SELF PROPELLING ARTIFICIAL INTELLIGENCE QUERY SYSTEM FOR WELD OBJECT DATABASE

摘要

The method to perceive two-dimensional objects such as a figure of weld joint have a commanding influence on the systematic way of identifying or recognizing targets being referred. The objects in two dimension are normally defined by a series of coordinate sets, however of which x and y values can not be placed for a direct comparison of matching. Therefore second derivatives of coordinate sets are required for further sophistication of identifying. In this sense, it is very crucial whether or not the inference engine for query system can directly assess the derivatives of the figure definitions in the course of inference for the network of meshed object like FEM model of a weld joint. Among the characteristic values of a two-dimensional object, we focused our attention only on the outer boundary or circumference, i.e. the total length of the periphery and the area enclosed by a curve were found to be significant. In the consecutive process of identifying each node and element, recurrent operations are prevailing and apparent in the inference. As of the recursion of inference procession, it have to be definitely iso-parametric, where a single form of operation is preferred to reside. In the progression of the inference, our rules in engine has successfully attained self propelling procession in the rule operation within a very short execution time.