Defense of the dissertation of Dosýmbekov Erlan Kadyrovich for the degree of Doctor of Philosophy (PhD) in the specialty «6D070400 - Есептеу техникасы және бағдарламалық қамтамасыз ету»
L.N. Gumilyov Eurasian National University, a dissertation defense for the degree of Doctor of Philosophy (PhD) by Dosýmbekov Erlan Kadyrovich on the topic «Design of reconfigurable intelligent blocks for performance analysis and prediction of engineering systems» by specialty «6D070400 - Computer Engineering and Software».
The dissertation was carried out at the «Сomputer and software engineering education department» of L.N. Gumilyov Eurasian National University.
The language of defense is kazakh
Official reviewers:
Kozbakova Ainur - Doctor of Philosophy (PhD), Associate Professor, Leading researcher at the Laboratory of Artificial Intelligence and Robotics of the RSE Institute of Information and Computing Technologies of the CS MSHE RK (Almaty, Republic of Kazakhstan).
Kuttubai Nurzhigit Bakytuly- Doctor of Philosophy (PhD), Department of Electronics and Astrophysics, Al-Farabi Kazakh national university (Almaty, Republic of Kazakhstan).
Temporary members of the Dissertation Committee:
- Barakhnin Vladimir Borisovich- doctor of Technical Sciences, Head of the Department of Mathematical Modeling, Faculty of Mechanics and Mathematics, Novosibirsk State University (Novosibirsk, Russian Federation).
- Merembaev Timur Zhumakanovich- doctor of Philosophy (PhD), Leading Researcher, RSE "Institute of Information and Computational Technologies" (Almaty, Republic of Kazakhstan).
- Mukazhanov Nurzhan Kakenovich- doctor of Philosophy (PhD), Associate Professor, Department of Software Engineering, Satbayev University (Almaty, Republic of Kazakhstan).
Scientific advisors:
Matkarimov Bakhyt Turganbayevich - doctor of Technical Sciences, Professor of the Department of Artificial Intelligence Technologies, L.N. Gumilyov Eurasian National University (Astana, Republic of Kazakhstan).
Paltashev Timur Tursunovich- doctor of Technical Sciences, Professor, Advanced Micro Devices, Member of the Artificial Intelligence Group (Santa Clara, California, USA)
Zyubin Vladimir Evgenievich- doctor of Technical Sciences, Associate Professor, Institute of Automation and Electrometry, Siberian Branch of the Russian Academy of Sciences (Novosibirsk, Russia)
The defense will take place on August 12, 2025, at 04:00 PM in the Dissertation Council for the training direction «8D061 - Information and communication technologies» in the specialty «6D070400 - Есептеу техникасы және бағдарламалық қамтамасыз ету» of L.N. Gumilyov Eurasian National University. The defense meeting is planned to be held online.
Link: https://clck.ru/3MyKPg
Address: Astana, A. Pushkin Street, 11, Educational Building 2, Room 222.
Abstract (English): ANNOTATION dissertation work of Dossumbekov Yerlan “Designing reconfigurable intelligent blocks for analyzing and predicting performance of engineering systems”, submitted for the degree of Doctor of Philosophy (PhD) in the specialty “6D070400 - Computer Engineering and Software” Relevance of the research topic. The State Program of Industrial and Innovative Development of the Republic of Kazakhstan is aimed at supporting and developing digital and intelligent technologies in industry. The focus is on improving the productivity, energy efficiency, and sustainability of domestic enterprises. The program promotes initiatives to introduce digital solutions, such as automation, robotics, and artificial intelligence (AI) systems, which allow for improved control and analysis of production processes. According to the legislation of the Republic of Kazakhstan, one of the goals of the industrial sector development is introduction of innovative technologies to increase the efficiency and competitiveness of production systems. This is reflected in such regulations as the state program of industrial and innovative development of the Republic of Kazakhstan, the energy strategy of the Republic of Kazakhstan for 2050. As international experience shows, the concept of designing reconfigurable intelligent blocks (FPGA) and artificial intelligence-based systems is actively supported and applied in industry. In particular, in the EU and the USA, intelligent FPGA and AI-based systems are being implemented in areas such as energy, heat transfer, and production process management to improve forecasting accuracy and energy conservation. EU countries are conducting research and implementing intelligent control technologies within the framework of the Horizon Europe program (the European Union Research and Innovation program), which aims to create high-tech solutions for the industrial sector. One of them is climate, energy and mobility. The energy dimension of this cluster is concentrated in the field of energy storage. International organizations such as the Institute of Electrical and Electronics Engineers (IEEE) and the International Electrotechnical Commission (IEC) are developing standards and recommendations aimed at implementing reconfigurable and intelligent systems for engineering solutions. These