METHOD FOR INTERPOLATION AND EXTRAPOLATION OF DATA WITH UNEVEN STEP OF CHANGING EXPERIMENTAL PARAMETERS

Abstract

The problem with using experimental data is the limited range and uneven measurement step for individual variables. This reduces the accuracy of determining the value of the desired function. The method of further research is to increase the number of values of the sought function based on limited experimental data. The research method involves developing a methodology for interpolation and extrapolation of experimental data using the functions of the Matlab package. The result of the investigation is that the technique allows you to create a uniform grid of reference values, fill it with the required step of changing variables, extrapolate data to the boundaries of the data space, where the influence of a given argument on the function is zero, isolate the contribution of individual arguments to the value of the function, synthesize the value of the function in new nodes reference grid of values The application of a standardized method was introduced to form a space of values of the relative rate of aging of cellulose insulation of power transformer windings as a function of temperature, humidity and oxygen access, which made it possible to increase the amount of data at least 4 times compared to the experiment.