Prediction Of Heat Transfer Coefficient Research Articles

Experimental comparison and optimal machine learning technique for predicting the thermo-hydraulic performance of Low-GWP refrigerants (R1234yf, R290, and R13I1/R290) during evaporation in plate heat exchanger

In recent times, employing machine learning techniques (MLTs) to forecast the evaporation/condensation performance of refrigerants, namely the heat transfer coefficient (HTC) and frictional pressure drop (FPD), has become increasingly significant. In the current investigation, an experimental comparison of the evaporation HTC and FPD of R1234yf, R290, and R13I1/R290 was explored in an offset strip fin-plate heat exchanger. Following this, six MLTs—support vector regressor (SVR), multilayer perceptron regressor (MLPR), gradient boosting regressor (GBR), AdaBoost regressor (ABR), ridge regressor (RR), and K-nearest neighbors regressor (KNNR)—were proposed to forecast the HTC and FPD of these refrigerants. The comparative experimental analysis revealed that the evaporation-HTC of R290 was 27.8–76.2 % and 46.2–73.8 % higher than that of R1234yf and R13I1/R290, respectively, while FPD declined by up to 74.8 % and 57.2 %, respectively. Under the same working conditions, the first transition from nucleation to convective boiling occurred in the order of R1234yf, R13I1/290, and R290. In addition, nucleate boiling dominated at lower vapor quality, while convective boiling prevailed at higher vapor quality. A total of 336 experimental data points corresponding to various test conditions from existing studies and current experiments were considered for the MLT prediction analysis. According to the results, GBR (without any enhancement approaches) was the best approach for forecasting the FPD and HTC, with mean absolute errors (MAE) of 0.125 % and 0.131 %, respectively. Additionally, to enhance the forecasting efficiency of the MLTs, feature selection, analysis of principal components, and hyperparameter tuning were employed. According to feature importance, mass flux, reduced pressure, saturation temperature, heat flux, and entry vapor quality were identified as the most influential parameters on the FPD and HTC. Ultimately, the most accurate predictions of HTC and FPD, with the lowest MAE of 0.111 %, were achieved using MLPR and SVR, respectively, with feature selection.

Experimental Investigation and Calculation of Convective Heat Transfer in Two-Component Gas–Liquid Flow Through Channels Packed with Metal Foams

This paper presents a study on heat transfer in two-phase mixtures (air–water and air–oil) flowing through heated horizontal channels filled with open-cell aluminum foams characterized by porosities of 92.9–94.3% and pore densities of 20, 30, and 40 PPI. The research included mass flux densities ranging from 2.82 to 284.7 kg/(m2·s) and heat flux densities from 5.3 to 35.7 kW/m2. The analysis examined the effects of flow conditions, fluid properties, and foam geometry on the intensity of heat transfer from the heated walls of the channel to the fluid. Results indicate that the heat transfer coefficient in two-component non-boiling flow exceeds that of single-phase flow, primarily due to fluid properties and velocities, with minimal impact from flow structures or foam geometry. An assessment of existing methods for predicting heat transfer coefficients in gas–liquid and boiling flows revealed significant discrepancies—up to several hundred percent—between measured and predicted values. To address these issues, a novel computational method was developed to accurately predict heat transfer coefficients for two-component non-boiling flow through metal foams.

Comparative Analysis of Heat Transfer Coefficient Using Experimental Data and Empirical Expressions

This study aims to determine the heat transfer coefficient by comparing results obtained from both experimental methods and empirical expressions. A controlled experiment involving the flow of water through a polyethylene hose was conducted, and the heat transfer coefficient was calculated based on inlet and outlet temperatures. These findings were then compared with values derived from empirical correlations, specifically the Sieder-Tate equation, to assess accuracy and reliability. The experimental setup maintained consistent boundary conditions, including water inlet temperature, airflow rate, and hose orientation. The results indicated that the experimentally determined heat transfer coefficient was 48.30 W/(m²·K), while the empirical calculation yielded a value of 53.44 W/(m²·K). The slight discrepancy between these values highlights minor experimental errors and assumptions inherent in empirical models. Overall, the close agreement between the experimental and empirical values validates the use of empirical correlations for predicting heat transfer coefficients in similar configurations. This study provides valuable insights for optimizing heat transfer processes in various industrial and engineering applications, emphasizing the importance of experimental validation. Future work could explore extended parameter ranges, advanced measurement techniques, and numerical simulations to further enhance the accuracy and applicability of the findings.

