Search NASA⌕ Search

SEARCH · Search NASA

Results for “Physics-based simulation”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

790 records · Page 44

REACTIVE TRANSPORT MODELING USING A PARALLEL FULLY-COUPLED SIMULATOR BASED ON PRECONDITIONED JACOBIAN-FREE NEWTON-KRYLOV

Systems of multicomponent reactive transport in porous media that are large, highly nonlinear, and tightly coupled due to complex nonlinear reactions and strong solution-media interactions are often described by a system of coupled nonlinear partial differential algebraic equations (PDAEs). A preconditioned Jacobian-Free Newton-Krylov (JFNK) solution approach is applied to solve the PDAEs in a fully coupled, fully implicit manner. The advantage of the JFNK method is that it avoids explicitly computing and storing the Jacobian matrix during Newton nonlinear iterations for computational efficiency considerations. This solution approach is also enhanced by physics-based blocking preconditioning and multigrid algorithm for efficient inversion of preconditioners. Based on the solution approach, we have developed a reactive transport simulator named RAT. Numerical results are presented to demonstrate the efficiency and massive scalability of the simulator for reactive transport problems involving strong solution-mineral interactions and fast kinetics. It has been applied to study the highly nonlinearly coupled reactive transport system of a promising in situ environmental remediation that involves urea hydrolysis and calcium carbonate precipitation.

58 GEOSCIENCES↗

Identifying Suitable Degradation Parameters for Individual-Based Prognostics

The ultimate goal of most prognostic systems is accurate prediction of the remaining useful life of individual systems or components based on their use and performance. Traditionally, individual-based prognostic methods use a measure of degradation to make lifetime estimates. Degradation measures may include sensed measurements, such as temperature or vibration level, or inferred measurements, such as model residuals or physics-based model predictions. Often, it is beneficial to combine several measures of degradation into a single parameter. Parameter features such as trendability, monotonicity, and prognosability can be used to compare candidate prognostic parameters to determine which is most useful for individual-based prognosis. By quantifying these features for a given parameter, the metrics can be used with any traditional optimization technique to identify an appropriate parameter. This parameter may be used with a parametric extrapolation model to make prognostic estimates for an individual unit. The proposed methods are illustrated with an application to simulated turbofan engine data.

Diagnostics↗

U.S. Transport Task Force Meeting - April 2014 Final Report

The ultimate goal of the U.S. Transport Task Force is to develop a physics-based understanding of confinement and particle, momentum and heat transport in magnetic fusion devices. This understanding should be of sufficient depth that it allows the development of predictive models of plasma transport that can be validated against experiment, and then used to anticipate the future performance of burning plasmas in ITER, as well as to provide guidance to the design of next-step fusion nuclear science facilities. To achieve success in transport science, it is essential to characterize local fluctuations and transport in toroidal plasmas, to understand the basic mechanisms responsible for transport, and ultimately to control these transport processes. These goals must be pursued in multiple areas, including ion and electron thermal transport, particle and momentum transport, the physics of H-modes and the edge pedestal, Internal Transport Barriers, energetic particle transport and 3D effects on all the underlying transport processes. Demonstrating our understanding requires multiple, successful, quantitative tests of theory, simulation and modeling using experimental results in fusion-relevant and basic plasmas (i.e., verification and validation). The 2014 U.S. TTF meeting was held in April 2014 in San Antonio TX to provide a forum for leading scientists focused on the study of transport of particles, momentum and heat in fusion plasmas. Approximately 110 scientists from the US and several from the EU and from China attended and heard oral talks on recent transport results. Several poster sessions were also held. One day of plenary talks were followed by Breakout sessions and poster sessions that were held on focused topics, including L-H transition physics, energetic particles, transport in high performance plasmas, divertor particle and heat flux management and innovative divertor designs, fundamental turbulence studies, end edge transport shortfall. Most of the invited oral talks are archived at the meeting website, see http://ttf2014.ucsd.edu/TTF_2014/Presentations.html A book of abtracts for all presentations and posters at the meeting is also available, see http://ttf2014.ucsd.edu/TTF_2014/Home_files/TTF%202014%20Abstract%20Book%20Final.pdf Finally the program for the meeting is also available, see http://ttf2014.ucsd.edu/TTF_2014/Home_files/TTF%202014%20Meeting%20Schedule.pdf

