Multi-fidelity Simulation Research Articles

Towards accelerating discovery via physics-driven and interactive multi-fidelity Bayesian Optimization

Abstract Both computational and experimental material discovery bring forth the challenge of exploring multidimensional and often non-differentiable parameter spaces, such as phase diagrams of Hamiltonians with multiple interactions, composition spaces of combinatorial libraries, processing spaces, and molecular embedding spaces. Often these systems are expensive or time-consuming to evaluate a single instance, and hence classical approaches based on exhaustive grid or random search are too data intensive. This resulted in strong interest towards active learning methods such as Bayesian optimization (BO) where the adaptive exploration occurs based on human learning (discovery) objective. However, classical BO is based on a predefined optimization target, and policies balancing exploration and exploitation are purely data driven. In practical settings, the domain expert can pose prior knowledge on the system in form of partially known physics laws and often varies exploration policies during the experiment. Here, we propose an interactive workflow building on multi-fidelity BO (MFBO), starting with classical (data-driven) MFBO, then expand to a proposed structured (physics-driven) sMFBO, and finally extending it to allow human in the loop interactive iMFBO workflows for adaptive and domain expert aligned exploration. These approaches are demonstrated over highly non-smooth multi-fidelity simulation data generated from an Ising model, considering spin-spin interaction as parameter space, lattice sizes as fidelity spaces, and the objective as maximizing heat capacity. Detailed analysis and comparison show the impact of physics knowledge injection and real time human decisions for improved exploration with increased alignment to ground truth.

Efficient inverse design optimization through multi-fidelity simulations, machine learning, and boundary refinement strategies

Efficient inverse design optimization through multi-fidelity simulations, machine learning, and boundary refinement strategies

An efficient multi-fidelity simulation approach for performance prediction of adaptive cycle engines

Adaptive cycle engine (ACE) with multi-stream, modulated bypass ratio, and high-thrust/high-efficiency mode provides the capability of multiple mission adaptation for the next-generation aviation propulsion system. The low-fidelity zero-dimensional (0D) performance simulation method is commonly adopted in conventional engines such as turbofan and turbojet. However, it is hard to accommodate well to ACE with complex variable geometry schemes and strong interaction between components. This paper presents an efficient multi-fidelity simulation approach, often referred to as component zooming, for predicting the performance of ACE with the three-stream configuration. The one-dimensional (1D) mean-line models of the adaptive fan and low-pressure turbine (LPT) are integrated into the 0D ACE model by the iteratively-coupled method. The performance predicted by the multi-fidelity ACE models with different component zooming strategies is compared. The maximum deviations of thrust, specific fuel consumption, turbine inlet temperature, and bypass ratio between the 0D/1D Adaptive Fan/1D LPT coupled model and 0D ACE model are −10.8%, −4.4%, −5.4% (−75 K), and 1.6%, respectively, which are considerable and non-negligible for the application in the design stage of ACE. The computing time of all the multi-fidelity models is less than 2 minutes. The effects of the key aerodynamic parameters of the adaptive fan and LPT on the engine performance are also evaluated. The proposed approach provides a generic and efficient solution for multi-fidelity ACE performance prediction with acceptable computing resource requirements and time cost, which is applicable in the engine conceptual and preliminary design stage.

A novel vortex-based velocity sampling method for the actuator-line modeling of floating offshore wind turbines in windmill state

The fluid-dynamic simulation of wind turbine aerodynamics is typically tackled by applying multi-fidelity computational tools. In this context, the so-called actuator line model combines a low-fidelity treatment of the rotor with a high-fidelity resolution of the wake. In this paper, a novel formulation of the actuator line model proposes a vortex-based method to sample the flow around the rotor to rigorously assign the forces imparted by the blades. This new technique is implemented into an in-house code developed within the OpenFOAM environment, and it is validated against wind-tunnel experiments on a laboratory-scale horizontal-axis wind turbine operated in fixed-bottom and floating conditions. The calculations are also compared against multi-fidelity simulations performed, on the same test case, in the frame of the OC6 Phase III project. The simulation results, obtained after a systematic analysis and selection of the model parameters, exhibit a remarkable agreement with the available experiments and place the present code in the proper ranking of fidelity levels, in-between momentum-balance methods and blade-resolved CFD models. Finally, the calculations for surge and pitch platform motions demonstrate the capability of the proposed technique to reliably predict the aerodynamics of turbine rotors in dynamic operation at affordable computational cost.

