Intelligent diagnosis of ROV hydraulic system based on Bayesian network

Hydraulic system is one of most important system of remote operated vehicles (ROVs). The hydraulic system of the “Sea Star” ROV consists of a 5-function manipulator, a 6-function manipulator, valve packs and submersible hydraulic pump. Bayesian network model has been built about faulty components and fault types. Firstly, Bayesian network model was built. Secondly, the network priori parameter was determined. Thirdly, parameter learning of Bayesian network was finished. Fourthly, the paper finished probabilistic inference of Bayesian network and model verification. Finally, given some fault types, the simulation on Bayesian network diagnosis model has good results. According to the above method, software has been developed by Microsoft Visual C++ (VC++), and the application result is good in processing logging data.