standards support the use of intelligent blocks and algorithms based on neural networks to improve the reliability of engineering systems. Modern research demonstrates that the design of reconfigurable systems, especially using FPGA-based neural networks, allows achieving high results in data analysis and forecasting. Heat exchangers (HE) play an important role in engineering systems, providing efficient heat transfer between different working media such as liquids and gases. They are used in various sectors and systems, including the food industry, oil refining processes, renewable energy systems, waste heat recovery, etc. Depending on the application and the specified requirements, the size and configuration of the heat exchanger may vary. Several factors affect the performance of the heat exchanger, including the type of heat exchanger used, the materials of the various components, the thermal characteristics of the flows, the operating conditions, etc. Improving maintenance performance is one of the most serious challenges for engineers, designers, and scientists seeking to create compact and reliable systems. In this regard, various methods and approaches have been proposed and tested in recent decades to assess their impact on enhanced heat transfer. Entering additional components such as fins, baffles, and vortex generators, modification of HE components, the use of liquids with improved properties, such as nanofluids, and optimization of size, configuration, and operating conditions are among the most common approaches and methods proposed to enhance heat transfer in these devices. Micro-heat exchangers, with significantly smaller sizes compared to traditional ones, have attracted attention in recent decades. For example, in the research of Saktivel et al., various nanoparticles, including Al2O3 (aluminum oxide), TiO2 (titanium dioxide), ZrO2(zirconium dioxide) and SiO2 (silicon oxide) were used in the working fluid of a shell-and-tube heat exchanger. They reported improved thermal performance when using nanofluids and noted better results when using TiO2 and Al2O3 compared to ZrO2 and SiO2. Javani et al. evaluated the impact of using nanofluids in a borehole heat exchanger and reported a decrease in thermal resistance and an increase in pressure drop compared to using pure water as a coolant. The size, shape, concentration, and type of nanoparticles, as well as the type of base fluid, play key roles in improving heat transfer using nanofluids. Zheng et al. evaluated various nanofluids for use in a plate heat exchanger. The nanoparticles under consideration were Fe3O4(magnetite), Al2O3 (aluminum oxide), SiC (silicon carbide), and CuO (copper oxide). The best thermal performance was observed for Fe3O4 nanofluids. Compared with deionized water, the convective heat transfer coefficient increased by 21.9% when using Fe3O4 /water with a concentration of 1% by weight. The idea of using nanofluids to improve thermal performance can be developed for micro heat exchangers. To facilitate the assessment of various factors affecting the performance characteristics of a heat exchanger, accurate modeling of these systems is of great importance, and it is important to propose appropriate models. In this regard, intelligent methods can be developed and used that have shown significant performance in modeling various thermal systems and tasks. Modeling the thermal properties of nanofluids as a heat carrier is one of the most common applications of these methods in the field of thermal engineering. Maleki et al. evaluated the accuracy of an artificial neural network (ANN) in estimating the thermal conductivity of nanofluids with silica particles (SiO₂, or silicon oxide) and various base fluids. The proposed ANN-based models are applicable for accurate modeling of thermal conductivity. Moreover, the type of learning function affects the accuracy of the estimated values. Other problems of thermal engineering, in addition to the properties of nanofluids, can be modeled using intelligent approaches. Abidi et al. used ANN methods to predict the performance of a nanofluid solar collector. Compared to traditional nonlinear regression, intelligent methods can be more accurate in modeling various characteristics of heat exchangers due to their ability to capture the underlying non-linearity of the data. Researchers have widely used various intelligent methods to simulate the performance of various heat exchangers. For example, Huang et al. used the support vector machine (SVM) method to evaluate the performance of a heat exchanger under cryogenic oscillating flow conditions. They applied standard SVM and SVM with the leave-one-out method and reported that the latter provides higher accuracy with a maximum error of 12.4%. In another research, Islamoglu used an artificial neural network (ANN) to predict the rate of heat transfer in a wire-on-pipe heat exchanger. The average absolute relative error of the proposed forecasting model was less than 3%. Wang et al. used ANN to estimate the heat transfer rate in shell-and-tube heat exchangers with continuous spiral baffles or