Abstract:This paper presents a daily peak load forecasting method using an analyzable structured neural network (ASNN) in order to explain forecasting reasons. In this paper, we propose a new training method for ASNN in order to explain forecasting reason more properly than the conventional training method. ASNN consists of two types of hidden units. One type of hidden units has connecting weights between the hidden units and only one group of related input units. Another one has connecting weights between the hidden u… Show more
“…10 and 11, the electric load prediction was performed well by the prediction method proposed in this paper. [6], the prediction error is 2.83% under the proposed method as compared to a prediction error of 2.53% for MEP, and although this error is 0.3% lower, it is within the prediction error of 3% for which electric power companies strive, so that the proposed method can be considered a good prediction method. However, the ∞ filter gives more precise results for the MEB prediction on average when looking at the predictions for 30 days in Fig.…”
Section: Predictions Resulting From Filter Differencesmentioning
confidence: 77%
“…In performing parameter estimation, we used the data for the past 30 days from the prediction start date as the past data for the simulations. [6]. In this paper, we use data for the 30 days before the prediction day, taking into consideration past data and the load difference due to season.…”
Section: Simulationsmentioning
confidence: 99%
“…Previously, the literature [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17] has included research on short-term electric load prediction. Refs.…”
Section: Introductionmentioning
confidence: 99%
“…Refs. [6] proposes a clarification of forecasting with a hierarchical NN and improves its explanatory capability. A feature of NNs is that they can make more precise forecasts than other forecasting methods and can readily reflect the intentions of their designers.…”
Section: Introductionmentioning
confidence: 99%
“…[17]. [6] and [17]. This prediction method has the advantages of assuming a state with the worst possible noise and improving prediction accuracy by setting the gain so that the prediction error is minimized.…”
This paper deals with ∞ filter-based short-term electric load prediction taking into consideration the characteristics of the load curve. We propose a predictive method to forecast the future electric load demand for 36 h from 12:00 PM, and evaluate the peak and bottom of the load curves on the next day. We propose a load model, estimate the unknown parameters of the model by means of an ∞ filter using the data separated for nonworking days and weekdays, with the same pattern of the previous data chosen and assigned to the model parameters. The simulation results show the effectiveness of the proposed prediction methodology. C⃝ 2014 Wiley Periodicals, Inc. Electron Comm Jpn, 97(12): 1-10, 2014; Published online in Wiley Online Library (wileyonlinelibrary.com).
“…10 and 11, the electric load prediction was performed well by the prediction method proposed in this paper. [6], the prediction error is 2.83% under the proposed method as compared to a prediction error of 2.53% for MEP, and although this error is 0.3% lower, it is within the prediction error of 3% for which electric power companies strive, so that the proposed method can be considered a good prediction method. However, the ∞ filter gives more precise results for the MEB prediction on average when looking at the predictions for 30 days in Fig.…”
Section: Predictions Resulting From Filter Differencesmentioning
confidence: 77%
“…In performing parameter estimation, we used the data for the past 30 days from the prediction start date as the past data for the simulations. [6]. In this paper, we use data for the 30 days before the prediction day, taking into consideration past data and the load difference due to season.…”
Section: Simulationsmentioning
confidence: 99%
“…Previously, the literature [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17] has included research on short-term electric load prediction. Refs.…”
Section: Introductionmentioning
confidence: 99%
“…Refs. [6] proposes a clarification of forecasting with a hierarchical NN and improves its explanatory capability. A feature of NNs is that they can make more precise forecasts than other forecasting methods and can readily reflect the intentions of their designers.…”
Section: Introductionmentioning
confidence: 99%
“…[17]. [6] and [17]. This prediction method has the advantages of assuming a state with the worst possible noise and improving prediction accuracy by setting the gain so that the prediction error is minimized.…”
This paper deals with ∞ filter-based short-term electric load prediction taking into consideration the characteristics of the load curve. We propose a predictive method to forecast the future electric load demand for 36 h from 12:00 PM, and evaluate the peak and bottom of the load curves on the next day. We propose a load model, estimate the unknown parameters of the model by means of an ∞ filter using the data separated for nonworking days and weekdays, with the same pattern of the previous data chosen and assigned to the model parameters. The simulation results show the effectiveness of the proposed prediction methodology. C⃝ 2014 Wiley Periodicals, Inc. Electron Comm Jpn, 97(12): 1-10, 2014; Published online in Wiley Online Library (wileyonlinelibrary.com).
In a distribution line, power system control and power equipment investment are planned based on a measured power system current. However, recently the mass introduction of photovoltaic (PV) make it difficult for us to precisely measure the demand curve that is a current consumed by electrical equipment because the reversal power flow from PV systems is superposed. Therefore, the prediction of demand curves of distribution line is indispensable for power system management. In addition, it is also necessary to estimate the reliability of the predicted values as well as predicted current itself. In this paper, we propose the estimation method of the prediction interval that is the index of reliability based on the past demand curve database. The feature of the proposed method based on Just-In-Time (JIT) modeling make it possible for us to accurately estimate the prediction interval by the normalized database of demand curve. In this paper, some numerical examples are presented, which demonstrate the effectiveness of the proposed method. C⃝ 2017 Wiley Periodicals, Inc. Electr Eng Jpn, 202(2): 12-23, 2018; Published online in Wiley Online Library (wileyonlinelibrary.com).
This paper deals with a residential co-generation system as an application of the hierarchically decentralized model. We assume the residential co-generation systems consist with autonomous co-generation units that behave autonomously. It is important in controlling co-generation systems to forecast quantity of energy demands, and thus we use Neural Network to forecast them. We propose a simulation model of the residential co-generation system and a model of agents which control co-generation unit autonomously. To simulate the model, we prepare artificial demand data. Through several computer simulations, effectiveness of sharing information between cogeneration units is investigated.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.