Design and implementation of an autonomous EGR cooling system using deep neural network prediction to reduce NOx emission and fuel consumption of diesel engine

Akolas H. I., Kaleli A., BAKIRCI K.

NEURAL COMPUTING & APPLICATIONS, cilt.33, sa.5, ss.1655-1670, 2021 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 33 Sayı: 5
  • Basım Tarihi: 2021
  • Doi Numarası: 10.1007/s00521-020-05104-1
  • Dergi Adı: NEURAL COMPUTING & APPLICATIONS
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, PASCAL, Applied Science & Technology Source, Biotechnology Research Abstracts, Compendex, Computer & Applied Sciences, Index Islamicus, INSPEC, zbMATH
  • Sayfa Sayıları: ss.1655-1670
  • Anahtar Kelimeler: Autonomous EGR cooling system, NOx, BSFC, Deep neural networks, Backpropagation, Optimal tuned PID controller, EXHAUST-GAS RECIRCULATION, IN-CYLINDER, PERFORMANCE, COMBUSTION, LP
  • Atatürk Üniversitesi Adresli: Evet

Özet

This study includes the design of an autonomous exhaust gas recirculation (EGR) cooling system and implementation of the system on diesel engine by using deep neural network (DNN)-based controller. The NOx formation and BSFC were optimized as output variables considering engine speed, load, EGR ratio and the exhaust gas temperature entering the intake manifold as input variables using deep and traditional NN modelling method. The DNN modelling method has stronger prediction capability than traditional NN and can deal with the modelling problem in the complex ICE data. The developed NN models were compared with the relative error analysis to verify the effectiveness of the deep NN modelling method. The activation function, the number of neurons, learning rate and backpropagation optimization method were considered as algorithm performance parameters for the design of optimum network by using the searching method.R-2, MSE, MAE and RAE are used to evaluate the output performance of the deep and traditional NN model. Then, the autonomous and independent EGR cooling system is designed apart from the main cooling system of the engine with electric pump and fan components, unlike conventional systems. PID controllers are designed for these components, and the parameters of the controller are determined by pattern search optimization method. In this way, the exhaust gas temperature entering the intake manifold and EGR ratio are set to the desired value according to the BSFC and NOx output parameters under different operating conditions of the engine. Experimental results showed that the designed DNN controller-based autonomous EGR cooling system outperforms compared to the conventional system in terms of reducing the NOx and BSFC, with a 5.751% and 2.997% reduction, respectively.