基于ISSA−LSTM的浓缩池溢流浓度预测

张洋洋; 樊玉萍; 马晓敏; 董宪姝; 金伟; 王大卫

doi:10.13272/j.issn.1671-251x.2022060084

基于ISSA−LSTM的浓缩池溢流浓度预测

张洋洋^1,,
樊玉萍^1, ,,
马晓敏^{1, 2},
董宪姝¹,
金伟¹,
王大卫³

1.
太原理工大学矿业工程学院, 山西太原　030024
2.
矿物加工科学与技术国家重点实验室, 北京　100160
3.
晋能控股集团马道头选煤厂, 山西太原　030006

基金项目: 国家自然科学基金重点国际(地区)合作研究项目（51820105006）；国家自然科学基金面上项目（52074189）；国家自然科学基金青年科学基金项目（52004178）。

详细信息

作者简介:
张洋洋（1997—）,男，安徽阜阳人，硕士研究生，研究方向为选煤智能化，E-mail：1278290544@qq.com

通讯作者:
樊玉萍（1988—），女，山西临汾人，副教授，硕士研究生导师，主要从事煤炭清洁高效利用方面的研究工作，E-mail：19880628fyp@163.com

中图分类号: TD94
计量
- 文章访问数: 325
- HTML全文浏览量: 76
- PDF下载量: 45
出版历程
- 收稿日期: 2022-06-21
- 修回日期: 2022-10-28
- 网络出版日期: 2022-08-08
- 刊出日期: 2022-11-24

Prediction of overflow concentration of thickener based on ISSA-LSTM

1.
College of Mining Engineering, Taiyuan University of Technology, Taiyuan 030024, China
2.
State Key Laboratory of Mineral Processing Science and Technology, Beijing 100160, China
3.
Madaotou Coal Preparation Plant, Jinneng Holding Group, Taiyuan 030006, China

摘要

摘要: 浓缩池溢流浓度监测是实现煤泥水智能加药的关键。针对基于传感器的溢流浓度监测方式会导致絮凝剂调节滞后的问题，提出了一种基于改进麻雀搜索算法（ISSA）−长短期记忆（LSTM）的浓缩池溢流浓度预测方法。首先，对浓缩生产过程中的多参数时间序列进行相关性分析和预处理，得到输入变量。其次，采用多策略联合改进麻雀搜索算法（SSA）：引入Tent混沌映射对麻雀种群进行初始化，以保证种群多样性，加快算法收敛速度；用螺旋捕食策略改进SSA的寻优过程，以兼顾局部开发和全局搜索能力；用萤火虫扰动策略对麻雀搜索结果进行扰动，以提高全局搜索能力，避免算法陷入局部最优。然后，采用ISSA优化双层LSTM网络模型的超参数。最后，构建基于ISSA−LSTM的浓缩池溢流浓度预测模型，进行在线监测。实验结果表明：① 选取Ackley函数和Rastrigin函数作为测试函数，得出ISSA的全局寻优性能和收敛速度均优于粒子群优化（PSO）算法、鲸鱼优化算法（WOA）和标准SSA。② 3种改进策略中，螺旋捕食策略对ISSA性能的提升起主导作用，混沌映射和萤火虫扰动策略协调算法的收敛速度和全局搜索能力，进一步提升算法寻优性能。③ 采用ISSA优化LSTM的超参数，解决了依靠主观经验取值时存在的欠拟合或过拟合问题，ISSA−LSTM模型的溢流浓度预测精度达97.26%，高于双层LSTM、SSA−LSTM、最小二乘支持向量机（LSSVM）等模型。④ 数据预处理可以提升模型的精度，降噪后溢流浓度预测精度比降噪前提升了30.25%。
- 选煤智能化 /
- 智能加药 /
- 溢流浓度预测 /
- 麻雀搜索算法 /
- LSTM /
- 混沌映射 /
- 螺旋捕食策略 /
- 萤火虫扰动策略
Abstract: The monitoring of the overflow concentration of the thickener is the key to realize intelligent dosing of coal slurry. The overflow concentration monitoring method based on the sensor will lead to the delay of flocculant regulation. In order to solve the above problem, a prediction method of overflow concentration of thickener based on improved sparrow search algorithm (ISSA) and long-short term memory (LSTM) is proposed. Firstly, the correlation analysis and pretreatment of multi-parameter time series in the process of concentration production are carried out to obtain the input variables. Secondly, the multi-strategies are combined to improve sparrow search algorithm (SSA). Tent chaotic map is introduced to initialize the sparrow population to ensure population diversity and speed up algorithm convergence. The optimization process of SSA is improved by using the spiral predation strategy to balance both local development and global search capabilities. The firefly perturbation strategy is used to perturb the sparrow search results to improve the global search performance and avoid the algorithm falling into local optimization. Thirdly, ISSA is used to optimize the hyperparameters of the two-layer LSTM network model. Finally, the overflow concentration prediction model based on ISSA-LSTM is established for on-line monitoring. The experimental results show the following points. ① The Ackley function and Rastigin function are selected as test functions. It is concluded that ISSA's global optimization capability and convergence speed are better than those of the particle swarm optimization (PSO) algorithm, whale optimization algorithm (WOA) and standard SSA. ② Among the three improved strategies, the spiral predation strategy plays a leading role in improving the performance of ISSA. The chaotic map and the firefly perturbation strategy coordinate the convergence speed and global search capability of the algorithm to further improve the optimization performance of the algorithm. ③ ISSA is used to optimize the hyperparameters of LSTM, which solves the problem of under-fitting or over-fitting when the values are determined by subjective experience. The prediction precision of overflow concentration of the ISSA-LSTM model reaches 97.26%, which is higher than that of double-layer LSTM, SSA-LSTM, and least square support vector machine (LSSVM) models. ④ Data pretreatment can improve the precision of the model, and the prediction precision of overflow concentration after noise reduction is improved by 30.25% compared with that before noise reduction.
- intelligent coal preparation /
- intelligent dosing /
- overflow concentration prediction /
- sparrow search algorithm /
- LSTM /
- chaotic map /
- spiral predation strategy /
- firefly perturbation strategy

