Abstract: A new intelligent anti-swing control scheme, which combined fuzzy neural network (FNN) and sliding mode control (SMC) with particle swarm optimization (PSO), was presented for bridge crane。 The outputs of three fuzzy neural networks were used to approach the uncertainties of the positioning subsystem, lifting-rope subsystem and anti-swing subsystem。 Then, the parameters of the controller were optimized with PSO to enable the system to have good dynamic performances。 During the process of high-speed load hoisting and dropping, this method can not only realize the accurate position of the trolley and eliminate the sway of the load in spite of existing uncertainties, and the maximum swing angle is only ±0。1 rad, but also completely eliminate the chattering of conventional sliding mode control and improve the robustness of system。 The simulation results show the correctness and validity of this method。82778

Key words: bridge crane; anti-swing control; fuzzy neural network; sliding mode control; particle swarm optimization

1Introduction

Bridge cranes are widely used in industry for transporting heavy loads and hazardous materials in shipping yards, construction sites, steel mills, nuclear power and waste storage facilities and other industrial complexes。 The crane should move the load as fast as possible without any excessive movement at the final position。 However, most of the common bridge crane results in a swing motion when the payload is suddenly removed after a rapid motion。 The swing motion can be reduced finally, but will be time consuming, i。e。, reducing the facility availability as well as productivity。 The failure of controlling crane also might  cause accident and may harm people and surrounding。 For this reason, there has been increasing interest in the design of an anti-swing control scheme for crane system [1−11]。

Sliding mode variable structure control has attracted intensive research due to its robustness to system parameter uncertainties and external disturbances on the sliding surface。 So far, the theory system in the field has been built perfectly and used widely for  practical systems [12−15]。 Some researchers used the variable structure control with sliding modes to control the crane system [14−15]。

A  fuzzy  anti-swing  control  scheme  was proposed

for a three-dimensional overhead crane [1]。 An anti-sway and tracking control by using the input/output linearization approach for harbor mobile cranes and a constant gain partial state feedback controller for rotary cranes were presented [2−3]。 For an overhead crane, a normal proportional and derivative (PD) regulator and a fuzzy cerebellar model articulation controller (CMAC) was designed, and this control can realize both position tracking and anti-swing [11], but the systematic model was simplified as the linear model in these references。 An adaptive sliding mode fuzzy control approach for a two-dimensional overhead crane was studied, and the system displayed high frequency chattering [14]。 Moreover, most researchers have treated the  crane system as a single input system, without considering the changes of lifting-rope。

The particle swarm optimization (PSO) algorithm, developed by KENNEDY and EBERHART, is an evolutionary algorithm which is inspired by the mechanism of biological swarm social behavior such as fish schooling and bird flocking。 It differs from the other evolutionary techniques in the adoption of velocity of inpiduals, and it can search more randomly than genetic algorithm (GA) and avoid falling into the local optimum with faster convergence speed [16−17]。

In order to solve the above problems, aiming at   the

Foundation item: Project(51075289) supported by the National Natural Science Foundation of China; Project(20122014) supported by the Doctor Foundation of Taiyuan University of Science and Technology, China