Improving turbulence modeling for gas turbine blades: A novel approach to address flow transition and stagnation point anomalies

Accurate prediction of temperature and Heat Transfer Coefficient (HTC) distributions over gas turbine blades is crucial for the design process and life assessment of these components. Numerical studies of flow over gas turbine blades face significant challenges in accurately simulating two complex phenomena: (1) the transition of flow from laminar to turbulent, and (2) stagnation point flow at the leading edge. Many turbulence models tend to overpredict the temperature on turbine blades, leading to incorrect identification of hot-spot regions and, consequently, erroneous estimations of blade life. This paper investigates the performance of various turbulence models in simulating flow and heat transfer over gas turbine vanes. The study includes three full turbulence models, i.e., Spalart-Allmaras (SA), Shear Stress Transport k−ω (SST-kw), and v2−f (V2F), as well as two transitional models, i.e., Transition SST (Trans-SST) and k−kL−ω (k-kl-w). Simulation results indicate that the v2−f, Trans-SST, and k−kL−ω models can detect flow transition. However, the transition length and onset location predicted by the Trans-SST and k−kL−ω models do not align with experimental data. Conversely, the v2−f model suffers from over-predictions at the leading edge due to stagnation point anomaly. To address these issues and due to capacities of the V2F model, this study proposes two modifications to enhance the performance of the V2F model. First, the production term of turbulent kinetic energy is redefined to mitigate the stagnation point anomaly. Second, the model is recalibrated to improve the prediction of flow transition. The new model, named the Production Modified V2F (PMV2F) model, shows promising results in predicting temperature and heat transfer coefficients.

Carbon dioxide evaporation heat transfer coefficient prediction in porous media using Machine learning algorithms based on experimental data

The prediction of the internal heat transfer coefficient during evaporation is vital for vapor-compression refrigeration and closed-loop power cycles. Accurate measurements and understanding of CO2 heat transfer in porous evaporators are essential for optimal system design across various operating conditions. This study utilizes a reference dataset derived from previous experiments that investigated the impact of porous evaporators on CO2′s internal heat transfer coefficient under sub-critical conditions, employing gravel sand as the porous medium. The dataset encompasses key factors: gravel sand porosities ranging from 39.8 % to 44.5 %, evaporator inlet pressures between 3700 and 4300 kPa, CO2 mass flow rates from 10.7x10-5–18x10-5kg.s−1, and porous tube effective diameters spanning 1.53 x10-3–3.4x10-3m. Employing three machine learning techniques (SVM, GPR, OBEM), the study predicts the internal heat transfer coefficient using regression models. The models’ predictions are analyzed and compared to expected values for validation, evaluating their performance using four statistical criteria. Results indicate SVM, GPR, and OBEM models achieved RMSEs of 1.5471, 1.8212, and 3.6978, respectively, while MAE errors were 1.1479, 1.2418, and 2.9787, respectively. Comparison with the dimensional analysis method reveals the effectiveness of the proposed models in accurately predicting internal heat transfer coefficients. The models exhibit low uncertainty and maintain prediction quality on an extended dataset without overfitting concerns. Overall, this research contributes valuable insights for designing heat exchangers and systems in vapor-compression refrigeration and closed-loop power cycles.

