兩種亞臨界機(jī)組負(fù)荷-式之一。在電力發(fā)電過程中，機(jī)組的負(fù)荷和壓力是兩個(gè)至關(guān)重要的參數(shù)，直接影響著機(jī)組的穩(wěn)定性和發(fā)電效率。因此，對機(jī)組負(fù)荷和壓力進(jìn)行研MATLAB軟件進(jìn)行WiththeincreasingdemandforelectricityinChina,subcriticalunitshavebecomeoneofthemainstreamelectricitygenerationmethods.Intheprocessofpowergeneration,theloadandpressureoftheunitaretwocriticalparametersthatdirectlyaffectthestabilityandefficiencyoftheunit.Therefore,thestudyofunitloadandpressureisofgreatpracticalsignificance.Thispapercomparesandanalyzestwosimplifiedmodelsofload-pressurerelationshipforsubcriticalunits.Oneisthetraditionallinearmodel,andtheotheristhenonlinearmodelbasedonneuralnetwork.Throughthecomparativeanalysisoftheaccuracy,robustness,reliabilityandapplicabilityofthetwomodels,itisfoundthatthenonlinearmodelhasbetterperformanceinpracticalapplications.Toverifytheanalysisresults,thispapercarriedoutexperimentalresearch,usingMATLABsoftwaretoestablishmodelsandsimulateexperiments.Bycomparingtheexperimentalresultswiththetheoreticalcalculationresults,thecorrectnessandpracticalityoftheanalysisconclusionsofthispaperareproved.Keywords:Subcriticalunits;Load;Pressure;Linearmodel;Neuralnetworkmodel;Modelcomparison;Experimentalanalysis$$p=aL+為了解決線性模型的缺陷，本文采用神經(jīng)網(wǎng)絡(luò)方法建立非線性模型。神經(jīng)網(wǎng)絡(luò)模型是一種類似于人腦神經(jīng)元結(jié)構(gòu)的模型，具備自適應(yīng)性，能本文采用了三層前饋神經(jīng)網(wǎng)絡(luò)，其中輸入層、隱層和輸出層分別包3個(gè)、61壓力和時(shí)間，隱層的神經(jīng)元數(shù)目根據(jù)經(jīng)驗(yàn)設(shè)置，輸出層的神經(jīng)元代表機(jī)組當(dāng)前的壓力。(MSE)MSE為1.34，而神經(jīng)網(wǎng)絡(luò)模型的MSE僅為0.67，兩種模