Time-series models guided by digital twin structure improve maintenance predictions
Grounding Time-Series Foundation Models in Digital Twin Topology for Predictive Maintenance
Artificial Intelligence
Summary
Predicting when machines might fail is important for fixing them in time. The authors studied computer models that learn from time-based data but usually focus only on predicting the next data point. They found that models which consider multiple related data streams work better for predicting machine health. By using information about how parts of a machine connect to each other, stored in a digital twin, the authors improved these models further. Their approach helps frozen (already trained) models predict machine failures more accurately without needing big changes.
What this means in practice
- •For maintenance engineering teams: Improve remaining useful life predictions by integrating digital twin structural data with pretrained time-series models for better maintenance scheduling.
- •For industrial iot developers: Build predictive maintenance tools that use topology-informed cross-attention to enhance model accuracy and handle complex machine operating conditions.
Tested on one dataset.
Authors
Sizhe Ma, Katherine A. Flanigan, Mario Bergés
Abstract
Digital twins increasingly support downstream analytical tasks that depend on time-series data, motivating interest in time-series foundation models (TSFMs) as scalable backbones. However, TSFMs are primarily pretrained for temporal continuation and often underperform on unseen tasks such as regression, and systematic empirical comparisons against state-of-the-art dedicated models in digital twin contexts remain limited. This paper makes three contributions. First, we benchmark five well-known TSFMs with frozen backbones on remaining useful life (RUL) prediction using the C-MAPSS dataset, finding that multivariate architectures substantially outperform univariate ones, particularly under varying operating conditions. This raises a deeper question: when cross-channel dependencies can be modeled through pretrained weights, target-task adaptation, and digital twin-derived representations, how much does each contribute, and are they complementary? Second, we propose a topology-informed fusion approach in which topological constraints, derived from the asset structure the digital twin stores among its information models, explicitly shape cross-attention, so that fused representations respect the physical system's local connectivity rather than relying on unconstrained all-to-all interactions. Third, we conduct an ablation study across C-MAPSS subsets of varying operational complexity that isolates the three sources and their interactions. The sources prove complementary rather than redundant, and topology-constrained attention outperforms unconstrained fusion, though by a small margin, enabling a frozen TSFM informed by digital twin representations to remain competitive or in some cases exceed state-of-the-art performance on this regression task.