COURSE OVERVIEW

 

This practical and highly-interactive course includes various practical sessions and exercises. Theory learnt will be applied using our state-of-the-art simulators. 

 

Planning the operation in modern power systems requires suitable anticipation of load evolution at different levels of distribution network. Under this perspective, load forecasting performs an important task, allowing optimization of investments and the adequate exploitation of existing distribution networks. 

 

Load forecasting is an essential element of power system involving prognosis of the future level of demand to serve as the basis for the supply side and demand side planning. The load requirements are to be predicted in advance so that the power system operates effectively and efficiently. It is done for planning, marketing, risk assessment, billing, dispatch or unit commitment purposes. 

 

Further, Load forecasting can help to estimate load flows and to make decisions that can prevent overloading. Timely implementations of such decisions lead to the improvement of network reliability and to the reduced occurrences of equipment failures and blackouts. 

 

 

 

                                                                                                                      

Load forecasting may be applied in the long, medium, short, and very short term time scale. Short-term forecasts (five minutes to one week ahead) are required to ensure system stability. Medium term forecasts (one week to six months ahead) are required for maintenance scheduling, while long term forecasts (six months to 10 years ahead) are required for capital planning. 

 

The long and medium term forecasting are used to determine the capacity of generation, transmission or distribution system additions and the type of facilities required in transmission expansion planning, annual maintenance scheduling, etc. The short-term load forecast is needed for control and scheduling of power system and also as inputs to load flow study or contingency analysis. The purpose of very short-term load forecasting (ranging from minutes to hours) is for real time control and security evaluation. 

 

Economic and reliable operation of an electric utility depends to a significant extent on the accuracy of the load forecast. The load dispatcher at main dispatch center must anticipate the load pattern well in advance so as to have sufficient generation to meet the customer requirements. Overestimation may cause the startup of too many generating units and lead to an unnecessary increase in the reserve and the operating costs. Underestimation of the load forecasts results in failure to provide the required spinning and standby reserve and stability to the system, which may lead into collapse of the power system network. Load forecast errors can yield suboptimal unit commitment decisions. 

 

Therefore, to reduce exposure risks, an accurate forecast is required. While system loads are predictable, they require analysis of many variables including day of the week, holidays, historical load patterns, and weather. Loads can react differently to the same weather conditions during different times of the year. Energy schedulers, portfolio managers, and grid security analysts need to understand how the load responds to weather changes during different times of the year, times of the day, and days of the week. 

 

Different forecasting models have been employed in power systems for achieving forecasting accuracy. Among the models are regression, statistical and spatial methods. In addition, artificial intelligence-based algorithms have been introduced based on expert system, evolutionary programming, fuzzy system, artificial neural network (ANN), and a combination of these algorithms. Among these algorithms, ANN has received more attention because of its clear model, easy implementation, and good performance. Most forecasting models and methods have already been tried out on load forecasting, with varying degrees of success. They may be classified into two broad categories: artificial intelligence based techniques and classical (or statistical) approaches. 

 

The former include expert systems, fuzzy inference, fuzzy neural models, and, in particular, artificial neural networks (ANN). The statistical methods differ from the previous approach in that they forecast the current value of a variable by using an explicit mathematical combination of the previous values of that variable and, possibly, previous values of exogenous factors (specially weather and social variables). Models that have been applied recently include autoregressive (AR) models, linear regression models, dynamic linear or nonlinear models, ARMAX models, threshold AR models, methods based on Kalman filtering, optimization techniques, and curve fitting procedures. The statistical models are attractive because some physical interpretation may be attached to their components, allowing engineers and system operators to understand their behavior. At the same time they offer relatively good performance.