SPARSE RECONSTRUCTION OF SURFACE LOADS ON AIRCRAFT USING POD AND RBFNN

loads based on the measurements from sensors in wind tunnel test or during flight are crucial for aerodynamic design, health monitoring and flight control. However, obtaining the complete surface loads in real-time is challenging due to the limitations of available measurements. To handle this problem, we propose a sparse reconstruction modeling method based on proper orthogonal decomposition (POD) and radial basis function neural network (RBFNN). This method leverages the dimensionality reduction capabilities of POD to extract the dominant modes of surface loads, and then use RBFNN to accurately predict the mode coefficients based on sparse measurements. Furthermore, to improve the accuracy of sparse reconstruction, we have developed an objective function that integrates the surface loads in optimizing the sensor locations. The case study of the DLR-F6 aircraft shows that a weight of 0.25 for the objective function provides the best reconstruction performance. Additionally, a minimum of 240 modeling samples is required to ensure the accuracy of the sparse reconstruction model. Specifically, with just 37 sensors, we can achieve real-time sparse reconstruction of surface pressure coefficients.