Data-driven Model Research Articles

The winding number of coronal flux ropes. I. Data-driven time-dependent magnetofrictional modelling

Magnetic flux ropes are key structures in solar and solar-terrestrial research. Their magnetic twist is an important quantity for understanding their eruptivity, their evolution in interplanetary space, and their consequences for planetary space environments. The magnetic twist is expressed in terms of a winding number that describes how many times a field line winds about the axis of the flux rope (FR). Due to the complexity of calculating the winding number, current methods rely largely on its approximation. We use a data-driven simulated FR to investigate the winding number $T_g$ in comparison to the commonly used twist proxy $T_w$, which describes a winding of two infinitesimally close field lines. We also estimate the magnetic flux enclosed in the resultant FR(s). We use the magnetic field analysis tools (MAFIAT) software to compute $T_g$ and $T_w$ for data-driven time-dependent magnetofrictional modelling of AR12473. We find that the FR boundaries can significantly differ depending on whether they are defined using the twist approximation $T_w$ or the winding number $T_g$. This also significantly affects the FR structure and the estimates of the enclosed magnetic flux. For the event analysed here, $T_g$ also reveals that the twisted flux system consists of two separate intertwined FRs. The results of this study suggest that the computation of the winding number ($T_g$) is important for investigations of solar FRs.

Computer vision-based real-time monitoring for swivel construction of bridges: from laboratory study to a pilot application

PurposeThe swivel construction method is a specially designed process used to build bridges that cross rivers, valleys, railroads and other obstacles. To carry out this construction method safely, real-time monitoring of the bridge rotation process is required to ensure a smooth swivel operation without collisions. However, the traditional means of monitoring using Electronic Total Station tools cannot realize real-time monitoring, and monitoring using motion sensors or GPS is cumbersome to use.Design/methodology/approachThis study proposes a monitoring method based on a series of computer vision (CV) technologies, which can monitor the rotation angle, velocity and inclination angle of the swivel construction in real-time. First, three proposed CV algorithms was developed in a laboratory environment. The experimental tests were carried out on a bridge scale model to select the outperformed algorithms for rotation, velocity and inclination monitor, respectively, as the final monitoring method in proposed method. Then, the selected method was implemented to monitor an actual bridge during its swivel construction to verify the applicability.FindingsIn the laboratory study, the monitoring data measured with the selected monitoring algorithms was compared with those measured by an Electronic Total Station and the errors in terms of rotation angle, velocity and inclination angle, were 0.040%, 0.040%, and −0.454%, respectively, thus validating the accuracy of the proposed method. In the pilot actual application, the method was shown to be feasible in a real construction application.Originality/valueIn a well-controlled laboratory the optimal algorithms for bridge swivel construction are identified and in an actual project the proposed method is verified. The proposed CV method is complementary to the use of Electronic Total Station tools, motion sensors, and GPS for safety monitoring of swivel construction of bridges. It also contributes to being a possible approach without data-driven model training. Its principal advantages are that it both provides real-time monitoring and is easy to deploy in real construction applications.

Capturing Local Temperature Evolution during Additive Manufacturing through Fourier Neural Operators

Abstract High-fidelity, data-driven models that can quickly simulate thermal behavior during additive manufacturing (AM) are crucial for improving the performance of AM technologies in multiple areas, such as part design, process planning, monitoring, and control. However, the complexities of part geometries make it challenging for current models to maintain high accuracy across a wide range of geometries. Additionally, many models report a low mean square error (MSE) across the entire domain of a part. However, in each time step, most areas of the domain do not experience significant changes in temperature, except for the regions near recent depositions. Therefore, the MSE-based fidelity measurement of the models may be overestimated. This paper presents a data-driven model that uses the Fourier Neural Operator to capture the local temperature evolution during the AM process. Beside MSE, the model is also evaluated using the R2 metric, which places great weight on the regions where the temperature changes significantly than MSE. The model was trained and tested on numerical simulations based on the Discontinuous Galerkin Finite Element Method for the Direct Energy Deposition AM process. The results shows that the model maintains 0.983-0.999 R2 over geometries not included in the training data, which is higher than Convolutional Neural Networks and Graph Convolutional Neural Networks we implemented, the two widely used architectures in data-driven predictive modeling.

