An investigation of an expert system for hydraulic circuit design with learning

Abstract

In this paper the major research objective is to study the possibility of using an expert system in the hydraulic circuit of a machine tool design (such as a drilling machine, milling machine, grinding machine, etc.) and also to develop an expert system of hydraulic circuit design on a PC/AT personal computer. In the designing of the hydraulic circuit, the expert system structure performs the inference and analysis of the qualitative data, and this can bring about efficiency, while introducing conventional programs (such as FORTRAN) to perform the calculation of the quantitative data. Finally, the Auto CAD drafting software is introduced to display the inferred results of the hydraulic circuit design in design drafts on the computer screen. In addition, a primary learning mechanism for hydraulic circuit design is developed. If the user is not satisfied with the inferred results or the system fails to infer any results, the system has designed a model for data acquisition which automatically invites the user to input a set of data but at this time the user must be an expert, and continues to reorganize and arrange the data, and even delete unnecessary or contradictory data. It then translates the data automatically into rules acceptable to the expert system for further inference. The learning does not come to an end until the user is satisfied.