本文選題：異步電機　切入點：最小二乘法　出處：《浙江大學(xué)》2015年碩士論文　論文類型：學(xué)位論文

【摘要】：近年來,隨著電力電子技術(shù)、計算機控制技術(shù)的發(fā)展,交流伺服控制在伺服控制領(lǐng)域逐漸占據(jù)了主導(dǎo)地位,異步電機也隨之得到了廣泛的應(yīng)用。因此,針對異步電機高性能控制的研究具有重要的意義。在很多異步電機應(yīng)用場合中,負(fù)載的轉(zhuǎn)動慣量是電機轉(zhuǎn)子的數(shù)倍甚至數(shù)十倍,其變化對電機系統(tǒng)的動態(tài)性能有較大的影響。本文從轉(zhuǎn)動慣量辨識出發(fā),將參數(shù)辨識與參數(shù)自整定結(jié)合,研究能夠適應(yīng)負(fù)載轉(zhuǎn)動慣量頻繁變化的參數(shù)自整定系統(tǒng)。 本文從最小二乘法出發(fā),采用了帶反饋的改進型遞推最小二乘辨識法,通過引入重初始化和邏輯判斷單元檢測辨識算法輸出,在待辨識量變化時觸發(fā)重初始化,徹底消除舊觀測數(shù)據(jù)影響。與原恒定遺忘因子的方法相比,一方面能夠大幅提高動態(tài)響應(yīng)速度,另一方面也能避免因遺忘因子選擇不當(dāng)而造成辨識結(jié)果波動。該算法可以更快地跟蹤時變參數(shù),縮短波動時間,提高辨識性能。 基于本文的轉(zhuǎn)動慣量辨識方法,從異步電機的數(shù)學(xué)模型出發(fā),可將電機以速度環(huán)為外環(huán)的雙閉環(huán)結(jié)構(gòu)簡化為典型Ⅱ型系統(tǒng),進而利用工程上常用的參數(shù)選擇法,建立系統(tǒng)轉(zhuǎn)動慣量與速度環(huán)PI參數(shù)的聯(lián)系,利用轉(zhuǎn)動慣量辨識結(jié)果自動調(diào)整PI控制參數(shù),實現(xiàn)參數(shù)自整定。對于負(fù)載轉(zhuǎn)動慣量頻繁變化的電機系統(tǒng),該自整定策略能夠顯著提高系統(tǒng)動態(tài)穩(wěn)態(tài)性能。 本文首先在Simulink環(huán)境下對異步電機轉(zhuǎn)動慣量參數(shù)辨識與參數(shù)自整定進行了仿真研究,在此基礎(chǔ)上采用dSPACE半實物仿真平臺,搭建外圍電路,進行了空載起動、突加突減負(fù)載等實驗,實驗結(jié)果表明所述理論正確,且在負(fù)載轉(zhuǎn)動慣量頻繁變化的場合具有較高的實用價值。
[Abstract]:In recent years, with the development of power electronics and computer control technology, AC servo control has gradually occupied a dominant position in the servo control field, and the asynchronous motor has been widely used. The research on high performance control of asynchronous motor is of great significance. In many applications of asynchronous motor, the moment of inertia of load is several times or even tens times of that of motor rotor. In this paper, starting from the moment of inertia identification, the parameter identification and parameter self-tuning are combined to study the parameter self-tuning system, which can adapt to the frequent change of the load moment of inertia. Based on the least square method, an improved recursive least squares identification method with feedback is used in this paper. By introducing reinitialization and logic judgement unit to detect the output of identification algorithm, the reinitialization is triggered when the identification quantity changes. Eliminate the influence of the old observation data completely. Compared with the original method of constant forgetting factor, on the one hand, it can greatly improve the dynamic response speed. On the other hand, it can avoid the fluctuation of identification result caused by the improper selection of forgetting factor, and the algorithm can track the time-varying parameters more quickly, shorten the fluctuation time and improve the identification performance. Based on the method of moment of inertia identification in this paper, starting from the mathematical model of asynchronous motor, the double closed loop structure of motor with speed loop as outer ring can be simplified into a typical 鈪，

本文編號：1568923