Modelling Identification and Controller Design for a Cooling Fan System Subject to Measurement Uncertainty

Abstract

This research proposes a full procedure of modelling system identification control for a cooling fan system First of all a physical model is established based on the aerodynamics Secondly to carry out the parameter identification discretize the model by the bilinear transform With the aid of the minimum difference filter(MDF) the noise effect to the least square regressors can be reduced Though the least square solution provides an optimal set of parameters there exists difference between the real and the fitting data Therefore seek the iteration of the Levenberg–Marquardt algorithm and the system identification will be improved With an optimal set of parameters the speed prediction and fault diagnosis can be realized By setting an upper and lower bound for the predicted speed it is convenient to monitor the status of a cooling fan system In this paper the covered inlet is assumed to be a fault and can be diagnosed by the abnormal speed Furthermore with the nonlinear PI type controller it is possible for the fan to reach a desired speed subject to the model uncertainty and disturbance Last but not least simulation and experiment are performed to verify the feasibility and applicability of the proposed approach

Date of Award	2020
Original language	English
Supervisor	Chao-Chung Peng (Supervisor)