Regression modeling of Bödewadt slip flow dynamics involving Reiner-Rivlin nanofluid based on a modified Buongiorno approach

Regression models are useful in analyzing rotational flows as they enable accurate predictions of wall shear and heat transfer coefficient. In addition, Bödewadt flow is of paramount importance in fluid dynamics of rotating systems such as turbomachinery and geophysical flows. Moreover, nanofluid’s enhanced heat transfer properties can improve cooling efficiency in applications involving turbines and electronic systems. This study delves into the Bödewadt boundary layer flow of a Reiner-Rivlin fluid containing nanoparticles over a stationary porous disk under slip conditions. The two-phase Buongiorno model is employed, incorporating temperature-dependent diffusion coefficients for enhanced accuracy. To facilitate numerical simulations, the transport equations are converted into an ordinary differential system comprising four unknowns. In the present work, a highly reliable Keller-Box methodology is adopted which agrees very well with the MATLAB built-in program ‘bvp4c’. The computed 2-D and 3-D streamlines vividly capture the Bödewadt flow scenario with Reiner-Rivlin nanofluid. The principle aim to investigate the impact of non-Newtonian behaviour and slip on the flow pattern, while also examining the behavior of temperature/concentration field for nanoparticle working fluids. As thermophoretic diffusion increases, the thermal boundary layer thickens considerably, leading to a notable decrease in the cooling rate of the disk. In contrast, Brownian diffusion has only a minimal impact on the heat transport. In addition, wall suction effect is observed to significantly boost the disk’s cooling rate, though at the expanse of increasing skin friction coefficients. This study introduces linear and quadratic regression models designed to precisely predict both the surface drag and disk cooling rate, which are crucial factors in engineering processes.

Analysis of the aerothermal performance of modern commercial high-pressure turbine rotors using different levels of fidelity

In the field of design and optimization of sophisticated geometries such as film-cooled turbine blades of modern jet engines, Computational Fluid Dynamics (CFD) simulation is commonly employed. As the pursuit for higher turbine entry temperatures intensifies, it becomes crucial for computational analyses to offer accurate predictions of metal temperatures and heat transfer coefficients on these critical components. This study investigates the impact of key physical factors that characterize the operation of these components on the accuracy of the computational model. In particular, the effects of including the exchange of thermal energy between the fluid and solid domains, as well as modelling the unsteady interaction between the rotor and stator are studied. The research focuses on a fully featured one-and-a-half stage high-pressure turbine of a commercial jet engine, utilizing proprietary software for conducting 3D Reynolds-Averaged Navier-Stokes flow simulations. To model the fluid-solid thermal interaction, steady-state Conjugate Heat Transfer (CHT) simulations are performed. The CHT results are then compared with experimental data obtained from a thermal paint test, achieving good levels of agreement. Additionally, phase-lag simulations are executed under adiabatic-wall conditions to evaluate the influence of stator-rotor interaction on near-wall gas temperatures. This work shows that, by modelling the periodic unsteadiness at the stator-rotor interface with a phase-lag technique, the maximum near-wall gas temperature prediction is significantly increased, sometimes more than 100K, with respect to the steady-state model one. Findings from this work also suggest that simplified “strip” source terms used to model the presence of film cooling hole rows on the surface of a blade can be used with satisfactory accuracy only for nominal conditions. The mass flows delivered by each source strip should not be linearly scaled based on mainstream quantities for use in different conditions, but they should be recalculated from scratch for the new conditions.

