Innovative Projects Realized

Improved unsteady aerodynamic influence coefficients

Advanced statistical signal processing algorithms for precise radar positioning

Research on data and operation transformation techniques

Development for bioresponsive building envelopes

X-rAID: computer aided diagnosis of x-ray images

Comparing standard and extreme VaR models during highly volatile periods

FiX – Flash-in-XNA (Xbox New Architecture)

Separating mixtures using single-file and dual-mode diffusion

Low-cost Machine Type Communication User Equipments for LTE (Part 2)

Investigating the structural behaviour of Hybrid Fibre Reinforced Concrete (HFRC) plates

Failure of reinforced concrete plates is caused by the decrease in its stiffness as a result of cracking. This leads to corrosion of the steel reinforcement. Conventional concrete mixtures may not be sufficient to provide the tensile resistance required to minimize cracking. Using fibre reinforced concrete (FRC) can be considered a practical and economical solution to increase concrete tensile strength, and to protect the steel reinforcement against corrosion in harsh environments. Both steel and synthetic fibres are now used together to enhance concrete toughness with minimal impact on concrete workability and constructability; this is referred to as hybrid fibre reinforced concrete (HFRC).

The proposed research project includes the assessment of existing research data available in the literature to evaluate the effects of hybrid fibres on the behaviour of concrete members. The proposed project involves development of the HFRC mixture and testing of eight HFRC plates to evaluate the effects of the hybrid fibres on the ultimate load capacity and crack propagation in HFRC plates.

The capacity of the tested HFRC concrete plates will be measured and compared to identical samples cast with normal concrete. This will enable quantification of the effects of fibres on the behaviour and capacity of the plates.

This research project was undertaken and completed with a grant from and the financial assistance of Petroleum Research Newfoundland & Labrador.

Lifecycle Analysis of the Nutrient Cycle of Biofuels

If Canada or any other nation is to address its long-term transportation energy needs by massively expanding the use of biofuels, we must be sure that such an approach is truly sustainable. There are two main routes for biofuels: biological (such as fermentation of corn to produce ethanol) and thermochemical (such as the gasification of biomass to produce syngas). Each technique has its own strengths and weaknesses.
Gasification is the primary technology used in the thermochemical route, in which pulverized biomass is gasified at high temperatures into syngas, which contains H2, CO, and wastes such as CO2 and H2O. After cleaning, the H2 and CO can then be combined into more useful energy products such as diesel, gasoline, methanol, dimethyl ether, fuel-grade hydrogen gas, or burned for electricity. Since biomass is a renewable fuel, this process has very low net CO2 emissions and can even have net negative CO2 emissions if CO2 capture and sequestration techniques are employed.
However, many gasifiers (such as the entrained downward-flow variety) produce a solid waste called slag, containing the non-volatile components of the biomass (primarily metals in their oxide form). Currently, this is primarily used as an ingredient in asphalt. Although the carbon (i.e. CO2) in the biofuel is renewable since it perpetually cycles between plant and atmosphere, the metals (valuable plant nutrients such as phosphorus, iron, magnesium, etc.) are not returned to the earth in a useable form since they are sequestered in the slag. As a result, any massive effort to use slagging gasifiers to extract energy from biomass for any purpose will necessarily result in the gradual depletion of soil nutrients. These can be replaced with fertilizers which will then have some other affect, such as increased algal blooms, depletion of some other mineral resource, etc.
The question, then, is how much of an effect will this be? Is this a meaningful amount of soil nutrition being depleted from the biosphere, or is it vanishingly small such that there is no real concern? To answer these questions, a life cycle analysis on a biomass-to-liquids process must be performed which considers all of the cradle-to-the-grave effects and their secondary impacts on other contributions. For example, the increased amounts of processing for fertilizers in the long run may cause increased CO2 emissions, pollution, or resource depletion which could potentially offset the environmental gains achieved by a thermochemical biomass-to-liquids approach to our energy infrastructure.

Probability-based black-box branch-and-bound optimization for biofuel system design

We are developing new ways of producing sustainable biofuels from non-food-competitive biomass feedstocks grown in Canada, such as switchgrass, forestry bi-products, and other forms of lignocellulose. However, in order to produce enough biofuels for transformative change to our nation’s energy infrastructure, there are many systems issues in the supply chain and chemical processing which must be overcome. To address this, we are currently developing “semicontinuous” approaches to producing biofuels such as biobutanol (gasoline and ethanol substitute) and bio-dimethyl-ether (diesel substitute) to overcome these challenges at lower costs.

Although the semicontinuous approach has significant promise, it is incredibly complex, and as such traditional approaches to process design no longer apply. Although attempts at designing the process can be made “by hand”, a formal mathematical optimization technique is required to determine the key process parameters. Batch sizes, distillation parameters, heat duties, flow rates, transition behavior, controller tuning parameters, set-points, and many other parameters must be determined simultaneously in order to discover a configuration which achieves quality, sustainability, and profitability constraints. However, this is a significant challenge due to the high dimensionality of the problem and the character of the computer models on which our analyses are based. As such, we have found that existing optimization solvers (specifically, those which solve the class of problems known as “black-box”) are wholly inadequate for our needs.

Therefore, in order to assist in our work in biofuels process, we propose the development of a new optimization algorithm which is suitable not only for our particular biofuels problem, but for a large class of black-box optimization problems. The proposed approach is to combine well known stochastic solvers such as particle swarm optimization (PSO) or differential evolution with branch-and-bound techniques that systematically reduce the size of the problem to improve convergence toward a global optimum.

Branch-and-bound techniques work by systematically dividing the optimization problem’s “search” space into regions, and mathematically proving that the global optimum cannot be in one region or another, thus eliminating it from consideration. However, branch-and-bound requires explicit knowledge of the model equations in order to do this, which are unavailable for our problem and other black-box problems. Therefore, we propose a new probability-based algorithm gets around the problem of requiring explicit knowledge of model equations by creating implicit, approximations of the model using the knowledge gained by particle swarm optimization runs.

With this technique, we cannot completely eliminate one region of the search space, but should be able to estimate the probability that the global optimum should exist in one region or another. For black-box problems, this should be significantly faster and more likely to converge upon a true global optimum than the current state-of-the-art.