Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/2295
DC FieldValueLanguage
dc.contributor.advisor王國禎zh_TW
dc.contributor.advisorGou-Jen Wangen_US
dc.contributor.author游政雄zh_TW
dc.contributor.authorHSUNG, YU.CHENGen_US
dc.date2001zh_TW
dc.date.accessioned2014-06-05T11:42:57Z-
dc.date.available2014-06-05T11:42:57Z-
dc.identifier.urihttp://hdl.handle.net/11455/2295-
dc.description.abstractCMP製程有著不穩定性與製程性能偏移之特性,為能有效控制CMP製程,本研究將以發展出之類神經網路式Run-to-Run學習控制系統應用於CMP製程控制上。 本研究提出結合輻射基底函數式類神經模糊網路與Run-to-Run製程控制技術之新的Run-to-Run製程控制系統架構。相較於傳統之EWMA控制器,此一新架構除控制器增益值將由類神經模糊網路提供外,一階線性預測模式亦將由非線性之類神經模糊網路取代,而非線性之類神經模糊網路也更有助於製程參數靈敏度分析,有效提供預測製程狀態之機制。 電腦模擬與實際實驗結果皆證明本研究所發展之新控制架構,較傳統之EWMA法則有更佳之誤差偏移抑制效果。而本論文所提出之新架構,亦無需更改機械結構與新增額外之感測元件,在產業應用面上有高度可行性與經濟效益,對半導體產業之效益提昇應有所貢獻。zh_TW
dc.description.abstractThe goal of this thesis is to propose an intelligent process control strategy that combines the RBF based neural fuzzy network and the run-to-run process control techniques for the erratic and unstable CMP processes. In this new run-to-run control scheme, two RBF based neural fuzzy networks are employed to replace the 1st order linear process predicting model and the linear controller of the conventional EWMA scheme, respectively. The learning and nonlinear mapping essences of the neural network provide this new control structure with more power in nonlinear process modeling and control. Computer simulations and experimental results demonstrate that the proposed new control scheme can suppress the drifting problem better than the conventional EWMA model. In addition, no extra sensor and machine modification is required to employ the new control method. Consequently, it is highly feasible for industrial implementation.en_US
dc.description.tableofcontents第一章 緒論………………………………………1 1.1 研究動機與目標……………………………1 1.2 論文大綱……………………………………5 第二章 化學機械研磨之Run-to-Run控制器……6 2.1 化學機械研磨………………………………6 2.2 Run-to-Run控制器…………………………10 2.3 化學機械研磨之Run-to-Run製程控制……13 第三章 以類神經網路為基礎之Run-to-Run學習控制系統(NNRTRC)………………………………………………………15 3.1 前言………………………………………………15 3.2 控制系統架構……………………………………16 3.3 類神經網路式Run-to-Run控制器學習法則……19 3.4 控制器訓練………………………………………27 3.5 CMP製程電腦模擬控制…………………………29 第四章 實驗驗證………………………………………43 4.1 參數選取…………………………………………43 4.2 實驗規劃與進行…………………………………44 4.3 實驗結果…………………………………………45 第五章 結論與未來展望………………………………51 5.1 結論………………………………………………51 5.2 未來展望…………………………………………53 參考文獻………………………………………………54zh_TW
dc.language.isoen_USzh_TW
dc.publisher機械工程學系zh_TW
dc.subjectCMPen_US
dc.subject化學機械研磨zh_TW
dc.subjectRun-to-Run process controlen_US
dc.subjectneural networken_US
dc.subjectRun-to-Run製程控制zh_TW
dc.subject類神經網路zh_TW
dc.title化學機械研磨之類神經網路式Run-to-Run製程控制zh_TW
dc.titleNeural Network Based Run-to-Run CMP Process Controlen_US
dc.typeThesis and Dissertationzh_TW
item.openairecristypehttp://purl.org/coar/resource_type/c_18cf-
item.openairetypeThesis and Dissertation-
item.cerifentitytypePublications-
item.fulltextno fulltext-
item.languageiso639-1en_US-
item.grantfulltextnone-
Appears in Collections:機械工程學系所
Show simple item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.