Search

It is difficult to implement modeling for overall construction due to the complex structure and distinct environmental influences for the high-speed elevator so researchers rarely achieve end-to-end control for vibration suppression. Meanwhile, the dynamic parameters of the hoist component (the hoist rope) contain more environmental disturbance characteristics than the traction machine. In this paper, a sensitivity analysis of data channels of hoist rope that can help to clarify the effective control loop based on experimental results is implemented. The authors combined a neural network-based (NN-based) regression model and attention module for predicting hoisting rope vibration to achieve over 95% accuracy, and determine the channel data that contributes most to vibration prediction.

Modeling and characterizing high-dimensional aerodynamic pressures, on the elevator in the hoistway, is very challenging. An accurate model is helpful to understand and analyze the pressure characteristics, which is a vital consideration in the design of a competitive elevator. The current full-order models are usually governed by the Navier–Stokes equations and have low computational efficiency. A reduced-order model is thus necessary to improve computational efficiency. This work aims at investigating two data-driven approaches, for modeling and characterizing the aerodynamic pressures, i.e., proper orthogonal decomposition (POD) and dynamic mode decomposition (DMD) that are able to extract temporal–spatial structures from the data. A three-dimensional (3D) model of the realistic elevator is first built and simulation data of the aerodynamic pressures during the motion of the elevator in the hoistway is generated via computational fluid dynamics (CFD). Then, POD and DMD are employed to analyze the simulation data. It is found that through clustering techniques and since there exist local aerodynamic pressure pulses during the operation of the elevator, the aerodynamic pressure on the elevator has some distinct patterns. Therefore, cluster-based POD and DMD are further employed in the analysis. The results demonstrate that the cluster-based POD and DMD can achieve lower reconstruction errors than POD and DMD.