【摘要】：本文主要研究伺服系統(tǒng)的精確建模問(wèn)題,通過(guò)分析機(jī)理建模的復(fù)雜與不精確的問(wèn)題,指出引入神經(jīng)網(wǎng)絡(luò)建模能帶來(lái)的快速性、精確性以及簡(jiǎn)便性的提升。而目前針對(duì)神經(jīng)網(wǎng)絡(luò)辨識(shí)的研究雖然有很多改進(jìn)方案,但是大多都只是在一些特定的仿真模型下效果較好,缺乏實(shí)際系統(tǒng)的驗(yàn)證,有些算法甚至并不適用于實(shí)際系統(tǒng)辨識(shí),因此本文研究基于神經(jīng)網(wǎng)絡(luò)的伺服系統(tǒng)精確模型辨識(shí)問(wèn)題,主要的研究成果可歸納為:首先,對(duì)一類以永磁同步電機(jī)為執(zhí)行元件的位置伺服系統(tǒng),進(jìn)行了標(biāo)稱模型分析與詳細(xì)的攝動(dòng)項(xiàng)環(huán)節(jié)分析,分析了不同非線性環(huán)節(jié)以及攝動(dòng)項(xiàng)會(huì)對(duì)神經(jīng)網(wǎng)絡(luò)辨識(shí)造成哪些影響,為優(yōu)化設(shè)計(jì)神經(jīng)網(wǎng)絡(luò)辨識(shí)方法提供了理論依據(jù)。其次,對(duì)比分析了神經(jīng)網(wǎng)絡(luò)辨識(shí)的基本結(jié)構(gòu)、神經(jīng)網(wǎng)絡(luò)的結(jié)構(gòu)特征與選型依據(jù)、神經(jīng)網(wǎng)絡(luò)訓(xùn)練的基本方法,通過(guò)對(duì)比指出選擇徑向基神經(jīng)網(wǎng)絡(luò)辨識(shí)的選型依據(jù),通過(guò)對(duì)比訓(xùn)練方法的優(yōu)劣為神經(jīng)網(wǎng)絡(luò)的參數(shù)訓(xùn)練方法提供了改進(jìn)方向。然后,結(jié)合伺服系統(tǒng)的特點(diǎn),提出了適用于伺服系統(tǒng)的兩點(diǎn)差分式串-并聯(lián)辨識(shí)結(jié)構(gòu),優(yōu)化了神經(jīng)網(wǎng)絡(luò)的結(jié)構(gòu),改進(jìn)了神經(jīng)網(wǎng)絡(luò)參數(shù)的訓(xùn)練算法,提出將正交最小二乘法(OLS)與梯度下降法(GD)相結(jié)合,能夠有效地改進(jìn)減少神經(jīng)網(wǎng)絡(luò)中心節(jié)點(diǎn)數(shù)量以及降低對(duì)初始位置選取的依賴,然后結(jié)合伺服系統(tǒng)工作的頻段,給出樣本數(shù)據(jù)、測(cè)試數(shù)據(jù)的選擇方法以及給出神經(jīng)網(wǎng)絡(luò)模型的評(píng)價(jià)方法。最后得到一個(gè)一步預(yù)測(cè)的模型結(jié)構(gòu),該結(jié)構(gòu)使用前幾時(shí)刻的實(shí)際數(shù)據(jù)作為輸入能夠準(zhǔn)確預(yù)測(cè)出下一時(shí)刻的輸出,并且通過(guò)仿真實(shí)驗(yàn)驗(yàn)證了改進(jìn)結(jié)構(gòu)和訓(xùn)練算法的有效性。最后,結(jié)合提出的針對(duì)伺服系統(tǒng)改進(jìn)的神經(jīng)網(wǎng)絡(luò)辨識(shí)方案,在實(shí)際轉(zhuǎn)臺(tái)伺服系統(tǒng)當(dāng)中采集開環(huán)的訓(xùn)練樣本與測(cè)試數(shù)據(jù),訓(xùn)練出其神經(jīng)網(wǎng)絡(luò)模型,再通過(guò)與傳統(tǒng)的掃頻方案得到的模型進(jìn)行對(duì)比,驗(yàn)證了神經(jīng)網(wǎng)絡(luò)用于實(shí)際系統(tǒng)建模的可行性。
[Abstract]:In this paper, the exact modeling of servo system is studied. By analyzing the complexity and imprecision of mechanism modeling, the paper points out the improvement of rapidity, accuracy and simplicity brought by the introduction of neural network modeling. However, although there are many improved methods for neural network identification, most of them only work well under some specific simulation models, lacking the verification of the actual system, and some algorithms are not even suitable for the actual system identification. Therefore, in this paper, the exact model identification of servo system based on neural network is studied. The main research results can be summarized as follows: firstly, for a class of position servo system with permanent magnet synchronous motor as the actuator, The nominal model analysis and the detailed analysis of perturbation terms are carried out, and the effects of different nonlinear links and perturbation terms on neural network identification are analyzed, which provides a theoretical basis for the optimization design of neural network identification methods. Secondly, the basic structure of neural network identification, the structural characteristics and selection basis of neural network, the basic training method of neural network, and the selection basis of selecting radial basis function neural network identification are pointed out. By comparing the advantages and disadvantages of the training methods, the improvement direction of the neural network parameter training method is provided. Then, combined with the characteristics of servo system, a two-point differential series-parallel identification structure is proposed for servo system. The structure of neural network is optimized, and the training algorithm of neural network parameters is improved. The combination of orthogonal least square method (OLS) and gradient descent method (GD) can effectively reduce the number of neural network center nodes and reduce the dependence on initial position selection, and then combine with the frequency band of servo system. The sample data, the selection method of test data and the evaluation method of neural network model are given. Finally, a one-step prediction model structure is obtained, which can accurately predict the output of the next moment by using the actual data from the previous time as input, and the effectiveness of the improved structure and the training algorithm is verified by simulation experiments. Finally, combined with the improved neural network identification scheme for servo system, the open-loop training samples and test data are collected in the actual turntable servo system, and its neural network model is trained. Compared with the model obtained by the traditional frequency sweeping scheme, the feasibility of the neural network used in the practical system modeling is verified.
【學(xué)位授予單位】：哈爾濱工業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP183;TM921.54

