我国是煤炭生产大国，煤炭作为国家能源安全稳定供应的“压舱石”，其重要性不言而喻。随着社会经济的高速发展，对能源的需求日益增大，煤炭资源开采的强度和深度逐渐加大，使得工作面矿压显现愈发剧烈，严重威胁井下人员和设备安全。因此，精准预测工作面矿压变化趋势，对保障煤炭资源安全开采、防治矿压灾害发生和为确定合理的支护参数提供数据参考具有极其重要的意义。

论文以梅花井煤矿232205工作面及其回采巷道的动态矿压监测和来压预测预警研究为背景，综合运用工程调研、理论分析、数学建模和工程实践等研究方法，对梅花井煤矿工作面矿压预警预测方法行了系统性研究。论文主要取得以下几方面研究成果：

提出了基于免疫粒子群混合算法优化BP网络(IA-PSO-BP)的工作面矿压预测模型。通过制定232205工作面及其回采巷道的矿压监测方案，以获取原始矿压数据，并对其进行预处理和归一化处理。以支架工作阻力数据为例，划分用于训练和验证模型的数据集。应用免疫粒子群混合(IA-PSO)优化算法对BP神经网络进行超参数优化，建立了基于IA-PSO-BP的工作面矿压预测模型。选取平均绝对误差(MAE)、均方误差(MSE)和相关系数(R²)作为评价指标来量化评价BP模型、PSO-BP模型和IA-PSO-BP模型的预测性能。实验结果表明：(1)IA-PSO-BP模型的收敛速度相较于BP和PSO-BP模型分别提高约8倍和2倍；(2)相较于BP和PSO-BP模型，IA-PSO-BP模型在3组测试集上的MAE和MSE均最小，R²均最大，预测精度显著提高，预测结果符合矿压数据的周期性变化规律，符合煤矿现场实际。

将所提出的IA-PSO-BP矿压预测模型在梅花井煤矿232205工作面上进行工程实际应用，构建智能矿压预测预警系统，实现了工作面及回采巷道矿压的动态监测及智能预警预测功能。应用结果表明，系统运行稳定，能够实现临灾告警和超前预警功能，对预防煤矿井下矿压灾害发生有着一定的指导意义。

China, being a major producer of coal, relies on this resource as a crucial element of its national energy security and a stable energy supply. The significance of coal is indisputable in this context. As the social economy experiences rapid development, the demand for energy is surging. Consequently, there is a gradual escalation in the intensity and depth of coal mining, placing increasing pressure on the working face of the mines. This mounting pressure poses a significant threat to the safety of underground personnel and equipment. Therefore, accurately predicting the trend of mine pressure at the working face holds immense importance. Such predictions ensure the safe extraction of coal resources, prevent and manage mine pressure disasters, and furnish valuable data for establishing appropriate support parameters.

This thesis focuses on the dynamic monitoring of mine pressure and the study of periodic pressure prediction and early warning in the working face of Meifuajing coal mine (232205) and its extraction tunnel. Drawing upon engineering research, theoretical analysis, mathematical modeling, and practical engineering, the thesis systematically investigates the methods for early warning and prediction of mine pressure in the working face of Meifuajing coal mine. The key research findings of this paper are summarized as follows:

A working face mine pressure prediction model, based on IA-PSO-BP, is proposed in this study. To obtain the raw mine pressure data, a monitoring program is developed for the working face 232205 and its backing lanes. The obtained data is pre-processed and normalized. As an example, the stent work resistance data is used to create a data set for training and validating the model. The IA-PSO optimization algorithm is employed to optimize the BP neural network with hyperparameters, thereby establishing the IA-PSO-BP-based working face mine pressure prediction model. Evaluation of the prediction performance of the BP model, PSO-BP model, and IA-PSO-BP model is conducted using mean absolute error (MAE), mean square error (MSE), and correlation coefficient (R²) as quantitative indices. The experimental results demonstrate that: (1) the IA-PSO-BP model exhibits a convergence speed approximately 8 times and 2 times faster than the BP and PSO-BP models, respectively; (2) in comparison to the BP and PSO-BP models, the IA-PSO-BP model demonstrates the smallest MAE and MSE, as well as the largest R², across the three test sets, indicating a significant improvement in prediction accuracy. The predictions align with the cyclic variation pattern of mine pressure data and are consistent with real-world observations at the coal mine site.

The IA-PSO-BP mine pressure prediction model, as proposed, was implemented in the working face of Meifujing coal mine (232205) to establish an intelligent system for mine pressure prediction and early warning. This system enables dynamic monitoring and intelligent prediction of mine pressure in the working face and extraction lane. The application results demonstrate the system's stable operation and its capability to provide timely warnings for impending disasters and over-warnings. These findings hold significant guidance in preventing mine pressure disasters in underground coal mines.