# Probabilistic digital twins for geotechnical design and construction

**Authors:** Dafydd Cotoarbă, Daniel Straub, Ian FC Smith  
**arXiv:** [2412.09432](https://arxiv.org/abs/2412.09432)  
**Date:** 2024-12-12  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2412.09432v1/

## Key points

- The digital twin approach is a promising solution for AECOM industries but faces limitations due to deterministic models.
- Conventional digital twins lack accuracy in considering uncertainties in AECOM projects, especially in geotechnical design and construction.
- The Probabilistic Digital Twin (PDT) framework has been proposed to address these challenges by integrating uncertainties tailored for geotechnical projects.
- The PDT framework incorporates various sources of uncertainty such as aleatoric, data, model, and prediction uncertainties throughout the modeling process.
- Bayesian methods are used in the PDT for model updating to ensure accurate reflection of site-specific conditions with additional information.
- The probabilistic digital twin framework aims to enhance decision-making processes and improve project outcomes despite substantial uncertainties in AECOM projects.

## Summary

The digital twin approach has emerged as a promising solution to the challenges faced by the Architecture, Engineering, Construction, Operations, and Management (AECOM) industries. However, its widespread application across AECOM sectors has been limited due to certain obstacles. One major hindrance is that conventional digital twins rely on deterministic models that necessitate deterministic input parameters. This limitation compromises their accuracy as they fail to consider the significant uncertainties inherent in AECOM projects, especially in geotechnical design and construction. To overcome this challenge, a groundbreaking Probabilistic Digital Twin (PDT) framework has been proposed. This innovative framework extends traditional digital twin methodologies by integrating uncertainties and tailoring them to meet the specific requirements of geotechnical design and construction projects. The PDT framework offers a structured approach for incorporating various sources of uncertainty such as aleatoric, data, model, and prediction uncertainties throughout the entire modeling process. Moreover, the PDT leverages Bayesian methods for model updating to ensure that site-specific conditions are accurately reflected as additional information becomes available. By addressing uncertainties in a systematic manner, the probabilistic digital twin framework demonstrates its potential to enhance decision-making processes and improve project outcomes even in the face of substantial uncertainties. This refined summary provides an insightful overview of how the Probabilistic Digital Twin (PDT) framework offers a sophisticated solution to address uncertainties in geotechnical design and construction projects within the AECOM industries. Through its structured approach and integration of Bayesian methods for model updating, the PDT shows promise in improving decision-making processes and enhancing project outcomes despite significant uncertainties inherent in these complex projects.