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 覃業(yè)梅;彭輝;阮文杰;;基于線性函數(shù)型權(quán)重的RBF-ARX模型的磁懸浮球系統(tǒng)預(yù)測(cè)控制[J];中南大學(xué)學(xué)報(bào)(自然科學(xué)版);2016年08期

2 段艷杰;呂宜生;張杰;趙學(xué)亮;王飛躍;;深度學(xué)習(xí)在控制領(lǐng)域的研究現(xiàn)狀與展望[J];自動(dòng)化學(xué)報(bào);2016年05期

3 鄒友龍;胡法龍;周燦燦;李潮流;李長(zhǎng)喜;Keh-Jim Dunn;;徑向基函數(shù)插值方法分析(英文)[J];Applied Geophysics;2013年04期

4 王鵬;關(guān)宇東;沈逢京;杜克;提純利;;一種力矩電機(jī)轉(zhuǎn)矩波動(dòng)系數(shù)自動(dòng)化檢測(cè)方法[J];微電機(jī);2013年11期

5 胡巖;關(guān)朕;吳偉;;異步起動(dòng)永磁同步電動(dòng)機(jī)的轉(zhuǎn)子槽漏抗計(jì)算[J];微特電機(jī);2012年09期

6 陳進(jìn)東;張相勝;潘豐;;基于Wiener模型的非線性預(yù)測(cè)函數(shù)控制[J];吉林大學(xué)學(xué)報(bào)(工學(xué)版);2011年S1期

7 吳德會(huì);;非線性動(dòng)態(tài)系統(tǒng)的Wiener神經(jīng)網(wǎng)絡(luò)辨識(shí)法[J];控制理論與應(yīng)用;2009年11期

8 田一鳴;黃友銳;高志安;黃宜慶;;基于GA與CSA-RBF神經(jīng)網(wǎng)絡(luò)辨識(shí)的自適應(yīng)PID控制器[J];系統(tǒng)仿真學(xué)報(bào);2008年17期

9 張從鵬;劉強(qiáng);;直線電機(jī)定位平臺(tái)的摩擦建模與補(bǔ)償[J];北京航空航天大學(xué)學(xué)報(bào);2008年01期

10 曲東才;;增強(qiáng)神經(jīng)網(wǎng)絡(luò)辨識(shí)模型泛化能力的研究[J];海軍航空工程學(xué)院學(xué)報(bào);2007年01期

相關(guān)博士學(xué)位論文 前2條

1 鄭偉峰;交流伺服系統(tǒng)無(wú)時(shí)滯反饋高響應(yīng)驅(qū)動(dòng)控制研究[D];哈爾濱工業(yè)大學(xué);2010年

2 黃進(jìn);含摩擦環(huán)節(jié)伺服系統(tǒng)的分析及控制補(bǔ)償研究[D];西安電子科技大學(xué);1998年

相關(guān)碩士學(xué)位論文 前5條

1 陳宇飛;機(jī)電伺服系統(tǒng)魯棒控制設(shè)計(jì)與實(shí)現(xiàn)[D];哈爾濱工業(yè)大學(xué);2014年

2 張斯倫;機(jī)電伺服系統(tǒng)低速性能分析與控制設(shè)計(jì)[D];哈爾濱工業(yè)大學(xué);2013年

3 楊旭;基于RBF神經(jīng)網(wǎng)絡(luò)的工業(yè)過(guò)程建模與優(yōu)化研究[D];哈爾濱理工大學(xué);2009年

4 葛翔;基于重復(fù)控制方法的高精度速率伺服控制系統(tǒng)設(shè)計(jì)[D];哈爾濱工業(yè)大學(xué);2007年

5 張新良;非線性系統(tǒng)神經(jīng)網(wǎng)絡(luò)辨識(shí)與控制的研究[D];南京航空航天大學(xué);2004年

，

本文編號(hào)：2415066