Energetic assessment of a precalcining rotary kiln in a cement plant using process simulator and neural networks

[Display omitted] •This study assessed the energetic efficiency of a precalcining rotary kiln process in a cement industry.•Energetic efficiency was estimated as 61.30 % using Aspen Plus process simulator and artificial neural network (ANN) model predict with correlation coefficient (R2) of 0.991.•Optimal energy efficiency of 61.5 % was established for the operation parameters; kiln feed of 205050 kg/hr, kiln fuel gas of 2821 kg/hr, calciner fuel gas of 5648 kg/hr, clinker cooling air of 247463 kg/hr and primary air of 7309 kg/hr. Cement production has been increasing rapidly leading to energy consumption, with serious cost implications and environmental challenges. Energy efficiency is a key component required to maintain the cement company’s environmental strategy. In this study, Aspen Plus process model and neural networks are used to assess the energetic efficiency of a precalcining rotary kiln in a cement production process. Aspen Plus process simulator estimated energy efficiency at 61.30 % using the first law of thermodynamic. Further, for the ANN model, kiln feed, kiln gas, calciner gas, clinker cooling air, and primary air were the operation parameters inputs. ANN model is validated and demonstrated it is capable of predicting cement rotary kiln energy efficiency accurately with a correlation coefficient (R2) of 0.991. In conclusion, the Bootstrap aggregated neural network (BANN) was used to search the optimal operational parameters in achieving the lowest mean square error (MSE) of the energy efficiency. The MSE for training, testing, and validation data sets were 3.64 × 10-5, 3.70 × 10-5, and 5.00 × 10-5 for in the estimation of rotary kiln system energy efficiency. To achieve this optimal condition of 61.5 % energy efficiency, the optimal parameters as determined by ANN (BANN) were kiln feed of 205050 kg/hr, kiln fuel gas of 2821 kg/hr, calciner fuel gas of 5648 kg/hr, clinker cooling air of 247463 kg/hr and primary air of 7309 kg/hr. Consequently, it is recommended that ANN should be combined with Bootstrap aggregated neural network (BANN) for effective prediction and monitoring of energy efficiency for precalcining rotary kiln system.