SEARCH · Search NASA

Results for “flexibility metric”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 307 records · Page 17

All sources NASA NTRS DOE OSTI

DOE OSTI2021

VFA Biorefinery Design for Informed Production of Ground and Aviation Fuels

The rising demand for low-carbon intensity fuel options requires evaluation of a system which can support their production economically and sustainably. Volatile fatty acids (VFAs) ranging from C2 to C8 can be derived in high yield from arrested anaerobic digestion of biomass. This talk provides an overview of our work in which we upgrade wet waste-derived VFAs using ketonization to elongate their carbon backbone to reach a range relevant to jet or diesel fuels. Kinetic models were used to predict ketone profiles from varying VFA profiles, thereby informing potential carbon flow to either (a) mixed paraffins for use in diesel or aviation fuel, or (b) ethers for use in diesel fuel. To do this, the anticipated products were screened for critical fuel characteristics using predictive tools. The criteria for neat and blended bioblendstocks were chosen based on conventional petrofuel requirements, and they were applied to inform fuel targets, identify limiting characteristics, and guide conversion development. While paraffins can serve as either diesel or aviation fuels, the latter has stringent criteria which include a precise distillation curve range tied to carbon distribution. Fuels which fall outside this range typically fail to meet other property metrics, and as such flashpoint and viscosity were identified as potentially limiting properties of this VFA aviation fuel. However, paraffin fractions which fall outside aviation criteria may meet diesel fuel requirements, which are looser except for the flashpoint minimum. Flashpoint was also a concern for smaller VFA diesel ethers, which posed an additional oxygenate risk of high water solubility. This fuel-informed conversion design process was demonstrated through production of paraffins with reduced sooting as compared to petrodiesel (~34%), and increased renewable aviation fuel blend levels (>70%). VFA-derived ethers improved petrodiesel by increasing fuel autoignition quality by 56% and reducing sooting by 86%. These VFA aviation and diesel fuels could enable greenhouse gas (GHG) reductions of over 165% and 50%, respectively, relative to their purely fossil-derived counterparts. Collectively, this work provides (i) an overview of how we leveraged the flexibility of VFAs for fuel production; and (ii) insight into the potential of a VFA biorefinery to accommodate varied feed and fuel applications, while also considering the merit of tailored fuel production pathways and GHG reduction potential.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

DOE OSTI2024

Review of data-driven models for quantifying load shed by non-residential buildings in the United States

Shifting and shedding power demand in buildings can be cost-effective techniques for grids to function reliably and for end users to earn compensation. Grid operators reimburse customers in proportion to the quantity of load shed. Simple data-driven methods are used to quantify this shed, which is the difference between a measured load during the event and modeled "baseline" that would have occurred in absence of the event. These methods have evolved over the years and in many cases have been integrated with building physics, to make them a hybrid between physics based and empirical models. However, there is no comprehensive analysis that provides guidance to building operators, grid operators and researchers in selecting appropriate models based on their specific needs and available data. Here, this work aims to fill this gap by critically assessing the performance of baseline models put forward from the year 2000 through 2023. The literature reviewed includes reports generated by grid operators, reports from national laboratories and academic journal articles. The work outlines modeling features like the inputs, training period, estimation method, adjustments to fine tune the predictions and metrics to evaluate the performance. A comprehensive list of 50 models has been provided. For each model, the study explores the applicability of the model to weather sensitive buildings, variability in the building profile, timing of the event, and whether the building reduces energy consumption before an event. The work identifies the situations in which a particular model works and draws lessons based on evidence of performance. Finally, recommendations to aid in model selection are given.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

A Stochastic Multi-Criteria Decision-Making Algorithm for Dynamic Load Prioritization in Grid-Interactive Efficient Buildings

Increasing deployment of advanced sensing, controls, and communication infrastructure enables buildings to provide services to the power grid, leading to the concept of grid-interactive efficient buildings. Since occupant activities and preferences primarily drive the availability and operational flexibility of building devices, there is a critical need to develop occupant-centric approaches that prioritize devices for providing grid services, while maintaining the desired end-use quality of service. In this paper, we present a decision-making framework that facilitates a building owner/operator to effectively prioritize loads for curtailment service under uncertainties, while minimizing any adverse impact on the occupants. The proposed framework uses a stochastic (Markov) model to represent the probabilistic behavior of device usage from power consumption data, and a load prioritization algorithm that dynamically ranks building loads using a stochastic multi-criteria decision-making algorithm. The proposed load prioritization framework is illustrated via numerical simulations in a residential building use-case, including plug-loads, air-conditioners, and plug-in electric vehicle chargers, in the context of load curtailment as a grid service. Suitable metrics are proposed to evaluate the closed-loop performance of the proposed prioritization algorithm under various scenarios and design choices. Scalability of the proposed algorithm is established via computational analysis, while time-series plots are used for intuitive explanation of the ranking choices.