Hybrid AI-Analytical Modeling of Droplet Dynamics on Inclined Heterogeneous Surfaces

This work presents a novel approach for the study of the movement of droplets on inclined surfaces under the influence of gravity and chemical heterogeneities. The developed numerical methodology uses data-driven modeling to extend the applicability limits of an analytically derived reduced-order model for the contact line velocity. More specifically, while the reduced-order model is able to capture the effects of the chemical heterogeneities to a satisfactory degree, it does not account for gravity. To alleviate this shortcoming, datasets generated from direct numerical simulations are used to train a data-driven model for the contact line velocity, which is based on the Fourier neural operator and corrects the reduced-order model predictions to match the reference solutions. This hybrid surrogate model, which comprises of both analytical and data-driven components, is then integrated in time to simulate the droplet movement, offering a speedup of five orders of magnitude compared to direct numerical simulations. The performance of this hybrid model is quantified and assessed in different wetting scenarios, by considering various inclination angles and values for the Bond number, demonstrating the accuracy of the predictions as long as the adopted parameters lie within the ranges considered in the training dataset.

The Methodology for Evaluating the Operating State of SF6 HVCBs Based on IDDA

To enhance the precision of evaluating the operational status of SF6 high-voltage circuit breakers (HVCBs) and devise judicious maintenance strategies, this study introduces an operational state assessment method for SF6 HVCBs grounded in the integrated data-driven analysis (IDDA) model. The relative degradation weight (RDW) is introduced as a metric for quantifying the relative significance of distinct indicators concerning the operational condition of SF6 HVCBs. A data-driven model, founded on critical factor stability (CFS), is formulated to convert environmental indicators into quantitative computations. Furthermore, an optimized fuzzy inference (OFI) system is devised to streamline the system architecture and enhance the processing speed of continuous indicators. Ultimately, the efficacy of the proposed model is substantiated through validation, and results from instance analyses underscore that the presented approach not only attains heightened accuracy in assessment compared to extant analytical methodologies but also furnishes a dependable foundation for prioritizing maintenance sequences across diverse components.

Streamflow prediction using support vector regression machine learning model for Tehri Dam

Accurate and reliable streamflow prediction is critical for optimising water resource management, reservoir flood operations, watershed management, and urban water management. Many researchers have published on streamflow prediction using techniques like Rainfall-Runoff modelling, Time series Models, Data-driven models, Artificial intelligence, etc. Still, there needs to be generalised method practise in the real world. The resolution of this issue lies in selecting different methods for a particular study area. This paper uses the Support vector regression machine learning model to predict the streamflow for the Tehri Dam, Uttarakhand, India, at the Daily and Ten Daily time steps. Two cases are considered in predicting daily and ten daily time steps. The first case includes four input variables: Discharge, Rainfall, Temperature, and Snow cover area. The second case comprises only three input variables: Rainfall, Temperature, and Snow cover area. Radial Kernel is used to overcome the space complexity in the datasets. The K-fold cross-validation is suitable for prediction as it averages the prediction error rate after evaluating the SVR model’s performance on various subsets of the training data. The streamflow data for daily and ten daily time steps have been collected from 2006 to 2020. The calibration period is from 2006 to 2016, and the validation period is from 2017 to 2020. Nash Sutcliffe Efficiency (NSE) and Coefficient of determination (R2) are used as the accuracy indicator in this manuscript. The lag has been observed in the daily prediction time series when three input variables are considered. For other scenarios, the respective model shows excellent results at both the temporal scale and the parametres, which play a vital role in prediction. The study also enhances the effect on the potential use of input parametres in the machine learning model.

Diffusion-Based Generative Network for de Novo Synthetic Promoter Design.

