Doctor of Philosophy, The Ohio State University, 2006, Industrial and Systems Engineering

The safety of the American workplace has improved dramatically over the past 30 years. This improvement is directly correlated with the adoption and enforcement of OSHA regulations (OSHA, “OSHA Facts”). However, despite the great strides that have been achieved, some industry sectors continue to produce unnecessarily high numbers of serious and preventable injuries. Machine-related injuries are responsible for nearly half of the thousands of amputation injuries that occur each year. Most machine injuries are preventable through known methods that are well documented. For most machines, OSHA provides guarding and operational requirements that are very general and broadly applicable. However, in the case of mechanical power presses the codes are quite specific and intended to address the specific hazards associated with such presses. This study proposes that the OSHA codes related to mechanical power presses are adequate and address most of the guarding concerns, but employers often fail to comply with the codes, apparently out of a lack of understanding of their implementation. It is hypothesized that an effective tool to help guide personnel through the evaluation of press safety hazards will improve the likelihood of an individual in accurately identifying press hazards. Based on the perceived need, a software tool was developed to assist in the hazard identification process. This tool was tested experimentally to determine its effectiveness. The hazard evaluation performance of a software-assisted group of novices was compared with the performances of a peer group and a group of press professionals, both comparison groups using traditional evaluation methods (specifically ANSI B11.TR3). Each of the experimental groups evaluated three different mechnical power presses. The hazards identified by each experimental group were to address the specific requirements of the applicable OSHA codes for guarding of mechanical power presses (29CFR1910.212 and 29CFR1910.217). Th (open full item for complete abstract)

Committee: Gary Maul (Advisor) Subjects:

Master of Science (MS), Ohio University, 2024, Mechanical Engineering (Engineering and Technology)

In recent years, lithium-ion batteries (LIBs) have played an essential role in nowadays energy storage system, especially electric vehicles (EVs) and portable electronics because of its high energy density and long cycle life [1, 2]. However, one of the biggest challenges is how to guarantee their dependability and trustworthiness. In the present investigation, Acoustic Emission (AE) and Ultrasound Testing (UT) techniques are systematically employed to verify probable critical defects in the LIBs. Where AE technology is able to record the stress waves produced by the growth of the defects, UT uses high-frequency sound waves to penetrate the batteries and provide an indication of the internal voids. The performances of these approaches were systematically tested on as-received, pre-damaged and cold-soaked batteries. Different AE and UT activity patterns were shown in the results under various environmental conditions that influenced battery performance. Combining Acoustic Emission (AE) and Ultrasound Testing (UT) with clustering and outlier analysis machine learning algorithms improved defect detection effectiveness. Such research highlights that AE and UT can be robust noninvasive techniques for on-line health monitoring of LIBs that should aid in maintaining the longevity and operability of LIBs.

Committee: Brian Wisner (Advisor) Subjects: Acoustics; Mechanical Engineering

Doctor of Philosophy, University of Toledo, 2021, Engineering

Autonomous vehicles are the next revolution in transportation. They are capable of recognizing their surroundings, navigating, and avoiding obstacles without human intervention. Autonomous vehicles rely on advanced technologies such as Artificial Intelligence (AI) to become fully automated. In this dissertation, methods to improve autonomous vehicles' safety on roads are presented. A collision warning system is used to assist drivers. An application of the Naive Bayes classifier model – a supervised machine learning model based on Bayes' theorem – to determine the potential for rear-end collisions between highway vehicles is proposed. Two vehicles are utilized, with one vehicle following the other. The parameters studied are speed, distance, and acceleration–deceleration. A set of training examples involving over 100 potential collision scenarios have been analyzed. This dissertation also proposes the integration of artificial neural networks into the safety programmable logic controller (fail-safe PLC) to create an algorithm that controls a robotic vehicle and ensures safety on the roads. Artificial neural iv networks (ANNs) are a supervised machine learning model based on a computing system built to simulate the way the human brain processes and analyzes information. A fail-safe PLC offers a safety concept in the field of machine and personnel protection. A set of training examples involving more than 30 data was evaluated to train the artificial neural networks. In addition, a fail-safe PLC program was designed to perform under special conditions. Indoor obstacle avoidance courses were taken as examples to examine the effectiveness of the obstacle avoidance system. Simulation results show that the systems are successfully predicted and responded correctly to different driving scenarios.

