Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Critical heat flux”

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.

At least 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Physics-based hybrid machine learning for critical heat flux prediction with uncertainty quantification

Critical heat flux (CHF) is a key quantity in nuclear system modeling due to its impact on heat transfer, safety margins, and reactor performance. This study develops and validates an uncertainty-aware hybrid modeling approach that combines machine learning with physics-based models to predict CHF in cases of dryout. The Biasi and Bowring empirical correlations were paired with three ML uncertainty quantification (UQ) techniques: deep neural network (DNN) ensembles, Bayesian neural networks (BNNs), and deep Gaussian processes (DGPs). A pure ML model without a base model was evaluated for comparison. Model performance was assessed under plentiful (7,350 points) and limited (9 points) training data scenarios using parity, uncertainty distributions, and calibration curves. Results show that the Biasi hybrid DNN ensemble achieved the best overall performance, with a mean absolute relative error of 1.846%, and well-calibrated uncertainty estimates. The BNN-based hybrids showed slightly higher error (2.14%) but superior uncertainty calibration. DGP models underperformed, with over 6% error and poor uncertainty calibration. All hybrid models outperformed pure machine learning configurations, demonstrating resistance against data scarcity. These findings indicate that hybrid modeling significantly improves predictive accuracy, interpretability, and resilience to data scarcity. The integration of uncertainty awareness provides actionable confidence in CHF predictions, which is vital for safety-critical decisions in nuclear applications. This hybrid approach offers a viable pathway for deploying ML models in reactor analysis tools while preserving domain knowledge and physical consistency.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2025

Aided Active Learning (AAL) for Enhanced Critical Heat Flux Prediction

Accurate prediction of critical heat flux (CHF) is crucial for the safe and efficient operation of nuclear reactors. Traditional CHF modeling methods often require extensive experimental data, which are hard to obtain. This study introduces the Aided Active Learning (AAL) framework, which strategically minimizes data requirements without sacrificing model accuracy. Unlike conventional Active Learning (AL), AAL introduces an additional step of randomly selecting a subset from the sample pool before applying the query strategy. To evaluate the performance of AAL, two query strategies—uncertainty-based sampling and error-reduction sampling—were evaluated across the following models: random forest (RF), feedforward neural network (FNN), and variational feedforward neural network (vFNN). The proposed framework demonstrated that AAL effectively reduces the number of training samples needed to achieve comparable predictive accuracy. For the RF model, AL required only 710 samples to achieve an R2 score of 0.98, as compared to the 4,785 samples needed by random sampling. Similarly, the FNN model achieved the same R2 score with just 355 samples when using AL, a significant improvement over the 825 samples required by random sampling. In case of uncertainty-based sampling strategy, vFNN attained an R2 of 0.98 with 3,420 samples, reducing the sample requirement by 47% relative to the 6,440 samples needed for random sampling. Its performance suggests that larger training data are required to fully leverage its uncertainty quantification capabilities.

22 - GENERAL STUDIES OF NUCLEAR REACTORS↗
NASA NTRS1972

Combined buoyancy and flow direction effects on saturated boiling critical heat flux in liquid nitrogen

Buoyancy effects on the critical heat flux and general data trends for a liquid nitrogen internal flow system were determined by comparison of upflow and downflow data under identical test conditions. The test section had a 1.28 cm diameter flow passage and a 30.5 cm heated length which was subjected to uniform heat fluxes through resistance heating. Test conditions covered a range of pressures from 3.4 to 10.2 atm, inlet velocities from 0.23 to 3.51 m/sec, with the liquid nitrogen temperature at saturated inlet conditions. Data comparisons showed that the critical heat flux for downflow could be up to 36 percent lower than for upflow. A nonmonotonic relationship between the critical heat flux and velocity was determined for upflow but not for downflow. A limiting inlet velocity of 4.12 m/sec was determined to be the minimum velocity required to completely suppress the influence of buoyancy on the critical heat flux for this saturated inlet flow system. A correlation of this limiting fluid velocity is presented that was developed from previously published subcooled liquid nitrogen data and the saturated data of this investigation.