Computer-aided promoter design is a major development trend in synthetic promoter engineering. Various deep learning models have been used to evaluate or screen synthetic promoters, but there have been few works on de novo promoter design. To explore the potential ability of generative models in promoter design, we established a diffusion-based generative model for promoter design in Escherichia coli. The model was completely driven by sequence data and could study the essential characteristics of natural promoters, thus generating synthetic promoters similar to natural promoters in structure and component. We also improved the calculation method of FID indicator, using a convolution layer to extract the feature matrix of the promoter sequence instead. As a result, we got an FID equal to 1.37, which meant synthetic promoters have a distribution similar to that of natural ones. Our work provides a fresh approach to de novo promoter design, indicating that a completely data-driven generative model is feasible for promoter design.

Data-Driven Inline Leak Detection for Pipelines Using Flow-Induced Acoustics Analysis

Fluid and water distribution networks are essential to the modern world. However, these systems are prone to leaks, which can lead to significant water loss, damage to infrastructure, and environmental pollution. The proposed solution makes use of Acoustic Emission sensors placed in discrete locations in the pipeline which measures the sound in the pipeline caused by the flow of fluids. Computation models are used to deduce the location from the input provided by the sensors. In case of leak, the leak is localized through cross correlation and TDOA methods. This solution is particularly developed for water distribution pipelines. Keyword: Acoustic data analysis, Data-driven models, Cross-Correlation, Time Difference of Arrival (TDOA)

A framework for the practical development of condition monitoring systems with application to the roller compactor

Implementing a condition-based maintenance strategy requires an effective condition monitoring (CM) system that can be complicated to develop and even harder to maintain. In this paper, we review the main complexities of developing condition monitoring systems and introduce a four-stage framework that can address some of these difficulties. The framework achieves this by first using process knowledge to create a representation of the process condition. This representation can be broken down into simpler modules, allowing existing monitoring systems to be mapped to their corresponding module. Data-driven models such as machine learning models could then be used to train the modules that do not have existing CM systems. Even though data-driven models tend to not perform well with limited data, which is commonly the case in the early stages of pharmaceutical process development, application of this framework to a pharmaceutical roller compaction unit shows that the machine learning models trained on the simpler modules can make accurate predictions with novel fault detection capabilities. This is attributed to the incorporation of process knowledge to distill the process signals to the most important ones vis-à-vis the faults under consideration. Furthermore, the framework allows the holistic integration of these modular CM systems, which further extend their individual capabilities by maintaining process visibility during sensor maintenance.

Cross-tokamak Disruption Prediction based on domain adaptation

Abstract The high acquisition cost and the significant demand for disruptive discharges for data-driven disruption prediction models in future tokamaks pose an inherent contradiction in disruption prediction research. In this paper, we demonstrated a novel approach to predict disruption in a future tokamak using only a few discharges. The approach aims to predict disruption by finding a feature space that is universal to all tokamak. The first step is to use the existing understanding of physics to extract physics-guided features from the diagnostic signals of each tokamak, called physics-guided feature extraction (PGFE). The second step is to align a few data from the future tokamak (target domain) and a large amount of data from existing tokamak (source domain) based on a domain adaptation algorithm called CORrelation ALignment (CORAL). It is the first attempt at applying domain adaptation in the task of cross-tokamak disruption prediction. PGFE has been successfully applied in J-TEXT to predict disruption with excellent performance. PGFE can also reduce the data volume requirements due to extracting the less device-specific features, thereby establishing a solid foundation for cross-tokamak disruption prediction. We have further improved CORAL (supervised CORAL, S-CORAL) to enhance its appropriateness in feature alignment for the disruption prediction task. To simulate the existing and future tokamak case, we selected J-TEXT as the existing tokamak and EAST as the future tokamak, which has a large gap in the ranges of plasma parameters. The utilization of the S-CORAL improves the disruption prediction performance on future tokamak. Through interpretable analysis, we discovered that the learned knowledge of the disruption prediction model through this approach exhibits more similarities to the model trained on large data volumes of future tokamak. This approach provides a light, interpretable and few data-required way by aligning features to predict disruption using small data volume from the future tokamak.

