Research On Anti-Sway And Positioning Control Of Rotary Crane System Based On Sliding Mode Control

In recent years,with the rapid development of our country,infrastructure construction has become more important.Cranes,as large-scale transport equipment,play an indispensable role in factories,docks and other places.Because of its simple structure and high flexibility,rotary crane is more and more used in narrow space and heavy tasks.However,the operation of the rotary arm often results in unnecessary oscillations of the hooks and loads in the condition of increasing or decreasing speed,which not only affects the efficiency of the project,but also may cause unexpected injury events.In addition,the weight of the load and the length of the sling can not be ignored in practice,which means that the sling and the load can cause complex secondary swing characteristics and bring severe test to the operator.For the safe and efficient operation of the crane system,this paper takes the double-pendulum rotary crane as the object,and studies the suppression of two-stage swing angle and the positioning of the rotary arm based on the sliding mode control theory.First,considering the double pendulum characteristics,the dynamic model of the rotary crane is constructed by Lagrange method.In order to simplify the dynamic equation,some reasonable assumptions are put forward,and a disturbance observer is introduced.While compensating for external disturbances,the complex dynamic equation is simplified to a linear matrix form,which also makes the crane system decomposed into two subsystems with similar structure.Subsequently,based on the subsystem model,this paper studies from three aspects.A fast terminal sliding mode control strategy based on radial basis function neural network is presented for double-pendulum rotary crane system with uncertain model parameters.In order to improve the transportation efficiency of the crane,the terminal sliding mode control strategy is used to obtain the limited-time convergence.In the design of sliding surface,a linear term is added to improve the convergence rate of the system state.Radial basis function neural network is used to approximate the parameter uncertain part of the system model,and a compensation controller is designed.The hyperbolic tangent function is used instead of the symbolic function switching control to obtain smoother input performance.The stability of the crane system under this strategy is proved by Lyapunov theory,and the method is proved to be effective in angle suppression and arm tracking by simulation.Further,a fast non-singular terminal sliding mode control strategy based on extreme learning machine is presented for the above crane system to avoid singular problems.In order to simplify the parameter design process of the neural network,the extreme learning machine algorithm with random parameter assignment is used,which has excellent generalization ability and training speed.The proposed fast nonsingular sliding mode control not only improves the convergence speed of the state variable,but also effectively eliminates the singular problem of infinite input moment.At the same time,the double power reaching law is used to speed up the approximation of state variables away from the sliding surface.The simulation verifies that the control strategy has better dynamic performance,can quickly locate the rotary arm,eliminate the oscillation of swing angle,and is robust to uncertain disturbances.Finally,an adaptive fast integral terminal sliding mode control strategy with asymmetric input constraints is presented for a double-pendulum rotary crane system with uncertain model parameters and saturated input.In order to satisfy the engineering requirements flexibly,an asymmetric restriction function is designed for the actuator of the rotary crane.At the same time,the Gaussian error function is used to obtain smooth control input after restriction.The control precision is greatly improved by incorporating an integral term into the controller design,which eliminates the steady-state error and avoids singular problems.The coefficients of each item in the controller are adjusted by an adaptive strategy,which simplifies the artificial tuning process.The unknown upper bounds of the limited error and the parameter uncertain disturbance are also estimated by the adaptive method.The method is proved to be stable by Lyapunov theory,and simulation shows that the control input is limited between the upper and lower bounds of the settings.Under the input constraints,the control target of the sway position can be achieved.