• Journal of Wind Energy
• /
• v.15 no.2
• /
• pp.10-22
• /
• 2024

This research is a comprehensive analysis of wind power prediction sensitivity using a Long Short-Term Memory (LSTM) deep learning neural network model, accounting for the inherent uncertainties in wind speed estimation. Utilizing a year's worth of operational data from an operational wind farm, the study forecasts the power output of both individual wind turbines and the farm collectively. Predictions were made daily at intervals of 10 minutes and 1 hour over a span of three months. The model's forecast accuracy was evaluated by comparing the root mean square error (RMSE), normalized RMSE (NRMSE), and correlation coefficients with actual power output data. Moreover, the research investigated how inaccuracies in wind speed inputs affect the power prediction sensitivity of the model. By simulating wind speed errors within a normal distribution range of 1% to 15%, the study analyzed their influence on the accuracy of power predictions. This investigation provided insights into the required wind speed prediction error rate to achieve an 8% power prediction error threshold, meeting the incentive standards for forecasting systems in renewable energy generation.

• Journal of Intelligence and Information Systems
• /
• v.5 no.1
• /
• pp.95-101
• /
• 1999

In this paper, a comparison between different methods to combine predictions from neural networks will be given. These methods are bagging, bumping, and balancing. Those are based on the analysis of the ensemble generalization error into an ambiguity term and a term incorporating generalization performances of individual networks. Neural Networks and AI machine learning models are prone to overfitting. A strategy to prevent a neural network from overfitting, is to stop training in early stage of the learning process. The complete data set is spilt up into a training set and a validation set. Training is stopped when the error on the validation set starts increasing. The stability of the networks is highly dependent on the division in training and validation set, and also on the random initial weights and the chosen minimization procedure. This causes early stopped networks to be rather unstable: a small change in the data or different initial conditions can produce large changes in the prediction. Therefore, it is advisable to apply the same procedure several times starting from different initial weights. This technique is often referred to as training ensembles of neural networks. In this paper, we presented a comparison of three statistical methods to prevent overfitting of neural network.

• Nuclear Engineering and Technology
• /
• v.55 no.3
• /
• pp.839-849
• /
• 2023

The development of automation technology to reduce human error by minimizing human intervention is accelerating with artificial intelligence and big data processing technology, even in the nuclear field. Among nuclear power plant operation modes, the startup and shutdown operations are still performed manually and thus have the potential for human error. As part of the development of an autonomous operation system for startup operation, this paper proposes an action coordinating strategy to obtain the optimal actions. The lower level of the system consists of operating blocks that are created by analyzing the operation tasks to achieve local goals through soft actor-critic algorithms. However, when multiple agents try to perform conflicting actions, a method is needed to coordinate them, and for this, an action coordination strategy was developed in this work as the upper level of the system. Three quantification methods were compared and evaluated based on the future plant state predicted by plant parameter prediction models using long short-term memory networks. Results confirmed that the optimal action to satisfy the limiting conditions for operation can be selected by coordinating the action sets. It is expected that this methodology can be generalized through future research.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.19 no.4
• /
• pp.161-167
• /
• 2019

Recently, demand forecasting techniques have been actively studied due to interest in stable power supply with surging power demand, and increase in spread of smart meters that enable real-time power measurement. In this study, we proceeded the deep learning prediction model experiments which learns actual measured power usage data of home and outputs the forecasting result. And we proceeded pre-processing with moving average method. The predicted value made by the model is evaluated with the actual measured data. Through this forecasting, it is possible to lower the power supply reserve ratio and reduce the waste of the unused power. In this paper, we conducted experiments on three types of networks: Multi Layer Perceptron (MLP), Recurrent Neural Network (RNN), and Long Short Term Memory (LSTM) and we evaluate the results of each scheme. Evaluation is conducted with following method: MSE(Mean Squared Error) method and MAE(Mean Absolute Error).

