Predictive control of air-fuel ratio in aircraft engine on fuel-powered unmanned aerial vehicle using fuzzy-RBF neural network

Abstract

Air-fuel ratio control is important for optimizing the performance and reducing the exhaust emission of fuel-powered unmanned aerial vehicles 0(UAVs). However, previous studies on engine air-fuel ratio control neglect the fuel injection process and load of UAV propellers, and traditional methods could not satisfy the control requirement of an air-fuel ratio error of less than ± 2% when a UAV operates in different conditions. Here, to optimize the control performance, the mean value model of a fuel-powered aircraft engine is improved and an adaptive fuzzy radial basis function (RBF) neural network is used to perform predictive control. The simulation results are compared with the traditional control and some previous studies, and engine control experiments are implemented for demonstration. The simulation and experimental results indicate that, through predictive control using a fuzzy-RBF neural network, the air-fuel ratio of the aircraft engine can be controlled within ± 1% bounds of the stoichiometric value (14.7), and the highest error can be reduced by 68 to 75% compared with that in the previous work and the traditional neural network model, traditional PID method, and second-order sliding mode strategy. This research can be considered as a reference for intelligent algorithm applications on the power system of fuel-powered UAVs.