24 POWER TRANSMISSION AND DISTRIBUTION↗

NASA NTRS2009

Benefits Assessment of Algorithmically Combining Generic High Altitude Airspace Sectors

In today's air traffic control operations, sectors that have traffic demand below capacity are combined so that fewer controller teams are required to manage air traffic. Controllers in current operations are certified to control a group of six to eight sectors, known as an area of specialization. Sector combinations are restricted to occur within areas of specialization. Since there are few sector combination possibilities in each area of specialization, human supervisors can effectively make sector combination decisions. In the future, automation and procedures will allow any appropriately trained controller to control any of a large set of generic sectors. The primary benefit of this will be increased controller staffing flexibility. Generic sectors will also allow more options for combining sectors, making sector combination decisions difficult for human supervisors. A sector-combining algorithm can assist supervisors as they make generic sector combination decisions. A heuristic algorithm for combining under-utilized air space sectors to conserve air traffic control resources has been described and analyzed. Analysis of the algorithm and comparisons with operational sector combinations indicate that this algorithm could more efficiently utilize air traffic control resources than current sector combinations. This paper investigates the benefits of using the sector-combining algorithm proposed in previous research to combine high altitude generic airspace sectors. Simulations are conducted in which all the high altitude sectors in a center are allowed to combine, as will be possible in generic high altitude airspace. Furthermore, the algorithm is adjusted to use a version of the simplified dynamic density (SDD) workload metric that has been modified to account for workload reductions due to automatic handoffs and Automatic Dependent Surveillance Broadcast (ADS-B). This modified metric is referred to here as future simplified dynamic density (FSDD). Finally, traffic demand sets with increased air traffic demand are used in the simulations to capture the expected growth in air traffic demand by the mid-term.

Bloem, Michael↗

NASA NTRSDate not supplied

Streamline-Traced, External-Compression Supersonic Inlets for Mach 2

A computational study was performed to explore the aerodynamic design and performance of streamline-traced, external-compression (STEX) inlets for Mach 2.0. The performance metrics included inlet flow rates, total pressure recovery, and total pressure distortion. The study explored the use of round and flattened capture cross-sections and a bleed slot. The design of the inlet and generation of the inlet geometry was performed using the Supersonic Inlet Design and Analysis(SUPIN) Tool. Computational grids were generated, and methods of computational fluid dynamics (CFD) were applied to solve the three-dimensional, turbulent flow through the inlets using the Wind-US CFD flow solver. It was found that the use of a bleed slot with about 5% bleed resulted in an inlet total pressure recovery of 95% at Mach 2 with acceptable radial and circumferential total pressure distortion. It was also found that there was only a slight decrease in performance between a round and flattened capture cross-section. This suggests that the use of streamline tracing offers flexibility in shaping the inlet capture cross-section for a more favorable integration with an aircraft wing or fuselage. This flexibility and the good performance of the inlet provides a novel inlet design for future commercial supersonic aircraft.

Computational Fluid Dynamics↗

DOE OSTI2022

When Do Efficiency and Demand Flexibility Go Hand-in-hand?

Utilities have been experimenting with integrated demand side management (IDSM) programs since the 2000s. The potential benefits of improved program cost-effectiveness and customer engagement from combining energy efficiency (EE), demand flexibility (DF), and other distributed energy resources into an integrated customer offering have been recognized although there are several known regulatory and program administrative challenges. In addition, as buildings adopt EE measures, the baseline load profile change generally reduces the potential load that can be shed or shifted. This has been a significant technical barrier for customers and program implementers. However, it is a myth that EE always reduces DF. Load change from EE can be time-varying. Therefore, whether EE improves or reduces DF should be evaluated on an individual measure basis accounting for weather dependencies and interactions. For IDSM program design purpose, it is useful to understand how common EE features influence DF and the underlying building physics. In this paper, we use parametric simulations of a prototype medium office building to evaluate how various EE features influence DF, measured by a “demand decrease intensity” (W/ft2) metric. These EE features cover envelope characteristics, internal loads, and airside HVAC system. The parametric analysis shows that some efficient HVAC control measures will increase DF but not the traditional building envelope, lighting, and ventilation-related efficiency measures. These findings contribute to the technical basis for achieving enhanced energy benefits by packaging appropriate HVAC control measures in IDSM program design. Program developers should further validate these results in targeted pilot projects.

