Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “REINFORCEMENT”

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 1,567 records · Page 87

Reinforcement Learning in a Nonstationary Environment: The El Farol Problem

This paper examines the performance of simple learning rules in a complex adaptive system based on a coordination problem modeled on the El Farol problem. The key features of the El Farol problem are that it typically involves a medium number of agents and that agents' pay-off functions have a discontinuous response to increased congestion. First we consider a single adaptive agent facing a stationary environment. We demonstrate that the simple learning rules proposed by Roth and Er'ev can be extremely sensitive to small changes in the initial conditions and that events early in a simulation can affect the performance of the rule over a relatively long time horizon. In contrast, a reinforcement learning rule based on standard practice in the computer science literature converges rapidly and robustly. The situation is reversed when multiple adaptive agents interact: the RE algorithms often converge rapidly to a stable average aggregate attendance despite the slow and erratic behavior of individual learners, while the CS based learners frequently over-attend in the early and intermediate terms. The symmetric mixed strategy equilibria is unstable: all three learning rules ultimately tend towards pure strategies or stabilize in the medium term at non-equilibrium probabilities of attendance. The brittleness of the algorithms in different contexts emphasize the importance of thorough and thoughtful examination of simulation-based results.

Bell, Ann Maria↗

Orbiter Reinforced Carbon-Carbon Advanced Sealant Systems: Screening Tests

Oxidation protection for the Orbiter reinforced carbon-carbon (RCC consists of three components: silicon carbide coating, tetraethyl orthosilicate (TEOS) impregnated into the carbon substrate and a silicon based surface sealant (designated Type A). The Orbiter Type A sealant is being consumed each mission, which results in increased carbon-carbon substrate mass loss, which adversely impacts the mission life of the RCC components. In addition, the sealant loss in combination with launch pad contamination (salt deposit and zinc oxide) results in RCC pinholes. A sealant refurbishment schedule to maintain mission life and minimize affects of pin hole formation has been implemented in the Orbiter maintenance schedule. The objective of this investigation is to develop an advanced sealant system for the RCC that extends the refurbishment schedule by reducing sealant loss/pin hole formation and that can be applied to existing Orbiter RCC components. This paper presents the results of arc jet screening tests conducted on several sealants that are being considered for application to the Orbiter RCC.

Curry, Donald M.↗

In-plane and Interlaminar Shear Strength of a Unidirectional Hi-nicalon Fiber-reinforced Celsian Matrix Composite

In-plane and interlaminar shear strength of a unidirectional SiC fiber-reinforced (BaSr)Al2Si2O8 celsian composite were measured by the double-notch shear test method between room temperature and 1200 C. The interlaminar shear strength was lower than the in-plane shear strength at all temperatures. Stress analysis, using finite element modeling, indicated that shear stress concentration was not responsible for the observed difference in strength. Instead, the difference in layer architecture and thus, the favorable alignment of fiber-rich layers with the shear plane in the interlaminar specimens appears to be the reason for the low strength of this composite. A rapid decrease in strength was observed with temperature due to softening of the glassy phase in the material.

Uenal, O.↗

Optimal Reward Functions in Distributed Reinforcement Learning

We consider the design of multi-agent systems so as to optimize an overall world utility function when (1) those systems lack centralized communication and control, and (2) each agents runs a distinct Reinforcement Learning (RL) algorithm. A crucial issue in such design problems is to initialize/update each agent's private utility function, so as to induce best possible world utility. Traditional 'team game' solutions to this problem sidestep this issue and simply assign to each agent the world utility as its private utility function. In previous work we used the 'Collective Intelligence' framework to derive a better choice of private utility functions, one that results in world utility performance up to orders of magnitude superior to that ensuing from use of the team game utility. In this paper we extend these results. We derive the general class of private utility functions that both are easy for the individual agents to learn and that, if learned well, result in high world utility. We demonstrate experimentally that using these new utility functions can result in significantly improved performance over that of our previously proposed utility, over and above that previous utility's superiority to the conventional team game utility.