Multi-fidelity simulation of aeroengine for far-off-design conditions using iterative coupled method based on auxiliary maps

Iterative coupled methods are widely used in multi-fidelity simulation of rotating components due to the simple implementation, which iteratively eliminates the errors between the computational fluid dynamics models and approximate characteristic maps. However, the convergence and accuracy of the iterative coupled method are trapped in characteristic maps. In particular, iterative steps increase sharply as the operation point moves away from the design point. To address these problems, this paper developed an auxiliary iterative coupled method that introduces the static-pressure-auxiliary characteristic maps and modification factor of mass flow into the component-level model. The developed auxiliary method realized the direct transfer of static pressure between the high-fidelity models and the component-level model. Multi-fidelity simulations of the throttle characteristics were carried out using both the auxiliary and traditional iterative coupled methods, and the simulation results were verified using the experimental data. Additionally, the consistency between the auxiliary and traditional iterative coupled methods was confirmed. Subsequently, multi-fidelity simulations of the speed and altitude characteristics were also conducted. The auxiliary and traditional iterative coupled methods were evaluated in terms of convergence speed and accuracy. The evaluation indicated that the auxiliary iterative coupled method significantly reduces iterative steps by approximately 50% at the near-choked state. In general, the auxiliary iterative coupled method is preferred as a development of the traditional iterative coupled method in the near-choked state, and the combined auxiliary-traditional iterative coupled method provides support for successful multi-fidelity simulation in far-off-design conditions.

Leveraging deep reinforcement learning for design space exploration with multi-fidelity surrogate model

Design automation is undergoing a new generation of changes caused by artificial intelligence technologies represented by deep learning and reinforcement learning. Notably, the advantages of deep reinforcement learning in addressing solution optimisation and decision-making tasks with cognitive automation functionality have garnered attention in design. In the context of surrogate model-driven engineering design optimisation, this paper addresses current research challenges such as reliance on domain knowledge for local development, shortcomings in the self-learning and adaptive capabilities of optimisation algorithms for global exploration, etc. Centred around the deep reinforcement learning model, Deep Q-learning, and complemented by self-organising maps and neural network technologies, we propose a methodology considering multi-fidelity simulation data for design space exploration. This approach effectively reduces sampling costs and enables the optimisation model to learn the optimal direction for high-precision predictions and achieve rapid, accurate optimisation. Finally, the effectiveness of the proposed method is comprehensively validated through four typical optimisation scenarios and a case study involving the optimisation of a wheeled robot's suspension swing arm structure. This work will be a crucial reference for applying deep reinforcement learning in simulation-driven engineering design optimisation.

Multi-Fidelity Simulation of Gas Turbine Overall Performance by Directly Coupling High-Fidelity Models of Multiple Rotating Components

Multi-Fidelity Simulation of Gas Turbine Overall Performance by Directly Coupling High-Fidelity Models of Multiple Rotating Components

Central cone film cooling scheme of turbofan engine based on multi-fidelity simulation

To improve the computational efficiency of multi-fidelity simulation, an improved fully coupled method was proposed, which was applied to the coupled simulation of a 3-D model of an exhaust system and a 0-D model of a turbofan engine. Different from conventional approach, by “truncating” the engine model, the CFD calculation of the exhaust system can be decoupled from the solution of nonlinear equations in the engine model, and an outer iteration equation set was constructed to solve the multi-fidelity model. It was found that the improved fully coupled method has better computational efficiency than traditional methods while maintaining equivalent computational results. Then, the improved fully coupled method was applied to the design of the central cone film cooling system, the influence of cooling gas bleed mechanism, the hole open area ratio of the central cone and hole density, hole direction, hole shape on the infrared characteristics was investigated. The impact of central cone film cooling on the overall performance of the engine was quantitatively evaluated. Combined with various favorable factors, when maintaining the HP rotor speed unchanged, the integrated infrared radiation intensity of the central cone under supersonic cruise conditions is reduced by 93 % compared to the uncooled state, while the engine thrust increases by 0.3 % and fuel consumption rate increases by 0.18 %. The multi-fidelity simulation method proposed in this paper and the conclusions obtained are of great significance for the structural design of low-infrared exhaust systems.

