Data-driven Model Research Articles

The SAGA Survey. V. Modeling Satellite Systems around Milky Way–Mass Galaxies with Updated UniverseMachine

Abstract Environment plays a critical role in shaping the assembly of low-mass galaxies. Here, we use the UniverseMachine (UM) galaxy–halo connection framework and Data Release 3 of the Satellites Around Galactic Analogs (SAGA) Survey to place dwarf galaxy star formation and quenching into a cosmological context. UM is a data-driven forward model that flexibly parameterizes galaxy star formation rates (SFRs) using only halo mass and assembly history. We add a new quenching model to UM, tailored for galaxies with m ⋆ ≲ 109 M ⊙, and constrain the model down to m ⋆ ≳ 107 M ⊙ using new SAGA observations of 101 satellite systems around Milky Way (MW)–mass hosts and a sample of isolated field galaxies in a similar mass range from the Sloan Digital Sky Survey. The new best-fit model, “UM-SAGA,” reproduces the satellite stellar mass functions, average SFRs, and quenched fractions in SAGA satellites while keeping isolated dwarfs mostly star-forming. The enhanced quenching in satellites relative to isolated field galaxies leads the model to maximally rely on halo assembly to explain the observed environmental quenching. Extrapolating the model down to m ⋆ ∼ 106.5 M ⊙ yields a quenched fraction of ≳30% for isolated field galaxies and ≳80% for satellites of MW-mass hosts at this stellar mass. Spectroscopic surveys can soon test this specific prediction to reveal the relative importance of internal feedback, cessation of mass and gas accretion, satellite-specific gas processes, and reionization for the evolution of faint low-mass galaxies.

Development of a Generalizable Data-Driven Turbulence Model: Conditioned Field Inversion and Symbolic Regression

This paper addresses the issue of predicting separated flows with Reynolds-averaged Navier–Stokes (RANS) turbulence models, which are essential for many engineering tasks. Traditional RANS models usually struggle with this task, so recent efforts have focused on data-driven methods such as field inversion and machine learning (FIML) to correct this issue by adjusting the baseline equations. However, these FIML methods often reduce accuracy in attached boundary layers. To address this issue, we developed a “conditioned field inversion” technique. This method adjusts the corrective factor β (used by FIML) in the shear-stress transport (SST) model. It multiplies β with a shield function fd that is off in the boundary layer and on elsewhere. This maintains the accuracy of the baseline model for the attached flows. We applied both conditioned and classic field inversion to the NASA hump and a curved backward-facing step, creating two datasets. These datasets were used to train two models: SR-CND (symbolic regression-conditioned, from our new method) and SR-CLS (symbolic regression-classic, from the traditional method). The SR-CND model matches the SR-CLS model in predicting separated flows in various scenarios, such as periodic hills, the NLR7301 airfoil, the 3D SAE (Society of Automotive Engineers) car model, and the 3D Ahmed body, and outperforms the baseline SST model in the cases presented in the paper. Importantly, the SR-CND model maintains accuracy in the attached boundary layers, whereas the SR-CLS model does not. Therefore, the proposed method improves separated flow predictions while maintaining the accuracy of the original model for attached flows, offering a better way to create data-driven turbulence models.

A Multiscale Framework for Capturing Oscillation Dynamics of Autonomous Vehicles in Data-Driven Car-Following Models

A Multiscale Framework for Capturing Oscillation Dynamics of Autonomous Vehicles in Data-Driven Car-Following Models

Study on runoff forecasting and error correction driven by atmosphere–ocean-land dataset

