Research Data Engineering Machinery Remote Fault Diagnosis Process

With the development of science and technology, the level of intelligence,automation, integration of modern engineering machinery is constantlyincreased. As Engineering machinery generally works in harsh environment, andits fault mechanism is fairly complex and diverse, it could easily lead to themalfunction and the serious influence production operation efficiency. So it hasan important practical significance to monitor and diagnose the fault of theengineering machinery remotely. In the engineering machinery remote faultdiagnosis system, data is the basis of decision-making and diagnosis, so it isessential to process the data effectively. On one hand, when data was transferredfrom field devices to expert diagnostic center, in order to ensure the efficienttransmission of data and to avoid data loss or be tampered with, at the same time,to reduce the data storage space, it is needed to choose the appropriate methodfor data compression, encryption and transmission; on the other hand, whentroubleshooting in the diagnostic center, in order to avoid an inaccurate andunreliable diagnosis result which is caused by the too single failure state, it isneeded to take advantage of multi-sensor information fusion technology to fusethe multi-source information, so as to make a comprehensive decision.On the basis of researching the architecture of Internet of things, accordingto the actual demand for engineering machinery remote fault diagnosis system,the architecture of engineering machinery remote fault diagnosis system wasdesigned based on the Internet of things. And the relevant data processingtechnology in the network layer and application layer was studied.In the aspect of data compression, an improved compression algorithmbased on Huffman encoding algorithm was put forward to process binary streamdata. The algorithm respectively grouped binary stream as compressionprocessing units, and counted the frequency of binary code under each grouping,and then built the Huffman tree with heap sort algorithm to encode data. Finally,the minimum compression ratio value corresponded to the bit length of the packet was chosen to encode to achieve the optimal compression. Theintroduction of the heap sort algorithm reduced the complexity of thecompression algorithm and improved the response time of system. In terms ofdata encryption, RC4stream encryption, whose principle was simple andtechnology was mature, was used to encrypt the data; in the aspect of datatransmission, a communication mechanism of Socket was used for datatransmission. At last, in VC++environment, three direction vibrationacceleration data were realized a safe and reliable transmission when thehydraulic pump had a ball head looseness fault.In the terms of fault diagnosis in the application layer, on the basis ofresearching multi-sensor information fusion technology, from the sourceevidence itself, an improved algorithm based on the evidence weight wasproposed, and it overcame the disadvantages of dealing with high conflictevidence on D-S evidence theory. A numerical example showed that theimproved algorithm could fuse the conflict evidence effectively and it convergedfaster than other improvements, what’s more, the synthesis of the results couldbe better than others. Applying the algorithm to fuse the characteristicinformation of the three vibration signal of pump housing and the temperaturesignal of leakage outlet, ultimately, the fault of the hydraulic pump wasdiagnosed accurately.