segmented baffles. There are some limitations in using traditional heat exchangers, such as the size of the equipment and the required installation space. The development of heat exchangers with more compact dimensions and mini or micro sizes will be an effective solution to overcome these problems. Using traditional methods for heat exchanger modeling, such as computational fluid dynamics (CFD), will be time-consuming and computationally expensive. In this regard, it is advisable to use alternative approaches and methods with fast performance and lower computational costs. In addition to the computational cost, the accuracy of the predictive model is of significant importance, which must be taken into account when developing the model. Thus, the relevance of the thesis is determined by the need to develop intelligent methods for assessing the thermal characteristics of heat exchangers with high accuracy and relatively low computational costs. Using the proposed models based on intelligent methods, it is possible to optimize the design of heat exchangers, considering the applied input data in the model as variables of optimization algorithms and to design heat exchangers with better performance. Goal of dissertation research. The goal of the research is to develop models based on intelligent methods for estimating the Nusselt number in a microheater using a hybrid TiO2 nanofluid (titanium dioxide) and ZnO nanoparticles (zinc oxide), as well as to implement intelligent predictors for a heat exchanger on reconfigurable FPGA and to study reconfigurable intelligent blocks for analyzing and predicting the performance of engineering systems using neural networks aimed at to optimize the management of thermal processes. To achieve this goal, it was necessary to solve the following tasks: - study of the theoretical foundations, methods, and models of an artificial neural network (ANN) for accurate modeling of thermal conductivity and predicting the performance of various nanofluid heat exchangers; - perform a sensitivity analysis to assess the importance of the input data under consideration in the models; - comparison between the model output and the data presented in the experimental study; - development of an ANN model based on a multilayer perceptron (MLP) and neural networks of the group method of data handling (GMDH) for analyzing the performance of a microplate heat exchanger; - implementation and realization of intelligent predictors for heat exchangers on reconfigurable FPGAs; - to analyze the current state and trends in the field of reconfigurable technologies and their application in engineering systems; - development of architecture of an intelligent FPGA-based block that ensures efficient execution of analysis and forecasting algorithms; - research of data processing methods and their integration with neural networks to create adaptive control systems; The scientific novelty of the work. In this paper, two types of artificial neural network are used, namely the Group Method of Data Handling (GMDH) and multilayer perceptron (MLP), to model the Nusselt number of a microplate heat exchanger using a hybrid nanofluid as a new example. In addition, sensitivity analysis is performed to assess the importance and level of influence of the three variables considered in the experimental work, which is the novelty of this work. The performance of the proposed models is compared taking into account various statistical criteria such as R2 (coefficient of determination), average absolute relative deviation (AARD) and mean square error (MSE). The main aspects of the novelty include: integration of FPGA and neural networks; adaptive control algorithms; method of group consideration of arguments; experimental verification: Conducting experimental studies confirming the effectiveness of the proposed solutions using the example of microplate heat exchangers. The experimental results show a significant improvement in system performance and reliability compared to traditional control methods. Thus, this research contributes to development of engineering systems management technologies by offering new approaches to data analysis and forecasting that can be successfully applied in various industries. The main provisions submitted for defense: 1. The hardware architecture of the MLP neural network based on FPGA has been developed, which provides accelerated data processing by performing operations in parallel and reducing latency, which increases the overall performance of the system. 2. Each layer of the neural network is implemented using FPGA logic blocks, which allows flexible adaptation of the architecture to specific engineering system tasks and efficient use of hardware platform resources. 3. The process of data analysis and modeling using GMDH Shell software has been automated, which ensures high forecast accuracy and minimizes the influence of the human factor when building models. 