Application Of Genetic Algorithm And Neural Network In Deep-seabed Robot Fault Diagnosing

Deep-seabed robot is the key point to the whole system of deep-sea mining system and its fault diagnosis system is very important to the security of the deep-sea mining system. Deep-seabed robot is a quite complex mechanical system. And the traditional fault diagnosis methods always need to establish an accurate mathematical model, so these methods are not suitable for designing Deep-seabed robot fault diagnosis system. Multiple intelligent diagnosis methods were adopted to design the fault diagnosis system of Deep-seabed robot. Expert intelligent diagnosis system with neural network and genetic algorithms is combined into the fault diagnosis system for Deep-seabed robot conditions monitoring and fault diagnosis. It is significance for the reliability of the whole deep-sea mining system.Deep-seabed robot is composed of several subsystems, including hydraulic system, mining system, running agency and so on. Considering the characteristics of the system and the faults, Deep-seabed robot fault diagnosis system is studied and designed. Firstly, the structure and function of the robot system are studied, and a possible design of fault diagnosis system is given. Secondly, the fault diagnosis system of Deep-seabed robot is base on neural network. Thirdly, according to the shortcomings of neural network, it is combined with genetic algorithm to improve the fault diagnosis system of Deep-seabed robot. Finally, the whole fault diagnosis system is designed, and the human machine interface of fault diagnosis system is developed in the platform Visual Studio. The functions, such as real-time fault diagnosis, fault simulation, information inputting and data query and so on are implemented.The test results show that the Deep-seabed robot fault diagnosis system is working correctly, the algorithm is effective.