Aim: Is developing of short-term load forecasting math model of the electrical engineering complex of the district regional electric grid 6-35 kV with the use of artificial neural networks.
Methods: The tools of regression analysis and deep machine learning were used in the work.
Results: The neural network model for short-term load forecasting of the electrical engineering complex of section regional electric grid 6-35 kV, which considered factors of time, meteorological conditions, disconnections of individual power transmission lines, the operation mode of electricity consumers with a capacity of over 670 kW, the fact of the availability of central heating and water supply, has been obtained.
Conclusion: The developed neural network math model reduces the problem of short-term load forecasting to the search of matrix free coefficients through training on the available statistical data.