Application of BP neural network in failure diagnosis of electromagnetic direction valve

The paper discussed the application of BP Neural network in electromagnetic direction valve failure diagnosis. The models of failure simulation were made through damaging some parts of the electromagnetic direction valve. A three-layer BP neural network were constructed to diagnose the electromagnetic direction valve based on the signals measured by a non -intrusive detection device. After trained by a subset of the input-output dataset, the diagnosis errors of the application of the BP neural network in electromagnetic direction valve failure diagnosis were comparatively small.