Оцінювання імовірнісних характеристик генерування сонячних електростанцій в задачі інтелектуалізації локальних електричних систем

Abstract

В роботі розглядається питання аналізу імовірнісних характеристик генерування відновлювальних джерел електроенергії, зокрема сонячних електричних станцій (СЕС). Проаналізовано закон розподілу генерування сонячної електростанції для певного проміжку часу доби протягом року. Представлена математична модель гаусових сумішей, що дозволяють отримати основні характеристики імовірнісного характеру генерування сонячної електростанції. Оцінено можливість узгодження генерування сонячної електростанції з добовими графіками споживання. Визначено імовірність покриття заданого графіка навантаження генеруванням СЕС протягом певного часу доби впродовж всього року

Growth of the share of renewable sources of energy in electric energy balance of the country is stipulated both by state preferences and certain reduction of the prices at the world market for the elements of such sources. Solar power plants are developing at a rapid pace. Some consumers start to construct plants of direct conversion of solar energy as a means to provide certain level of energy independence. These consumers leave the group of “passive” and join the group of “active” users and are able to influence the operation mode of local energy system. In new conditions the coordination of the work of active consumers, sources of distributed generation and distributive electric network is impossible without intellectualization of the process of generation, distribution and consumption on the level of local electric system , using Smart Grid technologies. Greater part of renewable sources of energy is represented by solar power plants (SPP). Their operation mode depends on natural conditions of the region, where they are located. That is why the problem of SPP stability estimation is important in the problem of covering the daily schedule of energy consumption. The aim of the research is to evaluate the probability of providing the consumption schedule by means of corresponding source on the base of statistic data. The authors try to solve the given problem, using EM algorithm. The authors could not select the law of SPP generation distribution, applying direct analysis of statistical data, that would satisfy the character of data change, it is suggested to divide the data into the series of mixtures, each of which would correspond to normal law. EM algorithm enables to select optimal, by the criterion of maximum of probability, amount of mixtures. By the values of characteristics of these mixtures the probabilities of emergence both of certain generation and consumed power could be obtained. To reduce the negative impact of random factors, characteristic for certain periods of the day, statistic data were considered by half hour periods of the day. As a result, the authors succeeded in determining the general coefficient of source stability. Probabilities of covering the possible load in determined period of the day are defined. It enables further to evaluate necessary volume of reserve and start-stop characteristics of energy sources that can be used as the reserve..