Large And Complex Curved Hull Plate Forming Process Parameters Self-Leaming Expert System

In line heating of shipbuilding, large surface curvature of line heating forming mainly rely on the experience of workers, this way not only efficiency is very low, and the processing quality is volatile. Due to the large curvature plate, calculating forming status information needs more times, also flame parameter types are different insure forming status, too many factors affect the forming of plate and line heating plate intelligent machining system can not accurately predict process parameters, finally, they will hinder the development of modern shipbuilding. In above problems, domestic and foreign research have solved problems of technological parameters of line heating plate and shell shrinkage relationship model, prediction and surface matching of smaller curvature, but still exist some following problems, large curvature plate three-dimensional match, forecast of flame types and accurate three-dimensional fire road location prediction, has not solved well. This paper mainly aimed at above unsolved problems to developed expert systems to solve such problems.For such three problems, main work of my paper is as follow:(1)To do some research of the description of the expand of surface, coordinate transformation and matching technology, flame prediction algorithm and machine learning algorithms.(2)Studying the three-dimensional coordinate conversion technology which can solve the unification of coordinate system of measured surface and forming surface.(3)Studying the differences between forming surface and measured surface which could solve the problem of fitting detecting surface and forming surface.(4)Researching relationship between curvature surface forming state parameters and flame types, using machine learning techniques to study their relationship, enables expert system to identify forming state of large surface in specific forming state and predicts flame types.(5)Researching cross flame, longtitude flame and cross-long flame to solve the problem of accurate three-dimensional flame location information forecast.(6)Finally we developed expert system software to complete such function.Coming up with such problems, description and quantitative method of forming parameters and method of matching forming surface and measured surface has been proposed to decide measured surface is forming or not and used machine algorithms to learn relationship between parameters of description of forming surface and flame types, finally, proposed three dimensional flame location prediction algorithm. Finally, we improves degree of predicting accuracy of expert system, speeds up the process automation level of operation of the ship, plays a positive role to the development of defense force, and contributes to popularization and application.