Luna Innovations (United States)

Hydraulic actuators are used across the shipping industry and are installed in various environments performing a wide array of operations. Their widespread use necessitates a condition-based maintenance program, ideally utilizing machine learning models to provide real-time estimates of the health condition, to maintain an acceptable level of fleet readiness. This paper presents a hydraulic actuator dataset collected from six industrial actuators under various fault conditions and loads. The dataset is publicly available for use in other studies. This paper describes the actuator test stand and provides a series of baseline machine learning experiments. The numerical experiments demonstrate that a machine learning model can identify classes across a range of machine learning problems. However, these numerical experiments use data from each actuator during training. The final numerical experiment withholds data from individual actuators and fault types during training. The model is unable to correctly classify the withheld fault type during this experiment. This presents an opportunity for the data set to be used in further research on generalizing models.

ABSTRACT Galvanic corrosion can cause significant damage on aerospace structures due to mixed material components and aggressive operating environments under atmospheric conditions. Predicting both the galvanic corrosion magnitude and distribution is important to help inform the design and maintenance of new and existing structures, respectively. Prior efforts have developed atmospheric models for risk mapping, but validation has been in the form of simulated atmospheric environments in the lab, rather than actual corrosion processes in outdoor environments, with the latter being more representative of operational conditions for aerospace components. Therefore, the present work investigates the development and validation of an atmospheric model predicting the galvanic corrosion magnitude and distribution that is expected to occur in outdoor conditions.

Jet quenching is a phenomenon in heavy-ion collisions arising from jet interactions with the quark-gluon plasma (QGP). Its study is complicated by the interplay of multiple physics processes that affect jet observables. In addition, detector effects may influence the results and must be accounted for when identifying quenched jets. We employ a Long Short-Term Memory (LSTM) neural network trained on jet substructure, incorporating parton shower history, to predict jet-by-jet quenching levels. Using photon-jet samples from the \textsc{Jewel} event generator, we show that the LSTM predictions strongly correlate with true jet energy loss. This validates that the model effectively learns the features of jet-QGP interaction. We simulate detector effects using \textsc{Delphes} simulation framework and demonstrate that the method identifies quenching effects in a realistic environment. We test the approach with photon-jet momentum imbalance, jet fragmentation function, and jet shape, which were not included in the training, confirming its ability to distinguish true quenching features.

While supervised machine learning is prevalent in prognostics and health management applications, the success of these models is dependent upon being trained on accurate data. Training data collected with faulty sensors can degrade the performance of these models when deployed in an operational environment. This study investigates the impact of faulty data and the robustness of feature extraction methods and tree-based classifiers. This study also provides an open-source software package for injecting faults into time series data. The numerical experiments are performed on an open-source hydraulic actuator data set and demonstrate that certain features are robust to certain types of faults and that more complex models, such as ensemble techniques, are more robust to sensor faults than simple models. This work suggests that more complex models and larger (and possibly redundant) feature sets may be preferred in situations where sensor faults are likely. Furthermore, certain feature extraction techniques may be selected if certain faults are more likely than others.

While identification of epigenetic changes in individuals with psychiatric dysfunctions such as post-traumatic stress disorder (PTSD) is paramount to genomic research, there is no rapid and simplified way to detect an epigenetic marker such as DNA methylation in genes. Here, we introduce a faster, simpler method to detect methylation in the form of 5-methylcytosine (5mC, termed as PTSD-associated base) in known CpG sites using nanoenhanced surface plasmon resonance imaging-based epigenetic assay (EpiNanoSPRi). This assay platform simultaneously detects a panel of single, site-specific PTSD bases in target genes or regions using an anti-5mC antibody and a universal nanoenhancer on a gold-coated sensing chip. Not only can EpiNanoSPRi identify 5mC at the single-base level, but it also can quantify the extent of DNA methylation. Our method is superior and more practical to bisulfite-based DNA sequencing techniques as it will significantly reduce DNA methylation identification from 4 days (e.g., DNA Sequencing) to 9 h without massive analysis workflow. This platform can potentially be applied to diagnose other psychiatric disorders such as Alzheimer's, Parkinson's, dementia, schizophrenia, and Huntington's diseases.

A nosologia do Transtorno do Espectro Autista - TEA é historicamente muito recente, tendo sua primeira aparição como uma categoria própria em 1980, sendo antes tido como um tipo de esquizofrenia na infância. O autismo é hoje compreendido como uma condição do neurodesenvolvimento cuja característica mais marcante é um déficit persistente nas capacidades de comunicação e interação social, prevalecendo em 1 a cada 36 crianças, de acordo com o CDC. A definição e os critérios mais aceitos do transtorno repousam nos principais manuais, nomeadamente o Manual Diagnóstico e Estatístico de Transtornos Mentais (DSM) e o Código Internacional de Doenças (CID), que buscam se adequar às novas atualizações científicas e socioculturais, com o intuito de tornar os modelos diagnósticos mais claros e confiáveis. Em 2022 foi publicada a revisão da quinta edição do DSM (DSM-5-TR), que engendrou algumas mudanças no que diz respeito tanto ao critério, quanto ao texto da condição para o TEA. Nesse sentido, o objetivo do presente artigo é apresentar um estudo das alterações trazidas pelo DSM-5-TR na categoria do TEA, bem como discutir os contextos e motivos por trás dessas mudanças.

Abstract In this laboratory work, zero resistance ammeter measurements were used to collect galvanic corrosion measurements for scribed organic coating samples and e-coated samples in situ. For this work, a new sample design was derived from samples currently used in the automotive industry consisting of a base plate and a cover plate of another material to form a galvanic couple with a crevice between them. The samples were connected to a developmental test system that uses zero resistance ammeter (ZRA) measurements to collect galvanic corrosion rates for multiple samples in situ. E-coated samples were evaluated in accelerated corrosion tests for comparison with traditional organic coatings. A description of the samples, electrochemical measurement device, and methods for coating testing will be reported along with the analysis of the resulting data and its implications for evaluating various coating types.