Committee: Richard Molyet (Committee Chair); Mansoor Alam (Committee Member); William Evans (Committee Member); Mohammed Niamat (Committee Member); Junghwan Kim (Committee Member); Daniel Georgiev (Committee Member) Subjects: Engineering

Master of Science (MS), Ohio University, 2020, Industrial and Systems Engineering (Engineering and Technology)

A nontraditional decision tree approach was used to predict worker intent to use passive occupational exoskeletons in various manufacturing tasks. A dataset adapted from a previous study containing 33 records of participant, exoskeleton, and task combinations was used to create multiple decision tree models. Worker intent to use the exoskeleton was used as the target variable for all decision tree models. Data were collected during two separate sessions with fifteen participants at five manufacturing facilities in Ohio. Participants wore exoskeletons for under 30 minutes in each session and answered questions pertaining to personal characteristics, task characteristics, and personal preferences. Response data were used to create practitioner and research decision tree models. The practitioner models classified worker intent to use exoskeletons using only task characteristics and personal characteristics. Research models used personal characteristics, task characteristics, and personal preferences features to predict intent to use with all collected data. Both practitioner and research models may be useful for practitioners and exoskeleton developers for better understanding factors related to intent to use exoskeletons. All models created in the study yielded findings consistent with previous exoskeleton literature. This study demonstrated the ability of classification trees to identify nonlinear relationships in datasets relating to intent to use assistive technologies.

Committee: Diana Schwerha Ph.D (Advisor); Gary Weckman Ph.D (Committee Member); Dean Bruckner Ph.D (Committee Member); Timothy Ryan Ph.D (Committee Member) Subjects: Industrial Engineering

Doctor of Philosophy, University of Toledo, 2018, Engineering (Computer Science)

A strong emphasis on safety in commercial and military aviation is as old and as significant as the field of aviation itself. With the growing role of autonomy in aviation, the future of flight comprises of two general directions: manned and unmanned. Manned aircraft is the more established area, in which a human flight crew serves as the main driving force in ensuring an aircraft's safety and success. Within this time-tested concept, the most significant bottleneck of safety lies within a crew managing tasks of high mental workload. In recent years, autonomy has aided in easing cognitive workload. From there, the challenge lies within applying a seamless blend of human and autonomous control based on the needs of one's mental load. Meanwhile, the field of unmanned aerial vehicles (UAVs) poses its own unique challenges of integrating into a shared airspace and transitioning from remote human-centric control to fully autonomous control. In such a case, minimizing discrepancies between predicted UAV behavior and actual outcomes is an ongoing task to ensure a safe and reliable flight. While manned and unmanned flight safety may seem distinctly different in these regards, this dissertation proposes an overarching common theme that lies within the ability to effectively model inputs and outputs through machine learning to predict potential safety hazards and thereby improve the overall flight experience. This process is conducted by 1) evaluating different machine learning techniques on assessing cognitive workload, 2) predicting trajectories for autonomous UAVs, and 3) developing adaptive systems that dynamically select appropriate algorithms to ensure optimal prediction accuracy at any given time. The first phase of the research involves the manned side of flight safety and does so by examining effects of different machine learning techniques used for assessing cognitive workload. This begins by comparing the different algorithms on four different datasets i (open full item for complete abstract)

Committee: Vijay Devabhaktuni PhD (Committee Chair); Mansoor Alam PhD (Committee Member); Ahmad Javaid PhD (Committee Member); Devinder Kaur PhD (Committee Member); Weiqing Sun PhD (Committee Member); Lawrence Thomas PhD (Committee Member) Subjects: Computer Engineering; Computer Science

