Research & Awards

Control, Monitoring & Optimization

Our research work is concentrated on the development of advanced control formulations for distributed parameter systems modelled by linear and nonlinear PDE systems within the chemical, petroleum, environmental and biomedical engineering area. There is a variety of distributed parameter systems considered in our current research efforts:

Optimal and model predictive control (MPC) of chemical packed-bed and tubular reactors. In particular, realizable optimal control formulations have been developed in order to address important aspects of actuator constraints (constraints with respect to capacity of actuator and/or spatial location of actuator implementation), state/output constraints (ability of control configuration to satisfy constraints), state constraints and optimality of the control law with respect to actuator/sensor configuration. Classical and modern control controller realizations are merged and demonstrated in the process systems which are of industrial interest.[singlepic id=3 w=320 h=240 float=]
Optimal and model predictive control of crystal growth processes. The optimal control of crystal growth processes is one of the open problems in the semiconductor process industry. In particular, the CZ-Si crystal growth process is a prime example of multi-physics system in which pulling of the crystal and thermal effects affect the quality of the grown crystal. The fluctuations in the crystal temperature distribution and rate at which the crystal cools may induce large thermoelastic stresses leading to defect and dislocation generation.

In addition, to the crystal quality the optimal operating conditions are desirable for CZ crystal growth process. The current crystal growth processes operate at slow crystal growth rate (usually 3-5cm/h). Therefore, it is of the heigh importance to provide an optimal faster growth rate that will be integrate with the tight temperature regulation and pulling mechanism control.

Mixing in chemical reaction systems. An understanding of viscous mixing is of considerable technological importance in the context of materials processing, reactive and non-reactive polymer processing, food processing and/or stabilization of hazardous waste. In most cases mixing is to be controlled and enhanced, while in few cases suppressed. Mixing barriers are consequence of particular time varying flow features which can be identified by the framework of Lagrangian coherent structures.

Lagrangian Coherent Structures (LCS)

Within the area of dynamical systems, we consider fluid flow problems within the framework of dynamical systems and it is well-known that there exist special flow patterns within the fluid domain which influence the transport structure and mechanics of the flow. Such structures for a general time-dependent flow, as finite-time invariant manifolds in the extended phase space, and it can be captured in the Lagrangian framework of trajectories.

We consider analysis of the flow melt dynamics in two the most important crystal growth processes Czocharlski crystal growth and Bridgman crystal growth process. In the CZ-Si crystal growth process, as it can be shown in left figure, there is a formation of the LCS barrier (red line) which is placed in the vicinity of the melt-solid interface and identifies the pocket within the melt flow such that impurities captured within the pocket will not get back in bulk flow and inevitably will be incorporated in the grown crystal. In the case of Bridgman process, the LCS barriers vary with time as they drift with the time-varying flow emanating from the thermal and buoyancy driven instabilities.[nggallery id=2]

Finite element numerical realization of CZ crystal growth

Environmental monitoring. Identification of barriers to the transport of pollutants (fume, dust, or toxic chemicals released from the oil production plant) by investigation of the wind flow field by the method of Lagrangian Coherent Structures (LCS) in the territory of West and Central Canada is the most accurate and reliable way to determine the real dynamical distribution of the pollutant transported from the source to other regions of interest (case study: Wildfire in BC August 2010). In particular, we would like to develop a computing tool for petrochemical industry, which will 1.) integrate and utilize real-time wind weather data available from Canadian Daily Climate Data (CDCD); 2.) compute the flow map of the wind in the provinces and territories (BC,AB,SW,MN,YU,NW) and identify the barriers and patterns of possible pollutant transfer; 3.) provide an analysis on flow [singlepic id=1 w=320 h=240 float=right], patterns and pollution associated with the fume with respect to historically available data in the case of large scale BC wildfire in August 2010, 4.) provide the petrochemical industry with a reliable set of tools which can be used for hazardous events handling with respect to release of toxic vapor chemicals and monitoring of their dispersion through the domain of interest. Finally, the same tools can be used to estimate the transport barriers and determine the dynamic evolution of the oil spill and/or other chemicals on the Athabasca river, the Great Lakes, Hudson Bay and/or ocean. (see LCS contours represented as strong coloured red lines in the area of western Canada)

