【摘要】 为验证煤矿井下瓦斯浓度数据监测的准确性,实现对瓦斯传感器的远程状态评估,保证煤矿井下的安全工作,提出了一种基于小波包和分形理论的瓦斯传感器状态评估方法。首先构建了优化支持向量回归模型,利用输出的残差信号判断瓦斯传感器工作状态是否异常;针对瓦斯传感器的常见故障,结合小波包变换与分形理论对传感器的异常信号进行故障特征提取,构建故障特征向量;利用粒子群算法优化支持向量机的偏二叉树模型对故障进行分类识别,输出状态评估结果。与传统的BP神经网络法和SVM分类法对比,所提评估方法的整体准确率达95%,说明了其有效性和准确性。更多还原

【Abstract】 It proposes a gas sensor status assessment method based on wavelet packet and fractal theory to verify the accuracy of gas concentration data monitoring in coal mines, realize remote status assessment of gas sensors and ensure safe works in coal mines. In this paper, an optimized support vector regression model is constructed, and the output residual signal is used to determine whether the gas sensor’s working state is abnormal. For the common faults of the gas sensor, combined with wavelet packet transform and fractal theory, the abnormal signal of the sensor is extracted and the fault is constructed feature vector, use particle swarm algorithm to optimize the partial binary tree model of support vector machine to classify and identify faults and output result. Compared with the traditional BP neural network method and SVM classification method, the method proposed in this paper has a fault recognition rate of 95%, which shows its effectiveness and accuracy.更多还原