Enhancing Wind Farm Efficiency Through Active Control of the Atmospheric Boundary Layer’s Vertical Entrainment of Momentum

In contemporary wind farm design, the primary focus has traditionally been on reducing wake interference to optimize energy capture from horizontal wind flows. However, with the scaling up of wind farms, their interaction with the Atmospheric Boundary Layer (ABL) evolves, making vertical entrainment the main mechanism for the exchange of momentum and energy. This study introduces a methodical approach to augment the efficiency of large-scale offshore wind farms by actively controlling this vertical entrainment of momentum within the ABL. The strategy involves the precise engineering of advection fluxes to alter wind flow dynamics, utilizing turbines as effective vortex generators, toward a process of ”regenerative wind farming.” This setup aims to create a vorticity and vertical flux system akin to those observed in highly unstable ABLs. Expanding upon previous studies that focused on single Vertical Axis Wind Turbines (VAWTs), our research explores the implementation of multi-rotor systems equipped with lift-generating wings. These systems are designed to exert forces perpendicular to the prevailing wind direction, thus creating trailing vortices and directing the flow orthogonally for improved vertical advection. This research is part of a comprehensive investigative framework that combines experiments and multifidelity simulations. The current study extends those findings to wind farm simulations, aiming to assess the impact of ABL control on a full wind farm scale. The first part of the work validates an established analytical wind farm performance model against real wind farm data for thirty-one wind farms in the North Sea and Baltic Sea. The results confirm the predicted trend of decreased performance with increased wind farm size and density. The model is used to calculate the performance of a wind farm for varying regimes of vertical entrainment due to the creation of large-scale circulatory systems. The results are compared against 3D vortex simulations of the full wind farm in ”regenerative wind farming” mode. Our results demonstrate a notable improvement in wind speeds at the turbine hub height and the potential to double the feasible density of wind farms without compromising efficiency compared to traditional setups. These findings suggest a promising pathway towards a more sustainable and profitable future in wind energy, achieved through the strategic manipulation of ABL momentum, regenerating the energy in the wind farm.

SEAHOWL: Partitioned Multiphysics and Multifidelity Modelling of Wind Turbines with Monolithically Coupled Elastodynamics

We introduce SEAHOWL (Servo-Elasto-Aero-Hydro Offshore Wind Lab), a novel multiphysics multifidelity simulation framework for large-scale onshore, offshore, and floating wind turbines. For structural dynamics, multibody and finite element problems are coupled monolithically and solved within a single system of equations through Project Chrono, providing high numerical stability and optimal coupling accuracy. Combined with partitioned multiphysics, modularity is at the heart of the framework with various numerical methods and solvers (internal or external) available for each physics. This flexibility allows for the selection a target trade-off between numerical accuracy and computational efficiency on a use-case basis. Simulations showcased here have been selected through the prism of industrial R&D challenges, covering: controller response and loads over the operational wind range, 3P effect for different levels of fidelity for the blades, floating wind turbine response under irregular waves and turbulent wind, and innovative multibody application for 3P fatigue mitigation. SEAHOWL is cross-validated against the reference OpenFAST whole-turbine simulator from NREL, showing overall good agreement while differences between the two frameworks are highlighted.

