Research On Aluminum Alloy Welding Intelligent Knowledge System Of High-speed Train Body

With the development of welding technology, digital management and sharing knowledge of the welding increasingly highlighted its importance. Building intelligent knowledge system becomes more and more significant for improving welding knowledge system and accumulating and updating the knowledge of welding processes, accelerating the the progress of the welding digitization and providing better services for product project.Based on these requirements in welding industry, the paper studies a welding databases and expert systems to optimize the welding process and improve product reliability. The system achieves automatic welding process design, optimization of welding process qualification, management of welding data, and the check of the knowledge base.Firstly, the overall knowledge structure of the high-speed train body were studied, for aluminum welding expert system. A set of knowledge systems are established with the multi-level knowledge by combining development environments and operating environment Uniform methods of knowledge is used for processing knowledge duplicated coverage and theblank ranges of knowledge are covered by human experience of experts and engineers. In order to quantify the reliability of knowledge of the knowledge base, this paper introduces the concept of different confidence levels for knowledge source.Secondly, the paper investigates the acquisition mechanisms for different kinds of aluminum alloy welding knowledge by classifying them with uncertain knowledge, rules of theory and rules of thumb. The knowledge is represented with different presentation methods including production rules and first-order predicate logic and the knowledge base and comprehensive database is therefore established for, constructing the high-speed train body aluminum welding system.The paper further studies knowledge verification system based on B-P artificial neural network models and propose a checking processes for the mechanical properties of welded joints which can commonly used as an indicator of the knowledge base. The difference between the proportion of the predicting result and requirements of the mechanical properties in practical application is used to determine the actual value of the welding process knowledge floating confidence.Finally, a platform for the development of this system is built by means of appropriate technology and development tools, and the aluminum welding expert system is developed based on browser / server(Browser / Server) network model. Finally, the theory in the paper is validated by the system with designing welding procedures and welding procedure qualification.