Modification of the adaptive algorithm for solving the twodimensional problem of charge distribution in photovoltaic converters with the existence of localized bands with large values of target function gradient

The aim is to build a modified adaptive algorithm forsolving two-dimensional problems in modeling the chargedistribution of photovoltaic energy converters. In case of theexistence of localized bands with large values of a target functiongradient adaptive algorithm work will require significantthicke...

Повний опис

Збережено в:
Бібліографічні деталі
Дата:2017
Автори: Reztsov, V., Matyakh, S.
Формат: Стаття
Мова:Українська
Опубліковано: Institute of Renewable Energy National Academy of Sciences of Ukraine 2017
Теми:
Онлайн доступ:https://ve.org.ua/index.php/journal/article/view/34
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:Vidnovluvana energetika
Завантажити файл: Pdf

Репозитарії

Vidnovluvana energetika
Опис
Резюме:The aim is to build a modified adaptive algorithm forsolving two-dimensional problems in modeling the chargedistribution of photovoltaic energy converters. In case of theexistence of localized bands with large values of a target functiongradient adaptive algorithm work will require significantthickening of the grid to achieve a given accuracy of the calculationin these areas, which will significantly increase the number ofgrid nodes in neighboring areas because of restrictions on thedifference between the values of neighboring steps. However, ifthe target function in areas adjacent to areas with large gradientsare pretty smooth and has low gradient values, the algorithm willhave to spend considerably more computing resources thannecessary for a given accuracy. The extra amount of time andcomputing resources will be particularly great in a situationwhere high gradient areas are localized in the calculation areaand occupy a small part of it.For this case modification of an adaptive algorithm isproposed. During the transition to a new time layer algorithmdetermines the presence of localized areas with high gradients oftarget function, using a numerical method for a transition to a newtime layer calculation in every localized area separately to reducethe dimension of the global grid and reduce computing costs.A modified version of the adaptive algorithm identifiesareas of high gradient define the boundary conditions for theseareas from the results, where the error is acceptable and leadsnumerical method recursive launch separately for every selectedarea, which allows significant thickening of the grid to calculatewith high accuracy, where necessary, but without affecting thedistribution of nodes in neighboring areas of the global grid. Thisthe modification allows optimizing the spend of time andcomputational resources in numerical simulation of chargetransport in photovoltaic energy converters.