Direct multi-fidelity integration of 3D CFD models in a gas turbine with numerical zooming method

Multi-fidelity simulation improves the simulation accuracy and captures more detailed information about aero engines under limited computing resources, which is implemented by coupling different levels of models using numerical zooming methods. However, there is an obvious problem in traditional zooming methods such as the iterative coupled zooming method or mini-map method: both the convergence and accuracy depend highly on the component general characteristic maps. Based on the investigation of a micro gas turbine, a direct zooming method (Cycle with CFD in it, CWCFD) is developed. It directly embeds the 3D CFD compressor and turbine model into a 0D component-level model without component general characteristic maps. Then, the CWCFD zooming method is compared with the traditional 0D component-level model in terms of the throttle characteristics of the micro gas turbine, and the experimental data of the ground test is performed to verify the effectiveness of the CWCFD zooming methods. The results indicate that the CWCFD zooming method matches well with the test data better than the traditional 0D component-level model.

Efficient analysis of composites manufacturing using multi-fidelity simulation and probabilistic machine learning

This paper introduces an innovative approach for the efficient analysis of composites manufacturing processes and phenomena. The method combines low- and high-fidelity simulation schemes with limited amounts of experimental data to train surrogate machine learning (ML) models. Guided by a novel approach, Spatially Weighted Gaussian Process Regression (SWGPR), a predictive model is efficiently constructed and calibrated by assigning datapoint-dependent noise levels to simulation points, establishing a multi-scale data-driven uncertainty structure. This study demonstrates the effectiveness of the method in accurately predicting process-induced deformations (PIDs) for L-shaped cross-ply laminates using minimal experimental efforts. The presented method aims to provide a cost-effective and broadly applicable framework for understanding and improving the design, development, and manufacturing of composites.

Multi-fidelity simulation of aeroengine across wide operation range using auxiliary fully coupled method

Multi-fidelity simulations for the design point or equilibrium lines of aeroengines are implemented using fully coupled methods, simultaneously optimizing computational resources and speed. Fully coupled methods remove the requirement of characteristic maps by directly incorporating computational fluid dynamics models of subcomponents into the thermodynamic model of an aeroengine. However, poor convergence remains an urgent problem for fully coupled methods, particularly when performing multi-fidelity simulations in the near-boundary region. In response to this challenge, an auxiliary fully coupled method was developed in this study. This auxiliary method introduces a parametric priori knowledge and auxiliary coordinate transformations into the existing direct fully coupled method. An automated multi-fidelity simulation platform was established in-house and validated using the KJ66 turbojet engine. The simulation results of the equilibrium line were then compared with experimental data to validate the high accuracy and reliability of the multi-fidelity simulation. Moreover, two fundamental selection schemes for matching parameters (mass flow or static pressure, transferred between the computational fluid dynamics and thermodynamic models) were implemented and systematically compared. The results demonstrate that the adaptive selection of matching parameters extends the application of multi-fidelity simulations to a broader operation range. In addition, the auxiliary fully coupled method was applied to a multi-fidelity simulation of a wider operation range, such as an adjustable nozzle area and variable power extraction. It was also compared with the existing direct fully coupled method with respect to the stable convergence scope. For the mass-flow matching scheme, the maximum stable convergence scope of the direct method was extended from 105 % to 315 % by the auxiliary method. For the static-pressure matching scheme, the auxiliary method also broadened the maximum stable convergence scope of the direct method from 55 % to 190 %. The comparison reveals that the auxiliary fully coupled method exhibits a wider range of far-off-design conditions, including the near-boundary region.

Exploring the role of simulator fidelity in the safety validation of learning‐enabled autonomous systems

AbstractThis article presents key insights from the New Faculty Highlights talk given at AAAI 2023, focusing on the crucial role of fidelity simulators in the safety evaluation of learning‐enabled components (LECs) within safety‐critical systems. With the rising integration of LECs in safety‐critical systems, the imperative for rigorous safety and reliability verification has intensified. Safety assurance goes beyond mere compliance, forming a foundational element in the deployment of LECs to reduce risks and ensure robust operation. In this evolving field, simulations have become an indispensable tool, and fidelity's role as a critical parameter is increasingly recognized. By employing multifidelity simulations that balance the needs for accuracy and computational efficiency, new paths toward comprehensive safety validation are emerging. This article delves into our recent research, emphasizing the role of simulation fidelity in the validation of LECs in safety‐critical systems.

