Outputs Show only Author
Title
Category
Date
Authors
Projects
Transient Modeling of a Solid Oxide Fuel Cell using an Efficient Deep Learning HY-CNN-NARX Paradigm Control and monitoring systems are crucial for ensuring optimal performance, efficiency, and longevity of Solid Oxide Fuel Cells (SOFC). Developing a model to accurately capture the transient behavior of the SOFC under dynamic operation is essential for these applications. As electrochemical, mass-transfer, and thermal equations are involved in SOFCs’ dynamic, physics-based approaches to capture all the complexity of SOFC systems are often too computationally expensive for real-time implementation. In this paper, an innovative multi-input neural network is developed to build an accurate model of SOFC that is computationally efficient. The proposed algorithm uses experimental data and is a modified nonlinear autoregressive exogenous (NARX) network. An optimal sequence of the most recent observations (output voltages) that characterize the SOFC dynamics is passed through stacked 1D convolutional layers (CNN) to extract latent spatial/temporal information. Then the output is concatenated with current and past exogenous inputs. The fusion of learned features, representing the internal dynamics and the effect of exogenous inputs, is then fed to a fully connected network for one step ahead performance prediction. This hybrid structure, called HY-CNN-NARX, is capable of identifying the transient dynamics of the SOFCs by unrolling information from historical outputs and inputs within a feedforward framework. To evaluate the model performance, lab-scale tubular SOFCs are experimentally tested at 650–750 ◦C under various dynamic operations and initial conditions, and the transient and steady-state responses of the SOFC are collected. Once the model is validated using a wide operating range of data, the generalization and prediction performance of the proposed network is compared with a conventional NARX and a recurrent stacked Long short-term memory (LSTM) model. The comparative results demonstrate that the proposed HY-CNN-NARX model outperforms the NARX model on unseen data with an accuracy improvement of 59.2% for root mean square error and 53.7% for mean absolute percentage error. In addition, the prediction performance of the HY-CNN-NARX is comparable to its recurrent counterpart, the LSTM model, but with a 39.3% faster execution time.T06-A02 University of Alberta Publication 2024-04-14 T06-A02 Developing a Time-Efficient Model for Solid Oxide Fuel Cells Using Self-Supervised Convolutional Autoencoder and Stateful LSTM Network The development of an efficient dynamic model for solid oxide fuel cell (SOFC) systems is essential for control strategies, process optimization, and fault diagnostics - key steps toward maximizing power generation and extending their lifetime. Long-Short-Term-Memory (LSTM), a potent type of
recurrent neural network, has proven to be adept at modeling intricate physical dynamic systems using extensive input-output data. They are particularly well-suited for capturing longrange
dependencies and temporal patterns. Nevertheless, due to memory and computing constraints in practice, training an LSTM network to effectively learn long-term dependency is challenging, which stems from the sequential processing nature of LSTMs. In this study, a self-supervised convolutional
autoencoder (AE) is developed to learn a concise yet informative temporal representation of historical input data. These compressed hidden sequences are then input into an LSTM model,
which is trained using a truncated backpropagation through time (TBPTT) algorithm. Once trained, an identical stateful LSTM model is reconstructed using the learned parameters, enabling the LSTM to predict the voltage using only the most recent past input data, rather than a sequence of data over
which the model was trained. The efficacy of the proposed framework is validated utilizing diverse experimental data collected from a lab-scale SOFC. The results indicate while the model properly identifies the underlying dynamics of the SOFC, it significantly reduces the training and runtime prediction computation costs of a traditional LSTM model by 38% and 42% respectively. This enhancement holds promise for improving real-time control and diagnostics of SOFC systems,
ultimately contributing to their improved performance and reliability.T06-A02 University of Alberta Publication 2024-09-05 T06-A02 Control-oriented Modeling of a Solid Oxide Fuel Cell Affected by Redox Cycling using a Novel Deep Learning Approach T06-A02 University of Alberta Publication 2024-09-10 T06-A02 Performance Prediction of a Range of Diverse Solid Oxide Fuel Cells using Deep Learning and Principal Component Analysis T06-A02 University of Alberta Publication 2025-01-23 Zeynab Salehi, Mohamadali Tofigh, Sajad Vafaeenezhad, Charles Robert Koch,
Shahbakhti, M. T06-A02 SOFC Database Three tubular cells with different design parameters are fabricated, and each is tested under 18 different operating conditions. Moreover, a planar cell is tested under 10 operating conditions. Tubular and planar cells are substantially different in cell parameters and tested under a wide range of operating conditions. At each operating condition, fuel cell polarization curves are collected. A dataset of 23,820 samples from these cells, that could be useful for data-driven modeling is obtained and available as open source to the communityT06-A02 University of Alberta Publication 2024-03-01 T06-A02 Temporal Dilated Convolution and Nonlinear Autoregressive Network for Predicting Solid Oxide Fuel Cell Performance T06-A02 University of Alberta Publication 2024-10-01 Mohamadali Tofigh,
