Modeling and Optimization of Reflow Thermal Profiling Operation: A Comparative Study

迴焊溫度曲線建模與最佳化

蔡聰男(Tsung-Nan Tsai)

《工業工程學刊》 26卷6期 (2009/11) Pp. 480-492

https://doi.org/10.29977/JCIIE.200911.0006

摘要

本研究運用複合人工智慧技術與需求函數(desirability function)方法最佳化迴焊溫度參數(reflow thermal profiling)並比較兩者最佳化績效之良窳。迴焊溫度曲線爲一組或多組溫度－時間曲線，其應用於監控迴焊溫度輸出項與基板(printed circuited board, PCB)/電子零件處於迴焊爐內之熱效應。提升迴焊製程品質成爲改善電子組裝良率重要決定因素之一，不同產品組態之迴焊溫度需求迥異，若採用不適當迴焊溫度設定進行迴焊作業，將產出諸多焊接缺陷(soldering defect)，甚至導致可觀重工與報廢成本。本研究採用L18(2^1×3^7)田口實驗矩陣用以收集迴焊製程資訊，最佳化方法一係運用類神經網路訓練實驗數據用以探查迴焊製程之非線性關係，再運用基因演算法自已訓練類經網路中找尋最佳迴焊溫度參數。再者，方法二乃運用AHP分析各個迴焊輸出項之權重並將多重品質特性結合爲單一需求指標，再採需求函數方法最佳化之。利用DPMO，良率、及製程標準差作爲績效指標，經實務績效驗證後，顯示方法二提供較佳最佳化績效，但兩者方法皆可大幅改善現階段迴焊製程品質。

關鍵字

表面黏著技術 ； 迴焊製程 ； 基因演算法 ； 類神經網路 ； 製程最佳化

並列摘要

In this study, a comparative study of optimizing the reflow thermal profiling parameters using a hybrid artificial intelligence and the desirability function approaches without/with combining multiple performance characteristics into a single desirability is presented. Reflow soldering is the key determinant for the improvement of the first-pass yields of electronics assembly. A reflow thermal profile is a time-temperature contour with multiple performance characteristics utilized to monitor the heating effects on a printed circuit board (PCB) and surface mount components (SMCs) in the reflow oven. The use of an inadequate reflow thermal profile may not only produce a variety of soldering failures, but can also result in the needs for considerable reworking and waste. An L18 (2^1×3^7) Taguchi experiment design is conducted to collect the thermal profiling data. A quick propagation (QP) neural network is modeled based on experimental data to formulate the nonlinear relationship between the thermal profiling factors and responses, and a genetic algorithm (GA) is used in the optimization of thermal profiling factors with the fitness function based on the trained QP neural network model. Alternatively, the response columns for the experimental data can be transformed into a single measure of desirability which is then optimized by the desirability function approach with the response weightings derived from an analytic hierarchy process (AHP). The empirical evaluation results show that the desirability function approach with combining the multiple performance into a single desirability delivery superior soldering performance to that obtained by the hybrid artificial intelligence method without combining the multiple performance into a single desirability, as measured by the DPMO, yield rate, and process sigma.

並列關鍵字

surface mount assembly ； reflow soldering ； genetic algorithm ； neural network ； process optimization

參考文獻

Chang, H. H.(2006).Dynamic multi-response experiments by backpropagation networks and desirability functions.Journal of the Chinese Institute of Industrial Engineers.23,280-288.

Abouhamze, M.,M. Shakeri(2007).Multi-objective stacking sequence optimization of laminated cylindrical panels using a genetic algorithm and neural networks.Composite Structures.81,253-263.

Apell, M. C.(2004).Considerations for lead-free reflow soldering.Circuits Assembly Magazine.28-30.

Ayag, Z.(2005).A fuzzy AHP-based simulation approach to concept evaluation in a NPD euvironment.IIE Transactions.37,827-842.

Barbini, D.,G. Diepstraten,U. Marquez(2005).Process considerations for optimizing a reflow profile.Surface Mount Technology Magazine.28-31.