Design and Application of Fault Diagnosis Expert System for Missile Test Equipment

　　This test equipment is a very important component of the missile weapon system. It is responsible for testing the performance and parameters of the missile and determining whether the quality of the missile meets the technical requirements. It is an important gateway to determine whether the missile can be successfully launched. Its main test content includes a comprehensive inspection of the hardware and software of the missile system; checking whether the connection of each instrument in the missile system is correct and whether the work is matched; checking whether the connection between the missile system and other subsystems is correct and whether the work is matched; checking the reliability of the system flight software operation, etc. Its test results are an important basis for analyzing and judging the performance of the missile system.

　　The missile test equipment is mainly composed of a ground measurement microcomputer, an acquisition controller, a monitoring device, a power supply/simulator, etc., which are connected through the I/O bus and a dedicated bus interface. The analog quantity test circuit, the digital quantity test circuit, the switch quantity test circuit and other circuits are connected to the ground measurement microcomputer through the measurement and control bus. Its structural block diagram is shown in Figure 1.

　　expert system

　　This test equipment is a product from the 1970s. The instrument uses a large number of discrete components and has poor working reliability. As the equipment has been used for a long time and a lot of troubleshooting experience has been accumulated, the basic conditions for using the expert system to diagnose the equipment's faults are met. The expert system needs to achieve the following functions for the ground test equipment:

　　● Real-time monitoring of ground test equipment and response to fault signals;

　　● According to the fault phenomenon, use the expert diagnosis system to find out the cause of the fault;

　　● Propose troubleshooting methods or suggestions based on expert knowledge and experience.

　　Considering that the storage and organization of data should have the characteristics of reliability and association, the expert system adopts relational database technology and is developed with C++Builder (CB) front-end development tools to realize the application under Windows.

　　The knowledge-based fault diagnosis expert system is based on the knowledge in the knowledge base. A good knowledge base management system can facilitate the acquisition of knowledge and the maintenance of the knowledge base. In view of the characteristics of missile test equipment, such as the large variety of data, large data volume, heavy data management tasks, and the requirements for accuracy, real-time and scalability of various data, this fault diagnosis expert system considers using a relational database to store, manage and maintain various data. Its benefits are:

　　(1) Database-based knowledge base management is a new technology for knowledge base management. The tables in the database are not only convenient to create, but also easy for users to maintain; the development of software technology also makes it easier to implement the interface between the database and the application. Therefore, it is very appropriate and convenient to use a database, especially a relational database, to establish and maintain knowledge.

　　(2) Systematically organizing and storing knowledge and effectively managing it are issues that must be considered when building an expert system. Building a knowledge base based on a relational database and making full use of the functions of the relational database management system can facilitate the design of the knowledge base management system.

　　(3) The establishment and maintenance technology of the knowledge base based on the database provides a framework for the diagnostic knowledge base to be easily expanded and improved. A reasonable knowledge base organization and management system is also the basis for improving diagnostic knowledge in the future. By continuously accumulating experience and continuously improving the existing knowledge system, the intelligence level of the diagnostic system can be improved. System Design and Implementation

　　Hardware system architecture model

　　The hardware part of the system mainly collects and detects the analog test signals, digital test signals and timing test signals of the missile ground test equipment in real time, and uses serial communication with the computer to provide the expert system with real-time information collection. Its structural block diagram is shown in Figure 2.

　　Expert system structure design

　　Through the analysis of the possible failure of the ground test equipment, there are two different fault diagnosis methods. One is the precise fault diagnosis mode for reasoning on certain factors; the other is the fuzzy judgment fault diagnosis mode for multi-factor fuzzy judgment. Therefore, different methods are adopted in the mechanism of knowledge acquisition, representation and reasoning. Figure 3 shows the structural model of the expert system.

　　An important function of the expert system is the ability to learn new knowledge. The knowledge base of the expert system can be improved and updated according to the needs of different models. The fault library system in the expert system undertakes the important task of preparing learning samples for the learning system. The fault instances of the system can be recorded in the form of {original signal, diagnosis conclusion, maintenance measures}. Such recorded instances are more versatile and can be used by multiple expert systems in theory. The fault diagnosis expert system of different models of ground measurement equipment can use its own symptom recognition system for identification and convert the instances into a form suitable for learning. The expert system has a special application window for knowledge acquisition, including operations such as adding, modifying, and deleting. In response to the different requirements of precise fault reasoning knowledge and fuzzy fault judgment knowledge, different knowledge acquisition window applications are made respectively. Knowledge can be acquired from maintenance experts and recorded maintenance data, or through system self-learning.

　　In the expert system, the core link is the inference engine. The inference knowledge is stored in the corresponding dynamic database. Knowledge inference controls the operation of the entire system and enables it to solve problems according to certain inference strategies. Since there are deterministic knowledge and fuzzy knowledge in fault analysis, knowledge inference uses two different methods: state chain inference and fuzzy inference.

　　The user conducts on-site inspection according to the system prompts to determine the fault phenomenon. The system will then give various reasons for the phenomenon and give corresponding troubleshooting suggestions. If the fault has not been eliminated, other reasons will be found until the fault is eliminated. Fuzzy reasoning obtains the evaluation fuzzy matrix through weight determination, weighted average model and quantitative index synthesis, and makes a comprehensive decision on the evaluation results.

　　Conclusion

　　The main innovations of this fault diagnosis expert system are: it realizes load-free dynamic access and multi-channel intelligent scanning, and adopts full-online isolated hardware testing technology, which solves the technical problems that the detection and diagnosis equipment interferes with the equipment under test due to full-time access, and cannot be online in real time due to weak monitoring capabilities, and requires the cooperation of the equipment under test to complete the diagnosis; it adopts a reasoning method that combines state chain reasoning with fuzzy reasoning, and uses a fault self-learning adaptive algorithm to optimize the description method of state and fault, improve the efficiency and positioning accuracy of state identification and fault diagnosis, and ensure the simultaneous realization of online monitoring, dynamic tracking, and fault diagnosis; it adopts modular design technology, variable classification packaging technology, and reusability design technology to realize the separation of data processing and diagnostic logic in the monitoring and diagnosis software, which greatly improves the software's operating efficiency, reliability, and portability.