Sacramento Municipal Utility District (SMUD) | Solano 1 & 2 VAR-Control Load Flow Study at Russell Substation | Montezuma Hills, California
SMUD’s Solano Wind Project Phases 1 and 2 – 23 Vestas V47 turbines at 660 kW and 29 Vestas V90 turbines at 3,000 kW – deliver power through a 21 kV collection system to Russell Substation, where the 230:21 kV transformer ‘T2’ connects to the transmission system. To hold the 1.0-to-0.95 leading power factor SMUD must maintain at the 230 kV side of T2, and so avoid reactive-power usage fees, a two-stage 21 kV capacitor bank had been installed at the T2 bus.
That capacitor bank created a new problem. Switching it on drove the voltage across the 21 kV collection system high enough to trip the high-voltage alarms at the V90 turbines – so, in practice, SMUD could not fully use the very equipment it had installed for compliance. EETS was engaged to build a full load-flow model of the system and find a way to operate the capacitor bank without alarming the turbines.
The study’s answer was a single, targeted tap change at the turbines – arrived at only after modeling showed that the intuitive fix at the main transformer would have solved one turbine’s problem by creating another’s.
The capacitor bank is a two-stage, switched 21 kV installation of two 10.8 MVAR steps – 21.6 MVAR in total – that compensates for transformer T2’s reactive demand. The difficulty is voltage. The incoming 230 kV supply is frequently well above nominal, with levels of 242 to 244 kV recorded; T2’s no-load tap changer sits at 235.75 kV; and switching capacitive VARs onto the 21 kV bus raises the collection-system voltage further still. The V90 turbines, which generate at 1,000 volts through a 3,160 kVA nacelle transformer, alarm at 10% above nominal – and the modeling showed how little margin remained: with the T2 transformer at its present tap, the worst-case V90 sat at 23,758 volts against a 23,760-volt alarm threshold, just two volts short of tripping.
Complicating the picture, the two turbine types do not share the same voltage window. The V90 alarms symmetrically at ±10% of its nominal, while the V47 turbines, served by padmount transformers, alarm at +10% and −6% of their 690-volt rating. Tap adjustments are available at three levels – the main T2 transformer, the V90 nacelle transformers, and the V47 padmounts – so any change made to relieve one turbine type ripples through to the other.
The capacitor bank is not optional: without it, SMUD cannot hold the leading power factor its interconnection requires, and incurs reactive-power fees. Yet putting the capacitors into service over-voltages the collection system and trips the V90 turbines offline on high-voltage alarms. SMUD was effectively unable to use equipment it had already installed for compliance. Any solution had to preserve full use of both capacitor stages, not merely tolerate the bank at partial output.
The margin is squeezed from both directions and across a moving target. The incoming supply swings up to 242–244 kV, so high-voltage headroom at the V90s is thin; but pushing the voltage down to protect them threatens to drop the V47s below their low-voltage alarm, and the V47 window is asymmetric, allowing less undervoltage than overvoltage. The right answer had to keep both turbine types out of alarm – high and low – across the full incoming-voltage range and with zero, one, or two capacitor stages online, all at once.
Sacramento Municipal Utility District (SMUD)
Public / Municipal Utility – Wind Generation
Solano Wind Farm, Montezuma Hills, California
Load Flow Modeling │ Voltage / VAR Control Study │ Transformer Tap Coordination │ Collection System Analysis │ Interconnection Compliance
As part of this expansion, AWA identified an opportunity to recover energy that was previously being wasted.
Sacramento Municipal Utility District (SMUD)
Public / Municipal Utility – Wind Generation
Solano Wind Farm, Montezuma Hills, California
Load Flow Modeling │ Voltage / VAR Control Study │ Transformer Tap Coordination │ Collection System Analysis │ Interconnection Compliance
As part of this expansion, AWA identified an opportunity to recover energy that was previously being wasted.
SMUD’s Solano Wind Project Phases 1 and 2 – 23 Vestas V47 turbines at 660 kW and 29 Vestas V90 turbines at 3,000 kW – deliver power through a 21 kV collection system to Russell Substation, where the 230:21 kV transformer ‘T2’ connects to the transmission system. To hold the 1.0-to-0.95 leading power factor SMUD must maintain at the 230 kV side of T2, and so avoid reactive-power usage fees, a two-stage 21 kV capacitor bank had been installed at the T2 bus.
