運用資料探勘方法建立變壓器一、二次側間電容值預測模型;Using data mining to establish the transformer capacitance value between the primary and secondary side prediction model

NCUIR > School of Management at National Central University > Executive Master of Industrial Management > Electronic Thesis & Dissertation > Item 987654321/67446

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/67446

Title:	運用資料探勘方法建立變壓器一、二次側間電容值預測模型;Using data mining to establish the transformer capacitance value between the primary and secondary side prediction model
Authors:	顏錞靖;Yen,Chung-Jing
Contributors:	工業管理研究所在職專班
Keywords:	資料探勘;迴歸;決策樹;變壓器;電容
Date:	2015-05-18
Issue Date:	2015-07-30 17:51:41 (UTC+8)
Publisher:	國立中央大學
Abstract:	電源供應器為電子電器產品不可或缺的零件，電源供應器中的變壓器是主要供給電子產品轉換電壓的元件，也是影響電源供應器電磁干擾強度的主要因子。而電子電器產品上市販售前，都必通過電磁干擾的測試，所以如何測試變壓器的電磁干擾強度，是變壓器重要的課題，為此我們發展出測試變壓器電磁干擾度的方式TCPS(Turns Capacitance between Primary and Secondary)。由於變壓器結構複雜且存在著許多的不可管控的因素，所以變壓器結構的改變對於TCPS數值的預估，一直停留在嘗試錯誤的階段，工程師浪費了許多資源與時間重複試做變壓器樣品來確認TCPS數值上，沒有一個預估的方法與模型，只能依照先前的經驗以及電學理論來推估方向，至於調整變壓器結構對TCPS數值的增減，完全沒有把握，只能從樣品試做後確認。本研究使用資料探勘的手法，設定不同影響因子，以收集大量TCPS數據，運用統計分析來建立TCPS數值預估模型，導入設計驗證流程中，以減少嘗試錯誤的次數，節省開發時間與資源，使得節省的資源可以用在更有效率的地方，從節省的開發時間搶得市場佔有率，提高公司的獲利。 ;The power supplier is an essential part of electrical and electronic products. The transformer inside the power supplier is a main component for changing voltage, which also has key impacts on the power supplier’s electromagnetic interference intensity. Before the electrical and electronic products making market debut, all products must pass the electromagnetic interference testing. As the result, how to test transformer’s electromagnetic interference intensity is an important issue. Therefore, developing the testing method for transformer’s electromagnetic interference intensity: Turns Capacitance between Primary and Secondary (TCPS). Due to the structure of transformer is complicated and many uncontrollable factors, trial & error is the only way to predict TCPS. The engineers waste a lot of resources and times to test the TCPS value. This research use data mining to consider different factors in predicting TCPS value. Using statistics analysis to establish TCPS value prediction model in the process of design verification test reduces the times of trial & error. Therefore, the engineers could be more efficient in research and development, to increase the company’s profit.
Appears in Collections:	[Executive Master of Industrial Management] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	840	View/Open

社群 sharing

Loading...