short circuit current calculations - Measurement of –ve sequence voltage phase unbalance - Interactions between the Clapham SVC and the wind turbines - Impact of pulsating Bravo Dome compressor motors on the wind turbines - Harmonic studies - Impact of closing the Gladstone to Walsenburg 230 kV line.

2 BASIC ASSUMPTIONS The assumptions that apply for the power flow study are summarized as follows: 2.1 Originating Case for Study The originating case for this study is the 2008 hs2-sa (summer peak load) case prepared by member organizations of the Western Electricity Coordinating Council (WECC). This case included the proposed Gladstone to Walsenburg 230 kV transmission line. 2.2 TI-04-1214 Wind Ranch Representation Significant detail was included in the representation of the wind ranch to include the main substation transformer, and an equivalent 34.5 kV feeder to each cluster of wind turbine generators (WTGs) connected to that feeder. The detail of the layout and information of the feeder was taken from single line diagrams SE-1, for the 40, 80 and 120 MW wind ranch options (see Appendix X).


ELECTRANIX 3

May 31st, 2004

Each cluster of 12 to 15 WTGs was represented as a single WTG in the studies. There were 2 clusters and consequently 2 equivalent WTGs represented in the 40 MW option. The 80 MW injection was represented by 4 lumped WTGs for its 4 clusters, and the 120 MW injection by 6 lumped WTGs for 6 clusters. This information is shown in Figures 2, 3 and 4. The underground 34.5 kV cables applied for each feeder to its single equivalent lumped WTG consisted of two types of cables: 1000 MCM Aluminum between the main substation and the cluster. 4/0 Aluminum within each cluster. The parameters applied for each type of cable for the power flow program in per unit of 100 MVA base for 1000 feet length were:

Table 1: Parameters for Wind Ranch 33 kV collector system cables. Cable R/1000’ X/1000’ B/1000’ 1000 MCM Al 0.0034 0.005 0.00072 4/0 Al 0.01 0.007 0.00026


ELECTRANIX 4

May 31st, 2004

Customer 115kV Clapham 115kV Rosebud 115kV

Customer 34.5kV

TMVA 27/36/45 X = 8.5% (0.3148pu)

R = 0.0315 X = 0.0338 B = 0.007 1.68 mi.

R = 0.0265 X = 0.0189 B = 0.006 1.5 mi.

Pmax = 21MW MVA = 23.38 Qmax = 6.9MVAr Qmin = -10MVAr

14 Turbines

22 MVA X = 0.2760

23 MVA X = 0.257

34.5kV 34.5kV

0.575kV 0.575kV

Pmax = 19.5MW MVA = 21.71 Qmax = 6.4MVAr Qmin = -9.5MVAr

13 Turbines

Figure 2: Customer 2 Feeder Equivalent Model For 40 MW Wind Ranch used in Cases 1,2,5 and 6 of system impact study. All impedances on 100 MVA base.


ELECTRANIX 5

May 31st, 2004

Customer115kV Clapham 115kV Rosebud 115kV

Customer 34.5kV

TMVA 54/72/90 X = 8.5% (0.1574pu)

R = 0.0369 X = 0.0409 B = 0.0082 2.05 mi.

R = 0.0259 X = 0.0238 B = 0.0059 1.44 mi.

R = 0.0548 X = 0.0597 B = 0.0122 3.08 mi.

R = 0.0361 X = 0.0257 B = 0.0084 2.05 mi.

Pmax = 22.5MW MVA = 25.05 Qmax = 7.4MVAr Qmin = -11MVAr

15 Turbines

Pmax = 18MW MVA = 20 Qmax = 5.9MVAr Qmin = -8.8MVAr

12 Turbines

25 MVA X = 0.24

20 MVA X = 0.3

25 MVA X = 0.24

20 MVA X = 0.3

34.5kV 34.5kV 34.5kV 34.5kV

0.575kV 0.575kV 0.575kV 0.575kV


12 Turbines


15 Turbines

Figure 3: Customer 4 Feeder Equivalent Model For 80 MW Wind Ranch used in Case 3 and Case 7 of system impact study. All impedances on 100 MVA base.


ELECTRANIX 6

May 31st, 2004

Customer 115kV Clapham 115kV Rosebud 115kV

Customer 34.5kV

TMVA 81/108/135 X = 8.5% (0.1049pu)

R = 0.0369 X = 0.0409 B = 0.0082 2.07 mi.

R = 0.1302 X = 0.1807 B = 0.0279

7.27 mi.

R = 0.05 X = 0.0639 B = 0.0109 2.79 mi.

R = 0.013 X = 0.0111 B = 0.003 0.74 mi.

R = 0.0884 X = 0.1127 B = 0.0192 4.95 mi.

R = 0.0429 X = 0.0511 B = 0.0094 2.41 mi.


15 Turbines


13 Turbines


12 Turbines


15 Turbines

25 MVA X = 0.24

22 MVA X = 0.2764

20 MVA X = 0.3

22 MVA X = 0.2764

22 MVA X = 0.2764

25 MVA X = 0.24

34.5kV 34.5kV 34.5kV 34.5kV 34.5kV

0.575kV 0.575kV 0.575kV 0.575kV 0.575kV


13 Turbines


13 Turbines

Figure 4: Customer 6 Feeder Equivalent Model For 120 MW Wind Ranch, used in Case 4 and Case 8 of system impact study. All impedances on 100 MVA base.


