The Study Of Multi-effect Distillation Energy-saving Technology Based On Genetic Algorithm

In order to utilize energy more fully and reasonably, the residual heat (including the sensible heat of the heating steam condensate in the reboiler of the first effect, the heat of tops and bottoms product of each effect distillation tower)was used to preheat feed flow. The energy integration model of paralled-feed, forward-feed, backward-feed multi-effect distillation system was established, which was composed of two-subsystem multi-effect distillation and multilevel preheating. Material-balance and energy-balance equations of the system were expressed in matrix equation. The matrix equation has the advantage of clear structure, simplifying to different effect number easily or not having preheat. Thereupon, all parameters of distillation system could be obtained through solving Material-balance and energy-balance equations by iterative method and matrix method, calculating theory plate number by stage by stage method and correlative aided calculation such as bubble-point calculation, dew-point calculation, preheater calculation et. The optimum design models of paralled-feed, forward-feed and backward-feed multi-effect distillation system were established on the basis of conventional design to reach the optimum energy-saving effect. The models aim at the least annual cost which includes operating cost (including vapor consumption cost, vacuum power cost) , equipment depreciation cost and maintain cost(including equipment depreciation and maintain cost of distillation column, reboiler, preheater). The steam temperature of reboiler at the bottom of first effect, the vapor temperature at the top of last effect, the operating reflux ratio of vari-effect distillation column, the heat-transfer temperature difference of vari-effect reboiler and outlet temperature of multilevel preheater are both the decision variables. The objective function is a multi-variable, nonlinear and complex one with inequality constraints(including implicit constraints and explicit constraints) and equality constraints. So it is more difficult to solve this model. The method to solve this model had not been reported in the literature. Then a new algorithm genetic algorithm combined with Fibonacci method and Lagrangian multiplier method was presented and studied in this work. The process of solving objective function was divided into inner and outer layer by this new method. The outer decision variables t_s, t_Dn , t_j are solved by genetic algorithm. The inner decision variables Ri, ?ti are solved by Fibonacci method and Lagrangian multiplier method respectively.New algorithm was realized by Visual Basic 6.0. The results showed the annual cost of paralled-feed, forward-feed and backward-feed triple-effect distillation system with preheating system decreased by 48.1%,41.3%,33.8% respectively than without preheating system. It indicated that the preheating and optimum design are the more effective measure to save energy. This algorithm that possesses good convergent stability is a effective one for solving optimum problem of multi-effect distillation column quickly. The design software for multi-effect distillation system was developed, which have amity interface and easy maneuverability. This software may be applied to conventional and optimum design of paralled-feed, forward-feed and backward-feed multi-effect distillation system.