Modern fault diagnosis method for electrical equipment

With the development of economic construction and the increase in the degree of electrification, electrical equipment has been widely used in various fields of industrial production, and the timely and accurate detection of potential and existing faults in electrical equipment is an important measure to ensure the safe operation of production. Therefore, the research on fault diagnosis theory and technology of motor equipment under different operating conditions and different operating conditions is the guarantee for reliable operation of equipment. Electrical equipment is the basic component of the power supply and power supply system. During operation, it is subject to various factors such as electricity, thermal machinery, and the surrounding environment. Its performance gradually deteriorates, eventually leading to failure; once a failure occurs, even if it stops working Very short, it will also cause great losses. The occurrence of these faults is always manifested by various symptoms, and the types of faults are also varied: both gradual and abrupt faults, and electrical and mechanical faults; There is a linear system failure and a non-linear system failure. The relationship between them is complex and intricate, which brings certain difficulties to the effective and rapid diagnosis of motor equipment faults. This article systematically describes the basic principles of the fault diagnosis of electrical equipment and domestic and foreign Modern fault diagnosis method, and analysis of various diagnostic methods, pointed out the development trend of fault diagnosis technology of electrical equipment. 2 The principle of fault diagnosis of electrical equipment The operation of electrical equipment is affected by many factors, such as grid voltage, load properties, installation environment, products. Quality, etc., Harsh environment and operation in the ultra-technology range are the cause of faulty production. The main reason of life is that the working principle of the electrical equipment is based on the electromagnetic theory, which is mainly composed of two parts: a circuit (winding) and a magnetic circuit (core). The transformer is a stationary device, and the motor is a rotating device. Their fault formation process and The expressions are similar in many ways: overheating of the windings of the equipment, aging of the insulation, deformation of the iron core and eccentricity of the rotor of the motor, etc. These signs are gradually deteriorated and lose their original performance. Effectively pass various detection techniques and signal analysis theory in time. Separating abnormal status information and diagnosing hidden faults are important measures to achieve reliable operation, reduce maintenance, and increase production efficiency. Current basic principles of fault diagnosis of electrical equipment are: 1) Current analysis method. The waveform is tested to diagnose the cause and degree of failure of the motor equipment; 2) The insulation diagnosis method uses various electrical test devices and diagnostic techniques to judge whether the insulation performance of the motor equipment is defective or not, and to predict the insulation life. 3) Temperature detection method using various temperatures The method is used to monitor the temperature rise of various parts of the motor equipment. The temperature rise of the motor is related to various fault phenomena. 4) The vibration and noise diagnosis method detects the vibration and noise of the motor equipment and processes the acquired signal. The cause and location of the motor failure, especially the diagnosis of mechanical damage is particularly effective 3 The modern method of fault diagnosis of motor equipment 3.1 Diagnosis method based on signal transformation Many fault information of the motor equipment exists in the form of modulation in the monitored Among electrical signals and vibration signals, if these signals are demodulated by means of a certain transformation, failure characteristic information can be easily obtained to determine the types of failures occurring in the motor equipment. Commonly used signal transformation methods are Hilbert transforms. And wavelet transform> 9. The specific definition of Hilbert transform can be seen, using Hilbert transform to realize the motor bearing and induction motor rotor fault diagnosis. Wavelet transform is not only a time-scale analysis but also a time-frequency analysis. It has the characteristics of multi-resolution and has the ability to characterize the local characteristics of the signal in both time-frequency domains. It uses the singular points of the wavelet transform (such as zero-crossings and extreme points). ) In the multi-scale comprehensive performance to detect the local mutation point of the signal, so as to carry out fault diagnosis. By preprocessing the stator current and performing the second wavelet transform, the fault characteristics of the stator winding of the motor are effectively extracted, and the diagnosis result of the stator current change caused by the external