Search

This research paper presents part of a research project that develops the participative design of innovative solar control devices integrated in school architecture using low-cost tensed structures, composed of reused household waste and renewable energy systems. In consequence, these devices are designed to be built during workshops about waste management and sustainability by elementary and high school students and teachers. In general, this project aims to reduce the negative environmental impacts by educational architecture and, at the same time, increase students and teachers´ comfort levels and sustainability awareness. The specific case studies in this project occur in Spanish schools, which have a particular endemic problem in terms of user comfort due to lack of solar control. The main novelty of this project is the aforementioned design of solar control devices composed of household waste to be assembled during workshops by students and teachers. Although there have already been previous workshops in which students learn about school buildings, this project moves a step forward and involves students and teachers in the construction process of a solar control device. This article presents three solar control waste based device prototypes, the main steps followed to design and assess them and the first workshop.

This research investigates software failure and fault analysis through data-driven machine learning approaches. Faults can happen in any software system and may hugely impact system reliability and user experience. Log data, the machine-generated data that records the system status, is often the primary source of information to track down a fault. This study aims to develop automated systems that recognize recurring faults by analyzing the system log data. The methodology developed in this research applies to the Ford SYNC vehicle infotainment system as well as other systems that produce similar log data. Log data has been used in manual examination to help trace and localize a fault. This manual process can be effective and sometimes the only feasible way of troubleshooting software faults. However, as the amount of log data increases significantly with the growing complexity and scale of software, the manual workload can get overwhelming. During the system-level validation tests, all system components are producing log data, resulting in tens of thousands of lines of log messages in just a few minutes. Therefore, automated diagnosis has been a promising approach for log data analysis. Three machine learning approaches are investigated in this research to tackle the fault diagnosis problem: 1) the data mining approach; 2) the statistical feature approach; and, 3) the deep learning approach. The first method attempts to mimic human experts to examine log data. Log sequences representing a fault are extracted through data mining techniques and used to identify anomalies. The method is effective when applied to a small volume of data, but computational efficiency can be an issue when scaling to larger datasets. As its name suggests, the second method involves an examination of the log data’s statistical and numerical features and adapting a machine learning model for decision making. The use of numerical features to describe log data has significant computational efficiency improvement over working directly with sequences. The last approach adopts deep learning models that process the log data in sequential format, enabling more sophisticated feature extraction that often exceeds human capability. In this research, all three methods are implemented and evaluated in a controlled testing environment, and their strengths and weaknesses are comparatively evaluated. This study also reports on a novel finding that the time information in a log sequence plays an important role in distinguishing a faulty condition from a normal one. For most software systems, the log sequences are unevenly spaced, meaning that the timestamps associated with log data are nonuniform. Existing log analysis studies generally overlooked the time information while emphasizing log sequences. This research proposes a novel deep learning structure to unify the processing of timestamps and log sequences. The timestamps are integrated through interpolation at an intermediate layer of a neural network. Testing results demonstrate that the inclusion of timestamps makes a significant contribution to identifying a fault, and that models using time stamps can push the performance to a higher level. Dissertation Doctor of Engineering (DEng)

Lithium-ion (Li-ion) batteries are amongst the most commonly used types in Electric (EVs) and Hybrid Electric (HEVs) Vehicles due to their high energy and power densities, as well as long lifetime. A battery is one of the most important components of an EV and hence it needs to be monitored and controlled accurately. The safety, and reliability of battery packs must then, be ensured by accurate management, control, and monitoring functions by using a Battery Management System (BMS). A BMS is also responsible for accurate real-time estimation of the State of Charge (SoC), State of Health (SoH) and State of Power (SoP) of the battery. The battery SoC provides information on the amount of energy left in the battery. The SoH determines the remaining capacity and health of a pack, and the SoP represents the maximum available power. These critical battery states cannot be directly measured. Therefore, they have to be inferred from measurable parameters such as the current delivered by the battery as well as its terminal voltage. Consequently, in order to offer accurate monitoring of SoC, SoH and SoP, advanced numerical estimation methods need to be deployed. In the estimation process, the states and parameters of a system are extracted from measurements. The objective is to reduce the estimation errors in the presence of uncertainties and noise under different operating conditions. This thesis uses and provides different enhancements to a robust estimation strategy referred to as the Smooth Variable Structure Filter (SVSF) for condition monitoring of batteries. The SVSF is a predictor-corrector method based on sliding mode control that enhances the robustness in the presence of noise and uncertainties. The methods are proposed to provide accurate estimates of the battery states of operation and can be implemented in real-time in BMS. To improve the performance of battery condition monitoring, a measurement-based SoC estimation method called coulomb counting is paired with model-based state estimation strategy. Important considerations in parameter and state estimation are model formulation and observability. In this research, a new model formulation that treats coulomb counting as an added measurement is proposed. It is shown that this formulation enhanced information extraction, leading to a more accurate state estimation, as well as an increase in the number of parameters and variables that can be estimated while maintaining observability. This model formulation is used for characterizing the battery in a range of operating conditions. In turn, the models are integral to a proposed adaptive filter that is a combination of the Interacting Multiple Model (IMM) concept and the SVSF. It is shown that this combined strategy is an efficient estimation approach that can effectively deal with battery aging. The proposed method provides accurate estimation for various SoH of a battery. Further to battery aging adaptation, measurement errors such as sensor noise, drift, and bias that affect estimation performance, are considered. To improve the accuracy of battery state estimation, a noise covariance adaptation scheme is developed for the SVSF method. This strategy further improves the robustness of the SVSF in the presence of unknown physical disturbances, noise, and initial conditions. The proposed estimation strategies are also considered for their implementation on battery packs. An important consideration in pack level battery management is cell-to-cell variations that impact battery safety. This study considers online battery parametrization to update the pack’s model over time and to detect cell-to-cell variability in parallel-connected battery cells configurations. Experimental data are used to validate and test the efficacy of the proposed methods in this thesis. Thesis Doctor of Philosophy (PhD) To address the critical issue of climate change, it is necessary to replace fossil-fuel vehicles with battery-powered electric vehicles. Despite the benefits of electric vehicles, their popularity is still limited by the range anxiety and the cost determined by the battery pack. The range of an electric vehicle is determined by the amount of charge in its battery pack. This is comparable to the amount of gasoline in a gasoline vehicle’s tank. In consideration of the need for methods to address range anxiety, it is necessary to develop advanced algorithms for continuous monitoring and control of a battery pack to maximize its performance. However, the amount of charge and health of a battery pack cannot be measured directly and must be inferred from measurable variables including current, voltage and temperature. This research presents several algorithms for detecting the range and health of a battery pack under a variety of operating conditions. With a more accurate algorithm, a battery pack can be monitored closely, resulting in lower long-term costs. Adaptive methods for determining a battery’s state of charge and health in uncertain and noisy conditions have been developed to provide an accurate measure of available charge and capacity. Methods are then extended to improve the determination of state of charge and health for a battery module.

