Optimal Sensor Placement Method Based On Optimal System Reliability Criterion

Harsh working environment usually leads to the frequent failures of modern systems.The reliability of these systems can be greatly improved by using the sensor to monitor the working state.Due to structure and cost restrictions,quantity of sensors is limited in the monitoring system.It is important to focus on the effectively monitoring of the key components in the optimization of sensor placement.High reliability is the guarantee for the safe and long-term operation of these systems.Therefore,it is feasible to introduce reliability as a criterion for the optimal sensor placement.Taking the system reliability as the optimization criterion,the first step is to build the system fault model and analyze the system reliability.Application of some redundant technologies in these systems has greatly improved their performance.However,fault characteristics such as epistemic uncertainty,dynamic failure modes and common cause failure occur when these systems break down,which raises some new challenges in fault diagnosis.Dynamic fault tree is used to describe the dynamic failure behaviors.Furthermore,some reliability parameters are calculated by mapping a dynamic fault tree into a dynamic Bayesian network and a dynamic evidential network,which can deal with common cause failure and propose a solution for the dynamic fault tree with interval distribution parameters of the components.For the optimal sensor placement,it is necessary to obtain the possible placement scheme of the sensor,so the potential locations of sensors must be determined first.To this end,this paper proposes the method for determining the potential locations of sensors based on DIF and TOPSIS algorithm when the component failure rate is a certain value,and the method for determining the potential locations of sensors based on DIF and VIKOR algorithm under epistemic uncertainty.These methods use reliability parameters as component evaluation attributes to rank the importance of components to achieve the effective allocation of sensors.In order to consider the influence of sensors reliability on the system,sensors are supposed to be a system component,and a diagnostic sensor model is developed by using a priority AND gate.The priority AND gate can not only connect sensors and components into the system in parallel,but also take into account the failure sequence between a sensor and a component.In addition,when the number and the potential locations of sensors is determined,all placement schemes can be enumerated.System reliability of the top event in each placement scheme is calculated,and the largest reliability of the top event is the optimal scheme.Finally,a case study is used to prove the effectiveness of the proposed method.