ELECTRANIX 7

May 31st, 2004

For each wind ranch rating option, the 34.5 kV cable feeders to the equivalent lumped WTG is summarized as follows:

Table 2: Wind Ranch 33 kV feeder cable and unit transformer impedances Cluster Feeder

Length (miles)

R (pu) X (pu) B (pu) Unit Tfr Impedance (pu)

No. WTGs

40 MW Feeder #1 1.68 0.0315 0.0338 0.007 0.276 13 Feeder #2 1.50 0.0265 0.0190 0.006 0.257 14 80 MW Feeder #1 2.05 0.0368 0.0409 0.0082 0.2395 15 Feeder #2 1.44 0.0259 0.0238 0.0059 0.3 12 Feeder #3 3.08 0.0548 0.0597 0.0122 0.2395 15 Feeder #4 2.05 0.0361 0.0257 0.0084 0.3 12 120 MW Feeder #1 2.07 0.0369 0.0409 0.0082 0.2395 15 Feeder #2 2.79 0.05 0.0639 0.0109 0.276 13 Feeder #3 0.74 0.013 0.0111 0.003 0.3 12 Feeder #4 4.95 0.0884 0.1127 0.0192 0.2876 13 Feeder #5 2.41 0.0429 0.0511 0.0094 0.2764 13 Feeder #6 7.27 0.1302 0.1807 0.0278 0.2395 15

Each individual WTG is a GE Wind-VAR unit rated at 1.5 MW, and the reactive power range that can be applied for ac voltage control is +0.49 MVAR, and –0.73 MVAR. 2.3 SVC at Clapham Only one ±25 MVAR is modelled as in service at Clapham. Although a second ±25 MVAR is being planned at the same location, it was not considered in these studies. 2.4 Thermal Capacity of the Springer to Rosebud 115 kV Line The thermal capacity of the 115 kV transmission line conductor from Springer to Rosebud is applied at 146.6 MVA instead of the 119.9 MVA as it is presently rated. The 146.6 MVA rating was determined through the utilization of a statistical static thermal conductor rating, based on the weather conditions recorded in New Mexico and the surrounding region4

4 McElvain, F.R. and Mulnix, S.S, Statistically Determined Static Thermal Ratings of Overhead High Voltage Transmission Lines in the Rocky Mountain Region, IEEE Transactions on Power Systems, May 2000, p. 899.


ELECTRANIX 8

May 31st, 2004

2.5 Power Flow Acceptance Criteria For each of the power flow cases and each contingency, bus voltages and equipment ratings are to fall within the WECC and Tri-State system operation criteria. Tri-State criteria that differ from WECC criteria are identified below. These include:

• Bus Voltages: 0.95 < V steady state < 1.05 during normal system conditions, and must be between 0.90 < Vsteady state < 1.10 during single contingency outage conditions.

• Steady state voltage deviations during contingencies must be less than 5%. Although this is not a substitute for voltage collapse studies, this criterion is a screening test for potential voltage collapse, where voltages that drop by more than 5% are flagged for further review. Past P-V analyses in the area have determined that the region is operating near its point of collapse for contingencies to the west of Springer substation. Additional generation to the east of Springer substation improves the voltage stability of the region.

• Line and Transformer Ratings in normal steady state conditions must not exceed 80% of the full load rating.

• Line and Transformer Ratings during contingencies must be within 100% of the continuous rating.

• Load tap changers, phase shifters, and switched shunt devices can operate during contingencies. This is a variation from the WECC criteria since Tri-State’s system is primarily rural with very low consumer density.

• Fixed area power flow interchanges are enforced during all contingencies. The slack bus was a unit at San Juan.

2.6 Transient Stability Acceptance Criteria For each of the base cases and each contingency, transient bus voltages and frequency deviations are to fall within the WECC and Tri-State system operation criteria. These criteria are summarized in Appendix IV, and include:

• Bus frequencies at a load bus must not fall below 59.6Hz for longer than 6 cycles.

• Voltage dip at a load bus must not exceed 25% at load busses, or 30% a non-load busses.

• Voltage dip at a load bus must not exceed 20% for longer than 20 cycles. • Mechanical oscillations must be damped after 20 seconds.

2.7 Simulation Software The simulation software used for this study is the GE PSLF V14.1 program. This version of the program includes the latest GE PSLF wind turbine dynamics models. The data for the parameters for the PSLF wind turbine model are included in Appendix VII along with the results of the stability study (Appendices III and V).


ELECTRANIX 9

May 31st, 2004

2.8 Load Sink The power balance when wind power from TI-04-1214 was being generated was achieved with two operating conditions. Firstly, power generated by San Juan was reduced by an equivalent amount. Secondly, load was added at Hernandez by an equivalent amount. Results for each method are included in the Appendices II and VI and are summarized in this report.

3 LOAD FLOW STUDY METHODOLOGY This interim report considers the steady state system impact of the power flow during single contingencies. In this particular study, outages within the NEA were simulated as listed in Appendix I. For each contingency, Cases 1 to 4 are with the Bravo Dome compressor motors all in-service and fully loaded at 43.8 MW total.

• Case 1 is the base case with no wind power being generated. • Case 2 is with 40 MW of WTG capacity in place and fully loaded. • Case 3 is with 80 MW of WTG capacity in place and fully loaded. • Case 4 is with 120 MW of WTG capacity in place and fully loaded.

Cases 5 to 8 are for the Bravo Dome compressor motors all out of service.

• Case 5 is the base case with no wind power being generated. • Case 6 is with 40 MW of WTG capacity in place and fully loaded. • Case 7 is with 80 MW of WTG capacity in place and fully loaded. • Case 8 is with 120 MW of WTG capacity in place and fully loaded.

A summary of the results for each case studied is included in Appendix II for the load sink at San Juan, and in Appendix VI for the load sink at Hernandez.

4 RESULTS (Steady State) 4.1 Summary of Power Flow Results The results of the power flow study in Appendix II consider the power flow acceptance criteria (Section 3.5 above) for the NEA area of interest (Zone 121 in the WECC base case). Output results are only shown if the power flow approaches violation conditions. The conclusions for the operating condition when the load sink is at San Juan are:

• There were no overload violations as a consequence of the TI-04-1214 wind farm, for any power level up to 120MW. The transformer at the Rowe Tap was overloaded in the base case and throughout the contingency cases at 0.84 per unit.


