The Researches On Fault Analysis And Fault Forecast Technology For Large-caliber Guns

The fault forecast technology is an effective means to improve the combat readiness and mission reliability for large caliber guns that often carry out high-strength training and battle tasks. This paper combines with the army's key scientific research task-"Dynamic Simulation and Fault Forecast for Large Caliber Guns", takes large caliber guns as research objects, adopts the integration of advanced dynamic fuzzy comprehensive evaluation theory and multi-Agent synchronous reasoning technology, and investigates the theory of fault forecast technology and its application on large caliber guns systematically and deeply. Based on the system's dynamic characters in future tasks, a scientific fault forecast model is established and an application software package is developed.The main contents involved in this paper are as follows:1. The deficiencies of traditional maintenance technologies are pointed out. Considering the request of modern war to equipment's mission reliability, the definitions of fault forecast technology and the predictive maintenance strategy based on fault forecast technology are put forward. The fault statistic analysis and the fault tree analysis are performed for large caliber guns. The fault trees of large caliber guns are established, which are the basis to ascertain forecast objects, seek reasoning paths and obtain forecast knowledge.2. Considering the characteristics of large caliber guns, a hybrid-reasoning fault forecast model integrated with dynamic fuzzy comprehensive evaluation and multi-Agent synchronous reasoning is firstly presented. The forward reasoning process based on dynamic fuzzy comprehensive evaluation creates a candidate fault set. And the backward reasoning process based on multi-Agent synchronous reasoning confirms the actual faults and find out the corresponding fault reasons and fault components. The model combines forward reasoning and backward reasoning, different knowledge representations and reasoning mechanisms into a uniform reasoning framework, and resolve the puzzles of multi-fault comprehensive evaluation and the mixture fault forecast of quantitative and qualitative faults in complex systems.3. Aiming to the deficiencies of traditional fuzzy comprehensive evaluation method, the dynamic weight coefficient is adopted into fault forecast. And the definition of dynamic membership and dynamic fuzzy comprehensive evaluation method are put forward. The fault emergence and development process is reflected by the dynamic variety and the...