摘要
在現(xiàn)代工業(yè)生產(chǎn)中，高性能的多電機同步協(xié)調(diào)控制可以提高機械、冶金、造紙、紡織等行業(yè)產(chǎn)品的質(zhì)量和成品率。這些產(chǎn)品的生產(chǎn)中都有物料傳送或類似的過程，該過程中對速度和張力的精確控制是保證產(chǎn)品質(zhì)量的關(guān)鍵，而張力和速度又是相互耦合的，如何對張力和速度進行協(xié)調(diào)控制一直是我們關(guān)注的焦點問題。因此研究一種高效、準確的算法來分別控制速度和張力具有重要意義。
本文以由皮帶連接的三臺交流電機所構(gòu)成的同步系統(tǒng)為對象，進行分析研究，采用基于遺傳算法的對角遞歸神經(jīng)網(wǎng)絡(luò)的PID控制算法來實現(xiàn)速度和張力的協(xié)調(diào)控制。
首先在算法方面，本文對遺傳算法和對角遞歸神經(jīng)網(wǎng)絡(luò)的定義、原理及其特性分別進行了分析討論，針對一般遺傳算法早期收斂，參數(shù)選擇等不足之處，采用了一種改進的自適應(yīng)遺傳算法，給出了交叉率和變異率自適應(yīng)調(diào)整的計算公式；用改進的遺傳算法的迭代學習訓練來獲得對角遞歸網(wǎng)絡(luò)的初始參數(shù)值，并將初值代入到基于對角遞歸神經(jīng)網(wǎng)絡(luò)的PID算法中，對系統(tǒng)的速度張力進行協(xié)調(diào)控制。
然后是對三電機的速度張力系統(tǒng)進行分析，建立系統(tǒng)的數(shù)學模型；并構(gòu)建控制系統(tǒng)的仿真模型進行試驗。結(jié)果表明經(jīng)過初值優(yōu)化的對角遞歸神經(jīng)網(wǎng)絡(luò)PID算法在對多變量非線性系統(tǒng)的控制中，可以不斷地在線修正PID參數(shù)，使系統(tǒng)得到優(yōu)異的控制，顯著改善系統(tǒng)的動靜態(tài)特性，提高系統(tǒng)的穩(wěn)定性和魯棒性。

關(guān)鍵詞 對角遞歸神經(jīng)網(wǎng)絡(luò)；遺傳算法；實數(shù)編碼；張力控制；在線調(diào)整











ABSTRACT
In the modern industrial production, high performance synchronous coordinated control of multi-motors can improve the quality and the rate of finished products of products in machinery industry, metallurgical industry, paper industry, textile industry and so on. There is material conveying or similar process in the product manufacturing, in which process the accuracy control of speed and tension is the key to guarantee product quality. However, speed and tension are mutual coupled. The problem we focus on is how to control speed and tension coordinately. Therefore, it’s most importantly significant to study an efficient and accuracy algorithm to control speed and tension respectively.
This thesis chooses the synchronous system of three AC motors as study object, which are connected by conveyor belt, and has analysis and research on it, then DRNN– PID(Diagonal Recurrent Neural Network, DRNN ) algorithm based on GA (genetic algorithm) is adopted to realize the coordinated control of speed and tension.
Firstly on algorithm aspect, the definitions, principles and characters are analysed and discussed on GA and DRNN in the thesis. Then, an improved adaptive genetic algorithm is used aiming at the deficiencies of general genetic algorithm in the thesis., such as early convergence and parameter selection, the computational formula is provided in which cross rate and mutation rate are self-adaption. The initial parameter values of DRNN-PID algorithm are obtained from the iterative learning training of the improved genetic algorithm, which are used in DRNN-PID algorithm. At last the speed and tension of system can be coordinately controlled.
Secondly, three-motor speed tension system is analyzed, its mathematic model is established in the thesis, and the control module of the system is built to simulate. The results show that the DRNN – PID algorithm whose initial value is optimized can tune PID parameters on line, achieve excellent control effect, notably improve the system dynamic and static characteristics, increase the system stability and robustness in the control towards multivariable nonlinear system.

Key words Diagonal Recurrent Neural Network (DRNN); genetic algorithm(GA); real coding; tension control; online adjustment



目 錄
摘要 I
ABSTRACT III
第一章 緒論 1
1.1 研究的目的和意義 1
1.2 多電機同步系統(tǒng)與張力控制的發(fā)展概況 1
1.3 智能算法在多變量非線性系統(tǒng)中的應(yīng)用 5
1.4 本文研究思路的提出 7
1.5 本文內(nèi)容的安排 8
第二章 遺傳算法和對角遞歸神經(jīng)網(wǎng)絡(luò) 9
2.1 遺傳算法 9
2.1.1 遺傳算法的基本原理 9
2.1.2 遺傳操作 10
2.1.3 基于實數(shù)編碼的遺傳算法 12
2.2 對角遞歸神經(jīng)網(wǎng)絡(luò) 14
2.2.1 神經(jīng)網(wǎng)絡(luò)理論基礎(chǔ) 14
2.2.2 對角遞歸神經(jīng)網(wǎng)絡(luò) 16
2.3基于遺傳算法的DRNN-PID算法 18
2.3.1 基于DRNN的PID控制算法 18
2.3.2 基于遺傳算法的DRNN-PID算法 21
2.4 本章小結(jié) 23
第三章 三電機的速度張力系統(tǒng) 24
3.1 交流電機的矢量控制 24
3.2 張力控制 26
3.3 三電機的速度張力系統(tǒng) 28
3.4 本章小結(jié) 30
第四章 系統(tǒng)仿真模型的構(gòu)建 31
4.1 仿真工具 31
4.1.1 SIMULINK簡介 31
4.1.2 S-Function簡介 33
4.2 系統(tǒng)仿真模型的構(gòu)建 36
4.2.1 構(gòu)建三電機系統(tǒng) 36
4.2.2 三電機速度張力控制系統(tǒng)的設(shè)計 43
4.2.3 控制系統(tǒng)仿真模型的構(gòu)建 44
4.3 本章小結(jié) 46
第五章 仿真試驗分析 47
5.1 開環(huán)試驗分析 47
5.2 基于遺傳算法的對角遞歸神經(jīng)網(wǎng)絡(luò)PID控制試驗分析 49
5.3 本章小結(jié) 62
結(jié)論 63
參考文獻 65