ELECTRANIX 10

May 31st, 2004

• Undervoltage deviation violations greater than 5% were observed when the wind power from TI-04-1214 was both 80 and 120 MW for Contingency No. 5 (outage of the Black Lake -Taos 115 kV line), and for Contingency No. 8 (outage of the Colinas – Rowe Tap 115 kV line). Additional undervoltage deviations exceeding 5% were observed at the Taos 115 kV busbar for outage of the Ojo – Taos 345 kV line (Contingency 12) and for the outage of the Taos 345/115 kV transformer (Contingency 15) when the TI-04-1214 wind farm was generating at 120 MW and the compressor motors at Bravo Dome were out of service.

• There were no bus voltage violations for contingency conditions.

The ±25 MVAR SVC at Clapham remained operational for the cases studied, and was sized sufficiently to prevent any voltage violations. The switching of the shunt capacitor banks at Clapham and Springer substations was checked to ensure the SVC was operating with sufficient range. Results for the operating condition when the load sink is a load at Hernandez are found in Appendix VI, and are summarized as follows:

• Outage of the Taos – Hernandez 115 kV line causes severe overload of the Ojo 115/345 kV transformer and the Ojo – Hernandez 115 kV transmission line.

• Outage of the Ojo – Taos 345 kV transmission line causes severe overload of the Ojo 115/345 kV transformer and the Ojo – Hernandez 115 kV transmission line. There is a similar performance observed for the outage of the 115/345 kV transformer (as would be expected). There are also significant voltage drops on the 115 kV transmission system from Hernandez to Springer.

• There are cases where PSLF failed to converge, particularly when the wind ranch is generating 80 to 120 MW.

• The system impact around Taos – Ojo –Hernandez with the Bravo Dome compressor motor load out of service is slightly more severe compared with the compressor motor load being in service.

There is a difference in system impact between having the load sink being accommodated by backing off San Juan generation compared to a lumped load at Hernandez. When the San Juan generation is backed off, a 120 MW rated wind ranch for TI-04-1214 is possible with minor additional facilities required. With the load sink lumped as a single load at Hernandez, major additional facilities are required in the Taos – Ojo – Hernandez triangle for a rating of the TI-04-1214 Wind Ranch exceeding 40 MW. An increase in wind ranch rating above 40 MW would be possible with increased investment in facilities that have yet to be defined.


ELECTRANIX 11

May 31st, 2004

4.2 Discussion of Power Flow Results The results of the power flow study highlight how the definition of the load sink impacts system performance. If the load sink is not additional load to what is built into the WECC 2008 hs2-sa (summer peak load) base case used for this study, then the load sink is achieved by backing off scheduled generation. If San Juan generating station is backed off, then the results in Appendix II are valid and little in the way of additional transmission facilities are required, based on the results from this System Impact Study. If the loads being targeted from TI-04-1214 Wind Ranch have all or part of their generation requirements normally scheduled from a generating station other than San Juan, then those generating stations need to be identified. If new load at or near Hernandez is being supplied from TI-04-1214 so that no scheduled generation needs to be backed off, then the results in Appendix VI are valid, and additional transmission facilities will be required in the Taos – Ojo – Hernandez triangle.

5 STABILITY STUDY METHODOLOGY Stability studies were performed with the goal of evaluating the voltage stability response of the NEA 115 kV power system, and dynamic stability of the Bravo Dome synchronous compressor motors. Load flow base cases 1 to 4 (Appendix I) have the Bravo Dome compressor motors operating. These base cases were run with the disturbances listed in Table 3. These cases are not exhaustive, but have been selected to represent severe disturbances. Since disturbance S2 was not applicable to Case 1, a total of 19 simulations were run. For this study, results of the disturbance simulations were examined relative to the April 2003 NERC/WECC Planning Standards. These simulations represent disturbances based on the WECC 2008 heavy-summer configuration. It is noteworthy to mention that the transient stability representation of the compressor motors does not include the pulsating torque effect on the compressor motors due to piston action. This effect was included in PSCAD models, but only for breaker energization studies (Section 7). The Bravo Dome Compressor motors were modeled in both PSCAD and PSLF as two synchronous motors. One motor represents six 8000 hp motors and one to represent three 6000 hp motors. Exciter models were provided from an earlier study [2]. Impedance relays were represented on the 115 kV transmission line sections Springer to Gladstone, and Gladstone to Clapham. The settings on each based on the impedance Z of the transmission line section were:


ELECTRANIX 12

May 31st, 2004

Zone 1: 80% of Z, no delay Zone 2: 125% of Z, with 20 cycles delay Zone 3: 150% of Z, with 60 cycles delay. The cases were undertaken with the GE Wind -VAR wind turbine and generator model included in the PSLF program and with specific parameters provided by GE Wind, which contain the undervoltage settings and ride-through capability they expect to provide for TI-04-1214 Wind Project. This model also includes the reactive power control capability of the wind turbine.

Table 3: Transient Stability Disturbance List Test Stability Disturbance Descriptions

S1 3Ø permanent fault on the Customer 34.5 kV Transformer bus, transformer clears after 6 cycles

S2 Sudden loss of Power from the Wind Ranch due to wind gust

S3 3Ø permanent fault on the Gladstone-Walsenburg 230kV line, clears after 6 cycles

S4 3Ø permanent fault on the Springer-Black Lake 115kV line, clears after 6 cycles

S5 3Ø permanent on the Ojo-Taos 345kV line, clears after 6 cycles.

6 RESULTS (Transient Stability) 6.1 Summary of Transient Stability Results Results for all of the 19 disturbance simulations are shown in Appendix III, and met or exceeded the specified category of the NERC/WECC Performance Standards (See Appendix IV). For the condition where the load sink was lumped at Hernandez, the simulation results are assembled in Appendix V. 6.2 Discussion of Bravo Dome Compressor Motor Tripping Although the monitored system busses met the required stability criteria, several of the disturbance cases caused load tripping of the Bravo Dome compressor motors. During a three-phase fault applied to the middle of the Gladstone-Walsenburg 230 kV line (Test S3), Bravo Dome compressor motors tripped due to under-frequency for Cases 1 and 2 (0 MW and 40 MW of wind, respectively). As more wind generation was added in Cases 3 and 4 (80 MW and 120 MW respectively), the compressor motors at Bravo Dome were able to remain on through the disturbance. For Case 2 the 40MW TI-04-1214 wind ranch also tripped due to undervoltage.

