Surrogate Model Research Articles

Coupled flow and energy models with phase change in permafrost from pore- to Darcy scale: Modeling and approximation

We consider coupled models of flow and energy with phase change in thawing soils in permafrost at the pore- and Darcy scales. We develop the underlying models from first principles and define robust and conservative discretization algorithms based on cell-centered finite volume discretization on Cartesian grids and implicit time stepping for individual models. To couple the models we consider sequential approaches as well as time-lagging of some of the nonlinear properties. We also derive practical new surrogate models to obtain Darcy scale data from the pore-scale simulations. Finally, we study the sensitivity of models to the data in realistic scenarios. In particular we find that for the dynamics at the time scale of days, the nonlinear flow parameters have most significant impact on pressure, while temperature is less sensitive to data.

Design optimization of elastic metamaterials with multilayered honeycomb structure by Kriging surrogate model and genetic algorithm

Design optimization of elastic metamaterials with multilayered honeycomb structure by Kriging surrogate model and genetic algorithm

Towards a Digital Twin Framework in Additive Manufacturing: Machine Learning and Bayesian Optimization for Time Series Process Optimization

Laser directed-energy deposition (DED) offers notable advantages in additive manufacturing (AM) for producing intricate geometries and facilitating material functional grading. However, inherent challenges such as material property inconsistencies and part variability persist, predominantly due to its layer-wise fabrication approach. Critical to these challenges is heat accumulation during DED, influencing the resultant material microstructure and properties. Although closed-loop control methods for managing heat accumulation and temperature regulation are prevalent in DED literature, few approaches integrate real-time monitoring, physics-based modeling, and control simultaneously in a cohesive framework. To address this, we present a digital twin (DT) framework for real-time model predictive control of process parameters of the DED for achieving a specific process design objective. To enable its implementation, we detail the development of a surrogate model utilizing Long Short-Term Memory (LSTM)-based machine learning which uses Bayesian Inference to predict temperatures across various spatial locations of the DED-built part. This model offers real-time predictions of future temperature states. In addition, we introduce a Bayesian Optimization (BO) method for Time Series Process Optimization (BOTSPO). Its foundational principles align with traditional BO, and its novelty lies in our unique time series process profile generator with a reduced dimensional representation. BOTSPO is used for dynamic process optimization in which we deploy BOTSPO to determine the optimal laser power profile, aiming to achieve desired mechanical properties in a DED build. The identified profile establishes a process trajectory that online process optimizations aim to match or exceed in performance. This paper elucidates components of the digital twin framework, advocating its prospective consolidation into a comprehensive digital twin system for AM.

More Efficient Energy Management for Hybrid AC/DC Microgrids With Hard Constrained Neural Network-Based Conversion Loss Surrogate Models

More Efficient Energy Management for Hybrid AC/DC Microgrids With Hard Constrained Neural Network-Based Conversion Loss Surrogate Models

DyMMM-LEAPS: An ML-based framework for modulating evenness and stability in synthetic microbial communities

There have been a growing number of computational strategies to aid in the design of synthetic microbial consortia. A framework to identify regions in parametric space to maximize two essential properties, evenness and stability, is critical. In this study, we introduce DyMMM-LEAPS (Dynamic Multi-species Metabolic Modeling - Locating Evenness And stability in large Parametric Space), an extension of the DyMMM framework. Our method explores the large parametric space of genetic circuits in synthetic microbial communities to identify regions of evenness and stability. Due to the high computational costs of exhaustive sampling, we utilize adaptive sampling and surrogate modeling to reduce the number of simulations required to map the vast space. Our framework predicts engineering targets and computes their operating ranges to maximize the probability of the engineered community to have high evenness and stability. We demonstrate our approach by simulating five co-cultures and one three-strain culture with different social interactions (cooperation, competition, and predation) employing quorum-sensing based genetic circuits. In addition to guiding circuit tuning, our pipeline gives an opportunity for a detailed analysis of pockets of evenness and stability for the circuit under investigation, which can further help dissect the relationship between the two properties. DyMMM-LEAPS is easily customizable and can be expanded to a larger community with more complex interactions.

