1 Introduction
Overall design is a key step in the design process of machine tools. The final performance of machine tools and the manufacturing cost depend not only on the design quality of the parts and other details, but more importantly on the quality of the overall design.
In the increasingly fierce market competition, the product life cycle is getting shorter and shorter. The ability to quickly develop new products with high efficiency, high quality, and low cost has become a decisive factor in the occupation of the market. This makes people pay more and more attention to the new design and manufacturing mode of concurrent engineering. The core of concurrent engineering is to enable product developers to consider all factors in the product life cycle (including design, analysis, manufacturing, assembly, inspection, maintenance, cost, and quality, etc.) as early as possible in the design process. This requires that there must be a multi-objective evaluation system, specifications and methods (eg, performance-oriented, manufacturability, assemblyability, etc.) that are oriented to the entire product life cycle process for distributed parallel processing and collaborative decision making. Otherwise, the application of concurrent engineering is empty talk. From this perspective, the overall performance-oriented design evaluation is not only the need to improve product quality and reduce costs, but also lays the foundation for the implementation of concurrent engineering. It is a part of the multi-objective collaborative assessment system, and is the application of concurrent engineering in the design and manufacture of machine tools. Lay the foundation.
This article is based on the above ideas, established a system model to achieve machine tool product performance evaluation in the overall program design phase.
2 overall design ideas
In the working process of the machine tool, various static forces, dynamic forces, and temperature changes are generated, which can cause the machine tool to deform, vibration, noise, and unsteady motion. This has an adverse effect on machining accuracy and productivity. The performance of the machine tool mainly includes two major aspects: geometric accuracy and work accuracy. The former mainly assesses products according to national standards (parts), while the latter mainly refers to the influence of temperature changes on machining accuracy during the work of the machine tool. The overall design of the machine tool is based on the business decision-making and other aspects of the company's decision to produce the product type, performance and specifications, the initial program decision made on the overall plan of the machine tool, as a basis for future detailed design. Its purpose is to ensure the best overall, optimize design, reduce costs, and shorten the development cycle. It mainly includes the following three parts: technical parameter design, overall layout design and structural optimization design. The pros and cons of the overall design scheme play a very important role in ensuring the performance of the machine tool products. Through the analysis and evaluation of machine tool performance in the overall design stage of the design, the designer can help improve the design scheme continuously so that high-performance machine tool products can be manufactured in the future.
In the overall design phase of the machine tool, only a preliminary plan for the main components of the machine tool is determined, and it is not possible to provide very complete information for the evaluation system. Therefore, it is difficult to make an accurate quantitative assessment of all the properties of the machine tool at this stage. Most performance can only be pre-assessed based on past experience. Based on this feature, this model integrates the use of advanced computer technologies and integrates the expert system technology in artificial intelligence with a multi-level fuzzy comprehensive evaluation method to realize the intelligent design of the overall design of the machine tool. Evaluation to ensure the best overall design from a performance point of view.
3 System Architecture
The system model proposed in this paper uses the architecture shown in Figure 1. In the system, engineering experts are simulated by the expert system. Through analysis and inference of the overall design information obtained, the future product performance indicators are forecasted, and a multi-level fuzzy comprehensive evaluation method is used to achieve product performance based on the prediction results. Evaluation.
Figure 1 System Model Architecture
Inbound interface: mainly responsible for the interaction between the user and the system, and provide designers with real-time participation functions.
Information acquisition module: It is mainly responsible for obtaining all the information needed for evaluation from the external overall design system and storing it in the fact base. The description of the overall plan design information here refers to the idea of ​​the classification layer description method proposed in [2], but it has been improved based on its own requirements on specific details. If users use different information description methods, a corresponding conversion mechanism should be added in this module, that is, the information is converted into an internal agreed format that the system can recognize.
Expert system: The knowledge base is the core of the expert system. The knowledge base of the expert system includes the meta knowledge base and the domain knowledge base. The knowledge base is made up of rich experience in the past. These experiences include not only the empirical knowledge that can be extracted and sublimated into theory, but also the knowledge that can be deduced from existing theories, and the empirical knowledge that cannot be explained by existing theories and technologies.
Different types of machine tools not only have different evaluation indicators, but also have different influences on the evaluation targets. Therefore, in order to achieve the objective truth of the final assessment result, when the specific solution is made, the expert system firstly obtains the product general plan design information obtained according to the obtained meta-knowledge stored in the knowledge base, and obtains an indicator corresponding to the product in a tree structure. Plan the tree and assign corresponding weights to each indicator. The distribution of weights is accomplished by years of experience accumulated in practice. Users can visualize the listed index items and indicator weights until they are satisfied.
