https://www.mdu.se/

mdu.sePublications
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Machine Learning Techniques for Enhanced Heat Transfer Modelling
Mälardalen University, School of Business, Society and Engineering.
2024 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

With the continuous growth of global energy demand, processes from power generation to electronics cooling become vitally important. The role of heat transfer in these processes is crucial, facilitating effective monitoring, control, and optimisation. Therefore, advancements and understanding of heat transfer directly correlate to system performance, lifespan, safety, and cost-effectiveness, and they serve as key components in addressing the world's increasing energy needs.

The field of heat transfer faces the challenge of needing intensive studies while retaining fast computational speeds to allow for system optimisation. While advancements in computational power are significant, current numerical models lack in handling complex physical problems such as ill-posed. The domain of heat transfer is rapidly evolving, driven by a wealth of data from experimental measurements and numerical simulations. This data influx presents an opportunity for machine learning techniques, which can be used to harness meaningful insights about the underlying physics.

Therefore, the current thesis aims to the explore machine learning methods concerning heat transfer problems. More precisely, the study looks into advanced algorithms such as deep, convolutional, and physics-informed neural networks to tackle two types of heat transfer: subcooled boiling and convective heat transfer. The thesis further addresses the effective use of data through transfer learning and optimal sensor placement when available data is sparse, to learn the system behaviour. This technique reduces the need for extensive datasets and allows models to be trained more efficiently. An additional aspect of this thesis revolves around developing robust machine learning models. Therefore, significant efforts have been directed towards accounting for the uncertainty present in the model, which can further illuminate the model’s behaviour. This thesis shows the machine learning model's ability for accurate prediction. It offers insights into various parameters and handles uncertainties and ill-posed problems. The study emphasises machine learning's role in optimising heat transfer processes. The findings highlight the potential of synergistic application between traditional methodologies and machine learning models. These synergies can significantly enhance the design of systems, leading to greater efficiency.

Place, publisher, year, edition, pages
Västerås: Mälardalens universitet, 2024.
Series
Mälardalen University Press Dissertations, ISSN 1651-4238 ; 399
National Category
Engineering and Technology Energy Engineering
Research subject
Energy- and Environmental Engineering
Identifiers
URN: urn:nbn:se:mdh:diva-64898ISBN: 978-91-7485-625-5 (print)OAI: oai:DiVA.org:mdh-64898DiVA, id: diva2:1815756
Public defence
2024-02-13, Delta, Mälardalens universitet, Västerås, 09:00 (English)
Opponent
Supervisors
Available from: 2023-12-01 Created: 2023-11-29 Last updated: 2024-01-23Bibliographically approved
List of papers
1. A Data-Driven Approach for the Prediction of Subcooled Boiling Heat Transfer
Open this publication in new window or tab >>A Data-Driven Approach for the Prediction of Subcooled Boiling Heat Transfer
2020 (English)In: Proceedings of The 61st SIMS Conference on Simulation and Modelling SIMS 2020, 2020, p. 435-442Conference paper, Published paper (Other academic)
Abstract [en]

In subcooled flow boiling, heat transfer mechanism involves phase change between liquid phase to the vapour phase. During this phase change, a large amount of energy is transferred, and it is one of the most effective heat transfer methods. Subcooled boiling heat transfer is an attractive trend for industrial applications such as cooling electronic components, supercomputers, nuclear industry, etc. Due to its wide variety of applications for thermal management, there is an increasing demand for a faster and more accurate way of modelling. 

In this work, a supervised deep neural network has been implemented to study the boiling heat transfer in subcooled flow boiling heat transfer. The proposed method considers the near local flow behaviour to predict wall temperature and void fraction of a sub-cooled mini-channel. The input of the network consists of pressure gradients, momentum convection, energy con- vection, turbulent viscosity, liquid and gas velocities, and surface information. The output of the model is based on the quantities of interest in a boiling system i.e. wall temperature and void fraction. The network is trained from the results obtained from numerical simulations, and the model is used to reproduce the quantities of interest for interpolation and extrapolation datasets. To create an agile and robust deep neural network model, state-of-the-art methods have been implemented in the network to avoid the overfitting issue of the model. The results obtained from the deep neural network model shows a good agreement with the numerical data, the model has a maximum relative error of 0.5 % while predicting the temperature field, and for void fraction, it has approximately 5 % relative error in interpolation data and a maximum 10 % relative error for the extrapolation datasets. 

