CAJ | 학술논문

现有汽车衡不具备容错功能,为此提出了一种汽车衡智能容错方法.根据多路称重传感器之间的相关性,利用径向基函数神经网络建立称重传感器预估网络和自适应选择网络,完成称重传感器故障检测、失效传感器正常输出预估和汽车衡工作模式的自动切换,并将预估信号与其它正常传感器信号组合,通过称重结果融合网络完成任一传感器失效状态下汽车衡的正常称重.试验与现场检定表明,这种智能容错方法实现了汽车衡任一传感器失效状态下的正常称重,其各项指标均优于4级秤的要求,避免了称重系统失效.
현유기차형불구비용착공능,위차제출료일충기차형지능용착방법.근거다로칭중전감기지간적상관성,이용경향기함수신경망락건립칭중전감기예고망락화자괄응선택망락,완성칭중전감기고장검측、실효전감기정상수출예고화기차형공작모식적자동절환,병장예고신호여기타정상전감기신호조합,통과칭중결과융합망락완성임일전감기실효상태하기차형적정상칭중.시험여현장검정표명,저충지능용착방법실현료기차형임일전감기실효상태하적정상칭중,기각항지표균우우4급칭적요구,피면료칭중계통실효.
An intelligent fault-tolerant method for struck scale is proposed. An adaptive-choice network and an estimative network for load cells based on radial basis function network and the correlation of load cells are established. Using these networks, the faulty load cell is detected and its normal output is estimated, then the normal working mode and failure working mode are automatically switched. The estimative signal and the signals of other good cells are assembled, and then the weighing result is obtained by weighing fusion network. Verification in field and experiments show that with this proposed method, the truck scale can normally weigh load and its performance is better than that of the 4th class scale defined by Chinese National Standard.