动态模糊PLS法实现废水处理出水指标预测
|
摘要
针对废水处理过程普遍存在的时变性和非线性特征,提出动态模糊偏最小二乘法(DFPLS)实现废水出水指标预测。分别采用线性偏最小二乘(LPLS)、模糊偏最小二乘(FPLS)和DFPLS方法对比分析。结果表明:DFPLS方法预测均方误差相较于LPLS和FPLS分别下降了88.61%和77.50%;DFPLS在第3潜变量下的输出累计方差贡献率相较于FPLS提升了38.51%,显著提高了废水处理过程预测的准确性,验证了该方法的有效性。
In view of the time-varying and non-linear characteristics of wastewater treatment process, constructing a dynamic fuzzy partial least squares(DFPLS) model for wastewater treatment process was proposed to predict index of the effluent. Applying the linear partial least squares(LPLS), fuzzy partial least squares(FPLS) and DFPLS methods to comparative analysis shows that, compared with LPLS and FPLS methods, the predictive mean square error(MSE) of DFPLS method can decrease by 88.61% and 77.50% respectively; and the cumulative variance contribution rate of DFPLS method can be increased by 38.51% compared with FPLS method under the third principal component and the accuracy of predicting the wastewater treatment process can be increased obviously along with a verified validity of the method
In view of the time-varying and non-linear characteristics of wastewater treatment process, constructing a dynamic fuzzy partial least squares(DFPLS) model for wastewater treatment process was proposed to predict index of the effluent. Applying the linear partial least squares(LPLS), fuzzy partial least squares(FPLS) and DFPLS methods to comparative analysis shows that, compared with LPLS and FPLS methods, the predictive mean square error(MSE) of DFPLS method can decrease by 88.61% and 77.50% respectively; and the cumulative variance contribution rate of DFPLS method can be increased by 38.51% compared with FPLS method under the third principal component and the accuracy of predicting the wastewater treatment process can be increased obviously along with a verified validity of the method
引文
[1] 彭开香,马亮,张凯.复杂工业过程质量相关的故障检测与诊断技术综述[J].自动化学报,2017,43(3):349～365.
[2] 苏鑫,吴迎亚,裴华健,等.大数据技术在过程工业中的应用研究进展[J].化工进展,2016,35(6):1652～1659.
[3] Dong Y,Qin S J.Regression on Dynamic PLS Structures for Supervised Learning of Dynamic Data[J].Journal of Process Control,2018,68:64～72.
[4] Du H,Zhou G,Ge H,et al.Satellite-Based Carbon Stock Estimation for Bamboo Forest with a Non-Linear Partial Least Square Regression Technique[J].International Journal of Remote Sensing,2012,33(6):1917～1933.
[5] Shan P,Peng S,Tang L,et al.A Nonlinear Partial Least Squares with Slice Transform Based Piecewise Linear Inner Relation[J].Chemometrics & Intelligent Laboratory Systems,2015,143:97～110.
[6] 夏巧生.非线性偏最小二乘建模方法及在近红外光谱建模上的应用[J].计算机与应用化学,2014,31(1):109～112.
[7] 刘强,秦泗钊.过程工业大数据建模研究展望[J].自动化学报,2016,42(2):161～171.
[8] Bang Y H,Chang K Y,Leeb I B.Nonlinear PLS Modeling with Fuzzy Inference System[J].Chemometrics & Intelligent Laboratory Systems,2002,64(2):137～155.
[9] Lughofer E,Pollak R,Zavoianu A C,et al.Self-Adaptive Evolving Forecast Models with Incremental PLS Spaceupdating for On-Line Prediction of Micro-Fluidic Chip Quality[J].Engineering Applications of Artificial Intelligence,2018,68:131～151.
[10] Parastar H,Bazrafshan A.Fuzzy C-means Clustering for Chromatographic Fingerprints Analysis:A Gas Chromatography-Mass Spectrometry Case Study[J].Journal of Chromatography A,2016,1438:236～243.
[11] Yordanova S.An Approach to Observability and Controllability Analysis of Nonlinear Plants on the Basis of TSK Models[J].Information Technologies & Control,2015,13(1-2):35～45.
[12] Liu H,Huang M,Yoo C.A Fuzzy Neural Network-Based Soft Sensor for Modeling Nutrient Removal Mechanism in a Full-Scale Wastewater Treatment System[J].Desalination & Water Treatment,2013,51(31-33):6184～6193.
[2] 苏鑫,吴迎亚,裴华健,等.大数据技术在过程工业中的应用研究进展[J].化工进展,2016,35(6):1652～1659.
[3] Dong Y,Qin S J.Regression on Dynamic PLS Structures for Supervised Learning of Dynamic Data[J].Journal of Process Control,2018,68:64～72.
[4] Du H,Zhou G,Ge H,et al.Satellite-Based Carbon Stock Estimation for Bamboo Forest with a Non-Linear Partial Least Square Regression Technique[J].International Journal of Remote Sensing,2012,33(6):1917～1933.
[5] Shan P,Peng S,Tang L,et al.A Nonlinear Partial Least Squares with Slice Transform Based Piecewise Linear Inner Relation[J].Chemometrics & Intelligent Laboratory Systems,2015,143:97～110.
[6] 夏巧生.非线性偏最小二乘建模方法及在近红外光谱建模上的应用[J].计算机与应用化学,2014,31(1):109～112.
[7] 刘强,秦泗钊.过程工业大数据建模研究展望[J].自动化学报,2016,42(2):161～171.
[8] Bang Y H,Chang K Y,Leeb I B.Nonlinear PLS Modeling with Fuzzy Inference System[J].Chemometrics & Intelligent Laboratory Systems,2002,64(2):137～155.
[9] Lughofer E,Pollak R,Zavoianu A C,et al.Self-Adaptive Evolving Forecast Models with Incremental PLS Spaceupdating for On-Line Prediction of Micro-Fluidic Chip Quality[J].Engineering Applications of Artificial Intelligence,2018,68:131～151.
[10] Parastar H,Bazrafshan A.Fuzzy C-means Clustering for Chromatographic Fingerprints Analysis:A Gas Chromatography-Mass Spectrometry Case Study[J].Journal of Chromatography A,2016,1438:236～243.
[11] Yordanova S.An Approach to Observability and Controllability Analysis of Nonlinear Plants on the Basis of TSK Models[J].Information Technologies & Control,2015,13(1-2):35～45.
[12] Liu H,Huang M,Yoo C.A Fuzzy Neural Network-Based Soft Sensor for Modeling Nutrient Removal Mechanism in a Full-Scale Wastewater Treatment System[J].Desalination & Water Treatment,2013,51(31-33):6184～6193.