Estimation Of Thermal Conductivity Research Articles

New approach to reduce analysis time of thermal response test: Active thermal response test using linear heat optical fiber sensor

This study introduces an innovative approach to reduce the analysis time of thermal response tests (TRTs) by employing an active TRT using a linear heat optical fiber sensor. Conventional TRTs require prolonged heating periods, making them time-consuming and resource-intensive. This research focuses on comparing conventional and active TRTs using computational fluid dynamics (CFD) simulations to validate the accuracy and efficiency of the proposed method. The active TRT method demonstrated the potential to shorten heating time while maintaining accurate thermal conductivity measurements. The findings reveal that the difference in thermal conductivity estimation between 15 and 48 h is only 0.03 W/(m∙K), indicating stable results with shorter testing times compared to the conventional TRT. Although stable results were obtained, uncertainties persist due to field conditions and measurement errors. To address this, an approximate equation was proposed for estimating thermal conductivity using the Relaxation Time of Temperature (RTT) method, which has been validated in previous studies through ground recovery temperatures. Future work will involve conducting field Active TRT to verify the accuracy of this method in reducing TRT duration.

Impact of helium nanobubbles on tungsten fuzz nanostructure conductivity

Explosive electron emission splashes occurring in plasma interactions with nanostructured surfaces can be induced by Joule energy release under reduced conductivity. Distribution of helium nanobubbles inside tungsten at an elevated energy of implanted helium ions has been studied via numerical molecular dynamic modeling, and its impact on the nanostructure conductivity has been estimated. Average concentration and size of nanobubbles were calculated to be about n = 1020 cm−3 and r = 3 Å, respectively, at the helium fluence of about 2 × 1015 cm−2. The distribution of nanobubbles becomes more flat and extends deeper into the bulk material with increasing impact energy. At energies below about 300 eV, most nanobubbles form within a depth of 10 nm, corresponding to the characteristic size of the nanowires. At higher energies, a significant portion of the nanobubbles forms deeper in the material. An approach for estimating the electron scattering frequency in tungsten by helium nanobubbles has been proposed, taking into account the volume porosity of the W-fuzz nanostructure and scaling the helium content to the experimentally measured one. The resulting value for the tungsten fuzz nanostructure was 2.27 × 1016 s−1 and the resistivity was 1.92 mΩ cm that is 364 times larger than the normal tungsten resistivity, the conductivity of 521 S/cm was 0.275% from the normal tungsten one. The corresponding estimation of thermal conductivity via Wiedemann–Franz law gave 3.8 mW cm−1 K−1, which agrees well with recent measurements.

Estimation of Effective Thermal Conductivity of Spherical and Ellipsoidal Shaped Randomly Packed Mono-Sized, Binary-Sized, and Poly-Dispersed Ceramic Pebble Beds

Estimation of Effective Thermal Conductivity of Spherical and Ellipsoidal Shaped Randomly Packed Mono-Sized, Binary-Sized, and Poly-Dispersed Ceramic Pebble Beds

Estimation of heat source, specific heat, and thermal conductivity of insulation materials using modified conjugate gradient method

The use of insulation materials is crucial for minimizing thermal losses in various industries like automotive, building, solar, and aerospace. Plexiglass and Expanded Polystyrene (EPS) foam are popular choices due to their effective temperature regulation, lightweight nature, and cost-effectiveness. Accurate measurement of their thermal properties is vital for achieving conservation goals. While the parallel hot wire technique provides valuable insights, accurately determining heat source strength (HW) poses a challenge, impacting measurement accuracy. To address this, the modified Conjugate Gradient Method (CGM) is employed for simultaneous estimation of thermal conductivity (λ), specific heat (c) and HW of Plexiglass and EPS foam. The direct problem result has been validated with published experimental work that shows deviations less than 2 %. The study investigates the impact of sensor positioning relative to the hot-wire (rm) and introduces artificial Gaussian error with standard deviations (θ) of 0.01 °C, 0.02 °C and 0.1 °C. At θ = 0.1 °C the percentage relative errors in the estimation of λ, c and HW with rm = 5 mm for Plexiglass are 5 %, 1.442 %, and 4.651 % respectively. For EPS foam the corresponding errors are 2.44 %, 1.26 %, and 1.74 %. The modified CGM emerges as a reliable algorithm for the measurement of thermal properties of insulation materials.

