Optimal design of fault tolerant sensor networks

The selection of measurements is one of the most important problems in the design of process instrumentation This paper is concerned with the design of sensor networks such that the observability of the variables which are necessary for the process control remains satisfied in the presence of sensor failures. Pseudo-minimal and minimal sensor sets are organized into an oriented graph which contains all the possible reconfiguration paths for which those variables remain observable. A bottom-up analysis of this graph allows to compute reliability functions which evaluate the robustness of the observability property with respect to sensor failures. The design of optimal sensor networks thus resumes to finding pseudominimal sensor sets such that the mean time before loosing the observability property is larger than a predefined value.