Mechanical Engineering
Permanent URI for this communityhttp://hdl.handle.net/1903/2263
Browse
10 results
Search Results
Item Denoising the Design Space: Diffusion Models for Accelerated Airfoil Shape Optimization(2024) Diniz, Cashen; Fuge, Mark D; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Generative models offer the possibility to accelerate and potentially substitute parts of the often expensive traditional design optimization process. We present Aero-DDM, a novel application of a latent denoising diffusion model (DDM) capable of generating airfoil geometries conditioned on flow parameters and an area constraint. Additionally, we create a novel, diverse dataset of optimized airfoil designs that better reflects a realistic design space than has been done in previous work. Aero-DDM is applied to this dataset, and key metrics are assessed both statistically and with an open-source computational fluid dynamics (CFD) solver to determine the performance of the generated designs. We compare our approach to an optimal transport GAN, and demonstrate that our model can generate designs with superior performance statistically, in aerodynamic benchmarks, and in warm-start scenarios. We also extend our diffusion model approach, and demonstrate that the number of steps required for inference can be reduced by as much as ~86%, compared to an optimized version of the baseline inference process, without meaningful degradation in design quality, simply by using the initial design to start the denoising process.Item INTERPRETABLE AND SPEED ADAPTIVE CONVOLUTIONAL NEURAL NETWORK FOR PROGNOSTICS AND HEALTH MANAGEMENT OF ROTATING MACHINERY(2023) Lee, Nam Kyoung; Pecht, Michael; Azarian, Michael H; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Faulty rotating machines exhibit vibrational characteristics that can be distinguished from healthy machines using prognostics and health management methods. These characteristics can be extracted using signal processing techniques. However, these techniques require certain inputs, or parameters, before the desired characteristics can be extracted. Setting the parameters requires skill and knowledge, as they should reflect the component geometries and the operational conditions. Using convolutional neural networks for diagnosing faults on a rotating machine eliminates the need for parameter setting by replacing signal processing with mathematical operations in the networks. The parameters that affect the outcomes of the operations are learned from data during the training of the neural networks. The networks can capture characteristics that are related to the health state of a machine, but their operations are not interpretable. Unlike signal processing, the internal operations of the networks have no constraints that guide the networks to transform vibrations into certain information, that is, vibrational characteristics. Without the constraints, there is no basis for understanding the characteristics in terms that can be associated with the physics of failure. The lack of interpretability impedes the physical validation of vibrational characteristics captured by the networks.This dissertation presents a method for changing the internal operations of a convolutional neural network to emulate a specific type of signal processing known as envelope analysis. Envelope analysis demodulates vibrations to extract vibrational signatures associated with mechanical impact on a defective rolling component. An understanding of envelope analysis, along with knowledge of the geometries of machine components and operational speeds, allows for a physical interpretation of the signatures. The dissertation develops speed adaptive convolutional layers and a rotational speed estimation algorithm to identify defect signatures whose frequency components change as the speed changes. The characteristics that are captured by the developed convolutional neural network are verified through a feature selection process that is designed to filter out physically implausible features. Case studies on three different systems demonstrate the feasibility of using the developed convolutional neural network for the diagnosis.Item DYNAMIC PROGNOSTIC HEALTH MANAGEMENT FOR RESPONSE TIME BASED REMAINING USEFUL LIFE PREDICTION OF SOFTWARE SYSTEMS(2022) Islam, Mohammad Rubyet; Sandborn, Dr. Peter A; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Prognostics and Health Management (PHM) is an engineering discipline focused on predicting the future point at which systems or components will no longer perform as intended. The prediction is often