完整後設資料紀錄
DC 欄位語言
dc.contributor.author涂宗廷en_US
dc.contributor.authorTsung-Ting Tuen_US
dc.contributor.author洪士林en_US
dc.contributor.authorShih-Lin Hungen_US
dc.date.accessioned2014-12-12T02:24:29Z-
dc.date.available2014-12-12T02:24:29Z-
dc.date.issued2000en_US
dc.identifier.urihttp://140.113.39.130/cdrfb3/record/nctu/#NT890015058en_US
dc.identifier.urihttp://hdl.handle.net/11536/66441-
dc.description.abstract由於類神經網路具有可訓練性及容錯性,其已成功地被應用於不同領域,包含自動控制、最佳化問題、語言及影像之判識、氣象預測、結構反應模擬。從文獻中可發現類神經網路應用於結構系統之模態識別則不多見。 本研究首度提出類神經網路式之系統識別模式(System Identification using Artificial Neural Network)。其原理即擬利用類神經網路分析結構物地震反應量測數據,藉由訓練所得之權重矩陣直接估算結構系統之動態特性:自然振動頻率、阻尼比及模態。另外,研究中亦對類神經網路於結構物反應作預測及識別上之迷思作一闡述,並提出一網路回饋預測模式(Artificail Neural Network Feedback Modal)與傳統網路直接預測模式(One-Step-Ahead Prediction)作一比較。 研究中將利用數值模擬方式,對本研究模式之可行性作初步之驗證。文中亦對於敏感度分析之加入(剔除不必要之輸入變數,以減少網路之訓練時間)、訓練資料多寡、時間延滯、量測自由度減少之影響,作一詳細之探討。最後,本研究中所提之兩模式將實際應用於一實測結構物地震反應。zh_TW
dc.description.abstractDue to the capacity of training and the high tolerance to partially inaccurate data, neural networks have been successfully applied to various fields such as automatic control, optimization, speech and image recognition, weather prediction, and prediction of structural responses. Nevertheless, it is hardly found the applications of neural networks to determine the dynamic characteristics of structures in the published work. The System Identification Using Artificial Neural Networks(SIANN)Model will be firstly proposed in this study. The basic concept of this model is to extend the application of neural networks to identify the dynamic characteristics of building structures from their earthquake responses. The natural frequencies, modal damping, and mode shapes can be directly determined from the weighting matrices of a neural networks. Furthermore, for the sake of illustrating the myth of predicting and identifying using artificial neural networks, the Artificial Neural Networks Feedback Model(ANNFM)will be proffered. And this model will be compared with the traditional prediction model such as One-Step-Ahead Prediction Model. For verify the preliminary feasibility of this study, the numerical simulation will be preceded. In additions, the influence of utilizing sensitivity analysis, amount of training data, steps of time delay, and the lack of measurement in DOF will be discussed in detail. Eventually, the two model proposed in this study will be applied to a real building structure.en_US
dc.language.isozh_TWen_US
dc.subject類神經網路zh_TW
dc.subject系統識別zh_TW
dc.subject預測模式zh_TW
dc.subject敏感度分析zh_TW
dc.subject時間延滯zh_TW
dc.subject實測結構物地震反應zh_TW
dc.subjectArtificial Neural Networksen_US
dc.subjectSystem Identificationen_US
dc.subjectPrediction Modelen_US
dc.subjectSensitivity Analysisen_US
dc.subjectTime Delayen_US
dc.subjectReal Building Structureen_US
dc.title類神經網路於房屋結構系統識別之應用zh_TW
dc.titleApplication of Artificial Neural Networks for System Identification in Building Structuresen_US
dc.typeThesisen_US
dc.contributor.department土木工程學系zh_TW
顯示於類別:畢業論文