Profile
| Keywords: | Control, Internal Combustion Engines, alternative fuels, machine learning |
Dr. C.R. (Bob) Koch is a Professor in the Department of Mechanical Engineering at the University of Alberta, Edmonton Alberta (since 2001). He was the Associate Chair Research in the Mechanical Engineering department (2013- 2015). In 2007, he was a visiting professor at the Institute of Systems Theory at the University of Stuttgart, Stuttgart, Germany. In 2016 he was a visiting professor at the Technical University of Berlin, Berlin Germany.
Before joining the University of Alberta, Koch worked for 10 years in the automotive industry. From 1991 to 1992 and from 1994 to 2001 he worked at Daimler-Benz - DaimlerChrysler in Stuttgart Germany in advanced internal combustion engines. During 1992 to 1994 he worked for General Motors in the Detroit area on automotive powertrain control. From this work he holds seven US patents and five German patents.
Koch received his B.S. degree in mechanical engineering from the University of Alberta, Edmonton, Canada in 1985, and his M.S. and Ph.D. degrees from Stanford University, Palo Alto, CA, in 1986 and 1991, respectively.
Improving fluid mechanical and combustion systems using closed loop control methods is the general focus of Koch’s research. His research interests include combustion engines, advanced powertrains and control of both reacting and non-reacting fluids and combining machine learning and controls.
Koch is currently a technical editor for Mechatronics Journal and Control Engineering Practice and has been on the technical committee for a variety of conferences and reviews. He is also active providing peer review for journal papers, grant applications and conferences. He has chaired sessions in conferences and been on the program committee for conferences. FES Funded ProjectsOutputs
| Title |
Category |
Date |
Authors |
| Performance and Emission Investigation of Hydrogen Diesel Dual Fuel CombustionT01-P04 University of Alberta | Publication | 2023-01-01 | | | Deep Neural Network based Performance and Emission ModellingT01-P04 University of Alberta | Publication | 2023-01-01 | | | MPC implementation for HCCI Combustion using a Deep Neural Network based modelT01-P04 University of Alberta | Publication | 2023-01-01 | | | Journal of Aerosol Science Morphology and volatility of particulate matter emitted from a gasoline direct injection engine fuelled on gasoline and ethanol blendsT01-P04 University of Alberta | Publication | 2018-04-23 | | | Investigating the effect of temperature on NOx sensor cross sensitivity to ammonia using a physic-based model"Investigating the effect of temperature on NOx sensor cross sensitivity to ammonia using a physic-based model", T01-P04 University of Alberta | Activity | 2017-05-16 | Masoud Aliramezani, "Khashayar Ebrahimi ", Koch, C., "Robert Hayes " | | Symmetric Negative Valve Overlap effects on energy distribution of a single cylinder HCCI engineT01-P04 University of Alberta | Publication | 2018-04-02 | | | Phenomenological model of a solid electrolyte NOx and O2 sensor using temperature perturbation for on-board diagnosticsAbstract Amperometric \NOx\ sensors are increasingly used in automotive industry to meet the stringent emission measurement regulations. These sensors measure O2 and NOx concentration using two different sensing cells. In this work, a physics-based model was developed and then employed to predict the sensor output for oxygen as a function of sensor temperature and oxygen concentration. A temperature perturbation method was also developed based on the model to calibrate the sensor output with respect to oxygen concentration. The model accurately matched the experimental results for steady state and transient conditions. A two step sensor diagnostics procedure based on the sensor temperature perturbation method was then proposed. The first diagnostics step evaluates the sensor output to check if it is within the acceptable range. The second diagnosis step checks the plausibility of the sensor output based on the physics based model and temperature perturbation. A self-calibration procedure was also implemented inside the diagnostics procedure using temperature perturbation at engine-off. This self-recalibration only requires an external relative humidity measurement. T01-P04 University of Alberta | Publication | 2018-04-15 | Masoud Aliramezani, Koch, C., Ron Patrick | | Real-time Control of HCCI Engine Using Model Predictive