Effective sensor placement in a steam reformer using gappy proper orthogonal decomposition

In this study, a gappy proper orthogonal decomposition (POD) method was applied to a steam reformer model with combustion, flow, and catalysis to determine the optimal number and placement of sensors. The dominant POD modes were identified based on a limited number of snapshots obtained from a spatial domain simulation, and the POD modal content was calculated from the corresponding gappy data. This information was used to estimate the differences between the sensor measurements and actual fields. The estimation results were utilized to verify the accuracies of gappy POD projections of 20 snapshots of the positions of six, three, and two sensors, and the sensor arrangements determined by using a proposed objective function were compared to those resulting from applying the conventional method. In addition, reconstructions based on gappy data were evaluated in four validation cases, and the accuracy and robustness of the sensor positions in various situations were confirmed. Consequently, this paper optimized sensor placement for the steam reformer in terms of temperature prediction and proposed modified the objective function which maintains orthogonality of the mask matrix.