基于GD-FNN算法的危化品气体传感器阵列数据融合

Data fusion of hazardous chemical gas sensor array based on GD-FNN algorithm

基于以ZnO2为敏感材料的半导体气体传感器构建2×2和2×3危化品气体传感器阵列,搭建危化品气体检测平台,对CO和CH4组成的混合气体进行气体检测,建立融合模型对混合气体进行定量分析。通过对气体传感器进行标定,发现气体传感器的响应与气体浓度之间具有非线性关系,选择可以通过自学习确定模糊规则的广义动态模糊神经网络(GD-FNN)作为定量分析工具;利用GD-FNN算法构建2×2阵列和2×3阵列的数据融合模型,设置初始参数并进行网络训练,最后利用GD-FNN算法融合模型对混合危化品气体测试样本数据进行定量分析。结果表明:融合结果中两种阵列数据融合模型的融合精度相近,总体最大相对误差为10.63%,平均精度可以达到90%以上,可靠性强,具有较好的参考价值。

Based on the semiconductor gas sensor using ZnO2 as the sensitive material,2×2 and 2×3 hazardous chemical gas sensor arrays are constructed,and the hazardous chemical gas detection platform is built.The mixed gas composed of CO and CH4 is detected,and the fusion model is established for quantitative analysis of the mixed gas.By calibrating the gas sensor,it is found that there is a nonlinear relationship between the response of the gas sensor and the gas concentration.The generalized dynamic fuzzy neural network(GD-FNN)which can determine the fuzzy rules by self-learning is selected as the quantitative analysis tool.GD-FNN algorithm is used to construct the data fusion model of 2×2 array and 2×3 array,set the initial parameters and conduct network training.Finally,GD-FNN algorithm fusion model is used to conduct quantitative analysis on the test sample data of mixed hazardous chemical gases.The results show that the fusion accuracy of the two array data fusion models is similar,the maximum relative error is 10.63%,the average precision can reach more than 90%,the reliability is strong,and it has a good reference value.