A temperature sensor fault detector as an artificial neural network application

The paper presents a fault detection method based on artificial neural network (ANN) capabilities applied in a particular case: a temperature virtual measurement system (VMS). A study concerning the ANN architectures and the ANN training rules involved in a fault detection procedure are also presented. The results obtained for the VMS based on a data acquisition board or on a GPIB controlled instrumentation highlight the general applicability of neural processing.