Prediction of optimum operating conditions of a furnace under uncertainty: An integrated framework of artificial neural network and genetic algorithm

Khan M., Ahmad I., Ahsan M., Kano M., ÇALIŞKAN H.

Fuel, cilt.330, 2022 (SCI-Expanded) identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 330
  • Basım Tarihi: 2022
  • Doi Numarası: 10.1016/j.fuel.2022.125563
  • Dergi Adı: Fuel
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, PASCAL, Aerospace Database, Biotechnology Research Abstracts, Chemical Abstracts Core, Communication Abstracts, INSPEC, Metadex, Pollution Abstracts, Civil Engineering Abstracts
  • Anahtar Kelimeler: Artificial intelligence, Exergy analysis, Fired heater, Furnace, Machine learning, Optimization
  • Uşak Üniversitesi Adresli: Evet

Özet

The current work developed an integrated framework of artificial intelligence and genetic algorithm (GA) for the furnace of a petroleum refinery to predict the optimum amount of excess air and mass flow rates of crude oil and fuel stream in the presence of uncertainty in process conditions. A furnace model was built in Aspen Exchanger Design and Rating (EDR) using industrial data. Then interface between the furnace model and MATLAB was established using the com server. Data points were generated by varying the crude oil composition as well as the inlet temperature and pressure of cold crude oil, fuel, and air stream through ±1%, ±2%, ±3%, ±4%, and ±5% change in their original values. Then GA was used to determine the optimal amount of excess air and mass flow rates of crude oil and fuel for each data point. In total, 360 data points were generated and 70 % of the data set was utilized for training an Artificial Neural Networks (ANN) model while the remaining data points were evenly distributed for its validation and testing. The proposed ANN model achieved high performance with a correlation coefficient of 0.99984. Furthermore, exergy analysis of the standalone furnace model, the GA based optimized model, and the ANN integrated model was performed. The GA based approach outperformed the standalone model in terms of exergetic efficiency. On the hand, the ANN based framework achieved comparable exergetic efficiency to the GA based approach within significantly lesser period of time.