Yu, Lili↗

NASA NTRS2014

Cost Analysis In A Multi-Mission Operations Environment

Spacecraft control centers have evolved from dedicated, single-mission or single missiontype support to multi-mission, service-oriented support for operating a variety of mission types. At the same time, available money for projects is shrinking and competition for new missions is increasing. These factors drive the need for an accurate and flexible model to support estimating service costs for new or extended missions; the cost model in turn drives the need for an accurate and efficient approach to service cost analysis. The National Aeronautics and Space Administration (NASA) Huntsville Operations Support Center (HOSC) at Marshall Space Flight Center (MSFC) provides operations services to a variety of customers around the world. HOSC customers range from launch vehicle test flights; to International Space Station (ISS) payloads; to small, short duration missions; and has included long duration flagship missions. The HOSC recently completed a detailed analysis of service costs as part of the development of a complete service cost model. The cost analysis process required the team to address a number of issues. One of the primary issues involves the difficulty of reverse engineering individual mission costs in a highly efficient multimission environment, along with a related issue of the value of detailed metrics or data to the cost model versus the cost of obtaining accurate data. Another concern is the difficulty of balancing costs between missions of different types and size and extrapolating costs to different mission types. The cost analysis also had to address issues relating to providing shared, cloud-like services in a government environment, and then assigning an uncertainty or risk factor to cost estimates that are based on current technology, but will be executed using future technology. Finally the cost analysis needed to consider how to validate the resulting cost models taking into account the non-homogeneous nature of the available cost data and the decreasing flight rate. This paper presents the issues encountered during the HOSC cost analysis process, and the associated lessons learned. These lessons can be used when planning for a new multi-mission operations center or in the transformation from a dedicated control center to multi-center operations, as an aid in defining processes that support future cost analysis and estimation. The lessons can also be used by mature serviceoriented, multi-mission control centers to streamline or refine their cost analysis process.

Newhouse, M.↗

NASA NTRS2014

Cost Analysis in a Multi-Mission Operations Environment

Spacecraft control centers have evolved from dedicated, single-mission or single mission-type support to multi-mission, service-oriented support for operating a variety of mission types. At the same time, available money for projects is shrinking and competition for new missions is increasing. These factors drive the need for an accurate and flexible model to support estimating service costs for new or extended missions; the cost model in turn drives the need for an accurate and efficient approach to service cost analysis. The National Aeronautics and Space Administration (NASA) Huntsville Operations Support Center (HOSC) at Marshall Space Flight Center (MSFC) provides operations services to a variety of customers around the world. HOSC customers range from launch vehicle test flights; to International Space Station (ISS) payloads; to small, short duration missions; and has included long duration flagship missions. The HOSC recently completed a detailed analysis of service costs as part of the development of a complete service cost model. The cost analysis process required the team to address a number of issues. One of the primary issues involves the difficulty of reverse engineering individual mission costs in a highly efficient multi-mission environment, along with a related issue of the value of detailed metrics or data to the cost model versus the cost of obtaining accurate data. Another concern is the difficulty of balancing costs between missions of different types and size and extrapolating costs to different mission types. The cost analysis also had to address issues relating to providing shared, cloud-like services in a government environment, and then assigning an uncertainty or risk factor to cost estimates that are based on current technology, but will be executed using future technology. Finally the cost analysis needed to consider how to validate the resulting cost models taking into account the non-homogeneous nature of the available cost data and the decreasing flight rate. This paper presents the issues encountered during the HOSC cost analysis process, and the associated lessons learned. These lessons can be used when planning for a new multi-mission operations center or in the transformation from a dedicated control center to multi-center operations, as an aid in defining processes that support future cost analysis and estimation. The lessons can also be used by mature service-oriented, multi-mission control centers to streamline or refine their cost analysis process.