Papell, S. S.↗
NASA NTRS1993

The effects of buoyancy on the critical heat flux in forced convection

The critical heat flux (CHF) in forced convection over a flat surface at relatively low flow velocities has been found, not unexpectedly, to depend upon the orientation of the buoyancy. The CHF for R-113 was measured at various heating surface orientations for test section Reynolds numbers ranging between 3000 and 6500. In this flow range, the buoyancy force acting on the vapor generally dominates over the flow inertia, yet the inertia would still be substantial were gravity to be reduced. In the experiments of this study, the CHF is determined for heating surface orientations ranging from 0 deg to 360 deg, for flow velocities between 4 cm/s and 35 cm/s, and for subcoolings between 2.8 C and 22.2 C. The results presented here demonstrate the strong influence of buoyancy at low flow velocities, which diminishes as the flow velocity and subcooling are increased.

Brusstar, Matthew J.↗
NASA NTRS1998

Forced Convection Boiling and Critical Heat Flux of Ethanol in Electrically Heated Tube Tests

Electrically heated tube tests were conducted to characterize the critical heat flux (transition from nucleate to film boiling) of subcritical ethanol flowing at conditions relevant to the design of a regeneratively cooled rocket engine thrust chamber. The coolant was SDA-3C alcohol (95% ethyl alcohol, 5% isopropyl alcohol by weight), and tests were conducted over the following ranges of conditions: pressure from 144 to 703 psia, flow velocities from 9.7 to 77 ft/s, coolant subcooling from 33 to 362 F, and critical heat fluxes up to 8.7 BTU/in(exp 2)/sec. For the data taken near 200 psia, critical heat flux was correlated as a function of the product of velocity and fluid subcooling to within +/- 20%. For data taken at higher pressures, an additional pressure term is needed to correlate the critical heat flux. It was also shown that at the higher test pressures and/or flow rates, exceeding the critical heat flux did not result in wall burnout. This result may significantly increase the engine heat flux design envelope for higher pressure conditions.

Meyer, Michael L.↗
DOE OSTI2026

Deployment of Traditional and Hybrid Machine Learning for Critical Heat Flux Prediction in the CTF Thermal-Hydraulics Code

Critical heat flux (CHF) marks the transition from nucleate to film boiling, where heat transfer to the working fluid can rapidly deteriorate. Accurate CHF prediction is essential for efficiency, safety, and preventing equipment damage, particularly in nuclear reactors. Although widely used, empirical correlations frequently exhibit discrepancies when compared to experimental data, limiting their reliability in diverse operational conditions. Traditional machine learning (ML) approaches have demonstrated potential for CHF prediction but often suffer from limited interpretability, data scarcity, and insufficient knowledge of physical principles. Hybrid model approaches, which combine data-driven ML with base models, mitigate these concerns by incorporating prior knowledge of the domain. This study integrates an externally trained purely data-driven ML model and two hybrid models (using the Biasi and Bowring CHF correlations) within the CTF subchannel code via a custom Fortran framework. Performance was evaluated using two validation cases: a subset of the Nuclear Regulatory Commission (NRC) CHF database and the Bennett dryout experiments. In both cases, the hybrid models demonstrated significantly lower error metrics compared to conventional empirical correlations, with the best models often reducing relative error by about 5 percentage points. The pure ML model achieved comparable accuracy, outperforming the hybrid Biasi model in the NRC test case (3.3% versus 5.5% relative error) but exhibiting slightly higher error against the hybrid Bowring model in the Bennett test case (7.7% versus 6.1%). Trend analysis of error parity indicated that ML-based models reduced the tendency for CHF overprediction, improving overall accuracy. These results demonstrate that ML-based CHF models can be effectively integrated into subchannel codes and could potentially increase performance compared to conventional methods.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2026

Development of physics-consistent conditional diffusion model to overcome data scarcity in critical heat flux

Deep generative modeling provides a powerful pathway to overcome data scarcity in energy-related applications where experimental data are often limited. By learning the underlying probability distribution of the training dataset, deep generative models, such as the diffusion model, can generate high-fidelity synthetic samples that statistically resemble the training data. Such synthetic data generation can significantly enrich the size and diversity of the available training data, and more importantly, improve the robustness of downstream machine learning models in predictive tasks. The objective of this paper is to investigate the effectiveness of diffusion models for overcoming data scarcity in nuclear energy applications. By leveraging a public dataset on critical heat flux which covers a wide range of commercial nuclear reactor operational conditions, we developed a diffusion model that can generate an arbitrary amount of synthetic samples. Since a vanilla diffusion model can only generate samples randomly, we also developed a conditional diffusion model capable of generating targeted critical heat flux data under user-specified thermal-hydraulic conditions. The performance of the diffusion model was evaluated based on its ability to capture empirical feature distributions and pair-wise correlations, as well as to maintain physical consistency. The results showed that both the diffusion model and conditional diffusion model can successfully generate realistic and physics-consistent critical heat flux data. Furthermore, uncertainty quantification results demonstrate that the conditional diffusion model is highly effective in augmenting critical heat flux data while maintaining acceptable levels of uncertainty.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2026