ControlT01-P04 University of Alberta | Publication | 2018-06-01 | | | SCC 2018 in AachenT01-P04 University of Alberta | Activity | 2019-04-14 | | | Experimental Investigation and Analysis of Natural Gas RCCI on a Modifed GDI Engine using NVOT01-P04 University of Alberta | Publication | 2018-05-01 | | | Axial insulation rings - testing and simulation of pressure drop and temperature transients in engine exhaust catalystsT01-P04 University of Alberta | Publication | 2018-05-01 | Giffin Symko, Masoud Aliramezani, Koch, C., R E Hayes | | Meet with collaboratorsMeet with Drs. Pischinger and Andert RWTH Aachen in Aachen and discuss on going collaboration.T01-P04 University of Alberta | Activity | 2018-06-05 | | | Development and experimental validation of a field programmable gate array–based in-cycle direct water injection control strategy for homogeneous charge compression ignition combustion stabilityT01-P04 University of Alberta | Publication | 2020-03-01 | Gordon, D., Christian Wouters, Maximilian Wick, Bastian Lehrheuer, Jakob Andert, Koch, C., Stefan Pischinger | | In-cycle control for stabilization of homogeneous charge compression ignition combustion using direct water injectionHomogeneous charge compression ignition offers a high potential for the reduction of CO2 and NOx raw emissions; however, its use entails problems that are associated with low combustion stability, especially at the limits of the operating range. The recirculation of exhaust gases inside the combustion chamber by using a negative valve overlap leads to a strong coupling of consecutive cycles. The cyclic coupling induces phases of unstable operation after the occurrence of stochastic outlier cycles with misfire or incomplete combustion. These unstable phases are marked by reduced efficiency and increased emissions. Two in-cycle closed-loop control algorithms, which focus on the heat release in the intermediate compression, are presented in this article. To control the combustion process, direct water injection is used to ensure a direct influence on the temperature level in the combustion chamber; subsequently this influences combustion phasing. The decoupling of consecutive cycles serves to reduce deviations in the indicated mean effective pressure and crank angle position of 50% mass fraction burned. To develop a suitable controller, a first-order autoregressive model of homogeneous charge compression ignition combustion is split into intermediate compression and main combustion phases. Moreover, unstable sequences are analyzed in the time domain to identify appropriate in-cycle control concepts. The control concepts are developed based on the heat release in the intermediate compression as a strong correlation factor for consecutive cycles. To realize fast control interventions, a real-time cylinder pressure analysis as well as the control algorithms are implemented on a field-programmable gate array. The control algorithms are validated on a single-cylinder research engine and compared with conventional operation without in-cycle control. Results show a significant increase in the stability of combustion phasing and load by means of in-cycle control.T01-P04 University of Alberta | Publication | 2019-01-15 | Maximilian Wick, Julian Bedei, Jakob Andert, Gordon, D., Koch, C., Christian Wouters, Bastian Lehrheuer, Eugen Nuss | | Development and Experimental Validation of an FPGA Based In-Cycle Direct Water Injection Control Strategy for HCCI Combustion StabilityT01-P04 University of Alberta | Publication | 2018-01-01 | Gordon, D., Christian Wouters, Maximilian Wick, Bastian Lehrheuer, Jakob Andert, Koch, C., Stefan Pischinger | | An electrochemical model of an amperometric NOx sensorT01-P04 University of Alberta | Publication | 2019-01-01 | | | Active control of vortex shedding from a blunt trailing edge using oscillating piezoelectric flapsT14-P05 University of Alberta | Publication | 2019-05-01 | | | Symposium for Combustion Control (SCC) 2019SCC 2019 in Aachen
The Symposium for Combustion Control directed by Prof. Stefan Pischinger (Institute for Combustion Engines), Prof. Jakob Andert (Institute for Combustion Engines), Prof. Dirk Abel (Institute of Automatic Control), Dr. Thivaharan Albin (Institute of Automatic Control) and Prof. Heinz Pitsch (Institute for Combustion Technology) of RWTH Aachen University has already taken place for the fourth time. In this year, again over 90 participants from 13 different countries visited the conference in Aachen.