If the 3-phase fault on the Gladstone-Walsenburg line is cleared after 4 cycles instead of 6 (a realistic scenario), the Bravo Dome motors do not trip in Case 2,


ELECTRANIX 13

May 31st, 2004

which has 40 MW of wind generation. Additionally, the faster fault removal is sufficient to prevent the wind ranch from tripping during this disturbance. The reduced fault clearing time is not sufficient to prevent the Bravo Dome compressor motors from tripping during Case 1, where no power is available from wind. Further tests with simulated 1 phase faults instead of three phase faults with 6 cycle clearing time show there is no tripping of the Bravo Dome compressor motors.

It is worth noting that the under-frequency resulting from the Gladstone-Walsenburg 230kV line fault was reported in the WECC 2003 Study Program Annual Report – Case D0310, published Feb. 2004. This is an indication that the addition of larger amounts of wind generation in the immediate vicinity of the Bravo Dome compressor motors provides system support including offsetting the required power flow through the radial lines to the motors. 6.3 Discussion of Tripping of TI-04-1214 Wind Ranch During the 3 phase fault on the Walsenburg-Gladstone 230kV line, the turbines at the TI-04-1214 wind ranch also trip off for the 40 MW case. As discussed in Section 6.2, this condition may be alleviated by reducing the fault clearing time to 4 cycles from 6. Additionally, a fault on the main substation transformer at the wind farm site causes the wind ranch to be islanded from the network, and the wind turbines to trip. 6.4 Sympathetic Tripping The outage determined to be the most likely to cause sympathetic tripping on the Springer – Rosebud transmission system was the loss of the Gladstone – Walsenburg 230 kV line. This disturbance was studied for the following four loadings of TI-04-1214 Wind Ranch:

1. 0 MW wind generation 2. 40 MW wind generation 3. 80 MW wind generation 4. 120 MW wind generation

For case 4 with 120 MW wind generation, additional 3 phase faults were placed on nearby 115 kV transmission lines to see if sympathetic tripping occurred. These disturbances were located on:

1. Springer – York Canyon 115 kV transmission line 2. Springer – Storrie 115 kV line 3. Springer – Black Lake 115 kV line.

All faults were cleared after 6 cycles. The results of the stability study showed that for the 4 cases studied, the impedance characteristic entered the 3rd zone of protection for the distance relay located on the


ELECTRANIX 14

May 31st, 2004

Springer – Gladstone transmission line section for a very short time following a Gladstone – Walsenburg line fault. With the load sink located at Hernandez, the impedance characteristic entered the 3rd zone only for case 4, with 120 MW of generation. None of these cases resulted in sympathetic line tripping.

7 SHORT CIRCUIT AND TRANSIENTS STUDY The following studies were performed with PSCAD:

- Short Circuit Current Calculations - Measurement of –ve sequence voltage phase unbalance - Interactions between the Clapham SVC and the Wind Turbines - Impact of pulsating Bravo Dome compressor motors on the wind turbines - Harmonic Studies - Impact of Closing the Gladstone to Walsenburg 230 kV line

7.1 Model Development All models were developed in PSCAD V4.0.3 (the latest release at the time of this study). The system representation is shown in Figure 5 (most objects in the single line diagram (SLD) in Figure 5 are page components, with detailed models on sub-pages). The E-TRAN program is used to translate loadflow information into PSCAD and to initialize loads, transformer tap settings and to set machine initial conditions. Most of the system 3 or 4 busses away from the study area is represented in detail, with multi-port system equivalents being calculated at San Juan 345 kV, Comanche 230 kV, Storrie, Walsenburg and Hernandez 115 kV busses. E-TRAN calculates system equivalents by re-creating the admittance matrix from the loadflow information and uses network reduction techniques to reduce the network to a size manageable in PSCAD. Data from the dynamics file (machine reactances) was imported and used to update the loadflow generator impedances so that accurate system equivalents can be created. In addition:

- All transformers are represented using detailed 2 or 3 winding transformer models, including wye-delta transformations, neutral grounding, tap settings and non-linear saturation.

- The 115 kV transmission system from Rosebud back to Springer is represented with detailed phase-domain frequency dependent line models, whereas remaining transmission lines are represented based on loadflow RXB data (traveling wave Bergeron models for longer lines and pi-sections for short lines).

- The SVC at Clapham is modeled with a thyristor switching model of the TCR. The model was developed for a different project [2]. Although in the


ELECTRANIX 15

May 31st, 2004

future there will be 2 - 25 MVAR SVC units, only a 25 MVAR single unit was included in this study.

- The compressor motors at Rosebud and York Canyon were also modeled in a previous study [2] and re-used here. They are represented with DQ0 machines, and include the multi-mass and piston torsional effects. The models were validated against site measurements.

- The wind farm units are modeled using the latest PSCAD model from GE Wind, with a recommended number of units lumped together as per Section 2.2.

SAN_JUAN

N10292 T-Line

E_10292_10232_1

E

I

OJO N10232

T_10232_12050_1

345 115

OJO 345/13.8/115

OJO N12050 T-Line

E_12050_12038_1

E

HERNANDZN12038 T-Line

E_12038_12081_1

E

TAOS N12081 T-Line

E_12081_12011_1

E

BLACKLAKN12011

T_12011_12010_1

115 69

Blacklak 115/69

BLACKLAKN12010 P,Q

Load6.98

E2.294

T-Line

E_12011_12077_1

E

SPRINGERN12077 T-Line

E_12077_12020_1

E

CLAPHAM N12020

VN12020

18.425 kmB_12020_10394_1

FORSIGHTN10394 18.425 km

B_10394_12062_1

ROSEBUD N12062

T_12062_12063_1

115 13 .8

Rosebud 115/13.8

ROSEBUD N12063

VN12063Shunt

C0.0

E-12.5SVC Cap

CLAPHAM

SVC

G_12063_0_s1Bravo Dome Compress or 2 (6 units)