That capacitor bank created a new problem. Switching it on drove the voltage across the 21 kV collection system high enough to trip the high-voltage alarms at the V90 turbines – so, in practice, SMUD could not fully use the very equipment it had installed for compliance. EETS was engaged to build a full load-flow model of the system and find a way to operate the capacitor bank without alarming the turbines.
The study’s answer was a single, targeted tap change at the turbines – arrived at only after modeling showed that the intuitive fix at the main transformer would have solved one turbine’s problem by creating another’s.
The capacitor bank is a two-stage, switched 21 kV installation of two 10.8 MVAR steps – 21.6 MVAR in total – that compensates for transformer T2’s reactive demand. The difficulty is voltage. The incoming 230 kV supply is frequently well above nominal, with levels of 242 to 244 kV recorded; T2’s no-load tap changer sits at 235.75 kV; and switching capacitive VARs onto the 21 kV bus raises the collection-system voltage further still. The V90 turbines, which generate at 1,000 volts through a 3,160 kVA nacelle transformer, alarm at 10% above nominal – and the modeling showed how little margin remained: with the T2 transformer at its present tap, the worst-case V90 sat at 23,758 volts against a 23,760-volt alarm threshold, just two volts short of tripping.
Complicating the picture, the two turbine types do not share the same voltage window. The V90 alarms symmetrically at ±10% of its nominal, while the V47 turbines, served by padmount transformers, alarm at +10% and −6% of their 690-volt rating. Tap adjustments are available at three levels – the main T2 transformer, the V90 nacelle transformers, and the V47 padmounts – so any change made to relieve one turbine type ripples through to the other.
The capacitor bank is not optional: without it, SMUD cannot hold the leading power factor its interconnection requires, and incurs reactive-power fees. Yet putting the capacitors into service over-voltages the collection system and trips the V90 turbines offline on high-voltage alarms. SMUD was effectively unable to use equipment it had already installed for compliance. Any solution had to preserve full use of both capacitor stages, not merely tolerate the bank at partial output.
The margin is squeezed from both directions and across a moving target. The incoming supply swings up to 242–244 kV, so high-voltage headroom at the V90s is thin; but pushing the voltage down to protect them threatens to drop the V47s below their low-voltage alarm, and the V47 window is asymmetric, allowing less undervoltage than overvoltage. The right answer had to keep both turbine types out of alarm – high and low – across the full incoming-voltage range and with zero, one, or two capacitor stages online, all at once.
EETS built a full load-flow model spanning the individual turbines, the 21 kV collection system, the T2 bus, the 230:21 kV T2 transformer with all of its no-load taps, and the two-stage capacitor bank, updating the V90 and V47 turbine models EETS had developed for SMUD in an earlier wind-project study. It ran a large set of load-flow cases – more than twenty – varying the incoming 230 kV voltage, the T2 tap setting, the wind-farm generation level, and the number of capacitor stages switched on, and in each case tracked the highest- and lowest-voltage turbine of each type against its alarm thresholds. That sensitivity analysis first confirmed the field experience: with both capacitor stages on and incoming voltage approaching 242 kV, the V90 turbines go into high-voltage alarm.
The intuitive remedy is to raise the main T2 no-load tap one step, to 241,500 volts, which lowers the 21 kV secondary voltage. Modeling confirmed it works – for the V90s: it permits both capacitor stages and eliminates every V90 high-voltage alarm across the 230-to-242 kV incoming range. But it pushes the V47 turbines into low-voltage alarm when the incoming voltage is at 230 kV. Solving the problem at the main transformer did not remove it; it simply relocated it from the V90s to the V47s.
