Search Results

Data mining is defined as the discovery of useful, possibly unexpected, patterns and relationships in data using statistical and non-statistical techniques in order to develop schemes for decision and policy making. Data mining can be used to discover the sources and causes of problems in complex systems. In addition, data mining can support simulation strategies by finding the different constants and parameters to be used in the development of simulation models. This paper introduces a framework for data mining and its application to complex problems. To further explain some of the concepts outlined in this paper, the potential application to the NASA Shuttle Reinforced Carbon-Carbon structures and genetic programming is used as an illustration.

The aim of this work is to apply an innovative adaptive ℒ1 techniques to control flutter phenomena affecting highly flexible wings and to evaluate the efficiency of this control algorithm and architecture by performing the following tasks: i) adaptation and analysis of an existing simplified nonlinear plunging/pitching 2D aeroelastic model accounting for structural nonlinearities and a quasi-steady aerodynamics capable of describing flutter and post-flutter limit cycle oscillations, ii) implement the ℒ1 adaptive control on the developed aeroelastic system to perform initial control testing and evaluate the sensitivity to system parameters, and iii) perform model validation and calibration by comparing the performance of the proposed control strategy with an adaptive back-stepping algorithm. The effectiveness and robustness of the ℒ1 adaptive control in flutter and post-flutter suppression is demonstrated.

Although a multitude of anomaly detection and fault isolation programs can be found in the research, there does not appear to be any work published on architectural templates that could take advantage of multiple programs and integrate them into the desired systems. More specifically, there is an absence of a methodological process for generating anomaly detection and fault isolation designs to either embed within new system concepts, or supplement existing schemes. This paper introduces a new approach based on systems engineering and the System Modeling Language (SysML). Preliminary concepts of the proposed approach are explained. In addition, a case study is also mentioned.

Operator training using a weapon in a real-world environment is risky, expensive, time-consuming, and restricted to the given environment. In addition, governments are under intense scrutiny to provide security, yet they must also strive for efficiency and reduce spending. In other words, they must do more with less. Virtual simulation, is usually employed to solve these limitations. As the operator is trained to maximize weapon effectiveness, the effectiveness-focused training can be completed in an economical manner. Unfortunately, the training is completed in limited scenarios without objective levels of training factors for an individual operator to optimize the weapon effectiveness. Thus, the training will not be effective. For overcoming this problem, we suggest a methodology on guiding effectiveness-focused training of the weapon operator through usability assessments, big data, and Virtual and Constructive (VC) simulations.

Connected and Autonomous Vehicles (CAV) rely on information obtained from sensors and communication to make decisions. In a Cooperative Vehicle Safety (CVS) system, information from remote vehicles (RV) is available at the host vehicle (HV) through the wireless network. Safety applications such as crash warning algorithms use this information to estimate the RV and HV states. However, this information is uncertain and sparse due to communication losses, limitations of communication protocols in high congestion scenarios, and perception errors caused by sensor limitations. In this paper we present a novel approach to improve the robustness of the CVS systems, by proposing an architecture that divide application and information/perception subsystems and a novel prediction method based on non-parametric Bayesian inference to mitigate the detrimental effect of data loss on the performance of safety applications.

This paper presents the concept of a dual latent heat sink for thermal management of pulse heat generating electronic systems. The focus of this work is to verify the effectiveness of the concept during charging through experimentation. Accordingly, custom components were built and a prototype version of the heat sink was fabricated. Experiments were performed to investigate the implementation feasibility and heat transfer performance. It is shown that this heat sink is practicable and helps in arresting the system temperature rise during charging (period of pulse heat load).

This paper discusses the ground and range operations for a Shuttle derived Heavy-Lift Vehicle being launched from the Kennedy Space Center on the Eastern range. Comparisons will be made between the Shuttle and a heavy lift configuration (SLS-ETF MPCV - April 2011) by contrasting their subsystems. The analysis will also describe a simulation configuration with the potential to be utilized for heavy lift vehicle processing/range simulation modeling and the development of decision-making systems utilized by the range. In addition, a simple simulation model is used to provide the required critical thinking foundations for this preliminary analysis.

