Study on forecasting method of power engineering cost based on BIM and DynGCN

Area of Science:

• Engineering Management
• Artificial Intelligence
• Data Science

Background:

• Traditional power engineering cost management faces challenges in precision and dynamism within big data environments.
• Existing methods struggle to provide real-time monitoring and dynamic cost control throughout a project's lifecycle.

Purpose of the Study:

• To propose a novel approach for precise and dynamic cost management in power engineering using BIM and DynGCN.
• To enhance the accuracy of cost prediction for individual engineering stages and optimize overall project construction schemes.

Main Methods:

• Utilizing Building Information Modeling (BIM) for comprehensive whole-life cycle cost management.
• Applying spatiotemporal modeling-based dynamic graph convolutional neural networks (DynGCN) for accurate cost prediction.
• Integrating BIM data with DynGCN for dynamic cost control and scheme optimization.

Main Results:

• BIM technology significantly improves real-time cost monitoring and adjustment capabilities throughout the project lifecycle.
• The DynGCN method achieves a high prediction accuracy of 96% for engineering cost, closely aligning with actual values.
• The combined approach optimizes construction schemes by providing accurate cost predictions for each engineering link.

Conclusions:

• The integration of BIM and DynGCN offers a robust solution for addressing the limitations of traditional power engineering cost management.
• This methodology enhances cost control precision and dynamic adaptability in big data environments.
• The study demonstrates the effectiveness of advanced AI and data management techniques in optimizing large-scale engineering projects.