兰州工业学院学报
蘭州工業學院學報
란주공업학원학보
Journal of Lanzhou Institute of Technology
2015年
2期
66-69
,共4页
故障时间%复合预测%多次残差修正%等维新息动态神经网络
故障時間%複閤預測%多次殘差脩正%等維新息動態神經網絡
고장시간%복합예측%다차잔차수정%등유신식동태신경망락
failure time%combined forecasting%several times residual errors modification%equal dimension and new-information dynamic ANN
为了提高数控机床的可靠性，需要对其工作故障时间进行预测。基于等维新息观点，分别用灰色系统多次残差修正模型和神经网络等2种单一预测方法和等维新息递补神经网络组合预测方法对机床故障观测数据进行了预测，结果显示复合预测误差小于单一预测误差，模型有较高的预测精度。
為瞭提高數控機床的可靠性，需要對其工作故障時間進行預測。基于等維新息觀點，分彆用灰色繫統多次殘差脩正模型和神經網絡等2種單一預測方法和等維新息遞補神經網絡組閤預測方法對機床故障觀測數據進行瞭預測，結果顯示複閤預測誤差小于單一預測誤差，模型有較高的預測精度。
위료제고수공궤상적가고성，수요대기공작고장시간진행예측。기우등유신식관점，분별용회색계통다차잔차수정모형화신경망락등2충단일예측방법화등유신식체보신경망락조합예측방법대궤상고장관측수거진행료예측，결과현시복합예측오차소우단일예측오차，모형유교고적예측정도。
To improve the reliability of CNC machine tool, it is necessary to predict its failure time. Based on e?qual dimension and new?information viewpoint, failure time of CNC machine tools is predicted by using individu?al and hybrid forecasting method respectively. The former includes several times residual errors modification with grey system and Artificial Neural Network ( ANN) , and the latter contains hybrid ANN method. The results show that the forecasting error of hybrid method is lower than individual ones and model has a higher forecasting accu?racy.