The perception system is the key for autonomous vehicles to sense and understand the surrounding environment. As the cheapest and most mature sensor, monocular cameras create a rich and accurate visual representation of the world. The objective of this thesis is to investigate if camera-based deep learning models with transfer learning technique can achieve 2D object detection, License Plate Detection and Recognition (LPDR), and highway lane detection in real time. The You Only Look Once version 3 (YOLOv3) algorithm with and without transfer learning is applied on the Karlsruhe Institute of Technology and Toyota Technological Institute (KITTI) dataset for cars, cyclists, and pedestrians detection. This application shows that objects could be detected in real time and the transfer learning boosts the detection performance. The Convolutional Recurrent Neural Network (CRNN) algorithm with a pre-trained model is applied on multiple License Plate (LP) datasets for real-time LP recognition. The optimized model is then used to recognize Ontario LPs and achieves high accuracy. The Efficient Residual Factorized ConvNet (ERFNet) algorithm with transfer learning and a cubic spline model are modified and implemented on the TuSimple dataset for lane segmentation and interpolation. The detection performance and speed are comparable with other state-of-the-art algorithms. Thesis Master of Applied Science (MASc)

Predictive maintenance is extremely important to fleet owners. On-duty automobile engine failures add cost of extra towing, gas and labor expenses which can add up to millions of dollars every year. Early knowledge of upcoming failures helps reduce these expenses. Thus, companies invest considerably in fault detection and diagnosis (FDD) systems to reduce unnecessary costs. Artificial Intelligence (AI) is getting increasingly used in the data driven signal based FDD industry because it requires less labor and equipment. It also results in higher productivity since it can operate continuously. This research offers Artificial Intelligence based solutions to detect and diagnose the degradation of three Internal Combustion Engine (ICE) parts which may cause on-duty failures: lead-acid accessory battery, spark plugs, and Exhaust Gas Recirculation (EGR) valve. Since the goal behind most FDD systems is cost reduction, it is important to reduce the cost of the FDD test. Therefore, all the FDD solutions proposed in this research are based on three types of built-in sensors: battery voltage sensor, knock sensors and speed sensor. Furthermore, the engine database, the Machine Learning (ML) and Deep Learning (DL) models, and the virtual operating machines were all stored and operated in the cloud. In this research, eight Machine Learning (ML) and Deep Learning (DL) models are proposed to detect degradations in three vehicle parts mentioned above. Additionally, novel advanced pre-processing approaches were designed to enhance the performance of the models. All the developed models showed excellent detection accuracies while classifying engine data obtained under artificially and physically induced fault conditions. Since some variant data samples could not be detected due to experimental flaws, defective sensors and changes in temperature and humidity, novel pre-processing methods were proposed for Long Short-Term Memory Networks (LSTM-RNN) and Convolutional Neural Networks (CNN) which solved the data variability problem and outperformed the previous ML/DL models. Thesis Doctor of Philosophy (PhD) Predictive maintenance is extremely important to fleet owners. On-duty automobile engine failures add cost of extra towing, gas and labor expenses which can add up to millions of dollars every year. Early knowledge of upcoming failures helps reduce these expenses. Thus, companies invest considerably in fault detection and diagnosis (FDD) systems to reduce unnecessary costs. Artificial Intelligence (AI) is getting increasingly used in the data driven signal based FDD industry because it requires less labor and equipment. It also results in higher productivity since it can operate continuously. This research offers Artificial Intelligence based solutions to detect and diagnose the degradation of three Internal Combustion Engine (ICE) parts which may cause on-duty failures: lead-acid accessory battery, spark plugs, and Exhaust Gas Recirculation (EGR) valve. Since the goal behind most FDD systems is cost reduction, it is important to reduce the cost of the FDD test. Therefore, all the FDD solutions proposed in this research are based on three types of built-in sensors: battery voltage sensor, knock sensors and speed sensor. Furthermore, the engine database, the Machine Learning (ML) and Deep Learning (DL) models, and the virtual operating machines were all stored and operated in the cloud. In this research, eight Machine Learning (ML) and Deep Learning (DL) models are proposed to detect degradations in three vehicle parts mentioned above. Additionally, novel advanced pre-processing approaches were designed to enhance the performance of the models. All the developed models showed excellent detection accuracies while classifying engine data obtained under artificially and physically induced fault conditions. Since some variant data samples could not be detected due to experimental flaws, defective sensors and changes in temperature and humidity, novel pre-processing methods were proposed for Long Short-Term Memory Networks (LSTM-RNN) and Convolutional Neural Networks (CNN) which solved the data variability problem and outperformed the previous ML/DL models.

