Original scientific paper

https://doi.org/10.17559/TV-20251031003105

Intelligent Earthwork Identification and Automated Engineering Quantity Estimation in Power Grid Construction with Multi-Source Remote Sensing and BIM

Ou Pu ; Guangxi Power Grid Co., Ltd, Power Grid Planning Research Center, Guangxi Power Grid Co., Ltd, Nanning, Guangxi 530023, China *
Zhen Li ; Guangxi Power Grid Co., Ltd, Power Grid Planning Research Center, Guangxi Power Grid Co., Ltd, Nanning, Guangxi 530023, China
Jihui Tan ; Guangxi Power Grid Co., Ltd, Power Grid Planning Research Center, Guangxi Power Grid Co., Ltd, Nanning, Guangxi 530023, China
Hua Qin ; Guangxi Power Grid Co., Ltd, Power Grid Planning Research Center, Guangxi Power Grid Co., Ltd, Nanning, Guangxi 530023, China
Yi Li ; Guangxi Power Grid Co., Ltd, Power Grid Planning Research Center, Guangxi Power Grid Co., Ltd, Nanning, Guangxi 530023, China
Xianfu Liu ; Guangxi Power Grid Co., Ltd, Power Grid Planning Research Center, Guangxi Power Grid Co., Ltd, Nanning, Guangxi 530023, China

* Corresponding author.

Abstract

Traditional earthwork calculation of power grid construction relies on manual measurement and 2D drawings, and there are some problems such as low data collection efficiency, difficult information integration and lagging calculation results. This paper proposes an earthwork intelligent identification and automatic estimation method based on multi-source remote sensing and BIM (Building Information Model) data fusion. In this method, multi-source data are collected mainly by UAV (Unmanned Aerial Vehicle) tilt photography, supplemented by satellite remote sensing and airborne LiDAR, and pixel-level adaptive fusion is realized by convolutional neural network (CNN) with attention mechanism, and centimeter-level "super" terrain model with better accuracy than single data source is generated, which solves the defects of digital elevation model (DEM) distortion in vegetation area and sparse laser point cloud in exposed area. Build a one click modeling process of "DEM/Digital Terrain Model (DTM) →BIM Terrain Surface → Parameterized Earthwork Family" to achieve 1:1 digital restoration of real terrain, and design parameter changes can drive real-time reconstruction of the model; Establish a remote sensing BIM dynamic linkage mechanism of "data update model reconstruction engineering quantity recalculation", combined with the triangulation volume method to automatically complete the calculation of fill and excavation volume, and embed operations research algorithms to generate the optimal allocation plan with the goal of "minimizing transportation volume/cost". Taking the tower foundation construction area of an ultra-high voltage (UHV) transmission project as the experimental area, the results show that the error between the total earthwork calculation and the reference value of the traditional method is only −0.14%, and the error of a single tower foundation is ≤ 3.5%, which meets the engineering accuracy requirements. The total time consumption of the whole process is shortened from 10 person-days in the traditional method to 2.7 person-days in the traditional method; the efficiency is improved by 73%, and the on-site earthwork transportation volume is reduced by about 35%. This method provides a full-chain solution of "data acquisition-model construction-intelligent calculation-dynamic optimization" for the earthwork of power grid construction, which helps the project to be refined and intelligent management.

Keywords

automatic estimation; BIM; earthwork; engineering quantity; intelligent identification; multi-source remote sensing; power grid construction

Hrčak ID:

350406

URI

https://hrcak.srce.hr/350406

Publication date:

31.8.2026.

Visits: 0 *