Daily Peak Load Research Articles

タグチのT法を用いた翌日最大電力需要予測

In this paper, a new technique of daily peak load forecasting is proposed. At first, we propose Non-linear correction T method as improvement technique of T method that is one of the multivariate analysis techniques, called MT system, suggested in Quality Control. Moreover, we examine the daily peak load forecasting technique, which is based on data in the last year on a forecasting day, by T method and Non-linear correction T method. Furthermore, we examine the technique, which is based on the data of most recent several weeks of a forecasting day, by the T method and Non-linear correction T method. In addition, the effectiveness of the proposed method by comparing the multiple regression analysis is discussed.

Development of Low Voltage Network Templates—Part II: Peak Load Estimation by Clusterwise Regression

This paper proposes a novel contribution factor (CF) approach to predict diversified daily peak load of low voltage (LV) substations. The CF for each LV template developed in part I of the paper is determined by a novel method—clusterwise weighted constrained regression (CWCR). It takes into account the contribution from different customer classes to substation peaks, respecting the natural difference in time and magnitude between LV substation peaks and the variance within the templates. In CWCR, intercept and coefficients are constrained to ensure that the resultant coefficients do not lead to reverse load flow and can respect zero-load substations. Cross validation is developed to validate the stability of the proposed method and prevent over fitting. The proposed method shows significant improvement in the accuracy of peak estimation over the current status quo across 800 substations of different mixes of domestic, industrial and commercial (I&C) customers. The work in the two parts of the paper is particularly useful for understanding the capabilities of LV networks to accommodate the increasing penetration of low carbon technologies without large-scale monitoring.

Modelling Summer Daily Peak Loads in South Africa Using Discrete Time Markov Chain

Electricity demand exhibits a large degree of randomness in South Africa, particularly in summer. Its description requires a detailed analysis using statistical methodologies, in particular stochastic processes. The paper presents a Markov chain analysis of peak electricity demand. The data used is from South Africa's power utility company Eskom, for the period 2000 to 2011. This modelling approach is important to decision makers in the electricity sector particularly in scheduling maintenance and refurbishments of power-plants. The randomness effect is accountable to meteorological factors and major electricity appliance usage. Aggregated data on daily electricity peak demand is used to develop the transition probability matrices, steady-state probabilities, mean return- and the first passage times. Such analysis are important to Eskom and other energy companies in planning load-shifting, load analysis and scheduling of electricity particularly during peak period in summer.

非線形補正T法を用いた翌日最大電力需要予測

非線形補正T法を用いた翌日最大電力需要予測

Modelling and forecasting of electric daily peak load movement based on the elliptic‐orbit model with weekly periodic extension: a case study

Electric load movement analysis is an important task for effective operation and planning of power systems. It is shown that a weekly quasi-periodicity is enclosed in electric daily peak load movement. On the basis of weekly periodic extension for electric daily peak load movement, the so-called elliptic-orbit model is introduced to describe its movement. Electric daily peak load movement as time series is mapped into the polar coordinates to form the elliptic-orbit model, in which each 7-day-movement is depicted as one elliptic orbit. Experimental results of the Great Britain National Grid and Analysis indicate workability and effectiveness of the proposed method. It is shown that the electric daily peak load movement with weekly periodic extension is well described by the elliptic-orbit model, which presents a vivid description for analysing electric daily peak load movement in a concise and intuitive way, and others may benefit from it.

계절성과 온도를 고려한 일별 최대 전력 수요 예측 연구

계절성과 온도를 고려한 일별 최대 전력 수요 예측 연구

A Study on the Summer and Winter Load Forecasting by using the Characteristics of Temperature Changes in Korean Power System

– In Korean power system, the daily load variation is in accordance with the temperature changes in summer and winter time. However, it is not in accordance with temperature changes in spring and autumn time. The cooling load in summer and the heating load in winter contribute to the increasing of the load. Daily temperature changes will cause the load changes. Therefore, the temperature sensitivity is very important in improving the accuracy of the load forecasting during the summer and winter. This paper proposes the summer and winter load forecasting by using the characteristics of temperature changes in Korean power system. Based on daily peak load of summer and winter, the relationships between characteristics of temperature changes and daily peak load are analyzed in this paper. The characteristics of temperature changes are obtained by using daily peak load and high temperature data in summer and low temperature data in winter during eight years from 2001 to 2008. The sensitivity of daily load in accordance with the increase or decrease in temperature are calculated, respectively. Summer and winter load are forecasted based on the sensitivity between unit temperature changes and daily load variations. Load forecast data with better accuracy are obtained by using the proposed sensitivity than by using the ordinary daily peak load.

