化工过程监测系统的传感器网络设置问题研究

Study of Sensor Network Locations for Chemical Process Monitoring System

化工过程监测系统的性能与其传感器网络的设置方式有直接联系。通过建立化工过程有向图模型,结合故障传播模式分析和传感器网络搜索算法,从过程监测的角度研究了传感器设置问题。给出了考虑故障可观性与分辨率下的传感器设置方法。并以CSTR-heatexchanger过程为例,说明了统计过程监测方法与传感器网络设置问题相结合的重要性。

The effectiveness of on-line monitoring system for chemical processes is strongly concerned with the location of sensors which are used to collect fault information. However, most researches in the field of statistical process monitoring (SPM) focus on improving the performance of the monitoring algorithm for a given set of sensors. By using the digraph model of chemical process, the issue of sensor network locations (SNL) in the monitoring system was explored. With the consideration of characteristics of SPM algorithms, a new approach to the SNL issue was presented by a way different from the conventional methods, and both fault observability and resolution were guaranteed. The validity of the proposed approach was verified by monitoring a simulated CSTR-heat exchanger process. The importance of the integration of SPC and SNL was highlighted.