Investigation on heat transfer of supercritical CO2/Xe mixture crossing pseudo-critical temperature cooled in a horizontal circular tube

Supercritical CO2/Xe mixture has a potential application prospect in the Brayton cycle system from the perspective of thermodynamic. Supercritical cooler is one of the key components there. However, the cooling heat transfer characteristics and mechanism of supercritical CO2/Xe mixture is still unclear, which restricts the development and design of supercritical cooler. The heat transfer of supercritical CO2/Xe mixture crossing Tpc cooled in a horizontal tube with inner diameter 8 mm is numerically explored in this study. The influences of operating parameters on supercritical mixture heat transfer are thoroughly examined. Buoyancy effect is evaluated through the buoyancy criterion Gr/Re2. Further, the mechanism of supercritical CO2/Xe mixture cooling heat transfer is revealed. It is found that heat transfer enhancement may occur during supercritical CO2/Xe cooled process, and heat transfer is weakened when mass fraction of Xe increases. Considering the bulk specific heat and buoyancy dominating jointly the behavior of supercritical cooling heat transfer, a modified Dittus-Boelter correlation is newly developed to predict heat transfer coefficient of supercritical CO2 and its mixture with Xe. The correlation matches well with the experimental and numerical data, and the average relative deviations of the new correlation are 18.27 % and 14.14 %, respectively. The investigation provides insight into the characteristics and mechanism of supercritical CO2/Xe mixture cooling heat transfer. The correlation can provide significant theoretical guidance for accurate design and optimization of supercritical cooler.

Accurate machine-learning-based prediction of aerodynamic and heat transfer coefficients for cylindrical biomass particles

Cylindrical biomass (straw) particles play a crucial role in the numerical simulations of biomass co-firing in coal-fired boilers, wherein the aerodynamic and heat transfer coefficients of these particles are particularly important. Despite the existence of models proposed thus far for various specific conditions, developing a universal model for cylindrical particles remains challenging in multiphase flow systems. In this paper, a total of 1092 validated computational fluid dynamics (CFD) datasets of cylindrical drag, lift, torque coefficients and average Nusselt number are acquired based on direct numerical simulations of OpenFOAM body-fitted meshes, and a hybrid algorithmic framework of N+1 convolutional neural network (CNN) machine learning optimality is established. where multiple single-tree models (referred to as N models) are combined with an additional CNN layer (referred to as +1 model).The framework inherits the advantages of a single-tree models while significantly improving the overall algorithmic generalisation performance, with the aim of allowing efficient use of time and minimal expertise in the fluid modelling process. The mean relative absolute errors for drag, lift, torque coefficient, and the average Nusselt number was evaluated, and the results demonstrated a reduction of 75.19 %, 89.48 %, 89.53 %, and 85.70 %, respectively, in the mean relative absolute errors compared to the extremely randomised tree single-tree model. We extended the model scope to different aspect ratios (1 ≤ Ar ≤ 30), angles of incidence (0° ≤ θ ≤ 90°), and Reynolds numbers (1 ≤ Re ≤ 4000), and conducted random predictions. Compared with traditional mathematical relationships, the proposed framework reduced the average relative error to 0.99 %. The simultaneous application of models for the prediction of unknown parameters was validated against relevant early literature, which demonstrated that the average relative error of drag coefficients for the predicted values remained within 2.07 %. The N+1 (CNN) hybrid algorithm framework exhibited high accuracy and excellent adaptability, furnishing a substantial quantity of data to support the Euler–Lagrange model of non-spherical biomass particle multiphase flow in industrial applications.

Data-driven prediction of convective heat transfer coefficients in internal walls of aero-engine bearing chambers using Mind Evolution Algorithm-Enhanced Bayesian regularization neural networks

In the bearing chambers of aero-engines, the complex nature of the oil–gas flow complicates the precise calculation of heat transfer coefficients. Addressing computational discrepancies, firstly, this study establishes a correlation analysis method, which considers the relationship between several independent variables including the geometric shape of the bearing chamber, rotational speed, oil flow rate, thermal sealing air mass, circumferential angle, axial position, and the dependent variable, which is the logarithmic Nusselt number on the internal wall. Secondly, four models are developed using Multiple Linear Regression, Back Propagation Neural Network, Deep Belief Neural Network, and Convolutional Neural Network with Long Short-Term Memory Network, yielding Mean Relative Percentage Errors (MRPE) of 5.14%, 2.80%, 4.66%, and 4.12%, respectively. Due to the limited predictive performance of these models, the neural network topology is further optimized. After repeated trial-and-error calculations, the neural network with three hidden layers was found to be the most effective. Finally, a novel BP Neural Network optimized by the Mind Evolution Algorithm-enhanced with Bayesian regularization (MEA-BPNN-Bayesian) model is proposed forpredicting the local convective heat transfer, achieving an MRPE of 1.71%. The model accurately predicted heat transfer coefficients, opening up new ways to improve heat management in bearing chambers.