HTML全文

图 1 煤泥水浓缩工艺流程

Figure 1. Concentration process flow of coal slurry

下载: 全尺寸图片幻灯片

图 2 Spearman相关性分析热力图

Figure 2. Thermodynamic diagram of Spearman correlation analysis

下载: 全尺寸图片幻灯片

图 3 降噪前后入料流量对比

Figure 3. Comparison of feed flow before and after noise reduction

下载: 全尺寸图片幻灯片

图 4 降噪前后溢流浓度对比

Figure 4. Comparison of overflow concentration before and after noise reduction

下载: 全尺寸图片幻灯片

图 5 双层LSTM网络结构

Figure 5. Structure of double-layer LSTM network

下载: 全尺寸图片幻灯片

图 6 Tent映射混沌值与随机值对比

Figure 6. Comparison of chaotic value and random value of Tent map

下载: 全尺寸图片幻灯片

图 7 ISSA流程

Figure 7. Flow of improved sparrow search algorithm

下载: 全尺寸图片幻灯片

图 8 基于ISSA−LSTM的浓缩池溢流浓度预测模型

Figure 8. Prediction model of overflow concentration of thickener based on ISSA-LSTM

下载: 全尺寸图片幻灯片

图 9 Ackley函数

Figure 9. Ackley function

下载: 全尺寸图片幻灯片

图 10 Rastrigin函数

Figure 10. Rastrigin function

下载: 全尺寸图片幻灯片

图 11 不同优化算法的Ackley函数收敛曲线

Figure 11. Convergence curves of Ackley function of different optimization algorithms

下载: 全尺寸图片幻灯片

图 12 不同优化算法的Rastrigin函数收敛曲线

Figure 12. Convergence curves of Rastrigin function of different optimization algorithms

下载: 全尺寸图片幻灯片

图 13 不同改进策略的Ackley函数收敛曲线

Figure 13. Convergence curves of Ackley function of different improvement strategies

