Ash Accumulation Prediction in Coal-Fired Power Plant Boilers Based on GWO-VMD-BiLSTM

The presence of soot in coal-fired power plant boilers has long been recognized as a significant factor leading to reduced heat transfer efficiency and compromised safety. A precise and scientific prediction of ash fouling on the heat transfer surface plays a crucial role in developing effective strategies for soot-blowing, thereby improving energy efficiency. To tackle the challenges associated with fluctuation and randomness in predicting ash fouling, we propose an innovative approach integrating the Grey Wolf Optimization (GWO) algorithm, The implementation of Variational Modal Decomposition (VMD) and Bi-directional Long Short-Term Memory (Bi-LSTM) neural network is undertaken. Initially, the Grey Wolf Optimization (GWO) algorithm is employed to automatically ascertain the crucial parameters of VMD, thereby aiding in the decomposition of the original time series. Subsequently, GWO-VMD is employed to obtain the Intrinsic Mode Functions (IMFs), forming the basis for constructing prediction models using Bi-LSTM. Each IMF corresponds to a sub-series whereby a distinct Bi-LSTM model is created. In the end, the projected values of each sub-series are combined to produce the ultimate prediction outcomes. By comparing our proposed model with three alternative models, we demonstrate that our model achieves the highest accuracy in predicting the cleanliness factor sequences, thus confirming its credibility and precision.