Performance prediction and optimization strategy for LNG multistage centrifugal pump based on PSO-LSSVR surrogate model

The Multistage LNG cryogenic submersible pump is one of the crucial power equipment in LNG transport process. However, the high hydraulic losses and low efficiency of this pump make it necessary to optimize the structure of its impeller in order to reduce energy losses. In this study, an optimization strategy based on particle swarm optimization (PSO) and least squares support vector regression (LSSVR) machine is proposed to optimize the structure of impeller. The optimization of the impeller structure with the PSO-LSSVR method increased the head and efficiency of the LNG cryogenic submersible pump by 0.57% and 2.72%, respectively. Comparing the calculation results of PSO-LSSVR method with the CFD results, we found that the relative error between them don’t exceed 3%, which verified the calculation accuracy of the method. Then, an enstrophy dissipation theory is introduced to quantitatively analyze the pump energy loss. Comparing the impellers before and after optimization revealed that the maximum reduction in impeller energy loss was 18%.. The distribution and mechanism of vortex generation inside the impeller were analyzed with the Q-criterion and the relative vortex transport equation. The relative vortex transport equation revealed that the horseshoe vortex structure at the leading edge of the optimized impeller blade was suppressed, and the relative vortex stretching term and the Coriolis force term, which dominate the evolution of the vortex structure, were reduced. This study will provide related reference for the optimal design of Multistage LNG cryogenic submersible pumps.

Towards spatio-temporal prediction of cavitating fluid flow with graph neural networks

In this article, we present a deep learning-based surrogate model for spatio-temporal prediction of cavitating fluid flow. Specifically, we introduce a finite element-inspired rotation equivariant hypergraph neural network for inferring and predicting dynamical behaviors of cavitating flow. We generate ground-truth spatial–temporal data by simulating a full-order variational system based on homogeneous mixture-based cavitation theory. We consider the flow past a NACA0012 hydrofoil to examine the predictive ability of the proposed graph neural network for cavitation dynamics. Results demonstrate that the network achieves stabilized and accurate temporal predictions of the system states, successfully forecasting the evolution patterns of individual cavitation events. Additionally, comparisons of predicted fluid loading coefficients are in good agreements with the ground-truth values. We also discuss some challenges encountered in the long-term prediction of flow patterns across multiple cavitation events. The proposed framework has implications for design optimization, active control and development of a physics-based digital twin of a cavitating marine propeller.

Harmonic Current Optimization for Torque Ripple Reduction in Permanent Magnet Synchronous Machine Drives Based on Torque Ripple Surrogate Model

Harmonic Current Optimization for Torque Ripple Reduction in Permanent Magnet Synchronous Machine Drives Based on Torque Ripple Surrogate Model

Wastewater Disinfection with Photodynamic Treatment and Evaluation of its Ecotoxicological Effects

Research has demonstrated the presence of viruses in wastewater (WW), which can remain viable for a long period, posing potential health risks. Conventional WW treatment methods involving UV light, chlorine and ozone efficiently reduce microbial concentrations, however, they produce hazardous byproducts and microbial resistance that are detrimental to human health and the ecosystem. Hence, there is a need for novel disinfection techniques. Antimicrobial Photodynamic Inactivation (PDI) emerges as a promising strategy, utilizing photosensitizers (PS), light, and dioxygen to inactivate viruses. This study aims to assess the efficacy of PDI by testing Methylene Blue (MB) and the cationic porphyrin TMPyP as PSs, along a low energy consuming white light source (LED) at an irradiance of 50 mW/cm2, for the inactivation of bacteriophage Phi6. Phi6 serves as an enveloped RNA-viruses surrogate model in WW. PDI experiments were conducted in a buffer solution (PBS) and real WW matrices (filtered and non-filtered). Considering the environmental release of the treated effluents, this research also evaluated the ecotoxicity of the resulting solution (post-PDI treatment effluent) on the model organism Daphnia magna, following the OECD immobilization technical 202 guideline. Daphnids were exposed to WW containing the tested PS at different concentrations and dilutions (accounting for the dilution factor during WW release into receiving waters) over 48 h. The results indicate that PDI with MB efficiently inactivated the model virus in the different aqueous matrices, achieving reductions superior to 8 log10 PFU/mL, after treatments of 5 min in PBS and of ca. 90 min in WW. Daphnids survival increased when subjected to the PDI-treated WW with MB, considering the dilution factor. Overall, the effectiveness of PDI in eliminating viruses in WW, the fading of the toxic effects on daphnids after MB’ irradiation and the rapid dilution effect upon WW release in the environment highlight the possibility of using MB in WW PDI-disinfection.

Predicting biomass comminution: Physical experiment, population balance model, and deep learning

An extended population balance model (PBM) and a deep learning-based enhanced deep neural operator (DNO+) model are introduced for predicting particle size distribution (PSD) of comminuted biomass through a large knife mill. Experimental tests using corn stalks with varied moisture contents, mill blade speeds, and discharge screen sizes are conducted to support model development. A novel mechanism in the extended PBM allows for including additional input parameters such as moisture content, which is not possible in the original PBM. The DNO+ model can include influencing factors of different data types such as moisture content and discharge screen size, which significantly extends the engineering applicability of the standard DNO model that only admits feed PSD and outcome PSD. Test results show that both models are remarkably accurate in the calibration or training parameter space and can be used as surrogate models to provide effective guidance for biomass preprocessing design.

