MO-3LM-CA And Its Application In Operation Optimization Methods For Texaco Gasifier

Researches and practical project applications on new multi-objective cultural intelligent optimization algorithms which can be used to solve production unit operation optimization problems of complex process industry are of great significance for enhancing the core competence of enterprise in market economy and improving the economic benefit and social benefits of enterprise. In this paper, a new3-layer cultural intelligent optimization algorithm framework is presented. Standard genetic algorithm (GA), particle swarm optimization algorithm, and (PSO)differential evolution algorithm (DE) are embedded into this framework to construct3-layer mixed cultural differential evolution algorithm (3LM-CDE),3-layer mixed cultural genetic algorithm (3LM-CGA), and3-layer mixed cultural particle swarm optimization algorithm (3LM-CPSO) which can solve signal and multiple objective non-linear programming(NLP) optimization problems effectively. At the same time, the proposed algorithms are utilized to the signal and multiple objectives Operation Optimization to online optimize3control parameters of the Texaco gasifier, and the simulation results verify the effectiveness and excellent application value of the proposed algorithms. The contents in this dissertation can be summarized as follows:(1) A new3-layer cultural intelligent optimization algorithm framework is first presented. There are three layers in this framework structure:standard GA, PSO and DE can be embedded into the bottom population space to evolve multi-population and independently; the medium space extracts and develops the evolution knowledge of each population from the bottom population space; in the belief space, effective knowledge is further abstracted from the medium and is utilized back to medium space and bottom space to guide its searching process. Simulation experiments of20classical NLP problems show that this framework has excellent feasibility, convergence and universality.(2)3-layer cultural intelligent optimization algorithms (3LM-CGA,3LM-CPSO and3LM-CDE) presented in (1) are utilized to online operation optimize3control parameters of the Texaco gasifier in a real-world chemical plan, and the objective is to improve effective gas yield of gasfier. On the premise of meeting the gasfier’s process requirement, we can get better optimization results than standard GA, PSO and DE, and this can prove the effectiveness of3LM-CGA,3LM-CPSO and3LM-CDE.(3) Multi-objective3-layer mixed cultural differential evolution algorithms are presented. Based on3LM-CDE presented in (2) and combined with main objective method, linear weighted method, max-min method, ideal point method,2-norm method and Pareto-based methods, six kinds of multi-objective3-layer mixed cultural differential evolution (MO-3LM-CDE) algorithms are presented. Simulation experiments of10classical test functions show that MO-3LM-CDE has much better effectiveness and superiority than MO-CDE.(4) MO-3LM-CDE algorithms presented in (3) are utilized to the problem of online operation optimization shown in (2) which optimizes the three objectives including Outlet flowrate of syngas, CO Volume%in the Outlet syngas and H2Volume%in the Outlet syngas simultaneously. We can get better optimization results through multi-objective evolutionary algorithm presented in (3) than signal-objective evolutionary algorithm shown in (2). MO-3LM-CDE has better effectiveness and can provide a broader range of operation parameters of gasfier’s online operation optimization.