Model Development of Servo Drive System in Machine Tool

為了能有效的開發工具機，近年來許多國內外學者皆致力於發展虛擬工具機技術，目的便是為了減少工具機開發的成本與時間。虛擬工具機技術的基礎為建立準確的模型，欲準確的預測工具機的精度，須建立精確的伺服進給系統模型。本研究重點在於伺服控制系統之模型建立，首先建立連續模型以及考慮到迴路處理週期以及感測器取樣頻率的離散模型。根據頻域響應量測結果，利用粒子群最佳化(PSO)鑑別出模型中未知的參數。再來建立前饋模型，並根據馬達在時域之測試結果解析出前饋模型中的參數。模型驗證可分為頻域與時域部分。在頻域部分，速度迴路模型在預測頻寬時與實驗值相差在10%以內，位置迴路模型所預測之頻寬則沒有太大的誤差。時域部分，模型在未使用前饋時，預測的追隨誤差與實驗值相差在2%以內，加入前饋模型後，預測之追隨誤差與實驗值相差最大為13.7%。比對模型差異性，使用連續模型鑑別，速度控制器積分項(K_vi)的值將會偏大。將電流迴路簡化為1之建模方式僅適用於速度迴路頻寬與電流迴路頻寬匹配的情況下。建立完成的伺服控制模型，加上雙質量彈簧阻尼模型，亦可應用於負載慣量與聯軸器剛性之估測，估測之負載慣量誤差於1.5%以下，聯軸器剛性之估測誤差為10.8%。In order to design the machine tool effectively, many domestic and foreign scholars effort to the development of virtual machine technology in recent years, the purpose is to reduce the cost of machine tool design and time. The foundation of the virtual machine tool technology is to establish the accurate model, in order to accurately predict the accuracy of the machine tool, it is nec-essary to establish an accurate servo feed drive system model. This study fo-cuses on the establishment of a servo control system model, first of all, estab-lish a continuous model, and then establish a discrete model which consider the loop processing period and the sampling frequency of the sensor. According to the results of frequency response measurement, the parameters in the model are identified by particle swarm optimization(PSO). Then the feedforward model is established, and the parameters in the model are identified according to the test results of the servo motor in the time domain. Model validation can be divided into frequency and time domain parts. In the frequency domain, the velocity loop model is within 10% of the predicted bandwidth difference from the ex-perimental value, the position loop model predicted the bandwidth is accurate. In the time domain part, when the feedforward model is not used, the difference between the predicted position error and the experimental value is less than 2%. After the feedforward model is added, the maximum difference between the predicted position error and experimental values is 13.7%. Comparison of mod-el differences, the value of the integral term of speed controller will slightly larger when identify by continuous model. The simplification of the current loop to one is only applicable when the speed loop bandwidth matches the cur-rent loop bandwidth. The servo control model is established, coupled with dual mass spring damping model, also can apply to estimate the load inertia and coupling stiffness, the estimated load inertia error is less than 1.5%, and the es-timated error of the coupling stiffness is 10.8%.摘要 i Abstract ii 圖目錄 vii 表目錄 xi 符號表 xii 第一章 緒論 1 1.1前言 1 1.2文獻回顧 2 1.3研究目的 6 1.4論文架構 7 第二章 伺服系統模型推導 8 2.1伺服控制系統 8 2.2伺服系統模型建立 10 2.2.1電流迴路模型建立 10 2.2.2連續模型建立 13 2.2.3離散模型建立 15 2.3前饋控制模型建立 20 2.4系統鑑別理論 22 2.4.1粒子群最佳化原理 22 2.4.2粒子群最佳化更新公式 22 2.4.3系統鑑別應用 23 2.5雙質量彈簧阻尼模型 26 第三章 實驗規劃 28 3.1實驗架設 28 3.1.1伺服系統鑑別架設 29 3.1.2負載慣量與聯軸器剛性測試架設 30 3.2實驗規劃與流程 32 3.2.1伺服系統鑑別 32 3.2.2 負載慣量與聯軸器剛性 35 第四章 實驗結果與分析 37 4.1伺服迴路鑑別 37 4.1.1調機軟體掃頻機制 38 4.1.2速度迴路頻域響應量測 41 4.1.3速度迴路系統鑑別 43 4.1.4位置迴路系統鑑別 48 4.1.5位置迴路時域測試 51 4.1.6馬達黏滯摩擦係數鑑別 53 4.2前饋控制測試 54 4.2.1前饋控制輸出百分比測試 54 4.2.2前饋時機測試 56 4.2.3前饋模型參數鑑別 58 4.3負載慣量與聯軸器剛性測試 61 第五章 模擬結果與驗證分析 64 5.1伺服迴路模型驗證 64 5.1.1速度迴路頻域響應驗證 64 5.1.2位置迴路頻域響應驗證 67 5.1.3位置迴路時域驗證 68 5.1.4前饋模型驗證 70 5.2模型差異性比較 72 5.2.1電流迴路模型差異分析 72 5.2.2前饋模型之差異分析 75 5.3伺服迴路模型應用 76 5.3.1虛擬伺服迴路增益優化 76 5.3.2負載慣量與聯軸器剛性估測 78 第六章 結論與未來展望 82 6.1結論 82 6.2未來展望 83 參考文獻 8