Data-driven surrogate modelling of multistage Taylor cone–jet dynamics

The Taylor cone jet is an electrohydrodynamic flow typically induced by applying an external electric field to a liquid within a capillary, commonly utilized in colloidal thrusters. This flow generation involves a complex multiphase and multiphysics process, with stability contingent upon specific operational parameters. The operational window is intrinsically linked to flow rate and applied electric voltage magnitude. High voltages can induce atomization instabilities, resulting in the production of an electrospray. Our study presents initially a numerical investigation into the atomization process of a Taylor cone jet using computational fluid dynamics. Implemented within OpenFOAM, our numerical model utilizes a volume-of-fluid approach coupled with Maxwell's equations to incorporate electric body forces into the incompressible Navier–Stokes equations. We employ the leaky-dielectric model, subjecting the interface between phases to hydrodynamic surface tension and electric stress (Maxwell stress). With this model, we studied the droplet breakup of a heptane liquid jet, for a range of operation of 1.53–7.0 nL s−1 and 2.4–4.5 kV of extraction. First, the developed high-fidelity numerical solution is studied for the jet breakup and acceleration of the droplets. Second, we integrate a machine learning model capable of extending the parametric windows of operation. Additionally, we explore the influence of extractor and acceleration plates on colloidal propulsion systems. This work offers a numerical exploration of the Taylor cone–jet transition and droplet acceleration using novel, numerically accurate approaches. Subsequently, we integrate machine learning models, specifically an artificial neural network and a one-dimensional convolutional neural network, to predict the jet's performance under conditions not previously evaluated by computationally heavy numerical models. Notably, we demonstrate that the convolutional neural network outperforms the artificial neural network for this type of application data, achieving a 2% droplet size prediction accuracy.

Energy absorption of a novel auxetic structure reinforced by embedding tubes

To enhance structural crashworthiness while achieving lightweight design, this paper proposes a novel configuration that embeds a thin-walled square tube into auxetic structure which can take full advantage of the interaction between the auxetic structure and embedding tubes. The finite element model of auxetic structures reinforced by embedding tubes is established and validated by the compression experiments. The crashworthiness performance of auxetic structure (AUX), auxetic structure reinforced by embedding tubes (AET), and auxetic structure-filled tubes (AFT) are compared to find that the AET have the highest energy absorption and specific energy absorption (SEA). Through parametric analysis, designing the angles between short and long beams (θ1, θ2) as 70° and 40° can contribute to the crashworthiness performance, respectively. A suitable range for long beam thickness lies between 1.4 mm and 1.6 mm, while a short beam thickness of around 1.2 mm is more effective. The specific configuration can enhance the SEA and reduces peak crushing force (PCF). Increasing the thickness of the embedded tube can improve the crashworthiness by sacrificing the weight. It is a good balance between the crashworthiness performance and weight by designing the tube thickness as 1.2 mm. An analytical model is also established to predict the energy absorption of auxetic structure reinforced by embedding tubes under axial impact loadings. The surrogate modeling technique and the NSGA-II algorithm are also employed to optimize the AET configuration. The results show that the SEA of the optimized structure can increase from 14.5 kJ/kg to 18.6 kJ/kg, while the PCF can be reduced from 164.6 kN to 124.3 kN. Therefore, the auxetic structure reinforced by embedding tubes can play a critical role in the engineering field to absorb energy.

Does architectural design require single-objective or multi-objective optimisation? A critical choice with a comparative study between model-based algorithms and genetic algorithms

Efficiency and accuracy have been challenging in the design optimisation process driven by building simulation. The literature review identified the limitations of previous studies, prompting this study to explore the performance of single-objective versus multi-objective efficiency and accuracy on equivalent problems based on control variables and to consider more algorithmic options for a broader range of designs. This study constructed a comparative energy-related experiment whose results are in the same unit, either as a single-objective optimisation or split into two objectives. The project aims to reduce annual energy consumption and increase solar utilisation potential. Our approach focuses on the use of a surrogate modelling algorithm, Radial Basis Function Optimisation Algorithm (RBFOpt), with its multi-objective version RBFMOpt, to optimise the energy performance while quickly identifying new energy requirements for an iterative office building design logic, contrast to traditional genetic-algorithm-driven. In addition, the research also conducted a comparative study between RBFOpt and Covariance Matrix Adaptation Evolutionary Strategies (CMAES) in a single-objective comparison and between RBFMOpt and Nondominated Sorting Genetic Algorithm II (NSGA-II) in a multi-objective optimisation process. The comparison of these sets of Opt algorithms with evolutionary algorithms helps to provide data-driven evidence to support early design decisions.