Wolpert, David H.↗

Analysis of Graphite Reinforced Cementitious Composites

This paper describes analytical methods that can be used to determine the deflections and stresses in highly compliant graphite-reinforced cementitious composites. It is demonstrated that the standard transform section fails to provide accurate results when the elastic modulus ratio exceeds 20. So an alternate approach is formulated by using the rule of mixtures to determine a set of effective material properties for the composite. Tensile tests are conducted on composite samples to verify this approach; and, when the effective material properties are used to characterize the deflections of composite beams subject to pure bending, an excellent agreement is obtained. Laminated composite plate theory is also investigated as a means for analyzing even more complex composites, consisting of multiple graphite layers oriented in different directions. In this case, composite beams are analyzed by incorporating material properties established from tensile tests. Finite element modeling is used to verity the results and, considering the complexity of the samples, a very good agreement is obtained.

Vaughan, Robert E.↗

Reinforcement Learning in Distributed Domains: Beyond Team Games

Distributed search algorithms are crucial in dealing with large optimization problems, particularly when a centralized approach is not only impractical but infeasible. Many machine learning concepts have been applied to search algorithms in order to improve their effectiveness. In this article we present an algorithm that blends Reinforcement Learning (RL) and hill climbing directly, by using the RL signal to guide the exploration step of a hill climbing algorithm. We apply this algorithm to the domain of a constellations of communication satellites where the goal is to minimize the loss of importance weighted data. We introduce the concept of 'ghost' traffic, where correctly setting this traffic induces the satellites to act to optimize the world utility. Our results indicated that the bi-utility search introduced in this paper outperforms both traditional hill climbing algorithms and distributed RL approaches such as team games.

Wolpert, David H.↗

Constitutive Modeling of Nanotube-Reinforced Polymer Composite Systems

In this study, a technique has been proposed for developing constitutive models for polymer composite systems reinforced with single-walled carbon nanotubes (SWNT). Since the polymer molecules are on the same size scale as the nanotubes, the interaction at the polymer/nanotube interface is highly dependent on the local molecular structure and bonding. At these small length scales, the lattice structures of the nanotube and polymer chains cannot be considered continuous, and the bulk mechanical properties of the SWNT/polymer composites can no longer be determined through traditional micromechanical approaches that are formulated using continuum mechanics. It is proposed herein that the nanotube, the local polymer near the nanotube, and the nanotube/polymer interface can be modeled as an effective continuum fiber using an equivalent-continuum modeling method. The effective fiber retains the local molecular structure and bonding information and serves as a means for incorporating micromechanical analyses for the prediction of bulk mechanical properties of SWNT/polymer composites with various nanotube sizes and orientations. As an example, the proposed approach is used for the constitutive modeling of two SWNT/polyethylene composite systems, one with continuous and aligned SWNT and the other with discontinuous and randomly aligned nanotubes.

Odegard, Gregory M.↗

Material Properties for Fiber-Reinforced Silica Aerogels

Ceramic fiber-reinforced silica aerogels are novel materials for high performance insulation, including thermal protection materials. Experimental data are presented for the thermal and mechanical properties, showing the trends exhibited over a range of fiber loadings and silica aerogel densities. Test results are compared to that of unreinforced bulk aerogels.

White, Susan↗

Mechanical Properties of Continuous Fiber Reinforced Zirconium Diboride Matrix Composites

Continuous fiber reinforced zirconium diboride matrix composites, SCS-9a-(RBSiCZrB2)matrix, are being developed for leading edge, rocket nozzle and turbine engine applications. Recently, the composite materials have been characterized for tensile properties to 1250 C, the highest temperature tested. The tensile properties are fiber dominated as the matrix is microcracked on fabrication, but favorable failure characteristic are observed. Compression and shear mechanical testing results will be reported if completed. The effects of fiber volume fraction and matrix density on mechanical properties will be discussed. The target applications of the materials will be discussed. Specific testing being performed towards qualification for these applications will be included.