G_12063_0_s2Bravo Dome Compress or 1 (3 units)

T_10394_12105_1

115 34.5

GEWind 115/tertiary/34.5

FORFEED N12105 <-- 2.07 [km] -->

T-LineE_12105_12106_1

E

FORFEED1N12106

T_12106_12107_1

34.5 .575

Forfeed1 34.5/.575

FORWTG1 N12107

VN12107

G_12107_0_1Forwtg1 (15 x GE 1.5 Wind)

<-- 2.79 [km] -->T-Line

E_12105_12108_1E

FORFEED2

N12108

T_12108_12109_1

34.5 .575

Forfeed2 34.5/.575

FORWTG2

N12109

VN12109

G_12109_0_1Forwtg2 (13 x GE 1.5 Wind)

<-- 0.74 [km] -->T-Line

E_12105_12110_1E

FORFEED3

N12110

T_12110_12111_1

34.5 .575

Forfeed3 34.5/.575

FORWTG3

N12111

VN12111

G_12111_0_1Forwtg3 (12 x GE 1.5 Wind)

<-- 4.95 [km] -->T-Line

E_12105_12112_1E

FORFEED4

N12112

T_12112_12113_1

34.5 .575

Forfeed4 34.5/.575

FORWTG4

N12113

VN12113

G_12113_0_1Forwtg4 (13 x GE 1.5 Wind)

<-- 2.41 [km] -->T-Line

E_12105_12114_1E

FORFEED5N12114

T_12114_12115_1

34.5 .575

Forfeed5 34.5/.575

FORWTG5 N12115

VN12115

G_12115_0_1Forwtg5 (13 x GE 1.5 Wind)

<-- 7.27 [km] -->T-Line

E_12105_12116_1

E

FORFEED6N12116

T_12116_12117_1

34.5 .575

Forfeed6 34.5/.575

FORWTG6 N12117

VN12117

G_12117_0_1Forwtg6 (15 x GE 1.5 Wind)

T_12020_12019_1

115 69

Clapham 115/13.8/69

CLAPHAM N12019 T-Line

E_12019_12006_1

E

AMISTAD N12006 P,Q

Load6.64

E2.182

T-LineE_12019_12021_1E

CLAYTON

N12021 T-LineE_12021_10358_1

E

VANBUREN

N1035812.569.0

E

:1

CLAYTON

N10068

P,QLoad

4.05E1.331

P,QLoad

0.726E0.415

4.1669.0

E

:1

T4

N10334 P,QLoad

2.334E-1.037

35.81 kmB_12020_12100_1

GLADSTONN12100

T_12101_12100_1

230115

Gladstone 230/13.8/115

GLADSTONN12101 T-Line

E_12101_70459_1E

WALSENBGN70459 T-Line

E_70459_70122_1E

COMANCHEN70122

72.714 kmB_12077_12100_1

115.0230.0

E

:1

WALSENBGN70458 P,Q

Load7.997

E2.629

T_12077_12076_1

115 69

Springe r 115/13.8/69

SPRINGERN12076 P,Q

Load4.57

E1.502

T-LineE_12077_12079_1E

STORRIE N12079 69.0115.0

E

:1

STORRIE N12078 P,Q

Load5 .38

E1 .768

T-LineE_12077_12091_1E

YORKCANYN12091

T_12091_12090_1

115 69

YorkCany 115/69

YORKCANYN12090 P,Q

Load4 .83

E1 .558

T_12091_12102_1

115 13.8

YorkCany 115/13.8

YORKCANY

N12102

VN12102

G_12102_0_s1York Canyon Compressor 1

~E-7.4

-1.0 E_12102_0_s2

T_12081_12080_1

115 69

Taos 115/69

TAOS N12080 P,Q

Load28.75

E9.45

T_12082_12081_1

345115

Taos 345 /13.8/115

TAOS N12082

T-LineE_10232_12082_1

E

N10292

N12079

N70122

N70458

N12038

Net-EquivSubPage 2

N10292

N12079

N70122

N70458

N12038

Figure 5 – PSCAD System Single Line Diagram

7.2 Short Circuit Current Calculations The wind farm manufacturer (GE Wind) specified that short circuit studies should be performed with a PSCAD simulation model (as opposed to traditional loadflow short circuit calculations). To verify the system representation in PSCAD, short circuit currents at the Customer 34.5 kV bus were calculated using PSLF and were compared to those calculated in the PSCAD model. The base case system (without any wind represented) was used for this test case. The comparative results are summarized in Table 4.

Table 4 – Comparison of Short Circuit Current Calculations from Loadflow Programs Compared to PSCAD

Case Description SCC (3ΦΦΦΦ Short

Circuit Current Level, kA, RMS)

Notes

1. Loadflow (PSLF) 3.69 Traditional SCC Calculations. 2. PSCAD (detailed models not used) 3.69 Good verification of system

equivalents.


ELECTRANIX 16

May 31st, 2004

3. PSCAD (detailed transformers used)

3.78 Some detailed transformer 3 winding data had lower impedances compared to loadflow data.

4. PSCAD (detailed line models added)

3.80 Fundamental frequency impedance/admittance of detailed line models match the loadflow.

5. PSCAD (DQ0 machines added, SCC at first cycle)

3.79 SCC calculations in the loadflow use sub-transient reactance.

5. PSCAD (DQ0 machines added, SCC at 4th cycle)

3.5 SCC levels will be lower at the time of opening breakers due to frequency dependent effects in machines (ie XD’’, XD’…)

6. PSCAD (detailed SVC replaces SVC modeled as a generator)

3.2 In the previous cases, the SVC is represented as a generator, with an impedance that causes the SCC level to be higher.

The test cases show that the approach using the PSLF program for short circuit analysis will tend to exaggerate the short circuit levels. In particular:

- The short current levels will be lower 3 or 4 cycles into the fault (when the breakers have to interrupt the fault current) compared to the first cycle current levels. The loadflow program often uses subtransient reactance (Xd’’) as a machine impedance, which is analogous to taking the first cycle fault current.