Overcoming the lack of distribution data for range‐restricted habitat specialist frogs

AbstractWith over 40% of species threatened with extinction, the distributional range of most amphibians is still unknown, generating uncertainty whether species are naturally rare, under‐sampled, or difficult to detect. We implemented a modelling approach that uses bromeliads as surrogates to predict the distribution of habitat specialist frogs that lack distributional data. We aimed to predict and survey potential new sites for the occurrence of a rare and microendemic bromeliad‐dwelling frog from the Atlantic Rainforest of Brazil. We used Maxent to predict suitable areas based on climate and topographic profiles, combined with 21 occurrence records of bromeliads within which our target frog species (Crossodactylodes itambe) is strictly restricted. The bromeliad‐based models identified four areas potentially suitable, and subsequent surveys revealed an entirely new species of a rare bromeliad‐dwelling frog in one of the areas. We demonstrate that using easy‐to‐survey surrogate species that have a strong relationship with species that are hard to detect has enormous potential to reveal crucial information on the potential ranges and distribution of cryptic taxa. Such a surrogate modelling approach could be extended to other habitat specialist species that lack distribution data, such as amphibians restricted to specific refugia and/or reproduction sites. This could improve the targeting of surveys and increase understanding of the patterns of rarity and the drivers of species distribution, especially for areas with high endemicity and range‐restricted frogs.

On the limit behavior of lattice-type metamaterials with bi-stable mechanisms

This paper explores 2D lattice-type metamaterials featuring bi-stable unit cells and predefined symmetry configurations. Our investigation delves into the emergence of phase-transition patterns and probes the limit behavior of these lattice arrangements. We develop a macroscale generalized standard material model based on a quasi-convexified free energy framework and validate it through a specific lattice configuration—a 1D chain of bi-stable elements—where the relaxed free energy is analytically derived.To assess the limit behavior of the analyzed lattices, we elucidate the connection between energy release and topology at the microscale. This insight aids in identifying lattice configurations yielding high extrinsic energy dissipation density. We introduce the concept of dissipation efficiency and quantify it across all examined lattices under different loading conditions. Transverse loads prevail in the studied configurations and exhibit a detrimental effect by diminishing extrinsic energy dissipation during metamaterial phase transitions. To facilitate a comprehensive numerical assessment of diverse lattice configurations, we employ a surrogate model of the bistable element. This approach enables an efficient evaluation of sampling volumes constituted by numerous unit cells.

Optimization Design of Rocket Deflector Channel Surface Based on Kriging Surrogate Model

This study aims to optimize the gas flow environment of a rocket’s deflector channel. Numerical calculations and analysis were conducted on the influencing factors of a certain type of single-sided deflector channel, with the injection coefficient used as the evaluation index for the flow clearance. The study identified the deflector channel’s impact angle, curvature diameter of the bending section, and distance from the top of the deflector surface to the bottom as important structural parameters. There exists a complex nonlinear relationship between these parameters and the gas flow guiding effect of the deflector channel. The kriging surrogate model was employed for the optimization of the deflector channel’s surface. The results showed that, while meeting the basic requirements of the deflector channel, the optimized design improved the gas flow guiding effect by 27.58%, significantly enhancing the gas flow environment. The optimization error was within 1%, confirming the effectiveness and accuracy of the optimization. This study provides a reference for the future design and optimization of gas flow deflector channels, with the potential to improve the performance and efficiency of rocket deflector channels.

Multi-objective optimization of a hybrid carbon capture plant combining a Vacuum Pressure Swing Adsorption (VPSA) process with a Carbon Purification unit (CPU)

