多级离心压缩机混合模型研究.docVIP

  多级离心压缩机混合模型研究# 董世建1，王福利2，梁倩1，黄文辉1* （1. 东北大学信息科学与工程学院，沈阳 110004； 5 10 15 20 25 30 2. 流程工业综合自动化国家重点实验室，沈阳 110004） 摘要：以某钢厂燃气_蒸汽联合循环发电机组低压端煤气压缩系统为背景，建立以控制阀和 离心式压缩机为核心的多级离心压缩机机理模型。采用改进自适应遗传算法，辨识机理型模 中难以确定的参数。多级离心压缩机存在较多的影响因素，机理模型预测性能不精确。基于 多最小二乘支持向量机(M-LS-SVM)，补偿机理模型的误差，建立预测多级离心压缩机性能 的混合模型，。将混合模型应用于实际的煤气压缩系统中，结果表明与机理模型相比混合模 型有更高的精度和更好的预测性能。为制定相应的压缩机防喘控制策略奠定了良好的基础。 关键词：离心压缩机；机理模型；改进自适应遗传算法；参数辨识；多最小二乘支持向量机； 混合模型 中图分类号：TP273 Research of Hybrid Model for the Multi-stage centrifugal compressor DONG Shijian1, WANG Fuli2, LIANG Qian1, HUANG Wenhui1 (1. School of Information Science Engineering,Northeastern University,Shenyang 110004; 2. State Key Laboratory of Integrated Automation for Process Industries, Northeastern University, Shenyang 110004) Abstract: Based on the gas system of gas-steam combined cycle power plant from a steelworks, this establish a mechanistic model for the multi-stage centrifugal compressor, which consists essentially of control valves and centrifugal compressors . Improved adaptive Genetic Algorithms is applied to estimate the important parameters of the model. Since there are many factors influencing on the performance of the multi-stage centrifugal compressor, the mechanism model may yield inaccurate results. Thus with multiple least squares support vector machine employed to correct the error of the mechanistic model, the hybrid model is established . The hybrid model is applied to the practical gas compression system and the results demonstrate that the precision of the hybrid mode is higher than that of the mechanism model. This lays a solid foundation for establishing the anti-surge control strategy. Key words: centrifugal compressor ;mechanism model; improved adaptive genetic algorithm; parameter identification; multiple least squares support vector machine ;hybrid model  35 40 0 引言 离心压缩机具有流量小、压比高、结构灵活等优点，在工业生产、军事和农业等领域得 到广泛应用。但是当输入气体流量减少时，压缩机容易发生喘振。喘振会影响离心压缩机的 性能，破坏整个气体压缩系统的稳定行[1]。虽然文献[2]提出了离心压缩机的混合模型，但该 模型只是通过数据补偿建立压缩机输出功率和入口流量之间的关