Study On The Method Of The Remote Fault Diagnosis Of Printing Press Based On Signal Test

With the content of printing technology being improved, the fault of printing press has been integrated and complex increasingly. Once the fault can’t be resolved in time, it will cause a large number of raw material loss and unnecessary pollution which will directly affect company. It has been unable to achieve the purpose of rapid and accurate diagnosis, because diagnosing the fault of printing press mainly relies on experience and amplitude threshold detection, and the operators lack diagnostic knowledge. Signal test is a common method to diagnose the fault of printing press because the test signal parameters of the location of the fault are closely related to the type and severity of the fault.Based on the test signal, the deep studying on the typical failure mechanism of the printing press will integrate Support Vector Machine(SVM), the multi-information fusion and the Internet technology into the diagnosis system of the remote fault of printing press to realize the function of the system demand. The main contents of the paper are as follows:(1) Adopting the TPA method, studying the applicability of the common TPA method on the printing press, deeply analyzing the main causes and paths of the problem about the overprinting and noise of the printing press and building the TPA model separately support the diagnostic fault of the printing press.(2) Analyze the principle of Support Vector Machine based on the theory of machine learning. Put forward a method of abnormal detection based on One-Class according to the characteristics of the printing press. Study the selection of support vector kernel parameter and obtain the best one.(3) Propose a method of multi-information fusion fault diagnosis based on SVM probabilistic output, according to the characteristics of multi-information sources in the remote fault diagnosis system of printing press. The method can make full use of redundant information from each sensor to improve the accuracy of the diagnostic recognition(4) Select the key technology according to the demand function of the remote fault diagnosis system of printing press. Select MySQL as the database and the B/C network system to expound the communication modes between the subsystems.(5) Embed the technology of Matlab Web Server in the server, and realize the feature extraction, diagnosis calculation and signal reproduction based on Matlab software to realize the remote fault diagnosis system of the printing press. Taking a certain type of six-color offset press as the test object, collect the data, select the feature and the preliminary diagnosis result is obtained. Through the field test, correct the results and get better diagnosis results.