aco.rar_MATLAB仿真_蚁群算法_人工智能/神经网络/深度学习

<html xmlns="http://www.w3.org/1999/xhtml"> <head> <meta charset="utf-8"> <meta name="generator" content="pdf2htmlEX"> <meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1"> <link rel="stylesheet" href="https://static.pudn.com/base/css/base.min.css"> <link rel="stylesheet" href="https://static.pudn.com/base/css/fancy.min.css"> <link rel="stylesheet" href="https://static.pudn.com/prod/directory_preview_static/625d4fb3be9ad24cfa7dec5a/raw.css"> <script src="https://static.pudn.com/base/js/compatibility.min.js"></script> <script src="https://static.pudn.com/base/js/pdf2htmlEX.min.js"></script> <script> try{ pdf2htmlEX.defaultViewer = new pdf2htmlEX.Viewer({}); }catch(e){} </script> <title></title> </head> <body> <div id="sidebar" style="display: none"> <div id="outline"> </div> </div> <div id="pf1" class="pf w0 h0" data-page-no="1"><div class="pc pc1 w0 h0"><img class="bi x0 y0 w1 h1" alt="" src="https://static.pudn.com/prod/directory_preview_static/625d4fb3be9ad24cfa7dec5a/bg1.jpg"><div class="t m0 x1 h2 y1 ff1 fs0 fc0 sc0 ls0 ws0">第 35 卷 第 6 期 自 动 化 学 报 Vol. 35, No. 6</div><div class="t m0 x1 h2 y2 ff2 fs1 fc0 sc0 ls0 ws0">2009 年 6 月 ACTA AUTOMATICA SINICA June, 2009</div><div class="t m0 x2 h3 y3 ff3 fs2 fc0 sc0 ls0 ws0">基于蚁群优化算法的精密伺服转台故障诊断方法</div><div class="t m0 x3 h4 y4 ff4 fs3 fc0 sc0 ls0 ws0">甄子洋</div><div class="t m0 x4 h5 y5 ff5 fs4 fc0 sc0 ls0 ws0">1</div><div class="t m0 x5 h4 y4 ff4 fs3 fc0 sc0 ls0 ws0">王道波</div><div class="t m0 x6 h5 y5 ff5 fs4 fc0 sc0 ls0 ws0">1</div><div class="t m0 x7 h4 y4 ff4 fs3 fc0 sc0 ls0 ws0">王志胜</div><div class="t m0 x8 h5 y5 ff5 fs4 fc0 sc0 ls0 ws0">1</div><div class="t m0 x9 h2 y6 ff3 fs0 fc0 sc0 ls0 ws0">摘 要 提出了一种基于蚁群优化算法的精密伺服转台故障诊断方法. 根据现场观测建立了转台系统故障特征模式库. 利用</div><div class="t m0 x9 h2 y7 ff1 fs0 fc0 sc0 ls0 ws0">蚁群优化算法求解故障特征模式的最优分类问题, 并定义敏感度和明确度来评价蚁群搜索到的诊断规则的分类性能, 以减少</div><div class="t m0 x9 h2 y8 ff1 fs0 fc0 sc0 ls0 ws0">故障特征信息中的冗余信息, 使诊断规则得到约简. 对某精密伺服转台的若干类故障诊断结果表明, 该方法具有收敛速度快、</div><div class="t m0 x9 h2 y9 ff1 fs0 fc0 sc0 ls0 ws0">鲁棒性强、诊断精度高和结果可靠等优点.</div><div class="t m0 x9 h2 ya ff3 fs0 fc0 sc0 ls0 ws0">关键词 精密伺服转台, 蚁群优化, 故障诊断, 数据挖掘, 分类算法</div><div class="t m0 x9 h6 yb ff3 fs0 fc0 sc0 ls0 ws0">中图分类号 TP273</div><div class="t m0 xa h7 yc ff6 fs5 fc0 sc0 ls0 ws0">Ant Colony Optimization for Fault Diagnosis of High Precision Servo Simulator</div><div class="t m0 xb h8 yd ff2 fs1 fc0 sc0 ls0 ws0">ZHEN Zi-Yang</div><div class="t m0 xc h5 ye ff5 fs4 fc0 sc0 ls0 ws0">1</div><div class="t m0 xd h8 yd ff2 fs1 fc0 sc0 ls0 