Low-temperature separation processes,such as the cold end of ethylene plant and LNG separation,consist of three correlative sections: the core process (separation sequence),the heat exchanger network (HEN) and the refrigeration system.Traditionally,the separation sequences were designed first,followed by the heat exchanger network and the refrigeration system that supported it.As the total energy cost is summation of all the three parts,such design methodology failed to consider the interrelation among the three sections,could not guarantee the optimal total energy.In order to cope with this drawback,total site optimization method is required,but too many variables and constraints make it quite a complex problem.In this study,the combined consideration of distillation column profiles and of low-temperature shaftwork targets in the design of low-temperature distillation-based process has been explored.Based on stochastic optimization strategy[1] and exergy and pinch technique[2~5],a mathematical model is formulated to design low temperature processes,in which the stochastic algorithm deal with the flowsheet structure and the levels of continuous variables,and the exergy and pinch technique identify the shaftwork target without knowing the exact heat exchanger network and refrigeration system.With this method,separation sequence structure and minimum total energy cost are connected directly,the operating pressure and separation structure will be the only independent variables,which makes programming solving much easier.Moreover,this model can also be used to design processes with complex columns and heat integration.Temperature levels of refrigerant can be optimized as well.Finally,a case study of ethylene cold-end separation is used to illustrate the application of the approach.Optimum solutions in different conditions (simple column and complex column,with or without heat integration and different refrigerant levels settings) are compared with the original scheme.