articulated as a Remaining Useful Life (RUL). PHM has been widely applied to hardware systems in the electronics and non-electronics domains but has not been explored for software applications. While software does not decay over time, it can degrade over release cycles. Software degradation is a common problem faced by legacy systems. Today, software health management is confined to diagnostic assessments that identify problems. In contrast, prognostic assessment potentially indicates what problems will become detrimental to the operation of the system in the future. Relevant research areas such as software defect prediction, software reliability prediction, predictive maintenance of software, software degradation, and software performance prediction, exist, but all of these represent diagnostic models built upon historical data – none of which can predict an RUL for software. This dissertation addresses the application of PHM concepts to software systems for fault predictions and RUL estimation. Specifically, this dissertation addresses how PHM can be used to make decisions for software systems such as version update/upgrade, module changes, rejuvenation, maintenance schedules, and abandonment. This dissertation presents a method to prognostically and continuously predict the RUL of a software system based on usage parameters (e.g., the numbers and categories of releases) and performance parameters (e.g., response time). The model developed in this dissertation has been validated by comparing actual data generated using test beds. Statistical validation (regression validation) has also been carried out. A case study is presented based on publicly available data for the Bugzilla application. Controlled test beds for multiple Bugzilla releases are prepared to formulate standard staging environments to populate relevant data. This case study demonstrates that PHM concepts can be applied to software systems, and RUL can be calculated to make decisions on software management.Item Application of Diagnostics and Prognostics Techniques to Qualification Against Wear-Out Failure(2022) Ram, Abhishek; Das, Diganta; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)QUALIFICATION IS A PROCESS THAT DEMONSTRATES WHETHER A PRODUCT MEETS OR EXCEEDS SPECIFIED REQUIREMENTS. TESTING AND DATA ANALYSIS PERFORMED WITHIN A QUALIFICATION PROCEDURE SHOULD VERIFY IF PRODUCTS SATISFY THOSE REQUIREMENTS, INCLUDING RELIABILITY REQUIREMENTS. MOST OF THE ELECTRONICS INDUSTRY QUALIFIES PRODUCTS USING PROCEDURES DICTATED WITHIN QUALIFICATION STANDARDS. A REVIEW OF COMMON QUALIFICATION STANDARDS REVEALS THAT THOSE STANDARDS DO NOT CONSIDER CUSTOMER REQUIREMENTS OR THE PRODUCT PHYSICS-OF-FAILURE IN THAT INTENDED APPLICATION. AS A RESULT, QUALIFICATION, AS REPRESENTED IN THE REVIEWED QUALIFICATION STANDARDS, WOULD NOT MEET OUR DEFINITION OF QUALIFICATION FOR RELIABILITY ASSESSMENT. THIS THESIS PROVIDES AN APPLICATION-SPECIFIC APPROACH FOR DEVELOPING A QUALIFICATION PROCEDURE THAT ACCOUNTS FOR CUSTOMER REQUIREMENTS, PRODUCT PHYSICS-OF-FAILURE, AND KNOWLEDGE OF PRODUCT BEHAVIOR UNDER LOADING. THIS THESIS PROVIDES A REVAMPED APPROACH FOR DEVELOPING A LIFE CYCLE PROFILE THAT ACCOUNTS FOR LOADING THROUGHOUT MANUFACTURING/ASSEMBLY, STORAGE AND TRANSPORTATION, AND OPERATION. THE THESIS ALSO DISCUSSES IDENTIFYING VARIATIONS IN THE LIFE CYCLE PROFILE THAT MAY ARISE THROUGHOUT THE PRODUCT LIFETIME AND METHODS FOR ESTIMATING LOADS. THIS UPDATED APPROACH FOR DEVELOPING A LIFE CYCLE PROFILE SUPPORTS BETTER FAILURE PRIORITIZATION, TEST SELECTION, AND TEST CONDITION AND DURATION REQUIREMENT ESTIMATION. ADDITIONALLY, THIS THESIS INTRODUCES THE APPLICATION OF DIAGNOSTICS AND PROGNOSTICS TECHNIQUES TO ANALYZE REAL-TIME DATA TRENDS WHILE CONDUCTING QUALIFICATION TESTS. DIAGNOSTICS TECHNIQUES IDENTIFY ANOMALOUS BEHAVIOR EXHIBITED BY THE PRODUCT, AND PROGNOSTICS TECHNIQUES FORECAST HOW THE PRODUCT WILL BEHAVE DURING THE REMAINDER OF THE QUALIFICATION TEST AND HOW THE PRODUCT WOULD HAVE BEHAVED IF THE TEST CONTINUED. AS A RESULT, COMBINING DIAGNOSTICS AND PROGNOSTICS TECHNIQUES CAN ENABLE THE PREDICTION OF THE REMAINING TIME-TO-FAILURE FOR THE PRODUCT UNDERGOING QUALIFICATION. SEVERAL ANCILLARY BENEFITS RELATED TO AN IMPROVED TESTING STRATEGY, PARTS SELECTION AND MANAGEMENT, AND SUPPORT OF A PROGNOSTICS AND HEALTH MANAGEMENT SYSTEM IN OPERATION ALSO ARISE FROM APPLYING PROGNOSTICS AND DIAGNOSTICS TECHNIQUES TO QUALIFICATION.Item INFERENCE-BASED MODELING, MONITORING, AND CONTROL ALGORITHMS FOR AUTONOMOUS MEDICAL CARE(2022) Tivay, Ali; Hahn, Jin-Oh; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Autonomous