Felton, Larry↗

DOE OSTI2024

Investigating resource-efficient neutron/gamma classification ML models targeting eFPGAs

There has been considerable interest and resulting progress in implementing machine learning (ML) models in hardware over the last several years from the particle and nuclear physics communities. A big driver has been the release of the Python package, hls4ml, which has enabled porting models specified and trained using Python ML libraries to register transfer level (RTL) code. So far, the primary end targets have been commercial field-programmable gate arrays (FPGAs) or synthesized custom blocks on application specific integrated circuits (ASICs). However, recent developments in open-source embedded FPGA (eFPGA) frameworks now provide an alternate, more flexible pathway for implementing ML models in hardware. These customized eFPGA fabrics can be integrated as part of an overall chip design. In general, the decision between a fully custom, eFPGA, or commercial FPGA ML implementation will depend on the details of the end-use application. In this work, we explored the parameter space for eFPGA implementations of fully-connected neural network (fcNN) and boosted decision tree (BDT) models using the task of neutron/gamma classification with a specific focus on resource efficiency. We used data collected using an AmBe sealed source incident on Stilbene, which was optically coupled to an OnSemi J-series silicon photomultiplier (SiPM) to generate training and test data for this study. We investigated relevant input features and the effects of bit-resolution and sampling rate as well as trade-offs in hyperparameters for both ML architectures while tracking total resource usage. The performance metric used to track model performance was the calculated neutron efficiency at a gamma leakage of 10 -3 . The results of the study will be used to aid the specification of an eFPGA fabric, which will be integrated as part of a test chip.

47 OTHER INSTRUMENTATION↗

NASA NTRS2022

Europa Clipper Payload Verification and Validation: Test and Analysis Program Design

NASA’s Europa Clipper mission – which is planned to launch in the mid-2020s – is designed to investigate the habitability of the jovian moon Europa with a state-of-the-art suite of over 10 sensors and instruments. Now past its project critical design review, Europa Clipper is starting the execution of its verification and validation (V&V) program. The payload V&V function in particular is responsible for designing and coordinating a thorough test and analysis program that will confirm the proper functioning and performance of our payload instruments. Given the large number of instruments to manage, we have developed a suite of standardized tests and analyses that are designed to cover all of the common interfaces across instruments, while allowing for flexibility to support the unique needs of individual instruments. This suite of activities, coordinated and/or performed by the payload team, forms the bulk of our payload verification and validation campaign. This paper provides a detailed summary of the full set of V&V activities in that suite, covering the four categories of activities: analysis of lower-level evidence, analysis via simulation, testing in the system testbeds, and testing in assembly, test, and launch operations (ATLO). We will describe the purpose of each activity type, how the activities are developed, and the review process for each type of activity. We then define the complete list of activities written for each instrument and how subtleties specific to certain instruments are managed. The discussion will provide a comprehensive understanding of the activities in the payload V&V purview and how various interfaces with different levels of the project are managed. This paper also describes key metrics the payload team uses to track V&V progress and some insights into the teaming design and division of labor we use to ensure consistency and completeness across the entire payload suite. Finally, the paper will offer a snapshot of the status of the Payload V&V program prior to instrument delivery and offer insights into the upcoming phase of instrument delivery and the start of ATLO.

Lee, Dennis↗

NASA NTRS2011

An Overview of NASA's Current Materials Development Efforts for Mars EDL

Current roadmaps point to landing heavy masses (cargo, followed by manned vehicles) on Mars in the 2030's and the existing entry, descent and landing (EDL) technology will not be sufficient to facilitate such missions. In 2009 the Exploration Technology Development Program (ETDP) established the Entry, Descent and Landing Technology Development Project (EDL TDP), to be managed programmatically at Langley Research Center (LaRC) and technically a Ames Research Center (ARC). The purpose of the project is to further the technologies required to land heavy (approximately 40 metric ton) masses on Mars to facilitate exploration. The EDL TDP contains three technical elements. They are: 1) Thermal Protection Systems (TPS) development 2) Modeling and Tools (MAT) development 3) Supersonic Retropropulsion (SRP) development The primary goals of the EDL TDP TPS element is to design and develop TPS materials capable of withstanding the severe aerothermal loads associated with aerocapture and entry into the Martian atmosphere while significantly decreasing the TPS mass fraction contribution to the entry system. Significant advancements in TPS materials technology are needed in order to enable heavy mass payloads to be successfully landed on the Martian surface for robotic precursors and subsequent human exploration missions. The EDL TDP TPS element is further divided into two different TPS concepts for Mars EDL those being: 1) Rigid TPS for a mid L/D aeroshell with the capability to withstand dual pulsed heating environments as high as 500 W/square cm for aerocapture and 130 W/square cm for entry 2) Flexible TPS for a deployable aerodynamic decelerator with the capability to withstand dual pulsed heating environments as high as 120 W/square cm for aerocapture and 30 W/square cm for entry NASA, along with its vendors, has begun developing and testing materials for each of the deceleration approaches. These include multi-layer rigid ablators and flexible ablative materials. In order to model the response of these types of materials, new and improved modeling techniques will be required. This presentation will outline the types of materials that are under development and illustrate the need for advancement in modeling of ablative materials.