Accurate runoff forecasting results can not only provide an important basis for flood control scheduling, but also provide scientific support for water resources optimization, which promotes the maximization of the overall benefits of the basin. To further explore the inherent mechanisms of the atmosphere–ocean-land factors driving runoff changes, this study proposes the factors dimension reduction and interpretation framework based on Pearson, eXtreme Gradient Boosting and SHapley Additive exPlanations (P-XGBoost-SHAP). Base on this, the Gaussian Process Regression (GPR), Long Short-Term Memory neural network (LSTM) and Support Vector Machine (SVM) models are used to construct the atmospheric-ocean-land data-driven runoff prediction model. Meanwhile, for the runoff prediction residuals, this paper proposes an error multi-step correction framework based on ensemble empirical mode decomposition and autoregressive model (EEMD-AR). The case study of Lianghekou hydrological station shows that the factor dimension reduction and interpretation framework can greatly reduce the input dimension of the model, and explain the factors globally and locally by using SHAP Value. Compared with the traditional Random Forest (RF) dimension reduction method, it shows higher prediction accuracy. The Nash-Sutcliffe efficiency coefficient (NSE) can be increased to about 0.93, which is 4.91 % and 1.97 % higher than the series–parallel coupling (AR-Parallel) and empirical mode decomposition-autoregressive (EMD-AR) correction methods, respectively. The accuracy of the runoff forecasting prediction is improved while reducing the input dimension of the model.

DeepPipe: A Multi-Stage Knowledge-Enhanced Physics-Informed Neural Network for Hydraulic Transient Simulation of Multi-Product Pipeline

In the chemical pipelining industry, owing to the high-pressure transportation process, an accurate hydraulic transient simulation tool plays a central role in preventing the slack line flow and overpressure from causing pipeline operation treacherous. Nevertheless, the current model-driven method often faces challenges in balancing computational efficiency with accuracy, and the existing data-driven models struggle to produce explainable results from the physics perspectives since insufficient theoretical principles are incorporated into the model training. Additionally, the existing physics-informed learning architecture fails to achieve a gradient-balanced training, resulting from the significant magnitude difference in outputs and multiple loss terms. Consequently, a Multi-Stage Knowledge-Enhanced Physics-Informed Neural Network (MS-KE-PINN) is proposed for the hydraulic transient simulation of multi-product pipelines. To enforce the neural network producing simulation results with high consistency to physical laws, the governing equations, boundary, and initial condition are incorporated into the training process for an efficient mesh-free simulation. Then, considering that the significant magnitude difference between outputs can easily lead to deficient performance in the gradient descent, the magnitude conversion on the outputs and the equivalent conversion of the governing equations are implemented to enhance the training effect of the neural network. Subsequently, to tackle the imbalanced gradient of multiple loss terms with fixed weights, a multi-stage hierarchical training strategy is designed to improve the approximation capacity of the neural network. Numerical simulation cases demonstrate a better approximation function of the proposed model than the state-of-art models, while the mean absolute percentage errors yielded by MS-KE-PINN are reduced by 77.4 %, 88.7 %, and 87.8 % in three simulation operation conditions for pressure prediction. Furthermore, experimental investigations from a real-world multi-product pipeline suggest that the proposed model can still draw accurate simulation results even under complex and dynamic hydraulic transient scenarios in practice, with root mean squared errors reduced by 94.8 % and 80 % than that of the physics-informed neural network. To this end, the proposed model can conduct accurate and effective hydraulic transient analysis, thus ensuring the safe operation of the pipeline.

Uncovering the features of industrial odors-derived environmental complaints and proactive countermeasures by using machine-learning

Industrial odor-derived environmental complaints pose an emerging and far-reaching challenge in cities worldwide with intensive industries. Developing effective odor complaint management strategies is essential for mitigating the public impact of industrial odors. Based on a typical case of persistent tire manufacturing odors affecting local communities, we proposed an environmental complaint risks (ECR) prediction model using machine-learning (ML) approaches, which combined complaints with temporal-resolution manufacturing-meteorology-environment data. Through intensive match-making between ML algorithms and multi-source parameters, Random Forest models can achieve a reliable ECR-prediction model performance with an average ROC-AUC of 0.79at a monthly timescale, indicating the effectiveness of ML-based ECR prediction models. The interpretable ML model quantitively depicted the underlying mechanisms of ECR prediction, driven by process emission behaviors, local wind direction, and historical high-risk period. Furthermore, to mitigate predictable ECR within a future period, we designed a model framework that integrated ECR prediction models with an adaptive optimization genetic algorithm. This enabled the proactive management by precisely and dynamically allocating limited resources of emission regulatory to high-ECR periods in advance. The strategy was proven effective, achieving a significant 24.7% average reduction in the overall ECR forecast during a period with intensive complaints. Overall, this study proposed a data-driven model framework that efficiently helps the multi-stakeholders shift from passive response to proactive ECR management, thereby enhancing the environmental and social sustainability.