Experimental and numerical estimation of thermal conductivity of bio-based building composite materials with an enhanced thermal capacity

The paper tackles the important problem of estimating and predicting the thermal conductivity of bio-based building materials that can help decarbonize the building sector. Knowledge and the possibility to optimize their insulating properties is the first criterion for determining the ability of their use in the building sector. The novel yet not well-studied hemp shives and magnesium binder composites with improved thermal mass by microencapsulated phase change material (PCM) were considered. The samples of composites without PCM having different densities as well as with different amounts of microencapsulated PCM and the same density were manufactured and tested experimentally in different states, i.e., dry and after conditioning at relative humidity (RH) of 50, 75, and 90 %, and in the case of samples with PCM also at different average measurement temperatures selected with respect to the phase change range of the PCM. A novel method of predicting bio-based composites' effective thermal conductivity tensor was also developed. The method is based on the numerical solution of the heat conduction equation at the micro-scale considering real composite microstructure. The method was tuned using micro-computed tomography (μCT) data with the microstructure of composites without PCM and then applied to predict the thermal conductivities of composites with PCM, utilizing computational domains with an artificially distributed PCM in the binder. The experimental testing of composites without PCM revealed an apparent effect of increasing thermal conductivity with rising sample density and RH applied during conditioning. With an increase in density from 394 to 576 kg/m3, their thermal conductivities varied from 0.105 to 0.159 W/m/K for the dry state and from 0.157 to 0.241 W/m/K for RH 90 %. A similar effect of sample state and RH was observed for composites with microencapsulated PCM, which had the highest thermal conductivity, equal to 0.265 W/m/K, for the lowest PCM amount and RH 90 %. Moreover, a decrease in composites' effective thermal conductivity with an increase in PCM fraction was observed, except for the lowest PCM fraction, for which it increased. It also rose with increasing the average measurement temperature. The developed micro-scale-based numerical method allowed for predicting the thermal conductivities of composites with accuracies below 4.1 and 7.2 % for composites without and with PCM, respectively, except for the composite with the lowest amount of PCM, which behaved differently. Thus, its suitability for designing and optimizing bio-based composites was shown.

Sea ice mass balance during the MOSAiC drift experiment: Results from manual ice and snow thickness gauges

Precise measurements of Arctic sea ice mass balance are necessary to understand the rapidly changing sea ice cover and its representation in climate models. During the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) expedition, we made repeat point measurements of snow and ice thickness on primarily level first- and second-year ice (FYI, SYI) using ablation stakes and ice thickness gauges. This technique enabled us to distinguish surface and bottom (basal) melt and characterize the importance of oceanic versus atmospheric forcing. We also evaluated the time series of ice growth and melt in the context of other MOSAiC observations and historical mass balance observations from the Surface Heat Budget of the Arctic (SHEBA) campaign and the North Pole Environmental Observatory (NPEO). Despite similar freezing degree days, average ice growth at MOSAiC was greater on FYI (1.67 m) and SYI (1.23 m) than at SHEBA (1.45 m, 0.53 m), due in part to initially thinner ice and snow conditions on MOSAiC. Our estimates of effective snow thermal conductivity, which agree with SHEBA results and other MOSAiC observations, are unlikely to explain the difference. On MOSAiC, FYI grew more and faster than SYI, demonstrating a feedback loop that acts to increase ice production after multi-year ice loss. Surface melt on MOSAiC (mean of 0.50 m) was greater than at NPEO (0.18 m), with considerable spatial variability that correlated with surface albedo variability. Basal melt was relatively small (mean of 0.12 m), and higher than NPEO observations (0.07 m). Finally, we present observations showing that false bottoms reduced basal melt rates in some FYI cases, in agreement with other observations at MOSAiC. These detailed mass balance observations will allow further investigation into connections between the carefully observed surface energy budget, ocean heat fluxes, sea ice, and ecosystem at MOSAiC and during other campaigns.