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

2014 US TTF Meeting Summary

The ultimate goal of the U.S. Transport Task Force is to develop a physics-based understanding of confinement and particle, momentum and heat transport in magnetic fusion devices. This understanding should be of sufficient depth that it allows the development of predictive models of plasma transport that can be validated against experiment, and then used to anticipate the future performance of burning plasmas in ITER, as well as to provide guidance to the design of next-step fusion nuclear science facilities. To achieve success in transport science, it is essential to characterize local fluctuations and transport in toroidal plasmas, to understand the basic mechanisms responsible for transport, and ultimately to control these transport processes. These goals must be pursued in multiple areas, including ion and electron thermal transport, particle and momentum transport, the physics of H-modes and the edge pedestal, Internal Transport Barriers, energetic particle transport and 3D effects on all the underlying transport processes. Demonstrating our understanding requires multiple, successful, quantitative tests of theory, simulation and modeling using experimental results in fusion-relevant and basic plasmas (i.e., verification and validation). The 2014 U.S. TTF meeting was held in April 2014 in San Antonio TX to provide a forum for leading scientists focused on the study of transport of particles, momentum and heat in fusion plasmas. Approximately 110 scientists from the US and several from the EU and from China attended and heard oral talks on recent transport results. Several poster sessions were also held. One day of plenary talks were followed by Breakout sessions and poster sessions that were held on focused topics, including L-H transition physics, energetic particles, transport in high performance plasmas, divertor particle and heat flux management and innovative divertor designs, fundamental turbulence studies, end edge transport shortfall.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Validation of the National Solar Radiation Database (NSRDB) (2005-2012): Preprint

Publicly accessible, high-quality, long-term, satellite-based solar resource data is foundational and critical to solar technologies to quantify system output predictions and deploy solar energy technologies in grid-tied systems. Solar radiation models have been in development for more than three decades. For many years, the National Renewable Energy Laboratory (NREL) developed and/or updated such models through the National Solar Radiation Data Base (NSRDB). There are two widely used approaches to derive solar resource data from models: (a) an empirical approach that relates ground-based observations to satellite measurements and (b) a physics-based approach that considers the radiation received at the satellite and creates retrievals to estimate clouds and surface radiation. Although empirical methods have been traditionally used for computing surface radiation, the advent of faster computing has made operational physical models viable. The Global Solar Insolation Project (GSIP) is an operational physical model from the National Oceanic and Atmospheric Administration (NOAA) that computes global horizontal irradiance (GHI) using the visible and infrared channel measurements from the Geostationary Operational Environmental Satellites (GOES) system. GSIP uses a two-stage scheme that first retrieves cloud properties and then uses those properties in the Satellite Algorithm for Surface Radiation Budget (SASRAB) model to calculate surface radiation. NREL, the University of Wisconsin, and NOAA have recently collaborated to adapt GSIP to create a high temporal and spatial resolution data set. The product initially generates the cloud properties using the AVHRR Pathfinder Atmospheres-Extended (PATMOS-x) algorithms [3], whereas the GHI is calculated using SASRAB. Then NREL implements accurate and high-resolution input parameters such as aerosol optical depth (AOD) and precipitable water vapor (PWV) to compute direct normal irradiance (DNI) using the DISC model. The AOD and PWV, temperature, and pressure data are also combined with the MMAC model to simulate solar radiation under clear-sky conditions. The current NSRDB update is based on a 4-km x 4-km resolution at a 30-minute time interval, which has a higher temporal and spatial resolution. This paper demonstrates the evaluation of the data set using ground-measured data and detailed evaluation statistics. The result of the comparison shows a good correlation to the NSRDB data set. Further, an outline of the new version of the NSRDB and future plans for enhancement and improvement are provided.

