WEST BETHESDA, Md. –
Alongside the Navy and Naval Sea Systems Commands (NAVSEA), other organizations like the U.S. Air Force and NASA are developing cyber-physical systems now known as digital twin.
Managing and processing data has become a necessity across every organization, according to Dr. Dave Drazen, the Digital Twin Program manager for Naval Surface Warfare Center, Carderock Division. Drazen described digital twin as a virtual representation of a physical system that utilizes data and physical models to provide deeper insight into platform capabilities.
Digital twin is a continuous analytical fusion of data, physics-based models and machine learning to prescribe multiple future instantiations of ships and ship systems, which enables the user to readily identify the optimum choices. Carderock’s Dr. Ben Grisso, a mechanical engineer in the Criteria and Assessment Branch (Code 654), and Dr. Aly Mondoro, an engineer in the Performance Evaluation Branch (Code 653), have been working to perfect their digital twin in relation to the structural-health monitoring (SHM). During a Jan. 16 brown bag at Carderock’s West Bethesda, Maryland, headquarters on the structural health monitoring model, they set out to explain how.
SHM is the process of implementing a data-driven damage identification and lifetime load-monitoring strategy achieved by distinguishing different characteristics of a structure as it operates. These characteristics include: the operational and environmental loads; mechanical damage caused by these loads; the growth of the damage during operation; and the future performance of the structure due to the cumulative effects.
The goal of Grisso and Mondoro’s SHM digital twin is to provide real-time operational, situational awareness, extend platform service life, improve maintenance planning and increase asset availability through prognostics. The SHM hardware and detection algorithms are working to do just that. Currently, the SHM is used to detect, locate, characterize and track changes to a structure instantaneously. Once detected, they can facilitate event monitoring and provide structural self-awareness for autonomous operations while also enhancing mission-readiness assessments.
However, Mondoro and Grisso said there are challenges in that digital twin and SHM are unique in scope and complexity, and it is unlikely that these issues can be solved in a single industry.
The first challenge that the SHM and digital twin undertook was in respect to fatigue damage on Navy ships. They used a three-step approach: use global strain gauges to estimate loads acting on the ship; use a physics-based model to determine the vulnerable stress locations; and use a damage model to estimate the current state. Their research determined that the assumption needed to be that “… the stress response is a linear combination of the response to global vertical and lateral bending are the only significant global loads,” Grisso said. Using this data, they could determine where the critical stresses were occurring in unmonitored locations. Once this has occurred, they can perform a damage assessment.
Using the SHM and digital twin in this matter poses several benefits, such as a better understanding of how to address uncertainties. The work of these programs is ongoing, but according to Grisso and Mondoro, the work they’ve already done has been validated by the results. Although, they say that the biggest limitations to the SHM and digital twin is implementation, data storage and security requirements. Through this research, they have also found alternative approaches to solving the same problems, including calibration factors through different tools and modal decomposition.
“This project is giving us a better idea of how the ship is being used, how it is responding and the largest benefit of it is an enhanced understanding of the operational loads. Assumptions and limitations of the digital twin are tied to the analysis tools used, as well as the additional assumptions being made when the data, multi-scale physics-based models and damage models are integrated,” Mondoro said, emphasizing the benefits of their project.
Their current approach provides an enhanced understanding of loads and enables the estimation response at unmonitored locations. This reduces the cost and weight of the SHM, but introduces some uncertainties in the data it produces.
“It is always better to measure the structures directly with a sensor like structural-health monitoring than to just assume. We need to do this until we have a better understanding of our models and the variability associated with them,” Grisso said. “Obviously, we are not doing this work in a vacuum, and we hope to eventually take it across the entire naval enterprise.”
In closing, Mondoro and Grisso mentioned the additional scope of their SHM digital twin, which includes using point-cloud data fused with finite element analysis to aid in accounting for distortions and estimating strength; temperature data coupled with sensitization models and strain data to aid in stress corrosion crack monitoring; local strain gauges for assessing probabilistic crack growth for critical locations; the integration of acoustic emission data with crack-growth models and inspection reports for estimating crack growth; motions data and radar data for the estimation of wave environment; and shipboard guidance.