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 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.