4. A comparative analysis of the proposed intelligent system with traditional solutions has been carried out, confirming its superiority in terms of forecasting accuracy, response time, and adaptive control. The object of the research is artificial neural networks and reconfigurable intelligent blocks used to analyze and predict the performance of engineering systems, in particular, in the context of thermal process control. Based on urgency of the task, the subject of the study is three input parameters, namely the height of the plate, concentration of nanomaterials and the Reynolds number, for creating a model based on a multilayer perceptron neural network (MLP) and the method of group consideration of arguments (GMDH), hybrid nanofluids titanium dioxide TiO2 and zinc oxide ZnO, Nusselt numbers. As well as methods and algorithms implemented on the basis of multilayer perceptron’s (MLP) and other neural networks that are used for data processing and forecasting in engineering systems such as brazed microplate heat exchangers (MPHE). In particular, the following aspects are being investigated: architecture of reconfigurable FPGA-based intelligent blocks and their capabilities for performing complex data analysis algorithms; machine learning methods, in particular, neural networks used to identify dependencies and patterns in data, as well as to predict performance of engineering systems; processes of adaptation and training of neural networks, allowing to improve accuracy and effectiveness of predictions in changing operating conditions; impact of applied intelligent methods on performance and reliability of engineering systems, as well as on optimization of thermal processes. Research methods include theoretical and experimental research, mathematical and computer modeling, algorithmization, and hardware device development in the VHDL hardware description language. The following research methods are also used in this work, providing an integrated approach to development and analysis of reconfigurable smart blocks: 1. Analytical method: Analysis of current state and trends in the field of reconfigurable technologies, neural networks and their application in management of engineering systems. This method allows us to identify the key problems and needs of the industry, as well as to justify the choice of areas for further research. 2. Experimental method: Used to verify the proposed solutions. The experimental research includes development of prototypes of intelligent units and their testing on real engineering systems such as microplate heat exchangers. A comparative performance analysis before and after implementation allows you to evaluate the effectiveness of developed solutions. 3. Modeling: The use of mathematical and computer modeling methods to create virtual models of engineering systems and analyze their behavior in various operating modes. Modeling allows to test various control and forecasting algorithms, as well as evaluate their impact on performance. 4. Machine learning methods: Used for development and training of multilayer perceptron’s (MLP) and other neural networks. The use of machine learning methods makes it possible to identify complex dependencies and patterns in data, which is the basis for creating adaptive control systems. 5. Method of group accounting of arguments is being implemented for data analysis and processing, which allows taking into account multiple sources of information and their interrelationships. This method helps to improve accuracy of forecasting and resilience of systems to noise. The use of these methods allows for a comprehensive approach to research, which contributes to a deeper understanding of the problem and development of effective solutions for managing thermal processes in engineering systems. Practical significance of the work: The practical significance of this work lies in development and implementation of reconfigurable intelligent units that enhance the efficiency of thermal process control in engineering systems. The main aspects of practical significance include: Improving the efficiency of production processes. Implementation of the proposed solutions allows for significant optimization of thermal process management, which leads to a reduction in energy costs and an increase in the overall performance of systems; Improving product quality. Intelligent units based on machine learning methods provide precise control of parameters affecting the quality of the final product, which is especially important in industries with high quality requirements; Reducing maintenance costs. Forecasting potential problems and automatic regulation of processes help reduce the number of unplanned downtimes and equipment maintenance costs, which helps save resources; Flexibility and adaptability of systems. The ability to quickly reconfigure intelligent units for new tasks allows enterprises to adapt to changes in the market and technology, which increases their competitiveness; wide range of applications: The developed solutions can be applied in various industries, including energy, oil and gas, industrial automation and HVAC (Heating, Ventilation, and Air Conditioning). This makes them universal tools for improving the efficiency of processes in a wide variety of areas; integration with existing