The Balance of Power System Peak Load Regulation Considering the Participation of Nuclear Power Plant

Nuclear power plants (NPP) have to face urgent requirement of participating in peak load regulation of power grid. The peak load regulation performances of NPP were researched. Based on the calculation and analysis on peak load regulation cost and benefit of NPP, the location, criterion and peak balance algorithm of NPP participating in peak regulation were proposed. Results of the actual research show that NPP possesses better ability of peak load regulation and the cost is lower, so NPP can participate in daily peak load regulation, which will improve both security and flexibility of power grid and enhance the efficiency of resource configuration. However, the depth and speed of peak load regulation by NPP are restricted by security and economy.

Feature selection for daily peak load forecasting using a neuro-fuzzy system

Accurate electrical daily peak load forecasting (DPLF) is essential for power system management in order to prevent overloading and grid failure. Fuzzy neural networks have been successfully applied to load forecasting due to their nonlinear mapping and generalized behavior. In this paper, a neuro-fuzzy based DPLF (N-DPLF) model with a feature selection method is proposed for DPLF. The load data is clustered into seven subsets according to the season and day type. For each subset, the four features with the highest salience ranks are selected. After training N-DPLF model, the formed BSWs (bounded sum of weighted fuzzy membership functions) in accordance with the selected features denote characteristics of these features. The N-DPLF model provides explicit BSWs in hyperboxes, instead of the uncertain black box nature of neural network models, so that the selected features can be interpreted by the visually constructed BSWs. The N-DPLF model with a feature selection method shows a mean absolute percentage error (MAPE) of 1.86 % using Korea Power Exchange data over 1-year period.

Real‐time scheduling of electric vehicles charging in low‐voltage residential distribution systems to minimise power losses and improve voltage profile

Unscheduled charging of plenty of electric vehicles (EVs) might exert an adverse effect on the existing power grid, especially when the charging coincides with daily peak load at distribution level. In this study, a scalable real-time scheduling scheme for EV charging in low-voltage residential distribution system is proposed. Since most often, voltage drop would become a binding constraint when a distribution feeder is subject to a high EV penetration level, a scheduling method is first developed to increase the voltage safety margin. Then, a novel factor is derived to allow the scheduling to shift from voltage-safety-oriented to loss-minimisation-oriented, or vice versa, on demand of the EV penetration level. A number of charging scenarios were simulated to evaluate the proposed scheduling scheme. Simulation results verified that the proposed scheduling scheme is fast and efficient with circuit losses close to optimal at a low EV penetration level and voltage drops maintained within the tolerable limit at a high EV penetration level. The high scalability and effectiveness of the proposed scheme has made it suitable for coordinating large number of EV charging activities in real-time.

Daily Nigerian peak load forecasting using artificial neural network with seasonal indices

A daily peak load forecasting technique that uses artificial neural network with seasonal indices is presented in this paper. A neural network of relatively smaller size than the main prediction network is used to predict the daily peak load for a period of one year over which the actual daily load data are available using one step ahead prediction. Daily seasonal indices are calculated as a ratio of the predicted load to the actual load. The i th index is used as an additional input to a main network that predicts the load for the i th day of the year following the one for which the indices were computed. Both neural networks are trained by the back propagation algorithm. The technique is illustrated with data derived from the Nigerian national electric power system. Results obtained are good enough to meet the requirements of practical systems and show appreciable improvement over the normal one step ahead prediction with neural network. Keywords: load forecasting, neural network, seasonal indices, back propagation, actual load

Advanced natural gas storage system and verification tests of lined rock cavern - ANGAS project in Japan -