Process industry monitoring and control. Our research work in this direction has led to the development of efficient model predictive algorithms that were able to account for the infinite-dimensionality of the distributed parameter system modeled by parabolic partial differential equations, and provide optimal stabilization of linear and nonlinear systems subject to spatially distributed input and state constraints. [singlepic id=5 w=320 h=240 float=right] Along the same line the challenging and industrially appealing problem of boundary controlled system modelled by parabolic PDEs is explored. We demonstrate that formulated synthesis of model modal predictive controllers enforces a closed-loop stabilization in the presence of state and input constraints, and enforces optimal evolution of the spatially distributed state, so that prespecified input and state constraints are not violated.
Development of in-pipe swim robot (swimbot) device which combines novel methods of observer designs for pipeline systems. In particular, we develop the prototypes of the swimbot which is cheap, reliable and we demonstrate that frequent employment of the swimbot monitors pipeline integrity and adds to the already existing monitoring methodologies. Current swimbot realization records sound and it is equipped with IMUnits (Inertial Measurement Units) which record the positions of the swimbot within the pipeline. We combine the information about pipeline integrity obtained from swimbot with our observers designed for the transport systems, which yields very successful monitoring and leak detection device.
Environmental monitoring of water resources quality by Autonomous Sailing Vessel (ASV). We are developing an autonomous sailing vessel to monitor the water quality in the lakes, bays and ponds. The ASV is integrated with GPS, IMU and imaging in order to provide a real time reliable monitoring of the water quality or surface conditions. In particular, we are looking in improvements of known designs which serve as tools for integrated design and control of moving environmental surveillance agents.

DMOC & MPC controller synthesis (Robotics & Process Industry)

Discrete mechanics optimal control (DMOC) and model predictive control (MPC) synthesis for reaciton-diffusion process system with moving actuator. In this research efforts we are interested in fusion of control designs that include the dynamics motion of rigid bodies (robotic arms) and industrial process systems described by the underlying reaction-diffusion or reaction-conduction systems. In particular, an industrial important welding process is treated and subsequent model modal predictive control (MMPC) formulation which includes constraints on the available input injection and on allowable temperature profile is developed. The actuator arm dynamics is mechanical system which is represented through a discrete Lagrange-d'Alembert variational principle suitable for robust numerical integration and optimization purposes. The resulting discrete mechanical optimal control (DMOC) problem is combined with the constrained optimization structure emerging from the MPC realization. The DMOC methodology is used since one can include the holonomic and nonholonomic constraints (position and arm velocity constraints) in the proposed framework.

INDUSTRIAL ACTIVITIES

Irrigation and farming proces optimization. We are working on providing the customer the optimal watering, and crop surveillance paradigm which includes maintenance of water resources, monitoring of crop development and seasonal forecast of environmental conditions.
Pharmaceutical and Biotechnology applications. We provide biochemical and pharmaceutical systems analysis, production and quality analysis and pilot plant studies which are conducted in the accordance with safety and environmental standards.
Food and Beverage Production Analysis. Our Lab provides production assessment of the industrial or/and pilot plant projects. We apply statistical process control techniques which include scheduling, optimization and other improvements of operational conditions.
Water and wastewater treatment services. Our Lab provides automation and monitoring solutions for water and/or wastewater treatments with process control analysis and product production improvements. We have laboratory equipment that can provide insight into the plant operational modes and provide a monitoring and performance assesments.
HVAC, building automation and district heating . Our Lab provides insight and performs analysis of the HVAC and other thermodynamical systems of interest in rigorous and accurate way with the novel insights in the performance and analysis of cost saving features.
Iron and steel production. Our Lab provides technical and software support for analysis and operational improvements of iron and steel production and manufacturing plants. We have experience in modelling and providing monitoring modules for the iron casting processes and improving the quality of the manufactured steel.
Power Generation.
Pulp and paper.
Mining and Ore extraction.
Regulation of Diesel Engine Emission

Heavy-duty diesel engines play an important role in transportation and power generation applications, power systems for vehicles and industrial equipments. However, essential disadvantages of diesel engines contain the emission of significant levels of particulate matter and oxides of nitrogen (NOx), which are known to have detrimental health and environmental effects. As a consequence, manufacturers have developed emission control technologies in order to meet or exceed mandated requirements. The main components of the diesel engine emission system include the diesel oxidation catalyst(DOC), which is oxidises carbon monoxide and hydrocarbons, a particulate filter (DPF) to capture soot and selective catalytic reduction(SCR) .Within the SCR section of the emission control system, NOx is catalytically reduced to nitrogen and water using ammonia. Developing reliable dynamic models and control techniques aimed at the operation of SCR has attracted attention of researchers in academia and industry. To balance high NOx reduction efficiency and low ammonia slip, several urea dosing control techniques have been recently proposed.