Investigation on the heat transfer estimation of subcooled liquid hydrogen for transportation applications using intelligent technique

Hydrogen, as a sustainable fuel for transportation sector as well as some other heavy industries, should be stored or transferred in liquid phase. To prevent the evaporation of hydrogen in the tanks and the consequent hazards, it is common to lower its temperature below the saturation point known as subcooling of the fluid. When the subcooled liquid hydrogen flows through pipes with higher wall temperatures, the hydrogen can evaporate and flow boiling occurs, resulting in a complex heat transfer behavior of the fluid. In this study, we propose an advanced intelligent model based on the Cascade Forward Neural Network (CFNN) to predict Heat Transfer Coefficient (HTC) of subcooled liquid hydrogen under various flow conditions. Our model leverages on a dataset that is nearly twice the size of the largest previously used in the literature, significantly enhancing its generalization, reliability, and accuracy. The proposed CFNN model incorporates 19 input features, including crucial dimensionless parameters, orientation and thermodynamic condition of the fluid, and characteristics of the pipe that hydrogen flow through to accurately estimate HTC in unseen fluid conditions. The results demonstrate that the proposed model achieves an R-squared value of 0.9978, substantially outperforming the best empirical methods, which have an R-squared value of only 0.8058 for the same testing data. This innovative approach provides a highly accurate and reliable tool for predicting HTC, offering significant improvements over conventional methods and contributing valuable insights for real-world applications in the energy and transportation sectors.

Deep learning-based prediction of velocity and temperature distributions in metal foam with hierarchical pore structure

Constrained by the substantial computational time required for numerical simulation, a deep learning technique is applied to investigate fluid flow and heat transfer processes in metal foam with a hierarchical pore structure. This work adopted a 3D convolutional neural network (CNN) combining U-Net architecture to predict velocity and temperature distributions, alongside corresponding permeability and overall heat transfer coefficient. This approach demonstrates excellent capability in intricate image segmentation. The training sets were acquired by lattice Boltzmann method (LBM) simulations. The CNN model, trained on a substantial amount of data, demonstrates remarkable precision, exhibiting mean relative errors of 0.57% for permeability prediction and 2.27% for overall heat transfer coefficient prediction. Moreover, in CNN prediction, a broader range of structure parameters and boundary conditions beyond those in the training set was used to evaluate the practicability of the trained CNN model. In contrast to numerical simulation, the CNN model economizes approximately 95.41% and 99.57% of computational time for velocity and temperature distribution prediction, respectively, providing a novel approach for exploring transport processes in metal foam with hierarchical pore structure.

Machine learning boiling prediction: From autonomous vision of flow visualization data to performance parameter theoretical modeling

Flow boiling is a highly efficient configuration for meeting the high heat dissipation demands of thermal management systems. However, the complex physics of two-phase flow has hindered its broader application, especially in terms of quantifying visual information. Recent advancements in machine learning vision tools have revolutionized the analysis of phase change phenomena by enabling the digitalization of physically meaningful features such as void fraction, vapor-liquid interfacial behaviors, and liquid-solid wall wetting front areas en masse. In this study, we systematically investigate two-phase models that compute void fractions, heat transfer coefficients, and critical heat flux using live bubble data streams under microgravity. The collected empirical bubble data is used to supplement and validate traditional control-volume-based theoretical modeling approaches. Void fraction data is first validated with analytical frameworks. This is followed by void fractions and wetting front areas being used to improve correlations predicting heat transfer coefficients. This work showcases the potential of using a new machine learning-based strategy to accelerate scientific formula discovery through the extraction of multi-level and physically meaningful features.

