14878 research outputs found
Sort by
A unified framework for multi-level control and multi-objective optimization of hybrid campus microgrids: integrating adaptive mpc, modified firefly algorithm, and homer simulations
No full textHybrid microgrids have become an essential solution in integrating renewable energy sources into campus energy systems, addressing sustainability, reliability, and energy efficiency challenges. In this dissertation, I explore a comprehensive framework for controlling and optimizing hybrid microgrids, specifically focusing on the combination of solar PV, wind turbines, Combined Heat and Power (CHP) systems, and Battery Storage Systems (BSS) in a university campus setting. Traditional control strategies often fall short when dealing with the dynamic and intermittent nature of these renewable sources, necessitating advanced optimization and control techniques. To address these limitations, this research first examines the design and operation of hybrid renewable energy systems through a case study at Oakland University, utilizing HOMER software for simulations. This case study demonstrates the effectiveness in balancing key performance indicators, such as Net Present Cost (NPC), Levelized Cost of Energy (LCOE), and environmental impact. The optimized system configuration illustrates how hybrid microgrids can significantly reduce energy costs while enhancing sustainability in campus applications. Furthermore, I propose a Modified Firefly Algorithm (MFA) to optimize hybrid microgrid operations by solving multi-objective problems, including cost minimization and greenhouse gas emissions reduction. The MFA was specifically adapted to improve the efficiency of energy management systems (EMS) in microgrids by dynamically adjusting the optimization parameters. The algorithm outperformed traditional optimization techniques, offering superior results for complex multi-objective problems in hybrid microgrids. Additionally, this dissertation develops a multi-level control framework for hybrid microgrids, incorporating Adaptive Model Predictive Control (MPC) communication. The Adaptive MPC leverages predictive modeling to dynamically adjust control actions based on real-time data, optimizing power dispatch across the various energy sources and storage systems. This integration ensures reliable and efficient communication between distributed energy resources, improving system stability and performance under fluctuating conditions. This research provides valuable insights into the optimization and control of hybrid microgrids, demonstrated through real-world case studies and modified algorithmic approaches. The proposed methodologies cover more efficient, reliable, and sustainable energy systems in campus microgrids, offering a robust framework for future developments in smart grid technologie
Minutes of the Meeting of the University Senate, January 18, 2024
No full text1. Informational items: Graduate Program Submissions (Graduate Certificate in Industrial Hygiene, Graduate Certificate in Human Factors and Ergonomics, Graduate Certificate in Environmental Protection); Undergraduate Program Submissions (Recreation Leadership Minor, Clinical and Diagnostic Sciences BS, Electrical Engineering BSE new concentrations, School of Business Administration pre-core requirements); Combined Graduate and Undergraduate Submissions (General Education catalog update, Culture waiver for international students); OU Advance Phase 2 Report; Affordable Course Materials Initiative; Provost's Updates | 2. Roll Call | 3. Approval of the Minutes of December 14, 2023 | 4. Unfinished Business: None | 5. New Business: Senate Standing Committee staffing | 6. Good and Welfare | 7. Adjour
Digital Shearography for Nondestructive Testing (NDT): Determination of Smallest Detectable Defect and Improvement of its Visibility
No full textSensitivity is a key parameter for using NDT technology. Determining what factors affect the sensitivity of NDT and establishing the model to facilitate engineers to select the appropriate system parameters and the loading magnitude are very important in the practical applications. In the past decade, digital shearography has been widely used as an NDT tool for detecting delaminations and debonding defects in various composite materials, such as glass fiber reinforced polymer (GFRP), carbon fiber reinforced polymer (CFRP), Honeycomb structures, etc. Digital shearography is a laser interferometric technique and able to measure the first derivatives of deformation, i.e. strain information. It is suited well for NDT because defects generate strain concentration after a loading. As an NDT method, the sensitivity of digital shearography is an important parameter to measure the technology’s defect detection capabilities. However, due to various limitations, the sensitivity research of digital shearography is still in its infancy. First, this technique lacks a numerical model to offer a theoretical foundation for determining the minimum detectable delamination/debonding limits and the detectable depth of defects. Secondly, because a shearogram is a fringe pattern which is composed of both global deformation and defect information, smaller defect information is easily lost in the fringe patterns from global deformation. This research conducts in-depth study around determining the sensitivity of digital shearography and improving the defect visibility to meet the practical needs of helping engineers quickly select loads and quickly identify defects. To solve these problems, the main research work and innovation results are as follows: (1) In response to the first problem, this research presents a methodology of digital shearography for determining the size of the smallest detectable