Monitoring of Manufacturing Systems: Applying Deep Learning Algorithms to Improve Gas Turbine Performance and Sustainability

Gas turbines are considered one of the most important energy-generating systems in the world, covering the energy shortage resulting from the increasing electricity demand. Although the energy they produce is clean, they face a major environmental challenge, and this is due to the pollution they cause through the gases emitted during their operation. Carbon Oxides (CO) and Nitrogen Oxides (NOx) gas are some of the most important of these gases, and measuring their proportions in emissions has become an urgent matter. This measure contributes to many positive improvements. First, preserving the environment and air pollution by reducing the emission rates of greenhouse gases, which affect warming. Second, knowledge of the proportions of these gases makes us fully aware of the system's status, which allows us to develop methods for monitoring it using modern technologies. A prognosis of the proportions of emitted gases is one of the good solutions that contribute to the sustainability of gas turbines. In this work, we have proposed an intelligent model to predict the proportion of (CO) and (NOx), based on the Long Short-Term Memory (LSTM) this algorithm gave good results compared to the convolution neural network (CNN 1D).