Steam is the major working fluid in the process industries, and the energy utilization of steam pipe network directly affects the energy consumption of the production plant. The steam piping network system is a major component of the ethylene plant energy consumption. Due to the lack of detection information and the unreasonable design of network structure, the configuration of steam stay in experience, and the adjustment of it is blindness, which causes the energy waste. While, there are some problems that the process of the energy utilization and the distribution of the steam consumption on different levels are not reasonable, resulting in the waste of steam.Firstly the paper proposes a new optimization algorithm of collaborative quantum differential evolution algorithm (CQGADE) for the energy utilization and structure optimization of the steam network is a problem of mixed integer nonlinear programming (MINLP). Compared to the algorithm of the improved GA and the hybrid PSO, the proposed algorithm has good robustness and better convergence rate through the test function. Secondly, according to the lack of the real-time inspection of turbine steam flow, this paper put forward the neural network with link switches as the soft sensor model. CQGADE is applied to tune both network structure and parameters of the neural network simultaneously to provide foundation model for the optimization of the steam piping network. Finally, according to the ethylene steam network system, this paper use the costs of operation of steam network as the objective function based on the material balance, energy conservation, product quality. Using the proposed algorithm (CQGADE) to optimize the objective function, we get a get a satisfactory optimal result. |