University of Alberta
  • Edmonton, Alberta, Canada
Recent publications
Artificial Neural Network (ANN) and Support Vector Machine (SVM) methods are designed to accurately predict Laminar Flame Speed (LFS) over the entire engine operating range for Ammonia (NH3), Hydrogen (H2), and Methanol (CH3OH). These are promising zero-carbon or low-carbon alternative fuels for the transportation sector but require combustion models to optimize and control the engine performance. These developed Machine Learning (ML) methods provide an LFS prediction that requires several orders of magnitude less computation time than the original thermo-kinetic combustion mechanisms but has similar accuracy. Then an SVM and an ANN LFS model for blends of the three fuels was developed by combining LFS datasets of different fuels. Results show that for single fuels, ANN shows better performance than SVM and can predict the LFS with a correlation coefficient Rtest2 higher than 0.999. For fuel blends, SVM has better performance with Rtest2 close to 0.999. These predictive ML LFS models can be integrated into 0D and 1D engine models and their low computation time makes them useful for engine development and for future model-based combustion control applications.
Developing practical, inexpensive catalysts for the partial upgrading of crude oil is essential for converting heavy oil to meet pipeline standards without diluent. Here we report that air-stable unsupported iron sulfide nanoparticles derived from a single-source homogeneous precatalyst, Fe2S2(CO)6, are highly effective catalysts for the partial hydrogenation of substituted anthracenes and aromatic nitrogen and sulfur heterocycles. The reactivity of the anthracene series is governed by stereoelectronic effects imposed by the phenyl substituents, decreasing in the order anthracene > 9-phenylanthracene > 9,10-diphenylanthracene. Nitrogen- and sulfur-containing heterocycles are partially hydrogenated, with limited heteroatom removal for benzothiophene, showing that the unsupported iron sulfide catalyst is well-suited for the selective partial hydrogenation and limited defunctionalization, which are key for catalytic partial upgrading of heavy petroleum. The unsupported iron sulfide nanoparticles also catalyze hydrogenation of mixed “feeds” comprising combinations of model compounds, demonstrating that carbocyclic and heterocyclic molecules can be processed simultaneously with little self-inhibition of overall activity, a key requirement for compositionally challenging bitumen feeds. Several techniques were used to characterize the iron sulfide nanoparticles, as well as the organic reaction products. The heterogeneous catalyst, generated in situ, was more active than a ball-milled commercial iron sulfide catalyst. The heterobimetallic precatalyst Co2FeS(CO)9 does not enhance the rates of hydrogenation.
The purpose of this study is to improve the quality of the multi-track laser cladding coating. The Taguchi-Grey relation method was selected to realize process parameter optimization. The Taguchi method is used to design an L16 orthogonal experiment. The influence of three important laser cladding parameters (laser power, powder feeding rate, and scanning speed) on the micro-hardness, maximum load value, yield strength, UTS, and elongation had been analyzed based on the analysis of variance (ANOVA) and signal to noise ratio (SNR) methods. The results showed that the yield strength, UTS, and elongation were significantly affected by the laser power; the powder feeding rate denoted a highly significant influence on the microhardness and yield strength; and the scanning speed was a highly significant factor that had an influence on the maximum load value. Then, grey relational analysis (GRA) was used to convert five response targets into a single grey relational grade (GRG) that could be quantified in order to optimize the parameters for maximum micro-hardness, maximum load value, yield strength, UTS, and elongation. Finally, the optimum cladding process parameters were obtained. Through analysis of microstructure, the reduction of the Laves phase might be the main reason for the improvement of coating performance after optimization.
Numerous experimental results demonstrated that graphene oxide (GO) nanosheets could not only stabilize but also destabilize the crude oil-in-water emulsion. The interfacial property of GO and its interaction with surface-active components in crude oil may be responsible for the contradictory results, which is rarely reported and may be hardly explored via the experiments. Then molecular dynamics simulation method is employed to analyze the interfacial interaction between GOn nanosheets with various surface charge densities and asphaltene molecules in crude oil. The radial distribution function (RDF), two-dimensional number density distribution, and molecular orientation analysis are calculated to systematically explore the effects of GOn on the asphaltene interfacial distribution. The GOn-asphaltene binary film stability is evaluated at the macroscopic level via the steered dynamic simulation, which is further investigated by reduced density gradient (RDG) approach to visualize their intermolecular interaction. The results show that the GO1 with higher surface charge density exhibits greater repulsion (∼5000 kJ/mol) with asphaltene-1 (ASP1), which can effectively disrupt ASP1 interfacial arrangement and destabilize the O/W emulsion. Whereas, the generated n-π*, π-π stacking, and cation-π molecular interaction between GOn with lower surface charge densities and asphaltene should be responsible for the great emulsion stability. In a word, this study firstly proposed the novel understanding about the underlying mechanism of GOn caused demulsification and emulsification.
