基于BP神经网络及其改进算法的织机效率预测

doi:10.13475/j.fzxb.20190402507

摘要/Abstract

摘要：

为准确预测纺织厂织布车间的织机效率,提出利用BP神经网络、主成分分析结合BP神经网络(PCA-BP)、遗传算法改进BP神经网络(GA-BP)3种模型预测织机效率,并将GA-BP预测模型与传统BP神经网络和PCA-BP预测模型的预测结果进行对比分析。结果表明:GA-BP对原始数据的拟合度最好,相关系数为0.946 87, 比BP增加了6.42%,比PCA-BP增加了2.61%;GA-BP、PCA-BP、BP这3种网络十万入纬的经停仿真值与期望值间的平均误差分别为0.341 2、0.303 1、0.234 1,误差百分率分别为8.63%、7.67%、5.92%,不同网络结构下织机效率仿真预测值与期望值间的平均误差分别为3.010 9、2.688 4、2.118 9,误差百分率分别为3.51%、3.13%、2.47%;3种模型的预测准确度顺序由大到小为GA-BP、PCA-BP、BP。

关键词: BP神经网络, 遗传算法, 主成分分析, 预测模型, 织机效率预测

Abstract:

In order to predict the loom efficiency more accurately in the weaving workshop of textile mills, three models, i.e. BP neural network, principal component analysis combined with BP neural network(PCA-BP) and genetic algorithm modified BP neural network model (GA-BP), were used to predict the loom efficiency. At the same time, the prediction results of the GA-BP were compared with that of the BP neural network and PCA-BP neural network. The results show that the GA-BP has the best fitting degree to the original data, the correlation coefficient is 0.946 87, which is 6.42% higher than BP and 2.61% higher than PCA-BP. The average absolute errors between the simulated output value and the expected loom stoppage values over 100 000 weft insertions are 0.341 2, 0.303 1 and 0.234 1, respectively, for GA-BP, PCA-BP and BP models, corresponding to error percentages 8.63%, 7.67% and 5.92%. The average errors between the predicted and the expected values of the loom efficiency with different network models are 3.010 9, 2.688 4 and 2.118 9, respectively, with error percentages of 3.51%, 3.13%, 2.47%. The order of prediction accuracy of the three models is GA-BP, PCA-BP and BP.

Key words: BP neural network, genetic algorithm, principal component analysis, prediction model, loom efficiency prediction

中图分类号:

TS104.2

张晓侠, 刘凤坤, 买巍, 马崇启. 基于BP神经网络及其改进算法的织机效率预测[J]. 纺织学报, 2020, 41(08): 121-127.

ZHANG Xiaoxia, LIU Fengkun, MAI Wei, MA Chongqi. Prediction of loom efficiency based on BP neural network and its improved algorithm[J]. Journal of Textile Research, 2020, 41(08): 121-127.