load mutation and the external current asymmetry is almost not affected by the load. Effective and reliable online diagnosis of motor faults. Fault diagnosis based on wavelet transform is used to accurately diagnose single-phase short-circuit faults of synchronous motor. The wavelet transform is used to realize the fault diagnosis of motor bearing of shearer haulage unit. The wavelet transform is used to realize online fault diagnosis of transformer. Fault diagnosis method based on signal transformation has achieved a lot of results in the practical application of fault diagnosis of electrical equipment; especially wavelet transform is very suitable for detecting transient anomalies in entrainment in normal signal analysis and displaying its components in mechanical failure of electrical equipment. Diagnosis plays an important role. However, the diagnosis method based on signal transformation lacks the learning function. 3.2 Diagnosis Method Based on Expert System The diagnosis method based on expert system is based on the past experience of the experts of the system being diagnosed, and it is summarized into rules, and rule of reasoning is used to conduct fault diagnosis by using the experience rules. . This method was used to establish a camera fault diagnosis expert system, set the cause of the malfunction of the camera, rotor imbalance, oil film oscillation, shaft misalignment, shaft cracks, frequency doubled resonance, frequency division resonance, stator voltage too high, stator current When the temperature is too high, the temperature is too high, and the signal of the fault sign uses a noise signal, and the spectrum of the noise signal is analyzed to extract the characteristic spectrum. When a certain fault occurs, it must be reflected in the spectrum chart but the correspondence relationship is from the book and It is difficult to find on site, for this purpose, by diagnosing the past “experience” of fault diagnosis experts and performing fault diagnosis. The system also uses self-learning control strategies to improve the knowledge base and increase the system's self-diagnosis capability, which greatly improves the diagnostic speed and accuracy Diagnostic methods based on expert systems have the advantages of simple and rapid diagnostic process, but there are also limitations The method based on expert system belongs to inversion reasoning, so it is not a reasoning method to ensure uniqueness. This method has the bottleneck of obtaining knowledge. The connection between the symptoms observed in the complex system and the corresponding failures is quite complex. Experts' experience is often not unique and is quite difficult. Therefore, this method is not suitable for troubleshooting complex motor equipment or new and unexperienced electrical equipment. In addition, rule-based methods are not only the rules that are used for the diagnosis conclusion but also the rules used repeatedly. Can not make further explanations outside, usually only diagnose a single fault, it is difficult to diagnose multiple hidden hidden 3.3 Diagnostic method based on fuzzy theory In the field of fault diagnosis, fuzzy attributes often appear, such as the description of signs: temperature "high", vibration " "Everything" has fuzzy characteristics; failures and signs Relationship is often blurred. Fuzzy theory is the best tool for dealing with such problems. There are two methods for fuzzy fault diagnosis. One is to establish a causality matrix R between symptoms and fault types first, and then establish a fuzzy relation equation between faults and symptoms, that is, F= At this time, F is a fuzzy fault vector; S is a fuzzy symptom vector: "." is a fuzzy synthesis operator, which is a diagnosis method based on the fuzzy relationship and the synthesis algorithm. Another method is to establish a fuzzy rule base for faults and symptoms first, and then perform a fuzzy logical reasoning diagnosis process. This is a diagnostic method based on knowledge processing. The method one implements the fault diagnosis of squirrel cage induction motor bearings. The method II is used to establish a fuzzy rule base for stator current and stator winding faults of the motor. On-line diagnosis of stator winding faults is achieved through fuzzy inference. The fuzzy linguistic variables are close to natural language, the readability of knowledge is strong, and the fuzzy reasoning logic is similar. The human thinking process is easy to explain. However, fuzzy diagnosis knowledge acquisition is difficult, especially the fuzzy relationship between failure and symptoms is more difficult to determine, and the system's diagnostic ability depends on fuzzy knowledge base, poor learning ability, easy to missed diagnosis and misdiagnosis. In addition, because fuzzy linguistic variables are represented by fuzzy numbers (ie, membership degrees), how to achieve the transformation between linguistic variables and fuzzy numbers is a difficult point in implementation. However, it has been a problem to introduce fuzzy theory into the field of fault diagnosis. In conformity with the inevitable trend of the nature of things.

