Legacy Turbine Forum, No. 8 in a series
By Luke Williams, PE, Consultant
www.geLegacyGasTurbineSupport.com
Theme and scope
More gas-turbine users may soon face the problem that landed on one Northeast plant’s desk: available fuel-gas pressure was going down, but the operating obligation was not. The supplier said line pressure would be reduced from 320 psig to 250 psig because of a local natural-gas storage failure. The customer still had to maintain dual-fuel capability under its contract.
The first reaction was predictable. Find a gas compressor and restore pressure to the level assumed in the control specification. But the unit, a dual-fuel GE MS6001B with Mark IV controls, already had fuel-transfer logic designed to move the machine from gas to liquid fuel on a loss of gas pressure. The question became whether that existing system, with revised control settings and proper testing, could solve the immediate problem without adding compression equipment.
Issue clusters
The compressor search did not look promising. Equipment for this application was not readily available, and the cost was high. More important, the proposed compressor was being considered to recover a pressure margin that might not be required for successful transfer if the control logic and fuel system response were understood in detail.
Under the existing logic, a transfer from gas to liquid fuel is initiated when fuel-gas pressure at the gas compartment falls to 265 psig. In a normal transfer, the logic holds for 30 seconds at 0.5 FSR to fill the liquid-fuel piping and check for liquid-system position servo faults. When that timer is complete, gas FSR transfers to liquid FSR at 3.3 FSR/sec. At base load, with FSR at 70, the transfer requires about 21 seconds.
The control specification minimum gas-fuel supply pressure was 270 psig. That value should not be treated as a number floating by itself. It is tied to the amount of gas in the piping downstream of a pressure-loss event, the load on the unit when transfer begins, the time needed to complete the transfer, and the likely cause of pressure loss, such as a piping leak, upstream valve failure, instrument failure, or curtailment by the supplier. The plant’s history of gas-pressure-loss transfers and the consequence of failing to complete a transfer before flame loss also belong in the review.
On a loss-of-gas-pressure transfer, the 30-second fill period is bypassed and transfer starts immediately. As supply pressure falls, the gas ratio valve opens to maintain P2 pressure at the gas control valve. The control algorithm calculates P2 as FPKGNG multiplied by TNH, minus FPKGNO. At 100% TNH, the existing P2 value was 271.6 psig, essentially the current minimum supply pressure. During transfer, the ratio valve may move to 100% position to maintain P2. The gas control valve also continues opening to hold temperature control at base load.
If the gas control valve reaches 100%, load will begin to decrease. The practical question is whether the transfer can finish before flame is lost. Ten seconds into the transfer, fuel mix is roughly 50% gas and 50% liquid. Increasing liquid-fuel flow helps offset the load loss caused by falling gas pressure. At part load in droop control, FSR is fixed by the 4% droop setting. If load starts to decrease and TNH falls, speed control sees the change and increases FSR to maintain TNH. The protection logic allows both ratio and control valves to operate at 100% position because 100% position is not treated as a fault alarm.
Contrasting viewpoints
The conservative answer is to buy compression and restore the supply pressure assumed by the original specification. That answer has merit when the unit cannot tolerate uncertainty, when transfer history is poor, or when testing cannot be performed safely. It also may be the only answer if the lower supply pressure cannot support normal gas operation without driving the ratio valve into saturation.
The other answer is to verify what the machine can actually do. In this case, fuel-nozzle and combustor pressure relationships mattered. Expected pressure ratio across the liner was about 1.05, and fuel-nozzle design had a minimum ratio of 1.15. On a cold day, compressor discharge pressure could reach 170 psig. With the liner ratio, liner pressure would be about 162 psig, and gas-nozzle pressure would need to be about 170 psig to make the required ratio. Those values supported a closer look at revised gas-pressure transfer settings rather than an immediate compressor purchase.
Synthesized guidance
The site had never experienced an emergency transfer to liquid fuel. Given the transfer logic, valve behavior, fuel mix during transfer, droop response, and pressure-ratio review, reducing the transfer pressure from 265 psig was judged reasonable. Because supply pressure would drop to 250 psig, the P2 control curve also had to be changed so the ratio valve would not saturate during normal operation. For a P2 pressure below 250 psig, the 100% TNH setpoint constant FPKGNG was set to 0.0452 V/% for a P2 pressure of 240 psig.
The revised transfer pressure was set at 220 psig and tested. The unit completed transfer successfully in the expected 21-second time frame. That result does not make 220 psig a fleet recommendation. It does show that, for this machine and this set of controls, the existing dual-fuel transfer capability provided a workable path around a lower gas-supply pressure.
Users facing curtailment or supplier-driven pressure reductions should not start with the purchase order. Start with the control specification, fuel-transfer sequence, P2 curve, valve limits, expected CPD, nozzle pressure ratio, droop response, and actual site history. Then test the revised setting under controlled conditions. For some legacy dual-fuel machines, the answer may already be in the Mark IV logic. CCJ
