Abstract
We propose a novel optimal sensor placement determination method for jet engines. In jet engine health monitoring, fault detection with a state estimation is frequently carried out. A straightforward approach to obtain accurate estimations is to directly observe all the physical quantities of interest. However, typically, only a subset of them is observable. Moreover, in practice, the number of sensors is often limited due to cost and weight. Therefore, optimizing sensor placement is an important problem. More specifically, we solve a problem to maximize the estimation accuracy by optimally choosing a given number of physical quantities that are measured by sensors. We validate the proposed method by extensive simulations.