Doctor of Philosophy, The Ohio State University, 2014, Mechanical Engineering

A number of issues can plague the reliability of results computed using any software. When software is used to make safety critical decisions it is imperative that the results be dependable and that either there be no errors in the computed results or the error in the results be known to the user. This dissertation addresses the issues that can affect the accurate computation of results of scientific software. Scientific software is defined as software that performs extensive computations to model some physical phenomenon and provide results that can support decision making. The primary issues that affect the results of scientific software can be broadly classified into three categories: (1) incorrect requirements (2) coding errors and (3) missing requirements. This dissertation addresses these issues by adapting the Hazard and Operability (HAZOP) method for application to scientific software. Before applying HAZOP method to scientific software, a representation of the system (scientific software) is developed from its requirements written in formal language (Z specification language). Using a formal notation in writing requirements reduces ambiguity in the specification and also offers an opportunity to mathematically verify them. Another advantage of using formal specifications is that test cases can be developed from the resulting representation of the system which tests the functionality of the system. Missing requirements pose a big threat since they cannot be identified from testing and therefore can reduce the dependability on the results without the knowledge of the user. Missing requirements are commonly observed to be related to operational environment of the system. HAZOP analysis helps in the identification of such requirements as it provides a structured approach for exploration of system failure modes by suggesting hypothetical failures. This dissertation provides details on (1) development of system representation from Z-specification language and (2) (open full item for complete abstract)

Committee: Carol Smidts Professor (Advisor); Tunc Aldemir Professor (Committee Member); Richard Denning Professor (Committee Member); Lei Cao Professor (Committee Member); Laura Lindsey Professor (Committee Member) Subjects: Engineering; Mathematics; Mechanical Engineering

Master of Science (MS), Ohio University, 2013, Civil Engineering (Engineering and Technology)

In this thesis, road safety assessment and prediction modeling for U.S. states fatal crashes are addressed. In the first part, a DEA-based Malmquist Index model was developed to assess the relative efficiency and productivity of U.S. states in decreasing the number of road fatalities. Even though the national trend in fatal crashes has reached to the lowest level since 1949 (Traffic Safety Annual Assessment Highlights, 2010), a state-by-state analysis and comparison has not been studied considering other characteristics of the holistic national road safety assessment problem in any work in the literature or organizational reports. The single output, fatal crashes, and five inputs were aggregated into single road safety score and utilized in the DEA-based Malmquist Index mathematical model. The period of 2002-2008 was considered due to data availability for the inputs and the output considered. According to the results, there is a slight negative productivity (an average of -0.2 percent productivity) observed in the U.S. on minimizing the number of fatal crashes along with an average of 2.1 percent efficiency decline and 1.8 percent technological improvement. The productivity in reducing the fatal crashes can only be attributed to the technological growth since there is a negative efficiency growth is occurred. It can be concluded that even though there is a declining trend observed in the fatality rates, the efficiency of states in utilizing societal and economical resources towards the goal of zero fatality is not still efficient. In the second part, a nonparametric prediction model, Artificial Neural Network, was developed to assist policy makers in minimizing fatal crashes across the United States. Seven input variables from four safety performance input domains while fatal crashes was utilized as the single output variable for the scope of the research. Artificial Neural Networks (ANN) was utilized and the best neural network model was developed out of 1000 n (open full item for complete abstract)

Committee: Deborah McAvoy Ph.D. (Advisor); Byung-Cheol Kim Ph.D. (Committee Member); Ken Walsh Ph.D. (Committee Member); M. Khurrum S. Bhutta Ph.D. (Committee Member) Subjects: Civil Engineering; Industrial Engineering; Transportation

Master of Science (MS), Ohio University, 2007, Computer Science (Engineering)

This thesis constitutes an effort of verifying and validating a safety system designed for a specific research facility. An initial comprehensive review of the system design is presented, detailing all the relevant aspects of the system and investigating the way its design development interrelates to the formal "safety analysis" procedures proposed in the literature. The verification process includes the development of a complete formal specification for the system and the investigation of how well the original design follows its formal requirements. The validation process details the system's hardware and software implementations, discusses the testing approach, and evaluates the final outcomes. In summary, this work can be considered as an effort to prove that the operation of the laboratory in question, within the designed safety system's scope, is safe.

Committee: Frank Drews (Advisor) Subjects: