A novel hybrid model based on combined preprocessing method and advanced optimization algorithm for power load forecasting

Research on short-term power load forecasting based on VMD and GRU

The traditional method for power load forecasting is susceptible to various factors, including holidays, seasonal variations, weather conditions, and more. These factors make it challenging to ensure the accuracy of forecasting results. Additionally, there is a limitation in extracting meaningful physical signs from power data, which ultimately reduces prediction accuracy. This paper aims to address these issues by introducing a novel approach called VCAG (Variable Mode Decomposition-Convolutional Neural Network-Attention Mechanism-Gated Recurrent Unit) for combined power load forecasting. In this approach, we integrate Variable Mode Decomposition (VMD) with Convolutional Neural Network (CNN). VMD is employed to decompose power load data, extracting valuable time-frequency features from each component. These features then serve as input for the CNN. Subsequently, an attention mechanism is applied to give importance to specific features generated by the CNN, enhancing the weight of crucial information. Finally, the weighted features are fed into a Gated Recurrent Unit (GRU) network for time series modeling, ultimately yielding accurate load forecasting results.To validate the effectiveness of our proposed model, we conducted experiments using two publicly available datasets. The results of these experiments demonstrate that our VCAG method achieves high accuracy and stability in power load forecasting, effectively overcoming the limitations associated with traditional forecasting techniques. As a result, this approach holds significant promise for broad applications in the field of power load forecasting.

A New Time Series Forecasting Model Based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise and Temporal Convolutional Network

In this paper, a new hybrid time series forecasting model based on the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and a temporal convolutional network (TCN) (CEEMDAN-TCN) is proposed. The CEEMDAN is used to decompose the time series data and the TCN is used to obtain a good prediction accuracy. The effectiveness of the model is verified in univariate and multivariate time series forecasting tasks. The experimental results indicate that compared with the long short-term memory model and other hybrid models, the proposed CEEMDAN-TCN model shows a better performance in both univariate and multivariate prediction tasks.

Deterioration of Electrical Load Forecasting Models in a Smart Grid Environment

Smart Grid (S.G.) is a digitally enabled power grid with an automatic capability to control electricity and information between utility and consumer. S.G. data streams are heterogenous and possess a dynamic environment, whereas the existing machine learning methods are static and stand obsolete in such environments. Since these models cannot handle variations posed by S.G. and utilities with different generation modalities (D.G.M.), a model with adaptive features must comply with the requirements and fulfill the demand for new data, features, and modality. In this study, we considered two open sources and one real-world dataset and observed the behavior of ARIMA, ANN, and LSTM concerning changes in input parameters. It was found that no model observed the change in input parameters until it was manually introduced. It was observed that considered models experienced performance degradation and deterioration from 5 to 15% in terms of accuracy relating to parameter change. Therefore, to improve the model accuracy and adapt the parametric variations, which are dynamic in nature and evident in S.G. and D.G.M. environments. The study has proposed a novel adaptive framework to overcome the existing limitations in electrical load forecasting models.

Ensemble system for short term carbon dioxide emissions forecasting based on multi-objective tangent search algorithm

Carbon emissions play a crucial role in inducing global warming and climate change. Accurate and stable carbon emissions forecasting is beneficial for formulating emissions reduction schemes and achieving carbon neutrality as early as possible. Although previous studies have concentrated on employing one or several methods for carbon emissions forecasting, the improvement in forecasting performance is limited because they ignore the importance of objectively selecting the models and the necessity of interval forecasting. In this paper, a novel ensemble prediction system, composed of data decomposition, model selection, phase space reconstruction, ensemble point prediction, and interval prediction, is proposed to conduct both point and interval predictions, which has been proven to be effective in prompting carbon emissions forecasting accuracy and stability. According to the empirical results, the mean MAPE results of our proposed forecasting strategy in point prediction are 1.1102% (in Dataset A) and 1.1382% (in Dataset B), and the mean CWC values in the interval forecasting are 0.3512 and 0.1572, respectively. Thus, the proposed forecasting system improves the forecasting performance relative to other models considerably, which can provide meaningful references for policymakers.

An Ensemble Model based on Deep Learning and Data Preprocessing for Short-Term Electrical Load Forecasting

Electricity load forecasting is one of the hot concerns of the current electricity market, and many forecasting models are proposed to satisfy the market participants’ needs. Most of the models have the shortcomings of large computation or low precision. To address this problem, a novel deep learning and data processing ensemble model called SELNet is proposed. We performed an experiment with this model; the experiment consisted of two parts: data processing and load forecasting. In the data processing part, the autocorrelation function (ACF) was used to analyze the raw data on the electricity load and determine the data to be input into the model. The variational mode decomposition (VMD) algorithm was used to decompose the electricity load raw-data into a set of relatively stable modes named intrinsic mode functions (IMFs). According to the time distribution and time lag determined using the ACF, the input of the model was reshaped into a 24 × 7 × 8 matrix M, where 24, 7, and 8 represent 24 h, 7 days, and 8 IMFs, respectively. In the load forecasting part, a two-dimensional convolutional neural network (2D-CNN) was used to extract features from the matrix M. The improved reshaped layer was used to reshape the extracted features according to the time order. A temporal convolutional network was then employed to learn the reshaped time-series features and combined with the fully connected layer to complete the prediction. Finally, the performance of the model was verified in the Eastern Electricity Market of Texas. To demonstrate the effectiveness of the proposed model data processing and load forecasting, we compared it with the gated recurrent unit (GRU), TCN, VMD-TCN, and VMD-CNN models. The TCN exhibited better performance than the GRU in load forecasting. The mean absolute percentage error (MAPE) of the TCN, which was over 5%, was less than that of the GRU. Following the addition of VMD to the TCN, the basic performance of the model was 2–3%. A comparison between the SELNet model and the VMD-TCN model indicated that the application of a 2D-CNN improves the forecast performance, with only a few samples having an MAPE of over 4%. The model’s prediction effect in each season is discussed, and it was found that the proposed model can achieve high-precision prediction in each season.