Chaotic time series prediction with residual analysis method using hybrid Elman-NARX neural networks

http://repository.usp.ac.fj/7551/1/ccrnntime.pdf

Cooperative coevolution of Elman recurrent neural networks for chaotic time series prediction

Very short-term load forecasting predicts the loads 1 h into the future in 5-min steps in a moving window manner based on real-time data collected. Effective forecasting is important in area generation control and resource dispatch. It is however difficult in view of the noisy data collection process and complicated load features. This paper presents a method of wavelet neural networks with data pre-filtering. The key idea is to use a spike filtering technique to detect spikes in load data and correct them. Wavelet decomposition is then used to decompose the filtered loads into multiple components at different frequencies, separate neural networks are applied to capture the features of individual components, and results of neural networks are then combined to form the final forecasts. To perform moving forecasts, 12 dedicated wavelet neural networks are used based on test results. Numerical testing demonstrates the effects of data pre-filtering and the accuracy of wavelet neural networks based on a data set from ISO New England.

Very Short-Term Load Forecasting: Wavelet Neural Networks With Data Pre-Filtering

The very short-term load forecasting (VSTLF) problem is of particular interest for use in smart grid and automated demand response applications. An effective solution for VSTLF can facilitate real-time electricity deployment and improve its quality. In this paper, a novel approach to model the very short-term load of individual households based on context information and daily schedule pattern analysis is proposed. Several daily behavior pattern types were obtained by analyzing the time series of daily electricity consumption, and context features from various sources were collected and used to establish a rule set for use in anticipating the likely behavior pattern type of a specific day. Meanwhile, an electricity consumption volume prediction model was developed for each behavior pattern type to predict the load at a specific time point in a day. This study was concerned with solving the VSTLF for individual households in Taiwan. The proposed approach obtained an average mean absolute percentage error (MAPE) of 3.23% and 2.44% for forecasting individual household load and aggregation load 30-min ahead, respectively, which is more favorable than other methods.

Household Electricity Demand Forecast Based on Context Information and User Daily Schedule Analysis From Meter Data

Very short-term load forecasting predicts the loads in electric power system one hour into the future in 5-min steps in a moving window manner. To quantify forecasting accuracy in real-time, the prediction interval estimates should also be produced online. Effective predictions with good prediction intervals are important for resource dispatch and area generation control, and help power market participants make prudent decisions. We previously presented a two level wavelet neural network method based on back propagation without estimating prediction intervals. This paper extends the previous work by using hybrid Kalman filters to produce forecasting with prediction interval estimates online. Based on data analysis, a neural network trained by an extended Kalman filter is used for the low-low frequency component to capture the near-linear relationship between the input load component and the output measurement, while neural networks trained by unscented Kalman filters are used for low-high and high frequency components to capture their nonlinear relationships. The overall variance estimate is then derived and evaluated for prediction interval estimation. Testing results demonstrate the effectiveness of hybrid Kalman filters for capturing different features of load components, and the accuracy of the overall variance estimate derived based on a data set from ISO New England.

Hybrid Kalman Filters for Very Short-Term Load Forecasting and Prediction Interval Estimation

It is indispensable to accurately perform short-term load forecasting of 10 minutes ahead in order to avoid undesirable disturbances in power system operations. The authors have so far developed such a forecasting method based on conventional chaos theory. However, this approach cannot give accurate forecasting results when the loads consecutively exceed the historical maximum or are less than the minimum. Electric furnace loads with steep fluctuations are another factor degrading the forecast accuracy.



This paper presents an improved forecasting method based on chaos theory. In particular, the potential of the Local Fuzzy Reconstruction Method, a variant of the localized reconstruction methods, is fully exploited to realize accurate forecasting as much as possible. To resolve the forecast deterioration due to suddenly changing loads such as by electric furnaces, they are separated from the rest and smoothing operations are carried out afterwards. The separated loads are forecasted independently from the remaining components. Several error correction methods are incorporated to enhance the proposed forecasting method. Furthermore, a consistent measure of obtaining the optimal combination of parameters to be used in the forecasting method is presented. The effectiveness of the proposed methods is verified by using real load data for 1 year. © 2004 Wiley Periodicals, Inc. Electr Eng Jpn, 148(2): 55–63, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/eej.10322

Development of very‐short‐term load forecasting based on chaos theory

This paper proposes a new method for obtaining controlling unstable equilibrium point (CUEP) for transient stability analysis in electric power systems. A minimization problem is formulated to attain CUEP by applying the critical trajectory (CTrj) method to the boundary controlling unstable equilibrium point (BCU) method. CTrj method is a unique approach proposed by the authors to obtain critical condition for transient stability problem. The Proposed method simultaneously computes a trajectory on the stability boundary starting from the exit point reaching CUEP. The effectiveness of the method is demonstrated through simulations for various systems.

An Application of Critical Trajectory Method to BCU Problem for Transient Stability Studies

This paper presents a new computation method for transient stability analysis for electric power systems. Different from existing methods, a minimization problem is formulated for obtaining critical clearing time (CCT). The method is based on the computation of a trajectory on the stability boundary, which is referred to as critical trajectory in this paper. The critical trajectory is defined as the trajectory that starts from a point on a fault-on trajectory and reaches a critical point of losing synchronism. The new proposal includes the critical conditions for synchronism and the unified minimization formulation using a modified trapezoidal formulation for numerical integration. It will be demonstrated that the solution of the minimization problem successfully provides the exact CCT that agrees with the conventional numerical simulation method.

A new method for direct computation of critical clearing time for transient stability analysis

It is indispensable to accurately perform the short-term load forecasting of 10 minutes ahead in order to avoid undesirable disturbances in power system operations. The authors have so far developed such a forecasting method based on the conventional chaos theory. However, this approach is unable to give accurate forecasting results in case where the loads consecutively exceed than the historical maximum or lower than the minimum. Also, electric furnace loads with steep fluctuations have been another factor to degrade the forecast accuracy.This paper presents an improved forecasting method based on Chaos theory. Especially, the potential of the Local Fuzzy Reconstruction Method, a variant of the localized reconstruction methods, is fully exploited to realize accurate forecast as much as possible. To resolve the forecast deterioration due to sudden change loads such as by electric furnaces, they are separated from the rest and smoothing operations are carried out afterwards. The separated loads are forecasted independently from the remaining components. Several error correction methods are incorporated to enhance the proposed forecasting method. Furthermore, a consistent measure of obtaining the optimal combination of parameters to be used in the forecasting method is given. The effectiveness of the proposed methods is verified by using real load data for one year.

A Development of Very Short-Term Load Forecasting Based on Chaos Theory

This paper studies new techniques for critical trajectory method, a recent new method proposed by the authors for obtaining critical clearing time (CCT) for transient stability analysis. A specific feature of the proposed method lies in its ability to provide exact CCT without approximations since no such methods have existed so far. The method is based on the computation of the critical trajectory, which is defined as the trajectory that starts from a point on a fault-on trajectory at CCT and reaches an end point. There are a few possible methods for the treatment of the end point conditions, computational performances of the methods are investigated in terms of accuracy of CCT and computational efficiency. It is shown that the proposed methods successfully provide the exact CCT for Xd' generator model with and without controllers that agrees with the conventional numerical simulation method.

Hiroaki Sugihara

Papers

Development of very‐short‐term load forecasting based on chaos theory

An Application of Critical Trajectory Method to BCU Problem for Transient Stability Studies

A new method for direct computation of critical clearing time for transient stability analysis

A Development of Very Short-Term Load Forecasting Based on Chaos Theory

A novel method for direct computation CCT for TSA using critical generator conditions