" Ali Kharazmi
" ,
" Daniel Smith
" ,
Hanifi, A. , Charles Robert Koch
T06-A02 Transfer learning-based deep neural network model for performance prediction of hydrogen-fueled solid oxide fuel cells T06-A02 University of Alberta Publication 2025-01-01 T06-A02 A parallel Seq2Seq neural architecture for long-horizon performance forecasting and online condition monitoring of fuel cells T06-A02 University of Alberta Publication 2026-03-01 T06-A02 Dynamic Kolmogorov–Arnold networks for time-varying degradation modeling in solid oxide fuel cells T06-A02 University of Alberta Publication 2026-03-01 T06-A02 A deep Conv-Atten-KAN architecture to identify solid oxide fuel cell dynamics under fuel starvation T06-A02 University of Alberta Publication 2025-08-06 T06-A02 State-of-health forecasting of solid oxide fuel cells using physics-informed temporal graph convolutional network T06-A02 University of Alberta Publication 2026-02-01 T06-A02 Imitative Learning Control of a LSTM-NMPC Controller on PEM Fuel Cell for Computational Cost Reduction T01-P04 University of Alberta Publication 2023-01-01 Alireza Salahi, S Moghadasi, H Borhan, Charles Robert Koch,
Shahbakhti, M. T01-P04 Cold Climate Impact on Air-Pollution-Related Health Outcomes: A Scoping Review T01-P04 University of Alberta Publication 2022-01-01 T01-P04 Modeling, diagnostics, optimization, and control of internal combustion engines via modern machine learning techniques: A review and future directions T01-P04 University of Alberta Publication 2022-01-01 T01-P04 Hybrid Machine Learning Approaches and a Systematic Model Selection Process for Predicting Soot Emissions in Compression Ignition Engines T01-P04 University of Alberta Publication 2021-11-01 Saeid Shahpouri, Armin Norouzi, Christopher Hayduk, Reza Rezaei,
Shahbakhti, M. , Charles Robert Koch
T01-P04 Model Predictive Control of Internal Combustion Engines: A Review and Future Directions T01-P04 University of Alberta Publication 2021-10-01 Armin Norouzi, Hamed Heidarifar,
Shahbakhti, M. , Charles Robert Koch, Hoseinali Borhan
T01-P04 Safe deep reinforcement learning in diesel engine emission control T01-P04 University of Alberta Publication 2023-02-01 T01-P04 Integrating Machine Learning and Model Predictive Control for automotive applications: A review and future directions T01-P04 University of Alberta Publication 2023-04-01 Armin Norouzi Yengeje, Hamed Heidarifar, Hoseinali Borhan,
Shahbakhti, M. , Charles Robert Koch
T01-P04 Laminar Flame Speed modeling for Low Carbon Fuels using methods of Machine Learning T01-P04 University of Alberta Publication 2023-02-01 Saeid Shahpouri, Armin Norouzi Yengeje, Christopher Hayduk, Alexander Fandakov, Reza Rezaei, Charles Robert Koch,
Shahbakhti, M. T01-P04 Deep learning based model predictive control for compression ignition engines T01-P04 University of Alberta Publication 2022-10-01 Armin Norouzi Yengeje, Saeid Shahpouri,
Gordon, D. , Alexander Winkler, Eugen Nuss, Dirk Abel, Jakob Andert,
Shahbakhti, M. , Charles Robert Koch
T01-P04 End-to-End Deep Neural Network Based Nonlinear Model Predictive Control: Experimental Implementation on Diesel Engine Emission Control T01-P04 University of Alberta Publication 2022-12-01 Gordon, D. , Armin Norouzi Yengeje, Alexander Winkler, Jakub Tyler McNally, Eugen Nuss, Dirk Abel,
Shahbakhti, M. , Jakob Andert, Charles Robert Koch
T01-P04 Machine Learning Integrated with Model Predictive Control for Imitative Optimal Control of Compression Ignition Engines T01-P04 University of Alberta Publication 2022-01-01 Armin Norouzi, Saeid Shahpouri,
Gordon, D. , Alexander Winkler, Eugen Nuss, Dirk Abel, Jakob Andert,
Shahbakhti, M. , Charles Robert Koch
T01-P04 Hybrid emission and combustion modeling of hydrogen fueled engines T01-P04 University of Alberta Publication 2023-01-01 T01-P04 Performance and Emission Investigation of Hydrogen Diesel Dual Fuel Combustion T01-P04 University of Alberta Publication 2023-01-01 T01-P04 Thermodynamic and Economic Analysis of a Novel Concept for Methane Pyrolysis in Molten Salt Combined with Heliostat Solar Field Methane pyrolysis has emerged as a promising low-carbon hydrogen production method, serving as a bridge from conventional fossil fuels to renewable energies (REs). A recent technique for sustainable hydrogen production involves the process of methane pyrolysis in molten salt. To better understand the potential of sustainable turquoise hydrogen production, this paper presents a technoeconomic assessment of methane pyrolysis in molten salt combined with a high-temperature heliostat solar field. In addition, a high-temperature thermal energy storage (HTES) unit is utilized to ensure uninterrupted hydrogen production during periods of solar unavailability, thereby contributing to the economic viability of the overall system. The proposed system can produce 9 tons of H 2 /day and 27 tons of solid carbon/day with a round-trip efficiency of 49.8% and a levelized cost of hydrogen (LCOH) of 1.93 $/kg. To further decarbonize the process, the feasibility of