Optimization of an Oil and Gas Separation Plant for Different Reservoir Fluids Using an Evolutionary Algorithm

The optimization of an oil and gas separation plant operating revenue has been performed for different characteristic reservoir fluid types (gas condensate, volatile oil, and black oil) using an evolutionary algorithm. A process simulation model mimicking a typical plant has been used as a black-box model and optimized with respect to nine design variables using the covariance matrix adaptation evolution strategy (CMA-ES) algorithm. The plant studied has three separation stages, including gas recompression for each stage as well as a final gas boosting step before export. Each compression stage includes gas cooling and partial condensation upstream compressors. All condensate streams from the recompression system are recycled back into the separation system for increased liquid recovery. The results indicate the following common optimal settings among others: the first-stage temperature is optimal at the high bound, the third-stage temperature is optimal at the low bound, the temperature of the gas from the scrubber receiving gas from the middle stage separator is optimal at the lower bound. Some of the settings are different between the three fluids investigated, but with a clear trend among the fluids. One example is the optimal middle-stage separator pressure, which increases with decreasing gas–oil ratio of the fluid. Benchmarking the optimization potential indicates that an increased operating revenue of close to 1% may be realized for the gas condensate and the volatile oil, whereas the optimization potential is less for a black oil fluid type. It is also noted that this optimization may come at a significant penalty in terms of the energy required, especially for the volatile oil case.