- The SVC will add short circuit current contribution if represented as a generator with 0 MW output). The real SVC does not contribute to the short circuit current.

The PSCAD result is a time domain transient simulation program, so there will also be a decaying dc offset in the fault current waveforms. The fault currents are computed by taking the currents in the first cycle of the fault (maximum current minus the minimum current) /2, and then converted to RMS. A typical short circuit current waveform is shown in Figure 6.


ELECTRANIX 17

May 31st, 2004

Case SC6 showing decaying DC offset

0.400 0.450 0.500 0.550 0.600 0.650 0.700 0.750 0.800 ... ... ...

-8.0

-6.0

-4.0

-2.0

0.0

2.0

4.0

y

Ic

Figure 6: Short Circuit Current (SCC) showing decaying dc offset

An additional test case was performed to approximate the short circuit impedance of the GE 1.5 MW wind turbines. The PSCAD model of the turbine was run feeding a 1 pu resistance on the .575 kV terminals of the turbine. The unit was then faulted (0 resistance 3 phase fault) and the fault current determined. From the first cycle fault current (averaged over 3 phases), the short circuit impedance is approximately 0.4 pu on a 0.575 kV and 1.5 MVA base. If a fault is prolonged, the Rosebud compressor motors will slow down and the fault currents can exhibit pole-slipping beating characteristics. In reality these motors would trip on under-frequency for prolonged faults. Table 5 shows the short circuit current results from the full PSCAD model with the dc offset ignored:


ELECTRANIX 18

May 31st, 2004

Table 5: AC Component of Fault Current (kA, RMS)

(120MW Generation at TI-04-1214 Wind Ranch)

Case Fault Bus Fault Description SC Current (1st Cycle)

SC Current (3rd Cycle)

SC Current (6th Cycle)

SC1 12062 Rosebud 115kV 2.22 1.70 0.82 SC2 12063 Rosebud 13.8kV 14.90 11.37 4.24 SC3 10394 Customer 115kV 2.65 1.82 0.92 SC4 12105 **Customer 34.5kV 8.31 3.34 4.69 SC5 12020 Clapham 115kV 2.96 2.17 1.31 SC6 12019 Clapham 69kV 2.38 2.34 2.05 SC7 12101 Gladstone 230kV 2.05 1.58 1.43 SC8 12100 Gladstone 115kV 4.04 3.25 2.52 SC9 12077 Springer 115kV 3.50 3.11 2.82

SC10 12076 ***Springer 69kV 1.77 1.74 1.74

** In some cases, beat frequencies due to machine speed changes cause the current to increase from the third cycle to the sixth cycles, or cause the two measurements to be very different. *** Due to conflicting transformer impedance data, the load flow data was used for the Springer transformer for case SC10. This caused the short circuit current on this bus to rise slightly. 7.3 Measurement of Negative (-ve) Sequence Voltage Levels The PSCAD model was run in steady state at 0, 40, 80 and 120 MW wind levels. The instantaneous A, B and C voltages were recorded, processed by Fourier analysis, and then expressed in terms of fundamental frequency sequence components. The –ve sequence compensation in the SVC controls was turned off. The –ve sequence is predominantly caused by the use of non-transposed transmission lines through 115 kV system. The results are shown in Table 6. Although the wind farm increases the –ve sequence unbalance, levels are quite small. The positive (+ve) sequence voltages are controlled (either by the SVC or by the excitation system on the compressor motors) so are constant.


ELECTRANIX 19

May 31st, 2004

Table 6: Measurement of Voltage Sequence Components (pu)

Clapham

(SVC 115 kV Bus) Rosebud

(Compressor 13.8 kV Bus) Case +ve Seq -ve Seq 0 Seq +ve Seq -ve Seq 0 Seq

1 (0 MW Wind) 1.02 0.0023 0.00019 1.02 0.0036 0 2 (40 MW Wind) 1.02 0.0038 0.00046 1.02 0.0026 0 3 (80 MW Wind) 1.02 0.0058 0.00058 1.02 0.004 0 4 (120 MW Wind) 1.02 0.0074 0.0006 1.02 0.006 0 Note that the loads in the PSCAD model are symmetrical on all 3 phases, whereas the loads in the real system can be unbalanced and will add to the levels shown in Table 6. 7.4 Interactions between the Clapham SVC and the Wind Turbines The PSCAD models were run over a 10 second period to look for interactions between the SVC and the Wind Turbines units. In all cases (0, 40, 80 and 120 MW of wind) there were no apparent oscillations or control interactions. The SVC controls use Proportional-Integral controllers, but fortunately they are considered with a droop characteristic that prevents multiple operating points and minimizes hunting problems with other nearby fast-acting control devices such as the TI-04-1214 WTGs. 7.5 Impact of Pulsating Bravo Dome Compressor Motors on the Wind

Turbines The compressor motors at Bravo Dome (Rosebud) and York Canyon are piston driven, which generate pulsations and flicker into the electrical network. The concern is that these frequencies can coincide with mechanical modes of resonances on the wind turbine blades and shaft system. The GE wind model in PSCAD does not include details on the mechanical modes of oscillation. To address this concern, the model was run in steady state and the instantaneous electrical power into a wind turbine unit was calculated. The frequencies that can appear on the shaft (due to the electrical system) were then computed by Fourier analysis. The Fourier results are summarized in Table 7. The only significant interactions between the electrical grid and the wind turbine occurred near 5 Hz and at the 2nd harmonic (120 Hz). The 2nd harmonic component of electrical power is due to the presence of –ve sequence unbalanced voltages and currents (the –ve sequence transforms into the 2nd harmonic on the electrical torque of the machine). The 5 Hz component was much smaller (.16%) whereas the 2nd harmonic component was larger (up to .96%). From system measurements and calculations, it is known that the mechanical modes of resonances on the compressor motors at Bravo Dome are 5.025 Hz and 5.4054


ELECTRANIX 20

May 31st, 2004

Hz. The Fourier analysis (reported in Table 7) was performed over a 1 second steady state period (which results in a frequency resolution of 1 Hz), so these two frequencies combine into the 5 Hz value. To provide more details, Case 4 was run for 100 seconds in steady state (which results in 0.01 Hz resolution in the frequency domain), and the Fourier analysis repeated. This time the electrical power showed clear 5 Hz and 5.45 Hz components (5 Hz: 0.001258 pu, 5.45 Hz: 0.0002174 pu, 120 Hz, 0.009649 pu). It was observed that the 5 Hz component was larger then the 5.45 Hz component.

