At present, the research on detection and diagnosis technology in China mainly focuses on the following aspects:
(1) Research on sensing technology: Sensing technology is an instrument technology that reflects the state parameters of equipment. Various types of sensors have been developed in China, such as eddy current sensors, speed sensors, acceleration sensors, temperature sensors and so on. Recently developed sensing technologies include optical fiber, laser and acoustic emission.
(2) Research on signal analysis and processing technology: From traditional spectrum analysis, time series analysis and time domain analysis, some advanced signal analysis methods are introduced, such as fast Fourier transform, Wigner spectrum analysis and wavelet transform. The introduction of this new method makes up for the shortcomings of traditional analysis methods.
(3) Research on artificial intelligence and expert system: This research has become the mainstream of diagnosis technology. At present, there is an expert system for scheduled mechanical fault diagnosis, but the research on this technology in engineering has not reached the expected level.
(4) Research on neural network: For example, the research on neural network classification system of rotating machinery has been applied and achieved satisfactory results.
(5) About the development and research of diagnosis system: From single machine detection and diagnosis to the master-slave structure of upper and lower computers, and then to the network-based distribution system, the structure is becoming more and more complex and the real-time performance is getting higher and higher.
(6) Develop special and portable diagnostic instruments and equipment. At present, the fault diagnosis technology in metallurgy, electric power, chemical industry and other industries in China is very mature and has been widely used.
2 Modern fault diagnosis methods
The running state of construction machinery is very different, and the faults are also varied, and the diagnostic methods used are also different. Among many diagnostic methods, the commonly used diagnostic methods are vibration monitoring diagnosis, nondestructive testing technology, temperature diagnosis and ferrography analysis. In recent ten years, new diagnostic technologies such as fuzzy diagnosis, fault tree analysis, expert system and artificial neural network have appeared continuously, and fault diagnosis technology has gradually developed to intelligence.
(1) fault tree diagnosis method
Fault tree diagnosis method is based on the most undesirable fault state (result) in the research system, and it is gradually refined from the whole to the part according to a certain logical relationship, and the causes of the fault are analyzed by reasoning, and finally the initial basis of the fault is determined.
This reason, the degree of influence, the probability of occurrence. It is a graphic deduction method, which vividly draws the system fault and various factors leading to the fault into a fault diagram, which can intuitively reflect the relationship between the fault, components, system, factors and reasons, and can also quantitatively calculate the degree, probability and reasons of the fault. This method is intuitive and fast, and the knowledge base is easy to be dynamically modified. However, the disadvantage is that it is greatly influenced by subjective factors, and the diagnosis result depends heavily on the correctness and completeness of fault tree information, so it is impossible to diagnose unpredictable faults.
(2) Fault diagnosis expert system
Expert system is an artificial diagnosis system based on knowledge, and it is an artificial intelligence program that uses the knowledge and reasoning methods of a large number of human experts to solve complex practical problems. Fault diagnosis expert system is a kind of intelligent diagnosis technology which is researched and applied most. It is mainly used for complex systems that have no accurate mathematical model or are difficult to establish mathematical model. The main problems of expert system are the difficulty in obtaining knowledge and the slow running speed. The fault diagnosis expert system developed on the basis of advanced sensing technology and signal processing technology, which combines the advantages of modern science with the advantages of rich experience and thinking mode of experts in this field, has become the main direction of fault diagnosis technology development.
③ Fault diagnosis method based on fuzzy mathematics.
The propagation path of state signals of construction machinery is complex, the mapping relationship between faults and characteristic parameters is vague, the uncertainty of boundary conditions and the variability of operating conditions make it difficult to establish an accurate corresponding relationship between fault symptoms and fault causes. It is obviously unreasonable to use traditional binary logic, so we choose membership function and describe the tendency of these symptoms with corresponding membership degree. The fault diagnosis method based on fuzzy mathematics is to get the membership degree of various fault causes through the membership degree of some symptoms and fuzzy relation matrix, so as to express the tendency of various faults, thus reducing the difficulties brought by many uncertain factors to the diagnosis work. However, it is very difficult to establish correct fuzzy rules and membership functions for complex diagnosis systems, and it takes a lot of time.
(4) Fault diagnosis method based on neural network.
Neural network is an information processing system, which aims to imitate the working mode of human brain. It has a large number of processors connected in some way and distributed in parallel. Fault features are extracted from the information of various systems of construction machinery, and fault decision rules are determined by learning training samples, so as to carry out fault diagnosis. Neural network used for fault diagnosis can constantly adjust the weights of new faults through self-learning, improve the correct detection rate of faults, and reduce the false alarm rate and false alarm rate. Neural network has the ability of associative memory, pattern matching and similar induction, and realizes the complex nonlinear mapping relationship between faults and symptoms. For complex construction machinery with multiple faults and processes, as well as sudden faults or other abnormal phenomena, the causal relationship between the causes and symptoms of faults is complex, so it is effective to solve them with the help of neural network system.
(5) Support vector machine fault diagnosis method.
The serious lack of typical fault data samples is one of the main reasons that restrict the development of fault intelligent diagnosis technology. Support Vector Machine (SVM) is a new machine learning method based on statistical learning theory. Its goal is to get the optimal solution under the existing information, not just the optimal solution when the number of samples tends to infinity. This is especially suitable for solving practical problems such as small sample fault diagnosis.