Search NASA⌕ Search

SEARCH · Search NASA

Results for “identification problem”

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.

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

At least 1,657 records · Page 92

Cognitive engineering models in space systems

NASA space systems, including mission operations on the ground and in space, are complex, dynamic, predominantly automated systems in which the human operator is a supervisory controller. Models of cognitive functions in complex systems are needed to describe human performance and form the theoretical basis of operator workstation design, including displays, controls, and decision aids. Currently, there several candidate modeling methodologies. They include the Rasmussen abstraction/aggregation hierarchy and decision ladder, the goal-means network, the problem behavior graph, and the operator function model. The research conducted under the sponsorship of this grant focuses on the extension of the theoretical structure of the operator function model and its application to NASA Johnson mission operations and space station applications. The initial portion of this research consists of two parts. The first is a series of technical exchanges between NASA Johnson and Georgia Tech researchers. The purpose is to identify candidate applications for the current operator function model; prospects include mission operations and the Data Management System Testbed. The second portion will address extensions of the operator function model to tailor it to the specific needs of Johnson applications. At this point, we have accomplished two things. During a series of conversations with JSC researchers, we have defined the technical goal of the research supported by this grant to be the structural definition of the operator function model and its computer implementation, OFMspert. Both the OFM and OFMspert have matured to the point that they require infrastructure to facilitate use by researchers not involved in the evolution of the tools. The second accomplishment this year was the identification of the Payload Deployment and Retrieval System (PDRS) as a candidate system for the case study. In conjunction with government and contractor personnel in the Human-Computer Interaction Lab, the PDRS was identified as the most accessible system for the demonstration. Pursuant to this a PDRS simulation was obtained from the HCIL and an initial knowledge engineering effort was conducted to understand the operator's tasks in the PDRS application. The preliminary results of the knowledge engineering effort and an initial formulation of an operator function model (OFM) are contained in the appendices.

Mitchell, Christine M.↗

CRITICA: coding region identification tool invoking comparative analysis

Gene recognition is essential to understanding existing and future DNA sequence data. CRITICA (Coding Region Identification Tool Invoking Comparative Analysis) is a suite of programs for identifying likely protein-coding sequences in DNA by combining comparative analysis of DNA sequences with more common noncomparative methods. In the comparative component of the analysis, regions of DNA are aligned with related sequences from the DNA databases; if the translation of the aligned sequences has greater amino acid identity than expected for the observed percentage nucleotide identity, this is interpreted as evidence for coding. CRITICA also incorporates noncomparative information derived from the relative frequencies of hexanucleotides in coding frames versus other contexts (i.e., dicodon bias). The dicodon usage information is derived by iterative analysis of the data, such that CRITICA is not dependent on the existence or accuracy of coding sequence annotations in the databases. This independence makes the method particularly well suited for the analysis of novel genomes. CRITICA was tested by analyzing the available Salmonella typhimurium DNA sequences. Its predictions were compared with the DNA sequence annotations and with the predictions of GenMark. CRITICA proved to be more accurate than GenMark, and moreover, many of its predictions that would seem to be errors instead reflect problems in the sequence databases. The source code of CRITICA is freely available by anonymous FTP (rdp.life.uiuc.edu in/pub/critica) and on the World Wide Web (http:/(/)rdpwww.life.uiuc.edu).

Non-NASA Center↗

Liquid Cooling/Warming Garment

The NASA liquid cooling/ventilating garment (LCVG) currently in use was developed over 40 years ago. With the commencement of a greater number of extra-vehicular activity (EVA) procedures with the construction of the International Space Station, problems of astronaut comfort, as well as the reduction of the consumption of energy, became more salient. A shortened liquid cooling/warming garment (SLCWG) has been developed based on physiological principles comparing the efficacy of heat transfer of different body zones; the capability of blood to deliver heat; individual muscle and fat body composition as a basis for individual thermal profiles to customize the zonal sections of the garment; and the development of shunts to minimize or redirect the cooling/warming loop for different environmental conditions, physical activity levels, and emergency situations. The SLCWG has been designed and completed, based on extensive testing in rest, exercise, and antiorthostatic conditions. It is more energy efficient than the LCVG currently used by NASA. The total length of tubing in the SLCWG is approximately 35 percent less and the weight decreased by 20 percent compared to the LCVG. The novel features of the innovation are: 1. The efficiency of the SLCWG to maintain thermal status under extreme changes in body surface temperatures while using significantly less tubing than the LCVG. 2. The construction of the garment based on physiological principles of heat transfer. 3. The identification of the body areas that are most efficient in heat transfer. 4. The inclusion of a hood as part of the garment. 5. The lesser consumption of energy.

