PRISM :: Browsing by Author "Westwick, David"

Browsing by Author "Westwick, David"

Now showing 1 - 20 of 37

Open Access
A Pre-Placement Individual Net Length Estimation Model and an Application for Modern Circuits
(2011) Farshidi, Amin; Behjat, Laleh; Westwick, David
Open Access
Algorithms for non-convex system identification
(2007) Jazayeri, Seyed Pouyan; Rosehart, William Daniel; Westwick, David
Open Access
Algorithms for the reduction of clutter in tissue sensing adaptive radar (TSAR) signals
(2007) Kurrant, Douglas John; Westwick, David; Fear, Elise
Open Access
A Biologically Inspired Supervised Learning Rule for Audio Classification with Spiking Neural Networks
(2021-06-15) Peterson, Dylan George; Leung, Henry; Westwick, David; Uddin, Gias; Wang, Xin
Audio classification has many practical applications such as noise pollution monitoring, wildlife monitoring, audio surveillance, speech recognition, and more. For many of these applications, deploying classifiers on low powered devices for persistent monitoring is desirable. Artificial neural networks (ANN) have achieved significant success for audio classification tasks. However, it may not always be feasible to deploy current state-of-the-art ANNs to embedded devices due to their memory footprint and power consumption. Biologically inspired neural networks, also known as spiking neural networks (SNN), have been shown to significantly reduce power consumption during inference when compared with equivalent ANNs. They have also been theoretically proven to be more computationally powerful per unit than ANNs. These two properties make SNNs an attractive solution for machine learning tasks on low powered embedded devices, such as at the edge in an Internet of Things (IoT) sensor network. However, SNNs tend to lag behind in performance when compared to ANNs. This is partially because training SNNs is difficult since the standard backpropagation algorithm is not directly applicable due to the non-differentiable spiking nature of SNNs. Encoding data into spike trains compatible with SNNs is also an unresolved question when applying SNNs. This work compares different spike encoding schemes for audio data, and a learning algorithm for multilayer SNNs inspired by biologically plausible learning rules is developed. The proposed learning rule is then successfully applied to simple pattern recognition and audio classification tasks.
Open Access
Blind identification of a Wiener-Bose structure using expectation maximization
(2011) Bu bshait, Abdullah Saad; Westwick, David
Open Access
Blind Identification of the Electromechanical Modes of a Power System using a Wiener Model
(2016) Hossain, Tasnim; Westwick, David; Rosehart, William; Zareipour, Hamidreza; Nowicki, Edwin Peter
In this thesis, a blind system identification technique using a Wiener model is used to estimate the power system modes. A Wiener model, a universal approximator consisting of a dynamic linear system followed by a memoryless nonlinear element, is used to estimate the power system nonlinearities. It also constructs a set of intermediate data, which can be used by a linear estimation technique such as subspace identification for estimating the power system electromechanical modes. In this research, a blind variant of a subspace method, Numerical Algorithm for Subspace State Space System IDentification (N4SID), is used. The algorithm is tested with simulation data from the Kundur two-area network. The accuracy and reliability of these estimates are accessed by carrying out Monte Carlo simulations. The estimated results obtained from the simulated system using a Wiener model were very accurate with reduced prediction errors and was a good fit for the power system.
Open Access
Data-Driven Modeling of Wind Turbine Structural Dynamics and Its Application to Wind Speed Estimation
(2016-01-27) Saberi Nasrabad, Vahid; Sun, Qiao; Westwick, David; Wood, David; Mohamad, Abdulmajeed
In wind turbine control systems, the wind speed measurement is used in order to derive the optimal shaft speed for achieving the Maximum Power Point Tracking (MPPT) and to adjust the pitch angle optimally for protecting the turbine from excessive loading. In this thesis, a tower detection based effective wind speed estimation method is proposed. The tower dynamics is identified using subspace system identification method. To estimate the effective wind speed, an online model-based aerodynamic thrust force estimator is designed and implemented using Kalman filter and recursive least square algorithm. The estimated aerodynamic thrust force is used as an input to a neural network estimator to solve the inverse aerodynamic thrust force equation and estimate the effective wind speed. Finally, the simulation results for effective wind speed estimation for a turbulent wind field are presented and an evaluation method based on correlation coefficient is used to validate the results.