Computed tomography (CT) in combination with advanced image processing can be used to non-invasively and non-destructively visualize complex interiors of living and non-living media in 2 and 3-dimensional space. In addition to medical applications, CT has also been widely used in soil and plant science for visual and quantitative descriptions of physical, chemical, and biological properties and processes. The technique has been used successfully on numerous applications. However, with a rapidly evolving CT technologies and expanding applications, a renewed review is desirable. Only a few attempts have been made to collate and review examples of CT applications involving the integrated field of soil and plant research in recent years. Therefore, the objectives of this work were to: (1) briefly introduce the basic principles of CT and image processing; (2) identify the research status and hot spots of CT using bibliometric analysis based on Web of Science literature over the past three decades; (3) provide an overall review of CT applications in soil science for measuring soil properties (e.g., porous soil structure, soil components, soil biology, heat transfer, water flow, and solute transport); and (4) give an overview of applications of CT in plant science to detect morphological structures, plant material properties, and root-soil interaction. Moreover, the limitations of CT and image processing are discussed and future perspectives are given.
One-way photonic devices are based around the concept of directionally-dependent propagation of light that can be used in the design of optical circuits. We demonstrate a straightforward method arising from cascade lasing transfer using a pair of coupled microcavities. The concept is based on consecutive down-converted laser transmission leading to net cross-cavity propagation in one direction. To describe this system, we first develop a simple, generalized coupled-cavity rate equation model to calculate the efficiency of the lasing cascade process. The basic idea is then demonstrated experimentally using a coupled resonator geometry. The two-cavity lasing design forms a basic three-wavelength microdevice that can change the original lasing wavelength or act as a one-way down-converting micro-lasing system.
The near-wellbore permeability loss caused by the release and migration of naturally fine particles impairs the well productivity in oil reservoirs. The damage would be intensified in the presence of restrictions such as sand control screens due to retardation and accumulation of fine particles resulting from flow convergence and high volumetric flux. This study focuses on the experimental investigation and mathematical modeling of the permeability impairment near the sand control screens under a representative fines migration process for the SAGD well conditions. For this purpose, the single-phase flow experiments were conducted on unconsolidated sand packs in a sand retention testing (SRT) setup. The testing used different sand screens and considered the chemical effect of the fines migration process and different salinity for the saturating brine and flowing fluid. A 3D numerical model incorporating governing fluid and particle transport equations in porous media and an empirical permeability loss equation was established to simulate the fines migration process. The results show that the sand screens with a low open flow area and narrow aperture size cause higher permeability loss under the same flow conditions. Higher salinity reduction yields higher mobilized fines concentration, causing high near-screen permeability loss. For the salinities above 800 ppm, no significant pressure drops were observed. Furthermore, changing salinity to deionized water (DI) caused almost complete plugging of the sand pack, masking the flow performance of the sand control screen. A good match was obtained between the measured and calculated dimensionless pressure drops from the experimental and numerical models. The calibrated model could accurately enough predict the dimensionless pressure drops for the near-screen interval. It was confirmed that the porous medium's model parameters were nearly independent of the sand screen. However, different results were obtained for different salinities, confirming the 'parameters' dependency on not just porous medium but also fluid properties.