Koscheyev, Victor S.↗

Scintillation properties of Bi{sub 4}Ge{sub 3}O{sub 12} down to 3 K under {gamma} rays

Bismuth germanate (BGO) has been widely used as a room-temperature scintillator in many applications for decades. Interest in it has recently increased as a low-temperature scintillator to be used in bolometers for rare-event detection. We present our time-resolved-scintillation studies of BGO down to 3 K under {gamma}-ray excitation. Our multiple-photon-counting-coincidence-based setup allows clear identification of {gamma}-line energies at least as low as 122 keV down to base temperature and the measurement of the light yield and decay-time constants as a function of temperature. We also discuss the time structure of the pulses and report a previously unappreciated but significant, very slow component assigned to afterglow. Finally, we demonstrate that nonlinearity of the light yield as a function of energy persists at low temperatures.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Neutralino dark matter in BMSSM effective theory

We study thermal neutralino dark matter in an effective field theory extension of the MSSM, called ''Beyond the MSSM'' (BMSSM) in Dine, Seiberg and Thomas (2007). In this class of effective field theories, the field content of the MSSM is unchanged, but the little hierarchy problem is alleviated by allowing small corrections to the Higgs/higgsino part of the Lagrangian. We perform parameter scans and compute the dark matter relic density. The light higgsino LSP scenario is modified the most; we find new regions of parameter space compared to the standard MSSM. This involves interesting interplay between the WMAP dark matter bounds and the LEP chargino bound. We also find some changes for gaugino LSPs, partly due to annihilation through a Higgs resonance, and partly due to coannihilation with light top squarks in models that are ruled in by the new effective terms.

79 ASTRONOMY AND ASTROPHYSICS↗

Waste disposal technology transfer matching requirement clusters for waste disposal facilities in China

Highlights: Black-Right-Pointing-Pointer We outline the differences of Chinese MSW characteristics from Western MSW. Black-Right-Pointing-Pointer We model the requirements of four clusters of plant owner/operators in China. Black-Right-Pointing-Pointer We examine the best technology fit for these requirements via a matrix. Black-Right-Pointing-Pointer Variance in waste input affects result more than training and costs. Black-Right-Pointing-Pointer For China technology adaptation and localisation could become push, not pull factors. - Abstract: Even though technology transfer has been part of development aid programmes for many decades, it has more often than not failed to come to fruition. One reason is the absence of simple guidelines or decision making tools that help operators or plant owners to decide on the most suitable technology to adopt. Practical suggestions for choosing the most suitable technology to combat a specific problem are hard to get and technology drawbacks are not sufficiently highlighted. Western counterparts in technology transfer or development projects often underestimate or don't sufficiently account for the high investment costs for the imported incineration plant; the differing nature of Chinese MSW; the need for trained manpower; and the need to treat flue gas, bunker leakage water, and ash, all of which contain highly toxic elements. This article sets out requirements for municipal solid waste disposal plant owner/operators in China as well as giving an attribute assessment for the prevalent waste disposal plant types in order to assist individual decision makers in their evaluation process for what plant type might be most suitable in a given situation. There is no 'best' plant for all needs and purposes, and requirement constellations rely on generalisations meaning they cannot be blindly applied, but an alignment of a type of plant to a type of owner or operator can realistically be achieved. To this end, a four-step approach is suggested and a technology matrix is set out to ease the choice of technology to transfer and avoid past errors. The four steps are (1) Identification of plant owner/operator requirement clusters; (2) Determination of different municipal solid waste (MSW) treatment plant attributes; (3) Development of a matrix matching requirement clusters to plant attributes; (4) Application of Quality Function Deployment Method to aid in technology localisation. The technology transfer matrices thus derived show significant performance differences between the various technologies available. It is hoped that the resulting research can build a bridge between technology transfer research and waste disposal research in order to enhance the exchange of more sustainable solutions in future.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Design of analytical failure detection using secondary observers