Stuffle, Kevin↗

Fuzzy Q-Learning for Generalization of Reinforcement Learning

Fuzzy Q-Learning, introduced earlier by the author, is an extension of Q-Learning into fuzzy environments. GARIC is a methodology for fuzzy reinforcement learning. In this paper, we introduce GARIC-Q, a new method for doing incremental Dynamic Programming using a society of intelligent agents which are controlled at the top level by Fuzzy Q-Learning and at the local level, each agent learns and operates based on GARIC. GARIC-Q improves the speed and applicability of Fuzzy Q-Learning through generalization of input space by using fuzzy rules and bridges the gap between Q-Learning and rule based intelligent systems.

Berenji, Hamid R.↗

Constitutive Modeling of Nanotube-Reinforced Polymer Composites

In this study, a technique is presented for developing constitutive models for polymer composite systems reinforced with single-walled carbon nanotubes (SWNT). Because the polymer molecules are on the same size scale as the nanotubes, the interaction at the polymer/nanotube interface is highly dependent on the local molecular structure and bonding. At these small length scales, the lattice structures of the nanotube and polymer chains cannot be considered continuous, and the bulk mechanical properties can no longer be determined through traditional micromechanical approaches that are formulated by using continuum mechanics. It is proposed herein that the nanotube, the local polymer near the nanotube, and the nanotube/polymer interface can be modeled as an effective continuum fiber using an equivalent-continuum modeling method. The effective fiber serves as a means for incorporating micromechanical analyses for the prediction of bulk mechanical properties of SWNT/polymer composites with various nanotube lengths, concentrations, and orientations. As an example, the proposed approach is used for the constitutive modeling of two SWNT/polyimide composite systems.

Odegard, G. M.↗

Collaborating Fuzzy Reinforcement Learning Agents

Earlier, we introduced GARIC-Q, a new method for doing incremental Dynamic Programming using a society of intelligent agents which are controlled at the top level by Fuzzy Relearning and at the local level, each agent learns and operates based on ANTARCTIC, a technique for fuzzy reinforcement learning. In this paper, we show that it is possible for these agents to compete in order to affect the selected control policy but at the same time, they can collaborate while investigating the state space. In this model, the evaluator or the critic learns by observing all the agents behaviors but the control policy changes only based on the behavior of the winning agent also known as the super agent.

Berenji, Hamid R.↗

Nanotube Reinforcement of Adhesively Bonded Joints

Over the past five years there has been much excitement about the development of nanotubes and nanofibers and the potential that these materials may offer in enhancing electrical and mechanical properties of systems. The purpose of this paper is to present research into improving the mechanical performance of polymers by using nanofibers as a reinforcement to make high performance composite materials. This paper will present theoretical predictions of the composite modulus and then present the actual performance of the composite. Fabrication details will be given along with photos of the microstructure. The matrix material is polymethylmethacrylate (PMMA) and the nanofibers are vapor-grown carbon nanofibers produced by Pyrograph Products, Inc.

Johnson, W. S.↗

Rapid Prototyping of Continuous Fiber Reinforced Ceramic Matrix Composites

For ceramics to be used as structural components in high temperature applications, their fracture toughness is improved by embedding continuous ceramic fibers. Ceramic matrix composite (CMC) materials allow increasing the overall operating temperature, raising the temperature safety margins, avoiding the need for cooling, and improving the damping capacity, while reducing the weight at the same time. They also need to be reliable and available in large quantities as well. In this paper, an innovative rapid prototyping technique to fabricate continuous fiber reinforced ceramic matrix composites is described. The process is simple, robust and will be widely applicable to a number of high temperature material systems. This technique was originally developed at the University of Delaware Center for Composite Materials (UD-CCM) for rapid fabrication of polymer matrix composites by a technique called automated tow placement or ATP. The results of mechanical properties and microstructural characterization are presented, together with examples of complex shapes and parts. It is believed that the process will be able to create complex shaped parts at an order of magnitude lower cost than current CVI and PIP processes.