ws0">WANG Dao-Bo</div><div class="t m0 xe h5 ye ff5 fs4 fc0 sc0 ls0 ws0">1</div><div class="t m0 xf h8 yd ff2 fs1 fc0 sc0 ls0 ws0">WANG Zhi-Sheng</div><div class="t m0 x10 h5 ye ff5 fs4 fc0 sc0 ls0 ws0">1</div><div class="t m0 x9 h8 yf ff7 fs1 fc0 sc0 ls0 ws0">Abstract A new fault diagnosis method for high precision servo simulator is presented in this paper. Based on the</div><div class="t m0 x9 h8 y10 ff2 fs1 fc0 sc0 ls0 ws0">ﬁeld observation, a fault feature pattern base is built up. The ant colony optimization (ACO) is used for the optimal</div><div class="t m0 x9 h8 y11 ff2 fs1 fc0 sc0 ls0 ws0">classiﬁcation problem of the fault feature patterns, in which an index function based on the sensitivity and the speciﬁcity is</div><div class="t m0 x9 h8 y12 ff2 fs1 fc0 sc0 ls0 ws0">deﬁned to evaluate the performance of the diagnostic rule obtained by ACO, in order to decrease the redundant information</div><div class="t m0 x9 h8 y13 ff2 fs1 fc0 sc0 ls0 ws0">of the fault features so that the diagnosis rules can be reduced. Diagnosis results of several types of faults of a real high</div><div class="t m0 x9 h8 y14 ff2 fs1 fc0 sc0 ls0 ws0">precision servo simulator show that the diagnosis method based on ACO is characterized by fast convergence speed, strong</div><div class="t m0 x9 h8 y15 ff2 fs1 fc0 sc0 ls0 ws0">robustness, high diagnosis accuracy, and reliable results.</div><div class="t m0 x9 h8 y16 ff7 fs1 fc0 sc0 ls0 ws0">Key words High precision servo simulator, ant colony optimization (ACO), fault diagnosis, data mining, classiﬁcation</div><div class="t m0 x9 h8 y17 ff2 fs1 fc0 sc0 ls0 ws0">algorithm</div><div class="t m0 x11 h9 y18 ff1 fs6 fc0 sc0 ls0 ws0">故障诊断技术的发展主要经历了人工诊断、现</div><div class="t m0 x9 h9 y19 ff1 fs6 fc0 sc0 ls0 ws0">代诊断和智能诊断等三个阶段</div><div class="t m0 x12 ha y1a ff8 fs7 fc0 sc0 ls0 ws0">[1]</div><div class="t m0 x13 h9 y19 ff9 fs8 fc0 sc0 ls0 ws0">. 发展至今, 故障诊</div><div class="t m0 x9 h9 y1b ff1 fs6 fc0 sc0 ls0 ws0">断方法可划分为基于解析模型的方法、基于信号处</div><div class="t m0 x9 h9 y1c ff1 fs6 fc0 sc0 ls0 ws0">理的方法和基于知识的方法等. 近年来, 基于人工智</div><div class="t m0 x9 h9 y1d ff1 fs6 fc0 sc0 ls0 ws0">能的数据挖掘技术在复杂系统故障检测与诊断研究</div><div class="t m0 x9 h9 y1e ff1 fs6 fc0 sc0 ls0 ws0">中得到广泛应用, 包括神经网络</div><div class="t m0 x14 ha y1f ff8 fs7 fc0 sc0 ls0 ws0">[2]</div><div class="t m0 x15 h9 y1e ff1 fs6 fc0 sc0 ls0 ws0">、支持向量机</div><div class="t m0 x16 ha y1f ff8 fs7 fc0 sc0 ls0 ws0">[3]</div><div class="t m0 x17 h9 y1e ff1 fs6 fc0 sc0 ls0 ws0">、</div><div class="t m0 x9 h9 y20 ff1 fs6 fc0 sc0 ls0 ws0">粗糙集</div><div class="t m0 x18 ha y21 ff8 fs7 fc0 sc0 ls0 ws0">[4]</div><div class="t m0 x19 h9 y20 ff1 fs6 fc0 sc0 ls0 ws0">、智能优化算法</div><div class="t m0 x1a ha y21 ff8 fs7 fc0 sc0 ls0 ws0">[5]</div><div class="t m0 x1b h9 y20 ff1 fs6 fc0 sc0 ls0 ws0">以及相互之间的结合</div><div class="t m0 x1c ha y21 ff8 fs7 fc0 sc0 ls0 ws0">[6−8]</div><div class="t m0 x9 h9 y22 ff1 fs6 fc0 sc0 ls0 ws0">等方法. 