Research on the architecture of the most popular c

Research on the architecture of case-based integrated CAPP system

Introduction

at present, CAPP system can be divided into four basic application modes according to its working principle, namely interactive, retrieval, derivation and creation. In addition, there are systems between derivative or retrieval and generative, which are usually called semi generative systems. Among them, the retrieval system can only play a role when new parts can find parts with good similarity; The derivative system also depends on the similarity of parts. At the same time, due to the complexity of parts, it is extremely difficult to establish an effective standard process with large coverage, and the workload of manual modification is often large when it is used; The interactive system mainly provides convenience for printing. These methods rely on manual decision-making, and the degree of automation is low, and the efficiency and applicability are often not ideal, especially they are difficult to integrate with CAD and NC programming systems. At present, generative or semi generative systems generally use feature technology to describe parts, realize the integration with CAD and NC programming systems, and greatly improve the work efficiency. However, due to the complexity of parts and the experience of process design, on the one hand, it is difficult to regularize process knowledge; On the other hand, it is a correct process decision for a single feature, but it is not necessarily appropriate for the overall situation of the whole part. Therefore, the practicality of generative or semi generative systems remains to be solved

therefore, how to comprehensively utilize various working principles and make them organically combine, how to establish a good CAPP architecture, so that the CAPP system can meet the higher requirements of current and future development in function, integration, automation, practicality and versatility, and how to seek a CAPP system with powerful function, good practicability and special mechanism naturally becomes an urgent problem to be solved

1 working principle of case-based integrated CAPP system

the case-based integrated CAPP system in this paper comprehensively utilizes CAPP systems with different working principles. At the same time, case-based reasoning method is embedded in CAPP expert system, and feature technology is used to complete the unified description of part information to realize integration. The process document library of existing parts based on relational database is established, and the feature coding matrix is established Knowledge base and inference engine

it should be pointed out that the proposed case-based integrated CAPP system is an organic combination of four working principles: interactive CAPP, retrieval CAPP, derivative CAPP and generative CAPP system. It uses the principle to operate directly against the process documents and standard process documents of existing parts. Instead of the simple addition of the four working principles, the system can be divided into two parts, namely, the system based on examples and the system based on knowledge rules. What is stored in the case-based system is the standard process that has been verified by practice. Using the case-based design method can make good use of previous experience, while the expert system based on knowledge rules is completely a creative decision-making process, which can solve some problems with high complexity that the case-based system cannot solve. The advantages and disadvantages of the two parts of the system complement each other and inherit each other, forming a comprehensive CAPP system with strong function, good practicability and strong adaptability

the diversity of process design, process decision-making knowledge and production environment constraints of complex shape parts requires that CAPP system must seek an organic combination to adapt to the diversity of various conditions and re make filaments. On the basis of adapting to the type of work pieces, the system is required to adopt a variety of decision-making methods and adopt a multi-level distributed system structure. Therefore, in the re integrated CAPP system, according to the specific situation of parts, based on the similarity of parts, select the working principle that is most conducive to improving the automation of the design process and the practicality of the design results. To this end, the workflow of an example based integrated CAPP system is designed as shown in Figure 1

Figure 1 the workflow diagram of the case-based integrated CAPP system

as shown in Figure 1, the system first determines the type of parts according to the part information, then establishes the feature coding correlation matrix with the existing parts according to the feature layer information, calculates the similarity coefficient, and after judgment, carries out the case retrieval, and after successful retrieval, calls and modifies the case; If the retrieval fails or the similarity is too poor, the system will use the method based on knowledge rules for process design. The detailed steps are as follows:

(1) firstly, the CAD geometric description information, process information and other necessary information of parts are extracted from the dynamic information base to meet the functional requirements of the integrated CAPP system and complete the task of process design

(2) after forming a feature-based CAPP part model through information conversion and modeling functions, calculate the values of each element of the part feature code incidence matrix, that is, the calculation of similarity coefficient, in which the judgment of similarity and the calculation of similarity coefficient are the key to what decision-making method to adopt in the subsequent process

(3) after judging the similarity, if the old and new temperature: 25 ℃ parts have good similarity, take out the process documents corresponding to the original parts with good similarity to the new parts or the standard process documents of the part family, and determine the positive and negative difference characteristics of the new parts at the same time

(4) delete the part of the process or work step corresponding to the negative difference feature in the process document or standard process document of the part, and the process of the positive difference feature can be inferred by the inference engine in the expert system. The reasoning process results are combined with the previous automatically modified examples, and new process documents are formed after the combination, sequencing and optimization of work steps. At the same time, NCP interface documents are generated and modified as needed and stored in the information base. Finally, the modified process documents or standard process documents are used as examples of process design of new parts

