Study On The Automatic Modeling Method Of 3D Information Model For Substations

As a critical intermediate link of power generation,transmission and distribution,substations play a crucial role in normal and stable operations of the power system.Currently,the information management levels at various substations are unbalanced.Most of them use a management approach based on a two-dimensional plane diagram,which cannot fully and truthfully reflect the complex environment of the substations,the crossing of lines and equipment,and the three-dimensional space attributes of the equipment.Currently,manual modeling is carried out by referring to point cloud data using modeling software,constructing the 3D model,followed by the information management of substations in conjunction with the 3D model of the substation.However,the manual modeling process is time-consuming,laborious,and tedious.Therefore,it is a general trend to transform from manual modeling to automatic modeling.However,the traditional automatic mesh generation results based on point cloud data are often unsatisfactory.In response to the shortcomings of existing research methods,the specific works done in this thesis are as follows.(1)This paper proposes an automated modeling method for point cloud data of substations.First,an algorithm is designed to automatically generate a component model library by referring to the standard structure of the substation equipment.Then,the paper improves the Euclidean clustering algorithm to segment disconnected point cloud data in the scene point cloud.Next,by using the SHOT descriptor to find the corresponding point pairs between the model and the scene point cloud and combining the Hough voting algorithm,the method identifies each component and obtains its position information in the scene.Finally,the model is moved into the scene and replaces the point cloud to complete the automated modeling.This method overcomes the time-consuming,laborious,and tedious nature of manual modeling while ensuring the integrity and authenticity of the model.(2)This paper proposes a point cloud registration method for substations based on an improved SAC-IA algorithm,which supplements and improves the recognition and registration module of the automated modeling method.The method imposes constraints on the selection of random points in the SAC-IA algorithm,requiring that the triangle formed by the points selected in the model and scene point clouds must satisfy certain similarity criteria.Next,the error judgment of the SAC-IA algorithm is improved.In addition to using Euclidean distance between the registered model and the scene point cloud as the judgment standard,the SHOT difference and the mean curvature difference are also used as the judgment standard.Only when all three criteria are satisfied can a successful match be declared.Compared with the Hough voting method and the traditional SAC-IA algorithm,this method improves recognition accuracy and registration efficiency.(3)Finally,an automated modeling system for 3D point clouds of substations is developed in this paper,and the automatic modeling method for substation cloud data and the pose recognition method based on the improved SAC-IA algorithm proposed in this paper are applied to the system,which verifies the feasibility and effectiveness of the proposed method.Through practical application cases,it has been proven that the proposed methods are an effective solution to the modeling problem of point cloud data for substations.