Artur Piasecki

Infrared thermography provides non-contact access to spatially resolved wall-temperature information, but large flow-boiling archives require traceable separation of physically meaningful variations from measurement artefacts. This study develops a physics-guided workflow for quality assessment of longitudinal infrared-temperature profiles extracted from the central line of an externally observed heated foil in rectangular minichannels. The frozen dataset comprised 143,821 complete profiles, 35,666,447 spatial points, 447 source files, and 20 grid families. Leakage-safe source-file-level partitioning was used. Controlled evaluation covered eight local synthetic anomaly morphologies at three intensities and a separate global-offset challenge. A Random Forest using 64 engineered profile descriptors was compared with a five-channel masked residual one-dimensional convolutional neural network operating directly on complete profiles. On the locked test set, the Random Forest achieved an area under the precision–recall curve of 0.915, sensitivity of 0.731, and false-positive rate of 0.082, compared with 0.871, 0.452, and 0.063 for the neural model. Both models deteriorated under external grid-family transfer and remained near chance for uniform temperature offsets. The Random Forest is retained as the preferred supervised benchmark; geometry and grid-family transfer remain the principal unresolved limitation.

A unified experimental database and a documented Python (3.12.6)-based harmonisation workflow are presented for local heat-transfer analysis in rectangular-minichannel flow boiling. The database integrates 449 experimental source files and 64,385,791 point-level records covering six working fluids, multiple surface conditions, channel configurations, and orientations. The workflow combines template-based files, central-line infrared wall-temperature data, geometric and operating metadata, local pressure and saturation temperature reconstruction, bulk fluid temperature interpolation, heat-loss correction, local heat-transfer coefficient and Nusselt number calculation, operational branch labelling, and auditable quality-control flags. The database is characterised at the point and source-file levels to identify branch imbalance, unequal source-file sizes, and heterogeneous coverage. A source-file-grouped pilot benchmark is conducted on streaming-sampled subcooled and saturated subsets using Random Forest regressors and five group-based train/test partitions. Across the five splits, the nine-predictor RF achieved R² = 0.887 ± 0.015 for subcooled Nu and R² = 0.810 ± 0.043 for the subcooled heat-transfer coefficient; saturated performance was weaker and more split-sensitive (R² = 0.340 ± 0.134 for Nu and 0.317 ± 0.091 for the heat-transfer coefficient). These results are treated as feasibility screening rather than final model ranking or evidence of campaign- or configuration-independent transfer. The database architecture and group-aware pilot validation provide a documented foundation for controlled physical interpretation and further validation.

Infrared thermography provides non-contact access to spatially resolved wall-temperature information, but large flow-boiling archives require traceable separation of physically meaningful variations from measurement artefacts. This study develops a physics-guided workflow for quality assessment of longitudinal infrared temperature profiles extracted from the central line of an externally observed heated foil in rectangular minichannels. The frozen dataset comprised 143,821 complete profiles, 35,666,447 spatial points, 447 source files, and 20 grid families. Leakage-safe source-file-level partitioning was used. Controlled evaluation covered eight local synthetic anomaly morphologies at three intensities and a separate global-offset challenge. A Random Forest using 64 engineered profile descriptors was compared with a five-channel masked residual one-dimensional convolutional neural network operating directly on complete profiles. On the locked test set, the Random Forest achieved an area under the precision-recall curve of 0.915, sensitivity of 0.731, and false-positive rate of 0.082, compared with 0.871, 0.452, and 0.063 for the neural model. Both models deteriorated under external grid-family transfer and remained near chance for uniform temperature offsets. The Random Forest is retained as the preferred supervised benchmark; geometry and grid-family transfer remain the principal unresolved limitation.

Subcooled flow boiling of distilled water was investigated in a vertical annular minichannel with smooth and surface-modified heated tubes. Copper and stainless-steel substrates were tested with ZnO and ZnO/PMHS coatings; Al₂O₃ was additionally tested on stainless steel. A simplified one-dimensional cylindrical model provided local effective heat transfer coefficients, and modified surfaces were compared pointwise with smooth references at matched operating conditions and axial positions. A modification was considered favourable only when the heat transfer coefficient increased without an increase in wall temperature. ZnO on stainless steel was the only modification meeting this criterion at both nominal mass flow rates: the mean pointwise coefficient increased by 44.2% at 7 kg/h and 42.9% at 10 kg/h, while mean wall temperature decreased by 36.9 and 32.8 K, respectively. Al₂O₃ and ZnO/PMHS on stainless steel reduced the coefficient and increased wall temperature, whereas copper modifications showed no robust improvement relative to the designated references. A separate model-sensitivity assessment did not alter the qualitative ranking. ZnO-modified stainless steel was therefore the best-performing configuration within the tested matrix; no broader superiority is claimed beyond the present geometry, fluid, flow rates, and heat-flux range.

A unified experimental database and a reproducible Python-based harmonisation workflow are presented for local heat transfer analysis in rectangular-minichannel flow boiling. The database integrates 449 experimental source files and 64,385,791 point-level records covering six working fluids, multiple surface conditions, channel configurations, and orientations. The workflow combines template-based files, central-line in-frared wall-temperature data, geometric and operating metadata, local pressure and saturation-temperature reconstruction, bulk-fluid temperature interpolation, heat-loss correction, local heat transfer coefficient and Nusselt number calculation, operational regime labelling, and auditable quality-control flags. The database is characterised at point and source-file levels to identify regime imbalance, unequal experiment sizes, and heterogeneous coverage. A leakage-safe pilot benchmark is then conducted on streaming-sampled subcooled and saturated subsets using Random-Forest regressors with source-file-level train/test separation. The strongest result is obtained for subcooled Nusselt number (R² = 0.88), followed by subcooled heat transfer coefficient (R² = 0.80); saturated targets are more difficult (R² = 0.46 for Nu and 0.41 for heat transfer coefficient). These results are treated as feasibility screening rather than final model ranking. The proposed database architecture and validation strategy establish a repro-ducible foundation for controlled physical interpretation and extended machine-learning assessment.