Vaidyanathan, R.↗

Synthesis and Characterization of Carbon Nanotubes for Reinforced and Functional Applications

Many efforts have been engaged recently in synthesizing single-walled and multi-walled carbon nanotubes due to their superior mechanical, electrical and thermal properties, which could be used to enhance numerous applications such as electronics, sensors and composite strength. This presentation will show the synthesizing process of carbon nanotubes by thermal chemical vapor deposition and the characterization results by using electron microscopy and optical spectroscopy. Carbon nanotubes were synthesized on various substances. The conditions of fabricating single-walled or multi-walled carbon nanotubes depend strongly on temperatures and hydrocarbon concentrations but weakly on pressures. The size, growth modes and orientations of carbon nanotube will be illustrated. The advantages and limitations of several potential applications including sensor, heat pipe, field emission, radiation shielding, and reinforcements for composites by using carbon nanotubes will be discussed.

Zhu, Shen↗

Constitutive Modeling of Nanotube-Reinforced Polymer Composites

In this study, a technique is presented for developing constitutive models for polymer composite systems reinforced with single-walled carbon nanotubes (SWNT). Because the polymer molecules are on the same size scale as the nanotubes, the interaction at the polymer/nanotube interface is highly dependent on the local molecular structure and bonding. At these small length scales, the lattice structures of the nanotube and polymer chains cannot be considered continuous, and the bulk mechanical properties can no longer be determined through traditional micromechanical approaches that are formulated by using continuum mechanics. It is proposed herein that the nanotube, the local polymer near the nanotube, and the nanotube/polymer interface can be modeled as an effective continuum fiber using an equivalent-continuum modeling method. The effective fiber serves as a means for incorporating micromechanical analyses for the prediction of bulk mechanical properties of SWNT/polymer composites with various nanotube shapes, sizes, concentrations, and orientations. As an example, the proposed approach is used for the constitutive modeling of two SWNT/LaRC-SI (with a PmPV interface) composite systems, one with aligned SWNTs and the other with three-dimensionally randomly oriented SWNTs. The Young's modulus and shear modulus have been calculated for the two systems for various nanotube lengths and volume fractions.

Odegard, G. M.↗

Carbon Fiber Reinforced Ceramic Composites for Propulsion Applications

This report presents a critical review of the processing techniques for fabricating continuous fiber-reinforced CMCs for possible applications at elevated temperatures. Some of the issues affecting durability of the composite materials such as fiber coatings and cracking of the matrix because of shrinkage in PIP-process are also examined. An assessment of the potential inexpensive processes is also provided. Finally three potential routes of manufacturing C/SiC composites using a technology that NC A&T developed for carbon/carbon composites are outlined. Challenges that will be encountered are also listed.

Shivakumar, Kunigal↗

Long-term Isothermal Aging Effects on Weight Loss, Compression Properties, and Dimensions of T650-35 Fabric-reinforced PMR-15 Composites-data

A cooperative program was conducted with the General Electric Aircraft Engines plant in Evendale, Ohio, to study the effects of long-term isothermal aging at elevated temperatures on compression and thermal durability properties of T650 35 fabric-reinforced PMR 15 composites. This degradation study was conducted over an approximate time period of 3 1/2 yr. The aging temperatures were 204, 260, 288, 316, and 343 C. Specimens of different dimensions were evaluated. Specimens with ratios of the cut edge to total surface area of 0.03 to 0.89 were fabricated and aged. The aged and unaged specimens were tested in compression as specified in Test Method for Compressive Properties of Rigid Plastics (ASTM D695M). Thickness changes, degraded surface layer growth, weight loss, and failure modes were monitored and recorded. All property changes were thickness dependent.

Bowles, Kenneth J.↗