Research On The Elevator Fault Diagnosis Based On The BP Neural Network

Under the drive of urbanization in our country, the elevator industry progressesday by day, the elevator becomes the indispensable equipment in high-rise buildings.However, elevator faults happen occasionally, causing greatly threat to people’s lifeand property thus raised a research hotspot of the trouble-free operation of theelevator. With the development of science and technology, the elevator system hasbecome a typical computer system, develop and apply the elevator intelligent faultdiagnosis to make management and maintenance of the elevator tend to scientizationand standardization become the aim of this research. The elevator status’s real-timemonitoring and fault prediction can not only reduce the burden of personnelmaintenance, but also can efficiently and accurately identify the failure position andthe possible failure reasons, it has importantly practical value. However, there aresome bottlenecks in presently applied information acquisition and fault diagnosis ofthe elevator, such as intelligent elevator system with perfect security and function isstill scarce, the result of fault diagnosis algorithm that has put into use is poor, it isdifficult to acquire data at the scene, etc.Make the wireless sensor network use to the elevator data monitoring, its uniqueadvantage is suitable for the site data acquired largely and timely, and transmittedsafely and stably. Wireless sensor network is to allocate abundant sensor nodes in anunattended monitoring area, sensor nodes via routing nodes deliver data to thecoordinator be in the form of self organized or multiple hops. Monitoring centeradopts simple but efficient BP algorithm to complete the elevator fault identification.It can automatically identify whether the elevator operates safely or not withoutcomplex elevator model, then make reasonable decisions and act correctly.Aim at assessing the overall performance of the elevator, this paper studies theelevator fault diagnosis based on BP intelligent algorithm.The first step is to solve theproblem of informational sources: design a wireless sensor network based on Zigbee protocol to collect pressure signal of the elevator car. The second step is to study therealization of fault diagnosis algorithm: comprehend the theories and the defects ofBP algorithm, and then propose a hybrid algorithm SA-CG–BP to perfect thealgorithm’s performance. Finally, design a fault supervision platform: in thesupervision center, manage the monitoring data by Labview software, mainly tostorage and display data, and realize fault diagnosis by the hybrid intelligentalgorithm.Ultimately proved that the hybrid intelligent algorithm based on BPalgorithm can achieve the elevator fault diagnosis quickly and accurately, and thegeneralization performance is wonderful.