Table 7: Effect of Pulsating Motor Load on Wind Turbines (WTG #1)

Case

Harmonic Frequencies Present on Wind Turbine Electrical Power (pu on Total WTG Rated Power (N*1.5MW)) (FFT Scan of Instantaneous Electrical Power 1 Hz to 10 kHz)

2 (40 MW Wind) 5 Hz: 0.00163 | 120 Hz: 0.0035 3 (80 MW Wind) 5 Hz: 0.00154 | 120 Hz: 0.0066 4 (120 MW Wind) 5 Hz: 0.00149 | 120 Hz: 0.0096 Note: FFT (Fourier analysis) performed over 1 second period, so frequency resolution is 1 Hz. The electrical power that appears at the WTG terminals will result in an electrical torque with similar frequency components. The frequencies on the electrical power match those of the Rosebud motors (5.025 Hz and 5.4054 Hz). 7.6 Harmonic Studies The GE wind model for PSCAD is not a detailed switching based model (it is possible to represent the Voltage Sourced Converter used in the double wound generator and thereby model the switching of the electronic devices) so it cannot inherently reproduce the level of harmonics in the system. However, GE provided a table of measurements of voltage and current harmonic magnitudes. This information can be used to estimate how the addition of the wind turbine units will affect the Total Harmonic Distortion (THD) levels in the nearby system. To study this, it was assumed that the harmonics that appear on the 115 kV system at Customer substation can be represented as a current injection onto the 115 kV bus (this is reasonable because these harmonic currents must pass through the inductance of the 115/34.5 kV transformers and 34.5 kV unit transformers, thus approximating a current source). The harmonic currents (measured from the 575 volt terminals of a single GE wind turbine) were output to a text file, and then read in by a custom-developed component in PSCAD. The harmonic currents on each phase were then re-constructed (essentially an inverse Fourier transform) and injected onto the 115 kV system (scaling factors were applied for the number of wind turbines in service and the different voltage level). A validation test of the new model was made by performing a Fourier analysis from the time domain current injections and comparing them to the original measured harmonics in the frequency domain (with excellent results).


ELECTRANIX 21

May 31st, 2004

The resulting THD levels are shown in Table 8 and graphically in Figure 7 for varying levels of wind output.

Table 8: THD Level of Customer 115 kV Bus (%) Wind Farm Harmonics Added Wind Farm Harmonics Removed Wind MW Phase A Phase B Phase C Phase A Phase B Phase C

0 1.476 1.404 1.435 1.476 1.404 1.435 40 1.327 1.26 1.291 1.321 1.258 1.288 80 1.278 1.212 1.239 1.247 1.196 1.223 120 1.316 1.257 1.254 1.229 1.19 1.218

Maximum Individual Harmonic: 1.064 (phase B of 0 MW case, 65 Hz) (all other cases had all individual harmonics below 1%)

THD Levels at Foresight 115 kV

1

1.1

1.2

1.3

1.4

1.5

1.6

0 40 80 120

Wind MW

TH

D (

%)

Phase A (without wind harmonics)

Phase B (without wind harmonics)

Phase C (without wind harmonics)

Phase A (with wind harmonics)

Phase B (with wind harmonics)

Phase C (with wind harmonics)

Figure 7: THD Level of Customer 115 kV Bus (%)

It is interesting to note that the largest single component of frequency in these cases was a 65 Hz component (+/- 1 Hz), which was 1.064%. This occurs because of the pulsation effect of the compressor motors (the 5 and 5.4 Hz mechanical mode of oscillation appears on the electrical grid as 65 and 65.4 Hz as seen on the instantaneous voltage waveshape, and which reflects through in the RMS measurement as an approximate 5 Hz component). This 65/5 Hz component of 115 kV voltage can possibly be minimized with flicker reduction controls on the SVC (this was not represented in the SVC model).


ELECTRANIX 22

May 31st, 2004

The THD levels increase as wind power levels increase, however this is compared to a “theoretical” wind generator that does not produce harmonics (ie the table shows the results with the wind turbine harmonic currents on and off). The results also show that the base case (without any wind production) has higher THD levels. The representation of the wind turbine harmonics as a current source (as seen from the 115 kV side) is an approximation. If THD levels are at high levels, a more detailed GE wind model for PSCAD is required which represents the VSC as switching devices. Another test was done using linear harmonic impedance scanning… A graph showing the harmonic impedance as a function of frequency (as seen from the Customer 115 kV bus) is shown in Figure 8. In this case, 120 MW of GE wind turbines are in service and are approximated with a 0.4 pu inductive source impedance.

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 300

200

400

600

800

1000

0 Sequence+ve Sequence

+ve and 0 Seq Harmonic Impedance

Harmonic #

Oh

ms

Figure 8: Harmonic Impedance at Customer 115 kV Bus

(120 MW Wind Case 4) 7.7 Impact of Closing the Gladstone to Walsenburg 230 kV line When the new 230 kV line from Gladstone to Walsenburg is out of service, a large phase shift will exist between the systems on either side of the breaker (from PSLF loadflow studies, the phase shift was more then 35 degrees in the base case). The concern was that closing the line breaker could cause a large transient effect on the compressor motors at Rosebud or on the Wind Farm units.


