Using the Taguchi method for optimization of the powder metallurgy forming process for Industry 3.5. (October 2020)

• Record Type: Journal Article
• Authors: Wang, Hung-Kai; Wang, Zih-Huei; Wang, Ming-Chi
• Is Part Of: Computers & industrial engineering. Volume 148(2020)
• Publication Date: 2020-10
• Subjects: Industry 3.5 -- Powder metallurgy -- Parallelism -- Taguchi method -- Analysis of variance -- Parameter optimization; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2020.106635
• Languages: English
• ISSNs: 0360-8352

Abstract

Graphical abstract: This study uses the Taguchi method to analyze various factors that affect parallelism in the forming stage of the powder metallurgy process and to determine the optimal parameter setting and verification. Highlights: This study uses the Taguchi method to determine the optimal setting for powder metallurgy process. Two experiments are validated that the performance of products increased 55.73% and 49.44%. Values of the S/N ratio in confidence interval to verify whether the experiment is reproducible. The product yield and production cost were simultaneously improved for this case company. Abstract: In recent years, powder metallurgy has developed rapidly and has been used to produce mechanical parts with complex structures. High precision and material utilization with low production cost can be achieved through the powder metallurgy process. In order to reduce the cost of finishing processes and product scrap rate, quality management methods must be used to improve processes and increase the product yield. This study uses the Taguchi method to analyze various factors that affect parallelism in the forming stage of the powder metallurgy process and to determine the optimal parameter setting and verification. In the orthogonal array of the Taguchi method, we selected six factors with two levels as the machine setting, which would affect the product's parallelism. The current machine setting, and possible setting were determined as the first and second levels, respectively, and several experiments were designed to ascertain the configuration of the orthogonal array. After performing the designed experiments, the signal noise ratio (S/N ratio) was calculated for estimating variability of products. New samples were manufactured based on the optimal setting and compared with the original data. The experimental result revealed that the average value of parallelism data was considerably reduced for two different products and decreased approximately 55.73% and 49.44%. Values of the S/N ratio should be in the confidence interval to verify whether the whole experiment is reproducible. Thus, the product yield and production cost were simultaneously improved for this case.