Adopting the proportion-integration-differentiation neural network (PIDNN) control algorithm

the advantages of both the tradition PID and neural network were combined. It was used to control the engine air-fuel ratio under different operating conditions

and when the condition changed

it could control the air-fuel ratio to the target value. An engine model was built in software AMESim. The PIDNN control algorithm was built in software MATLAB. Through the model-in-the-loop

the simulation results show that: at different altitudes

the PIDNN could control the air-fuel ratio to the targets accurately. When the engine operating condition changes

the PIDNN could control the air-fuel ratio to the target within 0.5s and ensure the overshoot under 0.2

which results in the improvements of the engine dynamic property

fuel economy and rapid responsibility.