LM-CNN-based Automatic Cost Calculation Model for Power Transmission and Transformation Projects

The cost calculation of power transmission and transformation project is the core part of cost control technology. The quality of the cost calculation model directly affects the efficiency and reliability of the cost management of power transmission and transformation projects. However, the existing...

Full description

Saved in:
Bibliographic Details
Main Authors: Xiaolin WU, Ling LUAN, Lianwu PAN, Hailong LI
Format: Article
Language:zho
Published: State Grid Energy Research Institute 2023-02-01
Series:Zhongguo dianli
Subjects:
Online Access:https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202103063
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The cost calculation of power transmission and transformation project is the core part of cost control technology. The quality of the cost calculation model directly affects the efficiency and reliability of the cost management of power transmission and transformation projects. However, the existing models cannot reconcile the computational speed, accuracy and stability. Considering above-mentioned problems, firstly, a convolutional neural network model is constructed with its input and output determined according to the practical cost requirements of the power transmission and transformation projects. Then, the historical cost data are input into the network model as samples to calculate the network output. Finally, in view of the big difference between the expected output and the actual output, the Levenberg-Marquart algorithm is utilized to optimize the weight parameters of the convolutional neural network to complete the model training. Compared with the BP neural network and GD-CNN, the proposed model with higher prediction accuracy and stability combines the advantages of Levenberg-Marquart algorithm and convolutional neural network model to improve the calculation effect of power transmission and transformation project cost.
ISSN:1004-9649