Research on New Energy Consumption Data Combination Forecast Technology Based on Improved TCN-BiLSTM

Currently, the new energy consumption problem in regional power grids is worsening. Evaluating it with traditional optimization algorithms is challenging, inefficient, and time consuming. First, factors affecting consumption are analyzed and prediction data prepared. Then, an improved TCN - BiLSTM m...

Full description

Saved in:
Bibliographic Details
Published in2025 2nd International Conference on Electrical Technology and Automation Engineering (ETAE) pp. 224 - 228
Main Authors Liu, Jiale, Deng, Weisi, Liu, Xianzhuo, Gao, Weidong, Chen, Hao
Format Conference Proceeding
LanguageEnglish
Published IEEE 23.05.2025
Subjects
Bidirectional long short term memory
Computational modeling
Data models
Distributed databases
dynamic spectrum management improved TCNBiLSTM
Energy consumption
Feature extraction
Load forecasting
Load modeling
Optimization
power system
Power system stability
stability assessment
Online AccessGet full text
DOI10.1109/ETAE65337.2025.11089767

Cover

More Information
Summary:Currently, the new energy consumption problem in regional power grids is worsening. Evaluating it with traditional optimization algorithms is challenging, inefficient, and time consuming. First, factors affecting consumption are analyzed and prediction data prepared. Then, an improved TCN - BiLSTM model is proposed. TCN captures long - term dependencies and boosts efficiency. BiLSTM extracts temporal features well. A cross - entropy loss function is used in training to tackle class imbalance. Experiments show it performs well and accurately in power system transient stability assessment.
DOI:10.1109/ETAE65337.2025.11089767