图/表 7

图1

图2

图3

图4

图5

表1

表2

参考文献 17

[1]	RAMESH M C, RAJAMANICKAM R, JAYARAMAN S. The prediction of yarn tensile properties by using artificial neural networks[J]. Journal of The Textile Institute, 1995,86:459-469. doi: 10.1080/00405009508658772
[2]	倪红, 潘永惠. 基于BP神经网络的织物斜向弯曲性能的预测[J]. 纺织学报, 2009,30(2):48-51.
	NI Hong, PAN Yonghui. Prediction of diagonal bending performance of fabrics based on BP neural network[J]. Journal of Textile Research, 2009,30(2):48-51.
[3]	董旭烨, 李竹君. 精益生产工艺提高喷气织机的织造效率[J]. 纺织导报, 2015 (10):97-100.
	DONG Xuye, LI Zhujun. Lean production technology improves weaving efficiency of air-jet looms[J]. China Textile Leader, 2015(10):97-100.
[4]	王少芳, 蔡金锭, 刘庆珍. 基于改进GA-BP混合算法的电力变压器故障诊断[J]. 电网技术, 2004,28(4):30-33.
	WANG Shaofang, CAI Jinding, LIU Qingzhen. Power transformer fault diagnosis based on improved GA-BP hybrid algorithm[J]. Power Grid Technology, 2004,28(4):30-33.
[5]	孙晓刚, 原桂彬, 戴景民. 基于遗传神经网络的多光谱辐射测温法[J]. 光谱学与光谱分析, 2007 (2):213-216. pmid: 12961851
	SUN Xiaogang, YUAN Guibin, DAI Jingmin. Multispectral radiation thermometry based on genetic neural network[J]. Spectroscopy and Spectral Analysis, 2007(2):213-216. pmid: 12961851
[6]	司学锋. 基于聚类的BP神经网络在织物染色计算机配色中的应用研究[D]. 青岛:青岛大学, 2009: 32-36.
	SI Xuefeng. Application of BP neural network based on clustering in computer color matching of fabric dyeing[D]. Qingdao: Qingdao University, 2009: 32-36.
[7]	LUO Jianfei, LIN Weitie, CAI Xiaolong, et al. Optimization of nitrobacterial fermentation medium based on neural network and genetic algorithms[J]. Chinese Journal of Chemical Engineering, 2012,20(5):950-957.
[8]	FENG Yan, CHEN Yimin. Rectification of magnetic force tracker using neural network inaugmented reality system[J]. Journal of Shanghai University(English Edition), 2006(5):431-435. doi: 10.1007/s11741-006-0086-6
[9]	周松林, 茆美琴, 苏建徽. 基于主成分分析与人工神经网络的风电功率预测[J]. 电网技术, 2011,35(9):128-132.
	ZHOU Songlin, MAO Meiqin, SU Jianhui. Wind power prediction based on principal component analysis and artificial neural network[J]. Power System Technology, 2011,35(9):128-132.
[10]	张文霖. 主成分分析在SPSS中的操作应用[J]. 市场研究, 2005(12):31-34.
	ZHANG Wenlin. Operational application of principal component analysis in SPSS[J]. Market Research, 2005(12):31-34.
[11]	雷彦森. 遗传算法优化的BP神经网络在多模式集成预报的应用研究[D]. 南京:南京信息工程大学, 2018: 47-54.
	LEI Yansen. Application of BP neural network optimized by genetic algorithm in multi-model integrated forecasting[D]. Nanjing: Nanjing University of Information Science & Technology, 2018: 47-54.
[12]	王冰玉, 孙威江, 黄艳, 等. 基于遗传算法的安溪铁观音品质快速评价研究[J]. 光谱学与光谱分析, 2017,37(4):1100-1104.
	WANG Bingyu, SUN Weijiang, HUANG Yan, et al. Rapid evaluation of Tieguanyin Anxi based on genetic algorithm[J]. Spectroscopy and Spectral Analysis, 2017,37(4):1100-1104.
[13]	李勇, 陈晓川, 汪军, 等. 基于BP神经网络的原棉短纤指数预测模型[J]. 纺织学报, 2014,35(11):35-39.
	LI Yong, CHEN Xiaochuan, WANG Jun, et al. Raw cotton staple fiber index prediction model based on BP neural network[J]. Journal of Textile Research, 2014,35(11):35-39.
[14]	刘贵, 于伟东. 基于遗传算法和BP网络的精毛纺粗纱质量预报[J]. 纺织学报, 2009,30(5):28-33.
	LIU Gui, YU Weidong. Prediction of worsted roving quality based on genetic algorithm and BP network[J]. Journal of Textile Research, 2009,30(5):28-33.
[15]	吴耀, 杨瑞峰, 郭晨霞, 等. 基于GA-BP神经网络的光纤位移传感器光强补偿研究[J]. 电光与控制, 2019,26(4):111-114.
	WU Yao, YANG Ruifeng, GUO Chenxia, et al. Research on light intensity compensation of optical fiber displacement sensor based on GA-BP neural network[J]. Electronics Optic and Control, 2019,26(4):111-114.
[16]	邓伟锋, 李振璧. 基于GA优化BP神经网络的微电网蓄电池健康状态评估[J]. 电测与仪表, 2018,55(21):56-60, 85.
	DENG Weifeng, LI Zhenbi. Assessment of battery health in microgrid based on GA optimized BP neural net-work[J]. Electrical Measurement and Instrument, 2018,55(21):56-60, 85.
[17]	查刘根, 谢春萍. 基于免疫遗传算法的BP神经网络在纱线条干预测上的应用[J]. 丝绸, 2019,56(2):19-26.
	ZHA Liugen, XIE Chunping. Application of BP neural network based on immune genetic algorithms in yarn intervention measurement[J]. Journal of Silk, 2019,56(2):19-26.