Thermal performance appraisal of hybrid and nanofluid flow between a cone and a disk with variable thermal conductivity, viscous dissipation, and Joule heating

In many engineering systems, hybrid and nanofluids are influenced by their respective thermophysical characteristics. Lately, several models have been envisaged to foresee the hybrid and nanofluid attributes. The properties of hybrid nanofluids (HNFs) as potential heat transfer fluids are controlled by numerous aspects such as solid part size, volume fraction, and temperature. Considering the interesting facts of the hybrid and nanofluid flows, we in this exploration examined the thermal performance comparison of both types of fluid flows. The nanofluid and HNF are composed of molybdenum disulfide (MoS[Formula: see text]/kerosene oil and silicon dioxide–molybdenum disulfide (SiO2–MoS[Formula: see text]/kerosene oil, respectively. The flows are taken in a canonical gap between the cone and the disk. Both the cone and the disk may be rotating or stationary. The novelty of the computational model is enhanced by discussing the effects of viscous dissipation, Joule heating, and the variable thermal conductivity with convective condition. The Tawari and Das model is designed to analyze the heat transfer performance of the assumed fluid flows. The assumed fluid model is transmuted into the set of differential equations that are dealt numerically with the bvp4c MATLAB approach. The results are displayed in tables and graphical forms. For elevating estimations of the Eckert number, the heat transmission rate is found to be more significant in the disk than cone. It is also learned that hybrid nanoliquid heat transfer performance outperforms nanoliquid. The fluid velocity increases by raising the nanoparticle volume fraction [Formula: see text], and is decreasing for a higher magnetic field parameter. With the elevating estimates of thermal conductivity, it is shown that the cone transmits heat more quickly whereas the disc transmits heat more slowly. The thermal conductivity parameter increases the probability of collision of the liquid particles, that ultimately upsurges fluid heat transmission rate.

Analysis of the effect of surface coating and texturization on the estimation of effective thermophysical properties

This article aims to investigate the influence of surface modifications on the effective thermophysical properties of textured carbon steel and coated hard metal. The novelty of this work lies in the simultaneous estimation of thermal diffusivity and thermal conductivity via measurements using one active heating surface and Bayesian inference to solve the inverse heat conduction problem. To achieve this, an experimental setup was developed where samples were partially heated on an active surface, and the temperature was measured at two different points on the surface to estimate the thermal properties. The method was applied to two different thermal models using the same set of experimental data. Thus, the thermophysical properties were determined with and without the Markov Chain Monte Carlo technique using the Metropolis-Hastings sampling algorithm. Both techniques proved suitable for estimating the thermophysical properties. It is noteworthy that texturization and coating did not significantly modify the effective thermal diffusivity of the analyzed samples. However, there were considerable changes in the effective thermal conductivity, with an increase of approximately 13 % in the textured carbon steel sample and a reduction of approximately 11 % in the coated hard metal sample. The expanded uncertainty of the estimated effective thermophysical properties was less than 14 %. These results reflect low dispersion and good accuracy of the experimental technique used.

Estimation of heat transfer and thermal conductivity of particle-reinforced hollow cylinder composites

With the expansion of seabed gas hydrate exploitation, particle-reinforced cement is widely used in well cementing. However, the estimation of heat transfer and thermal conductivity for hollow cylinder composites reinforced with particles remains ambiguous and requires further clarification. This paper employs the inclusion-based boundary element method (iBEM) to calculate both the local field distribution and effective thermal properties of particle-reinforced hollow cylinder composites (PRHCC). Particle interactions are simulated through the introduction of the eigen-temperature gradient, and the boundary effect is modeled using the boundary element method. Numerical simulations explore the influence of interparticle interactions, particle volume fraction, particle properties, and particle distribution on the thermal properties of PRHCC. The results show that the particle interactions have distinct effects in different directions, significantly altering the temperature spread in the central and lower regions within the particles, and then affecting the overall thermal properties of PRHCC. This method serves as a valuable reference for the engineering industry, providing insightful guidance for optimizing design, fabrication processes, and material selection, particularly in components with thermal considerations.

A-priori modelling of selected physico-chemical properties of Malbec V. Vinifera specie Alphonse Lavallee variety of black grape juice

Estimation of physico-chemical properties is essential for food processing operations ranging from equipment design to plant installation and extending towards packaging, storage and distribution applications. Among the various physico-chemical properties, density, heat capacity and thermal conductivity play an influential role. Herein, the physico-chemical properties such as estimation of density, heat capacity and thermal conductivity for Malbec V. Vinifera specie Alphonse Lavallee variety of black grape juice are estimated as a function of temperature (274 to 339 K) and concentration (13.6–45°Brix). A group contribution method approach coupled with the genetic algorithm (GA) formulation is employed to estimate the parameters and respective non-linear equations for important property parameter calculations are generated. The objective functions use experimental dataset of density, thermal conductivity and heat capacity over the same temperature and concentration range subtracted from the estimated dataset to minimize the objective function values in the genetic algorithm framework. Additionally, cubic equation of state is employed as constitutive relation for such estimation. The results follow a similar trendline and show a close resemblance in comparison of the experimental dataset with high accuracy indicating a potential use of similar methodology towards estimation of properties for other food matrices.Graphical