蚁群优化 (Ant colony optimization, ACO)</div><div class="t m0 x9 h9 y23 ff1 fs6 fc0 sc0 ls0 ws0">算法由 Dorigo 等通过模拟自然界蚂蚁集体出巢搜</div><div class="t m0 x9 h9 y24 ff1 fs6 fc0 sc0 ls0 ws0">索最短觅食路径的行为而提出</div><div class="t m0 x1d ha y25 ff8 fs7 fc0 sc0 ls0 ws0">[9]</div><div class="t m0 x1e h9 y24 ff9 fs8 fc0 sc0 ls0 ws0">, Parpinelli 等将</div><div class="t m0 x9 h9 y26 ff9 fs8 fc0 sc0 ls0 ws0">ACO 算法用于医学诊断应用中, 可获得比决策树分</div><div class="t m0 x9 h9 y27 ff1 fs6 fc0 sc0 ls0 ws0">类算法 CN2 更高的分类准确率和更简便的分类规</div><div class="t m0 x9 h9 y28 ff1 fs6 fc0 sc0 ls0 ws0">则</div><div class="t m0 x1f ha y29 ff8 fs7 fc0 sc0 ls0 ws0">[10]</div><div class="t m0 x20 hb y28 ff9 fs8 fc0 sc0 ls0 ws0">.</div><div class="t m0 x21 hc y2a ff1 fs9 fc0 sc0 ls0 ws0">收稿日期 2008-12-16 收修改稿日期 2009-03-11</div><div class="t m0 x21 hd y2b ffb fsa fc0 sc0 ls0 ws0">Received December 16, 2008; in revised form March 11, 2009</div><div class="t m0 x21 hc y2c ff1 fs9 fc0 sc0 ls0 ws0">国家自然科学基金 (60874037), 教育部博士点科研基金 (200702870-</div><div class="t m0 x9 hc y2d ffb fsa fc0 sc0 ls0 ws0">50), 江苏省普通高校研究生科研创新计划 (CX08B</div><div class="t m0 x22 he y2e ffc fsb fc0 sc0 ls0 ws0">−</div><div class="t m0 x23 hc y2d ffb fsa fc0 sc0 ls0 ws0">091Z), 南京航空</div><div class="t m0 x9 hc y2f ff1 fs9 fc0 sc0 ls0 ws0">航天大学博士学位论文创新与创优基金 (BCXJ08-06) 资助</div><div class="t m0 x21 hd y30 ffb fsa fc0 sc0 ls0 ws0">Supported by National Natural Science Foundation of China</div><div class="t m0 x9 hd y31 ffb fsa fc0 sc0 ls0 ws0">(60874037), Specialized Research Fund for the Doctoral Pro-</div><div class="t m0 x9 hd y32 ffb fsa fc0 sc0 ls0 ws0">gram of Higher Education of China (20070287050), Gradu-</div><div class="t m0 x9 hd y33 ffb fsa fc0 sc0 ls0 ws0">ate Student Research Innovation Program of Jiangsu Province</div><div class="t m0 x9 hd y34 ffb fsa fc0 sc0 ls0 ws0">(CX08B</div><div class="t m0 x24 he y35 ffc fsb fc0 sc0 ls0 ws0">−</div><div class="t m0 x25 hd y34 ffb fsa fc0 sc0 ls0 ws0">091Z), and Innovation and Excellence Foundation of</div><div class="t m0 x9 hd y36 ffb fsa fc0 sc0 ls0 ws0">Doctoral Dissertation of Nanjing University of Aeronautics and</div><div class="t m0 x9 hd y37 ffb fsa fc0 sc0 ls0 ws0">Astronautics (BCXJ08-06)</div><div class="t m0 x21 hc y38 ffb fsa fc0 sc0 ls0 ws0">1. 