Prediction of coke quality based on improved WOA-LSTM

doi:10.11949/0438-1157.20211351

Abstract

Abstract:

In the context of “double carbon”, improving coke quality is one of the key points to ensure the high quality development of the steel industry. The coking industry has the problem of on-line real-time monitoring and the generalization ability of the coke quality prediction model is relatively poor. An adaptive global search algorithm, namely improved whale optimization algorithm (WOA) and long short-term memory (LSTM) recurrent neural network integrated modeling method is proposed to solve this problem. We select the measurable parameters that can reflect the coke quality in the blended coal, and use principal component analysis (PCA) to remove the redundancy factors with small variability to obtain the prediction factors as the external input of LSTM network; add adaptive inertia weight and optimal disturbance update to improve WOA, so as to train the super parameters of LSTM network, and use root mean square error (RMSE) and R-squared to test the algorithm. The improved AGWOA-LSTM model is compared with the LSTM model and WOA-LSTM model to verify the superiority of the method. The results show that the AGWOA-LSTM model has high accuracy and fast operation speed, and has a guiding significance for coke production.

Key words: whale optimization algorithm, coke quality, prediction model, neural network, principal component analysis

摘要：

“双碳”背景下，提升焦炭质量是保证钢铁行业高质量发展的研究重点之一，而炼焦行业存在着在线实时监测难、焦炭质量预测模型泛化能力差等问题。为此，提出一种通过自适应全局搜索算法，即改进鲸鱼优化算法（WOA）与长短期记忆（LSTM）循环神经网络综合建模的方法来解决这一问题。首先选取出配合煤中可反映焦炭质量的可测参数，再运用主成分分析（PCA）去除变异性小的冗余因子后，得到预测因子，将其作为LSTM网络的外部输入；通过加入自适应惯性权重以及最佳扰动更新改进WOA，从而训练LSTM网络的超参数，采用均方根误差（RMSE）和R-squared 进行算法检验；最后将改进后的AGWOA-LSTM模型与典型的LSTM、WOA-LSTM模型进行对比，以验证本方法的优越性。结果表明AGWOA-LSTM模型预测焦炭质量具有精度高、运行速度快等特点。研究对焦炭生产具有一定的理论指导意义。

关键词: 鲸鱼优化算法, 焦炭质量, 预测模型, 神经网络, 主元分析

CLC Number:

TQ 533.9

Libang LIU, Song YANG, Zhijian WANG, Xinxin HE, Wenlei ZHAO, Shoujun LIU, Wenguang DU, Jie MI. Prediction of coke quality based on improved WOA-LSTM[J]. CIESC Journal, 2022, 73(3): 1291-1299.

刘立邦, 杨颂, 王志坚, 贺欣欣, 赵文磊, 刘守军, 杜文广, 米杰. 基于改进WOA-LSTM的焦炭质量预测[J]. 化工学报, 2022, 73(3): 1291-1299.