medical care systems are relatively recent developments in biomedical research that aim to leverage the vigilance, precision, and processing power of computers to assist (or replace) humans in providing medical care to patients. Indeed, past research has demonstrated initial promise for autonomous medical care in applications related to anesthesia, hemodynamic management, and diabetes management, to name a few. However, many of these technologies yet do not exhibit the maturity necessary for widespread real-world adoption and regulatory approval. This can be attributed, in part, to several outstanding challenges associated with the design and development of algorithms that interact with physiological processes. Ideally, an autonomous medical care system should be equipped to exhibit (i) transparent behavior, where the system’s perceptions, reasoning, and decisions are human-interpretable; (ii) context-aware behavior, where the system is capable of remaining mindful of contextual and peripheral information in addition to its primary goal; (iii) coordinated behavior, where the system can coordinate multiple actions in synergistic ways to best achieve multiple objectives; (iv) adaptable behavior, where the system is equipped to identify and adapt to variabilities that exist within and across different patients; and (v) uncertainty-aware behavior, where the system can handle imperfect measurements, quantify the uncertainties that arise as a result, and incorporate them into its decisions. As these desires and challenges are specific to autonomous medical care applications and not fully explored in past research in this area, this dissertation presents a sequence of methodologies to model, monitor, and control a physiological process with special emphasis on addressing these challenges. For this purpose, first, a collective variational inference (C-VI) method is presented that facilitates the creation of personalized and generative physiological models from low-information and heterogeneous datasets. The generative physiological model is of special importance for the purposes of this work, as it encodes physiological knowledge by reproducing the patterned randomness that is observed in physiological datasets. Second, a population-informed particle filtering (PIPF) method is presented that fuses the information encoded in the generative model with real-time clinical data to form perceptions of a patient’s states, characteristics, and events. Third, a population-informed variational control (PIVC) method is presented that leverages the generative model, the perceptions of the PIPF algorithm, and user-defined definitions of actions and rewards in order to search for optimal courses of treatment for a patient. These methods together form a physiological decision-support and closed-loop control (PCLC) framework that is intended to facilitate the desirable behaviors sought in the motivations of this work. The performance, merits, and limitations of this framework are analyzed and discussed based on clinically-important case studies on fluid resuscitation for hemodynamic management.Item A PHYSICS-INFORMED NEURAL NETWORK FRAMEWORK FOR BIG MACHINERY DATA IN PROGNOSTICS AND HEALTH MANAGEMENT FOR COMPLEX ENGINEERING SYSTEMS(2022) Cofre Martel, Sergio Manuel Ignacio; Modarres, Mohammad; Lopez Droguett, Enrique; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Big data analysis and data-driven models (DDMs) have become essential tools in prognostics and health management (PHM). Despite this, several challenges remain to successfully apply these techniques to complex engineering systems (CESs). Indeed, current state-of-the-art applications are treated as black-box algorithms, where research efforts have focused on developing complex DDMs, overlooking or neglecting the importance of the data preprocessing stages prior to training these models. Guidelines to adequately prepare data sets collected from CESs to train DDMs in PHM are frequently unclear or inexistent. Furthermore, these DDMs do not consider prior knowledge on the system’s physics of degradation, which gives little-to-no control over the data interpretation in reliability applications such as maintenance planning.In this context, this dissertation presents a physics-informed neural network (PINN) architecture for remaining useful life (RUL) estimation based on big machinery data (BMD) collected from sensor monitoring networks (SMNs) in CESs. The main outcomes of this work are twofold. First, a systematic guide to preprocess BMD for diagnostics and prognostics tasks is developed