论文部分内容阅读
突水水源判别是矿井防治水工作的重要环节。为准确有效地判别矿井突水水源,选取k++Na+、Ca2+、Mg2+、Cl-、HCO-3、SO2-46种离子作为判别突水来源的依据。对粒子群的惯性权重、两个学习因子进行非线性和线性改进,并引入变异算子对粒子的飞行速度进行改进,提高了标准粒子群(SPSO)算法的性能,确保了粒子群在进化过程中的多样性。运用改进的粒子群(MPSO)算法优化BP神经网络的初始权值与阈值,以某矿的20组水源样本作为训练样本进行学习与训练,并用建立好的MPSO-BP突水水源判别模型对6组待测样本进行判别,并对适应度进化曲线和判别结果进行详细的分析与推理。应用结果表明:相比于SPSO算法,MPSO算法在优化BP神经网络过程中具有全局搜索能力更强、收敛速度更快、精度更高的优点,能够有效克服SPSO算法易早熟收敛的缺点;MPSO-BP模型应用于矿井突水水源的判别具有可行性,其判别结果的精度及稳定性明显优于BP模型和SPSO-BP模型,其具有判别泛化性更强的特点。因此,该方法在判别矿井突水水源方面具有一定的实用价值与开发潜力。
Water inrush discrimination is an important part of coal mine prevention and control work. In order to judge the mine water inrush accurately and effectively, K ++ Na +, Ca2 +, Mg2 +, Cl-, HCO-3, SO2- 46 ions were selected as the basis for judging the sources of water inrush. The inertia weight of the particle swarm, the two learning factors are improved nonlinearly and linearly, and the mutation operator is introduced to improve the flying speed of the particle, which improves the performance of the standard particle swarm optimization (SPSO) algorithm and ensures the evolution of particle swarm In the diversity. The improved particle swarm optimization (MPSO) algorithm is used to optimize the initial weights and thresholds of BP neural network. Twenty water samples in a mine are used as training samples to study and train. The MPSO-BP water inrush discriminant model The samples to be tested are discriminated, and the evolution curves of fitness and the discrimination results are analyzed and reasoned in detail. The application results show that compared with SPSO algorithm, MPSO algorithm has the advantages of stronger global search ability, faster convergence rate and higher accuracy in the process of BP neural network optimization, which can effectively overcome the shortcoming that SPSO algorithm premature convergence; MPSO- It is feasible to apply BP model to discriminate water inrush from coal mines. The accuracy and stability of BP model are superior to those of BP model and SPSO-BP model, which are more discriminant and more generalized. Therefore, this method has certain practical value and development potential in judging mine water inrush.