Research On Fault Diagnosis Technology Of Electromechanical Equipment Based On Cloud And End Integration

Mechanical and electrical equipment in the manufacturing industry has been more widely used with the development of intelligent manufacturing.Motor and motor bearings are important mechanical and electrical components,but the failure rate of these important parts has remained high.Therefore,the real-time monitoring of important equipment in the production process is an essential link to ensure the normal production work.The fault types of electrical and mechanical components such as motor and bearing are various.The fault features extracted from different faults by means of effective methods and accurate classification are the problems that need to be studied in real-time fault diagnosis of mechanical and electrical equipment.The emergence of intelligent manufacturing mode has promoted the process of data digitization in the manufacturing industry,and has also opened up space for the application of cloud plus terminal data processing in the manufacturing industry.Most of the traditional fault diagnosis systems collect the running data of equipment firstly,and then process the data with a PC.In this way,there will be no real-time diagnosis and a lot of opportunities for timely find the faults will be lost.In order to solve this problem,an embedded intelligent terminal is designed to realize the data acquisition and data processing in the field.But the data processing capacity of the terminal is limited,so this paper proposes a detection algorithm merged together with cloud computing and dynamic nesting sliding window.The algorithm utilizes the advantage of the strong computing power of cloud computing to solve the problem that intelligent terminal lacks the capacity to solve the fault signal.In order to improve the architecture of Cloud Plus,this paper further developed related research and an improved Cloud Plus Terminal Support Vector Machine(CPTSVM)is proposed.The previous fault diagnosis systems are run on a single processor,while the fault feature extraction,data transmission,and classification are carried out in the terminal and cloud respectively in the CPTSVM system.The "pipelining" is introduced into the data processing of CPTSVM,which can significantly improve the parallel performance of the diagnostic system.In the cloud,the Cloud Feature Mode Library(CFML)is built,and fault features are added to the model library selectively.Offline training and online training are all employed in the CPTSVM structure,and the updated data in CFML can renew the online training model,which can strengthen CPTSVM adaptability and obtain the ability of "lifelong learning".Through theoretical calculation and experimental verification,it is found that the existence of intelligent terminal makes the cost of fault diagnosis system greatly reduced,and the possibility of directly arranging the diagnostic equipment in the actual project is greatly improved.The use of cloud plus terminal method improves the generality of fault detection method of electromechanical equipment and the accuracy of fault diagnosis,so that the diagnosis system has the advantages of both intelligent terminal and cloud computing.In addition,the practical application of real-time fault diagnosis for electromechanical equipment has been advanced.