Instead, EETS recommended tapping up only the V90 nacelle transformers by a single step, from 21,600 to 22,140 volts. Because the change acts only on the V90 side of the system, it lowers the V90s’ own voltage without disturbing the V47s. The modeling verified that this permits full use of both capacitor stages, clears the V90 high-voltage alarms, and – checked for both high- and low-voltage alarms on both turbine types, across the full 230-to-242 kV incoming range and with zero, one, or two capacitor stages online – introduces no new alarm anywhere. A second tap step, to 22,680 volts, offered no further benefit. EETS advised retapping first the V90 units that had actually experienced alarms, since they see the highest voltage, and leaving the main T2 tap, the V47 padmount taps, and the V90 0.98-leading power-factor control unchanged.
Parameter | Detail |
Wind Farm | Solano Phases 1 & 2 – 23 × 660 kW Vestas V47 (induction generator) and 29 × 3,000 kW Vestas V90 (double-wound induction generator) |
Collection / Feeders | 21 kV collection to Russell Substation, circuits #5–#8 (15.18 / 27 / 30 / 30 MW) |
Main Transformer | ‘T2’ 230:21 kV, 60/80/112 MVA; no-load taps 218.5–241.5 kV; set at 235.75 kV |
Capacitor Bank | 21 kV, two-stage switched – two 10.8 MVAR steps (21.6 MVAR total) |
Driver | Maintain 1.0–0.95 leading power factor at the 230 kV side of T2 to avoid reactive-power usage fees |
Problem | Switching the capacitors on over-voltages the 21 kV system and trips V90 high-voltage alarms (V90 alarm at +10% of 1,000 V; worst case 23,758 V vs. 23,760 V threshold at present tap) |
Incoming Voltage | 230 kV nominal, frequently high – 242–244 kV recorded; studied over 230–242 kV |
Turbine Step-ups | V90 nacelle 3,160 kVA, 21.6 kV:1,000 V (±5% taps); V47 padmount 21.6 kV:690/384 V, 660–690 kVA (±5% taps) |
Rejected Fix | Raising the T2 tap to 241.5 kV clears the V90 alarms but causes V47 low-voltage alarms at 230 kV incoming |
Recommended Fix | Tap up the V90 nacelle transformers one step (21,600 → 22,140 V) – no new alarms across 230–242 kV with 0, 1, or 2 capacitor stages |
Left Unchanged | T2 tap (235.75 kV); V47 padmount taps; V90 power-factor control at 0.98 lead |
EETS resolved the conflict between the capacitor bank and the turbines. Modeling confirmed that switching in the 21 kV capacitors over-voltages the collection system and trips the V90 turbines’ high-voltage alarms – worst near the 242 kV top of the incoming range, and within two volts of alarm even at the present transformer tap. The obvious remedy, retapping the main T2 transformer, merely traded V90 high-voltage alarms for V47 low-voltage alarms, so EETS instead recommended a one-step tap-up of only the V90 nacelle transformers, which lets both capacitor stages run across the full 230-to-242 kV incoming range with no alarms on either turbine type – a targeted setpoint change requiring no new equipment. With the main transformer tap, the V47 padmounts, and the V90 power-factor control all left as they were, SMUD can now operate its capacitor bank fully, holding its interconnection power factor and avoiding reactive-power fees.
EETS made installed equipment usable, found the fix that solved the problem instead of moving it, and did it at no capital cost.
The capacitor bank was already in place for compliance, but effectively unusable because switching it on tripped the turbines. EETS’s study is what let SMUD actually run it. By finding tap settings under which both capacitor stages can stay online without alarming the turbines, EETS unlocked the reactive-power-fee savings the bank had been installed to capture – turning stranded equipment into a working asset.
The tempting fix – retapping the main transformer – clears the V90 alarms and breaks the V47s. Only by modeling both turbine types across the entire incoming-voltage range and every capacitor-switching state could EETS see that trap and find the one adjustment, at the V90 turbines alone, that satisfies all of them simultaneously. Verifying that the chosen change introduces no new high- or low-voltage alarm anywhere is what separates a real solution from a relocated fault.
The recommendation is a single tap change on the V90 nacelle transformers – no new hardware, and sequenced to begin with the units already seeing the highest voltage – while the rest of the system, including the power-factor control an earlier EETS study had set, is deliberately left untouched. It is the least-intervention answer that fully meets the operating objective, which on an in-service wind farm is precisely the kind of solution worth having.
As part of this expansion, AWA identified an opportunity to recover energy that was previously being wasted.