Data-driven modeling of low-frequency hydrodynamic loads

Abstract This paper presents a data-driven approach for estimating the linear and quadratic transfer functions relating incoming waves to hydrodynamic loads on floaters. The procedure relies on constructing a nonlinear auto-regressive (NARX) surrogate model for forecasting the hydrodynamic loads and a harmonic probing algorithm for extracting the transfer functions of the system. The implemented harmonic probing method is of numerical nature and avoids the use of computationally expensive symbolic coding tools. The method was verified against synthetic data generated from a potential-flow-based numerical model of the INO WINDMOOR 12 MW floater. The main advantage of NARX models is the fexibility of their structure, meaning that they do not presuppose that the nonlinearity of the hydrodynamic load is quadratic, as it is commonly done.

Space-time modeling of cascading hazards: Chaining wildfires, rainfall and landslide events through machine learning

The current study sets out to explore yearly landslide susceptibility dynamics on slopes regularly affected by fires. To do so, two yearly inventories have been generated, one for the landslides and one for the wildfires, for an area of approximately 2 km2 and for a period of 24 years. It is important to stress that space–time data-driven models employed for susceptibility assessment are relatively new, and their application so far has mostly linked landslide occurrences to the precipitation trigger and the standard morphometric characteristics of the landscape at hand. Here we also consider an additional element of disturbance to the slope equilibrium, in the form of burnt areas tested from one up to three years priors to the reference landslide occurrence time. The relevance of the wildfire spatiotemporal signal is tested as part of a multi-variate modeling procedure. The results highlight at least a 10 % performance increase when these wildfire-related predictors are featured (from an average AUC of 0.75 to 0.85 in a random forest modeling framework). The associated yearly variations in the landslide occurrence probability are translated into individual maps, stressing the extent to which the standard and static definition of susceptibility does not hold especially in the context of multiple hazards and in an urban setting such as the Camaldoli hills (Naples, Italy) we chose for our test site.

Subsampling approach and data-driven models to predict silicate glass melt viscosity

This study focused on developing a predictive tool for glass melt viscosity calculation between 900 °C and 1500 °C. Database containing about 16000 silicate glass compositions was built from both data available in the literature and from a proprietary dataset. The approach combines statistical techniques of design of experiments, machine learning, and subsampling strategies for model training. Prediction accuracy was found to be very promising for the different types of silicate glasses studied. Relative error of viscosity prediction at 1200 °C was around 20 % for simple SBN compositions, and less than one order of magnitude for more complex compositions.

Simulation and optimisation of vacuum (pressure) swing adsorption with simultaneous consideration of real vacuum pump data and bed fluidisation

Pressure swing adsorption (PSA) is a promising technology for gas separation and purification, receiving considerable attention in the past decades. Existing studies on simulating PSA involving a vacuum pump, the vacuum (pressure) swing adsorption [V(P)SA] process, typically estimate vacuum pump energy consumption using an approximate constant energy efficiency or an empirical energy efficiency correlation, leading to inaccurate representation of realistic vacuum pump performance. In this work, we propose an enhanced computational approach for simulation and optimisation of V(P)SA processes through simultaneous incorporation of realistic vacuum pump performance prediction models and adsorption bed fluidisation limits. The vacuum pump prediction models are developed using data-driven modelling techniques with realistic vacuum pump performance data. The enhanced computational approach more accurately accounts for the V(P)SA performance, without relying on an estimated vacuum pump efficiency and assumed pressure/flow rate boundary conditions at the vacuum pump end of the adsorption bed. Furthermore, this approach prevents the bed fluidisation. The computational results show that the developed prediction models accurately represent the actual performance curves of the vacuum pump. It is also demonstrated that incorporating the vacuum pump prediction models and fluidisation constraints in PSA optimisation leads to significantly different optimal solutions compared to when these factors are not considered.