In 19 technical presentations and additional plenary speeches of Prof. Christian Schwarz (BMW Group), Prof. Per Tunestål (Lund University) and Prof. Gregory M. Shaver (Purdue University), the latest theoretical and application driven developments for the control of next generation combustion engines were presented. During the two conference days, varied topics were expounded.
Prof. Christian Schwarz showed detailed life cycle analysis of different electrified powertrains and highlighted that improved combustion engines can significantly contribute to a reduction of CO2 emissions. The use of e-fuels couples the production of electrical energy to the transport sector combines benefits like peak shaving in the energy grid with low-CO2 transport. Prof. Per Tunestål showed an outstanding presentation how combustion control can act as an enabler for clean combustion engines. Also the third keynote from Prof. Gregory Shaver focused on low emission engines, but more on the Diesel side. He showed very clearly that improved control methods could help to keep the exhaust aftertreatment system in the right temperature range.The dinner took place at the restaurant LivingRoom close to the historical city hall.T01-P04 University of Alberta | Activity | 2019-06-04 | | | AAC 20199th IFAC Int. Sym. on Advances in Automotive Control, Orleons, FranceT01-P04 University of Alberta | Activity | 2019-08-10 | | | Secondary instabilities in the wake of an elongated two-dimensional body with a blunt trailing edgeT14-P05 University of Alberta | Publication | 2018-01-01 | | | Integral discrete-time sliding mode control of homogeneous charge compression ignition (hcci) engine load and combustion timingConference Paper T01-P04 University of Alberta | Publication | 2019-08-10 | | | A Variable-Potential Amperometric Hydrocarbon SensorUsing the understanding of an inexpensive production NOx sensor, the operating parameters are changed to enable hydrocarbon measurement using the same sensor. A limiting-current-type amperometric hydrocarbon sensor for rich conditions (in the absence of O2) is developed in this work. To do this, an inexpensive three-chamber amperometric sensor with three separate electrochemical cells is parameterized to measure propane concentration. The sensor is tested using a controlled sensor test rig at different propane concentrations. The inputs to the sensor electrochemical cells have been modified to determine the best HC measurement parameters (HCMPs) for measuring propane at different concentrations. First, the transient performance and stability of the sensor are optimized by changing the sensor temperature, the reference cell potential, and the stabilizing cell potential at a high propane concentration (5000 ppm - balanced with nitrogen). Over the range tested, the sensor has the longest stable output duration at the temperature of 1009 K, the reference cell potential of 0.67 V and the stabilizing cell potential of 0.45 V. Using these sensor inputs for sensor temperature, reference cell potential and stabilizing cell potential, the sensor steady state behavior is studied to find the diffusion-rate-determined operating region. The sensor is shown to have a linear sensitivity to propane concentration from 0 to 3200 ppm. Finally, the sensor response time to different step changes from 0 up to 5000 ppm propane concentration are studied. It is shown that propane stepsize does not have a significant effect on the sensor response time. Consequently, using the working principles of an existing production amperometric NOx sensor and changing the sensor operating parameters, an amperometric hydrocarbon sensor that works in diffusion rate determining operating region is developed.T01-P04 University of Alberta | Publication | 2019-01-01 | | | Machine learning-based diesel engine-out NOx reduction using a plug-in PD-type iterative learning controlT01-P04 University of Alberta | Publication | 2020-08-26 | | | Combustion Institute Canadian Section (CICS) 2018T01-P04 University of Alberta | Activity | 2018-05-11 | | | Robotic Manipulator Control Using PD-type Fuzzy Iterative Learning ControlT01-P04 University of Alberta | Publication | 2019-01-01 | | | Symposium for Combustion Control (SCC) 2018The automotive world is facing rapid changes. Real world driving emissions are in the focus of the public, well-established technologies are reassessed and new players enter the global market. To achieve a sustainable and green mobility the development of efficient and clean combustion engines is one of the key requirements. Most of the promising and novel approaches require innovative closed-loop control approaches, detailed physical models, powerful control logics and new sensor concepts.