defect and its depth under various loading magnitudes for the purpose of nondestructive testing. First, a mechanical model based on the thin plate theory to calculate the expected bending of close-to-surface defects was proposed; the model built a relationship among the deformation caused by a defect, the size and the depth of the defect, as well as the load and the material properties. Second, the relationship between the relative deformation measured by shearography and the deformation induced by a defect was established based on the optimized shearing amount and the sensitivity of digital shearography. Based on these analyses, relationships between the size of the smallest detectable defect and the depth under different load amounts were established for different defect shapes. (2) A demonstration of the sensitivity limit of digital shearography is shown on the basis of the sensitivity model, and the search for strategies to improve digital shearography is undertaken. After research, while keeping the equipment consistent, the material unchanged, and the loading conditions the same, the best way to improve the sensitivity of digital shearography is to increase the contrast between defect information and background information. This method can make defects information clearer. Based on that, the second purpose was to examine methods for improving sensitivity found in previous studies and to discuss the advantages and disadvantages of all methods and developed the segment fitting method. According to the previous discussion, it can be found that the most common method is to make a fitting plane to represent the global deformation by unwrapping fringe pattern to build the continuous shearogram, and then subtracting the plane to increase the contrast. Secondly, the continuous shearogram of complex deformations makes it challenging to choose the fitting equation. Based on this, a piecewise fitting method is proposed. This method is based on the conventional fitting method, which is fitted based on the phase change between each fringes on the shearogram. Because the phase values between each fringe are linearly distributed, this method does need to consider the fitting equation selection. The new planes then need to be subtracted from the original image to remove the global deformation, thus preserving the defect. (3) The second innovation of this research is the development of a practical and effective method to experimentally remove fringe patterns caused by the global deformation that makes small defects directly visible, which improves the non-destructive testing capabilities of digital shearography, thereby simplifying defect detection and visualization. For shearographic Non-Destructive Testing (NDT), the phase distributions of two interferograms under different loads P1 and P2 are recorded. This novel approach involves recording one additional phase distribution of an interferogram at a load between P₁ and P₂, e.g. P₁’. Two phase maps of shearograms can be generated, corresponding to the two loads Δ₂ = (P₁’-P₁) and Δ₁ = (P₂-P₁), respectively. Because of the nondestructive nature of the testing, the magnitude of the loads P₁ and P₂ is small, and the 1st derivative of global deformation of the test part is assumed to be linear. Therefore, a linear coefficient C based on the two shearograms can be determined. The information from global deformation is then removed by subtracting the shearogram generated with the small load Δ₂ multiplied by the correlation coefficient C from the one obtained with the relatively large load Δ₁. This technique is further improved by calculating a complete surface linear coefficient Cij, which improves the detail processing of the deformation of samples with complex geometry and mechanical properties. Experimental verification was conducted based on specimens with prefabricated defects of different sizes and different loading conditions to verify the proposed mathematical model and experimental methods to eliminate global deformation. Experimental results show that the developed model can provide useful estimates for digital shearography NDT, and in particular can help test engineers estimate the size of the smallest detectable defect and the depth of the defect under corresponding load magnitude. Also, the experimental coefficient method can effectively evaluate a variety of structures and also be verified to remove global deformation, which improves defect detection capabilities and increases visualization of the digital shearograph
Development of Predictive and Nonlinear Control Designs for an Electric All Wheel Drive Powertrain System
No full textVehicle electrification is a strong trend in the automotive industry. The use of electric motors for propulsion offers many opportunities and some technical difficulties. This thesis addresses the challenge of controlling and coordinating two electric motors as part of an electric all wheel drive (eAWD) powertrain to maintain vehicle stability during longitudinal and lateral motion of the vehicle. The control problem is broken into two pieces, vehicle level control and control of the electric motors. At the vehicle level, several control strategies are developed and compared. First, a weighted one step ahead (WOSA) controller is developed. Next, a nonlinear model predictive controller (NLMPC) is developed. Next, a feedback linearization controller (FLC) is developed. Finally, a sliding mode controller (SMC) is developed. In each case, the vehicle level controller is integrated with fuzzy rules to translate the vehicle level control signal into reference targets for the motor controllers. The integration of a physical model and fuzzy rules into a single controller is a unique concept that is explored in this thesis. Each of the electric motors consists of a traction motor and torque vectoring motor. WOSA controllers are developed for each type of motor and integrated with the vehicle level controller for overall vehicle control. The WOSA is a simple but powerful control approach for the motor controllers. Key system states are estimated using the controller output observer (COO) concept. The COO provides a good estimate of the system states with sensor data that is already available on typical vehicles. The COO also uses the WOSA control strategy. For both longitudinal and lateral motion, the effectiveness of the control strategies are demonstrated for launch and a high speed double lane change maneuver on high friction and low friction road surfaces using Simulink and CarSi
