基于VMD-PSO-SVM的甲醇合成气压缩机状态预测  

Status Prediction of Methanol Synthesis Gas Compressor Based on VMD-PSO-SVM

作　　者：左学谦 夏楠 Zuo Xueqian;Xia Nan(School of Mechatronics and Automation,Wuchang Shouyi University,Wuhan 430064,China)

机构地区：[1]武昌首义学院机电与自动化学院,武汉430064

出　　处：《机电工程技术》2025年第2期143-147,共5页Mechanical & Electrical Engineering Technology

摘　　要：针对甲醇合成气压缩机由于运行环境恶劣,导致故障信号中含有大量噪声干扰等问题,提出一种基于VMD-PSO-SVM模型的甲醇合成气压缩机状态预测模型。首先利用变分模态分解将原始数据分解成不同频率的本征模态函数,其次,将各子函数引入到PSO-SVM中,实现对子函数的状态预测,最后再将子函数叠加得到原函数的状态预测。采用实验组与对照组对照的研究方法,以平均绝对误差、均方根误差以及决定系数等参数来评估预测的准确性。结果表明:VMD-PSO-SVM模型较PSO-SVM模型综合提升60%左右,且需要迭代次数减少了83.8%。该方法具有预测精度高、鲁棒性好、抗噪声性能优、迭代次数少等优点,可以为预防性维护提供可靠的理论基础,在实际工程中有很高的实用价值。A state prediction model for methanol synthesis gas compressors based on VMD-PSO-SVM model is proposed to address the problem of high noise interference in fault signals caused by harsh operating environments.Firstly,variational mode decomposition is used to decompose the original data into intrinsic mode functions of different frequencies.Secondly,various sub functions are introduced into PSOSVM to achieve state prediction of the sub functions.Finally,the prediction of the original function′s condition is obtained by superimposing the sub-functions.This study adopts a research method comparing the experimental group with the control group,using parameters such as mean absolute error,root mean square error,and coefficient of determination to evaluate the accuracy of the predictions.The results show that the VMD-PSO-SVM model has a comprehensive improvement of about 60%compared to the PSO-SVM model,and the number of iterations required has been reduced by 83.8%.This method has the advantages of high prediction accuracy,good robustness,excellent noise resistance,and fewer iterations.It can provide a reliable theoretical basis for preventive maintenance and has high practical value in practical engineering.

关 键 词：甲醇合成气压缩机 状态预测 PSO-SVM VMD-PSO-SVM 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]