Research On Robust Sliding Mode Control Method Of Particleboard Dosing System Based On State Observer

As one of the three kinds of wood-based panels,particleboard plays an important role in wood processing industry with its advantages of high production technology,uniform physical structure and good processing performance.In the production process of particleboard,the glue dosing link is one of the important processes,which marks the production technology of particleboard,and is also the key to affect the quality of finished products.However,in the process of particleboard glue dosing,the pressure change of the adhesive in the pipeline will produce a certain reaction force,and the sensitivity of the sensor,the stability of electrical signal transmission,the inertia of the actuator and other factors will cause the system to be affected by uncertainties.It is an urgent problem that apply glue accurately and quickly under complex working conditions.Furthermore,the state variables of the particleboard glue dosing system are composed of the glue flow rate and acceleration of the output glue.And the state variables available are often required for the glue flow control.Due to the physical characteristics of the adhesive itself(such as adhesive viscosity,gelation speed,etc.),the state variable of the glue dosing system are difficult to measure accurately,and the economic cost of the measurement sensor is too high.In this case,how to carry out robust control is also worth studying.In the category of system and control,the state observer estimates the variables through the fusion of known information.As a virtual sensor,the state observer ensures the performance of the closed-loop control algorithm.As one of the robust control methods,sliding mode control switches the structure of the controller according to the degree that the system state deviates from the sliding mode,and make the system operates according to the law specified by the sliding mode to ensure the robust performance.Therefore,this dissertation studies robust sliding mode control based on the state observer to solve the problem of timedelay,steady-state performance and transient performance.The main work is as follows:The mathematical model is established according to each link of glue dosing process,and the recursive least square identification method is used to identify the parameters of the model.Considering that "data saturation" will affect the identification accuracy with the increase of iteration times in the identification process,the fuzzy dynamic forgetting factor is introduced to erase the old data,and the identification ability is improved by adjusting the forgetting factor in real time.For the steady-state performance problem of particleboard dosing system,a backstepping sliding mode controller is designed.The virtual control law is designed by backstepping method to realize the global adjustment of the system.Considering the "differential term explosion" problem of backstepping method,the dynamic surface control method is used to reduce the complexity of system calculation.At the same time,the sliding mode control method is combined with the dynamic surface to improve the robustness of the system.In the process of glue dosing control,a full order state observer is introduced to observe and estimate the state of the system.Considering that the design of full order observer requires an accurate system model,a high gain observer is designed to observe the system state more quickly and accurately without relying on the system model.The simulation results show that the proposed method can observe the system state accurately and quickly,and the robustness of the glue dosing system can be effectively improved.For the problem of time-delay robust control in particleboard dosing system,a neural network state observer is designed to compensate the observation of time-delay approximate deviation,and then a dynamic surface sliding mode control method is introduced to treat the part that cannot be observed by the observer and disturbance as unmodeled dynamics to improve the control effect of glue dosing system on time-delay.Due to the large number of controller parameters,the adaptive genetic optimization algorithm is used to adjust the parameters to obtain the globally optimal controller parameters,which could improve the robustness of the controller,and the effectiveness of the proposed method is verified by simulation experiments.For the transient performance problem of particleboard glue dosing system,the prescribed performance control method is used to control the tracking error within the preset small boundary,so as to ensure the convergence for the speed and the range of tracking error.For the reason that the uncertainty can not be covered by the control input,the dual extended state observers are introduced to compensate the observation.Combined with the sliding mode dynamic surface control method,the transient performance can be improved while ensuring the steady-state performance and state observation of glue dosing system by the simulation results.