However, these techniques developed to date use models comprised of ordinary differential equations (ODEs) to control the SCR and use inlet and outlet sensors for the state and parameter estimation. The main drawback of these techniques is that models developed based on ODE cannot capture important SCR dynamics, so the highest performance of the SCR cannot be achieved; however, good performance can be achieved by using models involving a large number of ODEs at the expense of computational time. In the present work, we are employing the full complex SCR model which is characterized as distributed parameter systems. More precisely, the complex SCR model is consisting of coupled hyperbolic and parabolic partial differential equations (PDEs). In the present work, we are designing a feedforward boundary controller to address the reduction problem of the amount of NOx emissions.

Plug Flow Crystallization (PFC)

Plug Flow Crystallizers (PFCs) are one of the most widely employed forms of continuous crystallizers. PFCs are usually selected for processes with fast kinetics and short residence times. Salient limitation of PFCs is that normally they do not operate at equilibrium conditions as a consequence of the short residence times and in addition the resulting yield is less then for batch processes. Recycling of the stream might improve the yield and be the potential avenue of improvements. Stability analysis and cycling operation conditions are addressed for PFC. From the stand point of practical application of control, monitoring and regulation we consider and explored the system in order to provide path to implementable application realization.

Oil & gas well production monitoring and optimization

Oil well production monitoring, control and optimization. The oil and gas industry development in Canada in the time of increased oil production and/or transportation costs, limited and deprived natural available resources and constant demands for higher production at the national and international level, is the main gear of maintaining prosperity and wealth at the national level. Moreover, in the publication of the Canadian Association of Petroleum Producers, 2010 CAPP Crude Oil Forecast, Markets & Pipeline - Production and Supply Data, which is widely referred to by the public, industry and government alike as a benchmark for crude oil forecasts, it is noted that the conventional production from the matured oil fields is going to decline from projected 3.20 million barrel/day in 2015 to 2.98 million barrel/day in 2025. This motivates the oil & gas industry to take steps with the view of cost reduction through efficient oil & gas production line segments. In particular, artificial lift is integral to oil production operations, especially in mature producing oil well fields, and the sucker rod pump system is the most often operating oil well realization (rod pumps account for 85%-90% of the pumps used in Canada/USA. The artificial lift technology covers a broad spectrum involving several types of pumps, with each type having its optimum application environment, and requires additional specific operational expertise. In this research projects directions, we intend to provide a foundation steps for novel technological platform for the improvement of sucker-rod oil well production system.

Biomass pyrolysis and production

Biomass torrefaction converts raw biomass into a high-energy-density, hydrophobic, compactable and grindable solid, making it a suitable alternative to coal . In the torrefaction process, biomass is heated to 200 - 300 C at very slow rates in an inert environment, until hemicellulose and some of the lignin in the wood are devolatilized. Due to its complex kinetics, many parameters affect the torrefaction process, such as the temperature in the reactor, the heating rate and the particle size and composition of the feedstock.

An auger reactor consists of a stationary vessel and a rotating screw inside the stationary vessel. The auger reactor is especially useful when dealing with particulate flow, since it eliminates the need for a carrier gas. In our work, an auger reactor was considered for biomass torrefaction.

We developed a dynamic model for the auger reactor for biomass torrefaction, based on unsteady mass and energy balances of the distributed parameter system. The model was solved using the method of characteristics. Simulations of the model were used to evaluate the effect of different operating conditions in the quality of the biomass produced. In addition, we developed an observer capable of reconstructing the temperature profile in the reactor, based on the temperature measurement at the exit of the auger.

Cardiac system dynamics and control

In the are of biomedical and biological systems, our research considers investigation of excitable media which are the signature of a complex system behavior describing underlying mechanisms of inter and intra cell signaling and muscle contraction. In particular, a wide variety of geometrically complex spatio-temporal phenomena, of which the most recognizable and striking are the rotating spiral waves, are explored from the point of design and control of wave propagation patterns.[singlepic id=6 w=320 h=240 float=right] The feedback controlled dynamics of meandering spiral waves which is considered as one of the possible mechanisms of cardiac fibrillation undergoes thorough investigation in the light of cardiac muscle dynamics, as it can be viewed as a path to the prevention and annihilation of the reentering spiral waves in the heart muscle. Feedback structure is used to move the tip of randomly meandering spiral wave in the center of the medium. Large scale simulations of anatomical tissue can reveal the nature of propagation of the electrical wave in the myocardium. In this area, development of the FEM (Finite Element Method) and FDM (Finite Difference Method) is performed in order to obtain realistic evolution of electric activity in monodomain tissue. Other research efforts consider development of pacing protocols for the stabilization of detrimental arrhythmias. New research topics include: