A dynamic zonal method based on KD-tree for calibration of five-hole probe under large flow angles

With the development of aviation engines, the range of flow angles that the five-hole probe needs to measure in experiments has continuously increased. To reduce measurement errors and operational costs at large flow angles, a dynamic zonal calibration method based on KD-tree is developed using coarse-to-fine search strategy. This method addresses the inner boundary measurement issue and low-linearity in-put problem. It first performs a coarse search of the test sample using 4D calibration coefficients and the KD-tree, dynamically dividing the test point into the calibration neighborhood. Then, low-linearity data is discarded based on the maximum pressure, and 2D calibration coefficients from the zonal method are used to precisely predict flow parameters. Verification tests at 0.15 Ma and 0.3 Ma (α ± 50°, β ± 40°) show that the new method achieves flow angle prediction absolute errors below 1° and total pressure relative errors below 0.2 %, comparable to the accuracy of complex neural networks, without requiring additional data training models.