Optimal Energy Management in a Large Petrochemical Plant

Abstract Five steam boilers and four turbine generators produce the steam and a part of the electric power consumed by the various chemical processes of a large plant. The optimal management of these generating units is a dynamic optimization problem because of the presence of start-up costs and it is stochastic because of the uncertainties on future demands. The following steps of the study are described : gathering of technical data, establishment of a demand forecasting procedure, identification of a dynamic stochastic model of the forecast errors (Markov chain and ARMA process), static optimization (optimal allocation of the load among operatings units), dynamic optimization (start-up/stop policy), synthesis of the whole, simulation results.