This work presents a correlation-based framework for comparative assessment of single-phase forced convection of binary hydrofluoroether/ethyl acetate (HFE/EA) mixtures in rectangular minichannels. Density, kinematic viscosity, and thermal conductivity were measured at 293.1, 313.1, and 328.1 K for selected compositions of HFE-7100/EA, HFE-7300/EA, and HFE-73DE/EA. Because several DSC-derived mixture-specific heat values were not sufficiently reliable for direct use, the mixture-specific heat capacity was estimated from literature-supported pure-component values using an ideal-mixture, mass-fraction-weighted approximation and used only for evaluation of the Prandtl number. Heat transfer was then assessed using the Sieder–Tate correlation for laminar thermally developing flow in two representative minichannel geometries. The highest predicted values for the HFE-7100/EA and HFE-7300/EA families were obtained for the most EA-rich retained compositions, whereas in the retained HFE-73DE/EA subset, the 50/50 mixture performed best because of its higher thermal conductivity. Validation against an experimental dataset for pure HFE-7100 in the short module showed systematic overprediction, with a mean relative difference of −13.44% and a MAPE of 15.6%. The calculated values should, therefore, be used for relative comparison rather than treated as unbiased absolute predictions.

This paper reports thermophysical-property data for binary dielectric mixtures of hy-drofluoroether (HFE) fluids and ethyl acetate (EA) and applies a correlation-based workflow to compare their single-phase forced-convection performance in rectangular minichannels. Density, viscosity, thermal conductivity, and isobaric heat capacity were measured at three temperature levels (293.1, 313.1, and 328.1 K) for selected compositions of HFE-7100/EA, HFE-7300/EA, and HFE-73DE/EA. Using these meas-ured properties, Reynolds and Prandtl numbers were evaluated and a laminar ther-mally developing correlation was employed to obtain Nusselt numbers and corre-sponding heat transfer coefficients. The assessment was performed for two geometries representing a long reference minichannel module and a short multi-minichannel module. A validation dataset for pure HFE-7100 in the short module, derived from IR thermography and an energy-balance data reduction, indicates a systematic deviation between correlation-based estimates and experimental values, which should be con-sidered when interpreting absolute predictions. The presented dataset and workflow support transparent down-selection of candidate mixtures prior to extended experi-mental campaigns.

Binary mixtures of HFE-73DE and ethyl acetate are investigated as dielectric working fluids for laminar minichannel cooling. Thermophysical properties of the pure components and four mixtures (10/90, 25/75, 50/50 and 75/25 mass % HFE-73DE/ethyl acetate) were measured over the relevant temperature range. Single-phase convective heat transfer tests were then carried out in a heated 1 × 4 × 180 mm minichannel test section under constant heat-flux conditions for pure HFE-73DE. A three-dimensional conjugate CFD model with temperature-dependent liquid properties was developed in Simcenter STAR-CCM+ and validated against these measurements; the average relative temperature difference between CFD and experiment remained below 0.5%, while a grid-convergence study based on the Grid Convergence Index (GCI) confirmed that the numerical uncertainty is comparable to the experimental one. The validated model was subsequently used to predict the axial evolution of wall temperature, fluid-core temperature, velocity and heat transfer coefficient for the four mixtures under identical conditions. The mean Nusselt numbers obtained from CFD were further compared with the classical Shah and London fully developed laminar solution for rectangular ducts, revealing that the present configuration yields values about 35–42% higher than the theoretical prediction owing to asymmetric heating and conjugate heat transfer. The results show that increasing the HFE-73DE mass fraction strengthens convective heat transfer and reduces fluid-temperature rise, while intermediate compositions (50/50 and 75/25) provide a favourable compromise between enhanced heat transfer performance and moderate pressure drop. The study provides guidance for composition selection and the design of dielectric minichannel heat exchangers operating with HFE-73DE/ethyl acetate mixtures.

This study investigated FC-72 boiling heat transfer in minichannels using two non-contact temperature measurement techniques: liquid crystal thermography (LCT) and infrared thermography (IRT). These methods were applied simultaneously to measure temperature distributions on the heated wall surface of minichannels, formed by a thin metal foil. The temperature data facilitated the calculation of local heat transfer coefficients at the foil–working fluid contact surface. Calibration of the liquid crystal colour response to temperature was conducted prior to the use of LCT. According to a comparison of the heat transfer coefficients and Nusselt numbers determined using LCT and IRT measurements, comparable temperature distributions are provided, with the average relative differences in heat transfer coefficients determined using these techniques remaining below 15%. The findings highlight the advantages of non-contact temperature measurement in minimising system disturbances while providing precise data for understanding flow boiling heat transfer mechanisms. Such results can contribute to the design of minichannel heat exchangers.

Analizy porównawcze pomiarów temperatury powierzchni wykonano na podstawie wyników badań wykonanych na stanowisku badawczym z wymiennymi foliami ze stopów Haynes-230 i Hastelloy X, o różnych grubościach. Podczas eksperymentów realizowano pomiar temperatury metodą kontaktową (czujnikami rezystancyjnymi i termoelementami) oraz metodą bezkontaktową (dwie kamery termowizyjne). Tor pomiarowy zawierał termometr o wysokiej dokładności współpracujący z czujnikiem RTD 1/10 DIN (o klasie B) oraz kalibrator podczerwieni.