The speedometer on my car goes to 230 km/hr. I seriously doubt if I have the skill to drive my car that fast, but if I could (and did) my car woudl not last very long. As the load on the engine increases, the maintenance requirements also increase – sometimes drastically.
This is one of the interesting things about industries using combustion engines and combustion gas turbines – we have the more power / more maintenance cost trade off to constantly consider.
I recently delivered a training course regarding turbo-machinery to a group of operating and maintenance foremen, and I asked the question … “when is your engine or turbine at maximum sustainable load”? Everyone understood the question, but no one could give me a numerical answer. Because of the more power / more maintenance cost trade, no one could give me a definite answer of when the engine was “at maximum sustainable load”.
Because the performance of combustion machinery is often directly linked to total plant throughput, increasing the load on the engine / turbine will increase throughput, which will almost always increase income. But will it increase profit? Maybe I am being overly simplistic in thinking … “determine the point where the engine / turbine is full, then find a way to operate at that point because it will maximise profit.” If you can determine the point where the engine or turbine is at maximum sustainable load, then you can measure it. If you can measure it, then it becomes the basis for your control system.
Some in the training class said they used speed to control load. While speed is related, it is not the single contributor to a load on a combustion machine. Going back to the automobile, I can drive at 100 km/hr uphill or downhill, and the engine has different loads. And for those of you who say the engine speedometer is not the engine speed, I can make the same statement if I replace the word “speedometer” with “tachometer”. Different engine rpms does not equate to engine load. I can run my engine at high rpms with or without a load on the engine.
As we continued our discussion, it became clear that every good parameter (or combination of parameters) was a “soft” parameter (the point of 100% load was open to debate) instead of a “hard” parameter (like a setpoint to shut down a device). I felt the group was working in the right direction, because any parameter would need to confront the more power / more maintenance issue. Without statistical data, the point of being “fully loaded” would be subjective. This made me feel they were looking at the correct parameters.
When I was a young engineer, I worked in some natural gas processing plants. We confronted this problem, and determined that for our turbo-charged internal combustion engines, the measurement was engine manifold pressure. After long debate, data analysis, guessing, compromise, swearing, and other unpleasant times, we reached concensus … the engine was at 100% sustainable load when the engine manifold pressure was at +4-inches of mercury. While I felt this was low, I now had a number to do my calculations and develop an optimising control scheme. Of note – we had a different measurement for our naturally aspirated engines, and yet another different measurement for our turbines.
And the scheme worked well. It was a simple cascade control loop. The inner loop varied engine speed, and was based on a process parameter. For example, for a refrigeration compressor, the parameter was chiller temperature. The speed was varied until process temperature setpoint was achieved. But … for our operations, colder = more income (and as long as maintenance costs were under control, more profit). The outer loop of the cascade control loop would change the process setpoint to increase the load on the engine. The result was a refrigeration system that would always run as cold as possible on the day. If the engine was in need of any maintenance, it would run as cold as possible. If the engine was freshly overhauled, it would run as cold as possible. The difference between the two days was small but noticeable, and we could still say we made as much profit as was possible on the day. We were able to routinely drop our chiller temperature by about 1-2 degrees with the new scheme. It was enough of a success to be adopted to all of our engines in our facilities.
Our economic situation was quite simple – we never had to determine our marginal profit. Our cost of feedstock always ensured that if we made product, we could sell it for a net profit (with the exception of extremely high maintenance costs). Other organisations did not have this luxury, and optimisation is more complex – and potentially needs to be done daily (or maybe hourly). I have a great deal of respect and admiration for people that have successfully optimised against those constraints.