14 SOLAR ENERGY↗

A New Physics-Based Modeling of Multiple Non-Planar Hydraulic Fractures Propagation

Because of the low permeability in shale plays, closely spaced hydraulic fractures and multilateral horizontal wells are generally required to improve production. Therefore, understanding the potential fracture interaction and stress evolution is critical in optimizing fracture/well design and completion strategy in multi-stage horizontal wells. In this paper, a novel fully coupled reservoir flow and geomechanics model based on the dual-lattice system is developed to simulate multiple non-planar fractures propagation. The numerical model from Discrete Element Method (DEM) is used to simulate the mechanics of fracture propagations and interactions, while a conjugate irregular lattice network is generated to represent fluid flow in both fractures and formation. The fluid flow in the formation is controlled by Darcy’s law, but within fractures it is simulated by using cubic law for laminar flow through parallel plates. Initiation, growth and coalescence of the microcracks will lead to the generation of macroscopic fractures, which is explicitly mimicked by failure and removal of bonds between particles from the discrete element network. We investigate the fracture propagation path in both homogeneous and heterogeneous reservoirs using the simulator developed. Stress shadow caused by the transverse fracture will change the orientation of principal stress in the fracture neighborhood, which may inhibit or alter the growth direction of nearby fracture clusters. However, the initial in-situ stress anisotropy often helps overcome this phenomenon. Under large in-situ stress anisotropy, the hydraulic fractures are more likely to propagate in a direction that is perpendicular to the minimum horizontal stress. Under small in-situ stress anisotropy, there is a greater chance for fractures from nearby clusters to merge with each other. Then, we examine the differences in fracture geometry caused by fracturing in cemented or uncemented wellbore. Moreover, the impact of intrinsic reservoir heterogeneity caused by the rock fabric and mineralogy on fracture nucleation and propagation paths is examined through a three-layered reservoir. Finally, we apply the method to a realistic heterogeneous dataset.

02 PETROLEUM↗

Improving Lidar Turbulence Estimates for Wind Energy

Remote sensing devices (e.g., lidars) are quickly becoming a cost-effective and reliable alternative to meteorological towers for wind energy applications. Although lidars can measure mean wind speeds accurately, these devices measure different values of turbulence intensity (TI) than an instrument on a tower. In response to these issues, a lidar TI error reduction model was recently developed for commercially available lidars. The TI error model first applies physics-based corrections to the lidar measurements, then uses machine-learning techniques to further reduce errors in lidar TI estimates. The model was tested at two sites in the Southern Plains where vertically profiling lidars were collocated with meteorological towers. This presentation primarily focuses on the physics-based corrections, which include corrections for instrument noise, volume averaging, and variance contamination. As different factors affect TI under different stability conditions, the combination of physical corrections applied in L-TERRA changes depending on the atmospheric stability during each 10-minute time period. This stability-dependent version of L-TERRA performed well at both sites, reducing TI error and bringing lidar TI estimates closer to estimates from instruments on towers. However, there is still scatter evident in the lidar TI estimates, indicating that there are physics that are not being captured in the current version of L-TERRA. Two options are discussed for modeling the remainder of the TI error physics in L-TERRA: machine learning and lidar simulations. Lidar simulations appear to be a better approach, as they can help improve understanding of atmospheric effects on TI error and do not require a large training data set.

17 WIND ENERGY↗

Risk-Informed Applications and Science – Past, Present, & Future

This is a presentation to be made at the University of Tokyo. The Idaho National Laboratory is creating the next-generation reliability- and risk-assessment methods and tools that support risk-informed decision-making by combining physics-based models with probabilistic quantification approaches. Integrating these two worlds of physics and probability using a simulation framework leads us to predictions based upon an approach called “computational risk assessment” which will serve as the technical basis for the future of reliability and risk approaches. The driving factors for this new approach includes: temporal (timing issues), spatial (location issues), mechanistic (physics issues), and topological (complexity issues). By combining phenomenology directly with stochastic quantification, we are able to perform advanced uncertainty analysis directly on both parameters and models. In addition, new tools allow for the creation of reduced order models that mimic high-fidelity engineering models while still permitting the realization of thousands of simulation iterations in short run-times on current workstation-class computers. While these advanced methods and tools can provide increased realism in our engineering safety and risk approaches, their greater benefit is to provide a risk-informed engineering framework for design and operation. This talk will briefly review the history of risk assessment and risk-informed applications, describe some of the current research and development found at the Idaho National Laboratory, and discuss potential future applications and approaches for advanced methods and tools.

