Development and Analysis of an Intelligent Electroplating Control System Based on Associative Memory under Conditions of Predictable Uncertainty

Abstract

The analysis of ways to improve the electroplated coating uniformity, among which the most promising for a multiassortment production is the current control, is presented in the article. The electroplating is considered as a control object and its mathematical model is described (with distributed coordinates based on the laws of Faraday and Ohm, Laplace’s partial differential equation) to current control. The search for optimal control is impossible in real time and requires lengthy preliminary calculations due to the complexity of the electroplating mathematical model. The use of associative memory in an intelligent system will make it possible to find a control that corresponds to the electroplating current state and the specified quality criterion (coating uniformity) in real time. An operation algorithm of the knowledge formation unit, which implies the use of an intelligent control system under predictable uncertainty with matrix-type associative memory, is proposed. The associative memory rows are filled based on the uniformity criterion approximation by the function and the choice of its values for given quality losses. An operation algorithm of the controller, which uses the control from the associative memory row with the closest weighted metric to the normalized values of the input variables, is considered. An operation algorithm of the quality assessment unit, which involves determining the influence of input variables on the output variable by calculating their weighting coefficients of significance, to determine the values of which the maximum values, range and maximum local changes are used for the probability of non-negativity of the output variable deviation from its given value, is proposed. A computational experiment was carried out to analyze the effectiveness of the proposed operation algorithms for the units of an intelligent electroplating control system based on associative memory in terms of the control search duration and loss in coating uniformity.