The Symposium for Combustion Control was established in 2015 to foster the interaction between the scientific community and the automotive industry. Its focus are the latest theoretical and application driven developments for the control of next generation combustion engines. In the last years, the program was completed with presentations for example given by VW AG, DENSO, Daimler AG, Jaguar Land Rover Ltd., Ford, BMW Group, TNO Automotive, and many further international companies and universities.T01-P04 University of Alberta | Activity | 2018-06-27 | | | Symposium for Combustion Control (SCC) 2017Symposium for Combustion ControlT01-P04 University of Alberta | Activity | 2017-06-28 | | | A control oriented diesel engine NOx emission model for on board diagnostics and engine control with sensor feedbackT01-P04 University of Alberta | Publication | 2019-01-01 | Masoud Aliramezani, Robert E Hayes, Armin Norouzi Yengeje, Koch, C. | | Integration of PD-type Iterative Learning Control with Adaptive Sliding Mode ControlProportional-Derivative type Iterative Learning Controller (PD-ILC) is combined with an Adaptive Sliding Mode Controller (ASMC) using a plug-in structure to a rotary pendulum. The ASMC adaptation law is used to update a switching gain of sliding surface in Sliding Mode Control (SMC) controller. The proposed hybrid controller stability and convergence are mathematically shown and then experimentally demonstrates using two-degree-of-freedom (2-DOF) Quanser\textcopyright~QUBE$^TM$~Servo 2 Rotary Pendulum. Results illustrate that adaptation law helps the controllers to achieve higher accuracy tracking performance compared to the classic SMC controller. Based on the experimental results, the hybrid control of PD-ILC and ASMC has faster and more accurate tracking results than ILC controller indicating the combined controller has better performance than the individual controllers.T01-P04 University of Alberta | Publication | 2020-07-15 | | | Support vector machine for a diesel engine performance and NOx emission control-oriented modelT01-P04 University of Alberta | Publication | 2020-07-15 | Masoud Aliramezani, Armin Norouzi Yengeje, Koch, C. | | NVO peak pressure based in-cycle control for HCCI combustion using direct water injectionHomogeneous Charge Compression Ignition (HCCI), is a low temperature combustion method, which can significantly
reduce nitrogen oxides (NOx) emissions compared to current lean-burn spark ignition engines. The lack of direct ignition
control leads to high cyclic variation with HCCI combustion. A fully variable electromagnetic valve train is used to
provide the required thermal energy for HCCI through internal exhaust gas recirculation (EGR) using negative valve
overlap (NVO). This leads to an increase in the cyclic coupling as residual gas and unburnt fuel is transferred between
cycles through EGR. To improve combustion stability an experimentally validated feed-forward water injection controller
is presented. Utilizing the low latency and rapid calculation rate of a Field Programmable Gate Array (FPGA) a real-time
calculation of the cylinder pressure and the controller is implemented on a prototyping engine controller. The developed
and experimentally tested controller relates the upcoming combustion phasing to the peak NVO pressure. This control
strategy aims to prevent the early rapid combustion following combustion during the NVO period by using direct water
injection to cool the cylinder charge and counter the additional thermal energy from any residual fuel that burnt during
the NVO period. By cooling the trapped cylinder mass the upcoming combustion phasing can be delayed to the desired
setpoint. The controller was experimentally tested showed slight improvement in the combustion stability as shown by a
