Predictive Control for Mitigating Solar Power Variations in Isolated Steam Power Systems

Abstract: The need to reduce the amount of fossil fuel consumed by industry is vital. However, it is crucial to understand that industry is less inclined to shift its consumption from fossil fuels to renewable energy sources solely based on environmentalconcerns; to realistically make industry switch from fossil fuels to renewable energy sources, renewable energy sources must provide efficient results for industry.An aspect of verifying the efficiency of solar power in an industrial setting is bystudying how various control techniques can accommodate solar power alongside fossil fuels.This thesis aims to analyze whether a predictive control model is suitable forsupplying an industrial plant with solar power alongside its coal power. Thisthesis will study two main forms of predictive control: a PID controller featuringa feedforward component and MPC control.The work within the thesis resulted in modelling an industrial plant, bothphysical and mathematical, and an optimization framework to test the MPC. Theframework can be extended to include additional components as well as to introduce disturbances into the model.The results of this thesis project prove that there are great improvements inperformance when applying predictive features to control systems. The predictive control systems presented in this thesis decrease the usage of a buffer in theform of a battery - signifying economic advantages to the predictive methods,and decreasing the total generated power fluctuations by accommodating the incoming solar power before it enters the system.