A Study on The Assessment of Staircase Wear Problem Based on Hybrid Construction Of 1D-2D Models

Abstract

In this study, the scientific method is used to obtain and preprocess the data, and after comparing the results of data clustering visualization by EM and KMeans method, the DBSACN model, which is divided into six clusters, is selected for the clustering and classification of stairs, and the corresponding models are established according to the frequency of staircase use, the preference of direction of use, the use of multiple people, the wear and tear and historical information, and the existence of the time of repair and alteration, etc., such as the one-dimensional linear model, two-dimensional discrete model, multi-step spatial expansion model, and so on. The corresponding models are established, such as one-dimensional linear model, two-dimensional discrete model, and multi-step spatial expansion model, etc. The friction bending model is analyzed by assuming parameters. The friction bending depth and other data are analyzed by assuming the parameters, and differential evolutionary fitting is carried out to obtain the relevant parameters and residual paradigms, and the MSE of the fitting analysis iteration is 0.10, which is verified by the sensitivity analysis, indicating that the model has strong robustness in different scenarios.