reduction in the standard deviation of indicated mean effective pressure and reduced pressure rise rates.T01-P04 University of Alberta | Publication | 2019-05-16 | Gordon, D., Koch, C., "Christian Wouters ", "Bastian Lehrheuer ", "Stephan Pischinger ", "Maximilian Wick ", "Jakob Andert " | | CCECE 2019T01-P04 University of Alberta | Activity | 2019-05-03 | | | Combustion Institute Canadian Section (CICS) 2019Combustion Institute Canadian Section (CICS) 2019T01-P04 University of Alberta | Activity | 2019-05-14 | | | Investigating the effect of temperature on NOx sensor cross sensitivity to ammonia using a physics based modelT01-P04 University of Alberta | Publication | 2017-05-15 | Masoud Aliramezani, K Ebrahimi, R E Hayes, Koch, C. | | Industry Mixer Lightning PostersIndustry Mixer Lightning Posters
T01-P04 University of Alberta | Activity | 2020-02-20 | | | The Effect of Operating Parameters of an Amperometric NOx-O2 Sensor on the Sensor ResponseThe Effect of Operating Parameters of an Amperometric NOx-O2 Sensor on the Sensor Response conference paper.T01-P04 University of Alberta | Publication | 2019-06-15 | | | Production engine emission sensor modeling for in-use measurement and on-board diagnosticsFES poster session 2019T01-P04 University of Alberta | Activity | 2020-02-26 | | | Response characteristics of an amperometric NOx-O2 sensor at non diffusion-rate-determining conditionsT01-P04 University of Alberta | Publication | 2021-04-14 | | | Evaluation of ASTM D6424 standard for knock analysis using unleaded fuel candidates on a six cylinder aircraft engineT01-P04 University of Alberta | Publication | 2021-04-01 | | | A correlation-based model order reduction approach for a diesel engine NOx and brake mean effective pressure dynamic model using machine learningT01-P04 University of Alberta | Publication | 2021-07-01 | Armin Norouzi, Masoud Aliramezani, Koch, C. | | A grey-box machine learning based model of an electrochemical gas sensorT01-P04 University of Alberta | Publication | 2020-10-01 | Masoud Aliramezani, Armin Norouzi, Koch, C. | | Development and experimental validation of a real-time capable field programmable gate array\textendash based gas exchange model for negative valve overlapT01-P04 University of Alberta | Publication | 2018-07-01 | Gordon, D., Christian Wouters, Maximilian Wick, Feihong Xia, Bastian Lehrheuer, Jakob Andert, Koch, C., Stefan Pischinger | | Homogeneous charge compression ignition combustion stability improvement using a rapid ignition systemT01-P04 University of Alberta | Publication | 2020-06-01 | Gordon, D., Christian Wouters, Shota Kinoshita, Maximilian Wick, Bastian Lehrheuer, Jakob Andert, Stefan Pischinger, Koch, C. | | Model Predictive Control of Ginzburg-Landau EquationT01-P04 University of Alberta | Publication | 2018-08-01 | Mojtaba Izadi, Koch, C., Stevan S Dubljevic | | Discrete-time model-based output regulation of fluid flow systemsT14-P05 University of Alberta | Publication | 2021-01-01 | | | Model Predictive Control of Jacket Tubular Reactors with a Reversible Exothermic ReactionT14-P05 University of Alberta | Publication | 2020-10-01 | Lu Zhang, Junyao Xie, Koch, C., Stevan Dubljevic | | Discrete output regulator design for linear distributed parameter systemsT14-P05 University of Alberta | Publication | 2020-08-01 | | | Internal Model Controller Design of Linearized Ginzburg-Landau EquationT14-P05 University of Alberta | Publication | 2020-01-01 | | | Cold Climate Impact on Air-Pollution-Related Health Outcomes: A Scoping ReviewT01-P04 University of Alberta | Publication | 2022-01-01 | | | Modeling, diagnostics, optimization, and control of internal combustion engines via modern machine learning techniques: A review and future directionsT01-P04 University of Alberta | Publication | 2022-01-01 | | | Hybrid Machine Learning Approaches and a Systematic Model Selection Process for Predicting Soot Emissions in Compression Ignition EnginesT01-P04 University of Alberta | Publication | 2021-11-01 | | | Model Predictive Control of Internal Combustion Engines: A Review and Future DirectionsT01-P04 University