Soil physical quality is paramount for root growth, water and air movement, and for its subsequent effects on chemical and biological processes in the soil. Management practices and their legacies can impact soil physical quality, and perennial grain cropping has been proposed as a solution to maintain or improve soil physical quality in agroecosystems due to their provision of year-round ground cover and increased root growth. An alfalfa-brome perennial forage crop, a perennial rye crop (Secale cereale L. x S. montanum L.), and an annual rye crop (S. cereale L.) were evaluated at two sites in Central Alberta with contrasting management histories (Edmonton and Breton) over 3 years to determine the effects on soil physical and hydraulic properties. Compared to the annual crop, the perennial forage crop reduced the bulk density of the uppermost soil depth sampled (5–10 cm depth increment) (p < 0.05) at the Edmonton site and increased soil macroporosity (p < 0.05) and pore connectivity (p < 0.05) in the deeper subsurface soil layer (25–30 cm depth increment) at both sites. While moderate improvements in soil physical and hydraulic properties manifested under the perennial rye crop when compared to the annual rye crop, they did not do so to the extent of the perennial forage crop. We attribute this to the inclusion of tap-rooted alfalfa in the perennial forage, and the overarching beneficial influence of root mass density on soil properties. Root mass density from highest to lowest consistently ranked as perennial forage > perennial rye > annual rye for both sites. Root mass density was negatively correlated with bulk density at both Breton (r = −0.77, p < 0.05) and Edmonton (r = −0.69, p < 0.05) sites. Furthermore, at Breton, root mass density positively correlated with macroporosity (r = 0.88, p < 0.01). Notably, the perennial rye crop enhanced soil carbon mass density relative to the annual rye crop in the clayey topsoil of the Edmonton site (p < 0.05), but treatment effects were muted at the Breton site due to the influence of previous land use. Despite moderate improvements in soil physical quality, our results suggest that 3 years of perennial rye monocropping falls short of the major improvements seen under a perennial alfalfa-brome forage crop over the same timeframe.
For decades, the network-organized reaction-diffusion models have been widely used to study ecological and epidemiological phenomena in discrete space. However, the high dimensionality of these nonlinear systems places a long-standing restriction to develop the normal forms of various bifurcations. In this paper, we take an important step to present a rigorous procedure for calculating the normal form associated with the Hopf bifurcation of the general network-organized reaction-diffusion systems, which is similar to but can be much more intricate than the corresponding procedure for the extensively explored PDE systems. To show the potential applications of our obtained theoretical results, we conduct the detailed Hopf bifurcation analysis for a multi-patch predator-prey system defined on any undirected connected underlying network and on the particular non-periodic one-dimensional lattice network. Remarkably, we reveal that the structure of the underlying network imposes a significant effect on the occurrence of the spatially nonhomogeneous Hopf bifurcations.
Prognostics based on the deep learning model often assume that their training and testing data come from the same equipment with similar working conditions. However, a machine often has specific operating conditions for different tasks, which will cause significant divergence in its measurements. Generally, planned maintenance or model-based fault detections can be done first to collect very few suspension histories when the machine works in a new environment. Few suspension histories can help the prognostic model generalize to new environments. This paper proposes a transductive method to use limited suspension histories in transfer prognostics. Different from reported cross-domain prognostics that only align two-domain histories in a holistic manner, the proposed domain adaptation strategy simultaneously minimizes the distance between both marginal and conditional probability distributions in different domains. If the measurements have a clear degradation manifold, iterative learning will allow the model to get better and better pseudo predictions, thus guiding the prognostic model to learn generalized domain invariant features to deal with different working conditions. A heuristic method and a parallel framework are proposed to verify model parameters and uncertainties. The prognostic performance of the proposed approach is validated by using two case studies.
Sandwich wall panels are typically comprised of two concrete layers (or wythes) that surround a layer of rigid insulation. The most common insulation types used in precast wall panels are Expanded (EPS) and Extruded (XPS) polystyrene. Previous work showed that walls with EPS insulation have higher in-plane shear strength compared to an XPS system. This increase is due to the surface roughness of EPS but is often neglected in design as this bond is expected to fail over time. In this study, the use of notched insulation is investigated. The notches allow XPS to achieve higher capacities since bond failure is prevented as well as prevents bond degradation from reducing the insulation contribution to shear resistance in the long term. Direct shear push-through tests were completed on 15 specimens, with 9 notched and 6 un-notched samples. All specimens were constructed with GFRP bar shear connectors arranged in an X-shape. Test parameters include GFRP bar diameter (9.5 and 16.0 mm), while using XPS, and notch type (no notch, trapezoidal, rectangular). Deformation over the course of the push-through tests was tracked using displacement transducers. When the 9.5 mm connector was used, the the inclusion of rectangular notches resulted in 26.7% and 33.3% increase in the initial stiffness and peak load respectively, while the trapezoidal notches only gave respective increments of and 17.1 and 21.1%. The benefits in the placement of notches trend are not seen between samples made of 16 mm connectors were the respective parameters only saw −10.8 and 10.5% boost.