The imperative challenge posed by climate change requires urgent actions to counteract the harmful effects of greenhouse gas emissions, particularly CO2, which contributes to approximately 80 % of emissions responsible for global warming. A hybrid system combining Vacuum Pressure Swing Adsorption (VPSA) unit with a Cryogenic Carbon Purification Unit (CPU) is evaluated to enhance recovery and purity of CO2captured from flue gas containing CO2concentration ranging from 5 % to 20 %. VPSA preconcentrates the CO2and CPU completes the separation and purifies the CO2. The study uses surrogate models for multi-objective optimization, considering energy consumption, cost, and CO2recovery, providing a time-efficient approach for investigating computationally demanding processes. Results from the study indicate that the hybrid system achieves over 90 % recovery for flue gas concentration range considered, while ensuring the production of high-purity CO2(>99.99 %) suitable for transportation. A trade-off analysis reveals the balance between recovery, electricity consumption, and economic viability. A sensitivity analysis identifies parameters influencing recovery and energy consumption, providing guidance for future optimization efforts. The techno-economic analysis highlights the impact of electricity prices and carbon taxes on total costs, identifying an optimum towards higher recovery values under rising carbon taxes. Furthermore, the research underscores concentration-dependent economic feasibility, emphasizing the attractiveness of concentrations above 10 % compared other technologies, which require higher concentrations. For an electricity price of 75 €.MWh−1, the total cost of the CO2capture hydride system considering CO2emissions with carbon tax of 100 €.tCO2-1for concentrations ranging from 10 % to 20 % is from 123 to 80 €.tCO2-1, respectively. The analysis of the electricity source shows the importance of a low-carbon emission energy mix for optimal carbon emissions reduction.

6σ Robust Lightweight Design of the GIL/GIS Bus Based on Co‐Simulation of Insulation, Current Flow, and Mechanical Stress

GIL/GIS bus is widely used in the field of power transmission and distribution, and the demand for lightweight is becoming more and more prominent. To reduce the weight of the bus and ensure the safety and reliability of the structure, a new 6σ robust design method of the GIL/GIS bus based on electric‐thermal‐mechanical co‐simulation is proposed. Taking a 550 kV/5000A bus as the initial structure, considering the uncertainty of product size and performance, a 6σ robust lightweight model of bus insulation electric field strength, current temperature rise, and mechanical stress multi‐performance co‐simulation is established. The finite element numerical calculation method is used to calculate and analyze the performance parameters of the bus. The Kriging surrogate model is used to form the response surface, and the optimal bus structure is obtained by genetic algorithm. The results show that after the robustness optimization, the performance of the bus meets the design requirements, the weight per meter decreased from 65.11 to 39.12 kg, and the lightweight effect is remarkable. Compared with the traditional deterministic optimization design, the reliability of the structure is improved. The new method of the GIL/GIS bus lightweight design proposed in this study has high feasibility and application prospects, which can provide new ideas for the development of gas‐insulated electrical equipment. © 2024 Institute of Electrical Engineer of Japan and Wiley Periodicals LLC.

Stochastic finite element-based reliability of corroded pipelines with interacting corrosion clusters

The performance of corroded carbon steel pipelines over the course of their design life is generally assessed by probabilistic variables with explicit limit state functions, rather than the realistic representation with stochastic spatial variability and implicit failure considerations. This could be due to the complexities associated with the uncertainty quantification and performance estimation approaches. The consideration of random process representation of corrosion defect propagation and material properties, along with computationally effective implicit formulation is expected to lead to accurate reliability outcomes. This paper proposes a stochastic-based reliability framework considering suitable failure modes represented with surrogate models, that lead to time-variant reliability estimation. The approach combines surrogate computational model with scalar random variables and random field discretisation of underlying characteristics to generate experimental designs and corresponding surrogate models over a time period, which are used to derive reliability estimates. The outcomes of this approach are compared with the results of explicit time-dependent functions. It was observed that reliability estimates of the corroded pipeline change rapidly after the fifth year, providing a much lesser probability of failure (of 3.08×10−3 at 30th-year) compared to the existing models (of 1.80×10−2 at 30th-year), thereby providing an effective pathway for risk-based maintenance and management.

Statistical Verification using Surrogate Models and Conformal Inference and a Comparison with Risk-Aware Verification

Uncertainty in safety-critical cyber-physical systems can be modeled using a finite number of parameters or parameterized input signals. Given a system specification in Signal Temporal Logic (STL), we would like to verify that for all (infinite) values of the model parameters/input signals, the system satisfies its specification. Unfortunately, this problem is undecidable in general. Statistical model checking (SMC) offers a solution by providing guarantees on the correctness of CPS models by statistically reasoning on model simulations. We propose a new approach for statistical verification of CPS models for user-provided distribution on the model parameters. Our technique uses model simulations to learn surrogate models , and uses conformal inference to provide probabilistic guarantees on the satisfaction of a given STL property. Additionally, we can provide prediction intervals containing the quantitative satisfaction values of the given STL property for any user-specified confidence level. We compare this prediction interval with the interval we get using risk estimation procedures. We also propose a refinement procedure based on Gaussian Process (GP)-based surrogate models for obtaining fine-grained probabilistic guarantees over sub-regions in the parameter space. This in turn enables the CPS designer to choose assured validity domains in the parameter space for safety-critical applications. Finally, we demonstrate the efficacy of our technique on several CPS models.