Research On Anti-Swing Control Strategy Of Bridge Crane Based On RBFNN

Bridge crane is a kind of practical suspension transportation equipment,because its technology is mature,does not occupy the ground space,the job scope is large and so on,the various production links of many enterprises are widely used.With the improvement of industrial automation and production efficiency,the demand for working efficiency and safety of bridge crane is also becoming higher and higher,especially in metallurgy,aerospace,assembly and other fields which require higher precision of swing control.Due to the flexible wire rope connection between the crane and the object,the change of the object speed lags behind the change of the vehicle speed when the motor driven trolley is loading and decelerating.The effective suppression of swing is not only an important means to improve the working efficiency,but also a necessary measure to improve the safety of production.Based on this,the research on anti-swing control has important practical significance.Based on the sufficient study of domestic and overseas literatures and practical applications of anti-swing control,this paper proposes two kinds of intelligent control strategies with different ideas according to the characteristics of crane high-order nonlinear mathematical model.The main work and achievements are as follows:(1)Firstly,the two-dimensional dynamics model of the trolley-crane system is established by using analytical mechanics,and its state-space equation is deduced.Then,the stability,controllability and observability of the system are analyzed by using Lyapunov stability theorem,which provides theoretical support for the subsequent research.(2)Aiming at the problem that fuzzy PID controller parameter adjustment is not accurate enough,this paper designs fuzzy RBF network PID controller for crane anti-swing control.Fuzzy control is used to introduce the driver’s operating experience into PID controller to make it have the ability of parameter adjustment.However,due to the limited number of fuzzy control rules and the large subjectivity of the formulation,the control precision is relatively rough.RBF neural network(RBF Neural Network,RBFNN)is introduced to improve the fuzzy rules and enhance the adjustment ability of the controller.The simulation results indicate that the performance of the proposed control method is prominently improved compared with the fuzzy PID control.It can realize accurate positioning without overshoot,and the swing angle suppression effect is obvious,which certifies the high efficiency of the control method.(3)In view of the problem that improper initial value selection of RBFNN parameter is easy to affect the network performance,this paper designs an improved genetic optimization RBFNN supervisory controller for crane anti-swing control.In view of the excellent learning ability of RBFNN,the RBFNN supervisory control is introduced into the anti-swing control.Among them,PD controller has the characteristics of fast response speed and strong robustness,but the control accuracy is low.RBFNN is used to continuously learn its output until it is completely replaced,realizing supervision control and effectively improving the control accuracy.However,the performance of RBFNN is closely related to the selection of the initial value of the Gaussian basis function parameter.Genetic algorithm is introduced to search the initial value globally.Because the genetic algorithm may be slow in convergence and easy to fall into the local optimal when facing complex problems,this paper improves the genetic algorithm from many aspects.The simulation results indicate that the proposed method has faster response speed and better suppression effect on swing angle than fuzzy RBF network PID control,and the controller performance is verified under different rope lengths,which proves the effectiveness of the proposed method.(4)In order to further verify the effect of the designed anti-swing controller,this paper designed a software algorithm and built a simulation experiment platform to verify the anti-swing effect of the improved genetically optimized RBFNN supervisory control.The results indicate that this method can realize precise positioning of the trolley,and the swing angle control basically conforms to the design anticipation,which certifies the practicable of this control method.