جستجوی مقالات مرتبط با کلیدواژه "پیش بینی قیمت" در نشریات گروه "برق"

Today, the electricity market in the world is known scientifically that the competition in it is more every day than the previous day. Since the ability to store electrical energy is very small, therefore, forecasting the consumption load and the price of electricity helps the market participants to get more profit. The impact of the load pattern on various factors and the non-linearity of the electricity price signal make it difficult to accurately forecast the load and price; Therefore, the use of intelligent algorithms has found more use in forecasting problems compared to numerical and statistical methods. Therefore, in this thesis, the issues related to forecasting the load and electricity price are stated. Also, the electric load and electricity price have been predicted using the Adaptive neuro fuzzy inference system (ANFIS) combined with the Teaching Learning based Optimization (TLBO) and the effect of various factors on it has been investigated and simulated. In fact, by combining the evolutionary algorithms with the fuzzy neural system, the adjustment of the optimal values of the parameters of the adaptive fuzzy neural network should be assigned to the intelligent optimization algorithm of teaching and learning. The purpose of using this approach is to improve network performance and reduce computational complexity compared to gradient descent and least squares methods. The results of the implementation of the proposed algorithm show the better efficiency of this algorithm compared to previous algorithms for predicting load and electricity price.

The electricity market in the world today is scientifically recognized and it is becoming more competitive every day. In the meantime, the price forecasting tool helps market participants gain more profit. The purpose of the present study is to extend the neural network and optimize it by gray wolf algorithm to predict electricity market price. Since traditional and neural network models have always used probabilistic methods to increase prediction accuracy, this model has tried to ignore these by introducing a new method to require less time for prediction. One of the issues that always affects the prediction accuracy is the existence of critical and sudden cases in the system, which will be controlled in this research using efficient methods. The proposed model uses Nordic electricity market information, but it is obviously applicable to any other market. Finally, to illustrate the accuracy of this model, a comparison is made between it and one of the traditional methods.

In recent years, the exploitation of the energy storage systems has attracted many researchers in the electricity industry. Due to the relatively high cost of storage, usually many investors are not being prevented from installing and operating in a power system. In this paper, the effect of the type of operation of the energy storage system (the whole year or part of the year) on the profit of the energy arbitrage (shift) in the electricity market is investigated by an electrical energy storage system. For Evaluation the effect of the price prediction error on the energy arbitrage revenue, two types of prices (real price, 3 Hour Ahead Predispatch Prices) have been used. Due to the encouragement of private investors, it is proposed to use the loan for the initial investment cost so that the proposed plan for the investor is economically justified. Due to high investment costs and longer lifespan of the energy storage system, the number of charging and discharging energy is one, which is the time of charging at the lowest cost of energy and the time of discharge at the highest energy cost to increase energy returns.

Forecasting the electricity price is important for electricity market players. Time series of the electricity price -as an inherently random phenomenon- have high uncertainty relative to the load. On the other hand, the non-stationary and non-linear characteristics of this time series make its forecasting difficult. On markets like the stock market, one can somehow forecast future price movements using technical analyses along with testing past prices and the volume of transactions. Therefore, this paper uses technical analysis indices for analyzing the time series data of the electricity market to forecast the electricity price. These indices are used as features extracted from time series of electricity price and applied to a Support Vector Machine (SVM) regression, through which the electricity price is predicted on daily horizon on Ontario electricity market.