Technologies of an underground natural gas storage system, a Lined Rock Cavern (LRC) gas storage system called ANGAS (Advanced Natural GAs Storage), have been studied. The purpose of the project is to develop a suitable LRC system for Japan, and to contribute to expanding use of natural gas. It is necessary to study measures to shave off the daily peak load of natural gas through pipelines and it is expected that the ANGAS is one of key countermeasures for peak shaving. In the project, our goal is to ensure the design method of the LRC system under the Japanese geological conditions. One of the key technologies of the developed design method is based on plastic deformation of the steel liner considering the rock deformation appropriately. Using the developed design method and procedures, a small test cavern which is located at the Kamioka mine, the midland of Japan, was designed and constructed. The test cavern is surrounded by a sedimentary rock mass which consists of mainly sandstone and mudstone. The inner design pressure of the test cavern is 20MPa. In this paper, some results of the verification tests are presented and the validity of the LRC system for Japan is confirmed. Specifically, the steel liner was deformed up to its plastic zone, in order to be unified with the buffer material and the backfilled concrete as a hybrid structure. The displacements in the surrounding rock mass were stable under the inner design pressure 20MPa. Furthermore, the real crack intervals of the backfilled concrete are recognized to be less than the designed maximum interval and the real maximum crack width is inferred to be close to the designed one under 20 MPa pressure. Therefore, the design considering the crack properties of the back filled concrete has been verified.

A hybrid model for daily peak load power forecasting based on SAMBA and neural network

According to the significance of power load demand forecasting, this paper suggests a new hybrid method to reach more accurate model with fast response. The proposed model consists of two algorithms: Self Adaptive Modified Bat Algorithm (SAMBA) and Artificial Neural Network (ANN). In recent years, SAMBA has been used as a powerful tool in the optimization problems. On the other hand among the most popular methods, ANN has shown powerful performance in load prediction as the result of its ability to detect nonlinear mappings among different variables. In addition, the special ability of SAMBA in fast convergence, its low dependency to setting parameters and simple implementation make this algorithm more premiere than the other optimization algorithms. Therefore, in this paper for the first time we use SAMBA to regulate the weight matrix of ANN and optimize the degree of uncertainty which exist in load demand prediction.

A hybrid intelligent-based linear-nonlinear model for accurate daily peak load forecasting

The objective of this work is to suggest a new hybrid intelligent-based linear-nonlinear model to capture both the linearity and nonlinearity of load time series to reach more accurate forecasting results. In the proposed method, the autoregressive integrated moving average (ARIMA) model is used to forecast the linear part of the time series. After that, the ARIMA residuals as the nonlinear component are modeled by the support vector regression (SVR) forecaster. In order to reduce the nonlinearity of the residuals, the discrete wavelet transform is used to decompose the ARIMA residuals into its high and low frequency components. Moreover, a new adaptive modified optimization tool based on particle swarm optimization (PSO) algorithm is proposed to find the optimal values of the SVR parameters suitably. The proposed adaptive modified PSO algorithm makes use of an adaptive framework to update the inertia weighting factor and the acceleration coefficients during the optimization process. The accuracy of the proposed method is examined by the empirical load data of Fars Electric Power Company, Iran.

Short term load forecasting of distribution systems by a new hybrid modified FA-backpropagation method

Distribution systems as the final link between the production side and the consumers engage the most contributions to the unavailability of electrical services. In fact, in the operation of distribution systems, forecasting the short term load is a precious and critical task. With a more accurate load forecasting, the distribution system operation management would be improved and more economical trade off with the electricity market can be achieved. However, as the result of high nonlinearity and variation of the loads in distribution systems, short term load forecasting in these systems is hard and complex. In this regard, this paper proposes a new hybrid method based on firefly algorithm FA and artificial neural network ANN to reach a more reliable and accurate forecasting model. The proposed method makes use of both the learning ability of ANN and the powerful search ability of FA to create a nonlinear mapping between the input and output load pattern data. In contrast to the other evolutionary based on ANN training methods, this work preserves a good balance between ANN traditional training techniques such as back-propagation method and evolutionary random search ability of FA in a hybrid framework. Meanwhile, a new sufficient two-stage modification method is proposed for FA to improve its ability in both the local and global searches. The feasibility and satisfying performance of the proposed method is examined on the practical daily peak load of a part of Shiraz distribution system, Iran.