Data-driven models for predicting compressive strength of 3D-printed fiber-reinforced concrete using interpretable machine learning algorithms

3D printing technology is growing swiftly in the construction sector due to its numerous benefits, such as intricate designs, quicker construction, waste reduction, environmental friendliness, cost savings, and enhanced safety. Nevertheless, optimizing the concrete mix for 3D printing is a challenging task due to the numerous factors involved, requiring extensive experimentation. Therefore, this study used three machine learning techniques, including Gene Expression Programming (GEP), Multi-Expression Programming (MEP), and Decision Tree (DT), to forecast the compressive strength of 3D printed fiber-reinforced concrete (3DP-FRC). The dataset comprises 299 data points with sixteen variables gathered from experimental research studies. For training the model, 70 % of the dataset was used, while the remaining 30 % was reserved for model testing. Several statistical metrics were utilized to evaluate the accuracy and applicability of the models. In addition, SHapley Additive exPlanations (SHAP), partial dependence plots, and individual conditional expectations approach were employed for the interpretability of the models. The proposed GEP, MEP, and DT models indicated enhanced efficacy, exhibiting correlation coefficient (R) scores of 0.996, 0.987, and 0.990, with mean absolute errors (MAE) of 1.029, 4.832, and 2.513, respectively. Overall, the established GEP model demonstrated exceptional performance compared to MEP and DT, showcasing high prediction precision in assessing the strength of 3DP-FRC. Moreover, a simple empirical formulation has been devised using GEP to predict the compressive strength, offering a simplified and efficient approach for predicting 3DP-FRC strength. The SHAP approach identified water, silica fume, fiber diameter, curing age, and loading directions as leading controlling parameters in predicting strength of 3DP-FRC. In summary, the proposed models can potentially minimize both the computational workload and the need for experimental trials in formulating the mixed design of 3D-printed concrete.

Urban Building Energy Modeling: A Comparative Study of Process-Driven and Data-Driven Models

Urban Building Energy Modeling: A Comparative Study of Process-Driven and Data-Driven Models

Power Quality State Estimation for Distribution Grids Based on Physics-Aware Neural Networks—Harmonic State Estimation

In the transition from traditional electrical energy generation with mainly linear sources to increasing inverter-based distributed generation, electrical power systems’ power quality requires new monitoring methods. Integrating a high penetration of distributed generation, which is typically located in medium- or low-voltage grids, shifts the monitoring tasks from the transmission to distribution layers. Compared to high-voltage grids, distribution grids feature a higher level of complexity. Monitoring all relevant nodes is operationally infeasible and costly. State estimation methods provide knowledge about unmeasured locations by learning a physical system’s non-linear relationships. This article examines a new flexible, close-to-real-time concept of harmonic state estimation using synchronized measurements processed in a neural network. A physics-aware approach enhances a data-driven model, taking into account the structure of the electrical network. An OpenDSS simulation generates data for model training and validation. Different load profiles for both training and testing were utilized to increase the variance in the data. The results of the presented concept demonstrate high accuracy compared to other methods for harmonic orders 1 to 20.

A data-driven operations management model: Implementing MIS for strategic decision making in tech businesses

In today’s rapidly evolving technological landscape, data-driven decision-making is critical for business growth and competitiveness. This paper presents a comprehensive operations management model that integrates Management Information Systems (MIS) to facilitate strategic decision-making in technology firms. The proposed model emphasizes the transformation of raw data into actionable insights, enabling organizations to optimize their operations and respond effectively to market dynamics. By systematically incorporating MIS into operational frameworks, the model provides a structured approach to data collection, analysis, and dissemination. It identifies key performance indicators (KPIs) that are essential for monitoring business performance and aligning operational activities with strategic objectives. Through case studies of leading tech companies, the model illustrates how data-driven methodologies can lead to improved forecasting, resource allocation, and overall operational efficiency. The framework is designed to be adaptable, allowing technology firms to customize the MIS components based on their unique business needs and market conditions. Key features of the model include real-time data analytics, predictive modeling, and user-friendly dashboards that enhance visibility across departments. These components empower decision-makers with timely and relevant information, fostering a culture of informed decision-making at all organizational levels. Moreover, the model addresses common challenges faced by tech businesses, such as data silos and integration issues. By promoting a centralized data repository and standardized reporting mechanisms, it enhances collaboration among teams and ensures consistency in data interpretation. The implications of this model extend beyond operational efficiency, influencing strategic initiatives such as product development, market expansion, and customer engagement. In conclusion, this data-driven operations management model not only supports strategic decision-making in technology firms but also positions them for sustainable growth in a competitive marketplace. By leveraging MIS effectively, organizations can harness the power of data to drive innovation and enhance overall business performance.