3.4 Diagnostic methods based on artificial neural network Artificial neural network (ANN) is a complex, non-linear system that is widely connected by a large number of simple processing units. It has learning ability, self-adaptive ability, and nonlinear approximation capability. The task of troubleshooting from the perspective of mapping is the mapping from symptoms to failure types. ANN technology to deal with fault diagnosis problems can not only identify the complex fault diagnosis mode, but also assess the severity of the fault and make the system continuously acquire new knowledge in the process of operation, modify the rules: barrier prediction because lANgN can automatically acquire diagnostic knowledge Ml system has the ability to adapt.

The application of ANN to fault diagnosis of motor equipment is one of the hot spots in current motor equipment fault diagnosis. The basic idea of ​​using BP network for fault diagnosis of motor equipment is to use the sensor to obtain the characteristic signal that characterizes the fault of the motor equipment. The rotor current is used as the characteristic signal, the electrical equipment is mainly used for electrical fault diagnosis, and the motor equipment noise is used as the characteristic signal. Take mechanical fault diagnosis of electrical equipment as the main; then perform FFT transform on the acquired characteristic signal, the size and proportion relation of the feature quantity on the frequency characteristic band can reflect the corresponding fault type, so use several spectral peak energy in the spectrum of the characteristic signal. The value is taken as the input sample of the neural network, and the corresponding fault type is used as the output sample of the neural network. The network adopts the BI algorithm to learn and obtain the mapping relationship between the input sample (feature signal) and the output sample (failure type). Using neural network associative memory and distributed processing functions to diagnose faults in electrical equipment BP network has a strong ability of nonlinear approximation, can identify fault patterns, and can also assess and predict the severity of faults, so it is widely used. However, due to the gradient descent method used in BP algorithm, there is a problem of slow convergence, oscillation and local minimum. In addition, a prominent problem of BP algorithm for fault diagnosis is that it has low processing power and does not have the increment for abnormal faults. The learning function is mainly due to the fact that BP algorithm is essentially an interpolation method in mathematics, and its problem-solving ability is very dependent on the sample; when a new failure type with a large difference from the sample occurs, it is often attributed to a Known fault types or judged as normal conditions, misdiagnosis or missed diagnosis, affecting the reliability of the diagnosis Therefore, in the current use of BP network for motor equipment fault diagnosis, various BP improved algorithms are used. 3.5 Diagnosis based on integrated intelligent system Methods With the increasing complexity of motor equipment systems, it is difficult to meet the fault diagnosis requirements of complex motor equipments by relying on a single fault diagnosis technology. Therefore, the integrated intelligent diagnosis systems formed by the above-mentioned various diagnostic technologies have become the current research on fault diagnosis of motor equipment. Hot spot. The main integration technologies are: the combination of rule-based expert system and ANN, the combination of fuzzy logic and ANN, the combination of chaos theory and ANN, and the combination of fuzzy neural network and expert system. The combination of expert system and neural network can make full use of the expert experience of expert system and the strong nonlinear mapping ability of neural network.

The reasoning ability of the computer and the causality within the system, taking into account the existence of a large number of unknown conditions and the influence of knowledge inaccuracy, combined with the successful application of the fault diagnosis of the camera to the fuzzy neural network technology is to digitize the human experience and knowledge Fuzzy processing, the transformation of rules and reasoning into neural network mapping and the extraction of empirical rules directly from data samples, and then combining these two transformations for intelligent information processing techniques, it makes full use of the characteristics of neural network processing of digital knowledge and The characteristics of fuzzy logic processing structured knowledge are proposed. The fuzzy neural network is applied to fault diagnosis of brushless DC motor. The structure and learning algorithm of fuzzy neural network are given, and a threshold vector fault diagnosis method is proposed. The simulation results show that this method The effectiveness of the neural network consists of a neural network composed of chaotic neural network neurons. By studying its nonlinear dynamic characteristics, the chaotic attractor's trajectory and its sensitivity to initial conditions, a dynamic associative memory of the chaotic neural network is realized. Based on success Article breaking chaotic neural network fault asynchronous motor rotor and re-stored.

4 Development Trends of Fault Diagnosis Technology of Electrical Equipment The accurate and timely fault diagnosis of electrical equipment ensures the production is safe and stable, and it is of great significance to avoid the huge loss of personnel and property. Fault diagnosis technology is a cross-cutting science and has been operating since the 1960s. With the rapid development of traditional methods, new theories and modern fault diagnosis methods are constantly emerging: Wavelet transform expert systems, fuzzy systems, neural networks, etc. have all been successful in the field of fault diagnosis because of the fault symptoms and fault characteristics of electrical equipment. Complex nonlinear characteristics make fault diagnosis and identification more complicated. It is impossible to achieve accurate and timely fault diagnosis of electrical equipment in a complex environment based on only one theory or one method. Therefore, an integrated intelligent fault diagnosis system must be New Trends in Fault Diagnosis Technology of Electrical Equipment In addition, no matter what the diagnostic method is, the acquisition of real signals is the prerequisite for successful fault diagnosis. Multi-sensor data fusion theory will certainly play an important role in fault diagnosis, and this has begun. Research and obtain corresponding results

Komatsu Engine Parts

Komatsu Engine Parts,Komatsu Engine,Komatsu Engine Spare Parts,Engineering Machine Parts

JINING SHANTE SONGZHENG CONSTRUCTION MACHINERY CO.LTD , https://www.stszcmparts.com