A comparative analysis of spherically and cylindrically shaped nanoparticles containing nanofluid flows over a permeable rotating disk affected by nanoparticles radius and nanolayer

The aim of this paper is to present a comparative analysis of spherically and cylindrically shaped nanoparticles immersed separately in a graphene oxide/water nanofluid mixture over a rotating disk influenced by Hall current. The disk’s surface is permeable with suction/injection impact. In order to reveal the thermal integrity of the flow, the effect of nanoparticle diameter and the liquid–solid interfacial layer at the molecular level is also introduced. Estimates of thermal conductivity, interfacial layer thickness and the radius of both spherical and cylindrical particles are taken from experimental studies and utilized to investigate the thermal behavior of the flows. The Tiwari–Das model for nanofluid flow is considered that incorporates the thermal radiative impact in the energy equation. The equations obeying boundary layer theory are resolved to ordinary differential equations using a transformation, and further numerically computed with the bvp4c scheme. The velocities and the temperature profiles are graphically demonstrated versus the dimensionless parameters. Quantities of physical interest such as surface heat flux and surface drag coefficient are tabulated numerically. The results show that the surface drag is larger for nanofluids having cylindrical particles than for nanofluids containing spherical particles. Is also found that the heat transfer rate is greater for nanofluids containing cylindrical particles than spherical particles at higher interfacial layer thicknesses.

Numerical estimation of effective thermal conductivity of reconstructed 2D structures of cement-based building materials

Cement-based building materials are fundamental for constructing modern buildings, and accurate determination of their thermal conductivity is crucial for calculating building energy consumption. In this study, random growth theory is used to reconstruct the two-dimensional pore structures of cement-based building materials in different forms. Furthermore, heat transfer simulations and calculations of effective thermal conductivity based on the finite element method are conducted for these pore structures. The effect of pore structure parameters on the effective thermal conductivity of cement-based building materials is investigated by this method. The results indicate that porosity, pore size, and the degree of pore regularity can all influence the effective thermal conductivity of cement-based building materials. At a porosity of 0.3, the fluctuations in the effective thermal conductivity caused by variations in pore sizes and degrees of pore regularity are most drastic.

Heated fibre-optic cable installed in compacted bentonite layers: numerical evaluation of influences of the compaction-induced variability on thermal conductivity estimation

Active heating of fibre-optic (FO) cable combined with distributed temperature sensing has been applied as a quality control tool for emplacing granulated bentonite buffer. For calibrating the tool, the FO cable is installed in specimens with known dry densities. Layer-by-layer compaction is often employed to prepare the specimens. Although the average dry density is guaranteed, some studies report possible variations in the dry density within a layer. In this study, heat transfer from a heated FO cable installed in granulated bentonite with different compaction-induced variabilities was numerically simulated. The reference case used homogeneous bentonite with an average dry density of 1.5 g cm −3 . For three other cases the granulated bentonite was filled in two, three and four layers, respectively, each of which had a distinct dry density gradient ranging from 1.3 to 1.7 g cm −3 . For each case, the FO cable placed at the centre of the container was numerically heated, and the thermal conductivity was calculated using temperature changes. In the cases where the FO cable was placed on the layer interface at which the dry density was discontinuous, the calculated apparent thermal conductivity was found to have been underestimated with an error of as much as 4%. In the other two cases, the thermal conductivities obtained were nearly identical to those corresponding to the average dry density. The influence of the compaction-induced variability in the dry density was found to be insignificant and the estimated thermal conductivity was essentially that of the bentonite represented by its average dry density. Thematic collection: This article is part of the Sustainable geological disposal and containment of radioactive waste collection available at: https://www.lyellcollection.org/topic/collections/radioactive

Computational Model of Effective Thermal Conductivity of Green Insulating Fibrous Media.