Predicting critical heat flux using localized sensing at invisible vapor-liquid interfaces

Predicting critical heat flux (CHF) in two-phase electronics cooling systems remains a significant challenge due to the sudden onset of boiling crisis and the difficulty in directly visualizing vapor-liquid interfaces. Existing sensing methods rely on lagging temperature measurements, optically accessible systems, or spatially averaged signals that cannot pinpoint CHF initiation at localized high-heat-flux regions. Here, we report a planar capacitive sensing approach that enables real-time, localized detection of vapor-liquid interface dynamics for CHF prediction in boiling heat transfer. The capacitive sensor exploits the dielectric constant difference between liquid and vapor phases to capture bubble nucleation, growth, and departure dynamics with a temporal resolution down to 2 ms. The capacitive sensing reveals distinct signals across boiling regimes: from high-frequency fluctuations during strong nucleate boiling to low-frequency fluctuations with increased amplitudes when approaching CHF. The multi-sensor array experiments demonstrate real-time localized sensing, where each sensor responds exclusively to boiling in its immediate vicinity without crosstalk from neighboring regions. This non-intrusive sensing approach provides predictive rather than lagging sensing signals of CHF occurrence, offering predictive diagnosis of two-phase liquid cooling for the thermal management of high-power-density electronics.

CHF↗
NASA NTRS2002

Investigation of Body Force Effects on Flow Boiling Critical Heat Flux

The bubble coalescence and interfacial instabilities that are important to modeling critical heat flux (CHF) in reduced-gravity systems can be sensitive to even minute body forces. Understanding these complex phenomena is vital to the design and safe implementation of two-phase thermal management loops proposed for space and planetary-based thermal systems. While reduced gravity conditions cannot be accurately simulated in 1g ground-based experiments, such experiments can help isolate the effects of the various forces (body force, surface tension force and inertia) which influence flow boiling CHF. In this project, the effects of the component of body force perpendicular to a heated wall were examined by conducting 1g flow boiling experiments at different orientations. FC-72 liquid was boiled along one wall of a transparent rectangular flow channel that permitted photographic study of the vapor-liquid interface at conditions approaching CHF. High-speed video imaging was employed to capture dominant CHF mechanisms. Six different CHF regimes were identified: Wavy Vapor Layer, Pool Boiling, Stratification, Vapor Counterflow, Vapor Stagnation, and Separated Concurrent Vapor Flow. CHF showed great sensitivity to orientation for flow velocities below 0.2 m/s, where very small CHF values where measured, especially with downflow and downward-facing heated wall orientations. High flow velocities dampened the effects of orientation considerably. Figure I shows representative images for the different CHF regimes. The Wavy Vapor Layer regime was dominant for all high velocities and most orientations, while all other regimes were encountered at low velocities, in the downflow and/or downward-facing heated wall orientations. The Interfacial Lift-off model was modified to predict the effects of orientation on CHF for the dominant Wavy Vapor Layer regime. The photographic study captured a fairly continuous wavy vapor layer travelling along the heated wall while permitting liquid contact only in wetting fronts, located in the troughs of the interfacial waves. CHF commenced when wetting fronts near the outlet were lifted off the wall. The Interfacial Lift-off model is shown to be an effective tool for predicting the effects of body force on CHF at high velocities.