Battery Electric Vehicles (BEV) are gaining market share but still must overcome several engineering challenges related to the lithium-ion battery packs powering them. The batteries must be carefully managed to optimize safety and performance. The estimation of battery states, which cannot be measured directly, is an important part of battery management and remains an active area of research since small gains in estimation accuracy can help reduce cost and increase BEV range. This thesis presents several improvements to battery state estimation using different methods. Electrochemical Impedance Spectroscopy (EIS) is receiving increased attention from researchers as a method for state estimation and diagnostics for real-time applications. Due to battery relaxation behaviour, long rest times are commonly used before performing the EIS measurement. In this work, methods were developed to significantly shorten the required rest times, and a State of Health (SoH) estimation strategy was proposed by taking advantage of the relaxation effect as measured by EIS. This method was demonstrated to have an estimation error of below 1%. At low temperatures, the accuracy of the battery model becomes poor due to the non-linear battery response to current. By using an adaptive filter called the Interacting Multiple Model (IMM) filter, the next part of this work showed how to significantly improve low temperature State of Charge (SoC) estimation. Further reduction in estimation errors was achieved by pairing the IMM with the Smooth Variable Structure Filter (SVSF), for SoC estimation errors below 2%. The work presented in this thesis also includes the application of Deep Neural Networks (DNN) for SoC estimation from EIS data. Finally, an extensive aging study was conducted and an accelerated protocol was compared to a realistic drive cycle based protocol using EIS as a characterization tool. Thesis Doctor of Philosophy (PhD) Replacing conventional gasoline/diesel powered cars with battery powered vehicles is part of a solution to the climate crisis. However, the initial costs paired with range anxiety stops many from switching to electric cars. Both cost and range are related to the battery pack. To achieve the best possible range for the lowest possible cost, battery packs must be carefully controlled by sophisticated algorithms. Unfortunately, battery range or health cannot be measured directly, but must be inferred through measurable indicators. This thesis explores battery behavior under different operating conditions and develops improved methods which can be used to determine battery health and/or range. A powerful method usually used only in laboratory settings is studied and improved to make it more suitable for implementation in electric cars. In this work it is used for accurate battery health determination. Furthermore, a strategy for improving battery range determination at low temperatures is also proposed.

Modern Advanced Driver Assistant Systems (ADAS) focus more on daytime driving and primarily use daylight cameras as the main vision sources to detect, classify, and track objects. However, evidence has proved that autonomous driving using such a setup is compromised in the dark, and consequently, resulting in accidents. The hypothesis is that adding an infrared camera to the existing ADAS will boost the detection rate and accuracy, and further enhance the overall safety. This thesis investigates how well a standalone infrared camera performs onboard vehicle perception tasks such as object detection and classification using both machine learning and deep learning algorithms. Given a custom labeled infrared driving dataset that contains 4 classes of objects, “People”, “Vehicle”, “Bicycle”, and “Animal”, multiple attempts and improvements of training a supervised learning model, namely the linear multi-class Support Vector Machine (SVM) has been made by using various image preprocessing and feature extraction methods to detect the objects. During training, hard example mining is used to reduce the number of false classifications. This SVM employs a One-Against-All (OAA) styled approach and uses the image pyramid technique to enable multi-scale detection. On the deep learning side, a Convolutional Neural Network (CNN) based state-of-the-art detector, the YOLOv4 family including the full-sized and tiny YOLOv4 has been selected, trained, and tested at different input sizes using the same dataset. Labeling format conversion is performed to make this work. The results show that using bilateral filtering with the Histogram of Oriented Gradients (HOG) feature to train an SVM is preferable and is more accurate than the YOLOv4 family. However, the YOLOv4 networks are significantly faster. Overall, a standalone infrared camera cannot provide dominant detection results, but it can definitely supply useful information to the ADAS and complement other sensory devices for improved safety. Thesis Master of Applied Science (MASc)

