矿井瓦斯与矿尘气固两相流多参数在线测量方法研究

Weighty disasters and accidents frequently occurs because of the poor production environment of mines. Gas and mine dust are considered as the main factors of abnormal death of miners. Therefore, they are important indexes of mine safety detection. A measurement method based on polarized light absorption and scattering spectra was proposed in this project to eliminate the influence of gas and mine dust overlapping on test results in the process of the optical measurement, and implement the simultaneous detection of gas components and solid dust. The theoretical model of gas and mine dust spectrum measurement will be established by simulation using Monte Carlo method. Polarization ratio which is irrelevant to the gas absorption will be considered as dust detection signal in this model. The information of dust particle size and concentration will be extracted directly. Then, Then,non-mode inversion algorithms based on regularization method will be used to calculate the particle size and concentration of mine dust accurately. Gas absorption spectrum will be compensated by the use of calculation results of mine dust,thus eliminating the effects of mine dust on the detection of methane gas and detecting gas concentration with high precision. Experimental platform will be constructed. The feasibility of the method and the reliability of the system will be verified by calibrating the standard gas and particle samples under conditions of different temperature, humidity and pressure. The method can provide theoretical basis and practical test method for gas and dust early-warning of mine.

矿井生产环境恶劣，重大型灾害、事故时有发生。矿井中瓦斯和矿尘是引起矿工非正常死亡的主要因素，因此瓦斯和矿尘是矿井安全检测的重要指标。项目提出一种基于偏振光吸收、散射光谱的测量方法，消除光学测量过程中瓦斯和矿尘信号相互交叠对检测结果的影响，实现矿井气体组分和固体矿尘的同时检测。项目通过蒙特卡洛方法进行理论仿真,建立瓦斯、矿尘光谱测量理论模型，模型利用与气体吸收无关的偏振比作为矿尘检测信号，直接提取矿尘颗粒粒径和浓度信息，采用基于正则化的独立模式颗粒反演算法，实现矿尘颗粒粒径和浓度精确计算，利用矿尘计算结果对气体吸收光谱进行补偿，消除矿尘对瓦斯气体检测的影响，实现瓦斯浓度的高精度检测。搭建实验平台，在不同温度、湿度、压力的情况下采用标准瓦斯气体和颗粒样本进行标定和实验研究，对方法的可行性和系统的可靠性进行验证。该方法为矿井瓦斯和矿尘预警提供理论依据和可行的测试手段。

矿井生产环境恶劣，重大型灾害、事故时有发生。矿井中瓦斯和矿尘是引起矿工非正常死亡的主要因素，因此瓦斯和矿尘是矿井安全检测的重要指标。项目提出一种基于偏振光吸收、散射光谱的测量方法，消除光学测量过程中瓦斯和矿尘信号相互交叠对检测结果的影响，实现矿井气体组分和固体矿尘的检测。项目主要研究成果如下：颗粒检测方面，建立了基于光散射的偏振比发理论模型，模型利用与气体吸收无关的偏振比作为矿尘检测信号，直接提取矿尘颗粒粒径和浓度信息，利用Tikhonov算法建立正则化目标函数数学模型，并采用L曲线法优化正则参数，最后改进了基于鱼群算法的非独立颗粒粒径反演模型及非负最小二乘法的独立颗粒粒径反演模型对颗粒粒径进行反演，经过仿真和实验验证，本项目所采用方法颗粒粒径反演精度高于原方法。气体组分测量方面，本项目研究了基于小波变换的光谱数据降噪处理方法，并采用峰值提取法（PEM）和主成分分析法（PCA）两种特征提取方法对降噪后的数据进行特征提取和降维，然后建立支持向量机（SVM）及其改进模型，对SVM模型的预测精度进行分析，确定利用PSO算法对SVM的惩罚参数c和核函数参数g进行寻优，建立基于PSO算法的改进 PEM-SVM模型和改进PCA-SVM模型。最后整合两个模型的优点，建立了PEM-PCA-SVM集成模型。使用测试集样本对PEM-PCA-ELM模型进行测试验证前面模型进行比较，实现CH4、CO2、C2H6、C3H8和C4H10五种其他的高精度预测。通过仿真和实验，验证了本项目提出方法的可行性。该方法为矿井瓦斯和矿尘预警提供理论依据和可行的测试手段。通过项目的研究，发表相关论文16篇，申请专利2项，培养研究生8名，获国家质检总局科技兴检奖1项。