本文主要說明撓性機構伺服系統機械參數之鑑別程序與方法。結合線性動力方程與控制方程,構成狀態空間方程,模擬撓性機構伺服系統頻率響應之Bode圖,以實驗數據為基礎,利用敏感度技巧與遺傳演算法,降低分析模擬與實測之差異,鑑別關節參數。文中以電控式伺服系統為例,負載經電動馬達驅動,歷經齒輪、螺桿、連桿、位置回授感測,其頻率響應之反應值,經以本文發展之方法鑑別5個關節傳動勁度與6個關節阻尼參數,經模擬後,驗證實此種鑑別方法在電控式撓性機構伺服系統應用之有效性。針對具有多數且不同型態關節之撓性機械伺服系統,此方法可提供快速且適合工程應用之鑑別方法。本文針對遺傳演算與敏感度法之鑑別結果比較,對系統反應參考點之敏感度較大之設計變數,兩種演算法計算結果差異較小。本文亦針對遺傳演算在各種條件下之鑑別結果,說明其數據與組件設計之關連性。
A direct procedure is proposed for identifying the joint parameters in a linear servomechanism with multiple joints. The linear dynamical equations of a mechanical system with gains from feedback sensors are derived. The Bode graphs are used to compare the experimental and analytical models to identify the stiffness and damping coefficients. Both of the sensitivity techniques and genetic algorithms are applied to reduce the discrepancies in the frequency responses of the output in analytical and experimental models. The joint parameters are then identified. An industrial servomechanism with multiple joints, in which a payload is driven by a mechanism, which consists of a motor, a ball screw, gears, linkages, and a feedback sensor, is used to demonstrate the effectiveness of the proposed procedure. The values of five stiffness and six damping parameters for the joints of this servomechanism were identified successfully by this procedure. For a complicated servomechanism, with multiple and various joints, the identification methods which can be easily applied to engineering application are developed. The identified joint parameters were also verified in a different operating condition of the servomechanism. The identified results shows that the variables with larger sensitivity will keep the narrower identified ranges and stringent design tolerances.