Advancements in batteries, microprocessors as well as an extra emphasis being put on the environment has pushed electric vehicles to the forefront of today. Despite the many benefits of electric vehicles, range anxiety and long charge times are hurdles to overcome. These shortfalls are a result of the current battery technology regardless of the many breakthroughs over the last decade. Lithium-ion Batteries and other modern chemistries pose a number of challenges in testing and research when compared to the traditional lead acid batteries. Current test systems fall short in providing a complete testing solution with. The focus of this thesis is to develop a complete software framework for battery characterization: testing, modelling and characterization to accompany battery testing hardware developed by D&V Electronics. The first step in battery characterization, involves battery testing in order to obtain data. This required development of the test software and a number of battery tests, including: Charge and discharge, state of charge vs. open circuit voltage curve generation, Electro-Impedance Spectroscopy, and capacity test. Research was done in order to ensure developed test procedures lined up with that of other publications. All data from the testing data is logged to a central database, allowing for the second major development, the model framework. The model framework is composed of seven different battery models that can be parameterized with the touch of a button, using data collected from the tester. It is a software framework that is meant to be expandable by abstracting the details of a model from the tester. This allows for new models and parameterization techniques to be integrated into the software without the need of new software development. Lastly, all development was used to do a battery characterization of a prismatic battery cell. All tests were conducted on a battery over two hundred cycles, followed by battery parameterization using the mode framework. The battery models were then used to simulate a US06 drive profile and compared to the same profile with measurements taken from the tester. With an average root mean square error of 8 millivolts, the battery characterization using the framework proved to be a success. Thesis Master of Applied Science (MASc)

In this research, a fault detection and diagnosis strategy for internal combustion engine is developed using measurements that are readily available in engine testing environment to monitor abnormal combustion. The FDD strategy is designed to monitor the engine on a cycle-by-cycle basis using measurements that are accessible on a running vehicle. Pressure measurements are easily accessible in a testing facility that provide useful insight into the quality of the combustion occurring inside the engine. However, due to its cost and complex installation procedures, it is not feasible to obtain in-cylinder pressure measurements from an in-vehicle engine. Faults of a mechanical system are often investigated using vibration. Due to the low cost and non-invasive nature of accelerometers, vibration measurement is used to monitor the in-vehicle engine. However, as vibration behaviors of complex system such as an engine is hard to characterize, in cylinder pressure measurement is used during the development of the FDD strategy to assist in characterizing the vibration measurement. Upon data acquisition, features are extracted from the vibration measurements using Extended-MSPCA for better characterization and data reduction with a multi-baseline technique. Pressure measurements are analyzed using thermodynamic theories to assess the combustion quality of each cycle. The vibration measurements are labelled corresponding to the pressure analysis. An artificial neural network classifier is developed using the extracted and labelled features. Developed classifier detected the fault and its location with an overall accuracy of 96.3%. Thesis Master of Applied Science (MASc)

Modeling and Implementation of a Hardware Efficient Low-Voltage-To-Cell Battery Balancing Circuit for Electric Vehicle Range Extension One disadvantage of electric vehicles is their limited driving range when compared to internal combustion engine vehicles. Battery packs are also a significant cost to electric vehicle manufacturers, and lithium-ion battery cells must remain within controlled voltage limits. Thus, the requirements for the electric system are to be cost effective, perform battery management, and make it as efficient as possible to increase its range. Battery packs are typically constructed from around 100 battery cells in a series connection. During use of an electric vehicle, the battery cells become mismatched due to small differences in capacity. This effect is further amplified as the electric vehicle ages. Diverging cells cause issues during driving, since weak cells can limit the useable capacity of the vehicle. In order to use the whole capacity of the battery pack, and thus the entire range of the electric vehicle, the cells should be balanced. Strong cells should distribute their excess capacity to weaker cells during driving. The thesis presents the design, modeling and implementation of a novel hardware-efficient battery balancing circuit. First, the theory behind battery balancing is presented. Next, existing battery balancing circuits are compared. Finally, the proposed battery balancing circuit is discussed. The design of the proposed topology is examined in detail. Simulations show that the circuit transfers energy between non-adjacent cells throughout the entire pack. Experimental work is performed on two custom printed circuit boards, a 12 cell lithium-ion module, and a 12V lead acid battery. The results confirm the function of the prototype. The effect of the battery balancing circuit on driving range is examined with vehicle modeling simulations. A 2018 Chevrolet Bolt model is produced and capacity differences are given to each cell. The proposed topology balances the cells while driving, extending driving range on UDDS and HWFET drive cycles. Thesis Master of Applied Science (MASc) One disadvantage of electric vehicles is their limited driving range when compared to internal combustion engine vehicles. Thus, there is a requirement to make the electric system as efficient as possible in order to increase its range. A large piece of the electric system includes the battery pack. Battery packs are typically constructed from around 100 battery cells in a series connection. During use of an electric vehicle, the battery cells become mismatched. This effect is also amplified as the electric vehicle ages. In order to use the whole capacity of the battery pack, and thus the entire range of the electric vehicle, the cells should be balanced. The thesis presents the design, modeling and implementation of a novel hardware efficient battery balancing circuit. The effect of the battery balancing circuit on driving range is examined.

The electrification of transportation is gradually becoming more prominent as it is more efficient and sustainable than conventional transportation alternatives found today. At the centre of this growth is battery testing and research, as they are the primary energy storage devices used to power electric vehicles. With the growing complexity of battery systems, testing and monitoring their performance relies on highly specialized and precise equipment. Furthermore, the use of battery models helps researchers improve their research while reducing the time and costs involved in testing. As such, accurate battery modelling is a critical component in predicting how a battery will behave in specific applications and under various conditions. In this research, a lithium-ion cell is tested extensively, and its performance is characterized across a wide range of operating conditions including temperature, current rates and state of charge (SOC) values. An equivalent circuit model for impedance modelling is proposed, which utilizes constant phase elements represented in the time domain to improve fitting accuracy. This is done concurrently with the development of a state of the art, fully automated battery test system which is showcased throughout the course of the research. In addition to this, an analysis is conducted on the low frequency impedance data used during research, as well as its effect on model accuracy. To provide significance behind the results and relevance to real-world applications, all of the impedance modelling is experimentally validated using temporal drive cycle data. This research was able to demonstrate that the use of a ZARC element can improve the mid-frequency fitting of impedance data relative to a conventionally used modelling approach. It also showcases how the use of low frequency electrochemical impedance spectroscopy (EIS) data can negatively impact the accuracy of impedance modelling. Thesis Master of Applied Science (MASc)

