Predictive Control For Wind Tunnel Flow Field Based On Multiple Models

Wind tunnel is an important ground test equipment for the development of aircraft,because the complexity of the aircraft working,we put forward almost harsh requirements for wind tunnel flow field precision,a single Predictive Control is able to realize high-precision control problem,because its optimizes control law based on predictive future control deviation,multi-model control can show the superiority under the complicated condition.Therefore,this thesis takes an ejector type transonic wind tunnel as the plant,on the basis of multiple models,employing Predictive Control to realize high-precision wind tunnel flow field.Because the wind tunnel flow conditions are complicated,the step response data foreach condition are different,therefore the wind tunnel field prediction models are different with the step response models are different under each condition,so we need to establish their own prediction model and controller respectively under different conditions,this led to a multi-model problem.Considering the superiority of multi-model control in solving nonlinear problems,therefore on the base of multi-model,we design wind tunnel flow field controller using predictive control algorithm in this paper,so the multi-model and the predictive control can be combined effectively.Multi-model control is a method based on decomposition-synthesis strategy,specific steps are: the first step is the decomposition of the problem,divided the nonlinear wind tunnel system into several linear subintervals according to different working conditions,each subinterval needs to establish its own predictive model and DMC controller,next step is the synthesis of the solution,through the controller's weighting strategy to achieve the purpose of tracking system dynamics.In order to describe the dynamic characteristics of the wind tunnel system more accurately,after synthesizing the structure of the multi-model predictive control,combining characteristics with wind tunnel flow field of non-linear and short blowing time,we selected the controller weighted multi-model predictive control,there may be some kind of interference in the blowing test,so that the current time error is not accurate,so this paper uses the matching error method,this method involves the calculation of weight coefficients,the error matching method is adopted in this paper,it takes into account both the current timeerror and the historical moment of error,avoids the effects of random errors effectively,finally,take the product of calculated weighting coefficient and the control rate as the totalcontrol rate to control the wind tunnel flow field.Finally,multi-model control and single model control are simulated respectively,simulation results show that,multi-model controller can regulate the flow field of the windtunnel effectively,so that the blow test can achieve the accuracy requirements,while singlemodel predictive control can't meet the requirements of nonlinear system,illustrating the effectiveness of the multi-model DMC controller.