Analysis Of Dynamic Characteristics Of Multi-chamber Nitrogen Oxygen Sensor And Research On Decoupling Control Strategy

Despite the progress of human society promoted by the rapid development of the automobile industry,the boom inevitably leads to more environmental pollution and energy shortage due to the restrictions on power sources.Hence,various countries have formulated increasingly stringent laws and regulations on pollution emissions.However,traditional vehicle-mounted oxygen sensors can only measure the oxygen content in tail gas but fail to detect the nitrogen oxides in exhaust gas as required by laws.In such a context,the research and development of high-precision and fast-response NOx sensors have become a challenging research topic at home and abroad.NO_x sensors cannot achieve high-precision and fast measurement under complicated working conditions with traditional PID control methods due to its time-varying characteristics and coupling relationships among chambers.Moreover,it’s hard to build an accurate mechanism model because of its complex internal mechanism.To tackle the problem,a NOx sensor test platform and a semi-physical simulation platform were established by the thesis to determine the coupling relationships among pump units by an experimental analysis of the relationships among the pump units.In addition,a control algorithm was proposed for the NOx sensor pump units with medium-and-high-temperature multi-chambers as required.The proposed method is specified as follows: 1.The adaptive PID control of the main pump.The thesis examined the combination types of neuron and traditional PID control,namely single neuron PID adaptive control and neural network-based PID adaptive control.2.The decoupling control of the auxiliary pump and the measuring pump,which was studied based on the feedforward compensation decoupling,neural network PID decoupling,and dynamic matrix control decoupling.The proposed control algorithm with self-adaptability can track system characteristics and solve time-varying optimization problems;furthermore,it can eliminate the coupling among pump units to control the major atmospheric influencing units without building an accurate mathematical model on the controlled objects.In this thesis,simulation experiments were performed on a variety of algorithms using tools including MATLAB/Simulink and Speedgoat hardware-in-the-loop simulation platform.The analysis result suggests that in terms of the adaptive control and decoupling control of NOx sensor pump units,the proposed algorithm has significantly better control efficiency than the traditional single closed loop PID feedback control.