Figures/Tables 11

References 36

1	刘君贤. 焦炉生产过程焦炭质量与炼焦能耗智能预测模型[D]. 长沙: 中南大学, 2010.
	Liu J X. Intelligent prediction model of coke quality and energy consumption in coke oven production process [D]. Changsha: Central South University, 2010.
2	田英奇. 配煤炼焦试验优化与神经网络焦炭质量预测模型的研究[D]. 上海: 华东理工大学, 2016.
	Tian Y Q. Research on the optimization of coal-blending coking experiment and neural network coke quality prediction model[D]. Shanghai: East China University of Science and Technology, 2016.
3	Malyi E I. Modification of poorly clinkering coal for use in coking[J]. Coke and Chemistry, 2014, 57(3): 87-90.
4	高志芳, 朱书全, 崔荣健. 线性模型预测焦炭灰分[J]. 矿业快报, 2005, 21(10): 28-29.
	Gao Z F, Zhu S Q, Cui R J. Linear model predicting ash content in coke[J]. Express Information of Mining Industry, 2005, 21(10): 28-29.
5	张进春, 吴超. 基于多重多元回归的焦炭质量预测模型[J]. 科技导报, 2010, 28(12): 79-84.
	Zhang J C, Wu C. Coke quality prediction model based on multivariate regression[J]. Science & Technology Review, 2010, 28(12): 79-84..
6	谢海深, 刘永新, 吕庆, 等. 焦炭质量预测模型[J]. 东北大学学报(自然科学版), 2007, 28(3): 373-377.
	Xie H S, Liu Y X, Lü Q, et al. Coke quality prediction models[J]. Journal of Northeastern University (Natural Science), 2007, 28(3): 373-377.
7	赵树果, 单晓云, 陈少敏, 等. 人工神经网络在炼焦配煤质量预测模型中的应用[J]. 中国煤炭, 2006, 32(7): 52-54.
	Zhao S G, Shan X Y, Chen S M, et al. Application of artificial neural network in prediction model of coking coal blending qualityl[J]. China Coal, 2006, 32(7): 52-54.
8	周洪, 闵礼书, 邹祥林. 基于神经网络的特大型焦炉焦炭质量预测研究[J]. 系统仿真学报, 2009, 21(6): 1543-1547, 1552.
	Zhou H, Min L S, Zou X L. Research on prediction of blended coal for coking quality in oversized coke furnace based on neural network[J]. Journal of System Simulation, 2009, 21(6): 1543-1547, 1552.
9	冯立颖. 改进的BP神经网络算法及其应用[J]. 计算机仿真, 2010, 27(12): 172-175.
	Feng L Y. Optimized BP neural networks algorithm and its application[J]. Computer Simulation, 2010, 27(12): 172-175.
10	Zhou H, Yang C J, Liu W H, et al. A sliding-window T-S fuzzy neural network model for prediction of silicon content in hot metal[J]. IFAC-PapersOnLine, 2017, 50(1): 14988-14991.
11	Lemme A, Reinhart R F, Steil J J. Online learning and generalization of parts-based image representations by non-negative sparse autoencoders[J]. Neural Networks, 2012, 33: 194-203.
12	Wang X Z, Shao Q Y, Miao Q, et al. Architecture selection for networks trained with extreme learning machine using localized generalization error model[J]. Neurocomputing, 2013, 102: 3-9.
13	Hinton G E, Osindero S, Teh Y W. A fast learning algorithm for deep belief nets[J]. Neural Computation, 2006, 18(7): 1527-1554.
14	Tang Y J, Xu J F, Matsumoto K, et al. Sequence-to-sequence model with attention for time series classification[C]//2016 IEEE 16th International Conference on Data Mining Workshops (ICDMW). Barcelona, Spain:IEEE, 2016: 503-510.
15	Pinheiro P, Collobert R, Jebarat, et al. Recurrent convolutional neural networks for scene labeling[C]//Proceedings of the 31st International Conference on Machine Learning (ICML-14), 2014: 82-90.
16	Mirjalili S, Lewis A. The whale optimization algorithm[J]. Advances in Engineering Software, 2016, 95: 51-67.
17	Liu J W, Chi G H, Liu Z Y, et al. Predicting protein structural classes with autoencoder neural networks[C]//2013 25th Chinese Control and Decision Conference (CCDC). Guiyang, China: IEEE, 2013: 1894-1899.
18	Yang Z, Zeng Z, Wang K, et al. Modified SEIR and AI prediction of the epidemics trend of COVID-19 in China under public health interventions[J]. Journal of Thoracic Disease, 2020, 12(3): 165-174.
19	Anang, Hadisupadmo S, Leksono E. Model predictive control design and performance analysis of a pasteurization process plant[C]//2016 International Conference on Instrumentation, Control and Automation (ICA). Bandung, Indonesia: IEEE, 2016: 81-87.
20	Yin S, Li X W, Gao H J, et al. Data-based techniques focused on modern industry: an overview[J]. IEEE Transactions on Industrial Electronics, 2015, 62(1): 657-667.
21	Achanta S, Godambe T, Gangashetty S V. An investigation of recurrent neural network architectures for statistical parametric speech synthesis[C]//Interspeech 2015. ISCA, 2015: 859-863.
22	耿志强, 曾荣甫, 徐圆, 等. 融合灰狼优化算法在工控系统入侵检测中的应用[J]. 化工学报, 2020, 71(3): 1080-1087.
	Geng Z Q, Zeng R F, Xu Y, et al. Intrusion detection of industrial control system based on grey wolf optimization integrated random black hole[J]. CIESC Journal, 2020, 71(3): 1080-1087.
23	Zhang J Q, Sanderson A C. JADE: adaptive differential evolution with optional external archive[J]. IEEE Transactions on Evolutionary Computation, 2009, 13(5): 945-958.
24	庞晓琼, 王竹晴, 曾建潮, 等. 基于PCA-NARX的锂离子电池剩余使用寿命预测[J]. 北京理工大学学报, 2019, 39(4): 406-412.
	Pang X Q, Wang Z Q, Zeng J C, et al. Prediction for the remaining useful life of lithium-ion battery based on PCA-NARX[J]. Transactions of Beijing Institute of Technology, 2019, 39(4): 406-412.
25	安剑奇, 陈易斐, 吴敏. 基于改进支持向量机的高炉一氧化碳利用率预测方法[J]. 化工学报, 2015, 66(1): 206-214.
	An J Q, Chen Y F, Wu M. A prediction method for carbon monoxide utilization ratio of blast furnace based on improved support vector regression[J]. CIESC Journal, 2015, 66(1): 206-214.
26	石博文, 尹燕燕, 刘飞. 基于PSO-控制变量参数化混合策略的间歇化工过程优化控制[J]. 化工学报, 2019, 70(3): 979-986.
	Shi B W, Yin Y Y, Liu F. Optimal control strategies combined with PSO and control vector parameterization for batchwise chemical process[J]. CIESC Journal, 2019, 70(3): 979-986.
27	焦李成, 杨淑媛, 刘芳, 等. 神经网络七十年: 回顾与展望[J]. 计算机学报, 2016, 39(8): 1697-1716.
	Jiao L C, Yang S Y, Liu F, et al. Seventy years beyond neural networks: retrospect and prospect[J]. Chinese Journal of Computers, 2016, 39(8): 1697-1716.
28	Wang Z J, Yang N N, Li N P, et al. A new fault diagnosis method based on adaptive spectrum mode extraction[J]. Structural Health Monitoring, 2021, 20(6): 3354-3370.
29	Pang Z H, Liu G P, Zhou D H, et al. Data-based predictive control for networked nonlinear systems with network-induced delay and packet dropout[J]. IEEE Transactions on Industrial Electronics, 2016, 63(2): 1249-1257.
30	Wang T, Gao H J, Qiu J B. A combined adaptive neural network and nonlinear model predictive control for multirate networked industrial process control[J]. IEEE Transactions on Neural Networks and Learning Systems, 2016, 27(2): 416-425.
31	Vazquez S, Rodriguez J, Rivera M, et al. Model predictive control for power converters and drives: advances and trends[J]. IEEE Transactions on Industrial Electronics, 2017, 64(2): 935-947.
32	Wang Z J, He X X, Yang B, et al. Subdomain adaptation transfer learning network for fault diagnosis of roller bearings[J]. IEEE Transactions on Industrial Electronics, 2021, doi:10.1109/TIE.2021.3108726 . DOI
33	雷琪, 刘君贤, 何勇, 等. 基于PCA与RBF的焦炭质量预测模型[J]. 控制工程, 2010, 17(4): 513-516, 520.
	Lei Q, Liu J X, He Y, et al. A prediction model for coke quality based on PCA and RBF[J]. Control Engineering of China, 2010, 17(4): 513-516, 520.
34	Bazhin V Y, Titov O V. Analysis of the coking process on the basis of heat-flux calorimetry[J]. Coke and Chemistry, 2015, 58(7): 254-258.
35	Klenske E D, Zeilinger M N, Schölkopf B, et al. Gaussian process-based predictive control for periodic error correction[J]. IEEE Transactions on Control Systems Technology, 2016, 24(1): 110-121.
36	Cheng L, Liu W C, Hou Z G, et al. Neural-network-based nonlinear model predictive control for piezoelectric actuators[J]. IEEE Transactions on Industrial Electronics, 2015, 62(12): 7717-7727.