Series
Linköping Electronic Conference Proceedings, ISSN 1650-3686 ; 176:62
National Category
Energy Systems Energy Engineering Fluid Mechanics and Acoustics Applied Mechanics
Identifiers
urn:nbn:se:mdh:diva-54529 (URN)10.3384/ecp20176435 (DOI)
Conference
61st SIMS Conference on Simulation and Modelling SIMS 2020, September 22-24, Virtual Conference, Finland
Available from: 2021-06-07 Created: 2021-06-07 Last updated: 2023-11-29Bibliographically approved
2. Derivation and Uncertainty Quantification of a Data-Driven Subcooled Boiling Model
Open this publication in new window or tab >>Derivation and Uncertainty Quantification of a Data-Driven Subcooled Boiling Model
Show others...
2020 (English)In: Energies, E-ISSN 1996-1073, Vol. 13, no 22, article id 5987Article in journal (Refereed) Published
Abstract [en]

Subcooled flow boiling occurs in many industrial applications where enormous heat transfer is needed. Boiling is a complex physical process that involves phase change, two-phase flow, and interactions between heated surfaces and fluids. In general, boiling heat transfer is usually predicted by empirical or semiempirical models, which are horizontal to uncertainty. In this work, a data-driven method based on artificial neural networks has been implemented to study the heat transfer behavior of a subcooled boiling model. The proposed method considers the near local flow behavior to predict wall temperature and void fraction of a subcooled minichannel. The input of the network consists of pressure gradients, momentum convection, energy convection, turbulent viscosity, liquid and gas velocities, and surface information. The outputs of the models are based on the quantities of interest in a boiling system wall temperature and void fraction. To train the network, high-fidelity simulations based on the Eulerian two-fluid approach are carried out for varying heat flux and inlet velocity in the minichannel. Two classes of the deep learning model have been investigated for this work. The first one focuses on predicting the deterministic value of the quantities of interest. The second one focuses on predicting the uncertainty present in the deep learning model while estimating the quantities of interest. Deep ensemble and Monte Carlo Dropout methods are close representatives of maximum likelihood and Bayesian inference approach respectively, and they are used to derive the uncertainty present in the model. The results of this study prove that the models used here are capable of predicting the quantities of interest accurately and are capable of estimating the uncertainty present. The models are capable of accurately reproducing the physics on unseen data and show the degree of uncertainty when there is a shift of physics in the boiling regime.

Place, publisher, year, edition, pages
MDPI, 2020
Keywords
computational fluid dynamics (CFD), artificial neural network (ANN), subcooled boiling flows, uncertainty quantification (UQ), Monte Carlo dropout, deep ensemble, deep neural network (DNN)
National Category
Energy Engineering
Identifiers
urn:nbn:se:mdh:diva-52858 (URN)10.3390/en13225987 (DOI)000594195200001 ()2-s2.0-85106615320 (Scopus ID)
Available from: 2020-12-17 Created: 2020-12-17 Last updated: 2023-11-29Bibliographically approved
3. PREDICTION OF THE CRITICAL HEAT FLUX USING PARAMETRIC GAUSSIAN PROCESS REGRESSION
Open this publication in new window or tab >>PREDICTION OF THE CRITICAL HEAT FLUX USING PARAMETRIC GAUSSIAN PROCESS REGRESSION
Show others...
2021 (English)In: Proceedings of the 15th International Conference on Heat Transfer, Fluid Mechanics andThermodynamics (HEFAT2021), HEFAT , 2021, p. 1865-1870Conference paper, Published paper (Refereed)
Abstract [en]