ELECTRANIX 23

May 31st, 2004

To investigate this, loadflow cases were run with this line open at Gladstone but energized at Walsenburg. The case was then translated to PSCAD (using E-TRAN) and multiple runs were performed which closed the breaker sequentially every 20 degrees of one cycle (all three phases closed at the same time). The over-frequency and under-frequency protection systems of the compressor motors were modeled (the GE Wind model already had protection systems included in the model). The state of each compressor motor (Rosebud 1, Rosebud 2 and York Canyon) as well as each wind turbine unit was monitored to look for trips due to the breaker closing. The breaker closing times were chosen to occur at the maximum frequency of the smallest Rosebud compressor motor (the steady state motor frequency will change with a 5 Hz and 5.4 Hz component due to the piston pulsating effect) and then varied every 20 degrees over a full cycle (19 runs). This was repeated for the minimum frequency swing, and was repeated for all 4 wind turbine output levels (0, 40, 80 and 120 MW). In all cases the compressor motors and the wind turbine units did not trip. However, it is conceivable that a larger phase angle across the contacts of the open circuit breaker at Gladstone could occur for a particular power flow condition, and a greater deviation of frequency of the compressor motors will result when the circuit breaker is closed. The open circuit phase angles observed in the models for the respective power flow conditions for this study were as shown in Table 9: Table 9: Phase angles across open circuit breaker at Gladstone for different

generation levels at the TI-04-1214 Wind Ranch. Power Generated at the TI-04-1214 Wind Ranch

0 MW 40 MW 80 MW 120 MW Phase angle across the open Gladstone 230 kV breaker

55O 40O 28O 17O

The increase in generated power from the TI-04-1214 Wind Ranch reduces the phase angle across the open Gladstone 230 kV circuit breaker, and lessens the impact on the frequency swing of the Bravo Dome compressor motors when the circuit breaker is closed. It is noteworthy that site measurements (and PSCAD models) indicate the compressor motor frequencies change significantly due to the pulsating piston effect. Fault analysis is typically conducted in transient stability algorithms, but the pulsation effect of the compressor motors is not included in the WECC models. If future motor-trip protection studies are to be undertaken, the steady state pulsating effects of the motors should be included (either in transient stability models or in PSCAD).


ELECTRANIX 24

May 31st, 2004

8 CONCLUSIONS For the operating condition where power generated by the TI-04-1214 Wind Ranch, acceptable steady state performance during contingencies for the three acceptance criteria of line/transformer overload, undervoltage deviations and maximum/minimum bus voltage limits was generally achieved in the NEA for the operation of the wind ranch up to 120 MW. No violations of the WECC/Tri-State operational standards were observed except for under voltage deviations exceeding 5% when the wind power from TI-04-1214 was both 80 and 120 MW for Contingency No. 5, outage of the Black Lake -Taos 115 kV line and for Contingency No. 8, outage of the Colinas – Rowe Tap 115 kV line. For the operating condition where designated load at or around Hernandez substation is additional to the normal projected loads for the area, and is to be supplied from the wind ranch without backing off scheduled generation, then major additional facilities are required in the Taos – Ojo – Hernandez triangle for a rating of the TI-04-1214 Wind Ranch exceeding approximately 40 MW. An increase in wind ranch rating above 40 MW would be possible with increased investment in facilities that have yet to be defined in order to comply with acceptable steady state performance during contingencies. The undervoltage deviations were more severe when the Bravo Dome compressor motors were out of service and all the power generated from TI-04-1214 wind had to flow west instead of some of it displacing power to the compressors. Thus when the Bravo Dome compressors were out of service, Contingencies No. 5 and 8 result in more severe violations than when the compressors were in service. In addition, undervoltage deviations exceeding 5% were observed at the Taos 115 kV busbar for outage of the Ojo – Taos 345 kV line (Contingency 12) and for the outage of the Taos 345/115 kV transformer (Contingency 15) when the TI-04-1214 was generating at 120 MW. Additionally, a marginal undervoltage condition was observed at the York Canyon 115kV bus during the base (n-0) case while 120MW was being generated at the TI-04-1214 site. The transient stability analysis indicated that during a three phase fault and loss of the proposed Gladstone to Walsenburg 230kV line, the compressor motors at Bravo Dome would trip due to underfrequency for cases with 40 MW of wind generation or less. During this contingency (Case2, fault S3), the 40 MW of TI-04-1214 wind turbines also tripped due to undervoltage at the generator busses. For 80 MW or more, both the load at Bravo Dome and the the new wind turbines were able to remain in service. This indicates that the addition of large amounts of wind generation in the vicinity of the Bravo Dome load provides system support in the region by offsetting the required power flow through the radial lines to the motors. PSCAD studies were performed to determine short circuit current levels, determine the level of –ve sequence imbalance, study wind turbine and SVC interactions,


ELECTRANIX 25

May 31st, 2004

determine the effect of the pulsations from compressor motors on the wind turbines, look for amplification of wind turbine harmonics and to study the transient impact of closing the breaker on the proposed Gladstone to Walsenburg 230 kV line. From steady state, short circuit, PSCAD electromagnetic transients and also transient stability analysis considerations, the operation of the TI-04-1214 Wind Ranch up to 120 MW still appears a realistic objective. The facilities required to achieve operation of the wind ranch at 120 MW are minimal. These include replacing the current transformers on the 115 kV transmission lines that are presently limiting and to identify if conductors on the 115 kV transmission have to be raised. The facilities study will address these issues.

9 REFERENCES [1] Frank R. McElvain, P.E., “TI-04-1214 Wind Project – System Impact Study

Report, “ prepared by Tri-State Generation and Transmission Association, Inc. for Customer Energy Company, December 2002.

[2] “Plains Electric Static VAR Generator: Transient Studies”:, Report No. AST-

83-1771-REV, November 29, 1983, By Advanced Systems Technology. [3] www.PSCAD.com

10 ACKNOWLEDGEMENTS Appreciation is expressed for the guidance and information provided by Frank McElvain of Tri-State for this study.

http://www.pscad.com/

system impact study final report · system impact study – final report electranix ii may 31 st,...

Documents