Periodic pressure prediction of working face based on dynamic adaptive sailfish optimization BP neural network
-
摘要: 针对现有工作面周期来压预测方法精度不足、泛化性较差和算力要求高等问题，提出了一种基于动态自适应旗鱼优化BP神经网络（DASFO−BP）的工作面周期来压预测模型。通过分析工作面周期来压机理，得到与来压相关的影响因素，通过皮尔逊相关系数确定对来压具有显著影响的因素（推进速度、直接顶厚度、基本顶厚度、采高、煤层倾角和倾向长度）作为预测模型输入，并以下次来压强度和来压步距作为预测模型输出。针对旗鱼优化（SFO）算法鲁棒性不足的问题，提出了动态自适应优化策略对SFO算法进行改进，即在优化前期利用SFO达到快速收敛的目的，中期则借助秃鹰搜索（BES）跳出局部最优，后期发挥粒子群优化（PSO）深度搜索的优势来提高解的精度。通过改进后的动态自适应旗鱼优化（DASFO）算法对BP神经网络的超参数进行训练，构建了基于DASFO−BP的来压预测模型。实验结果表明：DASFO算法在单峰和多峰测试函数上均能实现快速收敛；与BP，SFO−BP和NCPSO−BP相比，DASFO−BP对周期来压强度和步距的预测值与真实值更为接近，具有更高的精度，拟合能力和泛化能力强，能够准确预测下一周期来压分布情况。Abstract: In order to solve the problems of insufficient precision, poor generalization, and high computational requirements of existing methods for periodic pressure prediction of working face, a periodic pressure prediction model of working face based on dynamic adaptive sailfish optimization BP neural network (DASFO-BP) is proposed. By analyzing the mechanism of working face periodic pressure, the influencing factors related to pressure are obtained. The Pearson correlation coefficient is used to determine the factors that have a significant impact on pressure (advance speed, direct roof thickness, basic roof thickness, mining height, coal seam dip angle, and dip length) as inputs for the prediction model. The subsequent pressure intensity and pressure step distance are used as outputs for the prediction model. A dynamic adaptive optimization strategy is proposed to improve the robustness of the sailfish optimization (SFO) algorithm. In the early stage of optimization, SFO is used to achieve fast convergence, while in the middle stage, bald eagle search (BES) is used to escape local optima. In the later stage, the advantage of particle swarm optimization (PSO) deep search is utilized to improve the precision of the solution. A dynamic adaptive sailfish optimization (DASFO) algorithm is improved to train the hyperparameters of the BP neural network, and a pressure prediction model based on DASFO-BP is constructed. The experimental results indicate that the DASFO algorithm can achieve fast convergence on both unimodal and multimodal test functions. Compared with BP, SFO-BP, and NCPSO-BP, DASFO-BP has higher precision in predicting the intensity and step distance of periodic pressure, and has strong generalization ability and fitting capability. It can accurately predict the pressure and its distribution in the next period.
-
图 2 基于DASFO−BP的工作面周期来压预测模型结构
Figure 2. Structure of periodic pressure prediction model of working face based on dynamic adaptive sailfish optimization（DASFO）-BP
表 1 周期来压与影响因素相关性分析结果
Table 1. Analysis results of correlation between periodic pressure and influencing factors
参数 皮尔逊相关系数 直接顶厚度 基本顶厚度 采高 煤层厚度 煤层倾角 倾向长度 推进速度 周期来压步距 0.32 0.41 0.55 0.13 0.48 0.37 0.58 周期来压强度 0.52 0.56 0.41 0.07 0.35 0.28 0.61 
下载: 导出CSV
表 2 不同模型评价指标结果
Table 2. Evaluation index results of different models
模型 周期来压强度 周期来压步距 RMSE MAE R2 RMSE MAE R2 BP 0.0116 0.0598 0.6441 0.0076 0.0374 0.7753 SFO−BP 0.0136 0.0854 0.4676 0.0174 0.0568 0.5462 NCPSO−BP 0.0026 0.0386 0.8962 0.0037 0.0387 0.6732 DASFO−BP 0.0023 0.0317 0.9092 0.0030 0.0254 0.9891