Deterioration of pre-stressed concrete (PC) bridge structures due to overloading or extreme weather conditions is of significant concern for bridge administrators and engineers. To investigate the residual performance of PC bridge girders after years of service, PC girders salvaged from an abandoned 27-year-old bridge in Alberta are studied experimentally and analytically. This study aims at understanding the degraded structural behavior through numerical simulations by (1) comparing the analytically predicted and experimental behavior of the girders and (2) performing a parametric analysis via considering various possible defects in the PC girder. 2D nonlinear finite element models are developed to predict the flexural behavior of the PC bridge girders with different types of deteriorations of various levels. The effects of different deteriorations on PC girders presented here help to identify the main possible causes of the performance degradation of the PC girders. The insights into the structural behavior of PC girders with various defects are potentially beneficial for bridge inspectors and evaluators.
The global tragedy of the COVID-19 pandemic devastated communities and societies. The pandemic also upended public transit and shared mobility, causing declines in ridership, losses in revenue sources, and challenges in ensuring social equity. Despite ongoing uncertainty, guidance can instruct recovery and build a more resilient, socially equitable, and environmentally friendly transportation future. This chapter summarizes a recent scenario planning exercise conducted by the University of California Institute of Transportation Studies in collaboration with the Transportation Research Board (TRB) Executive Committee in Spring to Fall 2020. The exercise convened 36 transportation experts in the United States who developed policy actions and research options crafted to guide near- and long-term public transit and shared mobility. Clear themes emerged from the study regarding key actions for public transit operators in the areas of: (1) innovation and technology, (2) planning and operations, (3) customer focus, and (4) workforce development. A second grouping of broader policy strategies for both public transit and shared mobility included: (1) immediate policy and actions across actors, (2) alignment of societal objectives, (3) federal transportation spending authorization, and (4) finance and subsidies. While the exercise reiterated the need for rapid actions, thoughtful planning and decision-making can prepare both sectors for a more cooperative, multimodal ecosystem.
Prognostics based on deep learning models usually assume that training and testing data come from the same equipment with similar working conditions. In practice, some new working conditions of test equipment may not be recorded in the training dataset, and thus the learned remaining useful life (RUL) prediction model may not work well. This work is the first to propose a novel prognostic model based on state-space modeling and reinforcement learning to predict the RUL of equipment operating under extrapolated new conditions without corresponding training data. Instead of directly constructing a supervised model that relates monitoring measurements to their RUL, a discrete-time state-space model is built using possible failure histories. The contribution of this paper is twofold. First, on the basis of the state-space model, an interpretable prognostic model that combines Lyapunov constraint and reinforcement learning is proposed to predict the equipment RUL. Second, an adversarial training method based on the idea of H∞ robustness is integrated to reduce the effect of state-space modeling errors. The proposed model aims to reduce the causality between RUL and operating parameters and increase the causality between RUL and unobserved degradation characteristics. Consequently, the proposed model is interpretable and has the potential to predict RUL for the equipment operating under conditions beyond the record. The proposed approach is demonstrated and validated using a simulation study and an experimental dataset.