焦炭						入炉煤
灰分/% （质量）	全硫/% （质量）	水分/% （质量）	M₂₅	挥发分/% （质量）	M₁₀	灰分/% （质量）	全硫/% （质量）	水分/% （质量）	黏结指数	挥发分/% （质量）	细度/% （质量）
12.71	0.80	7.90	90.00	1.04	7.60	9.88	1.03	12.95	51.50	25.59	88.40
12.90	0.86	11.40	88.50	1.14	10.00	9.69	1.01	11.45	50.50	23.34	88.40
12.76	0.67	9.55	86.30	1.04	11.70	10.23	0.89	10.50	52.00	25.42	82.65
14.15	0.79	4.45	91.40	1.35	6.40	10.19	0.84	10.30	53.00	24.39	82.85
13.75	0.56	7.25	90.90	1.29	7.60	10.80	0.75	11.65	56.00	26.03	80.85
14.25	0.59	10.85	87.60	1.16	10.40	10.58	0.75	10.55	65.00	26.76	81.05
13.43	0.62	7.30	90.40	1.23	7.10	10.33	0.76	9.35	65.50	27.23	79.75
13.73	0.62	7.20	86.70	1.22	12.10	10.08	0.73	10.00	57.00	25.78	84.65
13.14	0.63	7.25	88.50	1.06	9.50	10.45	0.79	9.50	55.50	26.16	81.05
13.87	0.52	9.20	88.70	1.29	8.60	10.12	0.71	10.40	54.50	25.98	80.80
13.40	0.64	4.50	90.30	1.08	7.80	10.07	0.80	9.60	54.00	24.85	83.05