下载: 全尺寸图片幻灯片

图 14 不同改进策略的Rastrigin函数收敛曲线

Figure 14. Convergence curves of Rastrigin function of different improvement strategies

下载: 全尺寸图片幻灯片

图 15 适应度曲线对比

Figure 15. Comparison of fitness curves

下载: 全尺寸图片幻灯片

图 16 不同模型预测结果对比

Figure 16. Comparison of prediction results of different models

下载: 全尺寸图片幻灯片

表 1 S−G算法参数

Table 1 Parameters of S-G algorithm

参数	平滑因子	多项式次数
入料流量	0.15	2
入料浓度	0.10	1
加药量	0.15	2

下载: 导出CSV

表 2 降噪前后预测结果

Table 2 Prediction results before and after noise reduction

实验数据	RMSE	MAPE/%
降噪前	492.711 5	32.99
降噪后	55.788 3	2.74

下载: 导出CSV

表 3 SSA改进策略

Table 3 Improvement strategies of sparrow search algorithm

算法	混沌映射	螺旋捕食	萤火虫扰动
ISSA1	否	是	是
ISSA2	是	否	是
ISSA3	是	是	否

下载: 导出CSV

表 4 设备硬件参数

Table 4 Equipment hardware parameters

设备名称	DESKTOP−3ER5HIS
处理器	11th Gen Intel(R) Core(TM) i7−11800H @ 2.30 GHz
机带RAM	32.0 GB
系统类型	64位操作系统

下载: 导出CSV

表 5 超参数设置

Table 5 Setting of hyperparameters

超参数	设定值
第1层神经元个数${M}_{1}$	[10，150]
第2层神经元个数${M}_{2}$	[10，60]
初始学习率$ {I}_{1} $	[10⁻⁵，1]
L2正则化因子$ {I}_{2} $	[10⁻⁶，1]
最大迭代次数$t_{\rm{max}}$	400

下载: 导出CSV

表 6 不同模型评价指标对比

Table 6 Comparison of evaluation indicator of different models

模型	RMSE	MAPE/%
ISSA−LSTM	55.788 3	2.74
SSA−LSTM	56.237 0	3.08
LSSVM	74.490 2	3.67
双层LSTM	81.259 7	3.98

下载: 导出CSV

参考文献(23)