Since its inception in the nineteenth century, the Internal Combustion Engine (ICE) remains the most prevalent technology in transportation systems to date. In order to minimize emissions, it is important that ICE is operated according to its optimized design conditions. As such, condition monitoring and Fault Detection and Diagnosis (FDD) tools can play an important role in detecting conditions that would affect the operability of the engine. In this research, different signal-based Fault Detection and Diagnosis (FDD) techniques are researched and implemented for fault condition monitoring of ICE. The implementation of prognostics for the engine in an automated form has important consequences that include cost savings, increased reliability, reduction of GHG emissions, better safety, and extended life for the vehicle. In this research, in order to carry out FDD onboard, a low-cost and flexible internet-based data-acquisition system (DAQ) was designed and implemented. The main part of the system is an embedded hardware running a full desktop version of Linux. This sensory system leverages the positive aspects of both real-time and general-purpose architectures to ensure engine monitoring at high sampling rates. Unlike other commercial DAQ systems, the software of this device is open-source, free of charge, and highly expandable to suit other FDD applications. In addition to data collection at high sampling rates, the FDD system includes advanced FDD strategies. The Fault Detection and Diagnosis strategies considered use a combination of Fourier Transforms (FT), Wavelet Transforms (WT), and Principal Component Analysis (PCA). Meanwhile, Fault Classification was carried using Neural Networks consisting of the Multi-Layer Perceptron (MLP). Three strategies were comparatively considered for the training of the Neural Network (NN), namely the Levenberg-Marquardt (LM), the Extended Kalman Filter (EKF), and the Smooth Variable Structure Filter (SVSF) techniques. The proposed FDD system was able to achieve 100% accuracy in classifying a set of engine faults. Thesis Master of Applied Science (MASc)

Deep Learning Networks (DLN) is a relatively new artificial intelligence algorithm that gained popularity quickly due to its unprecedented performance. One of the key elements for this success is DL’s ability to extract a high-level of information from large amounts of raw data. This ability comes at the cost of high computational and memory requirements for the training process. Estimation algorithms such as the Extended Kalman Filter (EKF) and the Smooth Variable Structure Filter (SVSF) are used in literature to train small Neural Networks. However, they have failed to scale well with deep networks due to their excessive requirements for computation and memory size. In this thesis the concept of using EKF and SVSF for DLN training is revisited. A New family of filters that are efficient in memory and computational requirements are proposed and their performance is evaluated against the state-of-the-art algorithms. The new filters show competitive performance to existing algorithms and do not require fine tuning. These new findings change the scientific community’s perception that estimation theory methods such as EKF and SVSF are not practical for their application to large networks. A second contribution from this research is the application of DLN to Fault Detection and Diagnosis. The findings indicate that DL can analyze complex sound and vibration signals in testing of automotive starters to successfully detect and diagnose faults with 97.6% success rates. This proves that DLN can automate end-of-line testing of starters and replace operators who manually listen to sound signals to detect any deviation. Use of DLN in end-of-line testing could lead to significant economic benefits in manufacturing operations. In addition to starters, another application considered is the use of DLN in monitoring of the State-Of-Charge (SOC) of batteries in electric cars. The use of DLN for improving the SOC prediction accuracy is discussed. Thesis Doctor of Science (PhD) There are two main ideas discussed in this thesis, both are related to Deep Learning (DL). The first investigates the use of estimation theory in DL network training. Training DL networks is challenging as it requires large amounts of data and it is computationally demanding. The thesis discusses the use of estimation theory for training of DL networks and its utility in information extraction. The thesis also presents the application of DL networks in an end-of-line Fault Detection and Diagnosis system for complex automotive components. Failure of appropriately testing automotive components can lead to shipping faulty components that can harm a manufacturer’s reputation as well as potentially jeopardizing safety. In this thesis, DL is used to detect and analyze complex fault patterns of automotive starters, complemented by sound and vibration measurements.

The Internal Combustion Engines (ICEs) have played a significant role in transportation system to date and are expected to retain a significant market share through to 2050, according to the U.S. Energy Information Administration. Improving the efficiency of the ICEs is one of the most promising and cost-effective approaches to increasing highways vehicle’s fuel economy. The tools to address critical barriers to commercializing higher efficiency, lower emissions, advanced ICEs for passenger and commercial vehicles are increasingly important in the rapidly evolving automotive sector. In this research, a model based optimization strategy is developed for trade-off analysis of parts in Spark Ignition Internal Combustion Engines (SI-ICE). The trade-off analysis tool has been used as a complement to engine mapping to determine the operating region of an engine where a new part could lead to improvements in fuel efficiency, performance, and emissions. To build the engine models, a Design of Experiment (DoE) was developed for performing the engine tests. For each spark plug set, the engine tests were conducted twice with an acceptable control of the parameters that affect engine outputs. The engine torque, Break Specific Fuel Consumption (BSFC) and break specific NOx emission were considered as the engine responses. Engine models were built according to the two-stage modeling strategy by means of black box modeling techniques. The accuracies of the models were 96%, 95% and 92% for the engine torque, BSFC and NOx outputs respectively. For the optimization part, determination of the optimal spark timing for each spark plug was formulated as a multi-objective optimization problem searching for compromises among opposing objectives, i.e. engine torque, emission and fuel consumption. The optimization outputs were in form of Pareto fronts, enabling the selection of the best solutions in terms of different objectives by considering the higher level information. The resulted Pareto fronts of the two spark plugs were compared at different operating points of the Ford Coyote engine and results showed that the two plugs are comparable. The marginal difference was at low load and low speed condition, where the newly designed spark plug was better than the conventional design. Thesis Master of Applied Science (MASc)

