Research on fault diagnosis and forecast system of forest harvester based on CAN-bus information

The fault diagnosis of forest harvester is updating with the application of CAN bus technology. Aiming at the CAN technology which was utilized on forest harvester currently, the complexity of the fault information and the difficulty of diagnosis, an USB-CAN intelligent interface card was designed in this paper. Based on the interface card, the software of Microsoft Visual C++6.0 is utilized to build the fault diagnosis system with BP neural network and Kalman filter. The fault diagnosis and forecast to the main systems of forest harvester were released online after incepting, filtering and removing the noise of the signal from the CAN bus. As the experiments show that Kalman filtering plays good on removal of noise from the complex fault signal, and the BP neural network trainings of the systems are effective to implement non-linear mapping from the fault phenomenon to the fault position of forest harvester.