A sound understanding of the critical heat flux is of prime importance for any industrial boiling system design and safety. From the literature, the majority of the critical heat flux studies are based on empirical knowledge, often supported by ex- perimental investigations which are performed under specific conditions difficult to be generalized. Consequently, most of the available correlations have ±30% predictive error when com- pared to measurement data. Hence, accurate prediction of this quantity remains an open challenge for the thermal engineering community. The present study aims to investigate the hidden features that exist in experimental data using a machine learning technique. Firstly, a literature survey is carried out to collect experimental data for boiling flows in tubes under low pressure and low flow conditions. These experimental data consist of the following parameters: system pressure, mass flux, characteristic dimensions, thermodynamic quality, inlet subcooling, and critical heat flux. A parametric Gaussian process regression model is used to predict the critical heat flux. The prediction obtained from the model is then compared with experimental measurements and the values obtained from the critical heat flux look-up table. The model used in this study is capable of predicting the critical heat flux with better accuracy along with the information of prediction uncertainty. Moreover, it provides insights on the relevance of the different input parameters to the prediction of the critical heat flux and aligns well with the underlying physics. The model used in this study shows a good level of robustness which can be further extended for other geometries, datasets, and operating conditions. 

Place, publisher, year, edition, pages
HEFAT, 2021
Keywords
Boiling flows, Critical Heat Flux, parametric Gaussian process, Machine Learning, Heat Transfer
National Category
Engineering and Technology Energy Engineering
Research subject
Energy- and Environmental Engineering; Energy- and Environmental Engineering
Identifiers
urn:nbn:se:mdh:diva-55650 (URN)978-1-77592-216-2 (ISBN)
Conference
THE 15th INTERNATIONAL CONFERENCE ON HEAT TRANSFER, FLUID MECHANICS AND THERMODYNAMICS, HEFAT, Virtual Conference, 26-28 July 2021
Available from: 2021-08-26 Created: 2021-08-26 Last updated: 2023-11-29Bibliographically approved
4. Inverse Flow Prediction Using Pinns In An Enclosure Containing Heat Sources
Open this publication in new window or tab >>Inverse Flow Prediction Using Pinns In An Enclosure Containing Heat Sources
2023 (English)In: Proc. Thermal Fluids Eng. Summer Conf., Begell House Inc. , 2023, p. 429-438Conference paper, Published paper (Refereed)
Abstract [en]

While simulating heat transfer problems using a numerical method, the thermal boundary conditions are never known precisely, which leads to ill-posed boundary conditions for the energy equation. The lack of knowledge of accurate thermal boundary conditions in a practical application makes it impossible to solve this problem no matter how sophisticated the conventional numerical method is. Hence, the current work addresses this ill-posed problem using physics informed neural network by assuming that the thermal boundary near the source is unknown and only a few measurements of temperature are known in the domain. Physics-informed neural network is employed to represent the velocity and temperature fields, while simultaneously enforcing the Navier-Stokes and energy equations at random points in the domain. This work serves as an inverse problem since the goal here is to reproduce the global flow field and temperature profile in the domain with few measurement data points. Furthermore, the work focuses on using transfer learning for different parameters such as the position and size of the source term inside the enclosure domain. These parameters are of particular interest while designing a thermal system and being able to predict the flow and thermal behaviour instantly will allow for better design of the system. For this study, the sensors' data are extracted from numerical simulation results. The placement of the sensors in the domain plays a vital role in accuracy hence, sensors were optimized using the residual of the energy equation. The results obtained from this work demonstrate that the proposed method is in good agreement with the underlying physics represented by the numerical results.

Place, publisher, year, edition, pages
Begell House Inc., 2023
Keywords
Convection, Heat Transfer, Machine learning, Physics informed neural network, Transfer Learning
National Category
Energy Engineering
Identifiers
urn:nbn:se:mdh:diva-64443 (URN)2-s2.0-85171300317 (Scopus ID)
Conference
Proceedings of the Thermal and Fluids Engineering Summer Conference
Available from: 2023-10-05 Created: 2023-10-05 Last updated: 2023-11-29Bibliographically approved
5. Application of deep learning for segmentation of bubble dynamics in subcooled boiling
Open this publication in new window or tab >>Application of deep learning for segmentation of bubble dynamics in subcooled boiling
Show others...
2023 (English)In: International Journal of Multiphase Flow, ISSN 0301-9322, E-ISSN 1879-3533, Vol. 169, article id 104589Article in journal (Refereed) Published
Abstract [en]

