應用智慧模糊控制模型預測電漿噴塗製程之研究

A Study of the Intelligent Fuzzy Control Approach for Prediction of the Plasma Spraying Processes

本論文應用部份因子設計與變異數分析實驗於電漿噴塗製程，主要運用較少實驗次數於眾多因子中挑選對製程有顯著性因子。依據模糊控制理論設計模糊規則庫及推論引擎之智慧模型模糊控制器。根據實驗設計法設計模糊控制器將開發出穩健電漿噴塗製程，將可進一步地探討有關操作條件對噴塗層厚度之變化，從而發展出一個穩健之噴塗製程。本實驗之結果顯示，應用統計設計策略於實驗設計中，對噴塗厚度品質特性變異有較佳效果，經模糊控制器預測出誤差百分率為6.53%。這說明提出模糊控制器預測系統可反應實際電漿噴塗製程之特性狀況。

The purpose of this paper is to describe the use of fractional factorial experiments and analysis of variance along with a robust design for a plasma spraying process. Another approach utilized is fuzzy logic theory, which suggests a design including an intelligent fuzzy logic controller with a rule base and an inference engine. The results suggest that a fuzzy logic controller for statistical measures is developed for use in a robust plasma spraying process. Specifically, we intend to develop a relationship between control factors and spraying reinforcement in spraying processes as well as produce desired characteristics in a treated layer.In this study, a fuzzy logic controller design is built and utilized. Using a variety of experiments, the variability in the quality characteristic is reduced. The error of percentage, calculated through a fuzzy logic controller model and inferred rule, is 6.53%. It is clear that a fuzzy logic controller demonstrates better accuracy than other methods when predicting spraying thickness values in plasma spraying process.