(5) after judging the similarity, if the similarity between the old and new parts is very poor, and the part information forms a composite feature, the process is completely created by using the expert system. After the optimization of the work step sequence, the process file is generated, and the tool position interface file is also generated for NCP to generate tool position trajectory and NC instructions. Among them, composite features include process composite features and function composite features. The introduction of composite features is also conducive to improving the consistency and practicality of design results and simplifying the working process of the system

(6) if the creation fails or is not satisfied with the result, it can be modified by human-computer interaction, and the result can be saved after modification

(7) after completing all the design tasks, the information base sorts out and outputs the design results according to the format required by each external system

(8) after the modification and maintenance of the case library and the inspection of the actual production, the process documents generated by the final new parts can be used as an example of the process design of new parts in the future, so that the design experience can be reused and the self-learning ability of the system is improved

the above aspects are the workflow of an integrated CAPP system based on examples. The selection of process design examples is as follows: Generally speaking, there can be no identical parts. Therefore, the choice of examples can only be as similar as possible, and the degree of similarity will directly affect the invocation and modification of examples. The purpose of using case-based reasoning CBR in CAPP expert system is to minimize the workload of process design and improve work efficiency. There are two ways of example selection in the system: one is that the system judges and makes automatic selection through the size of similarity coefficient; The other is a complementary human-computer interaction selection method. The former is automatically selected by the system, that is, to pick out the examples with the best similarity, and the latter can be implemented by the system operator. According to experience, select the appropriate examples from the examples retrieved by the system, so as to make up for the fuzziness in the process of automatic retrieval of examples

2 architecture of case-based integrated CAPP system

according to the needs of the working principle of the system described above, the architecture of case-based integrated CAPP is designed as shown in Figure 2. The architecture integrates the working principles of many kinds of CAPP, and successfully embeds the idea of case-based reasoning into the expert system of CAPP based on knowledge rules. It can make good use of the experience of previous process design. In the process of system implementation, if the examples of process design of new parts cannot be successfully selected, the advantages of rule-based expert system can be fully used for process creation

Figure 2 the architecture of an integrated CAPP system based on examples

(1) modeling of parts

the model of parts includes part definition model, manufacturing resource description model and process decision model. Among them, the description of part definition model is the key. The established part feature definition model must include part level information, feature level information and underlying geometric information. In addition to obtaining these information from CAD system, they can also be modified and added through human-computer interaction; Manufacturing resource description model includes manufacturing resource data and manufacturing resource knowledge description; The process decision-making model is a model that describes the strategy of process planning to generate process documents

(2) part similarity

the description of part similarity is mainly realized by calculating the similarity coefficient of each element in the part feature coding incidence matrix. If the similarity is good, the number of new and old parts with the same feature code is more, that is, the structure of new and old parts is very similar. The similarity judgment of parts is the key to the realization of the system

(3) case based reasoning

by using the idea of case-based reasoning, the system can make better use of previous process design experience. Among them, the retrieval of examples adopts the idea of retrieval and derivative CAPP system, which combines the retrieval and invocation of examples, and improves the overall work efficiency of the system; The modification of cases is expressed in the form of knowledge expression of the original knowledge-based system, which can be processed by using the reasoning mechanism in the original system, so it is also convenient for the development, maintenance and expansion of the system

(4) rule-based reasoning

the idea of rule-based reasoning is to use all kinds of knowledge bases to realize the process creation of new parts by calling the inference engine. Among them, in the process of reasoning, it is necessary to transfer all kinds of knowledge bases, databases and description models to solve new problems and new situations that have not been encountered in the process design of new parts through creative decisions

(5) other functions

knowledge acquisition function is to change the intermediate result of reasoning into new knowledge through the input and modification of management knowledge and man-machine dialogue; The reasoning explanation mechanism displays the reasoning process in a graphical way, which can display the design content used in the reasoning process and help users understand the generation process of process design results; The user interface interaction function modifies each design content through human-computer interaction, so as to achieve the purpose of reasonable design results

2.2 knowledge base

the knowledge base of the system is designed as follows according to the needs of the functions to be realized:

 feature processing chain generation knowledge base: used to store the rule base generated by the processing chain for features

 machine tool selection knowledge base: used to store the rule base of the selected machine tool

 tool selection knowledge base: used to store the rule base of tool selection

 tolerance conversion knowledge base: used to store various common values expressed in numerical form