Open Access
Development of an Artificial Pancreas: Control of Glucose in Type 1 Diabetics Using Model Predictive Control with a Low Order Model
(2024-06-13) Frayne, Mark Christopher Hill; Pieper, Jeff; Westwick, David; Bisheban, Mahdis
Type one diabetes is an autoimmune disorder that causes people to not produce hormones that govern glucose control, which can lead to many health issues. The current treatment for this disorder is the injection of insulin either manually or with a preprogramed pump. An artificial pancreas is a device that aims to improve this treatment by connecting the pump to a glucose sensor and feedback controller to automatically adjust the injection rate. This project proposes the use of Model Predictive Control (MPC) to close this loop. MPC was implemented using a low-order linear model of the human glucose system and applied to simple human kinetic models as well as a higher-order complex plant. In both cases the controller was able to significantly reduce the maximum glucose levels along with time-integrated excursion measures. In realistic simulations modelling human physiology, maximum hyperglycemia was reduced from 174 mg/dL to 120 mg/dL when a disturbance of 100g of carbohydrates was applied.
Open Access
Development of an Autotuner for Plants with Stochastic Disturbances
(2017-12-20) Brebnor, Dave; Foley, Michael; Clarke, Matthew; Svrcek, William; Westwick, David; Kwok, Ezra
The most universally exercised model which amply describes the transient behaviour of a wide range of chemical processes is the first-order-plus-deadtime model. An identification experiment must be conducted to generate the data required to fit an accurate first-order-plus deadtime model. Relay feedback procedures have superseded the conventional step response identification tests for the purpose of tuning proportional-integral-derivative controllers. However, these techniques are known to perform poorly in the presence of process disturbances. Developing an autotuner that performs well in the presence of stationary stochastic disturbances is the chief objective of this dissertation. This is accomplished by the inclusion of two separate first-order filters in the relay feedback loop. A few guidelines are provided for the selection of the filter time constants. After the identification procedure yields the process parameters, a model-based tuning rule can be employed for controller tuning. Simulations of a continuous stirred tank heater and experiments conducted on a pilot-plant stirred tank heater illustrate the feasibility of the proposed autotuner.
Open Access
Enhancing Efficiency in Residential PV Systems: Novel Maximum Power Point Tracking Strategy for Reduced DC Bus Capacitance in Differential Power Processing Architecture
(2024-07-11) Aguero Meineri, Adrian Nicolas; Galiano Zurbriggen, Ignacio; Gray, Philippe; Westwick, David; Tan, Benjamin
Tracking efficiency, cost, and reliability are important factors when selecting PV architectures and converter topologies. PV systems require power converters to maximize power extraction, for which DC-DC converters are a common choice. Differential Power Processing (DPP) architectures can achieve higher efficiencies and lower cost by reducing the amount of power passing through these converters, while still providing Maximum Power Point Tracking (MPPT) capabilities. Single-phase grid connected PV systems, which are the most popular choice in residential applications, require a large capacitance in the DC bus to minimize the voltage ripple caused by double-line pulsating power, which has impacts on the cost and reliability of the system. This work introduces a new MPPT mode of operation for flyback converters in DPP architectures. The proposed MPPT method shows extremely fast dynamic performance and is capable of maximizing power extraction, even for extreme variations in the bus voltage. In this way, the proposed method enables a significant reduction in the DC bus capacitance, which contributes to reducing costs and facilitating the use of ceramic capacitors, while maintaining excellent tracking efficiency. The analysis incorporates comprehensive models that characterize the large-signal dynamic behaviour of ideal and non-ideal flyback converters, and it is supported by detailed mathematical procedures. The system performance behaviour and limits are validated through simulation and experimental results.
