基于粗糙集的多传感器联合燃爆辨识技术

Deflagration identification of multi-sensor based on rough set theory

燃爆辨识是获取物质理化参数及化学品物理危害的关键技术。针对燃爆状态复杂多变、单一传感器辨识效果不佳的问题,引入多传感器联合辨识,并采用粗糙集法对多传感器数据进行融合。首先分析了燃爆参数间内在关联性,设计了多传感器联合燃爆检测装置;从燃爆检测数据形成的决策表出发,运用基于完全简化规则集的粗糙集理论进行属性约简与值约简;在保证燃爆辨识准确率的前提下挑选出最有效的特征子集,达到降维、简化辨识模型的目的,建立了基于温度与压力变化速率的燃爆辨识模型。实验结果表明,该方法有效提高了燃爆测试装置的鲁棒性和辨识的置信概率,燃爆辨识准确率从传统方式的95.56%提高到99.34%。

Deflagration identification is a key technology to obtain material physical and chemical parameters and physical hazards of chemicals. Aiming at the problems that the deflagration has complicated variable status and single sensor system can not obtain satisfied identification effect, the multi-sensor joint identification is introduced, and a new method based on the rough set theory is adopted to fuse the data of multi-sensors. In this paper, the inherent correlation among the deflagration parameters are analyzed, a joint deflagration identification device based on multi-sensors is designed. Starting from the decision table formed from the deflagration test data, the com- plete reduced rule set of the rough set theory is adopted to reduce the attribute table and the core value table. Under the promise of ensuring the accuracy of deflagration identification, the most effective characteristic subset is selected, and the goal of reducing dimensions of decision table and simplifying the identification model is achieved. A deflagration identification model based on the change rates of temperature and pressure was established; and experiment results show that this method effectively increases the robustness of the deflagration test device and the confidence probability of the identification. The deflagrating identification accuracy is improved from 95.56% obtained with traditional method up to 99.34%.