Beck, Robin A.↗

NASA NTRS1993

Mirror fusion propulsion system - A performance comparison with alternate propulsion systems for the manned Mars mission

The performance characteristics of several propulsion technologies applied to piloted Mars missions are compared. The characteristics that are compared are Initial Mass in Low Earth Orbit (IMLEO), mission flexibility, and flight times. The propulsion systems being compared are both demonstrated and envisioned: Chemical (or Cryogenic), Nuclear Thermal Rocket (NTR) solid core, NTR gas core, Nuclear Electric Propulsion (NEP), and a mirror fusion space propulsion system. The proposed magnetic mirror fusion reactor, known as the Mirror Fusion Propulsion System (MFPS), is described. The description is an overview of a design study that was conducted to convert a mirror reactor experiment at Lawrence Livermore National Lab (LLNL) into a viable space propulsion system. Design principles geared towards minimizing mass and maximizing power available for thrust are identified and applied to the LLNL reactor design, resulting in the MFPS. The MFPS' design evolution, reactor and fuel choices, and system configuration are described. Results of the performance comparison shows that the MFPS minimizes flight time to 60 to 90 days for flights to Mars while allowing continuous return-home capability while at Mars. Total MFPS IMLEO including propellant and payloads is kept to about 1,000 metric tons.

Deveny, M.↗

NASA NTRS1993

Mirror fusion propulsion system: A performance comparison with alternate propulsion systems for the manned Mars Mission

Schulze, Norman R.↗

NASA NTRSDate not supplied

Crew Scheduling Performance

As NASA considers long-duration exploration missions (LDEMs), it is envisioned that crew will behave more autonomously as compared to low-Earth orbit missions. The necessary shift of Ops Planners’ complete management of scheduling and planning to provide flexibility for crew to manage their own schedule in real-time requires significant research and investigation in evaluation of concepts of operations, of software tools to support tasks, and of crew performance to complete scheduling tasks. Our research objective is to characterize the human performance envelope for the task of planning and scheduling (crew self-scheduling), develop countermeasures to mitigate adverse performance effects due to plan complexity, and inform performance standards and guidelines based on research results. In our efforts to understand and characterize scheduling performance of crew members, we have conducted subject matter expert interviews to obtain feedback on scheduling plan complexity drivers and plan goodness. Following the pilot study conducted last year, we designed a study and have begun remote human subject testing to evaluate non-expert human performance, workload, and situational awareness for the task of planning and scheduling. This study investigates the effects of the number and type of constraints on human performance, and the differences in scheduling metrics between scheduling and rescheduling tasks. Subject testing is currently in progress, and we aim to collect data from approximately thirty subjects. We will summarize the methods used and impacts of the number and type of scheduling constraints. We will also present lessons learned as well as relevant results of the study.

crew scheduling↗

DOE OSTI2022

ITeM: Independent temporal motifs to summarize and compare temporal networks

We report networks are a fundamental and flexible way of representing various complex systems. Many domains such as communication, citation, procurement, biology, social media, and transportation can be modeled as a set of entities and their relationships. Temporal networks are a specialization of general networks where every relationship occurs at a discrete time. The temporal evolution of such networks is as important to understand as the structure of the entities and relationships. We present the Independent Temporal Motif (ITeM) to characterize temporal graphs from different domains. ITeMs can be used to model the structure and the evolution of the graph. In contrast to existing work, ITeMs are edge-disjoint directed motifs that measure the temporal evolution of ordered edges within the motif. For a given temporal graph, we produce a feature vector of ITeM frequencies and the time it takes to form the ITeM instances. We apply this distribution to measure the similarity of temporal graphs. We show that ITeM has higher accuracy than other motif frequency-based approaches. We define various ITeM-based metrics that reveal salient properties of a temporal network. We also present importance sampling as a method to efficiently estimate the ITeM counts. We present a distributed implementation of the ITeM discovery algorithm using Apache Spark and GraphFrame. We evaluate our approach on both synthetic and real temporal networks.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2026

