基于选择性集成核学习算法的固废焚烧过程二噁英排放浓度软测量

doi:10.11949/j.issn.0438-1157.20181354

摘要/Abstract

摘要：

城市固废焚烧(MSWI)过程排放的二噁英 (DXN)是被称为“世纪之毒”的持续性污染物。该过程的多阶段、多温度区间的物理化学特性导致DXN排放浓度的机理模型难以构建。工业实际中通常以月或季为周期耗时近1周时间在实验室以离线化验方式滞后检测。针对这些问题，提出了基于选择性集成(SEN)核学习算法的DXN排放浓度软测量方法。首先，基于先验知识给出候选核参数集和候选惩罚参数集，采用核学习算法构建基于这些超参数的候选子子模型；然后，耦合优化和加权算法对相同核参数的候选子子模型进行选择与合并，进而得到基于不同核参数的候选SEN子模型集合；最后，再次采用优化和加权算法获得结构与超参数自适应的多层SEN软测量模型。采用UCI平台水泥抗压强度和焚烧过程DXN数据验证了所提方法的有效性。

关键词: 城市固废焚烧, 过程系统, 二噁英, 参数估值, 选择性集成, 废物处理

Abstract:

Dioxin (DXN) emitted from the municipal solid waste incineration (MSWI) process is a persistent pollutant of the “century poison”. DXN is one of the highly toxic and persistent pollution. The principal model of DXN emission is difficult to obtained duo to the complex multi-stage and multi-temperature phase’s physical chemical characteristics. In practical, DXN emission concentration is off-line measured with month or quarter period by quantified national laboratory with long lag time delay. Aiming at these problems, a new DXN emission concentration soft measuring method based on selective ensemble (SEN) kernel learning algorithm is proposed. At first, candidate kernel parameters and regularization parameters are given based on prior knowledge. Then, candidate sub-sub-models based on these super parameters are constructed. Thirdly, coupled optimization and weighting algorithms are used to build SEN-sub-models. Finally, these SEN-sub-models are selective combined as final SEN model by using optimization and weighting algorithms again. Simulation results based on the concrete compression strength and incineration process DXN data validate effectiveness of the proposed approach.

Key words: municipal solid waste incineration, process systems, dioxin, parameter estimation, selective ensemble, waste treatment

中图分类号:

TQ 028.8

汤健, 乔俊飞. 基于选择性集成核学习算法的固废焚烧过程二噁英排放浓度软测量[J]. 化工学报, 2019, 70(2): 696-706.

Jian TANG, Junfei QIAO. Dioxin emission concentration soft measuring approach of municipal solid waste incineration based on selective ensemble kernel learning algorithm[J]. CIESC Journal, 2019, 70(2): 696-706.