of Alberta | Publication | 2021-10-01 | | | Safe deep reinforcement learning in diesel engine emission controlT01-P04 University of Alberta | Publication | 2023-02-01 | | | Integrating Machine Learning and Model Predictive Control for automotive applications: A review and future directionsT01-P04 University of Alberta | Publication | 2023-04-01 | | | Laminar Flame Speed modeling for Low Carbon Fuels using methods of Machine LearningT01-P04 University of Alberta | Publication | 2023-02-01 | Saeid Shahpouri, Armin Norouzi Yengeje, Christopher Hayduk, Alexander Fandakov, Reza Rezaei, Koch, C., Shahbakhti, M. | | Support vector machine based emissions modeling using particle swarm optimization for homogeneous charge compression ignition engineT01-P04 University of Alberta | Publication | 2021-11-01 | Gordon, D., Armin Norouzi, Gero Blomeyer, Julian Bedei, Masoud Aliramezani, Jakob Andert, Koch, C. | | Deep learning based model predictive control for compression ignition enginesT01-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., Koch, C. | | Constrained Receding Horizon Output Estimation of Linear Distributed Parameter SystemsT14-P05 University of Alberta | Publication | 2022-01-01 | Junyao Xie, Jukka-Pekka Humaloja, Koch, C., Stevan Dubljevic | | End-to-End Deep Neural Network Based Nonlinear Model Predictive Control: Experimental Implementation on Diesel Engine Emission ControlT01-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, Koch, C. | | Machine Learning Integrated with Model Predictive Control for Imitative Optimal Control of Compression Ignition EnginesT01-P04 University of Alberta | Publication | 2022-01-01 | Armin Norouzi, Saeid Shahpouri, Gordon, D., Alexander Winkler, Eugen Nuss, Dirk Abel, Jakob Andert, Shahbakhti, M., Koch, C. | | Hybrid emission and combustion modeling of hydrogen fueled enginesT01-P04 University of Alberta | Publication | 2023-01-01 | | | Approximate moving horizon estimation for switching conservative linear infinite-dimensional systemsT01-P04 University of Alberta | Publication | 2023-01-01 | Junyao Xie, Jukka-Pekka Humaloja, Koch, C., Stevan Dubljevic | | Imitative Learning Control of a LSTM-NMPC Controller on PEM Fuel Cell for Computational Cost ReductionT01-P04 University of Alberta | Publication | 2023-01-01 | | | Application of a Combinatorial Vortex Detection Algorithm on 2 Component 2 Dimensional Particle Image Velocimetry Data to Characterize the Wake of an Oscillating WingT14-P05 University of Alberta | Publication | 2024-01-01 | Mathew Bussière, Guilherme M Bessa, Koch, C., David S. Nobes | | Transient Modeling of a Solid Oxide Fuel Cell using an Efficient Deep Learning HY-CNN-NARX ParadigmControl 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 | | | Developing an Efficient Model for a SOFC System 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 | 2024-07-08 | | | A New Architecture based on Temporal Convolution and Nonlinear Autoregressive Exogenous for Performance Prediction of Solid Oxide Fuel Cells under Dynamic OperationT06-A02 University of Alberta | Publication | 2024-03-01 | | | Control-oriented Modeling of a Solid Oxide Fuel Cell Affected by Redox Cycling using a Novel Deep Learning ApproachT06-A02 University of Alberta | Publication | 2024-03-01 | | | Control-oriented Modeling of a Solid Oxide Fuel Cell Affected by Redox Cycling using a Novel Deep Learning ApproachT06-A02 University of Alberta | Publication | 2024-10-27 | | | Performance Prediction of a Range of Diverse Solid Oxide Fuel Cells using Deep Learning and Principal Component Analysis T06-A02 University of Alberta | Publication | 2024-10-30 | | | Invention disclosure: SOFC diagnostics and control with use of customized power electronicsWe 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 | | | Canadian Hydrogen ConventionT06-A02 University of Alberta | Activity | 2023-04-24 | | | SOFC DatabaseThree 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 | |
|
|