相关文章 15

[1]	张卓, 丛洪莲, 蒋高明, 董智佳. 基于交互式遗传算法的Polo衫快速款式推荐系统[J]. 纺织学报, 2021, 42(01): 138-144.
[2]	李亮, 倪俊芳. 绗缝机花样加工代码自动生成算法[J]. 纺织学报, 2020, 41(11): 162-167.
[3]	谢子昂, 杜劲松, 赵国华. 衬衫吊挂流水线的自适应动态调度[J]. 纺织学报, 2020, 41(10): 144-149.
[4]	黄珍珍, 莫碧贤, 温李红. 基于遗传算法及仿真技术的服装生产流水线平衡[J]. 纺织学报, 2020, 41(07): 154-159.
[5]	郑小虎, 鲍劲松, 马清文, 周衡, 张良山. 基于模拟退火遗传算法的纺纱车间调度系统[J]. 纺织学报, 2020, 41(06): 36-41.
[6]	莫帅, 冯战勇, 唐文杰, 党合玉, 邹振兴. 基于神经网络和遗传算法的锭子弹性管性能优化[J]. 纺织学报, 2020, 41(04): 161-166.
[7]	黄倩倩, 李俊. 环境温度突变时人体热感觉变化机制研究进展[J]. 纺织学报, 2020, 41(04): 188-194.
[8]	黄淇, 周其洪, 张倩, 王绍宗, 范伟, 孙会丰. 基于系统布置设计-遗传算法的纱线浸染生产线布局优化[J]. 纺织学报, 2020, 41(03): 84-90.
[9]	张旭靖, 王立川, 陈雁. 基于遗传算法的服装缝制生产线平衡优化[J]. 纺织学报, 2020, 41(02): 125-129.
[10]	王晓晖, 刘月刚, 孟婥, 孙以泽. 基于遗传算法和神经网络的3D增材印花工艺参数优化[J]. 纺织学报, 2019, 40(11): 168-174.
[11]	周捷, 马秋瑞. 基于BP神经网络的运动文胸肩带属性与乳房振幅的函数关系[J]. 纺织学报, 2019, 40(09): 186-191.
[12]	孟朔, 潘如如, 高卫东, 王静安, 周利军. 采用主目标进化遗传算法的织造排程研究[J]. 纺织学报, 2019, 40(08): 169-174.
[13]	万莉, 贡丽英, 贾敏瑞. 基于主成分分析的智能复合材料结构损伤类型识别[J]. 纺织学报, 2019, 40(05): 53-58.
[14]	周捷, 马秋瑞. BP神经网络在塑身内衣压力预测中的应用[J]. 纺织学报, 2019, 40(04): 111-116.
[15]	查刘根, 谢春萍. 应用四层BP神经网络的棉纱成纱质量预测[J]. 纺织学报, 2019, 40(01): 52-56.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

