Set-point Determination and Regualtion for Flue Gas Desulphurization Process by ELM-assisted MOPSO and LQIR controller

In China, coal accounts for over a half of the total energy consumption and it is primarily used in coal-fired power stations. Undoubtedly, the change of a dominant role that coal played in electricity generation appears to be impossible in the near future. The operation of coal-fired power stations involves combustion of coal, which would generate a vast amount of sulphur dioxide (SO2), the resulting SO2 in turn pose numbers of environmental and human-health hazards. In view of the above considerations, optimization and control techiniques should be incorporated in flue gas desulphurization (FGD) system, so that further improvement in overall performance of FGD can be made. In this paper, based upon historical operational data from FGD system at a 1000MW unit coal-fired power station, we formulate the set-point determination in FGD process into a constrained many-objective optimization problem. First of all, multiple objective particle swarm optimization (MOPSO) algorithm along with extreme learning machine (ELM) is utilized to yield Pareto-optimal front. Then, for control purpose, setpoint is determined from the obtained Pareto front. Lastly, with the use of linear quadratic integral controller (LQIR), the setpoint regualation task is fulfilled. By simulation studies, derived respresentative point is theoretically justified and FGD process can be regulated to it successfully.