基于最小二乘法改进的ARIMA模型的设备内部状态预测方法

打开文本图片集

中图分类号：TP391.9 文献标识码：A 文章编号：2096-4706（2025）19-0157-07＼

Abstract： With the increasing complexity of engineering systems，traditional physical models and experience are strugling tomeetthepredictionrequirementsof internalstateofequipment.Inthispaper，animprovedARIMAmodelbased onleastsquaresmethodis proposed topredicttheremaining usefulifeofequipmentandother parameterestimationmethods. Firstly，the leastsquares method isusedtodetrendthetraning setsequence，andthentheARIMA modelisestablishedto mine theinteralrelationshipof thedetrended sequenceand predicttheequipment status withhigh precision.Thealgorithmcan efectivelycapturetheinternalrelationshipandtrend betweendata，and hasstrongadaptabilitytodynamicdata.Taking the batteryasanexample，thispaperpredictsteresiualifeandchagequantityunderdiferentconditions，andthepredictioefect is good intheshrt term.Theaverageabsolute errrofthepredictedresiduallife isonlyOO231，andtheroot meansquareerror is 0.0305.Indirentdischargeenviroments，theaverageabsoluteerorofresidualchrgeis0l9，O25200o7ndthe rootmeansquareerroris0.0162，0.0359，0.O062，whichproves thatthe modelcanpredicthestateoftheequipmentwithhigh precision in the short term.

Keywords： least squares method; ARIMA; residual life prediction; equipment internal state prediction

0 引言

在现代工业生产中，工程设备的健康状态直接影响生产效率、产品质量和经济效益[1-3]。（剩余10084字）