Review of Methods for Synthesis of the Model of Class of Spiroid Gearboxes for Intelligent CAD Systems. Part 2. Development of Methodology for Synthesis of the Generalized Model

Abstract

The necessary condition for development of the intelligent system of computer-aided designing activity allowing not only to implement graphic works, individual engineering analysis or documenting, is the development of the generalized model of the class of designed objects keeping the designer’s experience as individual technical solutions of layouts for the products, their units, assemblies, sub-assemblies and parts. The more complex the object is, the greater number of possible modifications of its layout can be implemented. Application of the algorithm of combinatorial search as the basic algorithm of functioning of the CAD system requires generation of the initial set of design elements and features of their application to implement the search. Generation of the pointed set can be implemented by various methods differing by the level of demonstration visibility, protection from errors, and redundancy. The graph method described in the first part of the paper is certainly the basic one, for which the algorithm of developing the classifier of the class of spiroid gearboxes was initially created. Complexity of formalization of this method stimulated the author to modify it. Therefore, the table, graph-and-table, matrix methods of developing the model of the class of spiroid gearboxes were created. The feature of the first method is that the hierarchic structure of the layout of the object is represented as a two-dimensional table according to which the generalization of information becomes more correct, since singling of non-obligatory components of the layout becomes more evident. The graph-and-table method is a symbiosis of the graph and table methods; it keeps the demonstrative feature of structures and provides for the correct description of functional elements. Due to combination of its demonstrative features and simplicity of implementation, the graph and table method eliminates the loss of information to the most. The matrix method is especially urgent when the structure of initial layouts is more complex, since this method allows for correct processing the non-obligatory apexes of group belonging and phantoms.