Assessment of the thermodynamics efficiency of a cement vertical raw mill using Aspen Plus and artificial intelligence models

Abstract

Abstract Globally, cement plants are striving to improve their energy efficiency. Therefore, it is critical for cement plant operations to increase the monitoring and control of a vertical raw mill energy process. This technology has attracted the interest of the cement industry with its proven benefits in cement grinding applications. A process simulator was used to study an industrial-scale vertical raw mill (VRM) with 65.4% energy efficiency. The paper proposes further a new model based on grid partitioning, sub-clustering, and fuzzy c-means, which incorporates genetic algorithms (GAs) and particle swarm optimizations (PSOs). VRM data from a steady plant process operation, such as raw material output, material moisture, kiln hot gas, mill fan flow, grinding pressure, and separator speed, was used as input to the prediction model. ANFIS-based prediction models are compared with process simulator predictions to determine the most accurate based on prediction performance criteria. Based on the results, the ANFIS model with sub-clustering assimilated with PSO is the most accurate prediction model for VRM energy efficiency. The coefficient of regression (R2) and root mean square error (RMSE) obtained by this model are 0.945 and 1.3006. The results also showed that VRM's energy efficiency decreased from 65.4 to 64.2% when the separator speed increased from 50 to 75 rpm; product particle size on P90µm decreased from 18.2–10.8%. Finally, the proposed ANFIS based model can be considered to be an efficient technique for predicting the energy efficiency of VRM production processes.