Revolutionizing inventory management: A comprehensive automated data-driven model using power BI incorporating industry 4.0

This is a study on advanced inventory management approaches in the mechanical industry using data-driven models, accompanied by Power BI analytics. Drawing insights from broad collaboration with relevant industrial expertise at different organizational scales, our research work addresses the intricacies and challenges in maintaining optimal inventory levels concerning growing industries and operational risks. In this global mechanical firm, we analyzed historical data using descriptive and predictive analytics to understand the trends of the past and to forecast future inventory needs. Descriptive analytics provided base insights about the present status of inventory, while predictive analytics helped proactive management of stock levels classified into on-hand and critical level categories for greater efficiency in strategies related to planning at the coordination end with its suppliers and mitigating risks. Core to our methodology will be the implementation of an automated data-driven machine-learning approach that brings minimal intervention from humans to solve some of the most critical problems in inventory management. One major contribution this study adds is a major product: a Power BI dashboard that visualizes critical part numbers falling below threshold values and recognizes suppliers with historical shipment shortages. Armed with this intuitive dashboard, supplier-coordinating engineers are better placed to take proactive action on inventory shortages, greatly improving operational efficiency in supply chain resilience.

Mathematical approaches to fisheries quota management: ensuring sustainable practices in U.S. commercial fishing

Fisheries quota management plays a critical role in ensuring the sustainability of commercial fishing practices, particularly in the United States, where overfishing and resource depletion pose significant challenges to marine ecosystems. This review explores the mathematical approaches utilized to manage fisheries quotas effectively and sustainably. The paper examines key mathematical models, including population dynamics, bioeconomic models, game theory, and Catch-per-Unit-Effort (CPUE) models, highlighting their role in maintaining fish stock levels and regulating fishing activities. Through case studies from U.S. fisheries, such as the New England groundfish and Alaska pollock fisheries, the review assesses the effectiveness of these models in real-world applications. Additionally, it explores how advances in technology, such as satellite tracking, big data, and artificial intelligence, are enhancing the accuracy and adaptability of fisheries management models. The review concludes with insights into future directions for mathematical approaches in fisheries management, emphasizing the need for continuous innovation to support sustainable commercial fishing practices. By integrating advanced mathematical techniques and data-driven models, this paper aims to provide a pathway for policymakers and fisheries managers to achieve long-term sustainability in U.S. commercial fisheries.

Very short-term wind power forecasting considering static data: An improved transformer model

The randomness and fluctuations in wind power generation present significant challenges for grid and wind farm dispatching. Accurate very short-term wind power forecasting (WPF) is therefore essential for the efficient operation of modern power systems. Data-driven models, such as Transformers, have demonstrated their effectiveness in WPF due to their ability to efficiently capture global features in long sequences. However, limited research has examined the impact of incorporating static data into WPF, which may limit forecasting accuracy. This paper proposes a Temporal Fusion Transformer forecasting model to address this challenge. This approach employs static data as the input features for the model. The model includes feature selection through a variable selection network and employs a specialized temporal fusion decoder to learn effectively from these static features. The case results show that the results of the proposed model are more accurate than the state-of-the-art methods, reducing MAPE by at least 1.32%, RMSE by 0.0091, and improving R2 by 0.035 in case studies. Additionally, the model maintains a manageable computational burden, underscoring its practical applicability.

