應用類神經網路於隧道式烤爐製程最佳化

機構典藏 > College of Business and Management > Department of Management Sciences > Thesis > Item 987654321/105338

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/105338

Title:	應用類神經網路於隧道式烤爐製程最佳化
Other Titles:	Applying neural network to process optimization of a tunnel oven
Authors:	張晉瑋;Chang, Chin-Wei
Contributors:	淡江大學管理科學學系碩士班時序時;鄭啟斌;Shih, Hsu-Shih;Chen, Chi-Bin
Keywords:	倒傳遞類神經網路;反應曲面法;最佳化;烤漆製程;back-propagation neural network;Response surface method;Optimization;Coating process
Date:	2015
Issue Date:	2016-01-22 14:53:33 (UTC+8)
Abstract:	在自然界中所有結構材料皆會隨著外在環境的影響而產生衰變、劣化等材料性能退化的問題。材料性能退化的問題會影響其構成物的外觀還有強度，例如建築工程上的鋼骨結構，生鏽的時候會造成表面脫落影響美觀，耐久度也會變差。為了避免材料性能退化的現象，我們常對設備與結構物施予保護措施，例如：塗層、防蝕工程、遮雨設計及緩蝕劑等，其中又以烤漆的塗層技術較為成熟且被普遍使用。　　本研究探討隧道式烤漆系統製程最佳化問題，在不同的烤爐溫度、空氣濕度及溶劑比例之下進行試驗，收集產品塗層厚度之資料並且透過倒傳遞類神經網路建立隧道式烤漆系統反應曲面法模型，以了解以上加工條件對塗層厚度的影響。　　然後以數學規劃建置烤漆製程反應曲面法的最佳化模型，並透過數值方法求解最佳之加工條件設定。研究結果顯示，倒傳遞類神經網路所建立的烤漆製程模型具高準確性的預測能力；而反應曲面法最佳化方法則可在不同的空氣濕度條件下，找到適切的烤爐溫度與溶劑比例設定值，以獲得較佳的塗層厚度結果。　　The mechanics of materials can be weaken by the outside environment and then problems of material decay and deterioration occur. Material degradation not only affects its appearance but also its strength. For example, in construction engineering, rust on the steel structure of a building peels the surface layer and reduces the endurance of the steel as well. To avoid fast material degradation, protection such as coating, anti-erosion procedures, eaves designing and inhibitors, are often allied to equipment or construction structures. Among which, coating is particularly popular for being a sophisticated technology, where the coating process is usually done by a paint baking oven. 　　This research aims to find the optimum settings of a tunnel oven to produce desired coating quality. The factors considered in this study include oven temperature, ratio of solution to paint, and the environmental humidity. Experiments are carried out to obtain the resulting thickness of coating under different settings of the aforementioned three factors. The response surface of the coating process by the tunnel oven is modeled by training a back-propagation neural network with the collected data. 　　The optimization of the response surface of the coating process is formulated as a linear programming problem and solved by a numerical method. The result shows that the back-propagation neural network well models the surface response of the coating process, and optimization procedure is able to find reasonable settings of the factors to obtain desired coating quality.
Appears in Collections:	[Department of Management Sciences] Thesis

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	85	View/Open

Loading...