The dimension of models derived on the basis of data is commonly restricted by the number of observations, or in the context of monitored systems, sensing nodes. This is particularly true for structural systems, which are typically high-dimensional in nature. In the scope of physics-informed machine learning, this article proposes a framework—termed neural modal ordinary differential equations (Neural Modal ODEs)—to integrate physics-based modeling with deep learning for modeling the dynamics of monitored and high-dimensional engineered systems. In this initiating exploration, we restrict ourselves to linear or mildly nonlinear systems. We propose an architecture that couples a dynamic version of variational autoencoders with physics-informed neural ODEs (Pi-Neural ODEs). An encoder, as a part of the autoencoder, learns the mappings from the first few items of observational data to the initial values of the latent variables, which drive the learning of embedded dynamics via Pi-Neural ODEs, imposing a modal model structure on that latent space. The decoder of the proposed model adopts the eigenmodes derived from an eigenanalysis applied to the linearized portion of a physics-based model: a process implicitly carrying the spatial relationship between degrees-of-freedom (DOFs). The framework is validated on a numerical example, and an experimental dataset of a scaled cable-stayed bridge, where the learned hybrid model is shown to out perform a purely physics-based approach to modeling. We further show the functionality of the proposed scheme within the context of virtual sensing, that is, the recovery of generalized response quantities in unmeasured DOFs from spatially sparse data.

In various scenarios, the motion of a tracked object, for example, a pointing apparatus, pedestrian, animal, vehicle, and others, is driven by achieving a premeditated goal such as reaching a destination. This is albeit the various possible trajectories to this endpoint. This paper presents a generic Bayesian framework that utilizes stochastic models that can capture the influence of intent (viz., destination) on the object behavior. It leads to simple algorithms to infer, as early as possible, the intended endpoint from noisy sensory observations, with relatively low computational and training data requirements. This framework is introduced in the context of the novel predictive touch technology for intelligent user interfaces and touchless interactions. It can determine, early in the interaction task or pointing gesture, the interface item the user intends to select on the display (e.g., touchscreen) and accordingly simplify as well as expedite the selection task. This is shown to significantly improve the usability of displays in vehicles, especially under the influence of perturbations due to road and driving conditions, and enable intuitive contact-free interactions. Data collected in instrumented vehicles are shown to demonstrate the effectiveness of the proposed intent prediction approach.

As characterization and modeling of complex materials by phenomenological models remains challenging, data-driven computing that performs physical simulations directly from material data has attracted considerable attention. Data-driven computing is a general computational mechanics framework that consists of a physical solver and a material solver, based on which data-driven solutions are obtained through minimization procedures. This work develops a new material solver built upon the local convexity-preserving reconstruction scheme by He and Chen (2020) A physics-constrained data-driven approach based on locally convex reconstruction for noisy database. Computer Methods in Applied Mechanics and Engineering 363, 112791 to model anisotropic nonlinear elastic solids. In this approach, a two-level local data search algorithm for material anisotropy is introduced into the material solver in online data-driven computing. A material anisotropic state characterizing the underlying material orientation is used for the manifold learning projection in the material solver. The performance of the proposed data-driven framework with noiseless and noisy material data is validated by solving two benchmark problems with synthetic material data. The data-driven solutions are compared with the constitutive model-based reference solutions to demonstrate the effectiveness of the proposed methods.

This work is concerned with damage detection in a commercial 52-meter wind turbine blade during fatigue testing. Different artificial damages are introduced in the blade in the form of laminate cracks. The lengths of the damages are increased manually, and they all eventually propagate and develop into delaminations during fatigue loading. Strain gauges, acoustic emission sensors, distributed accelerometers, and an active vibration monitoring system are used to track different physical responses in healthy and damaged states of the blade. Based on the recorded data, opportunities and limitations of the different sensing systems for blade structural health monitoring are investigated.

Data-informed predictive maintenance planning largely relies on stochastic deterioration models. Monitoring information can be utilized to update sequentially the knowledge on model parameters. In this context, on-line (recursive) Bayesian filtering algorithms typically fail to properly quantify the full posterior uncertainty of time-invariant model parameters. Off-line (batch) algorithms are—in principle—better suited for the uncertainty quantification task, yet they are computationally prohibitive in sequential settings. In this work, we adapt and investigate selected Bayesian filters for parameter estimation: an on-line particle filter, an on-line iterated batch importance sampling filter, which performs Markov Chain Monte Carlo (MCMC) move steps, and an off-line MCMC-based sequential Monte Carlo filter. A Gaussian mixture model approximates the posterior distribution within the resampling process in all three filters. Two numerical examples provide the basis for a comparative assessment. The first example considers a low-dimensional, nonlinear, non-Gaussian probabilistic fatigue crack growth model that is updated with sequential monitoring measurements. The second high-dimensional, linear, Gaussian example employs a random field to model corrosion deterioration across a beam, which is updated with sequential sensor measurements. The numerical investigations provide insights into the performance of off-line and on-line filters in terms of the accuracy of posterior estimates and the computational cost, when applied to problems of different nature, increasing dimensionality and varying sensor information amount. Importantly, they show that a tailored implementation of the on-line particle filter proves competitive with the computationally demanding MCMC-based filters. Suggestions on the choice of the appropriate method in function of problem characteristics are provided.