Zhang, Hui↗
NASA NTRS1996

Study of Critical Heat Flux and Two-Phase Pressure Drop Under Reduced Gravity

The design of the two-phase flow systems which are anticipated to be utilized in future spacecraft thermal management systems requires a knowledge of two-phase flow and heat transfer phenomena in reduced gravities. This program was funded by NASA headquarters in response to NRA-91-OSSA-17 and was managed by Lewis Research Center. The main objective of this program was to design and construct a two-phase test loop, and perform a series of normal gravity and aircraft trajectory experiments to study the effect of gravity on the Critical Heat Flux (CHF) and onset of instability. The test loop was packaged on two aircraft racks and was also instrumented to generate data for two-phase pressure drop. The normal gravity tests were performed with vertical up and downflow configurations to bound the effect of gravity on the test parameters. One set of aircraft trajectory tests was performed aboard the NASA DC-9 aircraft. These tests were mainly intended to evaluate the test loop and its operational performance under actual reduced gravity conditions, and to produce preliminary data for the test parameters. The test results were used to demonstrate the applicability of the normal gravity models for prediction of the two-phase friction pressure drop. It was shown that the two-phase friction multipliers for vertical upflow and reduced gravity conditions can be successfully predicted by the appropriate normal gravity models. Limited critical heat flux data showed that the measured CHF under reduced gravities are of the same order of magnitude as the test results with vertical upflow configuration. A simplified correlation was only successful in predicting the measured CHF for low flow rates. Instability tests with vertical upflow showed that flow becomes unstable and critical heat flux occurs at smaller powers when a parallel flow path exists. However, downflow tests and a single reduced gravity instability experiment indicated that the system actually became more stable with a parallel single-phase flow path. Several design modifications have been identified which will improve the system performance for generating reduced gravity data. The modified test loop can provide two-phase flow data for a range of operating conditions and can serve as a test bed for component evaluation.

Abdollahian, Davood↗
DOE OSTI2024

Data Efficiency Assessment of Generative Adversarial Networks for Critical Heat Flux Synthetic Data Generation

This study investigates the application of generative artificial intelligence techniques, particularly conditional generative adversarial networks (cGAN), in real-world engineering contexts, with a specific focus on synthetic data generation for critical heat flux (CHF). Utilizing a dataset comprising more than 20,000 real experimental CHF measurements, we conduct a series of experiments to examine cGAN’s behavior. These experiments encompass varying sizes of the training dataset, training cGAN on data from diverse experimental sources to generate new data on unseen experimental setups, and assessing the impact of excluding various input features on cGAN’s data generation accuracy. Our findings underscore the pronounced data dependency of cGAN for reliable performance, with decreased efficacy observed with smaller training dataset sizes. Notably, cGAN exhibits varying performance when trained on data from different experiments, with superior predictive capabilities observed for certain experiment sources compared to others. For instance, when cGAN was trained on data from Smolin et al.’s experiments or Zenkevich et al., it exhibited relatively good performance in generating the data from Becker et al., Kirillov et al., and Alekseev et al. experiments. In contrast, when trained with Alekseev et al.’s data and tasked with generating other experimental setups, cGAN showed notably poor performance. In both scenarios, cGAN’s performance was inferior compared to training on samples from all experiments concurrently. A feature importance analysis highlights the significant influence of parameters such as mass flux and heated length on accurate CHF generation, while other parameters like diameter and pressure have less impact. Inlet temperature is identified as a moderating factor by cGAN.

22 - GENERAL STUDIES OF NUCLEAR REACTORS↗
NASA NTRS2003

Investigation of Critical Heat Flux in Reduced Gravity Using Photomicrographic Techniques

Experiments were performed to examine the effects of body force on flow boiling critical heat flux (CHF). FC-72 was boiled along one wall of a transparent rectangular flow channel that permitted photographic study of the vapor-liquid interface just prior to CHF. High-speed video imaging techniques were used to identify dominant CHF mechanisms corresponding to different flow orientations and liquid velocities. Six different CHF regimes were identified: Wavy Vapor Layer, Pool Boiling, Stratification, Vapor Counterflow, Vapor Stagnation, and Separated Concurrent Vapor Flow. CHF showed significant sensitivity to orientation for flow velocities below 0.2 m/s, where extremely low CHF values where measured, especially with downward-facing heated wall and downflow orientations. High flow velocities dampened the effects of orientation considerably. The CHF data were used to assess the suitability of previous CHF models and correlations. It is shown the Interfacial Lift-off Model is very effective at predicting CHF for high velocities at all orientations. The flooding limit, on the other hand, is useful at estimating CHF at low velocities and for downflow orientations. A new method consisting of three dimensionless criteria is developed for determining the minimum flow velocity required to overcome body force effects on near-saturated flow boiling CHF. Vertical upflow boiling experiments were performed in pursuit of identifying the trigger mechanism for subcooled flow boiling CHF. While virtually all prior studies on flow boiling CHF concern the prediction or measurement of conditions that lead to CHF, this study was focused on events that take place during the CHF transient. High-speed video imaging and photomicrographic techniques were used to record the transient behavior of interfacial features from the last steady-state power level before CHF until the moment of power cut-off following CHF. The video records show the development of a wavy vapor layer which propagates along the heated wall, permitting cooling prior to CHF only in wetting fronts corresponding to the wave troughs. Image analysis software was developed to estimate void fraction from the individual video images. The void fraction records for subcooled flow boiling show the CHF transient is accompanied by gradual lift-off of wetting fronts culminating in some maximum vapor layer mean thickness, following which the vapor layer begins to thin down as the transition to film boiling ensues. This study proves the Interfacial Lift-off Model, which has been validated for near-saturated flow boiling CHF, is equally valid for subcooled conditions.

