摘要 電動自行車因具有環保、安全、省力等諸多優點,是政府一直致力推廣的綠色產品。然而在產品造形設計過程中,評估設計師創造能力的條件主要有二,一是“造形構想展開數量”的表現,二是“造形多變性”的表現。“造形構想展開數量”是創造力表現的基本要件,想要得到具有創造性的構想,首先必須要有數量眾多的想法,然後才能從這些想法當中選擇出最具有創造性的構想,但是設計師的思考仍有可能遺漏,因此本研究將利用電腦輔助計算進行造形結構的發展,彌補設計思考的遺漏。本研究針對電動自行車之車架進行造形的展開,以結構最佳化設計進行演算,接著利用集群分析,篩選出代表車架驗證其修飾過結構強度,然後將影響電動自行車的主要形態變數車架形式、電池尺寸、電池位置與菜籃訂定後,進行語意差異法認知調查,以主成分分析進行影響認知之形態分析,最後以改良式倒傳遞類神經網路法進行認知預測,並針對各族群之需求性,分析兩族群最佳之車形及其差異;分析的結果也顯示兩者的確有些車架造形是相同的,但有些車架造形完全各異的,由此驗證得知起始形雖然不同,但是應該是有些車架造形是相同的,並且演化全新造形的車架。結果顯示兩族群對於幾部經由拓樸最架化設計演算的車形感受度很高,而且經由數量化I類研究出電動自行車零件各部分與意象關係幫助設計師進行造形的展開,然而經由倒傳遞類神經學習意象與此產品造形結果誤差值都合理範圍;對於到傳遞類神經所建立意象資料庫可以幫助設計師所提供意象詞對,進行搜尋符合詞對意象的車架造形。
Abstract With many advantages of Electric bicycles, they have been still not common and popular. The main reason is that their exterior design has not revealed its essence. Moreover, the entire bicycles give a sense of instability when the frame is fitted with the transmission components such as a battery and motor, and this make a consumer not buy it. Therefore, it is essential to design a premium Electric bicycle form that can meet the need of consumers. This study will regard the results of measurement(force, pressure), of riding posture measurement, and of positioning of components such as battery, motor, etc as the boundary and limit conditions and will use the topology optimization design, including the optimization algorithm, structural analysis tool, and the topology optimization method to derive the Electric bicycle’s frame forms. Using cluster analysis on those collected city entertainment bicycle frames and derived frames, we will also prove that the derived frames are innovative designs. The results of the electric bicycle shape combination that can affect each kind of awareness of the two groups, obtained from “Quantification I” can be combined with the awareness prediction database to be used as reference data and supplementary tools for those designers of electric bicycles. The prediction database learned from the BP neutral network can help the designers to obtain the electric bicycle frame form that meets the designers’ image through an image term and through the database system. This system can help the designers to get a searching bridge between the image and the bicycle frame form.