Cracking gas compressor is usually a centrifugal compressor. The information on the performance of a centrifugal compressor under all conditions is not available, which restricts the operation optimization for compressor. To solve this problem, two back propagation (BP) neural networks were introduced to model the performance of a compressor by using the data provided by manufacturer. The input data of the model under other conditions should be corrected according to the similarity theory. The method was used to optimize the system of a cracking gas compressor by embedding the compressor performance model into the ASPEN PLUS model of compressor. The result shows that it is an effective method to optimize the compressor system.
构造了一种基于Alopex(Algorithm of pattern extraction)和分布估计算法(Estimation of distribution algorithm,EDA)相融合的进化算法EDA-Alopex。该算法将分布估计算法嵌入到一种基于Alopex的群智能进化算法(Alopex-based evolutionary algorithm,AEA)中,利用分布估计算法收敛速度快及与传统进化算法进化模式不同的特点来改进AEA算法。新算法综合了AEA算法搜索得到的个体间相关性信息和EDA搜索过程中得到的全局概率信息,能够更好地指导种群向有利的区域进化。仿真结果表明:EDA改进的EDA-Alopex算法搜索性能与AEA算法的搜索性能相比有较大提高,特别是其收敛速度与AEA算法相比有明显提高。
In this work, focusing on the demerit of AEA (Alopex-based evolutionary algorithm) algorithm, an improved AEA algorithm (AEA-C) which was fused AEA with clonal selection algorithm was proposed. Considering the irrationality of the method that generated candidate solutions at each iteration of AEA, clonal selection algorithm could be applied to improve the method. The performance of the proposed new algorithm was studied by using 22 benchmark functions and was compared with original AEA given the same conditions. The experimental results show that the AEA-C clearly outperforms the original AEA for almost all the 22 benchmark functions with 10, 30, 50 dimensions in success rates, solution quality and stability. Furthermore, AEA-C was applied to estimate 6 kinetics parameters of the fermentation dynamics models. The standard deviation of the objective function calculated by the AEA-C is 41.46 and is far less than that of other literatures' results, and the fitting curves obtained by AEA-C are more in line with the actual fermentation process curves.