The molecular interaction behaviors of a model asphaltene, N-(1-Hexylheptyl)-N'-(5-carboxylicpentyl) perylene-3,4,9,10-tetracarboxylic bisimide (C5Pe), between two solid surfaces of varying polarity, were studied by molecular dynamics simulations. The C5Pe molecules were solvated in water, toluene, or heptane, and the two substrate surfaces are hydrophilic alumina and relatively more hydrophobic siloxane surfaces (as model surfaces of two basal planes of kaolinite). Distinct adsorption behaviors were revealed by the simulations, which were driven and caused by the interaction of C5Pe molecules and surfaces of varying polarity in solvents of different nature. In water, both C5Pe monomer and aggregates adsorbed on the alumina surface, instead of the siloxane surface or staying in the liquid medium. Interestingly, in the control systems where a C5Pe molecule was placed between two identical surfaces in water, it displayed weaker adsorption when placed between two alumina surfaces than between two siloxane surfaces. Potential of mean force (PMF) calculations demonstrated the interplay between enthalpy-driven adsorption on the alumina surface and entropy-driven adsorption on the siloxane surface in water. In the adsorbed C5Pe aggregates on the alumina surface, the hydrophobic parts of the molecules stacked in a parallel manner and aligned perpendicularly to the alumina surface, while the hydrophilic parts formed hydrogen bonds with the surface. In toluene, C5Pe adsorbed on the alumina surface, driven by van der Waals and Coulomb interactions, as well as hydrogen bonding. A multi-layered C5Pe aggregate on the alumina surface in toluene was observed, which resulted from coordination bonds established through Ca²⁺ between C5Pe molecules. In heptane, adsorption was found on both alumina and siloxane surfaces, and PMF calculations showed similar strength of binding to the two surfaces. The adsorbed aggregates were compact, with intermolecular π-π stackings that were parallel to the surfaces. This work provides a mechanistic understanding of the interaction behaviors of asphaltenes in solution media when different clay surfaces are simultaneously present, how such behaviors may be influenced by the nature of the solvent, and the molecular forces driving such behaviors.
The invasion patterns of immiscible fluids in porous media determine the swept volume and displace efficiency, which is complexly affected by fluids, reservoir, and injection parameters. Unfortunately, there is a sparse understanding of the crossover zone during the imbibition process, description of the whole displacement process, and the method of fingering suppression. This paper used the microfluidic model combined with the glass-packing and the plate model to evaluate the invasion patterns, displacement process, and improvement methods. Eighteen microfluidic experiments were conducted under different injection conditions using oil-wet and water-wet etching models designed by actual pore throat structures. Then the phase diagram of the invasion front was plotted with log10Ca-log10M, and the influence of the viscosity ratio, injection rate, interfacial tension (IFT), and wettability on the invasion front was evaluated. The development of the invasion front and the occurrence of remaining oil were characterized by the fingering factor and image recognition, respectively. Finally, the glass-packing and plate displacement experiments were conducted to evaluate the improvement of the invasion front by pulse injection. The results show that the invasion patterns of the imbibition process can also be divided into capillary fingering, viscous fingering, and crossover zone. The invasion phase with increased viscosity and lower injection rate will have delayed breakthrough time and transform the viscous fingering into the capillary fingering. The breakthrough time in the oil-wet model is earlier, and the final swept volume is significantly reduced. But the effect of low-IFT water flooding to expand the swept volume in the oil-wet model is more significant. The fingering factor can universally characterize the displacement process, and the influence degree of each factor on fingering is viscosity ratio > wettability > Injection rate > IFT. The primary remaining oil type under the two wettability is cluster remaining oil which can be dispersed by lower-IFT water flooding. The swept area can be divided into “strong swept” and “weak swept” regions in the oil-wet model. The glass-packing model verified that pulse injection could improve the invasion front and enhance oil recovery by 7.23 % in the plate model experiment.
To economically and environmentally recover oil from reservoirs and promote CO2 utilization (CU) project, CO2 responsive surfactants have been developed to undertake multiple tasks including emulsification and demulsification during different production stages. Understanding the switching mechanisms from molecular perspectives is of great importance to the choice and design of high-performance CO2-responsive surfactants. In this work, we performed molecular dynamics (MD) simulations to study the emulsification and demulsification processes of a heptane/water mixture in the presence of a typical CO2-responsive surfactant-lauric acid (LA). Before injecting CO2, the deprotonated lauric acids (DLA) can stabilize O/W emulsions in an aqueous solution due to strong electrostatic repulsions and high interfacial activity of DLA, whereas the protonation of lauric acid (PLA) arising from CO2 injection would result in the coalescence of emulsion droplets thanks to the greatly reduced hydrophilicity of the polar groups of lauric acids and surface charge neutralization, which is unfavorable for emulsion stabilization. The potential mean force (PMF) results show a high energy barrier preventing the fusion process when two emulsions approach each other in the absence of CO2, indicating high stability of the emulsions. However, when DLA turns to be PLA, the energy barrier disappears and an attraction force emerges due to the entropic effect if two emulsions are close enough. Our study provides important insights into the structural properties of emulsions before and after CO2-triggered switching and sheds light on the switching mechanisms which may assist in selecting and designing efficient CO2-responsive surfactants.
