International Scientific Journals
of Scientific Technical Union of Mechanical Engineering "Industry 4.0"

Radial Basis Function (RBF) interpolation is widely used for reconstructing scattered multidimensional data, but its accuracy and stability often depend heavily on the point distribution, kernel choice, and matrix conditioning. This work proposes an improved RBF framework that combines a whitening transformation, adaptive sampling, and cross-validated parameter tuning to obtain more reliable reconstructions of complex two-dimensional fields. Whitening reduces geometric distortions and improves conditioning, while the adaptive sampling strategy focuses points in regions with strong gradients. Tests on a deliberately challenging synthetic function show that the whitening-enhanced RBF model achieves higher stability and lower error (in L1, L2, and L∞ norms) compared to the standard RBF formulation. The method is implemented in Python using open-source libraries and performs efficiently even on modest hardware.

This paper presents a method for predicting the performance of compressors in HVAC/R systems using 2-D quadratic spline interpolation. Compared to traditional third-order polynomial interpolation, the spline model provides smoother and more accurate approximations, especially useful when limited manufacturer data is available. Using real-world compressor data, we constructed mathematical models and evaluated their performance through key statistical indicators. Results indicate that the spline-based model yields lower RMSE values with reduced model order, improving its integration in system simulations. This study also provides a comparative analysis against a polynomial-based model developed in earlier research.