下载: 导出CSV
表 3 不同模型的周期来压预测结果
Table 3. Periodic pressure prediction results of different models
日期/（年−月−日） BP SFO−BP NCPSO−BP DASFO−BP 真实来压
强度/MPa真实来压
步距/m预测来压
强度/MPa预测来压
步距/m预测来压
强度/MPa预测来压
步距/m预测来压
强度/MPa预测来压
步距/m预测来压
强度/MPa预测来压
步距/m2023−12−09 45.46 22.64 46.88 25.39 45.24 25.72 46.25 24.25 46.99 25.59 2023−12−15 28.32 21.44 26.72 21.55 26.62 18.69 28.20 18.78 28.51 19.96 2023−12−19 29.35 20.56 32.62 18.83 29.52 20.46 31.15 20.52 30.53 20.17 2023−12−24 32.80 20.30 32.25 18.88 32.76 19.33 32.27 19.47 32.21 21.29 2023−12−30 32.55 26.03 30.31 26.09 30.28 26.31 31.78 29.10 31.52 28.43 2024−01−03 29.90 18.51 33.52 18.03 32.53 17.99 31.11 18.24 30.60 16.78 2024−01−07 28.35 26.70 29.35 25.79 30.04 22.25 29.71 26.78 29.91 25.06 2024−01−14 48.78 21.43 45.92 17.85 46.97 19.89 47.11 20.77 47.78 20.34 2024−01−20 43.91 24.00 42.84 23.72 41.38 22.69 43.27 25.23 42.35 23.56 2024−01−26 30.96 27.16 32.96 25.19 32.69 25.99 32.07 28.10 32.13 27.88 均方误差 1.2905 2.5368 2.2279 3.1205 1.5147 2.2162 0.2698 1.5184 — —
下载: 导出CSV
-
[1] 段柱平. 基于煤矿采煤工作面顶板事故原因及其防治措施探析[J]. 当代化工研究,2023(13):87-89.DUAN Zhuping. Analysis on the causes and prevention measures of roof accidents in coal mining faces[J]. Modern Chemical Research,2023(13):87-89. [2] TAN Tingjiang,YANG Zhen,CHANG Feng,et al. Prediction of the first weighting from the working face roof in a coal mine based on a GA-BP neural network[J]. Applied Sciences,2019,9(19). DOI: 10.3390/app9194159. [3] 李泽西. 基于可变时序移位Transformer−LSTM的集成学习矿压预测方法[J]. 工矿自动化,2023,49(7):92-98.LI Zexi. Ensemble learning mine pressure prediction method based on variable time series shift Transformer-LSTM[J]. Journal of Mine Automation,2023,49(7):92-98. [4] 贾永杰. 野青灰岩坚硬顶板工作面来压步距与强度预测分析[J]. 煤炭与化工,2022,45(12):1-7,11.JIA Yongjie. Prediction and analysis of pressure step and strength for Yeqing limestone hard roof[J]. Coal and Chemical Industry,2022,45(12):1-7,11. [5] 冯银辉,宋阳,李务晋,等. 基于支架数据优化的工作面矿压预测模型研究[J]. 煤炭工程,2023,55(6):101-107.FENG Yinhui,SONG Yang,LI Wujin,et al. Mine pressure prediction model for fully mechanized working face based on data optimization of hydraulic support[J]. Coal Engineering,2023,55(6):101-107. [6] 路建军,周宏范,冯明,等. 综采工作面来压步距预测及修正方法研究[J]. 煤炭工程,2022,54(11):118-123.LU Jianjun,ZHOU Hongfan,FENG Ming,et al. Prediction and correction method of weighting interval for fully mechanized mining face[J]. Coal Engineering,2022,54(11):118-123. [7] 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. [8] 冀汶莉,田忠,张丁丁,等. 基于遗传算法−深度神经网络的分布式光纤监测工作面矿压预测[J]. 科学技术与工程,2022,22(24):10485-10492. doi: 10.3969/j.issn.1671-1815.2022.24.016JI Wenli,TIAN Zhong,ZHANG Dingding,et al. Mine pressure prediction by genetic algorithm-deep neural network based on distributed optical fiber monitoring[J]. Science Technology and Engineering,2022,22(24):10485-10492. doi: 10.3969/j.issn.1671-1815.2022.24.016 [9] 吕文玉,王海金,伍永平,等. 基于PSO−SVR预测模型的综采工作面周期来压研究[J]. 煤炭工程,2022,54(4):86-91.LYU Wenyu,WANG Haijin,WU Yongping,et al. PSO-SVR prediction model of periodic weighting in fully mechanized working face[J]. Coal Engineering,2022,54(4):86-91. [10] 高阳. 基于长短时记忆神经网络的工作面矿压预测方法研究[D]. 