We propose using fully Bayesian Gaussian process emulation (GPE) as a surrogate for expensive computer experiments of transport infrastructure cut slopes in high-plasticity clay soils that are associated with an increased risk of failure. Our deterioration experiments simulate the dissipation of excess pore water pressure and seasonal pore water pressure cycles to determine slope failure time. It is impractical to perform the number of computer simulations that would be sufficient to make slope stability predictions over a meaningful range of geometries and strength parameters. Therefore, a GPE is used as an interpolator over a set of optimally spaced simulator runs modeling the time to slope failure as a function of geometry, strength, and permeability. Bayesian inference and Markov chain Monte Carlo simulation are used to obtain posterior estimates of the GPE parameters. For the experiments that do not reach failure within model time of 184 years, the time to failure is stochastically imputed by the Bayesian model. The trained GPE has the potential to inform infrastructure slope design, management, and maintenance. The reduction in computational cost compared with the original simulator makes it a highly attractive tool which can be applied to the different spatio-temporal scales of transport networks.

Model order reduction (MOR) methods enable the generation of real-time-capable digital twins, with the potential to unlock various novel value streams in industry. While traditional projection-based methods are robust and accurate for linear problems, incorporating machine learning to deal with nonlinearity becomes a new choice for reducing complex problems. These kinds of methods are independent to the numerical solver for the full order model and keep the nonintrusiveness of the whole workflow. Such methods usually consist of two steps. The first step is the dimension reduction by a projection-based method, and the second is the model reconstruction by a neural network (NN). In this work, we apply some modifications for both steps respectively and investigate how they are impacted by testing with three different simulation models. In all cases Proper orthogonal decomposition is used for dimension reduction. For this step, the effects of generating the snapshot database with constant input parameters is compared with time-dependent input parameters. For the model reconstruction step, three types of NN architectures are compared: multilayer perceptron (MLP), explicit Euler NN (EENN), and Runge–Kutta NN (RKNN). The MLPs learn the system state directly, whereas EENNs and RKNNs learn the derivative of system state and predict the new state as a numerical integrator. In the tests, RKNNs show their advantage as the network architecture informed by higher-order numerical strategy.

Anomaly detection in asset condition data is critical for reliable industrial asset operations. But statistical anomaly classifiers require certain amount of normal operations training data before acceptable accuracy can be achieved. The necessary training data are often not available in the early periods of assets operations. This problem is addressed in this paper using a hierarchical model for the asset fleet that systematically identifies similar assets, and enables collaborative learning within the clusters of similar assets. The general behavior of the similar assets are represented using higher level models, from which the parameters are sampled describing the individual asset operations. Hierarchical models enable the individuals from a population, comprising of statistically coherent subpopulations, to collaboratively learn from one another. Results obtained with the hierarchical model show a marked improvement in anomaly detection for assets having low amount of data, compared to independent modeling or having a model common to the entire fleet.

In the present work, neural networks are applied to formulate parametrized hyperelastic constitutive models. The models fulfill all common mechanical conditions of hyperelasticity by construction. In particular, partially input convex neural network (pICNN) architectures are applied based on feed-forward neural networks. Receiving two different sets of input arguments, pICNNs are convex in one of them, while for the other, they represent arbitrary relationships which are not necessarily convex. In this way, the model can fulfill convexity conditions stemming from mechanical considerations without being too restrictive on the functional relationship in additional parameters, which may not necessarily be convex. Two different models are introduced, where one can represent arbitrary functional relationships in the additional parameters, while the other is monotonic in the additional parameters. As a first proof of concept, the model is calibrated to data generated with two differently parametrized analytical potentials, whereby three different pICNN architectures are investigated. In all cases, the proposed model shows excellent performance.