The present work focuses on designing a robust deep-learning model to track bubble dynamics in a vertical rectangular mini-channel. The rectangular mini-channel is heated from one side with a constant heat flux, resulting in the creation of bubbles. Images of the bubbles are recorded using a high-speed camera, which serve as the input data for the deep learning model. The raw image data acquired from the high-speed camera is inherently noisy due to the presence of shadows, reflections, background noise, and chaotic bubbles. The objective is to extract the mask of the bubble given all these challenging factors. Transfer learning is adopted to eliminate the need for a large dataset to train the deep learning model and also to reduce computational costs. The trained model is then validated against the validation datasets, demonstrating an accuracy of 98% while detecting the bubbles. The model is then evaluated on different experimental conditions, such as lighting, background, and blurry images with noise. The model demonstrates high robustness to different conditions and is able to detect the edges of the bubbles and classify them accurately. Moreover, the model achieves an average intersection over union of 85%, indicating a high level of accuracy in predicting the masks of the bubbles. The method enables accurate recognition and tracking of individual bubble dynamics, capturing their coalescence, oscillation, and collisions to estimate local parameters by proving the bubble masks. This allows for a comprehensive understanding of their spatial-temporal behaviour, including the estimation of local Reynolds numbers.

National Category
Energy Engineering
Identifiers
urn:nbn:se:mdh:diva-64495 (URN)10.1016/j.ijmultiphaseflow.2023.104589 (DOI)001070346100001 ()2-s2.0-85172483742 (Scopus ID)
Available from: 2023-10-11 Created: 2023-10-11 Last updated: 2023-11-29Bibliographically approved
6. Inverse flow prediction using ensemble PINNs and uncertainty quantification
Open this publication in new window or tab >>Inverse flow prediction using ensemble PINNs and uncertainty quantification
(English)Manuscript (preprint) (Other academic)
Abstract [en]

The thermal boundary conditions in a numerical simulation for heat transfer are often imprecise. This leads to poorly defined boundary conditions for the energy equation. The lack of accurate thermal boundary conditions in real-world applications makes it impossible to effectively solve the problem, regardless of the advancement of conventional numerical methods. 

This study utilises a physics-informed neural network to tackle ill-posed problems for unknown thermal boundaries with limited sensor data. The network approximates velocity and temperature fields while complying with the Navier-Stokes and energy equations, thereby revealing unknown thermal boundaries and reconstructing the flow field around a square cylinder. The method relies on optimal sensor placement determined by the QR pivoting technique, which ensures the effective capture of the dynamics, leading to enhanced model accuracy. In an effort to increase the robustness and generalisability, an ensemble physics-informed neural network is implemented. This approach mitigates the risks of overfitting and underfitting while providing a measure of model confidence. As a result, the ensemble model can identify regions of reliable prediction and potential inaccuracies. Therefore, broadening its applicability in tackling complex heat transfer problems with unknown boundary conditions.

Keywords
Heat transfer, mixed convection, physics informed neural network, optimal sensor placement, transient simulation, inverse method
National Category
Engineering and Technology Computational Mathematics
Research subject
Energy- and Environmental Engineering
Identifiers
urn:nbn:se:mdh:diva-64897 (URN)
Available from: 2023-11-29 Created: 2023-11-29 Last updated: 2023-12-11Bibliographically approved
7. Investigation of Machine Learning Regression Techniques to Predict Critical Heat Flux over a Large Parameter Space
Open this publication in new window or tab >>Investigation of Machine Learning Regression Techniques to Predict Critical Heat Flux over a Large Parameter Space
2023 (English)In: Proceedings 20th International Topical Meeting on Nuclear Reactor Thermal Hydraulics (NURETH-20) Washington, D.C., August 20-25, 2023, 2023, p. 4516-4529Conference paper, Published paper (Refereed)
Abstract [en]

