Research On Intelligent Monitoring System Of Wound Rotor Induction Motor

As is well known to us all,motor is vital and crucial in our work and life.The safe operation of motor is of the utmost importance on many occasions.Undoubtedly,when something is wrong with the motor,it does lead to a series of terrible problems concerning economy and even personal safety.What’s worse,we couldn’t predict what would happen in the long run with regards to society.Researches have shown that apart from a small part of the sudden fault of motor,a majority of the breakdown experience the latent period and they could be dealt with professional technology.In addition,the motor is still at work for a short time even they go wrong at that time.If they are able to give an alarm and the losses are supposed to be minimized.Thus,it is of great necessity to monitor the motor that is used in sinificant occasions.In this paper, we will study the wound rotor induction motor that have big market share and developed a motor monitoring systems based on artificial intelligence,furthermore launch a series of studies around the system.The main contents are as follows:First and foremost it analyzes the failure mechanism of wound rotor induction motor. In particular,it demonstrates how to carry out mathematical model of motor brush- slip ring’s dynamic resistance and quantitative analyze the cause of the slip ring spark fault.And then,the overall design of Asynchronous Motor monitoring system is depicted.It introduces the principle of operation of the monitoring system, and determine the function module need to achieve and the tools develop monitoring system must used.Furthermore, the design and development of monitoring systems’ communication module. It use C #.NET to write the Monitoring System’s data collection procedures based on MODBUS communication protocol and verify its reliability of the monitoring system by running data.Accordingly, the establishment of motor diagnostic model based on the extreme learning machine. Firstly,set up a wound rotor induction motor fault experimentalplatform, gathering the waves of the motor stator three-phase voltage and current under the normal conditions and fault conditions, provide the study basis for follow-upfault diagnosis. Then extract the fault characteristic vectors by the method of transform the collected voltage and current signals in the motor experiment thought wavelet packet algorithm.And then we try to use the fault characteristic vectors to train diagnosis model based on ELM for the establishment of the wavelet packet-extreme learning machine fault diagnosis model. Finally,test the reliability of wavelet packet- extreme learning machine fault diagnosis model.Accordingly, establish a motor diagnosis model based on support vector machine.Firstly,giving abrief introduction of SVM’s principle and characteristics.Then using wavelet packet algorithm to extract fault characteristic vectors of the motor and establish the motor fault diagnosis model based on the support vector machine.Taking into account whether the parameters of SVM is select correctly or not has a greater impact on the diagnosis model’s performance, we use multiple population GA to optimize support vector machines for the purpose of improving the performance of the diagnostic model. Finally, using mixed programming technology to embed the fault diagnosis model into monitoring system.Finally,Monitoring stage and monitoring plan are being set up to test the key function regarding the monitoring system.The experiment result makes it clear that the development of the system with regards to electric motor will denifitely diagnose the situation whether the motor is working or not and has good practical reference value.