Research On Anti-swing Of Crane Loading Based On Adaptive Fuzzy Neural Network Control

With the development of port cargo loading becoming more and more intelligent and automatic,the container crane control requirements are also increasing.Due to the dynamic characteristics of the crane-loading system,container loading will swing during the trolley operational process.This not only reduces the operating efficiency,but also increases the security risk.Therefore,it is worth studying to realize the fast arrival of crane trolley and reduce the swing angle.In order to solve the problem of loading swing,this paper has done the following parts.The dynamic model of crane-loading is built based on Lagrange equation.Due to many uncertainty factors and parameters of the model,accurate mathematical model is difficult to establish.In this paper,ignoring some external uncertainties,simplify the model.Aiming at the swing problem of loading,the adaptive fuzzy PID control method is introduced to implement anti-swing.While fuzzy rules are mostly dependent on expert experience and domain knowledge,it has good robustness.Using fuzzy rules to adjust PID parameters in real time,It can realize the function of easy and accurate control,and can realize real-time control of the time-varying system.Matlab/Simulink platform is used to build the anti-swing system model for simulation.The simulation results show that the control method is more effective than the simple PID control method.Aiming at the problem that the fuzzy rules are too large in the process of multi-dimensional fuzzy inference and PID control is not suitable for complex nonlinear systems,a Takagi-Sugeno(T-S)fuzzy neural control method based on spectrul conjugate gradient method is proposed.The SNPRP conjugate gradient method is used to train the premise parameters and the consequent parameters of the T-S model.This method has sufficient descent and global convergence under strong search conditions.In order to obtain the best controller,using the energy minimum as the index,the linear quadratic optimal control is used to obtain the optimal control matrix of the system,and proved that under this feedback control,it achieved good control effect.Therefore,the input and output group data under this condition is used as training sample data to train the adaptive fuzzy neural network controller.Finally,the trained adaptive fuzzy neural network controller is applied in the crane-loading system for simulation.The results show that this control method in this paper has better control effect and robustness under different rope lengths and working conditions.Since the fuzzy rules cannot be reasonably determined with T-S fuzzy neural network control based on spectrul conjugate gradient method,a T-S fuzzy neural network control method based on improved kernel fuzzy clustering algorithm.Using kernel fuzzy C-means algorithm to and determine the best cluster center.A Krill optimization algorithm,which is improved by introducing the adaptive caucy mutation,is used to optimize the premise parameters and the consequent parameters of the T-S neural network model.Then the anti-swing controller of crane system is designed.In order to improve the accuracy of clusting,a fitness function is constructed by using the distance between classes,before the optimal cluster center is obtained by improved Krill optimization algorithm.Simulation on three test functions verifies the performance of improved Krill optimization algorithm.Finally,the linear quadratic controller was approached by the proposed method.The control effect is good in achieving the anti-swing control of the crane-loading system.