Aiming at the fault diagnosis problem of the spectrometer amplifier in nuclear radiation detection system

a diagnostic method that can diagnose the fault of the spectrometer amplifier when the nuclear radiation detection system is working is designed and verified.The method uses the output pulse signal of the charge-sensitive amplifier at the front end of the spectrometer amplifier as the excitation signal of the diagnostic circuit

and uses wavelet packet decomposition to extract the characteristic information of the output voltage waveform of the spectrometer amplifier under different fault states.Support vector machine（SVM） optimized by gravity search algorithm（GSA） achieve fault classification based on fault feature information.The simulation results of the model show that when single and combined soft faults occur in the spectrometer amplifier

the GSA-SVM model can accurately classify the faults according to the characteristic information of the fault signal

and realize the efficient diagnosis of the spectrometer amplifier.