based on expert knowledge and data science techniques. Second, a PINN-inspired PHM framework is proposed for RUL estimation through an open-box approach by exploring the system’s physics of degradation through partial differential equations (PDEs). The PINN-RUL framework aims to discover the system’s underlying physics-related behaviors, which could provide valuable information to create more trustworthy PHM models. The data preprocessing and RUL estimation frameworks are validated through three case studies, including the C-MAPSS benchmark data set and two data sets corresponding to real CESs. Results show that the proposed preprocessing methodology can effectively generate data sets for supervised PHM models for CESs. Furthermore, the proposed PINN-RUL framework provides an interpretable latent variable that can capture the system’s degradation dynamics. This is a step forward to increase interpretability of prognostic models by mapping the RUL estimation to the latent space and its implementation as a state of health classifier. The PINN-RUL framework is flexible as it allows incorporating available physics-based models to its architecture. As such, this framework takes a step forward in bridging the gap between statistic-based PHM and physics-based PHM methods.Item QUANTILE-BASED LSTM REMAINING USEFUL LIFE PREDICTOR(2021) saadon, yonatan; McCluskey, Patrick PM; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Accurate prediction of the remaining useful life (RUL) of a degrading component is crucial to prognostics and health management for electronic systems, to monitor conditions and avoid reaching failure while minimizing downtime. However, the shortage of sufficiently large run-to-failure datasets is a serious bottleneck impeding the performance of data-driven approaches, and in particular, those involving neural network architectures. Here, this work shows a new data-driven prognostic method to predict the RUL using an ensemble of quantile-based Long Short-Term Memory (LSTM) neural networks, which represents the RUL prediction task to a set of simpler, binary classification problems that are amenable for prediction with LSTMs, even with limited data. This methodology was tested on two run-to-failure datasets, power MOSFETs and filtration system, and showed promising results on both datasets it demonstrates that this approach obtains improved RUL estimation accuracy for both the power MOSFETs and the filtration system, especially with a small training dataset that is characterized by a wide range of the RUL.Item Systematic Integration of PHM and PRA (SIPPRA) for Risk and Reliability Analysis of Complex Engineering Systems(2021) Moradi, Ramin; Groth, Katrina; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Complex Engineering Systems (CES) such as power plants, process plants, and manufacturing plants have numerous, interrelated, and heterogeneous subsystems with different characteristics and risk and reliability analysis requirements. With the advancements in sensing and computing technology, abundant monitoring data is being collected. This is a rich source of information for more accurate assessment and management of these systems. The current risk and reliability analysis approaches and practices are inadequate in incorporating various sources of information, providing a system-level perspective, and performing a dynamic assessment of the operation condition and operation risk of CES. In this dissertation, this challenge is addressed by integrating techniques and models from two of the major subfields of reliability engineering: Probabilistic Risk Assessment (PRA) and Prognostics and Health Management (PHM). PRA is very effective at modeling complex hardware systems, and approaches have been designed to incorporate the risks introduced by humans, software, organizational, and other contributors into quantitative risk assessments. However, PRA has largely been used as a static technology mainly used for regulation. On the other hand, PHM has developed powerful new algorithms for understanding and predicting mechanical and electrical device health to support maintenance. Yet, PHM lacks the system-level perspective, relies heavily on operation data, and its outcomes are not risk-informed. I propose a novel framework at the intersection of PHM and PRA which provides a forward-looking, model- and data-driven analysis paradigm for assessing and predicting the operation risk and condition of CES. I operationalize this framework by developing two mathematical architectures and applying them to real-world systems. The first architecture