The problem of designing analytical failure-detection systems (FDS) for sensors and actuators, using observers, is addressed. The use of observers in FDS is related to the examination of the n-dimensional observer error vector which carries the necessary information on possible failures. The problem is that in practical systems, in which only some of the components of the state vector are measured, one has access only to the m-dimensional observer-output error vector, with m or = to n. In order to cope with these cases, a secondary observer is synthesized to reconstruct the entire observer-error vector from the observer output error vector. This approach leads toward the design of highly sensitive and reliable FDS, with the possibility of obtaining a unique fingerprint for every possible failure. In order to keep the observer's (or Kalman filter) false-alarm rate under a certain specified value, it is necessary to have an acceptable matching between the observer (or Kalman filter) models and the system parameters. A previously developed adaptive observer algorithm is used to maintain the desired system-observer model matching, despite initial mismatching or system parameter variations. Conditions for convergence for the adaptive process are obtained, leading to a simple adaptive law (algorithm) with the possibility of an a priori choice of fixed adaptive gains. Simulation results show good tracking performance with small observer output errors, while accurate and fast parameter identification, in both deterministic and stochastic cases, is obtained.

Sisar, M.↗

Open problems in particle astrophysics

The search for an understanding of the nature of the Dark Matter is one of the fundamental problems in science. This contribution outlines the evidence for the existence of Dark Matter (DM), and discusses the possibility that it is formed by new type of Weakly Interacting Massive (m{approx} 100 GeV) elementary Particles (WIMPs). Possible methods to verify this hypothesis include the creation of these particles in high energy accelerators, the observation of the elastic scattering of ordinary nuclei with the DM particles that form the halo of the Milky Way, and the (indirect) observation of the products of the self annihilation of the DM particles in our Galaxy. Possible, controversial hints of the existence of WIMPs will be discussed.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

NIBBS-Search for Fast and Accurate Prediction of Phenotype-Biased Metabolic Systems

Understanding of genotype-phenotype associations is important not only for furthering our knowledge on internal cellular processes, but also essential for providing the foundation necessary for genetic engineering of microorganisms for industrial use (e.g., production of bioenergy or biofuels). However, genotype-phenotype associations alone do not provide enough information to alter an organism’s genome to either suppress or exhibit a phenotype. It is important to look at the phenotype-related genes in the context of the genome-scale network to understand how the genes interact with other genes in the organism. Identification of metabolic subsystems involved in the expression of the phenotype is one way of placing the phenotype-related genes in the context of the entire network. A metabolic system refers to a metabolic network subgraph; nodes are compounds and edges labels are the enzymes that catalyze the reaction. The metabolic subsystem could be part of a single metabolic pathway or span parts of multiple pathways. Arguably, comparative genome-scale metabolic network analysis is a promising strategy to identify these phenotype-related metabolic subsystems. Network Instance-Based Biased Subgraph Search (NIBBS) is a graph-theoretic method for genome-scale metabolic network comparative analysis that can identify metabolic systems that are statistically biased toward phenotypeexpressing organismal networks. We set up experiments with target phenotypes like hydrogen production, TCA expression, and acid-tolerance. We show via extensive literature search that some of the resulting metabolic subsystems are indeed phenotype-related and formulate hypotheses for other systems in terms of their role in phenotype expression. NIBBS is also orders of magnitude faster than MULE, one of the most efficient maximal frequent subgraph mining algorithms that could be adjusted for this problem. Also, the set of phenotype-biased metabolic systems output by NIBBS comes very close to the set of phenotype-biased subgraphs output by an exact maximally-biased subgraph enumeration algorithm ( MBS-Enum ). The code (NIBBS and the module to visualize the identified subsystems) is available at http:// freescience.org/cs/NIBBS.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

On the Nature of the mHz X-Ray QPOs from ULX M82 X-1: Evidence for Timing-Spectral (anti) Correlation

Using all the archival XMM-Newton X-ray (3-10 keV) observations of the ultraluminous X-ray source (ULX) M82 X-1 we searched for a correlation between its variable mHz quasi-periodic oscillation (QPO) frequency and its energy spectral power-law index. These quantities are known to correlate in stellar mass black holes (StMBHs) exhibiting Type-C QPOs (approx 0.2-15 Hz). The detection of such a correlation would strengthen the identification of its mHz QPOs as Type-C and enable a more reliable mass estimate by scaling its QPO frequencies to those of Type-C QPOs in StMBHs of known mass. We resolved the count rates of M82 X-1 and a nearby bright ULX (source 5/X42.3+59) through surface brightness modeling and identify observations in which M82 X-1 was at least as bright as source 5. Using only those observations, we detect QPOs in the frequency range of 36-210 mHz during which the energy spectral power-law index varied from 1.7-2.2. Interestingly, we find evidence for an anti-correlation (Pearsons correlation coefficient = -0.95) between the power-law index and the QPO centroid frequency. While such an anti-correlation is observed in StMBHs at high Type-C QPO frequencies (approx 5-15 Hz), the frequency range over which it holds in StMBHs is significantly smaller (factor of approx 1.5-3) than the QPO range reported here from M82 X-1 (factor of 6). However, it remains possible that contamination from source 5 can bias our result. Joint Chandra/XMM-Newton observations in the future can resolve this problem and confirm the timing-spectral anti-correlation reported here.

Pasham, Dheeraj R.↗

Progressive Aerodynamic Model Identification From Dynamic Water Tunnel Test of the F-16XL Aircraft

Development of a general aerodynamic model that is adequate for predicting the forces and moments in the nonlinear and unsteady portions of the flight envelope has not been accomplished to a satisfactory degree. Predicting aerodynamic response during arbitrary motion of an aircraft over the complete flight envelope requires further development of the mathematical model and the associated methods for ground-based testing in order to allow identification of the model. In this study, a general nonlinear unsteady aerodynamic model is presented, followed by a summary of a linear modeling methodology that includes test and identification methods, and then a progressive series of steps suggesting a roadmap to develop a general nonlinear methodology that defines modeling, testing, and identification methods. Initial steps of the general methodology were applied to static and oscillatory test data to identify rolling-moment coefficient. Static measurements uncovered complicated dependencies of the aerodynamic coefficient on angle of attack and sideslip in the stall region making it difficult to find a simple analytical expression for the measurement data. In order to assess the effect of sideslip on the damping and unsteady terms, oscillatory tests in roll were conducted at different values of an initial offset in sideslip. Candidate runs for analyses were selected where higher order harmonics were required for the model and where in-phase and out-of-phase components varied with frequency. From these results it was found that only data in the angle-of-attack range of 35 degrees to 37.5 degrees met these requirements. From the limited results it was observed that the identified models fit the data well and both the damping-in-roll and the unsteady term gain are decreasing with increasing sideslip and motion amplitude. Limited similarity between parameter values in the nonlinear model and the linear model suggest that identifiability of parameters in both terms may be a problem. However, the proposed methodology can still be used with careful experiment design and carefully selected values of angle of attack, sideslip, amplitude, and frequency of the oscillatory data.

Murphy, Patrick C.↗

The general relativity with conformal units

General Relativity rewritten in conformal units identifies conformal intervals with the real observational distances. This identification gives a base to explain all epochs of the Universe evolution including Ia supernova luminosity long distance-redshift relation by the dominance of the Casimir vacuum energy of all physical fields. A set of arguments is discussed in favor that SNe Ia data in the conformal units can be an evidence of the conformal twistor structure of the space-time as a nonlinear realization of the affine group, like the nonlinear realization of chiral symmetry and phenomenological Lagrangian is evidence of the quark structure of hadrons.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Identification and Evaluation of Integration and Cross Cutting Issues Across HRP Risks

The HRP Integrated Research Plan contains the research plans for the 32 risks requiring research to characterize and mitigate. These risks to human health and performance in spaceflight are identified by evidence and each one focuses on a single aspect of human physiology or performance. They are further categorized by aspects of the spaceflight environment, such as altered gravity or space radiation, that that play a major role in their likelihood and consequence. From its inception the "integrate" in the Research Plan has denoted the integrated nature of risks to human health and performance, the connectedness of physiological systems within the human body regardless of the spaceflight environment, and the integrated response of the human body to the spaceflight environment. Common characteristics of the spaceflight environment include altered gravity, atmospheres and light/dark cycles, space radiation, isolation, noise, and periods of high or low workload. Long term exposure to this unique environment produces a suite of physiological effects such as stress; vision, neurocognitive and anthropometric changes; circadian misalignment; fluid shifts, deconditioning; immune dysregulation; and altered nutritional requirements. Matrix diagraming was used to systematically identify, analyze and rate the many-to-many relationships between environmental characteristics and the suite of physiological effects. It was also to identify patterns in the relationships of common physiological effects to each other. Analyses of patterns or relationships in these diagrams help to identify issues that cut across multiple risks. Cross cutting issues benefit from a multidisciplinary approach that synthesizes concepts or data from two or more disciplines to identify and characterize risk factors or develop countermeasures relevant to multiple risks. They also help to illuminate possible problem areas that may arise when a countermeasure impacts risks other than those which it was developed to mitigate, or identify groupings of physiological changes that are likely to occur that may impact the overall risk posture.

Steinberg, S. L.↗

Stochastic modeling of sunshine number data

In this paper, we will present a unified statistical modeling framework for estimation and forecasting sunshine number (SSN) data. Sunshine number has been proposed earlier to describe sunshine time series in qualitative terms (Theor Appl Climatol 72 (2002) 127-136) and since then, it was shown to be useful not only for theoretical purposes but also for practical considerations, e.g. those related to the development of photovoltaic energy production. Statistical modeling and prediction of SSN as a binary time series has been challenging problem, however. Our statistical model for SSN time series is based on an underlying stochastic process formulation of Markov chain type. We will show how its transition probabilities can be efficiently estimated within logistic regression framework. In fact, our logistic Markovian model can be relatively easily fitted via maximum likelihood approach. This is optimal in many respects and it also enables us to use formalized statistical inference theory to obtain not only the point estimates of transition probabilities and their functions of interest, but also related uncertainties, as well as to test of various hypotheses of practical interest, etc. It is straightforward to deal with non-homogeneous transition probabilities in this framework. Very importantly from both physical and practical points of view, logistic Markov model class allows us to test hypotheses about how SSN dependents on various external covariates (e.g. elevation angle, solar time, etc.) and about details of the dynamic model (order and functional shape of the Markov kernel, etc.). Therefore, using generalized additive model approach (GAM), we can fit and compare models of various complexity which insist on keeping physical interpretation of the statistical model and its parts. After introducing the Markovian model and general approach for identification of its parameters, we will illustrate its use and performance on high resolution SSN data from the Solar Radiation Monitoring Station of the West University of Timisoara.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Aerosol sources in subway environments

Highlights: • PM{sub 2.5} concentrations varied widely among subway platforms (26–86 µg m{sup −3}). • Main PM{sub 2.5} component in metro microenvironment is iron oxide. • Subway sources contribution account for up to 91% of ambient subway PM{sub 2.5}. • From the five subway-specific sources identified, brakes contributed the most. • Elements related with health issues are apportioned mostly by subway sources (> 60%). Millions of people use rail subway public transport around the world, despite the relatively high particulate matter (PM) concentrations in these underground environments, requiring the identification and quantification of the aerosol source contributions to improve the air quality. An extensive aerosol monitoring campaign was carried out in eleven subway stations in the Barcelona metro system, belonging to seven subway lines. PM{sub 2.5} samples were collected during the metro operating hours and chemically analysed to determine major and trace elements, inorganic ions, and total carbon. The chemical compositions of subway components such as brake pads, rail tracks and pantographs were also determined. The mean PM{sub 2.5} concentrations varied widely among stations, ranging from 26 µg m{sup −3} to 86 µg m{sup −3}. Subway PM{sub 2.5} was mainly constituted by Fe{sub 2}O{sub 3} (30–66%), followed by carbonaceous matter (18–37%) for the old stations, while for new stations equipped with Platform Screen Doors (PSD) these percentages go down to 21–44% and 15–30%, respectively. Both the absolute concentrations and the relative abundance of key species differed for each subway station, although with common patterns within a given subway line. This is a result of the different emission chemical profiles in different subway lines (using diverse types of brakes and/or pantographs). The co-emission of different sources poses a problem for their separation by receptor models. Nevertheless, receptor modelling (Positive Matrix Factorization) was applied resulting in ten sources, five of them subway-specific: RailWheel, RailWheel+Brake, Brake{sub A}, Brake{sub B}, Pb. The sum of their contributions accounted for 43–91% of bulk PM{sub 2.5} for the old stations and 21–52% for the stations with PSD. The decrease of the activity during the weekends resulted in a decrease (up to 56%) in the subway-specific sources contribution to the –already lower– bulk PM{sub 2.5} concentrations compared to weekdays. The health-related elements are mainly apportioned (> 60%) by subway sources.

54 ENVIRONMENTAL SCIENCES↗

PING Gamma Ray and Neutron Measurements of a Meter-Sized Carbonaceous Asteroid Analog

Determining the elemental composition of carbonaceous (spectral type C) asteroids is still one of the basic problems when studying these objects. The only main source of elemental composition information for asteroids is from their optical, NIR and IR properties, which include their spectral reflectance characteristics, albedo, polarization, and the comparison of optical spectroscopy with meteorite groups corresponding to asteroids of every spectral type. Unfortunately, these sources reflect observations from widely contrasting spatial scales that presently yield a void in the continuum of microscopic and macroscopic evidence, a lack of in situ measurement confirmation, and require deeper sensing techniques to discern the nature of these asteroids. The Probing In situ with Neutrons and Gamma rays (PING) instrument is ideally suited to address this problem because it can be used to determine the bulk elemental composition, H and C content, the average atomic weight and density of the surface and subsurface layers of C-type asteroids, and can provide measurements used to determine the difference between and distinguish between different types of asteroids. We are currently developing the PING instrument that combines gamma ray and neutron detectors with a 14 Me V pulsed neutron generator to determine the in-situ bulk elemental abundances and geochemistry of C-type asteroids with a spatial resolution of 1 m down to depths of tens of cm to 1 m. One aspect of the current work includes experimentally testing and optimizing PING on a known meter-sized Columbia River basalt C-type asteroid analog sample that has a similar composition and the same neutron response as that of a C-type asteroid. An important part of this effort focuses on utilizing timing measurements to isolate gamma rays produced by neutron inelastic scattering, neutron capture and delayed activation processes. Separating the gamma ray spectra by nuclear processes results in higher precision and sensitivity elemental composition measurements. Using gated data acquisition techniques allows for the unambiguous identification of gamma ray lines from different isotopes and nuclear processes, especially in situations when limited detector resolution results in overlapping gamma ray lines that cannot be individually resolved. In this paper, we will present the PING basalt layering experimental data, taken at the test facility at NASA Goddard Space Flight Center using the time tagged event-by-event data analysis technique, compared to our MCNPX computer simulation results for the C-type asteroid and basalt layering simulant models. Comparison of these data will show the advantages, validity, and measurement sensitivity of PING's nuclear interrogation methods to obtain more precise and sensitive in situ bulk elemental composition and density measurements of the subsurface of asteroids.

Bodnarik, J.↗

Identification and control of plasma vertical position using neural network in Damavand tokamak

In this work, a nonlinear model is introduced to determine the vertical position of the plasma column in Damavand tokamak. Using this model as a simulator, a nonlinear neural network controller has been designed. In the first stage, the electronic drive and sensory circuits of Damavand tokamak are modified. These circuits can control the vertical position of the plasma column inside the vacuum vessel. Since the vertical position of plasma is an unstable parameter, a direct closed loop system identification algorithm is performed. In the second stage, a nonlinear model is identified for plasma vertical position, based on the multilayer perceptron (MLP) neural network (NN) structure. Estimation of simulator parameters has been performed by back-propagation error algorithm using Levenberg-Marquardt gradient descent optimization technique. The model is verified through simulation of the whole closed loop system using both simulator and actual plant in similar conditions. As the final stage, a MLP neural network controller is designed for simulator model. In the last step, online training is performed to tune the controller parameters. Simulation results justify using of the NN controller for the actual plant.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Optimization-based control using input convex neural networks

Input convex neural networks (ICNNs) are a family of deep learning models where the outputs are constructed to be convex functions of the inputs. By parameterizing system models using ICNNs, optimization-based control problems can be solved as convex optimization problems, leading to improved performance and robustness. This work proposes a novel framework where the control objective function and constraints are modelled using ICNNs. A case study of optimization-based control with output constraints is conducted on a process with input multiplicity and nonminimum phase behavior. The simulation results demonstrate improved economic yield compared with normal neural networks. Additionally, the input convexity formulation is compared with simple regularization techniques, and unique benefits such as improved data efficiency and robustness of the proposed formulation are shown. Finally, by explicitly incorporating prior knowledge about convexity, this framework provides a good balance between the universal approximation power of deep learning and computational feasibility required by control.

42 ENGINEERING↗