The resolution of the Poisson equation is usually one of the most computationally intensive steps for incompressible fluid solvers. Lately, DeepLearning, and especially convolutional neural networks (CNNs), has been introduced to solve this equation, leading to significant inference time reduction at the cost of a lack of guarantee on the accuracy of the solution.This drawback might lead to inaccuracies, potentially unstable simulations and prevent performing fair assessments of the CNN speedup for different network architectures. To circumvent this issue, a hybrid strategy is developed, which couples a CNN with a traditional iterative solver to ensure a user-defined accuracy level. The CNN hybrid method is tested on two flow cases: (a) the flow around a 2D cylinder and (b) the variable-density plumes with and without obstacles (both 2D and 3D), demonstrating remarkable generalization capabilities, ensuring both the accuracy and stability of the simulations. The error distribution of the predictions using several network architectures is further investigated in the plume test case. The introduced hybrid strategy allows a systematic evaluation of the CNN performance at the same accuracy level for various network architectures. In particular, the importance of incorporating multiple scales in the network architecture is demonstrated, since improving both the accuracy and the inference performance compared with feedforward CNN architectures. Thus, in addition to the pure networks’ performance evaluation, this study has also led to numerous guidelines and results on how to build neural networks and computational strategies to predict unsteady flows with both accuracy and stability requirements.

The rapid development of new infrastructure programmes requires an accelerated deployment of new materials in new environments. Materials 4.0 is crucial to achieve these goals. The application of digital to the field of materials has been at the forefront of research for many years, but there does not exist a unified means to describe a framework for this area creating pockets of development. This is confounded by the broader expectations of a digital twin (DT) as the possible answer to all these problems. The issue being that there is no accepted definition of a component DT, and what information it should contain and how it can be implemented across the product lifecycle exist. Within this position paper, a clear distinction is made between the “manufacturing DT” and the “component DT”; the former being the starting boundary conditions of the latter. In order to achieve this, we also discuss the introduction of a digital thread as a key concept in passing data through manufacturing and into service. The stages of how to define a framework around the development of DTs from a materials perspective is given, which acknowledges the difference between creating new understanding within academia and the application of this knowledge on a per-component basis in industry. A number of challenges are identified to the broad application of a component DT; all lead to uncertainty in properties and locations, resolving these requires judgments to be made in the provision of safety-dependent materials property data.

The superposition of data sets with internal parametric self-similarity is a longstanding and widespread technique for the analysis of many types of experimental data across the physical sciences. Typically, this superposition is performed manually, or recently through the application of one of a few automated algorithms. However, these methods are often heuristic in nature, are prone to user bias via manual data shifting or parameterization, and lack a native framework for handling uncertainty in both the data and the resulting model of the superposed data. In this work, we develop a data-driven, nonparametric method for superposing experimental data with arbitrary coordinate transformations, which employs Gaussian process regression to learn statistical models that describe the data, and then uses maximum a posteriori estimation to optimally superpose the data sets. This statistical framework is robust to experimental noise and automatically produces uncertainty estimates for the learned coordinate transformations. Moreover, it is distinguished from black-box machine learning in its interpretability—specifically, it produces a model that may itself be interrogated to gain insight into the system under study. We demonstrate these salient features of our method through its application to four representative data sets characterizing the mechanics of soft materials. In every case, our method replicates results obtained using other approaches, but with reduced bias and the addition of uncertainty estimates. This method enables a standardized, statistical treatment of self-similar data across many fields, producing interpretable data-driven models that may inform applications such as materials classification, design, and discovery.

Multicomponent polymer systems are of interest in organic photovoltaic and drug delivery applications, among others where diverse morphologies influence performance. An improved understanding of morphology classification, driven by composition-informed prediction tools, will aid polymer engineering practice. We use a modified Cahn–Hilliard model to simulate polymer precipitation. Such physics-based models require high-performance computations that prevent rapid prototyping and iteration in engineering settings. To reduce the required computational costs, we apply machine learning (ML) techniques for clustering and consequent prediction of the simulated polymer-blend images in conjunction with simulations. Integrating ML and simulations in such a manner reduces the number of simulations needed to map out the morphology of polymer blends as a function of input parameters and also generates a data set which can be used by others to this end. We explore dimensionality reduction, via principal component analysis and autoencoder techniques, and analyze the resulting morphology clusters. Supervised ML using Gaussian process classification was subsequently used to predict morphology clusters according to species molar fraction and interaction parameter inputs. Manual pattern clustering yielded the best results, but ML techniques were able to predict the morphology of polymer blends with ≥90% accuracy.

This article recasts the traditional challenge of calibrating a material constitutive model into a hierarchical probabilistic framework. We consider a Bayesian framework where material parameters are assigned distributions, which are then updated given experimental data. Importantly, in true engineering setting, we are not interested in inferring the parameters for a single experiment, but rather inferring the model parameters over the population of possible experimental samples. In doing so, we seek to also capture the inherent variability of the material from coupon-to-coupon, as well as uncertainties around the repeatability of the test. In this article, we address this problem using a hierarchical Bayesian model. However, a vanilla computational approach is prohibitively expensive. Our strategy marginalizes over each individual experiment, decreasing the dimension of our inference problem to only the hyperparameter—those parameter describing the population statistics of the material model only. Importantly, this marginalization step, requires us to derive an approximate likelihood, for which, we exploit an emulator (built offline prior to sampling) and Bayesian quadrature, allowing us to capture the uncertainty in this numerical approximation. Importantly, our approach renders hierarchical Bayesian calibration of material models computational feasible. The approach is tested in two different examples. The first is a compression test of simple spring model using synthetic data; the second, a more complex example using real experiment data to fit a stochastic elastoplastic model for 3D-printed steel.