On the characteristics of the wake of a wind turbine undergoing large motions caused by a floating structure: an insight based on experiments and multi-fidelity simulations from the OC6 project Phase III

Abstract. This study reports the results of the second round of analyses of the Offshore Code Comparison, Collaboration, Continued, with Correlation and unCertainty (OC6) project Phase III. While the first round investigated rotor aerodynamic loading, here, focus is given to the wake behavior of a floating wind turbine under large motion. Wind tunnel experimental data from the UNsteady Aerodynamics for FLOating Wind (UNAFLOW) project are compared with the results of simulations provided by participants with methods and codes of different levels of fidelity. The effect of platform motion on both the near and the far wake is investigated. More specifically, the behavior of tip vortices in the near wake is evaluated through multiple metrics, such as streamwise position, core radius, convection velocity, and circulation. Additionally, the onset of velocity oscillations in the far wake is analyzed because this can have a negative effect on stability and loading of downstream rotors. Results in the near wake for unsteady cases confirm that simulations and experiments tend to diverge from the expected linearized quasi-steady behavior when the rotor reduced frequency increases over 0.5. Additionally, differences across the simulations become significant, suggesting that further efforts are required to tune the currently available methodologies in order to correctly evaluate the aerodynamic response of a floating wind turbine in unsteady conditions. Regarding the far wake, it is seen that, in some conditions, numerical methods overpredict the impact of platform motion on the velocity fluctuations. Moreover, results suggest that the effect of platform motion on the far wake, differently from original expectations about a faster wake recovery in a floating wind turbine, seems to be limited or even oriented to the generation of a wake less prone to dissipation.

Sonic boom near field analysis of a Mach 5 wave-rider configuration

The following paper discuss the near field sonic boom results obtained with CFD simulations regarding a Mach 5 wave rider configuration derived from the STRATOFLY heritage. Sonic boom near field have emerged as a vital tool for the mitigation and understanding of the impact of supersonic and hypersonic aircraft configurations. The work done discusses the major consideration in the CFD approach for the flow field simulations in the vicinity of the aircraft, including grid generation techniques based on the NASA sonic boom predictions workshop carried out between 2014 and 2020 and the results obtained for different flight conditions and operations. The near field simulation necessitates of accurate predictions of shock wave propagation generated by the aircraft, employing the high-fidelity numerical schemes and turbulence models. These results will be the input for our propagation tool that manage to propagate the shocks from the aircraft to the ground. This work is carried out within the MORE&LESS project, that is an EU-funded project and has the aim to investigate the environmental impact of supersonic aircraft through multi-fidelity simulations and test campaigns for creating multidisciplinary holistic framework for the evaluation of the future supersonic aircraft, trajectories, and operations.

HydroChrono: An Open-Source Hydrodynamics Package for Project Chrono

In this paper we present the development and verification of HydroChrono, a hydrodynamics package for the Project Chrono physics engine. This package includes the implementation of hydrodynamics equations, the added mass for multibody systems, the development of I/O functions as well as a Python API, and comparison against standard reference cases and other existing tools. HydroChrono provides a flexible, fully open-source solution for simulating wave energy converters (WECs), floating offshore wind turbines (FOWTs) platforms, and other hydrodynamic systems.
 Here we show, via comparisons with existing tools for benchmark verification cases, that HydroChrono accurately models hydrodynamic forces - making it a useful tool for the design and optimization of these systems. Additionally, the integration of HydroChrono with Project Chrono offers access to finite element modeling capabilities and high-fidelity modelling – with Chrono’s existing coupling to CFD and SPH codes. This provides numerical modelers with a multifidelity simulation framework for designing and validating these systems.
 The development of HydroChrono provides a new, open-source solution for simulating hydrodynamic systems. Its compatibility with other simulation tools enables a more streamlined and efficient design process, advancing the field and providing new opportunities for innovation in this area.

