I am using the publication Current and Prospective Costs of Electricity Generation until 2050 to source ramping constraints for an economic model of three different gas power plants - CCGT, combustion turbine and steam turbine. As I'm not an engineer I have the following questions:

  1. The model is a little simplistic. For that reason, I cannot use three different time and cost parameters for hot, warm and cold start. For a gas power plant with load factors of between 10 and 60%, which start would you think is most relevant (i.e. most often done): hot (shut down 8h or less), warm (8-50h) or (probably not) cold start (more than 50h)? Would you apply something in between hot and warm start?
  2. Regarding start-up time, there are not many studies for steam turbines (p.71 in pdf numbering). Someone advised my to use a rather slower start-up time for steam than for combustion. So could I use a value similar to CCGT, for example?
  3. For the "ramping load gradient limit in %-Pn/min" (p.74 in pdf numbering), I have the same issue with steam turbines as described in 2).

The primary advantage of gas plants is much faster ramping of output than others - they are energy-inefficient by comparison, but they are meant to handle spikes in demand in the grid. All of them will be of hot start (<8h) as even the steam ones use relatively small boilers that are quick to get up to speed, compared to conventional power plants. As result, they are kept on stand-by or operating at very low output except for when the spike needs to be handled.

You can definitely use the longer times found with CCGT when describing ST power plants - although even longer times will be found (gas burners and big boilers are cheaper to build than gas turbines; the CCGT will usually be smaller than ST plant).

I'd be more concerned over the inflated numbers for CCGT; they reach ~40-60% of capacity in time typical to GT plants, and then ramp up the output to 100% over the time typical for ST. That's also why you find the wildly varying values for the gradient in the document - not only is there the slow climb of steam part trailing a long way after the gas part, the power output of the gas turbine depend strongly (and non-linearly) on RPM (which again climbs at rate dependent on that power output) - generally, first 30% or so will be rather slow, then the engine picks up and the 50%-80% ramp is really steep. Regardless, 5-10 minutes is an adequate estimate of the time from ignition to full power (not sure what preparations the operator must perform beforehand). And then water in the steam part only starts heating slowly, steam turbine a long way from budging...

Any kinds of averages or momentary measurements in CCGT are seriously misguiding due to that awfully non-linear output ramp.

Similarly, for GT, the 100% output time and climb rate will depend strongly on each other. If you allow the GT to spin up with coils off, all its power directed towards increasing RPM, it will reach peak torque RPM quickly, and provide 100% output in a really short time since. OTOH if you ramp it up while drawing power, the extra load will reduce climb rate of RPM and delay reaching 100% output. So, a choice between 5 minutes of nothing, then 100%, or 12 minutes of output that is maybe 20% over the first 7 minutes, and then climbs to 100% over the remainder of the time.

  • $\begingroup$ Thank you and sorry for the delay! So I take from your answer: 1. use hot start values. 2. Yes. 3. Yes. $\endgroup$ – LenaH Jun 13 '17 at 11:06
  • $\begingroup$ Additionally, you explain that CCGT goes to low capacity rapidly and then to full capacity slowly. How about I do this (unfortunately my model can only do hourly granularity)…: $\endgroup$ – LenaH Jun 13 '17 at 11:06
  • $\begingroup$ CCGT: 0 hours from 0 to 40% capacity, 1 hour to ramp up from 40 to 100%. Gas turbine: 0 hours to ramp up from 0 to 40% capacity, 0 hours to ramp up from 40 to 100%. Steam turbine: 1 hour to ramp up from 0 to 40% capacity, 1 hour to ramp up from 40 to 100%. Or what do you think? Unfortunately, that is about as exact as my model can go, and I cannot do non-linear ramp-up times and costs – it is a simplistic economic model with more focus on electricity and gas price modelling. $\endgroup$ – LenaH Jun 13 '17 at 11:06
  • $\begingroup$ @LenaH: The numbers look plausible enough and surely CCGTs that match these exist (although I can't say if they are typical; in particular I don't really know what is the proportion in power production between the GT and ST parts of CCGT - if that "40%" is accurate. Personally, I suspect GT part would account for some 60% of energy but I have zero data to support this, just my hunch) $\endgroup$ – SF. Jun 13 '17 at 13:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.