A Comparison between forecasting methods of daily peak load series for electrical power in Mosul

Time series analysis gives a great importance to predict and many of the methods proposed for the purpose of forecasting. In view of the importance of accession deviation it was different of the goal of this research for a comparison between the Box-Jenkins method and Exponential Smoothing methods and compare the standard based on the mean square error(MSE) and mean absolute error (MAE) and mean absolute percentage error (MAPE), to assume the future values of the seasonal time series models and the model of suggested and fitting for time series data is ARIMA (0,1,2)(0,1,1) and has been applied to the observations represent daily of peak load for electric power in Mosul.

시계열 모형을 이용한 일별 최대 전력 수요 예측 연구

최근 일별 최대 전력수요 예측은 전력설비 계획 및 운용에 매우 중요한 사안으로 주목받고 있다. 본 연구는 일별 최대 전력수요 예측을 위하여 대표적 시계열 모형을 소개하고, 예측의 성능 비교를 위하여 RMSE(Root mean squared error)와 MAPE(Mean absolute percentage error)를 사용한다. 연구결과로 보완된 Holt-Winters 모형과 Reg-ARIMA 모형이 다른 모형에 비하여 우수한 예측 성능을 보였다. Forecasting the daily peak load for electricity demand is an important issue for future power plants and power management. We first introduce several time series models to predict the peak load for electricity demand and then compare the performance of models under the RMSE(root mean squared error) and MAPE(mean absolute percentage error) criteria.

The Temperature Sensitivity of the Commercial Load in KOREA

This paper presents the temperature impacts on daily peak load of commercial load in KOREA. This research focuses on the overall summer and winter peak load characteristics and the impact of temperature change on the commercial load. Load data of the last five years are used to analyze the temperature sensitivity of commercial load sensitivity.

Daily Electric Load Forecasting Based on RBF Neural Network Models

This paper presents a method of improving the performance of a day-ahead 24-h load curve and peak load forecasting. The next-day load curve is forecasted using radial basis function (RBF) neural network models built using the best design parameters. To improve the forecasting accuracy, the load curve forecasted using the RBF network models is corrected by the weighted sum of both the error of the current prediction and the change in the errors between the current and the previous prediction. The optimal weights (called “gains” in the error correction) are identified by differential evolution. The peak load forecasted by the RBF network models is also corrected by combining the load curve outputs of the RBF models by linear addition with 24 coefficients. The optimal coefficients for reducing both the forecasting mean absolute percent error (MAPE) and the sum of errors are also identified using differential evolution. The proposed models are trained and tested using four years of hourly load data obtained from the Korea Power Exchange. Simulation results reveal satisfactory forecasts: 1.230% MAPE for daily peak load and 1.128% MAPE for daily load curve.

Adaptive Classification Algorithm for Concept Drifting Electricity Pricing Data Streams

Electricity is the main observation in our daily life. There are many parameters or a factors on which the electricity load is depends on, knowable load factors such as whether conditions, temporal factors, and customer characteristics etc. Daily peak load is an important factor in the planning the production and pricing of electricity. In a simple terms, it is essential to get the knowledge of the local system demand will be on the next minutes, hours and days so that the generators with various startup times, startup cost can be changes as per the requirement and knowledge gain from the previous data collected. This paper is intended for industry/ organization to optimize Electricity usage. Energy consumption and pricing analysis is a primary area in power systems planning and management. Recent developments in energy market deregulation and provision of sustainable energy have contributed to increase interest in this area. The prices are not fixed and are affected by demand and supply of the market. The prices in electricity market can be set every five minutes. With this motivation, an algorithm is proposed for efficient Classification of concept Drifting Electricity pricing data streams. Thus, it is a challenge to learn from concept drifting data streams. In proposed algorithm, a decision tree is built incrementally and also used to develop training set based on these methods, in order to improve the accuracy of classification and prediction models under concept drift. A base learner is adaptive, a decision tree can have its nodes included and deleted dynamically. Adaptivity can be achieved by manipulating training data (instance selection), instead of taking all training history, take a number of the latest instances (training window). The new proposed algorithms detect change faster, without increasing the rate of false positives. Extensive studies on both synthetic and real-world data demonstrate that proposed algorithm outperforms well compared to several state-of-the-art online algorithms. In this paper we have compared electricity datasets with three algorithms to find out the algorithms efficiency on type of dataset. This data is again tested for error value for a particular number of iteration. The experimentation is conducted. The experimental evaluation produced satisfactory results.