Soft Sensors in the Primary Aluminum Production Process Based on Neural Networks Using Clustering Methods

Abstract

Primary aluminum production is an uninterrupted and complex process that must operate in a closed loop, hindering possibilities for experiments to improve production. In this sense, it is important to have ways to simulate this process computationally without acting directly on the plant, since such direct intervention could be dangerous, expensive, and time-consuming. This problem is addressed in this paper by combining real data, the artificial neural network technique, and clustering methods to create soft sensors to estimate the temperature, the aluminum fluoride percentage in the electrolytic bath, and the level of metal of aluminum reduction cells (pots). An innovative strategy is used to split the entire dataset by section and lifespan of pots with automatic clustering for soft sensors. The soft sensors created by this methodology have small estimation mean squared error with high generalization power. Results demonstrate the effectiveness and feasibility of the proposed approach to soft sensors in the aluminum industry that may improve process control and save resources.

Sparse Deep Neural Networks for Modeling Aluminum Electrolysis Dynamics

“…The paper mainly addresses the training speed of neural networks using an extreme learning machine. In [26] dense neural networks were used to predict variables in the electrolysis cell. The study accounted for the changing properties of the electrolysis cells by collecting data over the course of their lifecycle.…”

RNN- and LSTM-Based Soft Sensors Transferability for an Industrial Process

“…SS implementation often requires the use of black-box nonlinear dynamical identification strategies, which uses data collected from the distributed control system [ 11 ] and stored in the historical database. To achieve this aim, machine learning (ML) techniques are mostly used, ranging from Support Vector Regression [ 12 ], Partial Least Square [ 13 ], and classical multilayer perceptrons [ 1 , 14 , 15 , 16 , 17 ] to more recent deep architectures, such as deep belief networks [ 9 , 18 , 19 , 20 ], long short-term memory networks (LSTMs) [ 21 , 22 ], and stacked autoencoders [ 23 , 24 , 25 , 26 ]. Bayesian approaches [ 27 ], Gaussian Processes Regression [ 28 ], Extreme Learning Machines [ 29 ], and adaptive methods, [ 30 , 31 , 32 ] are also used.…”

Industrial Semi-Supervised Dynamic Soft-Sensor Modeling Approach Based on Deep Relevant Representation Learning

“…Basically, soft-sensing uses secondary variables (i.e., easy-to-measure variables) to estimate primary variables (i.e., hard-to-measure variables) [ 4 , 5 ]. Countless soft sensors have been designed using traditional methods: principal component regression (PCR) [ 6 , 7 ], partial least square (PLS) [ 8 , 9 ], support vector machine (SVM) [ 10 , 11 ], gaussian process regression (GPR) [ 12 , 13 ], artificial neural network (ANN) [ 14 , 15 ], and so on.…”