Font Size: a A A

Distillation Separation Sequence Optimization Synthesis Based On Soft Computing Intelligence

Posted on:2005-09-02Degree:DoctorType:Dissertation
Country:ChinaCandidate:H G DongFull Text:PDF
GTID:1101360152975559Subject:Chemical Engineering
Abstract/Summary:PDF Full Text Request
Process system synthesis is the most challengeable and creationary work among the field of process system engineering. The fact of energy lack induces to that separation process subsystem that is characteristic of energy denseness becomes a focus increasingly. Separation sequence synthesis has been the important section of saving energy and reducing waste among the whole process system.The final approach of process system synthesis must realize the simultaneous optimization of flowsheet structure and operation parameters, and carrying out the united optimization of flowsheet structure and operation parameters is a kind of feasible method accomplishing the simultaneous optimization of process system. Simple column distillation separation sequence synthesis is a basic study task among the separation sequence synthesis, and the essential settlement of this problem will provide powerful support for realizing the simultaneous optimization of distillation separation sequences incorporating heat-integrate, non-sharp splitting and complex column etc.Generally all kinds of classical methods applied have to be faced with the problem that the space and time cost is unendurable with the increase of the problem size. To the feature of combination explosion of the number of all distinct distillation separation sequences, this paper researches systemically the intelligence methods combined with soft computing, in order to develop the strategies solving separation sequence synthesis problem. Combined with simple column distillation separation sequence optimization synthesis problem, this paper emphasizes developing the following concrete study jobs:1. The optimization synthesis problem of simple column distillation separation sequences includes several concepts that are characteristic of combination, among those the number of all distinct separation sequences has the feature of combination explosion increase with the increase of the problem size. Through combination mathematic analysis in this paper, that the synthesis problem of distillation separation sequences has the same essential characteristics as some classical combination mathematics problems such as the triangulation of a convex polygon, matrix chain multiplication etc. is found. In fact, there is a certain regular relation between the total number of all possible different schemes and the problem sizes, which is the famous Catalan number. And then, by using computing complexity theory, the computing complexity of the synthesis problem of distillation separation sequences is analyzed concisely, when enumeration algorithm is used to enumerate all distinct separation sequences, its algorithm complexity isn't a polynomial function.2. Abstracting reasonable data structure and then constructing valid neighborhood structure are the foundation of realizing effective search optimization algorithm. At first, the describing method of the problem is put forward systemically in this paper, that is to say, a binary tree of the complete mode and the shortcut mode is respectively used to express a simple column distillation separation sequence. And then, by using the method of graph theory, two kinds of neighborhood mechanism are brought forward, based on the change mechanism of neighboring split position and of internal and external nodes in the binary tree, moreover, the completeness of two kinds of neighborhood structures above-mentioned is proved with mathematics theory, as a result, evolutionary super structure has been constructed. At last, the ordered splitting algorithm is put forward in this paper, that's, Through recursion call by stack technique for the initial component group, all of general binary trees of a simple column distillation separation sequence has been generated from stack permutations in feasible region.3. Tabu search is a kind of global optimization algorithm that simulates the search process that is characteristic of human's intelligence. However, the search process of normal tabu search algorithm only is serial and move of a single...
Keywords/Search Tags:separation sequence synthesis, neighbourhood (super) structure, tabu search algorithm, simulated annealing algorithm, genetic algorithm, fuzzy logic inference, artificial neural network
PDF Full Text Request
Related items