Autonomous vehicles are equipped with systems that can detect and track the objects in a vehicle’s vicinity and make appropriate driving decisions accordingly. Infrared (IR) cameras are not typically employed on these systems, but the new information that can be supplied by IR cameras can help improve the probability of detecting all objects in a vehicle’s surroundings. The purpose of this research is to investigate how IR imaging can be leveraged to improve existing autonomous driving detection systems. This research serves as a proof-of-concept demonstration. In order to achieve detection using thermal images, raw data from seven different driving scenarios was captured and labelled using a calibrated camera. Calibrating the camera made it possible to estimate the distance to objects within the image frame. The labelled images (ground truth data) were then used to train several YOLOv2 neural networks to detect similar objects in other image frames. Deeper YOLOv2 networks trained on larger amounts of data were shown to perform better on both precision and recall metrics. A novel method of estimating pixel error in detected object locations has also been proposed which can be applied to any detection algorithm that has corresponding ground truth data. The pixel errors were shown to be normally distributed with unique spreads about different ranges of y-pixels. Low correlations were seen in detection errors in the x-pixel direction. This methodology can be used to create a gate estimation for the detected pixel location of an object. Detection using IR imaging has been shown to have promising results for applications where typical autonomous sensors can have difficulties. The work done in this thesis has shown that the additional information supplied by IR cameras has potential to improve existing autonomous sensory systems. Thesis Master of Applied Science (MASc)

Faulty manufactured alternators lead to commercial and safety concerns when installed in vehicles. Alternators have a major role in the Electrical Power Generation System (EPGS) of vehicles, and a defective alternator will lead to damaging of the battery and other important electric accessories. Therefore, fault detection and diagnosis of alternators can be implemented to quickly and accurately determine the health of an alternator during end of line testing, and not let faulty components leave the manufacturer. The focus of this research is to develop a Model Based Fault Detection and Diagnosis (FDD) strategy for detecting alternator faults during end of line testing. The proposed solution uses Extended Kalman Smooth Variable Structure Filter (EK-SVSF) to detect common alternator faults. A solution using the Dual Extended Kalman Filter (DEKF) is also discussed. The alternator faults were programmatically simulated on alternator measurements. The experimental results prove that both the EK-SVSF and DEKF strategies were very effective in alternator modeling and detecting open diode faults, shorted diode faults, and stator imbalance faults. Thesis Master of Applied Science (MASc)

Personal vehicles have made great contributions to our life and satisfy our daily mobility needs. However, they have also caused societal issues, such as air pollution and global warming. Further to the recent attention to low-carbon energy technologies and environmentally friendly mobility, hybrid electric vehicles play an important role in the current automotive industry. As a leading center and an educational institution in Canada, McMaster University wants to build a Hybrid Electric Vehicle Powertrain Laboratory for introducing undergraduate students to hybrid powertrain architectures, instrumentation and control. A phased development of the hybrid powertrain teaching laboratory is being pursued. The first phase is to design a electric motor laboratory, as a platform for demonstrating motor characteristics. A LabVIEW based interface is designed to enable electric motor characterization tests. This laboratory set-up is still under construction. Real experiments would be implemented, once finishing the utility connections. For the hybrid powertrain laboratory, an innovative design architecture is proposed to enable different hybrid architectures, such as series, parallel, and power-split modes to be investigated. Instead of a planetary gearbox, bevel gearboxes with a continuous variable transmission (CVT) are used for making the laboratory more compact and flexible for demonstrating hybrid functionalities. The additional generator provides the ability of input power-split for allowing the engine to operate at a narrow high efficiency region. After designing the hybrid laboratory, a novel rule-based energy management strategy is applied to a simplified simulation model. Thesis Master of Applied Science (MASc)

A number of non-invasive fault detection and diagnosis (FDD) techniques have been researched and have proven to have worked well in classifying faults in internal combustion engines (ICE) and other mechanical and electrical systems. These techniques are an integral step to creating more robust and accurate methods of determining where or how a fault has or will occur in such systems. These FDD techniques have the potential to not only save time avoiding a tear-down of a costly machine, but could potentially add another layer of safety in detecting and diagnosing a fault much earlier than was possible before. Looking at the previous research methods and the systems they used to acquire this data, it is a natural progression to try and make a system which is able to encapsulate all of these ideologies into one inexpensive module capable of integrating itself into the advanced set of FDD. This thesis follows along with the development of a new wireless sensor that is developed specifically for the use in FDD for ICE and other mechanical systems. A new set of software and firmware is created for the system to be able to be incorporated into previously designed algorithms. After creating and manufacturing the sensor it is put to the test by incorporating it into several Artificial Neural Networks (ANN) and comparing the results to previous experiments done with previous research equipment. Using vibration data acquired from a running engine to train a neural network, the wireless sensor was able to perform equally as well as its expensive counter parts. It proved to have the ability to achieve 100% accuracy in classifying specific engine faults. The performance of three ANN training algorithms, Levenberg-Marquardt (LM), extended Kalman Filter (EKF), and Smooth Variable Structure filter (SVSF), were tested and compared. Adding to the feasibility of a standalone system the wireless sensor was tested in a live environment as a method of instant ICE fault detection. Thesis Master of Applied Science (MASc)