utilizing renewable energies as the source of electricity for the HTES was analyzed, yielding a synergistic achievement of clean and economic hydrogen production. This configuration resulted in a LCOH of 1.25 $/kg, after including the United States (US) clean hydrogen production tax incentives. The proposed system shows the potential to achieve the US Department of Energy target of 1 $/kg of hydrogen with a 20% reduction in the cost of RE infrastructure.T06-A02 University of Alberta Publication 2024-06-01 T06-A02 Modeling and microstructural study of anode-supported solid oxide fuel cells: Experimental and thermodynamic analyses Developing novel solid oxide fuel cells (SOFCs) with high stability running at low temperatures is an important objective in SOFC science. In the current paper, a comprehensive physics-based microstructure modeling using scanning electron microscope (SEM) image analysis was performed on several anode-support SOFCs operating at low temperatures with high stability. To bridge the gap in the literature regarding an accurate and realistic modeling, a new model was developed based on the variable fuel and air utilization factors and updated microstructure values (e.g., tortuosity, porosity, pore size, grain size). The model accuracy was verified by a thorough point-to-point validation for eight different cells with the configuration of Ni-YSZ (anode), YSZ (electrolyte), and GDC/PNO (cathode). Different temperatures, hydrogen, and air mass flow rates were used, for which an average error of less than 3% in the I-V curves was achieved. The microstructure of the cells, including cathode thickness (15–26 μm), anode-support thickness (350–460 μm), porosity (39 and 43%), grain size (1.1–1.4 μm), and pore radius (0.9–1.1 μm) were varied. Moreover, the effects of the critical operational and design parameters on the overpotential losses and cell performance were studied. The results show that a hydrogen flow rate of 43 sccm was ideal when the cell operated at 0.9 A/cm2 and 700 °C. Moreover, an average anode-support pore radius of 1.75 μm resulted in the best cell performance. It was also concluded that the electrolyte thickness has a higher effect on the cell performance compared to the cathode thicknessT06-A02 University of Alberta Publication 2023-09-13 T06-A02 Design, thermodynamic, and economic analyses of a green hydrogen storage concept based on solid oxide electrolyzer/fuel cells and heliostat solar field Green hydrogen production is facing challenges in balancing economic feasibility with sustainability. Employing efficient hydrogen production designs and benefiting from the potential of hydrogen storage provide two promising strategies to mitigate the economic constraints associated with green hydrogen production. This paper proposes a novel hybrid design for green hydrogen production/utilization based on efficient high-temperature units, including heliostat solar field, solid oxide electrolyzer cell (SOEC), and solid oxide fuel cell (SOFC). The proposed system is comprehensively investigated from thermodynamic and economic perspectives, along with conducting a case study based on hourly electricity prices and actual solar data. The system demonstrates a hydrogen production rate of 7.76 ton/day using the SOEC and a power generation of 54.3 MWh in the SOFC for peak demand shaving, yielding an overall round trip efficiency of 74.2%. The case study results indicate that the economic feasibility is significantly compromised if all the produced hydrogen is sold at prices below 2.75 $/kg; while, implementation of hydrogen storage for peak shaving can yield a promising payback period of less than 2 years. The hydrogen and hourly electricity prices, along with the duration of peak times, are the other critical factors that affect economic viability.T06-A02 University of Alberta Publication 2023-07-08 T06-A02 Invention disclosure: SOFC diagnostics and control with use of customized power electronics We submitted Cummins Invention Disclosure Form for the usage of customized power electronics to increase observability and controllability of solid oxide oxide fuel cells in a stack.T06-A02 University of Alberta IP Management 2023-04-03 T06-A02 2024 Canadian Hydrogen Convention Attended 2024 Canadian Hydrogen Convention and discussed collaboration with potential partners in Strathcona County to test SOFC for combined heat and power (CHP) application in their planned facility to create a hydrogen fueled CHP testbed using their recent ERA funding.T06-A02 University of Alberta Activity 2024-04-23 T06-A02 Transient NOx emission modeling of a hydrogen-diesel engine using hybrid machine learning methods T02-A04 University of Alberta Publication 2024-01-01 T02-A04 Real-time vehicular fuel consumption estimation using machine learning and on-board diagnostics data T02-A04 University of Alberta Publication 2023-01-01 T02-A04 Tailpipe NOx Emissions Modeling of a Heavy-Duty Diesel Truck Using Deep Learning Methods T02-A04 University of Alberta Publication 2025-01-01 T02-A04 Systematic framework for deep learning-based predictive injection control with Bayesian hyperparameter optimization for a hydrogen/diesel dual-fuel engine T02-A04 University of Alberta Publication 2025-11-01 Hossein Mehnatkesh, Alexander Winkler, Edward Sperling, Javad Kheyrollahi,
Shahbakhti, M. ,
Gordon, D. , Charles Robert Koch
T02-A04