In order to clarify and visualize the real state of the structural performances of ships in operation and establish a more optimal, data-driven framework for ship design, construction and operation, an industry-academia joint R&D project on the digital twin for ship structures (DTSS) was conducted in Japan. This paper presents the major achievements of the project. The DTSS aims to grasp the stress responses over the whole ship structure in waves by data assimilation that merges hull monitoring and numerical simulation. Three data assimilation methods, namely, the wave spectrum method, Kalman filter method, and inverse finite element method were used, and their effectiveness was examined through model and full-scale ship measurements. Methods for predicting short-term extreme responses and long-term cumulative fatigue damage were developed for navigation and maintenance support using statistical approaches. In comparison with conventional approaches, response predictions were significantly improved by DTSS using real response data in encountered waves. Utilization scenarios for DTSS in the maritime industry were presented from the viewpoints of navigation support, maintenance support, rule improvement, and product value improvement, together with future research needs for implementation in the maritime industry.

Traffic congestion across the world has reached chronic levels. Despite many technological disruptions, one of the most fundamental and widely used functions within traffic modeling, the volume–delay function has seen little in the way of change since it was developed in the 1960s. Traditionally macroscopic methods have been employed to relate traffic volume to vehicular journey time. The general nature of these functions enables their ease of use and gives widespread applicability. However, they lack the ability to consider individual road characteristics (i.e., geometry, presence of traffic furniture, road quality, and surrounding environment). This research investigates the feasibility to reconstruct the model using two different data sources, namely the traffic speed from Google Maps’ Directions Application Programming Interface (API) and traffic volume data from automated traffic counters (ATC). Google’s traffic speed data are crowd-sourced from the smartphone Global Positioning System (GPS) of road users, able to reflect real-time, context-specific traffic condition of a road. On the other hand, the ATCs enable the harvesting of the vehicle volume data over equally fine temporal resolutions (hourly or less). By combining them for different road types in London, new context-specific volume–delay functions can be generated. This method shows promise in selected locations with the generation of robust functions. In other locations, it highlights the need to better understand other influencing factors, such as the presence of on-road parking or weather events.

In chemical process engineering, surrogate models of complex systems are often necessary for tasks of domain exploration, sensitivity analysis of the design parameters, and optimization. A suite of computational fluid dynamics (CFD) simulations geared toward chemical process equipment modeling has been developed and validated with experimental results from the literature. Various regression-based active learning strategies are explored with these CFD simulators in-the-loop under the constraints of a limited function evaluation budget. Specifically, five different sampling strategies and five regression techniques are compared, considering a set of four test cases of industrial significance and varying complexity. Gaussian process regression was observed to have a consistently good performance for these applications. The present quantitative study outlines the pros and cons of the different available techniques and highlights the best practices for their adoption. The test cases and tools are available with an open-source license to ensure reproducibility and engage the wider research community in contributing to both the CFD models and developing and benchmarking new improved algorithms tailored to this field.

Mathematical models are essential to analyze and understand the dynamics of complex systems. Recently, data-driven methodologies have gotten a lot of attention which is leveraged by advancements in sensor technology. However, the quality of obtained data plays a vital role in learning a good and reliable model. Therefore, in this paper, we propose an efficient heuristic methodology to collect data both in the frequency domain and the time domain, aiming at having more information gained from limited experimental data than equidistant points. In the frequency domain, the interpolation points are restricted to the imaginary axis as the transfer function can be estimated easily on the imaginary axis. The efficiency of the proposed methodology is illustrated by means of several examples, and its robustness in the presence of noisy data is shown.

We introduce an approach for damage detection in gearboxes based on the analysis of sensor data with the multi-resolution dynamic mode decomposition (mrDMD). The application focus is the condition monitoring of wind turbine gearboxes under varying load conditions, in particular irregular and stochastic wind fluctuations. We analyze data stemming from a simulated vibration response of a simple nonlinear gearbox model in a healthy and damaged scenario and under different wind conditions. With mrDMD applied on time-delay snapshots of the sensor data, we can extract components in these vibration signals that highlight features related to damage and enable its identification. A comparison with Fourier analysis, time synchronous averaging, and empirical mode decomposition shows the advantages of the proposed mrDMD-based data analysis approach for damage detection.