Application of artificial neural network to predict brake specific fuel consumption of retrofitted cng engine
In this paper the applicability of artificial neural networks (ANN) is investigated for a retrofitted compressed natural gas (CNG) fueled spark ignition (SI) internal combustion engine (ICE). A four cylinder carbureted petrol engine is converted to run with NG and used throughout the work. The neura...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2009
|
| Subjects: | |
| Online Access: | http://eprints.um.edu.my/6785/1/Application_of_artificial_neural_network_to_predict_brake_specific_fuel_consumption_of_retrofitted_cng_engine_%282%29.pdf http://eprints.um.edu.my/6785/ http://ejum.fsktm.um.edu.my/article/811.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | In this paper the applicability of artificial neural networks (ANN) is investigated for a retrofitted compressed natural gas (CNG) fueled spark ignition (SI) internal combustion engine (ICE). A four cylinder carbureted petrol engine is converted to run with NG and used throughout the work. The neural networks toolbox of Matlab 6.5 is used to develop and test the ANN model on a personal computer. An optimal design is completed for the 3 to 12 hidden neurons on single hidden layer with six different algorithms: batch gradient descent (GD), resilient back-propagation (RP), levenberg-marquardt (LM), batch gradient descent with momentum (GDM), variable learning rate (GDX), scaled conjugate gradient (SCG) in the back-propagation neural network model. The training data for ANN is obtained from experimental measurements. Engine speed (rpm), throttle position, fuel-air equivalence ratio (φ) and torque (N-m) were used in input layer while break specific fuel consumption (gm/kWh) was used as output layer. Statistical analysis in terms of Root-Mean-Squared (RMS), absolute fraction of variance (R2), as well as mean percentage error is used to investigate the prediction performance of ANN. LM algorithm with 10 neurons on single hidden layer in back-propagation of ANN model has shown best result in the present study. The degree of accuracy of the ANN model in prediction is proven acceptable in all statistical analysis and shown in results. So, it can be concluded that ANN provides a feasible method in predicting specific fuel consumption of CNG driven SI engine. |
|---|