A Novel Power Load Identification Strategy Based on CEEMD-PCA-AE-Shapelet Dictionary Learning Algorithm
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
The development of the Energy Internet under the China 'dual-carbon' strategy has bring new challenges in load identification and anomaly detection, particularly in industrial applications. Traditional methods often struggle with the sparse, noisy nature of industrial load data, limiting their generalization ability and accuracy. This paper proposes a novel load identification method based on the CEEMD-PCA-AE-Shapelet dictionary learning algorithm, which addresses these challenges by combining an improved unsupervised anomaly detection model (CEEMD-PCA-AE) with the Shapelet dictionary learning algorithm. Firstly, CEEMD-PCA-AE enhances the model’s ability to handle non-linear and noisy data, significantly improving the generalization and accuracy of industrial load anomaly detection. Secondly, the Shapelet dictionary learning algorithm reduces computational complexity and improves model performance by incorporating dictionary learning into time-series classification. The proposed method outperforms traditional models, as demonstrated by numerical experiments, offering enhanced load identification accuracy, and improved detection of anomalies. This method has significant potential for optimizing the efficiency of energy management in industrial settings and can be applied globally, as shown by its effectiveness on real-world data from Liaoning Province, China.
Fields of Study & Subjects
Time Series Analysis and Forecasting
Energy Load and Power Forecasting
Smart Grid Energy Management
Publication details
- Journal / source
- EAI Endorsed Transactions on Energy Web
- Volume
- 12
- Publication date
- 31 March 2026
- Language
- EN