97 MATHEMATICS AND COMPUTING↗

Simurgh: A Framework for Cad-Driven Deep Learning Based X-Ray CT Reconstruction

High-resolution X-ray computed tomography (XCT) is an important technique for the inspection of additively manufactured (AM) parts. While XCT is typically used off-line to inspect a subset of manufactured parts, significantly accelerating measurement speed while retaining accuracy would enable use of XCT for in-line inspection to rapidly identify defects in each part as it is manufactured. Here, we propose a deep learning (DL) based approach that uses computer aided design (CAD) models of the AM parts and physics-based information to rapidly produce high-quality reconstructions from sparse XCT measurements without high quality ground truth data. Our approach uses a generative adversarial neural network (GAN) to produced realistic training data from the CAD-based simulations and a deep neural network that is trained using data from the first stage to produce accurate 3D reconstructions. Using experimental XCT data of metal parts, we demonstrate enhanced defect detection capabilities while dramatically reducing the scan time.

Ziabari, Amir↗

DEMONSTRATION OF A DATA-DRIVEN PHYSICS-BASED APPROACH FOR COMPUTATIONALLY EFFICIENT CFD PREDICTION OF TWO-PHASE BUBBLY FLOW

To realize efficient computational fluid dynamics (CFD) prediction of two-phase flow, a multi-scale physics-guided data-driven approach, Feature Similarity Measurement (FSM) technique was developed for error estimation in two-phase flow simulation using coarse-mesh CFD, to achieve a comparable accuracy as fine-mesh simulations with fast-running feature. By defining physics-guided parameters and variable gradients as physical features, FSM has the capability to capture the underlying local patterns in the coarse-mesh CFD simulation. Massive low-fidelity data and respective high-fidelity data are used to explore the underlying information relevant to the main simulation errors and the effects of phenomenological scaling. By learning from previous simulation data, a surrogate model using deep feedforward neural network (DFNN) can be developed and trained to estimate the simulation error of coarse-mesh CFD. In a demonstration case of two-phase bubbly flow, the DFNN model well captured and corrected the unphysical “peaks” in the velocity and void fraction profiles near the wall in the coarse-mesh configuration, even for extrapolative predictions. The research documented supports the feasibility of the physics-guided deep learning methods for coarse mesh CFD simulations which has a potential for the efficient industrial design.

42 ENGINEERING↗

Group Structure Optimization Using the PyGroup Code

Group constants are used as coefficients in the energy domain discretized Linear Boltzmann Equation (LBE), in which the Continuous-Energy (CE) fluxes are also integrated over energy bins and transformed into group fluxes. This procedure, based on the Multi-Group (MG) theory, is widely applied in radiation transport and diffusion simulations Careful preparation of MG cross section data, or group constants, is thus a prerequisite for success in acquiring high-fidelity LBE solutions. Generation of MG cross section data itself is a complicated multi-step task, since group constants are indispensably coupled with the solution of the LBE, i.e. the flux spectrum in the problem to be solved. Theoretically, a MG cross section library is only valid for the problem that it is built upon. Nevertheless, in practice MG cross section libraries are prepared for similar type of problems, such as thermal reactors, fast reactors, etc., to facilitate the neutronic simulations. Generally, CE point-wise cross section data set, generated by both theoretical models and experimental data, is first converted to Fine-Group (FG) libraries using a pre-defined or precalculated flux spectra, then, Broad-Group (BG) libraries are collapsed from the fine-group libraries using problem-dependent flux spectra. Proper treatments of energy and spatial self-shielding and resonance cross section are required in the CE-to-MG converting process Previously, we have developed a computer code, YGROUP, for fine-to-broad group cross section collapsing. To maintain high accuracy of the resulting BG libraries, YGROUP provides users several options on collapsing strategies and weighting schemes for group binning and evaluation of BG cross section data in users' specific problems.. An approach based on Particle Swarm Optimization was applied to automatically determine the optimal group structure. We also developed a hybrid weighting scheme to further improve BG library performance for Special Nuclear Materials (SNM) detection applications. In this work, a new computer code, PyGroup, has been developed to implement the PSO approach with an improved simulation-driven fitness function design that includes both response and forward/adjoint flux errors. Together with YGROUP, and deterministic transport code TITAN/PENMSHXP, the PyGroup code package is tested on a Dual Range Coincident Counter (DRCC) model. In this paper, we demonstrated the capability of the PyGroup code for group structure optimization using the DRCC detector model. In the future, an iteration scheme can be implemented to determine the number of broad groups based on a user-defined tolerance. we will also apply PyGroup to more complicated reactor models, and investigate physics-based fitness function designs to further reduce computational cost. (authors)