In the past decade, automobile manufacturers have gone through the initial adoption phase of electric mobility. The increasing momentum behind electric vehicles (EV) suggests that electrified storage systems will play an important role in electric mobility going forward. Lithium ion batteries have become one of the most common solutions for energy storage due to their light weight, high specific energy, low self-discharge rate, and non-memory effect. To fully benefit from a lithium-ion energy storage system and avoid its physical limitations, an accurate battery management system (BMS) is required. One of the key issues for successful BMS implementation is the battery model. A robust, accurate, and high fidelity battery model is required to mimic the battery dynamic behavior in a harsh environment. This dissertation introduces a robust and accurate model-based approach for lithium-ion battery management system. Many strategies for modeling the electrochemical processes in the battery have been proposed in the literature. The proposed models are often highly complex, requiring long computational time, large memory allocations, and real-time control. Thus, model-order reduction and minimization of the CPU run-time while maintaining the model accuracy are critical requirements for real-time implementation of lithium-ion electrochemical battery models. In this dissertation, different modeling techniques are developed. The proposed models reduce the model complexity while maintaining the accuracy. The thermal management of the lithium ion batteries is another important consideration for a successful BMS. Operating the battery pack outside the recommended operating conditions could result in unsafe operating conditions with undesirable consequences. In order to keep the battery within its safe operating range, the temperature of the cell core must be monitored and controlled. The dissertation implements a real-time electrochemical, thermal model for large prismatic cells used in electric vehicles' energy storage systems. The presented model accurately predicts the battery's core temperature and terminal voltage. Thesis Doctor of Philosophy (PhD)

The internal combustion engine (ICE) is widely used in applications such as automobiles, motorcycles and ships. After its long-term use faults occur that degrade its performance or cause it to malfunction. Therefore ICE fault detection and diagnosis (FDD) research is important for preventing serious economic loss and even human injuries caused by undetected faults. The development of an ICE FDD experimental setup is described. The setup uses sensors to measure the ICE vibration and FDD techniques implemented as a computer program. One common ICE fault, the misfire fault, is deliberately induced in some of the experiments by disconnecting a spark plug. The objective of the FDD is to determine if there is a misfire fault (or not), and which spark plug is faulty. Several FDD algorithms are proposed, one category of which is based on data processing techniques such as the variational mode decomposition (VMD) and the wavelet transform. This category of FDD algorithms includes the VMD-based FDD algorithm, wavelet-based kernel principle component analysis (KPCA) and VMD-based KPCA. The VMD-based FDD algorithm introduces a new FDD index based on VMD and statistics. According to the included experimental results, all of these algorithms are capable of detecting and locating the misfire fault with 100% accuracy. A new SVSF-based training algorithm for the radial-basis-function (RBF) artificial neural network is also proposed. The running-averaged wavelet coefficients of vibration data are used as the network input. The included experimental results show these SVSF-trained networks above achieve 100% accuracy in classifying the misfire faults. The SVSF-based training algorithm also produces a faster convergence rate. Thesis Master of Applied Science (MASc)

In the strive to use electric energy as the sole source of traction force in the automotive industry, the electric storage unit poses serious challenges. Although Li-Ion batteries have shown promising features, there is still room to improve the shortcomings of the electric storage unit. The chemical process of energy extraction and storage in the batteries results in inherently limited power density and lifetime. The notion of pairing a Li-Ion battery cell with an ultracapacitor will invigorate power capability of the hybrid cell. Furthermore, enabling local processing capability in each hybrid cell makes it possible to run sophisticated control and power management algorithms that result in increased battery longevity. Also, hybridization at cell level provides greater access to each cell’s momentary states which results in more accurate analysis, improved safety, and a broad control possibility. In this research, a Li-Ion battery cell is combined with an ultracapacitor. An appropriate configuration is selected for this particular application after reviewing several hybridization schemes and converter structures. A comprehensive power management unit is designed and implemented that is able to execute several strategies to mitigate battery stresses while improving power capabilities. A current regulator is also designed and implemented to realize the proposed strategies by providing the power converter with proper duty cycle. Rather than simulation under a wide range of test loads, a prototype hybrid cell is built and tested in an experimental setup. Thesis Master of Applied Science (MASc)