Predicting permeation of compounds across the outer membrane of P. aeruginosa using molecular descriptors

The ability Gram-negative pathogens have at adapting and protecting themselves against antibiotics has increasingly become a public health threat. Data-driven models identifying molecular properties that correlate with outer membrane (OM) permeation and growth inhibition while avoiding efflux could guide the discovery of novel classes of antibiotics. Here we evaluate 174 molecular descriptors in 1260 antimicrobial compounds and study their correlations with antibacterial activity in Gram-negative Pseudomonas aeruginosa. The descriptors are derived from traditional approaches quantifying the compounds’ intrinsic physicochemical properties, together with, bacterium-specific from ensemble docking of compounds targeting specific MexB binding pockets, and all-atom molecular dynamics simulations in different subregions of the OM model. Using these descriptors and the measured inhibitory concentrations, we design a statistical protocol to identify predictors of OM permeation/inhibition. We find consistent rules across most of our data highlighting the role of the interaction between the compounds and the OM. An implementation of the rules uncovered in our study is shown, and it demonstrates the accuracy of our approach in a set of previously unseen compounds. Our analysis sheds new light on the key properties drug candidates need to effectively permeate/inhibit P. aeruginosa, and opens the gate to similar data-driven studies in other Gram-negative pathogens.

Data-driven modeling in digital twin for power system anomaly detection

Background: Power system anomaly detection is of great significance for realizing system situation awareness and early detection of system operating risks. In view of the complex operating conditions of the system, there are a large number of opaque links in the mechanism, and the anomaly detection approach based on physical mechanism modeling is prone to system errors due to assumptions, simplification, and transfer in the modeling process. This paper focuses on digital twin based data-driven approaches for power system anomaly detection to compensate for the limitation of physical methods in dynamical modeling. Methods: First of all, a digital twin framework for power system real-time analysis is constructed based on the concept of digital twin. Then, this paper conducts researches on the core of the designed framework, i.e., digital twin modeling. Considering the complexity of power system operating conditions, data-driven modeling is preferred and a random matrix and free probability theory based model for anomaly detection of system operating situation is constructed. Results: Simulation data with different spatiotemporal structure generated through a Monte Carlo experiment verified the sensitivity of the constructed model for data correlations. Meanwhile, the case on the system operating data generated through the IEEE 118-bus system validate the effectiveness of the proposed model for the system anomaly detection. Conclusions: The constructed data-driven model can accurately characterize the correlations among data elements, has good sensitivity to the variation of data spatial and temporal correlations, and can depict the data residuals better than the M-P law curve, which indicates the practicability and necessity of the constructed data-driven model for the digital twin modeling of power system anomaly detection.

Impacts of spatiotemporal resolutions of precipitation on flood event simulation based on multimodel structures – a case study over the Xiang River basin in China

Abstract. Accurate flood event simulation and prediction, enabled by effective models and reliable data, are critical for mitigating the potential risk of flood disaster. This study aims to investigate the impacts of spatiotemporal resolutions of precipitation on flood event simulation in a large-scale catchment of China. We use high-spatiotemporal-resolution Integrated Multi-satellite Retrievals for Global Precipitation Measurement (IMERG) products and a gauge-based product as precipitation forcing for hydrologic simulation. Three hydrological models (HBV, SWAT and DHSVM) and a data-driven model (long short-term memory (LSTM) network) are utilized for flood event simulation. Two calibration strategies are carried out, one of which targets matching of the flood events, with peak discharge exceeding 8600 m3 s−1 between January 2015 and December 2017, and the other one is the conventional strategy for matching the entire streamflow time series. The results indicate that the event-based calibration strategy improves the performance of flood event simulation compared with a conventional calibration strategy, except for DHSVM. Both hydrological models and LSTM yield better flood event simulation at a finer temporal resolution, especially in flood peak simulation. Furthermore, SWAT and DHSVM are less sensitive to the spatial resolutions of IMERG, while the performance of LSTM obtains improvement when degrading the spatial resolution of IMERG-L. Generally, LSTM outperforms the hydrological models in most flood events, which implies the usefulness of the deep learning algorithms for flood event simulation.

