[Paper Review] Multifidelity digital twin for real-time monitoring of structural dynamics in aquaculture net cages

Eirini Katsidoniotaki, Biao Su+2 more|arXiv (Cornell University)|Jun 6, 2024

Water Quality Monitoring Technologies5 citations

TL;DR

The paper develops a multifidelity surrogate modeling framework using nonlinear autoregressive Gaussian processes (NARGP) to build a digital twin for real-time monitoring of aquaculture net cage dynamics, validated at a full-scale SINTEF ACE farm, and compares GP-PCA surrogates with graph convolutional networks (GCNs).

ABSTRACT

As the global population grows and climate change intensifies, sustainable food production is critical. Marine aquaculture offers a viable solution, providing a sustainable protein source. However, the industry's expansion requires novel technologies for remote management and autonomous operations. Digital twin technology can advance the aquaculture industry, but its adoption has been limited. Fish net cages, which are flexible floating structures, are critical yet vulnerable components of aquaculture farms. Exposed to harsh and dynamic marine environments, the cages experience significant loads and risk damage, leading to fish escapes, environmental impacts, and financial losses. We propose a multifidelity surrogate modeling framework for integration into a digital twin for real-time monitoring of aquaculture net cage structural dynamics under stochastic marine conditions. Central to this framework is the nonlinear autoregressive Gaussian process method, which learns complex, nonlinear cross-correlations between models of varying fidelity. It combines low-fidelity simulation data with a small set of high-fidelity field sensor measurements, which offer the real dynamics but are costly and spatially sparse. Validated at the SINTEF ACE fish farm in Norway, our digital twin receives online metocean data and accurately predicts net cage displacements and mooring line loads, aligning closely with field measurements. The proposed framework is beneficial where application-specific data are scarce, offering rapid predictions and real-time system representation. The developed digital twin prevents potential damages by assessing structural integrity and facilitates remote operations with unmanned underwater vehicles. Our work also compares GP and GCNs for predicting net cage deformation, highlighting the latter's effectiveness in complex structural applications.

Motivation & Objective

• Motivate real-time monitoring and autonomous operation needs for offshore aquaculture net cages under harsh marine environments.
• Propose a multifidelity surrogate modeling framework that fuses low-fidelity simulations with scarce high-fidelity field data.
• Implement and validate a digital twin at a full-scale fish farm to predict cage displacements and mooring loads.
• Evaluate whether nonlinear autoregressive Gaussian processes (NARGP) can efficiently fuse fidelity levels for real-time predictions.

Proposed method

• Use NARGP to learn nonlinear cross-correlations between low-fidelity simulations and high-fidelity sensor data.
• Train low-fidelity surrogates from FhSim simulations mapping currents to mooring loads and cage displacements.
• Apply PCA to reduce net-cage deformation outputs for GP modeling, then reconstruct full deformation.
• Fuse low- and high-fidelity GP models in a recursive, multifidelity framework with real-time metocean inputs.
• Compare Gaussian process surrogates with a PCA-preprocessing against graph convolutional networks (GCNs) for net-cage topology prediction.
• Evaluate computational efficiency enabling on-the-fly digital twin operation.

Experimental results

Research questions

• RQ1 Can a multifidelity surrogate framework accurately predict net cage displacements and mooring loads under stochastic metocean conditions?
• RQ2 Do GCNs offer advantages over GP-based surrogates for predicting complex net-cage deformation topology?
• RQ3 Is real-time digital twin performance feasible with limited high-fidelity data and abundant low-fidelity simulations?
• RQ4 How does the integration of metocean data improve surrogate-based predictions in a full-scale aquaculture setting?

Key findings

• The multifidelity GP model significantly improves mooring line load predictions over the low-fidelity model.
• The multifidelity framework yields predictions that closely align with unseen field measurements for mooring loads.
• For net cage displacement, the multifidelity approach improves over the low-fidelity GP in matching sensor data over time.
• GCNs can effectively handle the complex topology of net cages and offer competitive deformation predictions compared to GP-based surrogates.
• The workflow enables real-time monitoring and supports remote autonomous operations with reduced sensor deployment and maintenance.