Motor condition monitoring the results of the evaluation methods of fuzzy logic
..........................................................................................................................................................................................................................

Abstract: This article describes the selection of the vibration motor fuzzy parameters as input parameters, using fuzzy logic method for small and medium-sized motor vibration monitoring results of a comprehensive assessment of the principles and methods to achieve. Lists some of the experimental data to verify the feasibility of the method.
Key words: condition monitoring; motor; fuzzy logic approach
Key words: TM343.2 Document code: B
 
Xinjiang Dushanzi petrochemical refinery with all kinds of small and medium-sized three-phase asynchronous motor of more than 2,000 units, mainly on the 90KW more than 200 Class A motor vibration test, the main use of Vm-63 portable vibration device, to determine the basis for the vibration rate And the value of the standard rate of value.
For the production of the actual day-to-day maintenance and management needs, we have taken place in recent years, the three-phase AC induction motor made of mechanical failure statistics, analysis of the causes of motor vibration and vibration monitoring of the value of the characteristics of the monitoring data to identify fuzzy, Be sure to monitor the operation of equipment for the operation of the electrical fault and the type of state to provide an effective criteria to ensure the reliable operation of the motor.
First, three-phase asynchronous motor and the characteristics of the fault diagnosis
Day-to-day maintenance of motor repair practice confirmed that the common electrical failure of the main reasons there are electromagnetic vibration, no motor, the rotor imbalance, bearing wear, lubrication poor basis and lack of rigid mechanical loosening.
Measurement of electric motor failures can run current, direct current motor winding resistance and insulation and other methods to judge; mechanical failure diagnosis often used to monitor vibration, shock pulse method, the current diagnostic method and spectrum analysis, and so on.
According to the vibration motor speed to the ISO2373 standard motor six points position, vibration acceleration, speed, displacement amplitude characteristics, as well as the use of the apparatus function to determine the vibration monitoring project for the monitoring of the vibration amplitude, the value of speed, acceleration Hi value of the high-frequency and L, according to the electrical structure of the abolition of small and medium-sized non-axial end of the axis stretching measurements.
Motor is running, there are many uncertainties and therefore their fault diagnosis there is a certain degree of ambiguity, the main problems: (1) diagnosis of fuzzy parameters. Diagnosis for the parameters of the vibration frequency range with a certain degree of ambiguity, different parameters, location and direction of excessive vibration value, as reflected by the existence vagueness of state failure; (2) ambiguity methods of diagnosis. Caused by electrical failure is often not a single, for example, bearing damage, and so on the basis of fault often loose by the electromagnetic vibration, and do not move or cause an imbalance, so what use or which of several methods of integrated fault diagnosis with ambiguity; ( 3) the diagnostic criteria of ambiguity. Diagnostic criteria for fault diagnosis on the basis of which criteria to use to diagnose more accurately, on-site temperature, load, and other operating conditions change, the diagnostic criteria for how the amendment would affect the accuracy of the diagnosis, and there is a big ambiguity.
As a result, multi-parameter vibration analysis of the fuzzy diagnosis with the motor fault diagnosis of actual cases can be proved to be effective.
Second, the fuzzy diagnosis model
Since the motor is set up for each diagnostic criteria, therefore, direct method of pattern recognition.
Multi-factor fuzzy comprehensive evaluation of the vector-calculation for the Y = R X ', of which: fuzzy relationship matrix of the R line, said vector vibration characteristics of the fault, vector out that the reasons for failure, "" fuzzy logic operator said.
As the general matrix multiplication rule to avoid the loss of fault information, fuzzy logic operators use an ordinary matrix multiplication of the law.
Membership function to reflect the extent of the possibility of failure. Points can be judged similar to the way of symptoms to determine the domain membership, is about to monitor the various parameters and their corresponding standards division fault diagnosis and appropriate amendments to provide for membership. The method is simple and practical, proven, in line with the common vibration motor fault diagnosis request. If the method of calculation formula (1)
Where: Ux = membership;
      Xi = failure characteristics of the parameters of the actual value;
      Sx = was diagnosed with motor monitoring the corresponding parameters of the diagnostic criteria;
      Mx = correction factor.
Fuzzy comprehensive diagnosis to choose the appropriate diagnostic criteria. As the motor rotor to be at work under a variety of complex and the alternating stress, and some of the fault diagnosis is absolutely not applicable standards for the accuracy of the diagnosis, or worse, the use of relatively more reasonable standard to judge. The state of their own motor for vertical before and after comparison, the parameters change to more accurately reflect the state of change, determine the reliability of some of the larger, final statistical method for each motor to establish the relative standard.
As the motor to monitor the site to monitor the different parameters of different parts of the operation, and the motor parts of the impact of varying degrees, are therefore required to set up corresponding to the threshold, then the largest membership in principle to determine the most likely reason.
The principle is the threshold, where membership is greater than the threshold of the corresponding failure is the reason why, and one of the largest membership is the most likely cause of the malfunction. Through various types of electrical fault monitoring aggregate statistical data to calculate their membership, obtained similar failure of the motor each corresponding to the value of the fault diagnosis, and combined with the impact of the size of the equipment to determine the threshold corresponding to failure.

The use of expert knowledge to build Knowledge is the key to fuzzy reasoning, fault diagnosis fuzzy matrix reflects the failure and the reasons for the relationship between the signs of failure. And the reasons for failure between the complex, from the sign in order to find out the reasons, to be set in advance with the sign of the reasons for the correlation between the degree, that is, fuzzy diagnosis matrix.
Through the analysis of day-to-day electrical maintenance information and the corresponding state monitoring data, for all the state's failure to monitor the electrical fault types of identification, calculated each motor attached to the diagnostic criteria of value attached to that standard, obtained under the similar failure of each Monitoring of parameters in the corresponding standards for membership than the corresponding number of the membership, will monitor the parameters of the percentage of the reasons and as a sign between the weights in order to constitute a fuzzy diagnosis matrix.

Third, diagnostic examples
A key equipment for the oil pump.
1. Analysis of the process. In April 2003, the violent vibration of the motor, bearing more noise, diagnosed as non-performing dynamic balance and the former bearing wear. Around BEARING added to the grease, the value of the vibration is still large. After the overhaul in June when the rotor was a balancing test and found that there were two ends of rotor 125g and 99.7g of imbalance, and 7.02g be adjusted upward to 19.5g, and before and after the replacement of the bearings to eliminate vibration, operation Good.
2. Diagnosis. Through the different stages of the monitoring data analysis, such as the value has been attached to the table 5, fuzzy combination of matrix and the threshold for diagnosis together.
IV, Conclusion
After on-site inspection, the fuzzy diagnosis method and fuzzy matrix and the correction factor can be more accurately diagnosed with common electrical failure, and the disintegration of the maintenance of basically the same as the actual situation, can be better for the electrical maintenance and management of day-to-day basis, To meet the device management requirements.