验证样本序号	期望值	BP		PCA-BP		GA-BP
验证样本序号	期望值	仿真值1	误差1	仿真值2	误差2	仿真值3	误差3
1	4.42	4.384 2	0.035 8	3.976 4	0.443 6	4.412 0	0.008 0
2	5.91	5.323 3	0.586 7	6.180 8	0.270 8	5.206 6	0.703 4
3	6.61	6.337 0	0.273 0	6.351 8	0.258 2	6.421 4	0.188 6
4	1.81	2.042 9	0.232 9	1.674 5	0.135 5	1.807 8	0.002 2
5	2.45	1.646 8	0.803 2	2.588 0	0.138 0	2.364 8	0.085 2
6	3.96	3.522 1	0.437 9	3.516 5	0.443 5	4.371 6	0.411 6
7	1.10	1.593 8	0.493 8	1.610 3	0.510 3	1.358 3	0.258 3
8	1.69	1.857 7	0.167 7	2.271 7	0.581 7	1.751 8	0.061 8
9	1.98	1.857 7	0.122 3	2.271 7	0.291 7	1.751 8	0.228 2
10	6.77	5.882 2	0.887 8	6.877 0	0.107 0	6.194 4	0.575 6
11	6.61	6.717 6	0.107 6	6.685 6	0.075 6	6.415 1	0.194 9
12	2.73	2.639 4	0.090 6	2.167 6	0.562 4	2.437 4	0.292 6
13	1.64	1.917 9	0.277 9	1.567 4	0.072 6	1.878 3	0.238 3
14	3.45	3.095 9	0.354 1	3.479 5	0.029 5	3.256 9	0.193 1
15	5.52	5.406 7	0.113 3	5.124 4	0.395 6	5.507 0	0.013 0
16	5.28	5.045 1	0.234 9	5.490 0	0.210 0	5.275 6	0.004 4
17	5.19	5.454 3	0.264 3	4.745 1	0.444 9	5.175 2	0.014 8
18	4.00	3.342 9	0.657 1	4.484 6	0.484 6	3.260 5	0.739 5
平均误差			0.341 2		0.303 1		0.234 1
误差百分率/%			8.63		7.67		5.92

验证样本序号	期望值	BP		PCA-BP		GA-BP
验证样本序号	期望值	仿真值1	误差1	仿真值2	误差2	仿真值3	误差3
1	91.93	88.504 0	3.426 0	90.006 8	1.923 2	94.073 9	2.143 9
2	79.48	76.990 0	2.490 0	85.418 5	5.938 5	76.359 6	3.120 4
3	76.38	77.141 8	0.761 8	77.747 7	1.367 7	75.495 9	0.884 1
4	93.41	88.099 5	5.310 5	85.948 9	7.461 1	92.445 8	0.964 2
5	81.02	88.169 2	7.149 2	79.229 8	1.790 2	85.757 5	4.737 5
6	78.88	77.702 4	1.177 6	79.547 8	0.667 8	80.209 5	1.329 5
7	89.90	87.943 2	1.956 8	89.116 3	0.783 7	90.230 6	0.330 6
8	85.83	88.259 7	2.429 7	86.305 4	0.475 4	88.629 8	2.799 8
9	91.29	88.259 7	3.030 3	86.305 4	4.984 6	88.629 8	2.660 2
10	77.47	81.749 9	4.279 9	79.337 2	1.867 2	80.084 9	2.614 9
11	91.63	88.011 9	3.618 1	88.622 6	3.007 4	90.704 6	0.925 4
12	88.17	88.619 3	0.449 3	86.600 5	1.569 5	89.649 2	1.479 2
13	88.81	88.349 7	0.460 3	85.620 3	3.189 7	88.014 9	0.795 1
14	87.57	82.818 7	4.751 3	86.269 0	1.301 0	82.425 0	5.145 0
15	82.33	77.342 4	4.987 6	85.298 4	2.968 4	83.573 5	1.243 5
16	87.01	86.629 4	0.380 6	86.856 1	0.153 9	87.300 5	0.290 5
17	88.11	82.447 3	5.662 7	89.289 8	1.179 8	92.079 7	3.969 7
18	84.48	82.606 1	1.873 9	92.242 3	7.762 3	81.772 6	2.707 4
平均误差			3.010 9		2.688 4		2.118 9
误差百分率/%			3.51		3.13		2.47