Open Access
Identification of a model of sound transmission in the human knee: vibroarthrographic signals as a diagnostic tool
(2005) Dempsey, Erika; Westwick, David
Open Access
Identification of linear, time-varying systems
(2005) Sanyal, Sreemoyi; Westwick, David
Open Access
Identification of time-varying nonlinear systems
(2007) Ikharia, Bashiru Isa; Westwick, David
Open Access
Identifying the Sources and Parameters of Disturbances in Power Systems
(2022-06-19) Mansouri Habibabadi, Mohammad; Knight, Andrew Michael; Westwick, David; Nielsen, Jorgen S; Pahlevani, Majid; Nowicki, Edwin Peter; Chen, Yu (Christine)
The high penetration of renewable energy in power systems makes power systems dynamically more complicated; hence, a power system faces frequent disturbances, including electromechanical oscillations, harmonics and subharmonics, and forced oscillations. Furthermore, because a power system is dynamically complicated, the classic methods that mainly rely on the system’s dynamic model cannot work as it is hard to have the details of the power systems dynamic or is impossible. Therefore, measurement-based methods, which do not need the system dynamic, have received significant attention in recent years. This dissertation offers measurement-based methods to identify and mitigate three disturbances, including electromechanical oscillations, harmonics and subharmonics, and forced oscillations. The contribution of this research can be divided into four parts.The first contribution proposes a measurement-based strategy to estimate the parameters of the multi-mode electromechanical oscillations. This strategy employs the cascade structure of Damped-SOGI to estimate the parameters of electromechanical oscillations. The advantages of the proposed method are the capability of online implementation, robustness against noise, no need for dynamics of the system, no need to pre-filtering and pre-processing, and simplicity.The second contribution of this thesis is an algorithm to mitigate electromechanical oscillations. In this algorithm, the dynamic of electromechanical oscillations is extracted by the Damped-SOGI; afterward, this dynamic is placed in the control loop. Thus, the electromechanical oscillations are rejected from the power systems output based on the Internal Model Principle (IMP). Computer simulation and experimental results confirm the effectiveness of the proposed method.An algorithm to identify the source of harmonics and subharmonics is proposed as the third contribution of this dissertation. In this algorithm, the N4SID, an identification method, extracts the power system’s dynamics and inputs. Afterward, the source of the subharmonic is identified by studying the correlations between various extracted inputs. Simplicity and robustness against noise are of the advantages of the proposed algorithm.Locating the source of forced oscillations is tackled by subspace identification in the fourth contribution of this thesis. In this algorithm, the dynamic model of the power system is identified by subspace identification. Using the identified model, the dynamic responses of the forced oscillation are extracted in each PMU location. Finally, the source of forced oscillation is deter- mined by studying the magnitude and phase or the correlations of extracted forced oscillations. The capability to distinguish between forced oscillations and electromechanical oscillations is of the advantages of the proposed algorithm.
Open Access
Integrated Expansion Planning of Electricity, Heat, and Gas in Presence of Demand and Wind Uncertainty
(2022-06-22) Mozafari Jovein, Yasaman; Rosehart, William; Bergerson, Joule; Westwick, David; Zinchenko, Yuriy; Yanushkevich, Svetlana; Gokaraju, Ramakrishna
Recent shift towards higher renewable penetration in power systems, has resulted in increasing gas-fired generation capacity in power systems, implying electricity and gas infrastructure interdependency. Furthermore, highly efficient combined heat and power (CHP) units lead to heat and electricity interdependency. It is crucial to consider these interdependencies in the expansion planning of energy systems for an effective and reliable investment planning and policy design. In this thesis, the problem of integrated expansion planning of electricity, heat, and gas in presence of demand and wind uncertainty is addressed. A comprehensive multi-area planning model including generation, CHPs, boilers, transmission network, and gas pipeline expansion is proposed. The advantages of integrated approach versus the non-integrated approaches in terms of cost and emission is illustrated through simulations on a simplified Alberta energy system model. Representative operating scenario selection method is used to model wind and demand uncertainties, and to address computational complexity of a central planning model. Load duration curve technique is compared with k-means and k-means++ clustering, the impact of initialization is investigated, and the impact of spatial data correlation on the solution is analyzed. To overcome drawbacks of k-means clustering, application of algebraic multi-grid clustering in scenario selection for integrated energy system expansion planning is explored. The simulation results on a modified IEEE-118 bus test system and a 14-node gas network shows that the algebraic multi-grid clustering outperforms the classical methods such as k-means by following the benchmark case more closely. Finally, to ensure robustness of the obtained investment plan, an adaptive robust optimization model including electricity demand, heat demand, and wind uncertainty is proposed. The model ensures reliability targets in both electricity and heat sector are met. The simulation results on the IEEE-118 bus test system verify the effectiveness of the proposed model in dealing with uncertainties and meeting future electricity and heat demand reliably.