Development of a Hybrid Finite Element Method for Solving Inverse Engineering Problems
No full textThe ill-posed boundary value problems, such as contact problems, adhesive joint analysis, and damage identification, are of foremost importance in the design and manufacturing of machines. The analysis of these problems has attracted considerable attention from engineers and researchers in various industries. This dissertation presents a novel hybrid finite element method for solving ill-posed boundary value problems through inverse engineering, with a focus on accurately determining contact stress, identifying damage, and analyzing adhesive joints in mechanical engineering. A significant aspect of this method involves integrating empirical measurements with numerical simulations to enhance both the accuracy and reliability of finite element analyses under insufficient boundary conditions. A constrained optimization framework is also employed in this study. The method is evaluated through seven case studies, which include assessments of plate bending, rigid contact problems, Hertzian contact problems, damage identification tasks, single lap joint, and T-peel joint evaluations. A novel constraint equation based on the gradient of the loading function is introduced as well. These case studies highlight the method’s comprehensive applicability and effectiveness across a range of complex engineering challenges. The dissertation outlines future work that aims to expand the methods application to three-dimensional problems, improve the optimization algorithms, and explore further applications. This work lays a foundation for advancing more complex and reliable modeling techniques in mechanical engineering, with significant implications for both research and industrial applications
Minutes of the Meeting of the Oakland University Board of Trustees Audit Committee, October 18, 2024
No full textA. Call to Order | B. New Items for Consideration: Approval of Minutes of the Audit Committee Meeting of April 12, 2024; Draft Financial Statements, June 30, 2024 and 2023; Schedule of Expenditures of Federal Awards, Year Ended June 30, 2024; Internal Audit Department Audit Charter; | C. Other Items for Consideration that May Come Before the Committee | D. Adjournmen
Neuromorphic Computing with Antiferromagnetic Artificial Neurons
No full textThe increasing focus on Artificial Intelligence (AI) presents notable challenges, especially in the power requirements essential for training AI models. Consequently, there is a growing emphasis on neuromorphic computing, which aims to construct Artificial Neural Networks (ANNs) from artificial neurons to replicate the speed and efficiency of the human brain. With their significantly lower power consumption, spintronic devices acting as artificial neurons offer the potential for ANNs that rival conventional components. This innovative approach not only tackles the energy efficiency challenges but also paves the way for advancements in neuromorphic computing by integrating magnetic materials and spin-dependent effects. Antiferromagnetic (AFM) materials, characterized by their inherent THz frequencies, provide a unique opportunity for spintronic devices with ultra-fast dynamics. There's a proposal to utilize AFM materials to craft ultra-fast spin-Hall oscillators. These oscillators emit spiking signals resembling those of biological neurons, indicating their potential as artificial neurons. AFM oscillators exhibit exceptionally fast characteristics, including picosecond-scale spike widths and unique features absent in conventional artificial neuron models. This research examines the utilization of AFM oscillators as artificial neurons and their significance in the realm of neuromorphic computing. This dissertation begins with a comprehensive overview of relevant topics, covering the principles of spintronics, AFM materials, and neuromorphic computing. It examines the unique dynamics of AFM neurons, such as response latency and refraction time, which arise from an effective internal inertia. Then, it explores innovative AFM neuron circuits that exhibit functionalities unattainable by conventional artificial neurons, such as non-monotonic inhibition. Additionally, conventional learning algorithms like backpropagation are use to train AFM ANNs for pattern recognition. Finally, it leverages the similarities between AFM and biological neurons to model the biological neural network responsible for the withdrawal reflex. With their simplistic design and high speeds, AFM neurons exhibit energy efficiencies several orders of magnitude higher than traditional artificial neurons and even other spintronic designs. This suggests that AFM ANNs will excel at training AI models while addressing the energy crisis impeding technological progress. As a result, AFM neurons drive forward the progress of spintronic neuromorphic computing, providing a promising alternative for future technological developmen