Publications
Papers with collaborators on scientific machine learning, industrial systems, and computer vision.
Publication archive
15 papers
Separating Artificial-Viscosity Annealing from the L-BFGS Viscosity in Physics-Informed Neural Networks for Two-Phase Flow
Reducing smoothing during training is a common fix for sharp fronts, but it also changes the smoothing in the final stage. With both controlled, the final level carried the effect and no schedule advantage was detected.
Saltanbek Mukhambetzhanov, Yerzhan Kenzhebek, Samson Dawit Bekele, Saida Tastanova, Dagmawi Lemma, Timur Imankulov, Big Data and Cognitive Computing
Identifiability and Conditioning of Flux-Parameter Inversion for the Buckley–Leverett Equation
Physics alone cannot fix the viscosity ratio in a water-displaces-oil model. We show which measurements can, how much their errors are amplified, and why a good fitted field does not guarantee a good estimate.
Samson Dawit Bekele, Yerzhan Kenzhebek, Saltanbek Mukhambetzhanov, Saida Tastanova, Dagmawi Lemma, Timur Imankulov, Mathematics
Physics-Informed Neural Networks for Coupled Non-Isothermal Two-Phase Flow in Porous Media: Training Pathologies, Remedies, and the Role of Gradient Conflict
Conflict between training goals is a leading explanation for why physics-informed networks fail. On a hot-water flood with an exact solution, removing the conflict did not make training reliable, and two methods worked with it still present.
Yerzhan Kenzhebek, Timur Imankulov, Samson Dawit Bekele, Saltanbek Mukhambetzhanov, Machine Learning and Knowledge Extraction
Performance and stability of physics-informed Kolmogorov-Arnold networks for two-phase transport in porous media: A comparative study with physics-informed neural networks
Kolmogorov–Arnold networks promise accuracy with fewer parameters. On a sharp water front, ours used a third of the parameters but needed over twice the smoothing and about 50 times the training time.
Timur Imankulov, Yerzhan Kenzhebek, Samson Dawit Bekele, Darkhan Akhmed-Zaki, Results in Engineering
A Systematic Analysis of Physics-Informed Neural Networks for Two-Phase Flow with Capillarity: The Muskat–Leverett Problem
Capillary pressure couples pressure and water movement in rock, and most neural-network studies leave it out. We trained one network on the coupled equations and mapped what it needs; the water front, not pressure or loss, decided model choice.
Timur Imankulov, Alibek Kuljabekov, Samson Dawit Bekele, Zhumabek Zhantayev, Bakytzhan Assilbekov, Yerzhan Kenzhebek, Applied Sciences
Coupled pressure and saturation prediction for two-phase flow in porous media using physics-informed neural networks (PINNs)
One physics-informed network trained on both the pressure and water-saturation equations got pressure right but misplaced the water front. Bigger networks and reweighting didn’t fix it; a small diffusion term with tuned weights did.
Yerzhan Kenzhebek, Timur Imankulov, Samson Dawit Bekele, Irina Panfilova, Darkhan Akhmed-Zaki, Computer Methods in Applied Mechanics and Engineering
Predicting oil recovery factor using dense and cascade neural networks
Cascade networks add direct connections from the inputs to every later layer, and an earlier chemical-flooding study ranked them most accurate. On 163,726 noisy simulated samples, a cascade network and a dense network scored almost the same.
Samson Dawit Bekele, Serik Aibagarov, Yerzhan Kenzhebek, Timur Imankulov, B. Matkerim, Bulletin of the National Engineering Academy of the Republic of Kazakhstan
Enhancing Oil Recovery Predictions by Leveraging Polymer Flooding Simulations and Machine Learning Models on a Large-Scale Synthetic Dataset
Proxy models can stand in for slow polymer-flooding simulations. We generated 163,726 simulated samples and compared nine methods; a cubic polynomial matched the neural networks.
Timur Imankulov, Yerzhan Kenzhebek, Samson Dawit Bekele, Erlan Makhmut, Energies
Development of an Intelligent Video Surveillance System Based on YOLO Algorithm
A detector finds people in single frames; an office needs events such as someone entering a closed lab. We built a web-based system that turns live detections into occupancy, line-crossing, and face-recognition alerts on a consumer GPU.
Nurtugan Azatbekuly, Aksultan Mukhanbet, Samson Dawit Bekele, 2024 IEEE 4th International Conference on Smart Information Systems and Technologies (SIST)
On the Effectiveness of Kolmogorov—Arnold Networks for Enhanced Oil Recovery Prediction in Polymer Flooding
Neural networks predict oil recovery well but are hard to inspect. A Kolmogorov–Arnold network with 1,885 parameters came close to a 43,265-parameter dense network and showed how it used each input.
Samson Dawit Bekele, Yerzhan Kenzhebek, Timur Imankulov, 2024 7th International Conference on Algorithms, Computing and Artificial Intelligence (ACAI)
Intelligent Control System for Efficient Energy Management in the Oil and Gas Industry
At oil and gas sites, heating and ventilation protect equipment and clear hazardous gas as well as keeping people comfortable. We built a test platform that puts sensing, control, gas alarms, and an operator interface in one loop.
Nurdaulet Tasmurzayev, Bibars Amangeldy, Yedil Nurakhov, Shona Shinassylov, Samson Dawit Bekele, 2023 IEEE Asia-Pacific Conference on Computer Science and Data Engineering (CSDE)
Additional publications
Development of a Comprehensive Software Solution for Processing High-Sulfur, Copper-Poor Concentrates in the Copper Smelting Industry (opens in a new tab)
Software for estimating the masses and compositions of copper-smelting products from feed mixtures.
B. K. Kenzhaliyev, S. Zh. Aibagarov, Samson Dawit Bekele, S. A. Kvyatkovskiy, M. A. Dyussebekova, Bulletin of Shakarim University. Technical Sciences
Development and Evaluation of an Intelligent Control System for Sustainable and Efficient Energy Management (opens in a new tab)
A system connecting HVAC equipment, environmental sensors, live monitoring, and prediction software.
Bibars Amangeldy, Nurdaulet Tasmurzayev, Yedil Nurakhov, Shona Shinassylov, Samson Dawit Bekele, WSEAS Transactions on Electronics
Development of an Intelligent Oil Field Management System based on Digital Twin and Machine Learning (opens in a new tab)
A laboratory oil-displacement setup connected to live monitoring, a digital twin, and a sensor-based prediction workflow.
Nurdaulet Tasmurzayev, Bibars Amangeldy, Yedil Nurakhov, Shona Shinassylov, Samson Dawit Bekele, WSEAS Transactions on Electronics
Smart control system for optimal energy use in the oil and gas sector (opens in a new tab)
An HVAC prototype linking sensors, control hardware, and operator software over wired and wireless networks.
Nurdaulet Tasmurzayev, Shona Shinassylov, Bibars Amangeldy, Yedil Nurakhov, Samson Dawit Bekele, Challenges of Science