南京航空航天大学自动化学院 南京 210016</div><div class="t m0 x21 hd y39 ffb fsa fc0 sc0 ls0 ws0">1. College of Automation Engineering, Nanjing University of</div><div class="t m0 x9 hd y3a ffb fsa fc0 sc0 ls0 ws0">Aeronautics and Astronautics, Nanjing 210016</div><div class="t m0 x21 hd y3b ffb fsa fc0 sc0 ls0 ws0">DOI: 10.3724/SP.J.1004.2009.00780</div><div class="t m0 x26 h9 y18 ff1 fs6 fc0 sc0 ls0 ws0">精密伺服转台是一种用于模拟飞行器姿态变化</div><div class="t m0 x27 h9 y19 ff1 fs6 fc0 sc0 ls0 ws0">的高复杂、高精度和高可靠的机电设备, 是保障惯</div><div class="t m0 x27 h9 y1b ff1 fs6 fc0 sc0 ls0 ws0">性导航系统地面仿真试验精度的关键设备. 为保证</div><div class="t m0 x27 h9 y1c ff1 fs6 fc0 sc0 ls0 ws0">各个电子、机械、软件及其他部件长期安全可靠地</div><div class="t m0 x27 h9 y1d ff1 fs6 fc0 sc0 ls0 ws0">运行, 提高故障诊断技术的研究具有非常重要的意</div><div class="t m0 x27 h9 y1e ff1 fs6 fc0 sc0 ls0 ws0">义</div><div class="t m0 x28 ha y1f ff8 fs7 fc0 sc0 ls0 ws0">[11]</div><div class="t m0 x29 h9 y1e ff9 fs8 fc0 sc0 ls0 ws0">. 针对转台系统的故障诊断问题, 文献 [12] 运</div><div class="t m0 x27 h9 y20 ff1 fs6 fc0 sc0 ls0 ws0">用粗糙集理论直接从故障样本中导出诊断规则, 并</div><div class="t m0 x27 h9 y22 ff1 fs6 fc0 sc0 ls0 ws0">对决策表进行约简处理; 接着, 文献 [13] 提出了粗糙</div><div class="t m0 x27 h9 y23 ff1 fs6 fc0 sc0 ls0 ws0">集、ACO 算法和 BP 网络相结合的在线诊断方法.</div><div class="t m0 x26 h9 y24 ff1 fs6 fc0 sc0 ls0 ws0">本文的研究对象是课题组自主研制的某精密伺</div><div class="t m0 x27 h9 y26 ff1 fs6 fc0 sc0 ls0 ws0">服转台, 首先建立故障特征模式库, 然后通过 ACO</div><div class="t m0 x27 h9 y27 ff1 fs6 fc0 sc0 ls0 ws0">算法寻优获得含较少条件项且较高诊断精度的故障</div><div class="t m0 x27 h9 y28 ff1 fs6 fc0 sc0 ls0 ws0">诊断规则, 最后进行仿真研究.</div><div class="t m0 x27 hf y3c ffd fs5 fc0 sc0 ls0 ws0">1 精密伺服转台系统及其故障特征</div><div class="t m0 x27 h10 y3d ff9 fs8 fc0 sc0 ls0 ws0">1.1 伺服转台系统</div><div class="t m0 x26 h9 y3e ff1 fs6 fc0 sc0 ls0 ws0">精密伺服转台系统主要由指令系统、控制系统、</div><div class="t m0 x27 h9 y3f ff1 fs6 fc0 sc0 ls0 ws0">执行器、反馈元件和被控对象组成, 其硬件组成框图</div><div class="t m0 x27 h9 y40 ff1 fs6 fc0 sc0 ls0 ws0">如图 1 (见下页) 所示. 图 1 中, 指令系统构成人机</div><div class="t m0 x27 h9 y41 ff1 fs6 fc0 sc0 ls0 ws0">交互界面, 可从信号发生器和开关按键输入指令, 通</div><div class="t m0 x27 h9 y42 ff1 fs6 fc0 sc0 ls0 ws0">过 LED 数码显示器和发光二极管指示转台工作状</div><div class="t m0 x27 h9 y43 ff1 fs6 fc0 sc0 ls0 ws0">态并监控转台. 控制系统由控制微机和控制器构成,</div><div class="t m0 x27 h9 y44 ff1 fs6 fc0 sc0 ls0 ws0">控制微机完成位置环的数字控制和数据处理等功能,</div><div class="t m0 x27 h9 y3b ff1 fs6 fc0 sc0 ls0 ws0">而控制器完成对转台工作状态的监控保护以及对电</div></div><div class="pi" data-data='{"ctm":[1.611984,0.000000,0.000000,1.611984,0.000000,0.000000]}'></div></div> </body> </html>