THE USE OF INVERSE PROBLEM NUMBERICAL SOLUTION TO IMPROVE THE BALLAST LAYER USING CONTROL BY LIGHT WEIGHT DEFLECTOMETER

The article describes the light weight deflectometer operation principle, which used to access the elastic characteristics of soil under impact load. The study proposes the use of numerical solutions to the inverse problem to investigate the influence of installation parameters and foundations on measurement results. The solution is based on optimizing the parameters of the developed mathematical model. The criterion for minimization was the mean square error between the measured values of acceleration, velocity, and displacement and the calculated values obtained from the mathematical model. Due to the presence of multiple local minima in the optimized function, a genetic algorithm was used as the solver Tests were conducted to measure the elasticity coefficients on bases with different stiffness and deformation modulus on gravel ballast. The test results confirm the applicability of using a numerical inverse problem-solving algorithm to find unknown system parameters based on experimental data.