科研之友
微信
新浪微博
Facebook
分享链接
摘要
Complex turbine (i.e., steam turbine with multiple controlled and uncontrolled extractions) is one of the most important components in industrial utility system. The function of complex turbine is to generate power electricity, provide steam for the processes, and heat boiler feedwater using uncontrolled extraction steam. Although many studies have focused on the modeling and optimization of utility systems of different scales, few works have been reported on the modeling and operation optimization for utility system containing multiple complex steam turbines. In this paper, a superstructure of utility network consists of multiple interconnected various types of complex turbines (ICSTUN) is constructed. The superstructure demonstrates not only an external utility network that supplies utility energy for the processes but also an internal utility network that supplies boiler feedwater. A systematic modeling approach is developed for robust simulation, validation and operation optimization of ICSTUN. A mixed-integer nonlinear programming (MINLP) model based on an improved modeling principle of complex turbine is formulated for the operation optimization of ICSTUN. The MINLP model is able to simultaneously optimize the external and internal utility networks. A solving strategy incorporating model decomposition, model relaxation and case initialization is proposed to find a feasible solution. A large utility network in a petrochemical complex is studied to test the accuracy of the complex turbine model and the effectiveness of the proposed operation optimization methodology. The validation results of the complex turbine model show that the mean relative error between field operation data and simulation data is less than 1.18% and is acceptable in engineering application. The optimization of a single utility plant yields a maximum coal reduction rate of 4.59%, and the optimization of the total utility system consisting of two utility plants yields a maximum coal reduction rate of 6.01%. The optimization contribution is discussed in detail, and several conclusions and suggestions are presented.