图/表 13

参考文献 31

1	Korai M S , Mahar R B , Uqaili M A . The feasibility of municipal solid waste for energy generation and its existing management practices in Pakistan[J]. Renew. Sustain. Energy Rev., 2017, 72: 338-353.
2	Zhou H , Meng A , Long Y Q . A review of dioxin-related substances during municipal solid waste incineration[J]. Waste Management, 2015, 36: 106-118.
3	Lu J W , Zhang S , Hai J , et al . Status and perspectives of municipal solid waste incineration in China: a comparison with developed regions[J]. Waste Management, 2017, 69: 170-186.
4	Zhou Z , Zhao B , Kojima H , et al . Simple and rapid determination of PCDD/Fs in flue gases from various waste incinerators in China using DR-EcoScreen cells[J]. Chemosphere, 2014, 102: 24-30.
5	Koloma H , Takeuchi S , Iida M , et al . A sensitive, rapid, and simple DR-EcoScreen bioassay for the determination of PCDD/Fs and dioxin-like PCBs in environmental and food samples[J]. Environmental Science & Pollution Research, 2015, 22: 1-12.
6	Stamore B R . Modeling the formation of PCDD/F in solid waste incinerators[J]. Chemosphere, 2002, 47: 565-773.
7	Bunsan S W , Chen Y , Chen H W , et al . Modeling the dioxin emission of a municipal solid waste incinerator using neural networks[J]. Chemosphere, 2013, 92: 258-264.
8	Gullett B K , Oudejans L , Tabor D ．Near-realtime combustion monitoring for PCDD/PCDF indicators by GC-ＲEMPI-TOFMS[J]. Environmental Science ＆ Technology, 2012, 46 (2): 923-928.
9	郭颖, 陈彤, 杨杰, 等 . 基于关联模型的二口恶英在线检测研究[J]. 环境工程学报, 2014, 8(8): 3524-3529.
	Guo Y , Chen T , Yang J , et al . Study on on-line detection of dioxins based on correlation model[J]. Chinese Journal of Environmental Engineering, 2014, 8(8): 3524-3529.
10	李阿丹, 洪伟, 王晶 . 激光解吸/激光电离-质谱法二口恶英及其关联物的在线检测[J]. 燕山大学学报, 2015, 39(6): 511-515.
	Li A D , Hong W , Wang J . Online detection of dioxin and dioxin-related substances using laser desoption/laser ionization-mass spectrometry[J]. Journal of Yanshan University, 2015, 39(6): 511-515.
11	Pandelova M , Lenoir D , Schramm K W , Correlation between PCDD/F, PCB and PCBz in coal/waste combustion . Influence of various inhibitors[J]. Chemosphere, 2006, 62: 1196-1205.
12	Gullett B K , Oudejans L , Tabor D , et al . Near-real-time combustion monitoring for PCDD/PCDF indicators by GC–REMPI–TOFMS[J]. Environmental Engineering Science, 2012, 46: 923-928.
13	Guo Z H , Tang J , Qiao J F . Mathematic simulation model of dioxin emission concentration of municipal solid waste incineration based on Aspen-plus software[C]//The 37th Chinese Control Conference. USA: IEEE Press, 2018.
14	Wang W , Chai T Y , Yu W , Modeling component concentrations of sodium aluminate solution via hammerstein recurrent neural networks [J]. IEEE Transactions on Control Systems Technology, 2012, 20: 971-982.
15	Tang J . Chai T Y, Yu W, et al . Modeling load parameters of ball mill in grinding process based on selective ensemble multisensor information[J]. IEEE Transactions on Automation Science & Engineering, 2013, 10: 726-740.
16	Kano M , Fujwara K . Virtual sensing technology in process industries: trends & challenges revealed by recent industrial applications[J]. Journal of Chemical Engineering of Japan, 2013, 46: 1-17.
17	Tang J , Liu Z , Zhang J , et al . Kernel latent feature adaptive extraction and selection method for multi-component non-stationary signal of industrial mechanical device[J]. Neurocomputing, 2016, 216(C): 296-309.
18	Tang J , Chai T Y , Yu W , et al . Feature extraction and selection based on vibration spectrum with application to estimate the load parameters of ball mill in grinding process[J]. Control Engineering Practice, 2012, 20(10): 991-1004.
19	Zhou Z H , Wu J , Tang W . Ensembling neural networks: many could be better than all[J]. Artificial Intelligence, 2002, 137(1/2): 239-263.
20	汤健，柴天佑，丛秋梅, 等 . 选择性融合多尺度筒体振动频谱的磨机负荷参数建模[J]. 控制理论与应用，2015, 32(12): 1582-1591.
	Tang J , Chai T Y , Cong Q M , et al . Modeling mill load parameters based on selective fusion of multi-scale shell vibration frequency spectrum[J]. Control Theory & Application, 2015, 32(12): 1582-1591.
21	汤健，柴天佑，丛秋梅，等 . 基于EMD和选择性集成学习算法的磨机负荷参数软测量[J]. 自动化学报, 2014, 40(9): 1853-1866.
	Tang J , Chai T Y , Cong Q M , et al . Soft sensor approach for modeling mill load parameters based on EMD and selective ensemble learning algorithm[J]. Acta Automatica Sinica, 2014, 40(9): 1853-1866.
22	Tang J , Yu W , Chai T Y , et al . Selective ensemble modeling load parameters of ball mill based on multi-scale frequency spectral features and sphere criterion[J]. Mechanical Systems & Signal Processing, 2016, 66/67: 485-504.
23	Tang J , Qiao J F , Wu Z W , et al . Vibration and acoustic frequency spectra for industrial process modeling using selective fusion multi-condition samples and multi-source features[J]. Mechanical Systems and Signal Processing, 2018, 99: 142-168.
24	Tang J , Zhang J , Wu Z W , et al . Modeling collinear data using double-layer GA-based selective ensemble kernel partial least squares algorithm [J]. Neurocomputing, 2017, 219: 248-262.
25	Lv Y , Liu J , Yang T . A novel least squares support vector machine ensemble model for NO _x , emission prediction of a coal-fired boiler[J]. Energy, 2013, 55(1): 319-329.
26	Padilha C A D A , Barone D A C , Neto A D D . A multi-level approach using genetic algorithms in an ensemble of least squares support vector machines[J]. Knowledge-Based Systems, 2016, 106(C): 85-95.
27	汤健，乔俊飞，柴天佑, 等 . 基于虚拟样本生成技术的多组分机械信号建模[J]. 自动化学报, 2018, 44(9): 1569-1590.
	Tang J , Qiao J F , Chai T Y , et al . Modeling multiple components mechanical signals by means of virtual sample generation technique[J]. Acta Automatica Sinica, 2018, 44(9): 1569-1590.
28	Tang J , Qiao J F , Liu Z , et al . Mechanism characteristic analysis and soft measuring method review for ball mill load based on mechanical vibration and acoustic signals in the grinding process[J]. Minerals Engineering, 2018, 128: 294-311.
29	Yeh I C . Modeling of strength of high performance concrete using artificial neural networks [J]. Cement and Concrete Research, 1998, 28(12): 1797-1808.
30	Tang J , Yu W , Chai T Y , et al . On-line principal component analysis with application to process modeling[J]. Neurocomputing, 2012, 82(1): 167-178.
31	Chang N B , Huang S H . Statistical modelling for the prediction and control of PCDDs and PCDFs emissions from municipal solid waste incinerators[J]. Waste Management & Research, 1995, 13(4): 379-400.