systems. Smart blocks can be easily integrated into existing control and automation systems, which simplifies the process of their implementation and increases the cost of investment. Thus, the practical significance of the work lies in its contribution to development of management technologies that improve efficiency, reliability and quality of engineering systems, which is important for modern production and business in general. Approbation of dissertation results. The main results of the dissertation work were presented at seminars of L.N. Gumilyov Eurasian National University and at international conferences: 1. International Scientific and practical conference “Intelligent information and communication technologies - a means of realizing the third industrial revolution in the light of the strategy “Kazakhstan - 2050”. Astana, 2013; 2. II International Scientific and Practical Conference “Intelligent information and communication technologies - a means of implementing the third industrial revolution in the light of the Strategy “Kazakhstan - 2050”. L.N.Gumilyov ENU, Astana, 2014. Articles published in scientific journals recommended by the Committee for Control in the Sphere of Education and Science of the Ministry of Education and Science of the Republic of Kazakhstan: 1. Dossumbekov Ye.K. Characteristics and features of crystals of programmable logic integrated circuits FPGA. // VESTNIK, Issue: L.N.Gumilyov ENU, Astana, 2014. -pp. 425-429. 2. Dossumbekov Ye.K. How to Solve the Meta-Persistence Problem in Digital Systems Based on Software-Defined Logic Integrated Circuits. // VESTNIK, Issue: L.N.Gumilyov ENU, Astana, 2014. -pp. 180-185. 3. Dossumbekov Ye.K. The use of high-level synthesis in the design of reconfigurable IP blocks. // VESTNIK, Issue: L.N. Gumilyov ENU, Astana, 2014. -pp. 205-208. Publications based on the results of the study, including in scientific journals indexed in the Web of Science and Scopus databases: 1. Dossumbekov Ye.K., Nurkhat Zhakiyev, Mohammad Alhuyi Nazari, Mohamed Salem, Bekzat Abdikadyr. Sensitivity Analysis and Performance Prediction of a Micro Plate Heat Exchanger by Use of Intelligent Approaches. // International Journal of Thermofluids, Volume 22, May 2024, 100601, https://doi.org/10.1016/j.ijft.2024.100601. The act of implementing the developed intelligent block based on neural networks and programmable logic integrated circuits (author - Dossumbekov Ye.K.) testifies to its practical application in JDM Group LLP in order to analyze and predict the performance of engineering systems (implementation date - 14.02.2025). The volume and structure of the dissertation. The thesis consists of an introduction, three chapters, conclusion, and 2 appendices. The work was printed on 72 pages, using computer-based focusing capabilities in the form of illustrations, diagrams and tables. The list of references consists of 78 titles. The author expresses his sincere gratitude to Matkarimov Bakhyt Turganbayevich, Scientific supervisor, Doctor of Technical Sciences, Professor of the Department of Computer and Software Engineering at the L.N. Gumilyov Eurasian National University, as well as to foreign consultants, Doctor of Technical Sciences, Associate Professor Vladimir Zyubin (Institute of Automation and Electrometry SB RAS, Novosibirsk, Russia) and Doctor of Technical Sciences Paltashev Timur Tursunovich (Advanced Micro Devices, Santa Clara, California, USA) for their invaluable help, valuable recommendations and support provided during the research process.
'/%3e%3cpath%20fill-rule='evenodd'%20clip-rule='evenodd'%20d='M17.5427%2012.684C17.1383%2011.7889%2016.7126%2011.771%2016.3278%2011.7553C16.0132%2011.7419%2015.6529%2011.7427%2015.2932%2011.7427C14.9333%2011.7427%2014.3483%2011.8774%2013.8537%2012.4152C13.3587%2012.9531%2011.9639%2014.2533%2011.9639%2016.8978C11.9639%2019.5426%2013.8987%2022.098%2014.1684%2022.4569C14.4383%2022.8153%2017.9033%2028.4156%2023.3907%2030.5701C27.9513%2032.3606%2028.8793%2032.0045%2029.8692%2031.9149C30.8591%2031.8253%2033.0633%2030.6149%2033.5133%2029.3597C33.9632%2028.1048%2033.9632%2027.0291%2033.8282%2026.8043C33.6932%2026.5803%2033.3333%2026.4459%2032.7933%2026.1771C32.2534%2025.9083%2029.5992%2024.6078%2029.1043%2024.4286C28.6093%2024.2494%2028.2494%2024.1598%2027.8894%2024.6979C27.5295%2025.2355%2026.4954%2026.4459%2026.1804%2026.8043C25.8655%2027.1635%2025.5505%2027.2083%2025.0105%2026.9395C24.4706%2026.6699%2022.7318%2026.1029%2020.6691%2024.2717C19.0641%2022.8472%2017.9806%2021.0877%2017.6656%2020.5496C17.3506%2020.012%2017.6318%2019.7209%2017.9027%2019.4529C18.1451%2019.2121%2018.4426%2018.8254%2018.7126%2018.5116C18.982%2018.1977%2019.072%2017.9737%2019.252%2017.6153C19.4319%2017.2564%2019.342%2016.9425%2019.207%2016.6737C19.072%2016.405%2018.0228%2013.7468%2017.5427%2012.684Z'%20fill='white'/%3e%3cdefs%3e%3clinearGradient%20id='paint0_linear_181_11959'%20x1='22.8525'%20y1='42.8177'%20x2='22.8525'%20y2='0.921701'%20gradientUnits='userSpaceOnUse'%3e%3cstop%20stop-color='%2320B038'/%3e%3cstop%20offset='1'%20stop-color='%2360D66A'/%3e%3c/linearGradient%3e%3c/defs%3e%3c/svg%3e)