Avoidance of disruptions on KSTAR due to vertical displacement events via novel real-time stability assessment

Disruption avoidance via the DECAF approach has been achieved on KSTAR using a novel real-time vertical stability assessment and a multiactuator feedback control strategy. The development of disruption avoidance strategies with reactor-relevant reliability is an urgent activity, enabling future fusion power plants. The stability metric employed is based on a new formulation of a vertical force gradient balance metric evaluated across the poloidal cross section of the plasma, with parameters tuned using historical data. Evaluation of this metric on a validation set of 400 recent KSTAR shots indicates >82% of Vertical displacement events can be avoided via feedback control. Essential to its calculation is the two-dimensional toroidal current density distribution in the plasma. Measurement of this profile faster than fully-converged equilibrium reconstructions can deliver is found to improve forecaster performance and is achieved with a surrogate model that takes as input magnetic diagnostic measurements and outputs the current profile on a basis comprising the top principal components of historical current profiles (from past equilibrium reconstructions). This method solves the non-uniqueness problem typically faced when reconstructing current profiles directly from diagnostics, while improving computational time and accuracy. On average, profiles produced by this model reach coefficients of determination of >0.99 with respect to those from equilibrium reconstructions. The avoidance actuators employed include poloidal field coils and an electron cyclotron current drive system. The multiactuator approach, as shown in this first demonstration, allows disruption avoidance while minimizing impact to operational performance. This ability, along with its flexibility and speed, makes this new approach an attractive option for avoiding these types of disruptions in reactors.

Tobin, Matthew [Columbia Univ., New York, NY (Unit↗

DOE OSTI2024

Predicting Instability and the Effect of Wind Loading on Single-Axis Trackers

As PV modules continue to trend toward larger, thinner, and more flexible forms they grow more susceptible to damage from dynamic wind loading. As a result, understanding the impact of wind on PV systems, particularly when mounted on solar-tracking hardware, and identifying robust, stable array layouts and stow strategies is becoming increasingly important for the PV community. In our ongoing DuraMAT project, we are developing an open-source software package, PVade (PV aerodynamic design engineering), to simulate the cascading fluid-structure interaction that occurs within solar-tracking arrays to enable researchers to test hardware, layout, and tracker control changes, leading to enhanced stability and a reduction in wind-driven damage. We will give an overview of the PVade software, highlighting recent user-interface and algorithm developments, before presenting the latest outcomes from our ongoing validation campaign in which we analyze and compare with experimental data obtained from a DuraMAT 1 project. From there, we will present simulated results from a larger, multi-row array and highlight relationships between varying tracker angles and stability as measured by different experimentally validated metrics.

fluid structure interaction↗

DOE OSTI2023

Fast and efficient identification of anomalous galaxy spectra with neural density estimation

ABSTRACT Current large-scale astrophysical experiments produce unprecedented amounts of rich and diverse data. This creates a growing need for fast and flexible automated data inspection methods. Deep learning algorithms can capture and pick up subtle variations in rich data sets and are fast to apply once trained. Here, we study the applicability of an unsupervised and probabilistic deep learning framework, the probabilistic auto-encoder, to the detection of peculiar objects in galaxy spectra from the SDSS survey. Different to supervised algorithms, this algorithm is not trained to detect a specific feature or type of anomaly, instead it learns the complex and diverse distribution of galaxy spectra from training data and identifies outliers with respect to the learned distribution. We find that the algorithm assigns consistently lower probabilities (higher anomaly score) to spectra that exhibit unusual features. For example, the majority of outliers among quiescent galaxies are E+A galaxies, whose spectra combine features from old and young stellar population. Other identified outliers include LINERs, supernovae, and overlapping objects. Conditional modelling further allows us to incorporate additional information. Namely, we evaluate the probability of an object being anomalous given a certain spectral class, but other information such as metrics of data quality or estimated redshift could be incorporated as well. We make our code publicly available.

Böhm, Vanessa↗