编码序号	核参数	模型类型	预测误差 (RMSE)	惩罚参数（编码序号）
1	0.1	EnAll-sub-sub	16.1059	all
		SEN-sub	15.0883	15～17
		Best-Sub-sub	15.0883	17
2	1	EnAll-sub-sub	12.8921	all
		SEN-sub	12.4234	2～17
		Best-Sub-sub	11.8064	6
3	100	EnAll-sub-sub	9.8368	all
		SEN-sub	7.3687	4～17
		Best-Sub-sub	7.7284	10
4	1000	EnAll-sub-sub	15.9703	all
		SEN-sub	8.1098	11～17
		Best-Sub-sub	8.1362	13
5	2000	EnAll-sub-sub	16.4346	all
		SEN-sub	8.1475	14～17
		Best-Sub-sub	8.1459	15
6	4000	EnAll-sub-sub	16.6256	all
		SEN-sub	8.1642	16～17
		Best-Sub-sub	8.1511	17
7	6000	EnAll-sub-sub	16.6813	all
		SEN-sub	8.3123	16～17
		Best-Sub-sub	8.2182	17
8	8000	EnAll-sub-sub	16.7074	all
		SEN-sub	8.5411	16～17
		Best-Sub-sub	8.3811	17
9	10000	EnAll-sub-sub	16.7225	all
		SEN-sub	8.7862	16～17
		Best-Sub-sub	8.5831	17
10	20000	EnAll-sub-sub	16.7514	all
		SEN-sub	9.7130	16～17
		Best-Sub-sub	9.4900	17
11	40000	EnAll-sub-sub	16.7652	all
		SEN-sub	10.4129	16～17
		Best-Sub-sub	10.2943	17
12	60000	EnAll-sub-sub	16.7697	all
		SEN-sub	10.6267	16～17
		Best-Sub-sub	10.5610	17
13	80000	EnAll-sub-sub	16.7719	all
		SEN-sub	10.7145	16～17
		Best-Sub-sub	10.6731	17
14	16000	EnAll-sub-sub	16.7752	all
		SEN-sub	10.8124	16～17
		Best-Sub-sub	10.7968	17

