Wavelet-Transformer Hybrid Model for Efficient Prediction of Condenser Vacuum in a Coal-Fired Power Plant

In this paper, we propose a hybrid model combining Discrete Wavelet Transform (DWT) and a Transformer-based neural network to predict the condenser vacuum in a thermal power plant. The DWT is used to decompose the operational data into multi-frequency components, capturing both short-term fluctuations and long-term trends. The Transformer model is then employed to effectively captures complex temporal dependencies and interactions between different features. By integrating these two techniques, the model provides more accurate and robust predictions of condenser vacuum under varying operational conditions. Experimental results demonstrate that the proposed model significantly improves prediction performance compared to traditional methods, making it a valuable tool for optimizing power plant operations and enhancing overall efficiency.