Attention-Based Deep Recurrent Neural Network to Forecast the Temperature Behavior of an Electric Arc Furnace Side-Wall

Abstract

Structural health monitoring (SHM) in an electric arc furnace is performed in several ways. It depends on the kind of element or variable to monitor. For instance, the lining of these furnaces is made of refractory materials that can be worn out over time. Therefore, monitoring the temperatures on the walls and the cooling elements of the furnace is essential for correct structural monitoring. In this work, a multivariate time series temperature prediction was performed through a deep learning approach. To take advantage of data from the last 5 years while not neglecting the initial parts of the sequence in the oldest years, an attention mechanism was used to model time series forecasting using deep learning. The attention mechanism was built on the foundation of the encoder–decoder approach in neural networks. Thus, with the use of an attention mechanism, the long-term dependency of the temperature predictions in a furnace was improved. A warm-up period in the training process of the neural network was implemented. The results of the attention-based mechanism were compared with the use of recurrent neural network architectures to deal with time series data, such as Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU). The results of the Average Root Mean Square Error (ARMSE) obtained with the attention-based mechanism were the lowest. Finally, a variable importance study was performed to identify the best variables to train the model.

Soft Sensor Modeling for 3D Transient Temperature Field of Large-Scale Aluminum Alloy Workpieces Based on Multi-Loss Consistency Optimization PINN

“…It is evident that the optimal approach for partitioning X into k clusters can be achieved by employing the partition strategy, which entails minimizing the correlation square sum cost function min ss(∏) in Equation (23). Given that Tr X T X is determined by the sample space, the problem of min ss(∏) is considered as an equivalent optimization problem described as max U T U=I k Tr A T X T XA (24) According to the Ky Fan theorem, when dealing with a real and symmetric matrix B = X T X with eigenvalues…”

“…In recent years, the rapid advancement of machine learning has spurred research into utilizing neural network models for solving PDEs [ 21 , 22 , 23 ]. One approach is data-driven PDE solutions.…”

Prediction of outpatients with conjunctivitis in Xinjiang based on LSTM and GRU models

“…Bahdanau et al [24] showed that based on their preliminary experiments on machine translation, the two methods performed comparably. However, it is not known whether this task applies to other areas, so many scholars have made empirical comparisons, such as stock prediction [25], traffic flow prediction [26], short-term runoff prediction [27], and temperature forecast [28].…”