Lukyanenko S. Adaptive methods of numerical modelling highly-gradient processes in objects with the allocated parameters

The dissertation is devoted to the decision of a scientific-applied problem of increase of efficiency of modelling objects with the allocated parameters at the expense of an optimum choice of junction of a differential grid. It is offered the new approach to the automated construction replaceable non-uniform differential grids adapted for dynamics of change of required function which is based on use of the enclosed grids and allows to reduce significally dimension of systems differential the equations and time of modelling. The way of a rating of an error of the decision which simultaneously with a margin error defines its sensitivity change of a configuration differential grids which reduces calculation to definition of character of its transformation is offered and proved. The way of definition of parameters of transformation of a grid is offered and proved on the basis of the data on sensitivity of an error to sizes of steps which allows to avoid decisions auxiliary differential equations and toreduce expenses of machine time for construction of a new grid.