Radial basis Neural Network Model to Prediction of Thermal Resistance and Heat Transfer Coefficient of Oscillating Heat Pipe Using Graphene and Acetone-Based Nanofluids

Radial basis Neural Network Model to Prediction of Thermal Resistance and Heat Transfer Coefficient of Oscillating Heat Pipe Using Graphene and Acetone-Based Nanofluids

Experimental study on heat transfer characteristics of dispersed flow during strong transient reflood process

The dispersed heat transfer correlations obtained in the quasi-steady-state heating experiment may have some limitations in the strong transient reflood process. Therefore, this article will conduct a visualization experiment on single-rod channel reflood cooling under different working conditions. By combining the transient hydrodynamic characteristics of liquid droplets and the wall temperature transient curve, the boiling heat transfer characteristics of the dispersed flow stage during the reflood process are studied. Through comparative analysis, this article found that the traditional post-CHF (Critical Heat Flux) heat transfer correlations are difficult to characterize the impact of droplet number and size changes. This article is based on the form of a divided path heat transfer mechanism model. In response to the strong transient characteristics of dispersed flow during the reflood process in a restricted channel, the single-phase steam heat transfer model and the steam-droplet interfacial heat transfer model have been modified, greatly improving the prediction accuracy. The RMS of the predicted heat transfer coefficient decreased from 0.68 in the original model to 0.24 in the new model.

Prediction of heat transfer coefficient and pressure drop of flow boiling and condensation using machine learning

Heat transfer coefficient and pressure drop are two critical parameters in diverse thermal engineering applications, and the prediction of these parameters plays a crucial role in the development of heat exchangers. Traditional methods are often reliant on empirical correlations based on limited refrigerants and experimental conditions. The present research explores the application of machine learning methods to predict the heat transfer coefficient and pressure drop of a wide range of pure and mixed refrigerants. The methodology involves the compilation of over 20, 000 experimental observations encompassing an extensive range of operating conditions, heat exchanger geometries, and refrigerants. A total of 27 machine learning algorithm models, such as regression tree, support vector machines, and deep neural networks, are trained on this diverse dataset to discern intricate patterns and dependencies. After dividing the dataset into 80% for training and the rest for validation, the machine learning prediction shows an ability to predict the heat transfer coefficient and the pressure drop with high accuracy in most of the refrigerants. Notably, the accuracy achieved by the machine learning models surpassed that of conventional correlations, highlighting their superior predictive capabilities. Furthermore, the accurate prediction of heat transfer enhancement methods added another layer of validation to the model’s effectiveness. It is noteworthy that the dataset employed in this study has been made publicly accessible online.

Heat transfer optimisation using novel biomorphic pin-fin heat sinks: An integrated approach via design for manufacturing, numerical simulation, and machine learning

With the availability of advanced manufacturing techniques, non-conventional shapes and bio-inspired/biomorphic designs have shown to provide more efficient heat transfer. Consequently, this research investigates the heat transfer performance and fluid flow characteristics of novel biomorphic scutoid pin fins with varying volumes and top geometries. Numerical simulations were conducted using four hybrid designs for Reynolds Number 5500–13500. The impact of pin fin 'top' geometrical features on the heat transfer coefficient (HTC) was evaluated by combining computational fluid dynamics (CFD), experimental data, and machine learning. The results highlighted that the new pin fins saved 6.3 % to 14.3 % volume/material usage but produced around 1.5 to 1.7 times more heat transfer than conventional square/rectangular fins. Also, manipulating pin fins via the top geometrical properties can lead to more uniform velocity and temperature distributions while demonstrating the potential for increased thermal efficiency with reduced thermal resistance. Furthermore, six machine learning models accurately predict HTC using volume and surface area as key variables, achieving less than 5 % mean absolute percentage error (MAPE). Overall, this research introduces innovative biomorphic designs with unconventional geometries, emphasising resource optimisation and efficient HTC prediction using machine learning. It simplifies design processes, supports agile product development, calls for re-evaluation of conventional heat sink geometries, and provides promising directions for future research.

