Research Of Fault Diagnosis Technology To The Electric Propulsion System

Electric propulsion as a kind of new type propulsion is different from the traditional way of conventional propulsion.Currently it has been widely used in the ships,however the safe and stable operation has become more and more significant.Therefore,it's important to get the fault features precisely,position the fault locations accurately and push aside all faults rapidly in order to ensure the safe operation of the promoting systems and ships.In this paper we use the electric propulsion system of a ship as the research background,researching on the faults diagnosis technology of electric propulsion system and realizing the faults diagnosis,in order to improve the stability and reliability of the whole system,so as to ensure the safe operation of the ships.First of all,based on the rotor flux oriented control method,the models in natural coordinate system,mathematical model of vector space decomposition and 3dq coordinate transformation are established.Then the topological structure and working principle of fifteen phase inverter is analyzed,After that the motors in the vector space and 3dq coordinates control strategy are analyzed,eventually the simulation model And analyze for the simulation model is established.Secondly,the paper researches the load characteristic of propeller,in order to establish the model and analyze the results of the simulation.The main parameters of the inlet ratio is modified to fit the modified curve getted by using Chebyshev polynomials,the establishment of mathematical model of ship engine propeller has been completed.Aiming at several typical working conditions of the navigation is simulated.Thirdly,based on the normal electric propulsion model,this paper establishes the system fault simulation model and analyzes the fault characteristics in-depth.The fault of the inverters,the propulsion motors phase failure and propeller trouble have been analyzed simply and necessarily.For the propulsion motor interturn short circuit fault,due to whose model is complex,so the modeling process is studied in detail.When the propulsion system fails,the fault characteristic curve is obtained.Finally,the fault diagnosis technology of electric propulsion system bases on BP neural network is studied.Aiming at the defects existing in the standard BP algorithm,puts forward three measures.In order to improve the network performances,putting forward using the genetic algorithm optimized the BP algorithm.Through analyzing the sample datas of common fault in the power propulsion system,building and training of the BPneural network fault diagnosis system bases on the genetic algorithm has been completed.The simulation results show that using the BP neural network fault diagnosis system based on the genetic algorithm can complete the fault diagnosis for electric propulsion system quickly and accurately and each phase performance meets the design requirements.