Advanced Control Of Overhead Crane

Overhead cranes are widely used in various industrial applications as one of modern logistics equipment.Eliminating and controlling the swing of the load is very important for increasing the work efficiency and safety of a crane. Anti-sway control of a crane is one necessary function as a modern logistics equipment. Dynamic analysis of the crane-load system is the foundation for solving the problem of the crane's fast contraposition and anti-swing.Dynamics equations of a overhead crane systems in three-dimension,two-dimension and one-dimension are established by Lagrange method. In a reasonable range, the nonlinear dynamics equation of the crane system can be simplified to the linear in three-dimension, two-dimension and one-dimension.They provide a theoretical foundation in studying anti-swing of crane.Considering the difficulty and cost of on-site measurement for variables such as the load's swing angle, in this thesis, the corresponding variables can be observed through setting a full-state observer. The full-state observer is designed to observe all variables including the position through collecting the position information.Then, all state variables information is provided to the anti-swing control system.In this study,the state feedback based on pole placement, linear quadratic regulator (LQR) optimal control and proportional integral derivative (PID) control were used for the simulation of the crane anti-swing issues. Simulation results show that the basic methods have the shortage in crane anti-swing. In the analysis of neural network research, an adaptive PID control strategy based on Radial Basis Function (RBF) neural network(NN) is used in anti-swing crane system. Two radial basis function neural network controllers are designed to control the trolley position and load swing separately.RBF neural network self-adaptive PID is based on adaptive learning of RBF neural networks to on-line adjusting PID controller parameters, achieving the best combination of PID control parameters. Simulation results show that the presented controller can realize positioning no static error, no overshoot and rapid elimination of the load swing simultaneously.