Open Access
Microelectrodes for Neural Stimulation: Effects of Geometry
(2013-10-08) Ghazavi Khorasgani, Atefeh; Westwick, David; Dalton, Colin
Multi-electrode arrays are non-invasive devices for neural stimulation in vivo and in vitro. Improving the efficiency of these devices is desired for stimulating neurons over extended periods of time. In this work, changing the geometry of the electrode used in the array is investigated. This approach will improve efficiency without requiring fundamental changes in the fabrication process, allowing for ease of implementation. A model is presented to study features that provide optimum stimulation threshold from different sizes and shapes of electrodes. Specially, a single neuron-electrode interface was modelled and cell depolarization generated by stimulating the cell by different electrode shapes were compared. The geometries investigated were star, spiral, serpentine, and circular electrode shapes. Based on the simulations, the electrode geometries were designed and then fabricated into a planar microelectrode array test device. Proof of principle in vitro experimentation was then conducted, with the results being compared to the simulations.
Open Access
Model Predictive Control of DFIG-Based Wind Power Generation Systems
(2013-04-12) Soliman, Mostafa; Malik, Om; Westwick, David
Novel control strategies that improve the cost effectiveness of wind energy conversion systems are proposed in this thesis. The main focus is on grid-connected variable-speed variable-pitch wind turbines equipped with doubly fed induction generators (DFIGs). At the wind turbine control level, a multivariable control strategy based on model predictive control techniques is proposed. The proposed strategy is formulated for the whole operating region of the wind turbine, i.e., both partial and full load regimes. The pitch angle and generator torque are controlled simultaneously to maximize energy capture, mitigate drive train dynamic loads, and smooth the power generated while reducing the pitch actuator activity. This has the effect of improving the efficiency and the power quality of the electrical power generated, and increasing the life expectancy of the installation. Extensive simulation studies show that the proposed control strategy provides superior performance when compared to classical control strategies commonly used in the litterature. For applications having fault tolerant control requirements, such as offshore wind farms, a new wind turbine control strategy based on adaptive subspace predictive control is proposed. In contrast with subspace predictive control algorithms previously proposed in the literature, the proposed strategy ensures offset-free tracking. The effectiveness of the proposed strategy is illustrated by simulating a wind turbine under normal operation and a fault in the hydraulic pitch system. Another control problem considered in this thesis is the design of the generator control system to ensure fault ride through for DFIG-based wind turbines. This requirement is dictated by recent grid codes, and it necessitates that the DFIG should be connected to the grid and capable of providing reactive power support during large voltage dips. This is challenging for DFIG-based wind turbines due to their partially rated power converters. In this thesis, a novel control strategy, based on using model predictive control and a dynamic series resistance protection scheme, is proposed to ensure fault ride through requirement.
Open Access
Modelling, simulation and identification of induction machines in continuous and discrete time
(2004) Pan, Juntao; Westwick, David; Nowicki, Ed P.
Open Access
A Multi-Purpose Continuum Robot for Minimally Invasive Surgery
(2023-01-25) Esfandiari, Mojtaba; Sutherland, Garnette; Tavakoli, Mahdi; Westwick, David; Goldsmith, Peter
This thesis is about ''A Multi-Purpose Continuum Robot for Minimally Invasive surgery'' which consists of five chapters. It starts with an Introduction and Literature Review that studies some of the most famous surgical robotic systems and analyses their pros and cons. Chapter two has to do with the problem statement and challenges that need to be addressed while designing our specific surgical tool, and a CAD design of a flexible continuum robot will be done for our brain surgery application. In chapter three, the kinematic modeling of the proposed robot is analyzed by a new model based on the Euler spirals and the results are compared with conventional constant curvature models. In chapter four, a model predictive control algorithm is proposed that considers the input saturation constraints on robot actuators. Finally, a discussion and conclusion will be provided in chapter five.
Open Access
Neural-adaptive control of coker-off boilers in the heavy-oil industry
(2011) Mahmoodi Takaghaj, Sanaz; Macnab, Chris; Westwick, David