A prediction model based on data-driven method for velocity and heat transfer coefficient of falling-film liquid on horizontal tube

The prediction of liquid film velocity and heat transfer coefficient of horizontal-tube falling-film flow is important for the investigation and enhancement of the flow and heat transfer behavior. To date, the prediction mainly relies on semi-empirical-semi-theoretical formulas based on data fitting. This prediction method has many limitations and assumptions on working conditions, and the error is large when it is extended to practical applications. In this paper, a support vector regression machine prediction method based on sine cosine algorithm optimization is proposed to improve the prediction accuracy of the liquid film velocity and heat transfer coefficient. And the database was established by conducting horizontal-tube falling-film flow and heat transfer experiments. Results show that the support vector regression model has the best prediction performance with small sample size in comparison with other four machine learning methods and traditional correlation based on Least Square method. The RMSEs of liquid film prediction model and heat transfer prediction model are only 0.0281 and 0.0117 respectively. And the MAPEs are 8.58 and 2.63 respectively. The liquid film velocity and heat transfer coefficient were accurately predicted based on relevant parameters. The importance of these parameters was also analyzed.

Prediction of Heat Transfer Coefficient of Two Phase Flow Sodium Reactor Based on Artificial Neural Network

Liquid sodium metal is one of the main coolants for fast neutron reaction in nuclear reactors with high heat load and high thermal conductivity. Sodium reactor is one of the most promising and commercialized reactor types in modern nuclear power system. Because the chemical properties of liquid sodium metal are very active, so in a sealed environment, it is easy to react with oxygen, water and other substances, resulting in an explosion. Therefore, it is necessary to design a reasonable experimental circuit and conditions under the appropriate experimental environment to carry out the experiment. With liquid metal sodium as the experimental medium, this paper studies the flow heat transfer characteristics of a two-phase flow sodium reactor by establishing a mathematical model. The key heat transfer parameters in a two-phase flow sodium reactor are predicted by building an artificial neural network model, and the parameters affecting the network performance are adjusted by the artificial neural network. The error of the final result predicted by the adjusted artificial neural network model is within the range of ±10%. The prediction results show that the prediction by the artificial neural network model has higher prediction accuracy.

Research on vacuum glass insulation performance prediction based on unsteady state multivariate data screening and multi-model fusion self-optimization

A self-optimization regression method based on multivariate data screening and multi-model fusion is proposed for the regression prediction of vacuum glass insulation performance. Firstly, we analyzed the positive correlation between the temperature change rate and heat transfer coefficient. Secondly, we combined the technique of multi-distribution overall diffusion to generate virtual sample data with multiple variables corresponding to the small neighborhood of the target variable, addressing the issue of insufficient sample size. We conducted quantity and threshold screening on the generated data to enhance the effectiveness of the virtual sample data. The screened data and original training set data are integrated to form new training samples. Then, the limited-memory Broyden Fletcher Goldfarb Shanno with boundary constraints (L-BFGS-B) algorithm is introduced to optimize the parameters of multiple models to improve the accuracy of model regression prediction. Finally, through the self-searching optimization framework structure, the virtual sample data screening and optimization of various model parameters are updated. Eventually, the optimal model for predicting the heat transfer coefficient of vacuum glass insulation performance is determined from multiple models. In order to validate the effectiveness of self-optimization regression method, we conducted insulation performance heat transfer coefficient predictions under 30 different ways for vacuum glass. Experimental results demonstrate that the self-optimization regression method based on multivariate data screening and multi-model fusion can obtain the most effective regression prediction model, achieving accurate, rapid and intelligent prediction of vacuum glass insulation performance.