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.6
• /
• pp.453-459
• /
• 2021

The system marginal price of electricity is the amount paid to all the generating units, which is an important decision-making factor for the construction and maintenance of an electrical power unit. In this paper, we suggest a long-term forecasting model for calculating the system marginal price based on prices of natural gas and oil. As most variables used in the analysis are nonstationary time series, the long run relationship among the variables should be examined by cointegration tests. The forecasting model is similar to an error correction model which consists of a long run cointegrating equation and another equation for short run dynamics. To mitigate the robustness issue arising from the relatively small data sample, this study employs various testing and estimating methods. Compared to previous studies, this paper considers multiple fuel prices in the forecasting model of system marginal price, and provides greater emphasis on the robustness of analysis. As none of the cointegrating relations associated with system marginal price, natural gas price and oil price are excluded, three error correction models are estimated. Considering the root mean squared error and mean absolute error, the model based on the cointegrating relation between system marginal price and natural gas price performs best in the out-of-sample forecast.

• Proceedings of the IEEK Conference
• /
• 2003.07c
• /
• pp.2569-2572
• /
• 2003

In this paper, we Proposed the inverted pendulum control method using single neuron neural network that have weights as PID parameters. The proposed method has three inputs(proportion, integration, differentiation term of the error), and uses weights as P, I, D parameters. In order to verify the effectiveness of the proposed method, we experimented on the rotary inverted pendulum with load effect disturbance. The results showed the effectiveness and robustness of the proposed pendulum controller.

• Journal of Power System Engineering
• /
• v.19 no.5
• /
• pp.45-51
• /
• 2015

Grade 91 steel is used for the major structural components of Generation-IV reactor systems such as a very high temperature reactor (VHTR) and sodium-cooled fast reactor (SFR). Since these structures are designed for up to 60 years at elevated temperatures, the prediction of long-term creep life is very important to determine an allowable design stress of elevated temperature structural component. In this study, a large body of creep rupture data was collected through world-wide literature surveys, and using these data, the long-term creep life was predicted in terms of three methods: Larson-Miller (L-M), Manson-Haferd (M-H) and Wilshire methods. The results for each method was compared using the standard deviation of error. The L-M method was overestimated in the longer time of a low stress. The Wilshire method was superior agreement in the long-term life prediction to the L-M and M-H methods.

• Proceedings of the KIEE Conference
• /
• 1993.07a
• /
• pp.29-32
• /
• 1993

A technique computing short-term load foadcasting is essential for monitoring and controlling power system operation. This paper shows the use of general exponential smoothing to develop an adaptive forecasting system based on observed value of hourly demand. Forecasts of hourly load with lead times of one to twenty-four hours are computed at hourly intervals throughout the week. Standard error for lead times of one to twenty-four hour range from three to four percent average load. Studies are planned to investigate the use of weather influence to increase forecast accuracy.

• Journal of applied mathematics & informatics
• /
• v.27 no.3_4
• /
• pp.475-487
• /
• 2009

This paper studies the iteration method of solving a type of second-order three-point boundary value problem with non-linear term f, which depends on the first order derivative. By using the upper and lower method, we obtain the sufficient conditions of the existence and uniqueness of solutions. Furthermore, the monotone iterative sequences generated by the method contribute to the minimum solution and the maximum solution. And the error estimate formula is also given under the condition of unique solution. We apply the solving process to a special boundary value problem, and the result is interesting.

• Journal of international Conference on Electrical Machines and Systems
• /
• v.2 no.3
• /
• pp.283-287
• /
• 2013

This paper presents a method for calculating iron losses in three-phase induction motors under the inverter supply through the field-circuit coupled time-stepping finite element method (FEM). Iron losses are calculated by using the three-term iron losses separated model and modifying the loss coefficients obtained by the iron losses curves which are provided by the manufacturer under the sinusoidal supply. Simulation results by the presented method are verified by the measured results with an error lower than 5%, confirming the validity of the proposed method. Finally, iron losses distribution of the inverter-fed three-phase induction prototype motor is shown.