编码序号	核参数	模型类型	预测误差 (RMSE)	惩罚参数（编码序号）
1	0.1	EnAll-sub-sub	16.1059	all
		SEN-sub	15.0883	15～17
		Best-Sub-sub	15.0883	17
2	1	EnAll-sub-sub	12.8921	all
		SEN-sub	12.4234	2～17
		Best-Sub-sub	11.8064	6
3	100	EnAll-sub-sub	9.8368	all
		SEN-sub	7.3687	4～17
		Best-Sub-sub	7.7284	10
4	1000	EnAll-sub-sub	15.9703	all
		SEN-sub	8.1098	11～17
		Best-Sub-sub	8.1362	13
5	2000	EnAll-sub-sub	16.4346	all
		SEN-sub	8.1475	14～17
		Best-Sub-sub	8.1459	15
6	4000	EnAll-sub-sub	16.6256	all
		SEN-sub	8.1642	16～17
		Best-Sub-sub	8.1511	17
7	6000	EnAll-sub-sub	16.6813	all
		SEN-sub	8.3123	16～17
		Best-Sub-sub	8.2182	17
8	8000	EnAll-sub-sub	16.7074	all
		SEN-sub	8.5411	16～17
		Best-Sub-sub	8.3811	17
9	10000	EnAll-sub-sub	16.7225	all
		SEN-sub	8.7862	16～17
		Best-Sub-sub	8.5831	17
10	20000	EnAll-sub-sub	16.7514	all
		SEN-sub	9.7130	16～17
		Best-Sub-sub	9.4900	17
11	40000	EnAll-sub-sub	16.7652	all
		SEN-sub	10.4129	16～17
		Best-Sub-sub	10.2943	17
12	60000	EnAll-sub-sub	16.7697	all
		SEN-sub	10.6267	16～17
		Best-Sub-sub	10.5610	17
13	80000	EnAll-sub-sub	16.7719	all
		SEN-sub	10.7145	16～17
		Best-Sub-sub	10.6731	17
14	16000	EnAll-sub-sub	16.7752	all
		SEN-sub	10.8124	16～17
		Best-Sub-sub	10.7968	17

集成尺寸	集成子模型(核参数)的编码序号	RMSE
2	2,1	12.8102
3	3,2,1	9.9116
4	4,3,2,1	8.9488
5	5,4,3,2,1	8.5528
6	6,5,4,3,2,	8.3599
7	7,6,5,4,3,2,1	8.2641
8	8,7,6,5,4,3,2,1	8.2247
9	9,8,7,6,5,4,3,2,1	8.2221
10	10,9,8,7,6,5,4,3,2,1	8.2942
11	11,10,9,8,7,6,5,4,3,2,1	8.4228
12	12, 11, 10,9,8,7,6,5,4,3,2,1	8.5617
13	13, 11, 10,9,8,7,6,5,4,3,2,1	8.6963

集成尺寸	集成子模型(核参数)的编码序号	RMSE
2	2,1	12.8102
3	3,2,1	9.9116
4	4,3,2,1	8.9488
5	5,4,3,2,1	8.5528
6	6,5,4,3,2,	8.3599
7	7,6,5,4,3,2,1	8.2641
8	8,7,6,5,4,3,2,1	8.2247
9	9,8,7,6,5,4,3,2,1	8.2221
10	10,9,8,7,6,5,4,3,2,1	8.2942
11	11,10,9,8,7,6,5,4,3,2,1	8.4228
12	12, 11, 10,9,8,7,6,5,4,3,2,1	8.5617
13	13, 11, 10,9,8,7,6,5,4,3,2,1	8.6963

方法	预测误差（RMSE）			备注 (学习参数)
方法	最大值	均值	最小值	备注 (学习参数)
PLS	—	10.92	—	LV=7
KPLS	—	8.179	—	LV=8
GASEN-BPNN	14.8580	12.0756	10.3971	隐含节点=17
GASEN-LSSVM	10.1777	10.0149	9.7490	(100,800)
本文方法	—	7.163	—