Feature Selection for Audio-Based Fault Detection in Pumps

Whilst extensive work has been done on fault detection in bearings using sound, very little has been accomplished with other machine components and machinery partly due to the scarcity of datasets. The recent release of the Malfunctioning Industrial Machine Investigation and Inspection (MIMII) dataset opens the opportunity for research into malfunctioning machines like pumps, fans, slide rails, and valves. In this paper, we compare common features from audio recordings to investigate which best support the classification of malfunctioning pumps. We evaluate our results using the Area Under the Curve (AUC) as a performance metric and determine that the log mel spectrum is a very useful feature, at least for this dataset, but that other features can enhance detection performance when ambient noise is present (improving AUC from 0.88 to 0.94 in one case). Also, we find that mel Frequency Cepstral Coefficients (MFCC) perform substantially poorer as features than a sampled mel spectrogram.

Mechanical Testing and Reliability Analysis for 3D Printed Cubic Lattices

3D printing has enabled new avenues to design and fabricate diverse structures for engineering applications, such as mechanically efficient lattices. Lattices are useful as implants for biological applications for supporting in vivo loads. However, inconsistencies in 3D printing motivates a need to quantify uncertainties contributing to mechanical failure using probabilistic analysis. Here, 50 cubic unit cell lattice samples were printed and tested with designs of 50% porosity, 500-micron beam diameters, and 3.5mm length, width, and height dimensions. The average length, width, and height measurements ranged from 3.47mm to 3.48mm. The precision in printing with a 95% confidence level was greater than 99.8%. Lattice elastic moduli ranged from about 270 MPa to 345 MPa, with a mean of 305 MPa. Probabilistic analyses were conducted with NESSUS software. The distributions of input parameters were determined using a chi-square test. The first-order reliability method was used to calculate the probability of failure and sensitivity of each input parameter. The elastic modulus was the most sensitive among all input parameters, with 57% of the total sensitivity. The study quantified printing inconsistencies and sensitives using empirical evidence and is a significant step forward for designing 3D printed parts for mechanical applications.

Assessment of Elevator Risks and Code Requirements Associated With Slip, Trip and Fall Hazards

In previous work, the hazards associated with elevator door closures were identified and analyzed. Using the National Electronic Injury Surveillance System (NEISS) database of the U.S. Consumer Product Safety Commission (CPSC), incidents associated with door strikes were identified between the years 1990 to 2017. This current effort focuses on elevator slip, trip and fall hazards. The ASME A17.1 Safety Code for Elevators and Escalators requires that elevator systems be equipped with leveling devices to vertically align the car platform sill relative to the hoistway landing sill to attain a predetermined accuracy. Even with the leveling safety requirements, slip, trip and fall incidents for passengers exiting or entering elevators are known to occur. This paper will analyze elevator slip, trip and fall hazards using injury records from the NEISS database from 1990 to 2019. Relevant elevator incidents were extracted from this dataset through manual inspection of the text-based description fields of all elevator-related incident records found in the NEISS dataset from this time period. National projections of elevator incidents were then calculated from this extracted dataset and trended for the entire time period of 1990 through 2019. The age and sex distributions of these national projections were also analyzed. These projections and trends are then discussed in the context of ASME A17.1 requirements intended to mitigate the risks of injuries when entering or exiting an elevator.

Risk Analysis of Live Fire Shoot House Containment Systems Against Multiple Projectile Impacts

Many analytical and numerical models exist that can describe the effect of single projectile impacts on steel targets. These models are not adequate for the evaluation of live fire shoot house containment systems, which are subjected to repeated impact loading from small caliber projectiles over the lifetime of the structure. Models assuming perfectly rigid projectiles over-predict penetration depths. Models assuming rigid targets cannot predict any penetration, and hydrodynamic models are best suited to high velocity impacts well above the ranges of conventional ordinance. Development of sufficient analytical or numerical tools using traditional techniques would be either intractable, empirically based and unique to a given scenario, require unique material properties that are not commonly available, or require significant computational effort. Due to the limited amount of empirical data on multiple impact failure, classical reliability methods are not suitable for assessing the probability of containment system perforation. Using existing experimental results of .223 caliber ammunition against AR500 steel panels with 2-inch ballistic rubber, a commonly found protective system in these facilities, the cumulative effects of multiple projectiles were quantified to estimate the number of impacts required to perforate the target material. Impacts were simulated from normal distributions of the x and y coordinates describing the impact point using a cartesian coordinate plane. The impact resistance of the steel was also simulated from a triangular distribution to account for the variability of the experimental results. Monte Carlo Simulation was then used to estimate the expected number of impacts to cause failure at a single point on the target. Using this collective model, it was possible to determine that the distribution of the number of rounds to cause target failure approached a normal distribution. The results indicated that the mean impacts at failure was 11800 with a standard deviation of 800 impacts. Finally, targeting the allowable risk level for structural failure from the JCSS probabilistic model code from the simulated normal distribution, it was determined that the safe number of impacts was approximately 7996. Decision makers can utilize the safe number of impacts to inform training guidance for the future use of facilities and to develop effective inspection requirements. This model can also be adapted to evaluate similar training facilities and to assess how other small caliber projectile impacts would affect live fire shoot house containment systems, providing a useful tool for the design and analysis of future and the assessment of existing facilities for use with ammunition that did not exist during its design.