Modelling effective thermal properties is crucial for optimizing the thermal performance of materials such as new green insulating fibrous media. In this study, a numerical model is proposed to calculate the effective thermal conductivity of these materials. The fibers are considered to be non-overlapping and randomly oriented in space. The numerical model is based on the finite element method. Particular attention is paid to the accuracy of the results and the influence of the choice of the representative elementary volume (REV) for calculation (cubic or rectangular parallelepiped slice). The calculated effective thermal conductivity of fibrous media under different boundary conditions is also investigated. A set of usual mixed boundary conditions is considered, alongside the uniform temperature gradient conditions. The REV rectangular slice and uniform temperature gradient boundary conditions provide a more accurate estimate of the effective thermal conductivity and are therefore recommended for use in place of the usual cubic representative elementary volume and the usual mixed boundary conditions. This robust model represents a principal novelty of the work. The results are compared with experimental and analytical data previously obtained in the literature for juncus maritimus fibrous media, for different fiber volume fractions, with small relative deviations of 7%. Analytical laws are generally based on simplified assumptions such as infinitely long fibers, and may neglect heat transfer between different phases. Both short and long fiber cases are considered in numerical calculations.

A holistic approach to understanding structural, magneto-electronic, thermoelectric, and thermodynamic properties of RhMnZ (Z = Si, Ge) half Heusler alloys.

This study thoroughly examines the structural, mechanical, thermal, electronic, optical, and thermoelectric properties of RhMnZ (Z = Si, Ge) half-Heusler compounds, which feature 18 valence electrons. Using density functional theory (DFT) within the WIEN2k computational framework, the ground-state properties of these compounds were determined to establish a foundational understanding of their physical characteristics. To further assess their thermoelectric potential, the Boltzmann transport equation was applied with the constant relaxation time approximation, allowing for precise calculations of thermal and electrical conductivity. Results indicate that the lattice constants of RhMnSi and RhMnGe span from 5.6394 Å to 5.7447 Å, highlighting consistent crystalline structures that lack band gaps, confirming their metallic nature. Detailed elastic and thermodynamic evaluations demonstrate that these compounds are mechanically stable, displaying ductile and anisotropic behavior. The study further reveals that thermal properties, including specific heat and entropy, tend to increase with the atomic number of Z, suggesting that RhMnGe may have a slightly higher heat capacity compared to RhMnSi. For thermal conductivity estimation, Slack's model was employed, indicating that these compounds possess high lattice thermal conductivity-a crucial factor for thermoelectric materials. The substantial figure of merit (ZT) observed in these compounds, especially at elevated temperatures, points to their potential efficiency in thermoelectric applications. The combination of high thermal conductivity, favorable mechanical stability, and robust thermoelectric properties identifies RhMnZ compounds as promising candidates for use in energy conversion technologies, particularly where efficient heat-to-electricity conversion is needed. This study thus lays the groundwork for future applications of RhMnSi and RhMnGe in thermoelectric devices.

Test conditions influence on thermal conductivity and contact conductance of sand at transient state

The proper assessment of soil thermal properties, such as thermal conductivity, assumes a particular relevance for the design of Ground Source Heat Pump systems, as it determines the heat transfer and the energy efficiency design. Multiple methods can be used for thermal conductivity estimation, however significant scatters in measurements are often reported when using different (or even the same) methods. This work presents a detailed study of the thermal conductivity and the thermal contact conductance of a reference sand tested in transient conditions, analyzing the effect of several factors, such as the heating time, degree of saturation, soil density, temperature and heat flux intensity, on thermal conductivity measurements for dry state or fully saturated conditions. To avoid measurement errors and heterogeneity effects, only three samples were prepared with different compaction ratios, and systematically tested in dry and fully saturated conditions, under different control variables (temperature, testing time, and injected heat flux). Two analytical methods based on the line source solution were used to estimate soil thermal conductivity and to assess the probe-to-soil thermal contact conductance. Significant relationships were obtained between thermal conductivity and soil state conditions and testing variables, highlighting that a heating time longer than the one usually recommended in the standards is clearly needed. Finally, this study uses measured temperature values to determine the probe-to-soil thermal contact conductance, indicating its relevance in soil thermal conductivity estimation.

