The Monitoring And Diagnosis Of Large Crankshaft Grinder Based On Blind Source Separation Method

The security and reliability of mechanical equipment are the key tasksof industrial equipment research. The condition monitoring is an importantmeans of maintaining normal operation of the equipment to work properly.So the condition monitoring of large crankshaft grinding machine, whichcan timely obtain the state characteristics of the equipment to diagnosismachine failure, is of great significance. For the implementation ofcrankshaft grinder condition monitoring and fault diagnosis, this paperpropose a new methods based on signal blind source separation, to achievemonitoring and diagnosis task through the collected equipment vibrationsignal data, and further give detailed studies in the aspects of theoreticalresearch and experimental methods of equipment monitoring and diagnosis,as well as the setup of research platform. Specific contents are as follow:（1）According to the principal characteristics of large crankshaftgrinder, the monitoring and diagnosis system is developed. In this paper,the spindle rotation component of device is taken as the research object,based on the analysis of the specific requirements of the monitoring anddiagnosis process, then an innovative research method is proposed, whichemploys the vibration signal as the dominant data to obtain useful signalupon the signal blind source separation mean. Furthermore raises theoverall architecture of system, taking into account the display and remotetransmission of collected data.（2） The blind source separation algorithm, especially the analysisof independent component analysis in the blind source separation method,is introduced. Then a new blind source separation algorithm based on Newton iterative method is proposed. Compared with the traditionalalgorithm, the new one makes the separation matrix well converged.Separation speed and effectiveness improved obviously.（3）Study the kurtosis algorithm based on statistical theory and thenegative entropy algorithm based on information theory is given, both areusually employed in the independent component analysis. According to theinner connection of the two algorithms, a new blind source separationalgorithm by combining the two methods as an optimization criterion isdeveloped. The experimental results proved that the proposed newalgorithm can complete the task of blind source separation effectively andefficiently.（4）The monitoring and diagnosis system of large crankshaftgrinder is implemented, including the determination of the signalcollecting scheme, the design of the hardware and software in this system,and the setup of system platform. Experiments are conducted as describedin the details as follows: first collecting grinding machine data in thenon-load and load condition, then processing those data with the newraised blind source separation method, the analyzing the main vibrationsource of the equipment, at last the monitor and diagnosis task iscompleted.