Improving propulsion capability and reducing fuel consumption in hybrid vehicles using artificial neural networks.
The massive emission of greenhouse gases from conventional cars with internal combustion engine has pushed governments around the world to seek alternatives and solutions to their problems. So, the idea of electric cars was first introduced. However due to the lack of proper infrastructures, the lack of funding for less developed communities, as well as the low cost of fuels in the world, has left off the use of all-electric vehicles and has led to less use of this technology. Therefore, the alternative solution is hybrid vehicles that, in addition to lower fuel consumption, advancement of science and fixing defects in this area, has helped to keep costs low and make it tangible to the people of the world. The goal, therefore, is to demonstrate improvements in fuel consumption, improving propulsion performance and more coordination among hybrid vehicle components. In order to achieve this goal, the artificial neural network model (ANN) has been used to increase the propulsion capability and reduce fuel consumption in hybrid vehicles and all stages have been implemented using MATLAB software. To get the initial information the Simulink of the MATLAB software is used and it was used to solve the problem of the neural network. Finally, all figures of MATLAB software are shown in different stages, which would help to understand the problem better. The result of this research is the ability to reduce fuel consumption and increase the useful life of the vehicle components, including the battery (due to reduced travel disturbances and soft moving of the vehicle at different speeds). It is also possible to optimize the problem for various vehicles and all types of roads by changing the parameters and conditions of the motion during the work (Neural Network Parameterization), to obtain the best and most intelligent condition.
KEYWORDS: Artificial Neural Network (ANN), Fuel consumption, Error propagation distances, Neural network efficiency