I believe that the current framework used to calculate the efficiency and fuel consumption rate of thermal units (gas turbines, coal plants, etc) has some limitations that are especially pronounced in the combined cycle case (see #232).
Current framework
In the current framework, the efficiency is provided through a lookup table that defines the efficiency as a function of load percentage. I assume this method was chosen since such tables/graphs are readily available online for different thermal power generators, enabling easy implementation. The fuel consumption is then calculated by dividing the power by the heating value (HHV) of the fuel times the efficiency.
Limitations
Although easy to implement, I believe this method is flawed, as it reverses the order of things as they happen in reality. The efficiency depends on the fuel consumption, not the other way around. As I see it, this method works reasonable well when the generator is on and producing any level of power between full load and minimum stable load. However, it doesn't accurately model the actual fuel consumption in start-up or stopping states. It assumes it just operates at or below its lowest efficiency point in the lookup table, still giving a relatively high efficiency and therefore relatively low fuel consumption. However, as far as I understand, especially the startup phase requires significant fuel consumption while still producing no to very limited power. In other words: the efficiency should be very low in this phase of the process. As a result, the current framework doesn't accurately model the (fuel) price of starting up and shutting down a thermal power plant.
Possible solutions
- Reverse the order to determining the efficiency based on the fuel consumption. This would require a major overhaul and more data on the fuel consumption of different types of generators in different phases of operation. For that reason, this might not be feasible.
- Implement a slightly more complicated efficiency / fuel consumption logic. We keep the current framework in place if the unit is on, but switch to different implementations when the turbine is revving up or down (and perhaps also when it's shut down). This would only require minor changes to the code, and only one or two additional input variables (something like a
startup_fuel_consumption_rate and shutdown_fuel_consumption_rate).
Perhaps this is too complicated / detailed for the current level of modeling we are trying to do, in which case, feel free to close the issue. But I would still like to hear your opinion on this matter.
Tagging @genevievestarke @misi9170 @paulf81 and @dzalkind.
I believe that the current framework used to calculate the efficiency and fuel consumption rate of thermal units (gas turbines, coal plants, etc) has some limitations that are especially pronounced in the combined cycle case (see #232).
Current framework
In the current framework, the efficiency is provided through a lookup table that defines the efficiency as a function of load percentage. I assume this method was chosen since such tables/graphs are readily available online for different thermal power generators, enabling easy implementation. The fuel consumption is then calculated by dividing the power by the heating value (HHV) of the fuel times the efficiency.
Limitations
Although easy to implement, I believe this method is flawed, as it reverses the order of things as they happen in reality. The efficiency depends on the fuel consumption, not the other way around. As I see it, this method works reasonable well when the generator is on and producing any level of power between full load and minimum stable load. However, it doesn't accurately model the actual fuel consumption in start-up or stopping states. It assumes it just operates at or below its lowest efficiency point in the lookup table, still giving a relatively high efficiency and therefore relatively low fuel consumption. However, as far as I understand, especially the startup phase requires significant fuel consumption while still producing no to very limited power. In other words: the efficiency should be very low in this phase of the process. As a result, the current framework doesn't accurately model the (fuel) price of starting up and shutting down a thermal power plant.
Possible solutions
startup_fuel_consumption_rateandshutdown_fuel_consumption_rate).Perhaps this is too complicated / detailed for the current level of modeling we are trying to do, in which case, feel free to close the issue. But I would still like to hear your opinion on this matter.
Tagging @genevievestarke @misi9170 @paulf81 and @dzalkind.