[1]	冯甜欣,张晓光,刘景勇,等. 基于云平台的选煤厂环境监测系统[J]. 工矿自动化,2021,47(10):121-126. FENG Tianxin,ZHANG Xiaoguang,LIU Jingyong,et al. Environmental monitoring system of coal preparation plant based on cloud platform[J]. Industry and Mine Automation,2021,47(10):121-126.
[2]	董永胜,陈为高,侯佃平,等. 智能化选煤厂研究与建议[J]. 工矿自动化,2021,47(增刊1):26-31. DONG Yongsheng,CHEN Weigao,HOU Dianping,et al. Research and suggestions on intelligent coal preparation plant[J]. Industry and Mine Automation,2021,47(S1):26-31.
[3]	李朋,张明远,王保强. 分级破碎机智能化技术现状与发展方向[J]. 煤炭工程,2022,54(1):133-136. LI Peng,ZHANG Mingyuan,WANG Baoqiang. Status and development trends of intelligent technology of sizing crusher[J]. Coal Engineering,2022,54(1):133-136.
[4]	陈铭. 基于深度学习的浮选过程智能控制研究[D]. 徐州: 中国矿业大学, 2020. CHEN Ming. Study on the intelligent control of flotation process based on deep learning[D]. Xuzhou: China University of Mining and Technology, 2020.
[5]	赵毅鑫,杨志良,马斌杰,等. 基于深度学习的大采高工作面矿压预测分析及模型泛化[J]. 煤炭学报,2020,45(1):54-65. ZHAO Yixin,YANG Zhiliang,MA Binjie,et al. Deep learning prediction and model generalization of ground pressure for deep longwall face with large mining height[J]. Journal of China Coal Society,2020,45(1):54-65.
[6]	董丽丽,费城,张翔,等. 基于LSTM神经网络的煤矿突水预测[J]. 煤田地质与勘探,2019,47(2):137-143. DONG Lili,FEI Cheng,ZHANG Xiang,et al. Coal mine water inrush prediction based on LSTM neural network[J]. Coal Geology & Exploration,2019,47(2):137-143.
[7]	刘立邦,杨颂,王志坚,等. 基于改进WOA−LSTM的焦炭质量预测[J]. 化工学报,2022,73(3):1291-1299. LIU Libang,YANG Song,WANG Zhijian,et al. Prediction of coke quality based on improved WOA-LSTM[J]. CIESC Journal,2022,73(3):1291-1299.
[8]	宋刚,张云峰,包芳勋,等. 基于粒子群优化LSTM的股票预测模型[J]. 北京航空航天大学学报,2019,45(12):2533-2542. DOI: 10.13700/j.bh.1001-5965.2019.0388 SONG Gang,ZHANG Yunfeng,BAO Fangxun,et al. Stock prediction model based on particle swarm optimization LSTM[J]. Journal of Beijing University of Aeronautics and Astronautics,2019,45(12):2533-2542. DOI: 10.13700/j.bh.1001-5965.2019.0388
[9]	付华,赵俊程,付昱,等. 基于量子粒子群与深度学习的煤矿瓦斯涌出量软测量[J]. 仪器仪表学报,2021,42(4):160-168. DOI: 10.19650/j.cnki.cjsi.J2107486 FU Hua,ZHAO Juncheng,FU Yu,et al. Soft measurement of coal mine gas emission based on quantum-behaved particle swarm optimization and deep learning[J]. Chinese Journal of Scientific Instrument,2021,42(4):160-168. DOI: 10.19650/j.cnki.cjsi.J2107486
[10]	吴磊,康英伟. 基于改进粒子群优化长短时记忆神经网络的脱硫系统SO₂预测模型[J]. 热力发电,2021,50(12):66-73. WU Lei,KANG Yingwei. Prediction model of SO₂ concentration in desulfurization system based on improved particle swarm optimization LSTM[J]. Thermal Power Generation,2021,50(12):66-73.
[11]	LIU Guiyun,SHU Cong,LIANG Zhongwei,et al. A modified sparrow search algorithm with application in 3D route planning for UAV[J]. Sensors,2021,21(4):1224. DOI: 10.3390/s21041224
[12]	ZHU Yanlong,YOUSEFI N. Optimal parameter identification of PEMFC stacks using adaptive sparrow search algorithm[J]. International Journal of Hydrogen Energy,2021,46(14):9541-9552. DOI: 10.1016/j.ijhydene.2020.12.107
[13]	贾澎涛,苗云风. 基于堆叠LSTM的多源矿压预测模型分析[J]. 矿业研究与开发,2021,41(8):79-82. JIA Pengtao,MIAO Yunfeng. Multi-source mine pressure prediction model analysis based on stacked-LSTM[J]. Mining Research and Development,2021,41(8):79-82.