Impact of physical and attention mechanisms on U-Net for SST forecasting

Sea surface temperature (SST) prediction is a subject of great significance to the marine environment and human society. Changes in SST not only impact marine ecosystems and fishery resources but also trigger extreme weather events and disastrous consequences. Therefore, the precise prediction of SST is essential to avoiding these problems. Although numerous data-driven SST prediction models have emerged in recent years, these models are characterized by a lack of physical mechanisms related to sea temperature changes as well as insufficient generalization capabilities and interpretability. In our work, attempts were made to integrate physics-related convection phenomena into deep learning models, and traditional deep learning models were improved by incorporating time and space attention modules. The results of a series of experiments showed that the incorporation of physical mechanisms enhanced the performance of data-driven models. Furthermore, attention mechanisms were similarly helpful, of which temporal attention proved to be more important. The modules proposed in this work also improved the baseline model’s accuracy by 22%. In addition, seven-day SST predictions were carried out for the world’s five major fishing grounds. The results demonstrated that the application of transfer learning strategies yielded superior performance, further improving prediction accuracy by 1%–5%.

Developing a Data-Driven AI Model to Enhance Energy Efficiency in UK Residential Buildings

Residential buildings contribute 30% of the UK’s total final energy consumption. However, with less than one percent of its housing stock being replaced annually, retrofitting existing homes has significant importance in meeting energy-efficiency targets. Consequently, many physics-based and data-driven models and tools have been developed to analyse the effects of retrofit strategies from various points of view. This paper aims to develop a data-driven AI model that predicts buildings’ energy performance based on their features under various retrofit scenarios. In this context, four different machine learning models were developed based on the EPC (Energy Performance Certificate) dataset for residential buildings and standard assessment procedure (SAP) guidelines in the UK. Additionally, an interface was designed that enables users to analyse the effect of different retrofit strategies on a building’s energy performance using the developed AI models. The results of this study revealed the artificial neural network as the most accurate predictive model, with a coefficient of determination (R2) of 0.82 and a mean percentage error of 11.9 percent. However, some conceptual irregularities were observed across all the models when dealing with different retrofit scenarios. All summary, such tools can be further improved to offer a potential alternative or support to physics-based models, enhancing the efficiency of retrofitting processes in buildings.

Control-oriented modelling of wind direction variability

Abstract. Wind direction variability significantly affects the performance and lifetime of wind turbines and wind farms. Accurately modelling wind direction variability and understanding the effects of yaw misalignment are critical towards designing better wind turbine yaw and wind farm flow controllers. This review focuses on control-oriented modelling of wind direction variability, which is an approach that aims to capture the dynamics of wind direction variability for improving controller performance over a complete set of farm flow scenarios, performing iterative controller development and/or achieving real-time closed-loop model-based feedback control. The review covers various modelling techniques, including large eddy simulations (LESs), data-driven empirical models, and machine learning models, as well as different approaches to data collection and pre-processing. The review also discusses the different challenges in modelling wind direction variability, such as data quality and availability, model uncertainty, and the trade-off between accuracy and computational cost. The review concludes with a discussion of the critical challenges which need to be overcome in control-oriented modelling of wind direction variability, including the use of both high- and low-fidelity models.

Self-supervised learning of shedding droplet dynamics during steam condensation

Knowledge of condensate shedding droplet dynamics provides important information for the characterization of two-phase heat and mass transfer phenomena. Detecting and segmenting the droplets during shedding requires considerable time and effort if performed manually. Here, we developed a self-supervised deep learning model for segmenting shedding droplets from a variety of dropwise and filmwise condensing surfaces. The model eliminates the need for image annotation by humans in the training step and, therefore, reduces labor significantly. The trained model achieved an average accuracy greater than 0.9 on a new unseen test dataset. After extracting the shedding droplet size and speed, we developed a data-driven model for shedding droplet dynamics based on condensation heat flux and surface properties such as wettability and tube diameter. Our results demonstrate that condensate droplet departure size is both heat flux and tube size dependent and follows different trends based on the condensation mode. The results of this work provide an annotation-free methodology for falling droplet segmentation as well as a statistical understanding of droplet dynamics during condensation.

