中国安全科学学报 ›› 2023, Vol. 33 ›› Issue (S1): 112-118.doi: 10.16265/j.cnki.issn1003-3033.2023.S1.1304

• 安全工程技术 • 上一篇    下一篇

基于AdaBoost的地下采掘工程高韧性安全评估方法

徐晓滨1,2(), 施凡1,2(), 冯静1,2, 徐晓健1,2, 常雷雷1,2,**()   

  1. 1 杭州电子科技大学 自动化学院, 浙江 杭州 310018
    2 杭州电子科技大学 中国-奥地利人工智能与先进制造“一带一路”联合实验室, 浙江 杭州 310018
  • 收稿日期:2023-01-24 修回日期:2023-04-20 出版日期:2023-06-30
  • 通讯作者:
    **常雷雷(1985—),男,河北沧州人,博士,副研究员,主要从事复杂系统建模和安全性评估理论与实践方面的研究。E-mail:
  • 作者简介:

    徐晓滨 (1980—),男,河南郑州人,博士,教授,主要从事复杂系统安全性评估、故障预测等方面的研究。E-mail:

    施凡 (1999—),男,江苏南通人,硕士研究生,研究方向为复杂系统建模、机器学习以及安全性评估。E-mail:

    冯静 助理研究员

    徐晓健 助理研究员

  • 基金资助:
    浙江省属高校基本科研业务费专项资助(GK239909299001-010); 国家重点研发计划资助(2022YFE0210700); 国家自然科学基金资助(61903108); 浙江省杰出青年基金资助(LR21030001)

High-resilience safety assessment method of underground coal mining using AdaBoost

XU Xiaobin1,2(), SHI Fan1,2(), FENG Jing1,2, XU Xiaojian1,2, CHANG Leilei1,2,**()   

  1. 1 School of Automation, Hangzhou Dianzi University, Hangzhou Zhejiang 310018, China
    2 China-Austria Belt and Road Joint Laboratory on Artificial Intelligence and Advanced Manufacturing, Hangzhou Dianzi University, Hangzhou Zhejiang 310018, China
  • Received:2023-01-24 Revised:2023-04-20 Published:2023-06-30

摘要:

为确保煤炭地下采掘工程的安全性,提出一种基于AdaBoost的高韧性安全评估方法。首先,将隧道设计参数、地质条件参数和施工参数作为模型的输入,矿井沉降值作为输出,将施工现场采集的输入参数数据划分为训练数据集和测试数据集;其次,从数据集中抽取多个训练子集,分别建立多个子模型;然后,分别计算子模型的误差和权重,以神经网络作为基准模型,将所有训练数据集划分为正常数据和异常数据,并更新权重;最后,根据加权后的数据集,构建AdaBoost预测模型,并开展案例分析,验证该方法的有效性。结果表明:相比于直接使用AdaBoost算法,基于正常数据和异常数据构建的AdaBoost高韧性安全评估方法所得的结果准确率更高,该方法具有一定的科学性和有效性。

关键词: 地下采掘工程, AdaBoost, 高韧性, 安全评估, 异常数据, 正常数据

Abstract:

In order to ensure the safety of coal mining engineering, a high-resilience safety assessment method based on AdaBoost was proposed. Firstly, the tunnel design parameters, geological condition parameters, and construction parameters were taken as the input of the model, and the mine settlement value was taken as the output. The input parameter data collected at the construction site were divided into a training data set and a test data set. Secondly, multiple data sets were extracted from the data set to establish multiple sub-models respectively. Then, the errors and weights of the sub-models were calculated respectively, and the neural network was used as the benchmark model. All the training data sets were divided into normal data and abnormal data, and the weights were updated. Finally, according to the weighted data set, the AdaBoost prediction model was built, and case analysis was carried out. The effectiveness of the method was verified. The results show that the high-resilience safety assessment method using AdaBoost established based on normal data and abnormal data has higher accuracy than the results obtained by directly using the AdaBoost algorithm, which proves that the proposed method is scientific and effective.

Key words: coal mining engineering, AdaBoost, high resilience, safety assessment, abnormal data, normal data