A New Integrated Model for Simulating Adaptive Cycle Engine Performance Considering Variations in Tip Clearance

The low-fidelity simulation method cannot meet the requirements for predicting the performance of an adaptive cycle engine (ACE), especially when considering tip clearance variations in the compression and expansion systems. The tip clearances of the components of an ACE, such as the adaptive fan and turbine, vary drastically under different operating conditions. Though the tip clearance significantly impacts the engine’s performance, including its thrust and fuel consumption, variations in tip clearance are not considered in traditional ACE simulation models. This paper developed a new integrated model for predicting ACE performance, including multi-fidelity simulation models of the components and a newly developed, simplified model for predicting tip clearance. Specifically, the integrated model consists of a zero-dimensional (0D) engine performance simulation model, a three-dimensional (3D) adaptive fan numerical simulation model, a one-dimensional (1D) low-pressure-turbine (LPT) mean line model, and a multi-dimensional (MD) tip clearance prediction model. The integrated model solved the problem of considering the impact of tip clearance on an ACE and further improved the accuracy of thrust and fuel consumption predictions. Specifically, considering variations in the tip clearances under the design conditions, the differences in the thrust and specific fuel consumption (SFC) of the ACE are 1% and 0.3%, respectively. In conclusion, the integrated model provides a useful tool for evaluating the performance of an ACE while considering tip clearance variations.

Multi-fidelity reinforcement learning framework for shape optimization

Deep reinforcement learning (DRL) is a promising outer-loop intelligence paradigm which can deploy problem solving strategies for complex tasks. Consequently, DRL has been utilized for several scientific applications, specifically in cases where classical optimization or control methods are limited. One key limitation of conventional DRL methods is their episode-hungry nature which proves to be a bottleneck for tasks which involve costly evaluations of a numerical forward model. In this article, we address this limitation of DRL by introducing a controlled transfer learning framework that leverages a multi-fidelity simulation setting. Our strategy is deployed for an airfoil shape optimization problem at high Reynolds numbers, where our framework can learn an optimal policy for generating efficient airfoil shapes by gathering knowledge from multi-fidelity environments and reduces computational costs by over 30\%. Furthermore, our formulation promotes policy exploration and generalization to new environments, thereby preventing over-fitting to data from solely one fidelity. Our results demonstrate this framework's applicability to other scientific DRL scenarios where multi-fidelity environments can be used for policy learning.

Multi-Fidelity Simulation Modeling for Discrete Event Simulation: An Optimization Perspective

Multi-fidelity simulation is an effective approach to balancing speed and accuracy in expensive simulation, and its performance is affected by the quality of multi-fidelity simulation models. Building high-quality simulation models is non-trivial, especially for complex systems, because current manual modeling methods require sufficient domain knowledge and experience, increasing the labor and time costs. Motivated by the issues, this paper focuses on one of the most crucial simulation types, discrete event simulation, and develops a computer-aid multi-fidelity simulation modeling method called Optimization-based Multi-fidelity Simulation Modeling (OMFSM). OMFSM formulates multi-fidelity simulation modeling as a bi-objective simulation optimization problem to optimize speed and accuracy. An efficient optimization algorithm called Multi-objective Simulation Optimization based on Hypervolume (MOSO-HV) is tailored to select a set of high-quality models. Experimental results in a digital twin emergency department demonstrate that the computer-aid modeling method builds more and better multi-fidelity simulation models than manual modeling and reveal the effectiveness of MOSO-HV for OMFSM. The utility of OMFSM in multi-fidelity simulation is also justified by a real-world optimization problem. Note to Practitioners—Multi-fidelity simulation is an essential technique to fulfill the demand for accuracy analysis and quick decision-making in Industrial 4.0, such as digital twins and virtual reality. The quality of multi-fidelity simulation models significantly influences the performance of multi-fidelity simulation. Developers currently build multi-fidelity simulation models manually, and their experience determines the model’s quality. To reduce the labor and time costs in constructing high-quality multi-fidelity simulation models, we propose a computer-aid method named OMFSM from optimization for the first time. Experiments on a real case prove that OMFSM lightens the burden of manual modeling and provides more and better models for multi-fidelity simulation.