The main idea of this thesis is to formulate the smooth variable structure filter (SVSF) for target tracking applications in the presence of measurement origin uncertainty. Tracking, by definition is the recursive estimation of the states of an unknown target from indirect, inaccurate and uncertain measurements. The measurement origin uncertainty introduces the data association problem to the tracking system. The SVSF estimation strategy was first presented in 2007. This filter is based on sliding mode concepts formulated in a predictor-corrector form. Essentially, the SVSF uses an existence subspace and smoothing boundary layer to bind the estimated state trajectory to within a subspace around the true trajectory. The SVSF is demonstrated to be robust to modeling uncertainties and provide extra measures of performance such as magnitude of the chattering signal. Therefore, with respect to specific nature of car tracking problems that involves modeling uncertainty, it was hypothesized that a robust estimation strategy such as the SVSF, would improve the performance of the tracking system and give more robust tracking results. Also, having the extra information provided by the SVSF strategy, i.e. the chattering magnitude signal, would lead to algorithms that could better account for measurement origin uncertainty in the context of the data association process. Further to these hypotheses, this research has focused on investigating the performance of the SVSF in the target tracking problems, advancing the development of the SVSF, and employing its characteristics to deal with data association problems. The performance of the SVSF, in its current form, can be improved when there is fewer measurements than states by using its error covariance in target tracking. As the first contribution in this research, the SVSF is formulated in the context of target tracking in clutter and combined with data association algorithms, resulting in the SVSF-based probabilistic data association (PDA) and joint probabilistic data association (JPDA) for non-maneuvering and maneuvering targets. The results are promising in the tracking scenarios with modeling uncertainties. Therefore, the thesis is then expanded by generalizing the covariance of the SVSF for the cases where the number of measurements is less than the number of states. The generalized covariance formulation is then used to derive a generalized variable boundary layer (GVBL) SVSF. This new derivation gives an estimation method that is optimal in the MMSE sense and in the meantime preserves the robustness of the SVSF. The proposed algorithm improves the performance measures and makes a more reliable tracking algorithm. This thesis explores the hypothesis that multiple target tracking performance can be substantially improved by including chattering information from SVSF-based filtering in the data association method. A Bayesian framework is used to formulate a new set of augmented association probabilities which include the chattering information. The simulation and experimental results demonstrate that the proposed augmented probabilistic data association improves the performance of the tracking system including maneuvering cars, in particular for highly cluttered environments. The derived methods are applied on simulations and also on real data from an experimental setup. This thesis is made up of a compilation of papers that include three conference papers and three journal papers. Thesis Doctor of Philosophy (PhD)

The field of energy storage has improved drastically within the last two decades. Batteries of various chemistries have been relied on to provide energy for numerous portable electronic devices. Lithium-ion cells, when compared to other chemistries have been known to provide outstanding energy-to-weight ratios and exhibit low self-discharge when not in use [1]. The aforementioned benefits in conjunction with decreasing costs have made lithium-ion cells an exceptional choice for use in electrical vehicles (EVs). Battery Management Systems (BMS) in EVs are responsible for providing estimates for values that are indicative of the battery pack’s present operating condition. The current operating condition could be described by State of Charge, power fade, capacity fade and various other parameters [2]. Importantly, it is essential for the estimation technique to adjust to fluctuating cell characteristics as the cell ages, in pursuance of having available accurate estimates for the life time of the pack. In order for the estimation technique to properly estimate the desired quantities, a mathematical model capable of capturing cell dynamics is desired. There are various proposed methods recommended for mathematically modeling a cell, namely equivalent Circuit modeling, electro-chemical modeling and impedance spectroscopy. Consequently, in order to ensure mathematical models are accurate and further to have the ability to compare the proposed models, it is essential to have available data gathered from a given cell at specific operating conditions. This Master’s thesis outlines the development of a lithium-ion cell tester that is capable of controlling, monitoring and recording parameters such as current, voltage and temperature. The Dual capability of obtaining data from standardized cell tests as well as high frequency cell tests is fascinating and intriguing. As this capability holds the possibility of reducing cost otherwise spent on man hours and equipment which are both paramount in any industrially automated process. Thesis Master of Applied Science (MASc)

This research investigates the effects of an electric turbocharger in a hybrid electric powertrain. First generic vehicle models are created and run to understand the overall powertrain requirements of torque, power and energy of a performance consumer vehicle. Then a low fidelity baseline model of a conventional vehicle is created in Simulink to serve as a baseline measure. To analyze an electric turbocharger system a high-fidelity model in AMESIM of a 4 cylinder turbocharged engine was modified. This engine model was analyzed using virtual dynamometer tests and a simplified look-up table based controller was developed for the electric motor within the electric turbocharger. Next this engine model was inserted within three different types of hybrid powertrain architectures models in AMESIM. Each hybrid powertrain required a unique supervisory controller which was developed using Stateflow in Simulink. These controller algorithms were imported into AMESIM and the model was simulated over standard drive cycles. Since a very wide variation of electrification level exists within hybrid powertrains the supervisory controllers are calibrated for charge-sustaining simulations. This allows for impartial comparisons across the hybrid architectures. Lastly a track drive cycle was developed to understand electric turbocharger effects under high performance loading conditions Thesis Master of Applied Science (MASc) Turbochargers on internal combustion engines can utilize a portion of waste exhaust energy to pump more air into the cylinder leading to greater power and efficiency. A modern high performance 4-cylinder turbocharged engine is capable of replacing a V6 engine of much higher cylinder displacement. However turbocharged engines suffer from ‘turbo lag’ when the engine cannot immediately produce power. An electric turbocharger can virtually eliminate this ‘turbo lag’ as well as generate electricity from excess energy the turbocharger does not use. Electric turbochargers have been development by researchers and various automotive manufacturers. However the potential effects of such a system within the framework of a hybrid electric powertrain in a consumer vehicle has not been quantified. The objective of this research is to use high fidelity models to investigate the effects of an electric turbocharger system within a hybrid powertrain.