[14]	XING Yin,YUE Jianping,CHEN Chuang,et al. Dynamic displacement forecasting of dashuitian landslide in China using variational mode decomposition and stack long short-term memory network[J]. Applied Sciences,2019,9(15):2951-2962. DOI: 10.3390/app9152951
[15]	XUE Jiankai,SHEN Bo. A novel swarm intelligence optimization approach:sparrow search algorithm[J]. Systems Science & Control Engineering,2020,8(1):22-34.
[16]	吕鑫,慕晓冬,张钧. 基于改进麻雀搜索算法的多阈值图像分割[J]. 系统工程与电子技术,2021,43(2):318-327. DOI: 10.12305/j.issn.1001-506X.2021.02.05 LYU Xin,MU Xiaodong,ZHANG Jun. Multi-threshold image segmentation based on improved sparrow search algorithm[J]. Systems Engineering and Electronics,2021,43(2):318-327. DOI: 10.12305/j.issn.1001-506X.2021.02.05
[17]	吴新忠,韩正化,魏连江,等. 矿井风流智能按需调控算法与关键技术[J]. 中国矿业大学学报,2021,50(4):725-734. DOI: 10.13247/j.cnki.jcumt.001316 WU Xinzhong,HAN Zhenghua,WEI Lianjiang,et al. Intelligent on-demand adjustment algorithm and key technology of mine air flow[J]. Journal of China University of Mining & Technology,2021,50(4):725-734. DOI: 10.13247/j.cnki.jcumt.001316
[18]	滕志军,吕金玲,郭力文,等. 一种基于Tent映射的混合灰狼优化的改进算法[J]. 哈尔滨工业大学学报,2018,50(11):40-49. DOI: 10.11918/j.issn.0367-6234.201806096 TENG Zhijun,LYU Jinling,GUO Liwen,et al. An improved hybrid grey wolf optimization algorithm based on Tent mapping[J]. Journal of Harbin Institute of Technology,2018,50(11):40-49. DOI: 10.11918/j.issn.0367-6234.201806096
[19]	DEMIDOVA L,GORCHAKOV A V. A study of chaotic maps producing symmetric distributions in the fish school search optimization algorithm with exponential step decay[J]. Symmetry,2020,12(5):784-801. DOI: 10.3390/sym12050784
[20]	MIRJALILI S,LEWIS A. The whale optimization algorithm[J]. Advances in Engineering Software,2016,95:51-67. DOI: 10.1016/j.advengsoft.2016.01.008
[21]	孙林,黄金旭,徐久成,等. 基于自适应鲸鱼优化算法和容错邻域粗糙集的特征选择算法[J]. 模式识别与人工智能,2022,35(2):150-165. DOI: 10.16451/j.cnki.issn1003-6059.202202006 SUN Lin,HUANG Jinxu,XU Jiucheng,et al. Feature selection based on adaptive whale optimization algorithm and fault-tolerance neighborhood rough sets[J]. Pattern Recognition and Artificial Intelligence,2022,35(2):150-165. DOI: 10.16451/j.cnki.issn1003-6059.202202006
[22]	柳长安,冯雪菱,孙长浩,等. 基于改进麻雀算法的最大2维熵分割方法[J]. 激光技术,2022,46(2):274-282. DOI: 10.7510/jgjs.issn.1001-3806.2022.02.020 LIU Chang'an,FENG Xueling,SUN Changhao,et al. Maximum two-dimensional entropy segmentation method based on improved sparrow algorithm[J]. Laser Technology,2022,46(2):274-282. DOI: 10.7510/jgjs.issn.1001-3806.2022.02.020
[23]	曹通,白艳萍. 基于梯度下降优化的LSTM对空气质量预测研究[J]. 陕西科技大学学报,2020,38(6):159-164. DOI: 10.3969/j.issn.1000-5811.2020.06.026 CAO Tong,BAI Yanping. Study on the prediction of air quality based on LSTM with gradient descent optimization[J]. Journal of Shaanxi University of Science & Technology,2020,38(6):159-164. DOI: 10.3969/j.issn.1000-5811.2020.06.026

施引文献(10)

期刊类型引用(6)

1.	陈书航，王世博，葛世荣，王赟，马广军. 综采工作面刮板输送机煤流时空分布研究. 工矿自动化. 2024(09): 98-107 . 本站查看
2.	王桂忠，叶隆浩. 基于煤流量识别的带式输送机节能控制系统设计与研究. 煤矿机械. 2023(01): 14-17 . 百度学术
3.	周爱民，叶飞，施旭东，赵培成. 基于超声波传感器的带式输送机烟丝瞬时流量监测系统的设计. 现代信息科技. 2022(03): 149-152 . 百度学术
4.	唐文杰. 带式输送机运行状态智能监控体系的研究. 机械管理开发. 2022(09): 300-301+304 . 百度学术
5.	毛清华，毛金根，马宏伟，张旭辉，李铮. 矿用带式输送机智能监测系统研究. 工矿自动化. 2020(06): 48-52+58 . 本站查看
6.	李瑶，王义涵. 带式输送机煤流量自适应检测方法. 工矿自动化. 2020(06): 98-102 . 本站查看