焦炭						入炉煤
灰分/% （质量）	全硫/% （质量）	水分/% （质量）	M₂₅	挥发分/% （质量）	M₁₀	灰分/% （质量）	全硫/% （质量）	水分/% （质量）	黏结指数	挥发分/% （质量）	细度/% （质量）
12.71	0.80	7.90	90.00	1.04	7.60	9.88	1.03	12.95	51.50	25.59	88.40
12.90	0.86	11.40	88.50	1.14	10.00	9.69	1.01	11.45	50.50	23.34	88.40
12.76	0.67	9.55	86.30	1.04	11.70	10.23	0.89	10.50	52.00	25.42	82.65
14.15	0.79	4.45	91.40	1.35	6.40	10.19	0.84	10.30	53.00	24.39	82.85
13.75	0.56	7.25	90.90	1.29	7.60	10.80	0.75	11.65	56.00	26.03	80.85
14.25	0.59	10.85	87.60	1.16	10.40	10.58	0.75	10.55	65.00	26.76	81.05
13.43	0.62	7.30	90.40	1.23	7.10	10.33	0.76	9.35	65.50	27.23	79.75
13.73	0.62	7.20	86.70	1.22	12.10	10.08	0.73	10.00	57.00	25.78	84.65
13.14	0.63	7.25	88.50	1.06	9.50	10.45	0.79	9.50	55.50	26.16	81.05
13.87	0.52	9.20	88.70	1.29	8.60	10.12	0.71	10.40	54.50	25.98	80.80
13.40	0.64	4.50	90.30	1.08	7.80	10.07	0.80	9.60	54.00	24.85	83.05

成分	贡献率/%
黏结指数	61.2206
硫分	18.3321
灰分	13.4472
挥发分	4.1202
水分	2.7072
细度	0.1772

成分	贡献率/%
黏结指数	61.2206
硫分	18.3321
灰分	13.4472
挥发分	4.1202
水分	2.7072
细度	0.1772

Model	Length of training/s	RMSE	R-squared
LSTM	738	5.03×10^-3	0.85
WOA-LSTM	508	4.18×10^-3	0.91
AGWOA-LSTM	366	2.09×10^-3	0.98