Research On Knowledge-driven Intelligent Design Method Of Ship Coating Process

In the process of ship construction,ship painting is an important part of "shell,outfitting,and coating".At present,the ship coating process design mostly relies on the experience of on-site technicians,and there is no scientific process design process and specifications.At the same time,the coating process card only specifies the coating matching used in the coating operation.The selection of equipment,coating process parameters,etc.depends entirely on the technical level of the construction personnel,and it is difficult to achieve scientific management of the coating process and coating quality control.Under such circumstances,it is urgent to carry out research on the intelligent design method of ship coating process and develop the corresponding process design system to promote the intelligent and scientific process of coating process design.The main research contents are as follows:(1)According to the characteristics of ship coating process and data,combined with the requirements of coating process design,carry out process knowledge modeling and classification technology research based on the acquired coating process knowledge,and use the three-tuple modeling method to paint The knowledge of the coating process is visually expressed,and the knowledge of the coating process is divided into the knowledge of the coating object,the knowledge of the coating matching and the knowledge of the coating process parameters.On this basis,the ship coating knowledge base is established,and the coating process knowledge is stored in the database.,Lay the foundation for knowledge-driven process design.(2)Propose a hybrid classification algorithm of random forest and comprehensive similarity for painting objects.First,pre-process the painting process data,adopt standardized processing to eliminate errors and inconsistent texts,and normalize the original data to eliminate the dimensional influence between the data in the ship painting process;then,aim at the target of the painting The data set is classified and studied.According to the characteristics of the painted object,the random forest algorithm is used to classify the characteristics of the object to be painted into a piece of knowledge,and the set of painted objects corresponding to the piece of knowledge is obtained;finally,the comprehensive similarity threshold is set to reduce For the range of painted objects,obtain a set of similar painting objects,and combine the knowledge of coating matching and painting process parameters to obtain a set of painting processes for similar painted objects.(3)Establish a multi-objective evaluation method for ship coating.Through the study of the analytic hierarchy process,and based on the analysis of ship coating process objectives and quality requirements,a multi-objective evaluation model for ship coating is established with the single goals of coating consumption,coating manhours,and coating quality.Calculate the target weight according to the ship coating hierarchy model,and calculate the score in combination with the label weight,realize the scoring of the coating process,and obtain the coating process score list.Finally,select the coating process through the coating process score list.(4)Developed ship coating process design and management system.The overall realization framework of coating process in the knowledge base is described,and functional modules such as knowledge management and process design of the coating design system are constructed,and actual data from a shipyard is used for example verification.The results verify the feasibility and effectiveness of the system.