Complex aerospace engineering systems require innovative methods for performance monitoring and anomaly detection. The interface of a real-time data stream to a system for analysis, pattern recognition, and anomaly detection can require distributed system architectures and sophisticated custom programming. This paper presents a case study of a simplified interface between Programmable Logic Controller (PLC) real-time data output, signal processing, cloud computing, and tablet systems. The discussed approach consists of three parts: First, the connectivity of real-time data from PLCs to the signal processing algorithms, using standard communication technologies. Second, the interface of legacy routines, such as NASA's Inductive Monitoring System (IMS), with a hybrid signal processing system. Third, the connectivity and interaction of the signal processing system with a wireless and distributed tablet, (iPhone/iPad) in a hybrid system configuration using cloud computing.

An innovative nonlinear simulation approach for high power density synchronous generator systems is developed and implemented. Due to high power density, the generator operates in nonlinear region of the magnetic circuit. Magnetic Finite Element Analysis (FEA) makes nonlinear simulation possible. Neural network technique provides nonlinear functions for system level simulation. Dynamic voltage equation provides excellent mathematical model for system level simulations. Voltage, current, and flux linkage quantities are applied in Direct-Quadrature (DQ) rotating frame. The simulated system includes main machine, exciter, rectifier bridge, bang-bang control, and PI control circuitry, forming a closed loop system. Each part is modeled and then integrated into the system model.

With the development and licensing of inland, state-owned spaceports, and the ongoing development of several new reusable launch vehicles (RLV), the space launch industry is undergoing a significant transformation. As a result, there is a need to reevaluate current launch risk analysis methodologies and practices, which so far have revolved around the conservative casualty expectation analysis developed in the 1950s. Furthermore, an important aspect of launch risk analysis which gives rise to its complexity is its multidisciplinary nature. In analyzing such risk, the physics of and interactions between the varieties of hazards produced by launch vehicles breakups must be captured, modeled and, their effects analyzed. In this paper we discuss how a well-designed multidisciplinary modeling and analysis platform could be a significant step toward reducing the complexity just mentioned.

This paper reviews the literature which identifies the causes, consequences and cures for engine knock as it affects high performance engines. The physical events of normal and abnormal combustion are described. The observed variations in combustion phenomenon are explained through chemical kinetics. A mathematical model of combustion which can predict knock in an engine cylinder is summarized. Several mechanisms of knock induced damage are outlined. Design and operating considerations which affect an engine's propensity to knock are discussed. Terms that have become associated with combustion in general and the knocking phenomenon in particular are collected and examined

A power electronics module was equipped with an evaporative spray cooling nozzle assembly that served to remove waste heat from the silicon devices. The spray cooling nozzle assembly took the place of the standard heat sink, which uses single phase convection. The purpose of this work was to test the ability of spray cooling to enable higher power density in power electronics with high temperature coolant, and to be an effective and lightweight system level solution to the thermal management needs of aerospace vehicles. The spray cooling work done here was with 95 °C water, and this data is compared to 100 °C water/ propylene glycol spray cooling data from a previous paper so as to compare the spray cooling performance of a single component liquid to that of a binary liquid such as WPG. The module used during this work was a COTS module manufactured by Semikron, Inc., with a maximum DC power input of 180 kW (450 VDC and 400 A).

The application of Statistical Process Control and Design of Experiment methods in the research laboratory can lead to significant gains in the Powertrain development process. Empirical methods such as Design of Experiments, Regression, and Neural Network techniques can be applied to help researchers gain better understanding of the cause and effect relationships of emission, alternative fuel source, performance, fuel economy, and engine management system - calibration studies. The use of these empirical modeling techniques along with model based Genetic Algorithm, Gradient, or Constraint based solution search methods will help identify the “process settings” that improve fuel economy, improve performance, and reduce pollutants. Since empirical methods are fundamentally based on the acquired test data, it is vitally important that the laboratory measurements are repeatable, consistent, and void of sources of variance that have a significant effect on the acquired test data.