After completing the index planning and the distribution of the weights, the expert system will simulate the engineering experts, according to the criteria and normalization requirements stipulated in the evaluation model, based on the domain knowledge stored in the knowledge base and the overall design information obtained for the above index planning tree. The indicators in the score (that is, the prediction of the quality of the future product performance indicators), the score results are the degree of membership of each index for the assessment of the concentration of factors, will be used as the basis for multi-level fuzzy comprehensive evaluation. In addition, this module also provides visual real-time comprehension of index scores.
Product performance evaluation: Based on the characteristics of machine tool product performance, the system has determined the corresponding multi-level fuzzy comprehensive evaluation model. After obtaining expert scoring results for the indicators, the module completes a multi-level fuzzy comprehensive evaluation of the overall design plan. It not only achieves the evaluation of the evaluation object itself, but also provides a basis for the horizontal comparison between different objects, so as to finally obtain an optimal overall design plan.
4 Establishment of indicator system
Due to the variety of machine tools, complex structures, and varying working conditions, it is difficult to establish a universal index system that is suitable for all machine tools. To this end, the respective index systems under various conditions (machine types, specifications, models, working conditions, etc.) are considered in the meta-knowledge base, and deterministic and fuzzy indicators coexist. In the concrete solution, according to the specific information of the evaluation object, the index system is used to determine the corresponding index system for the solution task.
Through the analysis of the performance indicators of the machine tool products, it is found that some indicators can be divided into several more detailed indicators. These detailed indicators can even be subdivided. In view of this feature, this system describes the performance index of the machine tool product using a hierarchical tree structure. Here, the performance evaluation index of the metal cutting machine tool is taken as an example to explain.
In Fig. 2, each node constitutes its next-level indicator by its sub-nodes. The evaluation of the product begins with the lowest-level leaf node, and goes up one-by-one. Finally, the first-class indicator completes the performance of the machine tool product. Evaluation.
Figure 2 Machine Performance Index Architecture
5 System Assessment Model
The evaluation set determined in this system is V = (v1, v2, v3, v4) = (excellent, good, medium, poor). It is assumed that the index indicator module shown in Figure 3 is obtained from the previous indicator planning module and their respective weights are obtained. The scores of all the leaf nodes are obtained by the expert scoring module (ie: Membership). Here, we call the index of the same parent node in the system as a cluster of indicators. The weight vector of the indicators in each cluster meets the requirement of normalization, and the first-level index can be considered as the evaluation of the overall target (root node) or product performance. Indicator clusters. The evaluation begins with the index cluster at the lowest level, for example, the index cluster consisting of d1, d2, d3, and d4 can determine the fuzzy evaluation matrix as shown in equation (1):
(1)
Figure 3 Hierarchical tree structure of index system
Among them, Ri=(ri1,ri2,ri3,ri4) is a single-factor fuzzy evaluation set corresponding to the i-th index, and rij is corresponding to the i-th index versus the evaluation set Vj(j=1,2,3,4) The degree of membership. Combining the weight vector of the index cluster W = (w1, w2, w3, w4), we can get the fuzzy comprehensive evaluation set Q of the index D, and use the universal matrix multiplication to calculate Q = WR, namely:
(2) In this way, we obtain the degree of membership of the indicator D to the evaluation set based on the comprehensive consideration of the indicators d1, d2, d3, d4, and complete the fuzzy evaluation of the indicator D. Using the same method, starting from the lowest level indicator and stepping up, it is possible to achieve a fuzzy evaluation of all non-leaf node indicators (leaf nodes have been directly given the evaluation results by the expert's score), and the realization of the root node index After the fuzzy evaluation, based on the maximum degree of membership method in fuzzy mathematics, a more comprehensive and reasonable evaluation of the merits of the evaluation target can be made.
In order to facilitate the horizontal comparison between different design schemes, the system also assigns values ​​to each element of the evaluation set: excellent = 1.0, good = 0.8, medium = 0.6, difference = 0.4, ie V = (1.0, 0.8, 0.6, 0.4) ). Then combined with the evaluation object's degree of membership of the evaluation set, ie the value of the fuzzy evaluation set Q (root) of the root node index, the comprehensive evaluation value M, M=QV of the evaluation object is calculated by the weighted average method. In this way, the designer can select an optimal overall design scheme by comparing the M values ​​of different design schemes and combining the values ​​in the Q (root).
6 Conclusion
The performance evaluation system model of the machine tool product for parallel engineering proposed in this paper can realize the performance evaluation of the machine tool product in the early stage of the overall program design, and provide guarantee for obtaining the best overall design scheme. It has the following features:
(1) Through the consideration of the description of the design information, the integration with the overall scheme design system is realized.
(2) has a strong scalability.
(3) Emphasis on the characteristics of the overall design of the design of the experience, the expert system simulation engineering experts, embodies the characteristics of intelligence.
(4) Focus on the combination of empirical knowledge and normative knowledge, with strong practicality.
(5) Conforms to the idea of ​​concurrent engineering as part of a multi-objective collaborative assessment system that is oriented toward the entire life cycle of the product.