Pop-cosmos: Scaleable Inference of Galaxy Properties and Redshifts with a Data-driven Population Model

We present an efficient Bayesian method for estimating individual photometric redshifts and galaxy properties under a pretrained population model (pop-cosmos) that was calibrated using purely photometric data. This model specifies a prior distribution over 16 stellar population synthesis (SPS) parameters using a score-based diffusion model, and includes a data model with detailed treatment of nebular emission. We use a GPU-accelerated affine-invariant ensemble sampler to achieve fast posterior sampling under this model for 292,300 individual galaxies in the COSMOS2020 catalog, leveraging a neural network emulator (Speculator) to speed up the SPS calculations. We apply both the pop-cosmos population model and a baseline prior inspired by Prospector-α, and compare these results to published COSMOS2020 redshift estimates from the widely used EAZY and LePhare codes. For the ∼12,000 galaxies with spectroscopic redshifts, we find that pop-cosmos yields redshift estimates that have minimal bias (∼10−4), high accuracy (σ MAD = 7 × 10−3), and a low outlier rate (1.6%). We show that the pop-cosmos population model generalizes well to galaxies fainter than its r < 25 mag training set. The sample we have analyzed is ≳3× larger than has previously been possible via posterior sampling with a full SPS model, with average throughput of 15 GPU-sec per galaxy under the pop-cosmos prior, and 0.6 GPU-sec per galaxy under the Prospector prior. This paves the way for principled modeling of the huge catalogs expected from upcoming Stage IV galaxy surveys.

An ADMM approach for unit commitment with considering dynamic line rating

To enhance the transmission capabilities in power system scheduling, this paper develops a unit commitment model that incorporates dynamic transmission line capacities and proposes an efficient solving algorithm. A multi-scenario unit commitment model that integrates dynamic transmission line capacities is introduced, using quantile regression to construct a data-driven capacity increase model based on historical environmental data. The model is solved using Lagrangian relaxation and the Alternating Direction Method of Multipliers (ADMM) to decouple dynamic constraints, allowing the dual problem to be decomposed into sub-problems and solved iteratively. The proposed model and algorithm are validated using the IEEE-118 and IEEE-300 test cases, demonstrating their effectiveness in handling dynamic transmission line capacities and improving scheduling performance.The approach provides a robust and flexible solution for power system scheduling, enhancing reliability and economic efficiency.

Data-driven AI for the automated classification of the isothermal heat-treated thermal barrier coatings using pulsed infrared thermography

Abstract Thermal barrier coating (TBC) is a multilayer coating applied to metallic structures exposed to high temperatures, such as aero-engine parts and gas turbine blades. These thermally insulating materials are used to extend the component life by minimizing the high-temperature exposure of structural components. As a result of the high-temperature oxidation tests, a thermally grown oxide (TGO) layer formed at the interface between the ceramic topcoat layer and the bond coat layer. The primary cause of TBC performance decline and structure failure is the growth of the oxide layer. Hence, it is crucial to regularly monitor the formation of the oxide layer and the degradation of coatings. This can be achieved by developing coating-life prediction models, which are vital for quality control, health monitoring of TBCs, maintenance decision-making, and the prevention of catastrophic failures. Existing methods for life estimation, which are dependent on empirical methods and microstructural analysis, often fail due to various material compositions and service conditions. This study proposes an automated classification of iso-thermal heat-treated TBC samples using temperature data, which helps in predicting the TBC life and monitoring the TBC degradation. TBC-coated samples are isothermal heat-treated at 1000 °C, and the initial growth of TGO is monitored using a non-destructive thermal imaging technique. Identifying distinctive features in each class corresponding to layer thickness measurement and service hours is challenging and time-consuming. So, we integrate data-driven AI models and feature extraction techniques to interpret complex thermal patterns. The performance of deep learning models and classical machine learning models demonstrates that our method can achieve maximum classification accuracy and F1-score of 89.2% and 89.0%, respectively. This automated classification framework promises a powerful tool for predictive maintenance and remaining useful lifetime estimation of hot section components reliant on TBCs.