Otimização da frequência do ventilador de um condensador evaporativo utilizando inteligência artificial

Abstract

The present work addresses the theme of reducing electricity consumption in the industrial sector, which represents a significant portion of the total consumption in Brazil. This is a crucial factor for the competitiveness and sustainability of companies. Furthermore, it focuses on optimizing the performance of evaporative condensers, widely used refrigeration equipment in this sector. The main objective is to conduct a detailed thermal analysis of these condensers, taking into account environmental conditions such as temperature and humidity, in order to determine the ideal operating frequency of the exhaust system. The methodology employs principles of psychrometrics and equations carried out in the Engineering Equation Solver (EES) software to calculate this frequency. Additionally, the application of machine learning models, including Random Forest, Artificial Neural Networks, and Support Vector Regression (SVR), is proposed to predict the frequency of the fans in response to changes in environmental conditions. The analysis of the machine learning models showed that the Neural Network had the best values in the evaluation metrics, R², MSE and MAE, with the 33,000 data set standing out for its maximum optimization. In addition, there was a significant influence of relative humidity on the predicted frequency, showing a direct relationship between these parameters. Finally, determining the ideal frequency of the exhaust system resulted in a 43% reduction in costs related to the power of the equipment, varying by up to 46% when calculated based on the frequency predicted by the machine learning models.