A neural network-based input shaping for swing suppression of an overhead crane under payload hoisting and mass variations. (July 2018)

• Record Type: Journal Article
• Authors: Ramli, Liyana; Mohamed, Z.; Jaafar, H.I.
• Is Part Of: Mechanical systems and signal processing. Volume 107(2018)
• Page Start: 484
• Page End: 501
• Publication Date: 2018-07
• Subjects: Hoisting -- Input shaping -- Neural network -- Overhead crane -- Swing suppression; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2018.01.029
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: We propose a neural network-based input shaping for control of payload swing. The algorithm predicts optimal parameters and update the shaper in real time. Experiments on a crane show the controller achieves the highest swing reductions. The shaper also has a less duration resulting in a satisfactory speed of response. Abstract: This paper proposes an improved input shaping for minimising payload swing of an overhead crane with payload hoisting and payload mass variations. A real time unity magnitude zero vibration (UMZV) shaper is designed by using an artificial neural network trained by particle swarm optimisation. The proposed technique could predict and directly update the shaper's parameters in real time to handle the effects of time-varying parameters during the crane operation with hoisting. To evaluate the performances of the proposed method, experiments are conducted on a laboratory overhead crane with a payload hoisting, different payload masses and two different crane motions. The superiority of the proposed method is confirmed by reductions of at least 38.9% and 91.3% in the overall and residual swing responses, respectively over a UMZV shaper designed using an average operating frequency and a robust shaper namely Zero Vibration Derivative-Derivative (ZVDD). The proposed method also demonstrates a significant residual swing suppression as compared to a ZVDD shaper designed based on varying frequency. In addition, the significant reductions are achieved with a less shaper duration resulting in a satisfactory speed of response. It is envisaged that the proposed method can be used for designing effective input shapers for payload swing suppression of a crane with time-varying parameters and for a crane that employ finite actuation states.