A Kriging Based Fast and Efficient Method for Defect Detection in Massive Pipelines Using Magnetic Flux Leakages

Systems in service continue to degrade with passage of time. Pipelines are among the most common systems that wear away with usage. For public safety it is of utmost importance to monitor pipelines and detect new defects within the pipelines. Magnetic flux leakage (MFL) testing is a widely used nondestructive evaluation (NDE) technique for defect detections within the pipelines, particularly those composed of ferromagnetic materials. Pipeline inspection gauge (PIG) procedure based on line-scans or 2D-scans can collect accurate MFL readings for defect detection. However, in real world applications involving large pipe-sectors such extensive scanning techniques are extremely time consuming and costly. In this paper, we develop a fast and cheap methodology that does not need MFL readings at all the points used in traditional PIG procedures but conducts defect detection with similar accuracy. We consider an under-sampling based scheme that collects MFL at uniformly chosen random scan-points over large lattices instead of extensive PIG scans over all lattice points. Based on readings for the chosen random scan points, we use Kriging to reconstruct MFL readings over the entire pipe-sectors. Thereafter, we use thresholding-based segmentation on the reconstructed data for detecting defective areas. We demonstrate the applicability of our methodology on synthetic data generated using popular finite element models as well as on MFL data collected via laboratory experiments. In these experiments spanning a wide range of defect types, our proposed novel MFL based NDE methodology is witnessed to have operating characteristics within the acceptable threshold of PIG based traditional methods and thus provide an extremely cost-effective, fast procedure with competing error rates that can be successfully used for scanning massive pipeline sectors.

Developing an Efficient Approach for Unmanned Aerial Vehicle Reliability Analysis

Unmanned Aerial Vehicles (UAV) are increasingly get popularity in many applications. Their operation requires high level of safety and reliability to accomplish successful missions. In this study, the reliability was comparatively analyzed by different available approaches to select the efficient method. First, failure model of the system is developed. Then, three different scenarios are considered to study the effect of redundancies on the system reliability results. In the first scenario, there is no redundancy where in the second scenario there is only one redundant component and in the third scenario, there are three redundant components. Static reliability analysis such as Fault Tree Analysis (FTA), Reliability Block Diagram (RBD), Markov Chain (MC), and Bayesian Networks (BN) are applied on proposed scenarios and results are obtained. Regarding to time dependencies between redundant components, a dynamic-based methodology is also developed in this study through applying Dynamic Fault Tree (DFT) analysis. Proposed static and dynamic approaches are applied on an UAV as a case study and results are discussed. Finally, characteristics of each methodology and related conditions are clarified for selecting the efficient reliability analysis approach.

Adequate Lubrication As the First (and Often Only) Line of Defense Against Dust Explosions in Grain Elevators

Finely divided solid materials (e.g., dusts and fines), when dispersed in the air, can fuel particularly violent and destructive explosions. In this paper we will discuss a case study involving a dust explosion in a grain elevator and how a careful bearing greasing policy could have avoided it. We present the most common conditions that lead to bearing overheating which can serve as the ignition source for a dust explosion. Additionally, we stress the need to raise awareness among operators about the wide variety of greases available, and given this wide variety, it is critical for facilities to ensure they use a grease with characteristics as close as possible as these recommended by the equipment manufacturer.

Numerical Simulation of Detonation Propagation in Flame Arrestor Applications

A detail numerical study of detonation propagation and interaction with a flame arrestor product was conducted. The simulation domain was based on the detonation flame arrestor validation test setup. The flame arrestor element was modeled as a porous zone using the Forchheimer equation. The coefficients of the Forchheimer equation were determined using experimental data. The Forchheimer equation was incorporated into the governing equations for axisymmetric reactive turbulent flow as a momentum sink. A 21-step elementary reaction mechanism with 10 species was used to model the stoichiometric oxyhydrogen detonation. Different cases of detonation propagation including inviscid, viscous adiabatic, and viscous with heat transfer and a porous zone were studied. A detail discussion of the detonation propagation and effect of the arrestor geometry, the heat transfer and the porous zone are presented. The inviscid numerical model solutions of the detonation propagation parameters are compared to one-dimensional analytical solution for verification. The viscous solutions are qualitatively compared to historical experimental data which shows very similar trend. The effect of the porous media parameters on shock transmission and re-initiation of detonation is presented.

Naturalistic Male Skateboard User Speed Study

Skateboards have been used as a means of transportation and extreme sport participation for decades. However, the prevalence of skateboards as a source of transportation is increasing. The laws that permit skateboard users to travel in roadways and in pedestrian walkways can vary by state, city, or county, allowing for a large variance in travel speed and user behavior. The amount of data available for the average speed of skateboard users during travel and trick initiation is limited. This study will preliminarily describe the natural travel and trick initiation speeds of skateboard users. The data that is presented in this study is beneficial to a vast audience including, but not limited to: traffic safety, road and intersection design, accident reconstruction, skateboard design, bearing design and useful life, and wheel design and useful life. This is an observational study of users on public spaces; no personal identification or biometric data was collected.

A Waveguide Ultrasonic Phased Array Radar for High Temperature Equipment Monitoring

This paper proposes a design for monitoring high temperature structures utilizing phased array waveguide transducers, where the active sensing elements work synthetically together to achieve damage detection. The waveguide transducer is comprised of a wave-generation piezo element and a waveguide bar conducting the wave energy into the host structure. A coupled-field local finite element model (FEM) is constructed to grasp an in-depth understanding of the wave generation behavior. Via the harmonic analysis, optimum wave generation frequency and mode type can be analyzed. Thereafter, an array of waveguide transducer elements are attached to the host structure to study their wave manipulating ability. The system works based on the principle of phased array theory; the excitation instant of each element is controlled to form an equiphasic surface. In this way, the wave propagation phenomenon such as focusing and directional steering can be realized. An ultrasonic radar for high temperature working condition can thus be realized. The proposed system possesses great application potential to enhance the performance of Lamb wave SHM and NDE systems for high temperature structures.