Manufacturing-induced stochastic constitutive behaviors of additive manufactured specimens: testing, data-driven modeling, and optimization

Purpose Fused deposition modeling (FDM) is an extensively used additive manufacturing method with the capacity to build complex functional components. Due to the machinery and environmental factors during manufacturing, the FDM parts inevitably demonstrated uncertainty in properties and performance. This study aims to identify the stochastic constitutive behaviors of FDM-fabricated polylactic acid (PLA) tensile specimens induced by the manufacturing process. Design/methodology/approach By conducting the tensile test, the effects of the printing machine selection and three major manufacturing parameters (i.e., printing speed S, nozzle temperature T and layer thickness t) on the stochastic constitutive behaviors were investigated. The influence of the loading rate was also explained. In addition, the data-driven models were established to quantify and optimize the uncertain mechanical behaviors of FDM-based tensile specimens under various printing parameters. Findings As indicated by the results, the uncertain behaviors of the stiffness and strength of the PLA tensile specimens were dominated by the printing speed and nozzle temperature, respectively. The manufacturing-induced stochastic constitutive behaviors could be accurately captured by the developed data-driven model with the R2 over 0.98 on the testing dataset. The optimal parameters obtained from the data-driven framework were T = 231.3595 °C, S = 40.3179 mm/min and t = 0.2343 mm, which were in good agreement with the experiments. Practical implications The developed data-driven models can also be integrated into the design and characterization of parts fabricated by extrusion and other additive manufacturing technologies. Originality/value Stochastic behaviors of additively manufactured products were revealed by considering extensive manufacturing factors. The data-driven models were proposed to facilitate the description and optimization of the FDM products and control their quality.

Patient-Specific, Mechanistic Models of Tumor Growth Incorporating Artificial Intelligence and Big Data.

Despite the remarkable advances in cancer diagnosis, treatment, and management over the past decade, malignant tumors remain a major public health problem. Further progress in combating cancer may be enabled by personalizing the delivery of therapies according to the predicted response for each individual patient. The design of personalized therapies requires the integration of patient-specific information with an appropriate mathematical model of tumor response. A fundamental barrier to realizing this paradigm is the current lack of a rigorous yet practical mathematical theory of tumor initiation, development, invasion, and response to therapy. We begin this review with an overview of different approaches to modeling tumor growth and treatment, including mechanistic as well as data-driven models based on big data and artificial intelligence. We then present illustrative examples of mathematical models manifesting their utility and discuss the limitations of stand-alone mechanistic and data-driven models. We then discuss the potential of mechanistic models for not only predicting but also optimizing response to therapy on a patient-specific basis. We describe current efforts and future possibilities to integrate mechanistic and data-driven models. We conclude by proposing five fundamental challenges that must be addressed to fully realize personalized care for cancer patients driven by computational models.

Data-driven prediction of aerodynamic noise of transonic buffeting over an airfoil

Accurately predicting buffet frequency and aerodynamic noise level is crucial in transonic buffet noise reduction studies. In this study, the Random Forest (RF) algorithm is employed to predict the Power Spectral Density (PSD) and Overall Sound Pressure Level (OASPL) distribution over the supercritical airfoil RAE2822. The study indicates that the RF algorithm exhibits greater advantages over the Multi-Layer Perceptron (MLP). This algorithm does not suffer from the problems of high-frequency divergence or inflection distortion, maintaining a reduction in prediction errors for OASPL and PSD by approximately three to four orders of magnitude. Additionally, this paper proposes a priori criterion for evaluating the accuracy of PSD prediction. When there is a strong correlation between PSD at adjacent points, a data-driven modeling approach can achieve higher prediction accuracy. However, when the root mean square error of the cross-correlation functions, auto-correlation functions, and the PSD between adjacent points are both high, the generalization ability of pure data-driven modeling is insufficient, necessitating the additional monitoring points to ensure prediction accuracy.