Mudawar, Issam↗
DOE OSTI2025

Predicting critical heat flux with uncertainty quantification and domain generalization using conditional variational autoencoders and deep neural networks

Deep generative models (DGMs) can generate synthetic data samples that closely resemble the original dataset, addressing data scarcity. In this work, we developed a conditional variational autoencoder (CVAE) to augment critical heat flux (CHF) data used for the 2006 Groeneveld lookup table. To compare with traditional methods, a fine-tuned deep neural network (DNN) regression model was evaluated on the same dataset. Both models achieved small mean absolute relative errors, with the CVAE showing more favorable results. Uncertainty quantification (UQ) was performed using repeated CVAE sampling and DNN ensembling. The DNN ensemble improved performance over the baseline, while the CVAE maintained consistent results with less variability and higher confidence. Both models achieved small errors inside and outside the training domain, with slightly larger errors outside. Altogether, the CVAE performed better than the DNN in predicting CHF and exhibited better uncertainty behavior.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
NASA NTRS2005

Flow-Boiling Critical Heat Flux Experiments Performed in Reduced Gravity

Poor understanding of flow boiling in microgravity has recently emerged as a key obstacle to the development of many types of power generation and advanced life support systems intended for space exploration. The critical heat flux (CHF) is perhaps the most important thermal design parameter for boiling systems involving both heatflux-controlled devices and intense heat removal. Exceeding the CHF limit can lead to permanent damage, including physical burnout of the heat-dissipating device. The importance of the CHF limit creates an urgent need to develop predictive design tools to ensure both the safe and reliable operation of a two-phase thermal management system under the reduced-gravity (like that on the Moon and Mars) and microgravity environments of space. At present, very limited information is available on flow-boiling heat transfer and the CHF under these conditions.

Hasan, Mohammad M.↗
DOE OSTI2025

A model to assess Zircaloy’s mechanical property changes following a transient beyond critical heat flux

Maintaining the integrity of nuclear fuel rods is essential for ensuring public health and safety in nuclear power generation. During reactor operation, this integrity is confirmed by demonstrating compliance with established regulatory acceptance criteria. For moderate-frequency events, such as limiting transients and anticipated operational occurrences (AOOs), the current fuel integrity criterion is based on preventing boiling transition. This criterion assumes that prevention of boiling transition will prevent excessive cladding heating and, thus, fuel failure during normal operations. While conservative, this approach places significant constraints on core design, fuel cycle economics, and a plant’s ability to perform major power uprates, leading to suboptimal fuel utilization and inefficient carbon-free energy production. A more efficient approach could be achieved by revising the failure criterion to a material-specific limit rather than strictly preventing the boiling transition, since boiling transition per se is not a cause of fuel cladding failure. Here, as a result, a new licensing framework based on material properties, termed time-at-temperature (t@T), is needed. This approach would allow for brief periods of post–critical heat flux operation during an AOO without compromising safety. Implementing the t@T licensing strategy requires a robust technical foundation in material properties, which must be established through comprehensive data collection on both unirradiated and irradiated fuel and cladding materials. This foundation would enable the development of a safety basis that ensures safe operation while providing greater flexibility and efficiency for reactor operation. This paper documents a thorough review of the available data to establish a baseline knowledge that can inform the development of cladding mechanical models, as well as identify experimental data gaps that need to be addressed in future research. Machine learning and data informatics were utilized to extract the importance of parameters on the t@T parameter. Industry tools were used to perform baseline analyses to define the relevant transient conditions for data analysis. The subsequent review successfully identified applicable experimental data, as well as sufficient data to evaluate changes in cladding mechanical properties following an AOO transient. Rather than developing new models, this work coupled existing irradiation annealing and recrystallization models to calculate changes in hardness, yield stress, and ultimate tensile stress following an AOO event. The findings from this review were summarized to highlight the experimental data needs required to fill remaining gaps and support the development of future t@T licensing methodologies.

Cladding performance↗