焦作:河南理工大学,2022.GAO Yang. Study on prediction method of mine pressure in working face based on long-short-term memory neural network[D]. Jiaozuo:Henan Polytechnic University,2022. [11] LIU Yaping,DONG Lihong,YE Ou. Mine pressure prediction model of fully mechanized mining face based on improved transformer[C]. IEEE International Conference on Signal Processing,Communications and Computing,Xi'an,2022:1-5. [12] RUMELHART D E,HINTON G E,WILLIAMS R J. Learning representations by back-propagating errors[J]. Nature,1986,323:533-536. doi: 10.1038/323533a0 [13] 张琦. 综放工作面周期来压步距和来压强度的定量预测方法研究[J]. 煤炭与化工,2017,40(2):37-39,42.ZHANG Qi. Study on quantitative prediction methods of ground pressure strength and periodic weighting length in fully mechanized caving face[J]. Coal and Chemical Industry,2017,40(2):37-39,42. [14] 王陈真. 孤岛综放工作面覆岩结构分析及矿压特征研究[J]. 煤矿现代化,2019(5):26-29. doi: 10.3969/j.issn.1009-0797.2019.05.009WANG Chenzhen. The analysis of overburden rock structure and the study of ore pressure characteristics in the fully mechanized caving face of isolated island[J]. Coal Mine Modernization,2019(5):26-29. doi: 10.3969/j.issn.1009-0797.2019.05.009 [15] 李云鹏,赵善坤,李杨,等. 复杂坚硬岩层条件下特厚煤层综放开采矿压分级预测[J]. 煤炭学报,2021,46(增刊1):38-48.LI Yunpeng,ZHAO Shankun,LI Yang,et al. Prediction on weighting classification of fully-mechanized caving mining under extremely thick coal seam[J]. Journal of China Coal Society,2021,46(S1):38-48. [16] 巩师鑫,任怀伟,杜毅博,等. 基于MRDA−FLPEM集成算法的综采工作面矿压迁移预测[J]. 煤炭学报,2021,46(增刊1):529-538.GONG Shixin,REN Huaiwei,DU Yibo,et al. Transfer prediction of underground pressure for fully mechanized mining face based on MRDA-FLPEM integrated algorithm[J]. Journal of China Coal Society,2021,46(S1):529-538. [17] SHADRAVAN S,NAJI H R,BARDSIRI V K. The sailfish optimizer:a novel nature-inspired metaheuristic algorithm for solving constrained engineering optimization problems[J]. Engineering Applications of Artificial Intelligence,2019,80:20-34. doi: 10.1016/j.engappai.2019.01.001 [18] 黄鹤,熊武,吴琨,等. 基于记忆传递旗鱼优化的K均值混合迭代聚类[J]. 上海交通大学学报,2022,56(12):1638-1648.HUANG He,XIONG Wu,WU Kun,et al. K-means hybrid iterative clustering based on memory transfer sailfish optimization[J]. Journal of Shanghai Jiao Tong University,2022,56(12):1638-1648. [19] 段洁,伍瑞泽,尤敬尧,等. 基于改进秃鹰算法优化LSSVM的变压器故障诊断[J]. 红水河,2024,43(1):96-101. doi: 10.3969/j.issn.1001-408X.2024.01.018DUAN Jie,WU Ruize,YOU Jingyao,et al. Transformer fault diagnosis based on improved vulture algorithm to optimize LSSVM[J]. Hongshui River,2024,43(1):96-101. doi: 10.3969/j.issn.1001-408X.2024.01.018 [20] 田劼,李阳,张磊,等. 基于PSO−BP神经网络的临时支架支撑力自适应控制[J]. 工矿自动化,2023,49(7):67-74.TIAN Jie,LI Yang,ZHANG Lei,et al. Adaptive control of temporary support force based on PSO-BP neural network[J]. Journal of Mine Automation,2023,49(7):67-74. [21] 顾丽丽,刘勇,甄佳奇. 基于改进粒子群算法优化BP神经网络的甜菜产量预测方法[J]. 新疆大学学报(自然科学版),2021,38(2):191-196.GU Lili,LIU Yong,ZHEN Jiaqi. Prediction of beet yield based on BP neural network optimized by improved particle swarm algorithm[J]. Journal of Xinjiang University(Natural Science Edition),2021,38(2):191-196. -
点击查看大图
图(6) / 表(3)
计量
- 文章访问数: 108
- HTML全文浏览量: 30
- PDF下载量: 9
- 被引次数: 0