A unifying and accurate model to predict Critical Heat Flux (CHF) over a wide range of conditions has been elusive since wall boiling research emerged. With the release of the data utilized in the development of the 2006 Groeneveld CHF lookup table (LUT), by far the most extensive public CHF database available to date (nearly 25000 data points), development of data-driven predictions models over a large parameter space in simple geometry (vertical, uniformly heated round tubes) can be performed. Furthermore, the popularization of machine learning techniques to solve regression problems has led to more advanced tools for analyzing large and complex databases. This work compares three machine learning algorithms to predict the entire LUT CHF test database. For each selected regression algorithm (ν-Support vector, Gaussian process, and neural network), an optimized hyperparameter set is fitted. The best-performing algorithm is the neural network, which can achieve a standard deviation of the predicted/measured factor of 12.3%, three times lower than the LUT. In comparison, the Gaussian process regression and the ν-Support vector regression achieve a standard deviation of 17.7%, about two times lower than the LUT. All considered algorithms hence significantly outperform the LUT prediction performance. The neural network model and training methodology are designed to prevent overfitting, which is confirmed by data analysis of the predictions. Finally, a feasibility study of transfer learning is presented and future development directions (including uncertainty quantification) are discussed. 

National Category
Engineering and Technology Energy Engineering
Research subject
Energy- and Environmental Engineering
Identifiers
urn:nbn:se:mdh:diva-64896 (URN)
Conference
20th International Topical Meeting on Nuclear Reactor Thermal Hydraulics (NURETH-20), Washington, D.C., August 20-25, 2023
Available from: 2023-11-29 Created: 2023-11-29 Last updated: 2023-12-11Bibliographically approved
8. Comparison of machine learning approaches for spectroscopy applications
Open this publication in new window or tab >>Comparison of machine learning approaches for spectroscopy applications
2022 (English)In: Proceedings of the 63rd International Conference of Scandinavian Simulation Society / [ed] Lars O. Nord; Tiina Komulainen; Corinna Netzer; Gaurav Mirlekar; Berthe Dongmo-Engeland; Lars Eriksson, 2022, p. 80-85Conference paper, Published paper (Refereed)
Abstract [en]

In energy production the characterization of the fuel is a key aspect for modelling and optimizing the operation of a power plant. Near-infrared spectroscopy is a wellestablished method for characterization of different fuels and is widely used both in laboratory environments and in power plants for real-time results. It can provide a fast and accurate estimate of key parameters of the fuel, which for the case of biomass can include moisture content, heating value, and ash content. These instruments provide a chemical fingerprint of the samples and require a calibration model to relate that to the parameters of interest.

A near-infrared spectrometer can provide point data whereas a hyperspectral imaging camera allows the simultaneous acquisition of spatial and spectral information from an object. As a result, an installation above a conveyor belt can provide a distribution of the spectral data on a plane. This results in a large amount of data that is difficult to handle with traditional statistical analysis. Furthermore, storage of the data becomes a key issue, therefore a model to predict the parameters of interest should be able to be updated continuously in an automated way. This makes hyperspectral imaging data a prime candidate for the application of machine learning techniques. This paper discusses the modelling approach for hyperspectral imaging, focusing on data analysis and assessment of machine learning approaches for the development of calibration models.

Series
Linköping Electronic Conference Proceedings, ISSN 1650-3686, E-ISSN 1650-3740 ; 192
Keywords
machine learning, calibration models, hyperspectral imaging, near-infrared spectroscopy
National Category
Energy Engineering
Identifiers
urn:nbn:se:mdh:diva-61125 (URN)10.3384/ecp192012 (DOI)978-91-7929-545-5 (ISBN)
Conference
Proceedings of the 63rd International Conference of Scandinavian Simulation Society, SIMS 2022, Trondheim, Norway, September 20-21, 2022
Available from: 2022-12-06 Created: 2022-12-06 Last updated: 2023-11-29Bibliographically approved

Open Access in DiVA

fulltext(62613 kB)80 downloads
File information
File name FULLTEXT02.pdfFile size 62613 kBChecksum SHA-512
0304ecc8903fcfdcc1106848a90da5dc377c10e1fe5ac164e4b7c7a854817ca9e99cbfc4d8a44a5efff9c0e0ddd681a92222c60e20c50b989735eacff11a991e
Type fulltextMimetype application/pdf

Authority records

Soibam, Jerol

Search in DiVA

By author/editor
Soibam, Jerol
By organisation
School of Business, Society and Engineering
Engineering and TechnologyEnergy Engineering

Search outside of DiVA

GoogleGoogle Scholar
Total: 80 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

isbn
urn-nbn

Altmetric score

isbn
urn-nbn
Total: 1118 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf