Open Knowledge Representation Techniques and Its Application in Precision Forming Design

Precision forming knowledge of precision forging parts can be acquired from various sources and generally requires an effective representation for its efficient reuse. According to the specific characteristics of the precision forming of car forging parts, the precision forming engineering knowledge has been classified into several kinds with the method of the open knowledge representation techniques for the different kinds of knowledge proposed. The open representation techniques of precision forming knowledge and its application are described in detail, and combined with the development of the precision forming design system of the automotive starter guiding cylinder based on knowledge. The results show that the precision forming knowledge is represented in an effective way will solve the problem of fusion of precision forming knowledge and precision forming manufacturing, and improve greatly the utilization efficiency of the precision forming knowledge.


Introduction
Knowledge-based Engineering (KBE) has been used widely, which integrates artificial intelligence with CAX system and connects engineering design without interruption [1].KBE application has become pivotal for enhancing manufacturing enterprises competitiveness.Product design is a highly involved, often ill-defined, complex and iterative process [2].An effective Knowledge driven design tool requires efficient knowledge representation schemes.Therefore, many scholars have carried out the research of knowledge representation.Lee and Jeong [3] developed a distributed and flexible knowledge representation model, and tested a prototype.Edy Portmann, Patrick Kaltenrieder and Witold Pedrycz [4] put forward knowledge representation through graphs and provided an overview of different types of graphs stressing their relationships and their essential features.Haining Yao and Letha Etzkorn [5] introduced a semantic converter which converts knowledge representation in Conceptual Graphs into representation in Resource Description Framework.W.M. Wang, X. Peng [6] presented a dynamic representation of fuzzy knowledge (DRFK) method.Chen, W.L [7] put forward a knowledge representation approach based on Parameter Flow Chart.Xi P [8] studied the open representation of design knowledge in the form of formaluas, 2D mapping graphs, procedures and rules.
The problem with these researchers is that few papers have addressed the open knowledge representation in intelligent precision forming design.In this paper, we provided an open knowledge representation and its application in precision forming design system.This paper is organized as follows: Section 2 describes the precision forming design process based on knowledge.Section 3 puts forward the open representation of precision forming knowledge.Section 4 presents an application implementation of the knowledge driven precision forming design combined with the automotive starter guiding cylinder.The conclusion and future work are discussed in Section 5.

The precision forming design process based on knowledge
Based on the characteristics of precision forming design, the precision forming design process based on knowledge is shown in Fig. 1.Precision forming design, especially the precision forming design of complex product is a highly involved, often ill-defined, complex process in which design engineers perform a large number of activities related to creating and evaluating solutions.Most of these activities depend upon the design knowledge obtained in previous projects.In order to improve the efficiency and quality of the precision forming design of car forging parts, precision forming knowledge should be properly fused with forming design process.

The open representation of precision forming knowledge
According to characteristics of the precision forming of car cold/warm forging parts, the precision forming knowledge is usually classified into four categories, namely, process design knowledge, die design knowledge, numerical simulation knowledge and forming processing knowledge.As shown in Fig. 2.

The open representation of formulas knowledge
There are many formula requires an effective representation for its efficient reuse in the process of precision forming design.The BNF (Backus Naur Form) representation is usually used to represent the knowledge.

The open representation of process knowledge based on fuzzy logic
The fuzzy sets and membership function in fuzzy mathematics can be used to represent the uncertainty process knowledge in the precision forming design development.
If U domain is numerical domain, and the process knowledge can be described by numerical values, the membership function of U domain fuzzy subset can be defined as a linear function with parameters, namely: Where, a, b, c and d denote respectively design parameters, and a value is less than b value, b value is less than or equals to c value, c value is less than d value.

The open representation of case knowledge
In the process of precision forming design, there are much forming design support knowledge, such as part name, part material, part shape and similar parts forming design etc., the open representation of case knowledge is as follows:

The open representation of object-oriented precision forming knowledge
The method of object-oriented knowledge representation provides the approach of integrating forming process knowledge into forming design.These methods include abstraction, encapsulation, inheritance and polymorphism.The open representation of object-oriented precision forming process knowledge is as follows: Object::=(object name, data structure, the method sets, message interface)

The open representation of precision forming knowledge based on artificial neural network
For the open knowledge representation of multiple inputs and multiple outputs precision forming process knowledge, the BP neural network is usually adopted, as shown in Fig. 3.

Application case
Automotive starter guiding cylinder (as shown in Fig. 4) is the critical part of automotive starter transmission system, it works in a complex condition, and it has to meet the requirements for the excellent mechanical properties.At present, most automotive starter guiding cylinders are manufactured by precision forming, in which is characterized by low cost, high productivity, high precision and excellent mechanical properties.As there is so much design knowledge involved in the forming process, therefore, it can greatly improve the efficiency of the precision forming technology and die design to incorporate design knowledge and forming process, and realize the reuse and sharing of knowledge.

The open representation of precision forming design formula
The complicated shape coefficient of extrusion F k is often checked in the precision forming design, and the formula of the complicated shape coefficient of extrusion is as follows: Where, V0 is the preformed blank volume, V1 is the volume of final forming part.

The open representation of graphs in the precision forming design
Engineering graphs are common forms of precision forming design knowledge.Fig. 6 is the Z-versus-stroke graph for the automotive starter guiding cylinder precision forming, where Z is the forming load, stroke is the extrusion displacement.The BNF representation of graphs knowledge is described as following: <graph knowledge>::=<query statement ><graph data > <query statement >::=<graph X>|<graph Y> <graph data >::=<Z_load_curve> <original data><0.0><0.0>

The open representation of table knowledge
in the precision forming design Table 1 is the parameters of the final forged part and the billet for the automotive starter guiding cylinder precision forming.The open representation of table 1 knowledge is as follows: The_forming_parameters=Table ("The parameters of the final formed part and the billet", the forming parameters, the value of the parameter)

Conclusions
In this study, an open representation of precision forming design knowledge is proposed, and it can help capture design knowledge from various sources in the precision forming design process, and can effectively realize the fusion of precision forming design knowledge and precision forming process.The methods proposed are shown to be useful for precision forming knowledge reuse.In our future work, we will study how to enable designers to formulate complex knowledge queries so that relevant precision forming design knowledge can be recommended automatically.

Figure 1 .
Figure 1.Precision forming design process based on knowledge.

Figure 2 .
Figure 2. The classification of precision forming knowledge.

Figure 5 .
Figure 5.The interface of precision forming design based on KBE
The case knowledge of automotive starter guiding cylinder mainly includes parts name, parts material, the extrusion speed of the punch, forming temperature, friction coefficient and the stress and strain distribution in the forming process.The open representation of case knowledge is as follows: CASE INSTANCE: Case_No.ASGC Name_of_part: ASGC Material _of _Part: 20CrMo MaterialType_of_part: Metal Shapetype_of_part: Axially Symmetrical Parts Punch_speed: the stress distribution of forming process in the forming case base.

Figure 7 .
Figure 7.The stress distribution of forming process.

The open representation of table knowledge
Engineering table knowledge is usually stored in relational database, and the designers can easily query some parameter values with the help of query statements.

Table 1 .
The parameters of the final formed part and the billet.