Sea ice heat and mass balance measurements from four autonomous buoys during the MOSAiC drift campaign

As part of the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC), four autonomous seasonal ice mass balance buoys were deployed in first- and second-year ice. These buoys measured position, barometric pressure, snow depth, ice thickness, ice growth, surface melt, bottom melt, and vertical profiles of temperature from the air, through the snow and ice, and into the upper ocean. Observed air temperatures were similar at all four sites; however, snow–ice interface temperatures varied by as much as 10°C, primarily due to differences in snow depth. Observed winter ice growth rates (November to May) were <1 cm day−1, with summer melt rates (June to July) as large as 5 cm day−1. Air temperatures changed as much as 2°C hour−1 but were dampened to <0.3°C hour−1 at the snow–ice interface. Initial October ice thicknesses ranged from 0.3 m in first-year ice to 1.2 m in second-year ice. By February, this range was only 1.20–1.46 m, due in part to differences in the onset of basal freezing. In second-year ice, this delay was due to large brine-filled voids in the ice; propagating the cold front through this ice required freezing the brine. Mass balance results were similar to those measured by autonomous buoys deployed at the North Pole from 2000 to 2013. Winter average estimates of the ocean heat flux ranged from 0 to 3 W m−2, with a large increase in June 2020 as the floe moved into warmer water. Estimates of average snow thermal conductivity measured at two buoys during periods of linear temperature profiles were 0.41 and 0.42 W m−1 °C−1, higher than previously published estimates. Results from these ice mass balance buoys can contribute to efforts to close the MOSAiC heat budget.

Mathematical modelling and experimental validation of thermal conductivity of outer layer of fire protective clothing

Motivation for development of this mathematical model is the difficulties related to the experimental measurement of the thermal conductivity like frequent sample preparation, experimental setup availability and lack of easier methods for quick estimation of thermal conductivity of outer layer of fire protective clothing. Considering this, various studies were devoted in the past to determine the thermal conductivity of fabrics. Two mathematical approaches for prediction of thermal conductivity of woven outer layer have been presented in the present study. Series-parallel thermal resistance configuration has been used for prediction of effective thermal conductivity of the fabric. First approach uses input values of structural parameters and can be useful in the development of new woven fabrics for outer layer of fire protective clothing. Second approach uses fibre volume fraction and thermal conductivity of fibre and air. It is suitable for predicting thermal conductivity of finished fabrics. The values obtained from mathematical model were validated with experimental values.

Determination of the longitudinal thermal conductivity of low‐ordered carbon fibers by using an optothermal Raman technique

AbstractThe usage of carbon fibers (CFs) for high‐temperature applications has been increasing in recent years. However, the determination of thermal properties at high temperatures is a challenging task. In this study, the thermal conductivity of two different types of CF having a diameter in the range from 5–7 m, as a function of temperature, was examined by using the optothermal Raman method. Raman spectroscopy was first used to obtain the structural organization and structural homogeneity of the fibers. Then, owing to the fact that Raman spectra are sensitive to laser excitation power and external temperature, Raman spectroscopy was used as a contactless thermometer to determine the local temperature rise of the fibers. A formula was derived by solving the heat diffusion equation for cylindrical fibers and a set of boundary conditions, similar to the experimental conditions, which allows accurate estimation of longitudinal thermal conductivity. The results are discussed in relation with the phonon scattering theory and can be attributed to the combined effect of scattering from defects. The radiative and convective heat losses were estimated, and their influence on thermal conductivity was also determined.

Effective thermal conductivity estimation using a convolutional neural network and its application in topology optimization

Topology optimization of heterogeneous structures can find significant use in a wide range of applications, and its fabrication has been made possible by recent advances in additive manufacturing. However, the optimization procedure is computationally expensive, as each structural update requires the re-evaluation of the properties. The computational time is the major limiting factor in large-scale and complex structural optimization. In this study, a convolutional neural network (CNN) model for predicting effective thermal conductivity inspired by the VGG networks is proposed. Trained using 130,000 unique binary images, the model achieves high predictive accuracy. Specifically, it shows a mean absolute percent error (MAPE) of 0.35% in testing when the thermal conductivity of the solid is ten times larger than the fluid, and when the thermal conductivities assigned are that of aluminum and water, the MAPE is 2.35%. The prediction time is 15 ms for a single image with 128 × 128 pixels, which is 3 to 5 orders of magnitude faster than a finite volume simulation. When employed in topology optimization, the CNN retains a MAPE between 0.67% and 11.8% for different cases. The CNN model correctly predicts trends in effective thermal conductivity and improves the structure to close proximity of a theoretical maximum in all cases.