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Kriging-Based Inverse Uncertainty Quantification of BISON Fission Gas Release Model

Nuclear reactor fuel performance analysis studies the thermo-mechanical behavior of fuel rods and verify their compliance with safety criteria under both normal operation and accidental conditions. One of the major concerns in fuel performance analysis is the behavior of the fission gases xenon and krypton in uranium dioxide fuel, which significantly affect the thermo-mechanical performance of the nuclear fuel rods employed in current LWRs. Previously, the modeling of fission gas release (FGR) in nuclear fuel performance codes have been relying on empirical models which are not applicable beyond their range of calibration. Recently a more efficient and flexible physics-based FGR and swelling model that can describe a wider range of reactor operation conditions have been implemented in fuel performance code BISON. In this work, we performed Bayesian calibration to inversely quantify the uncertainties associated with the 5 tuning parameters in BISON FGR model, based on the time-dependent FGR measurement data from Riso-AN3 benchmark. It takes around 1 hour and 40 minutes for BISON to simulate the Riso-AN3 benchmark with a moderate mesh size using 32 processors. As hundreds of thousands of BISON simulations are needed during Markov Chain Monte Carlo (MCMC) sampling, running BISON directly is not a viable solution. Therefore we propose to use Kriging-based metamodels to replace BISON during MCMC sampling. Surrogate model is an approximation of the input/output relation of a computer code/model. It is also called metamodel, response surface, emulator. Surrogate models usually take much less computational time than the original model (like BISON) while maintaining the input/output relation of the original model to a desirable accuracy. Constructing the metamodels normally requires small number of executions of the original model. Once validated, metamodels can be used to perform uncertainty and sensitivity analysis, Bayesian calibration, optimization, etc.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Fuel Performance Evaluation of THOR-C Experiments

The Temperature Heatsink Overpower Response Commissioning (THOR-C) and THOR-Metal (THOR-M) experiments will be performed as part of an ongoing project for testing sodium fast reactor fuels with the Japan Atomic Energy Agency (JAEA). The THOR-C experiments consist of fresh metallic fuel pins and have been analyzed using the ABAQUS, Ansys codes and the BISON fuel performance code. THOR-M-Loss of Flow-1 (THOR-M-LOF-1) is designed to test an EBR-II irradiated fuel pin under LOF conditions. Simulation of the THOR-MLOF-1 experiment required first simulating the base irradiation of the fuel pin in EBR-II. MFUEL module of SAS4A/SASSYS-1 [1] is a physics-based metallic fuel performance model applicable to the normal operation, transient scenarios and fuel failure modeling including scenarios with bulk fuel melting. The model has been validated using EBR-II normal operation, separate effect transient tests as well as TREAT M-Series transient tests [2]. In this study, MFUEL models has been utilized together with a new capsule heat transfer model developed in this project. The new heat transfer model was necessary due to (1) significant amount of heat losses that required 2D heat transfer, (2) the presence of a titanium heat sink, rejecting a significant amount of heat, and (3) stagnant coolant conditions, which are inconsistent with SAS4A/SASSYS-1 (SAS) heat transfer model. Updates to SAS4A/SASSYS-1 and MFUEL has been described below, followed by a preliminary validation effort using the results from THOR-C-2 fresh fuel capsule experiment. A previous study for THOR-C-2 analysis using BISON code is also utilized in this study to model this test [3]. [1] D. O’Grady, A. J. Brunett, L. Ibarra, A. Karahan, T. Kim, T. S. Sumner, R. Thomas, T. H. Fanning, “The SAS4A/SASSYS-2 Version 5.7 Safety Analysis Code System,” Argonne National Laboratory,ANL/NSE-SAS/5.7, (2023). [2] A. Karahan, T. Kim, T. Fanning, D. O’Grady, “Validation of MFUEL Metal Fuel Performance Models of SAS4A/SASSYS-1,” Argonne National Laboratory, ANL/NSE-23/11, (2023). [3] M. Mihelish, A. Zabriskie, K. Paaren, P. Medvedev, C. Jensen, “Fuel Performance Predictions for the TREAT THOR-C Experiments,” Idaho National Laboratory, INL/RPT-23-73397, Revision 0, (2023)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Thermo-Fluid Modeling Framework for Supercomputer Digital Twins: Part 1, Demonstration at Exascale

