WANG Yali*,YU Jiming,WANG Yi.Prediction of Wear Resistance of Self-Lubricating Coating by Genetic Algorithm-Fuzzy Radial Basis Function Neural Network Model[J].Plating & Finishing,2021,(7):30-34.[doi:10.3969/j.issn.1001-3849.2021.07.006]
遗传算法-模糊径向基神经网络模型预测自润滑镀层耐磨性
- Title:
- Prediction of Wear Resistance of Self-Lubricating Coating by Genetic Algorithm-Fuzzy Radial Basis Function Neural Network Model
- 关键词:
- 自润滑镀层; 摩擦因数; 遗传算法-模糊径向基神经网络模型; 径向基神经网络模型
- 文献标志码:
- A
- 摘要:
- 针对传统神经网络模型存在的缺陷，引入遗传算法和模糊运算建立遗传算法-模糊径向基神经网络模型（GA-FRBFNNM），介绍了模型结构和仿真思路。以自润滑镀层耐磨性为研究主题开展正交实验，在正交实验结果中任取10组数据作为训练样本用于模型训练，其余6组数据作为测试样本用于模型性能测试。结果表明：GA-FRBFNNM的预测值更接近于真实值，其预测精度明显高于相同结构的径向基神经网络模型，验证了该模型是有效的，能够更准确预测自润滑镀层耐磨性。主要归因于引入模糊运算使得径向基神经网络全部节点都具备特定意义，另外引入遗传算法优化了训练算法，避免了模型陷入局部极小点等问题，使得模型性能得到有效提升。
- Abstract:
- Aiming at the defects of traditional neural network model, the genetic algorithm-fuzzy radial basis function neural network model （GA-FRBFNNM） was established by introducing the genetic algorithm and fuzzy operation, and the structure and simulation thoughts of the model were introduced. Orthogonal experiment was carried out with the wear resistance of self-lubricating coating as the research theme, and any ten groups of data in the orthogonal experiment results were taken as training samples for model training, and other six groups of data were taken as test samples for model performance testting. The results showed that the predicted value of GA-FRBFNNM was closer to the real value, and its prediction accuracy was significantly higher than that of the radial basis neural network model with the same structure, which indicated that the model was effective and it can accurately predict the wear resistance of self-lubricating coating. It was mainly attributed to the introduction of fuzzy operation, which makes all nodes of RBF neural network have specific meaning. In addition, the introduction of genetic algorithm to optimize the training algorithm, which avoids the model falling into local minimum point and other problems, thus effectively improving the model performance.
参考文献/References:
[1] 杨志昆.纳米溶胶对Ni基自润滑复合镀层微观组织及性能影响的研究[D]. 哈尔滨:哈尔滨工程大学, 2014.
Yang Z K.The effect of nano-sol on the microstructure and properties of Ni-based self-lubrication composite coating[D]. Harbin:Harbin Engineering University,2014 (in Chinese).
[2] Cheng Q,Yao Z J,Zhang F,et al. Microstructure and tribological property of Ni-MoS2 composite coatings prepared by ultrasonic and mechanical stirring electrodeposition[J]. Materials Research Express, 2019, 6(12):126434.
[3] Sangeetha S,Kalaignan G P,Anthuvan J T.Pulse electrodeposition of self-lubricating Ni-W/PTFE nanocomposite coatings on mild steel surface[J]. Applied Surface Science,2015, 359:412-419.
[4] He Y,Wang S C,Walsh F C.Self-lubricating Ni-P-MoS2 composite coatings[J]. Surface and Coatings Technology,2016,307:926-934.
[5] Ni M,Wang S L,Li W D,et al.A novel self-lubricating Ni-P-AlN-WS2 nanocomposite coating[J]. Materials Research Express, 2019, 6(11):116413.
[6] Chen M,Cheng W,Zhao Z,et al. Synergistic effect between nano-ceramic lubricating additives and electroless deposited Ni-W-P coating[J]. Chinese Journal of Mechanical Engineering, 2013, 26:114-120.
[7] 宋昌熙,郑春花,车硕源,等.基于人工神经网络的固体氧化物燃料电池性能预测模型开发[J]. 集成技术, 2020, 9(5):27-33.
Song C H,Zheng C H,Cha S W,et al.Development of a simulation model for polarization characteristics of solid oxide fuel cells based on an artificial neural network[J]. Journal of Integration Technology,2020,9(5):27-33 (in Chinese).
[8] 韩建军,南少伟,王俊伟.基于支持向量机的有机成膜涂层混凝土碳化深度预测模型[J]. 施工技术,2020,49(2):94-98.
Han J J,Nan S W,Wang J W.Carbonization depth prediction model of organic film-forming coating concrete based on support vector machine[J]. Construction Technology, 2020, 49(2):94-98 (in Chinese).
[9] 杨武,张春燕,马超.基于GA-BP神经网络的镁合金微弧氧化膜层厚度预测[J]. 兵器材料科学与工程,2017,40(1):88-92.
Yang W,Zhang C Y,Ma C.Thickness prediction of micro-arc oxidation coating on magnesium alloy based on GA-BP neural network[J]. Ordnance Material Science and Engineerin, 2017, 40(1):88-92 (in Chinese).
[10] 柯超,甘屹,王俊,等.基于BP神经网络辨识模型的PEMFC系统建模[J]. 软件工程, 2020, 23(3):14-16.
Ke C,Gan Y,Wang J,et al.PEMFC system modeling based on BPNN identification model[J]. Software Engineer, 2020, 23(3):14-16 (in Chinese).
[11] 杨军华,李勇,程伟.基于最邻近相关系数的指纹室内定位新算法[J]. 西北工业大学学报, 2017, 35(4):676-682.
Yang J H,Li Y,Cheng W.A novel algorithm for fingerprinting indoor localization based on correlation coefficient[J]. Journal of Northwestern Polytechnical University, 2017, 35(4):676-682 (in Chinese).
[12] 贾晓芬,郭永存,黄友锐,等.基于斯皮尔曼等级相关性的彩色图像椒盐噪点检测算法[J]. 中国科学技术大学学报, 2019, 49(1):63-70.
Jia X F,Guo Y C,Huang Y R,et al.Salt&pepper noise detection algorithm for color images based on spearman rank correlation[J]. Journal of University of Science and Technology of China, 2019, 49(1):63-70 (in Chinese).
备注/Memo
收稿日期： 2021-01-13；修回日期： 2021-03-16
*通信作者： 王亚利， access082001@126.com
基金项目： 河南省科技厅科技攻关项目（202102210384）