Steam injection has been proven to be a technically reliable and economically successful heavy oil recovery technique. However, the large amount of energy that is required to generate the steam, the high operational cost, as well as the environmental concerns are hindering the future applications of steam flooding. Reducing steam consumption and the associated greenhouse gas emissions in thermal-based heavy oil recovery methods is a great challenge in the heavy oil industry. A possible alternative to steam injection is hot water injection with chemical additives, especially in heavy oil reservoirs with some degree of mobility. In this paper, we performed static and dynamic evaluations on both conventional and novel chemicals as hot water additives. First, laboratory screening and evaluation for chemicals as hot water additives were performed at both 21 °C and 60 °C. Chemicals tested include the following: Span 80, silicon dioxide, diethylenetriaminepentaacetic acid (DTPA), switchable hydrophilicity tertiary amine (SHTA), and deep eutectic solvents (DESs). The evaluation was based on four selection criteria: emulsion stability, viscosity change, thermal stability, and chemical cost. Second, selected chemicals were further tested as hot water additives in sand-pack flooding experiments. Based on the overall static evaluations, four chemicals were selected for sand-pack flooding experiments at hot water conditions: DES 4, DES 9, SHTA, and SiO2 c. DES 11 and DTPA can still be used as alternatives, at least for core scale experimentations, but the extremely high cost can be a serious limitation to consider when assessing these chemicals for field trials. Dynamic sand-pack flooding experiments showed that all selected chemical additives could significantly improve the recovery factor compared to the hot water or steam injection without chemicals. The recovery factors of hot water injection with selected chemical additives reached 67 %–80.5 % which were much higher than that of steam injection without chemicals (38.8 %). Emulsification was found to be an important recovery mechanism of chemical additives in hot water flooding. The experimental results provided evidence that hot water injection with chemical additives might work as a good alternative to steam injection in improving the heavy oil recovery, reducing the cost, and lessening the environmental impact.
Plastic pollution is considered one of the most threatening environmental issues of the 21st century. Microplastics (MPs) and nanoplastics (NPs) have significantly contributed to greenhouse gas (GHG) emissions and hence it has an impact on climate change. Methane and ethylene are the two main greenhouse gases that are produced from the most used plastics when exposed to ambient solar radiation. On the one hand, the pollution caused by the MP and NP can impact the gas exchange and CO2 circulation, resulting in higher greenhouse gas emissions. On the other hand, climate change has an impact on MPs and NPs. For example, the terrestrial plastic, windborne plastic, plastic resuspension from sediment, and plastic persistence have been increased because of the impact of the climate change. The interactions between plastic pollution and climate change have yet to be entirely understood as this topic has only recently gained attention. Literature showed that the interactions between plastic pollution and climate change are significant and cannot be overlooked. It has been proven that MPs have deleterious effects on the environment that cannot be ignored. There are some efforts of mitigating the potential impact of MPs on the environment such as recycling the plastics and ocean plastic clean-up.
Micro- and nanoplastics (MPs and NPs) permeate easily through ecosystems and food webs, transporting adsorbed contaminants and impeding metabolic and other essential processes in living organisms. Although not well studied, the cumulation of effects of MPs and NPs indicates potential harm to human health and well-being. The media has been instrumental in bringing plastic pollution awareness to the forefront of societal discussions. Still, the less visible nature of MPs and NPs has resulted in much less coverage and public understanding of these contaminants. By identifying the interrelationships between the impacts of MPs and NPs on the natural environment, ecosystems, and anthropogenic societies, better policy and actions can be taken to mitigate long-term adverse effects. Academic researchers and policymakers need to work closely to develop iterative approaches for MPs and NPs management strategies.
Institution pages aggregate content on ResearchGate related to an institution. The members listed on this page have self-identified as being affiliated with this institution. Publications listed on this page were identified by our algorithms as relating to this institution. This page was not created or approved by the institution. If you represent an institution and have questions about these pages or wish to report inaccurate content, you can contact us here.
24,822 members
Nelson Lee
  • Division of Infectious Diseases
Zaher Hashisho
  • Department of Civil and Environmental Engineering and the School of Mining and Petroleum Engineering
Stepan Hlushak
  • Department of Chemical and Materials Engineering
116 St. and 85 Ave., T6G 2R3, Edmonton, Alberta, Canada