A thermo-fluid modeling framework is being developed for ExaDigiT---an open-source framework for developing comprehensive digital twins of liquid-cooled supercomputers. The work is being conducted in two parts, and discussion is divided into two companion papers. The work documented in this paper focuses on the development of a cooling system library in Dymola for the Frontier supercomputer at Oak Ridge National Laboratory. The second part, outlined in a companion paper, focuses on a templating structure called Auto-CSM for easily creating model-agnostic, physics-based thermo-fluid cooling system models for liquid-cooled supercomputers using a text-based schema. The cooling model is being developed using primarily the open-source Transient Simulation Framework of Reconfigurable Models (TRANSFORM) library. The library follows the templating architecture developed within the TRANSFORM library for modeling subsystems. A full-system validation was performed to validate a very simple model that is integrated with the system controls, and the results are presented herein.

Kumar, Vineet↗

Computational Modeling of Molten Salt Infiltration and Oxidation in Nuclear Graphite

Graphite is utilized as a moderator and reflector in advanced nuclear reactor designs due to its high thermal conductivity, neutron moderation properties, and resistance to radiation damage. However, its longterm performance and reliability are challenged by degradation mechanisms such as molten salt infiltration in molten salt reactors (MSRs) and oxidation in gas-cooled reactors (GCRs). These mechanisms can compromise the structural integrity and operational lifetime of graphite components, necessitating a more detailed assessment of their physical behavior. This report focuses on the development of computational models for molten salt infiltration and oxidation of graphite to aid the design and performance analysis of graphite components. For molten salt infiltration, a computational framework is developed that couples incompressible Navier-Stokes and phase-field model to simulate the penetration of molten salt into graphite?s interconnected pore structure. Initial model verification is performed using two-phase flows in two dimensions, demonstrating the models ability to capture fundamental physical behavior and agree with analytical solution. This framework is then applied to a realistic IG110 nuclear graphite , where a computed tomography extracted pore geometry is used to analyse the infiltration behavior of FLiNaK molten salt. This model provides insights into how the microstructure and other relevant parameters influence the transport pathways of molten salt into graphite, potentially offering a means to rapidly evaluate a graphite grade?s resistance to infiltration. For oxidation, the report details pore-scale mass and heat transport models, describing the diffusion of gases, reaction kinetics, and thermal effects. Additionally, this report highlights inconsistencies in the existing volume-averaged macroscopic model, particularly in upscaling of reaction kinetics and flux terms, and surface to volume transformations. These inconsistencies suggest that current formulations may not accurately capture the experimentally observed graphite oxidation process, highlighting the need for improved model development. This work advances the development of physics-based computational models for graphite degradation, contributing to improved predictive models for next-generation nuclear reactor designs. Future efforts will focus on refining the infiltration model to address non-physical behaviors and enhance its robustness. Additionally, for oxidation, further studies will employ the principles of volume averaging to rigorously derive the upscaled equations, potentially in collaboration with subject matter experts.

Computational Modeling of Molten Salt Infiltration↗

A Physics-Informed Reinforcement Learning Framework for Economic-Thermal Co-Optimization of Crypto Mining Data Centers: Preprint

The rapid expansion of cryptocurrency mining has created a new class of high-density data centers characterized by extreme thermal flux and high sensitivity to volatile economic markets. Traditional thermal management strategies, typically reliant on rule-based control, maintain static setpoints that fail to account for fluctuating electricity prices and cryptocurrency values - factors critical to mining profitability. To address this, we present a physics-informed reinforcement learning (PIRL) framework for economic-thermal co-optimization in crypto mining data centers. This framework consists of a proximal policy optimization (PPO) agent, a virtual testbed powered by high-fidelity physics-based models, and an interactive frontend dashboard. The PPO agent is trained using the virtual testbed and strict hardware safety limits. This physics-informed approach allows the agent to learn a stochastic policy that dynamically balances mining revenue against operational costs by co-optimizing HVAC cooling setpoints and IT computational hashrate. The simulation results demonstrate that the integrated framework achieved an 8.62% increase in net operational profit compared to traditional baseline strategies while strictly adhering to safety-critical temperature constraints (coolant supply temperature < 32 degrees C). This work provides a scalable template for the deployment of reinforcement learning in mission critical facilities where economic volatility and physical safety must be managed simultaneously.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