Calibrating the engine control unit to satisfy pollutant and performance objectives can be a challenging task. Due to the large number of variables and their interactive complexities, many firms apply design of experiment methods and modeling techniques to the acquired test data. This establishes a “black box” or “gray box” simulation model that predicts power and emissions as a function of the engine parameters. An offline optimization procedure on the fitted model(s) will identify the engine control strategy that best satisfies pollutant and performance objectives. A review of the literature reveals that the General Linear Modeling method and Neural Network modeling architectures are widely used in the development of “black box” or “gray box” simulation models. While Neural Network methods are “assumption free”, the General Linear Model method is limited to those problems in which the errors, ε, are normally distributed and have constant variance, σ2.

Effective thermal management and removal of the waste heat generated at diode arrays is critical to the development of high-power solid-state lasers. Thermal design must be considered in the packaging of these arrays. Two different packages with heat dissipation through spray cooling are evaluated experimentally and numerically. Their overall performance is compared with other packaging configurations using different heat removal approaches. A novel packaging design is proposed that can fulfill the requirements of low thermal resistance, temperature uniformity among emitters in the diode array, low coolant flow rate, simplicity and low assembly cost. The effect of temperature uniformity on the pumping efficiency for gain media is examined for our novel packaging design. The thermal stress induced by temperature variation within an emitter is also considered.

The purpose of this study was to increase IC engine performance by “tuning” the exhaust system to different induction system pressures using an empirical based modeling approach. The two distinct induction pressures are atmospheric and 13-15 mmHg above atmospheric. The above atmospheric induction pressure occurs when the race car is in the lead; the atmospheric pressure occurs with the race car is following the lead or “in the draft.” Since it is ideal to achieve optimum performance for both induction pressures, the problem was formulated and optimized using an empirical Multiple Response Surface Method (MRSM) approach. MRSM is a process that “extracts” multiple objective performance information through carefully controlled experiments and data modeling techniques. An analysis of the experimental data will identify the ideal header length configuration that maximizes performance for both induction pressure extremes.

The future of human exploration in the solar system is contingent on the ability to exploit resources in-situ to produce mission consumables. Specifically, it has become clear that the success of a manned mission to Mars will likely depend on fuel components created on the Martian surface. While several architectures for an unmanned fuel production surface facility on Mars exist in theory, a simulation of the performance and operation of these architectures has not been created. In this paper, the framework describing a simulation of one such architecture is defined. Within this architecture, each component of the base is implemented as a state machine, with the ability to communicate with other base elements as well as a supervisor. An environment supervisor is also created which governs low level aspects of the simulation such as movement and resource distribution, in addition to higher-level aspects such as location selection with respect to operations specific behavior.

Currently there are no community noise computer programs commonly in use in the United States dedicated to the modeling of interrupted flow. Constant speed programs (such as STAMINA 2.0) have been used with modified input to predict noise levels at intersections, but they cannot directly simulate traffic signal operation, actual deceleration and acceleration of vehicles, or queues of vehicles at signals. Noise prediction procedures for intersections can be improved by simulating actual intersection movements. The American Automobile Manufacturers Association has produced a model that not only models continuous flow (Constant Speed Traffic program - CST), but also allows modeling of interrupted flow (Variable Speed Traffic program - VST). This model has been updated to improve user friendliness and accuracy and is discussed in this paper.

Optimizing IC engine performance currently requires an exhaustive experimental search to determine the combination of internal components that maximizes torque or power. An alternate and more structured approach using Response Surface Methods will lead the experimenter to the optimum combination with the least number of trials. Using simulation software to evaluate IC engine configurations, this method improved the estimated power from 439 to 516 KW. Results of the study indicate that Response Surface Methods are a viable and robust method of converging to an IC engine configuration which achieves optimum performance.

Preliminary investigations of nonlinear modeling of aircraft synchronous generators using neural networks are presented. Aircraft synchronous generators with high power density tend operate at current-levels proportional to the magnetic saturation region of the machine's material. The nonlinear model accounts for magnetic saturation of the generator, which causes the winding flux linkages and inductances to vary as a function of current. Finite element method software is used to perform a parametric sweep of direct, quadrature, and field currents to extract the respective flux linkages. This data is used to train a neural network which yields current as a function of flux linkage. The neural network is implemented in a Simulink synchronous generator model and simulation results are compared with a previously developed linear model. Results show that the nonlinear neural network model can more accurately describe the responsiveness and performance of the synchronous generator.