is focused on enabling online system-level condition monitoring. The second architecture improves upon the first and realizes the objectives of using various sources of information and monitoring operation condition together with operational risk.Item DATA-DRIVEN ANALYTICAL MODELS FOR IDENTIFICATION AND PREDICTION OF OPPORTUNITIES AND THREATS(2018) Mishra, Saurabh; Ayyub, Bilal; Reliability Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)During the lifecycle of mega engineering projects such as: energy facilities, infrastructure projects, or data centers, executives in charge should take into account the potential opportunities and threats that could affect the execution of such projects. These opportunities and threats can arise from different domains; including for example: geopolitical, economic or financial, and can have an impact on different entities, such as, countries, cities or companies. The goal of this research is to provide a new approach to identify and predict opportunities and threats using large and diverse data sets, and ensemble Long-Short Term Memory (LSTM) neural network models to inform domain specific foresights. In addition to predicting the opportunities and threats, this research proposes new techniques to help decision-makers for deduction and reasoning purposes. The proposed models and results provide structured output to inform the executive decision-making process concerning large engineering projects (LEPs). This research proposes new techniques that not only provide reliable timeseries predictions but uncertainty quantification to help make more informed decisions. The proposed ensemble framework consists of the following components: first, processed domain knowledge is used to extract a set of entity-domain features; second, structured learning based on Dynamic Time Warping (DTW), to learn similarity between sequences and Hierarchical Clustering Analysis (HCA), is used to determine which features are relevant for a given prediction problem; and finally, an automated decision based on the input and structured learning from the DTW-HCA is used to build a training data-set which is fed into a deep LSTM neural network for time-series predictions. A set of deeper ensemble programs are proposed such as Monte Carlo Simulations and Time Label Assignment to offer a controlled setting for assessing the impact of external shocks and a temporal alert system, respectively. The developed model can be used to inform decision makers about the set of opportunities and threats that their entities and assets face as a result of being engaged in an LEP accounting for epistemic uncertainty.Item Optimizing Mass Customization Through Interaction Variability and Manufacturing Trade-offs(2017) Cage, Kailyn; Vaughn-Cooke, Monifa; Fuge, Mark D; Reliability Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Design methods that consider the complete physical system (human interfaces and functional capacities of human interfaces) and incremental distinctions in humans are not widely applied. Human beings vary from a cognitive and physical standpoint. Manufacturing approaches have attempted to implement mass customization to provide end users with personalized products. However, these approaches are limited since (1) mass customization is orthogonal to human variability and (2) manufacturing costs are increased, through additional time and parts, required when mass-producing customized products. This research facilitates the integration of traditional engineering performance metrics and biomechanics creating manufacturable innovations in customized design that target population accommodation. The present method captures (1) human and product interface interactions, (2) interaction accommodation, (3) the impact of interaction accommodation on engineering performance metrics (thermal, structural, fluid, reliability, etc.), and (4) number of products required to accommodate the population. Engineering design techniques provide a structured method for reducing product and performance metrics which provide the foundational framework for the optimization model(s) integrating this method. Optimization enables optimal performance metrics constrained by population accommodation, producing the product metrics and the number of products required to accommodate the population. This work is a novel approach for addressing complex questions for interaction variability in mass production targeting population accommodation while maintaining product performance, which facilitate addressing larger problems of mass customization in mass production.