Jorge Bardina

Training data requirement for a neural network to predict aerodynamic coefficients

Basic aerodynamic coefficients are modeled as functions of angle of attack, speed brake deflection angle, Mach number, and side slip angle. Most of the aerodynamic parameters can be well-fitted using polynomial functions. We previously demonstrated that a neural network is a fast, reliable way of predicting aerodynamic coefficients. We encountered few under fitted and/or over fitted results during prediction. The training data for the neural network are derived from wind tunnel test measurements and numerical simulations. The basic questions that arise are: how many training data points are required to produce an efficient neural network prediction, and which type of transfer functions should be used between the input-hidden layer and hidden-output layer. In this paper, a comparative study of the efficiency of neural network prediction based on different transfer functions and training dataset sizes is presented. The results of the neural network prediction reflect the sensitivity of the architecture, transfer functions, and training dataset size.

Distributed web-based expert system for launch operations

The simulation and modeling of launch operations is based on a representation of the organization of the operations suitable to experiment of the physical, procedural, software, hardware and psychological aspects of space flight operations. The virtual test bed consists of a weather expert system to advice on the effect of weather to the launch operations. It also simulates toxic gas dispersion model, and the risk impact on human health. Since all modeling and simulation is based on the Internet, it could reduce the cost of operations of launch and range safety by conducting extensive research before a particular launch. Each model has an independent decision making module to derive the best decision for launch

Simulation modeling of space missions using the high level architecture

This paper discusses an environment being developed to model a mission of the Space Launch System (SLS) and the Multipurpose Crew Vehicle (MPCV) being launched from Kennedy Space Center (KSC) to the International Space Station (ISS). Several models representing different phases of the mission such as the ground operations processes, engineered systems, and range components such as failure tree, blast, gas dispersion, and debris modeling are explained. These models are built using different simulation paradigms such as continuous, system dynamics, discrete-event, and agent-based simulation modeling. The High Level Architecture (HLA) is the backbone of this distributed simulation. The different design decisions and the information fusion scheme of this unique environment are explained in detail for decision-making. This can also help in the development of exploration missions beyond the International Space Station.

Micro-Flying Robotics in Space Missions

The Columbia Accident Investigation Board issued a major recommendation to NASA. Prior to return to flight, NASA should develop and implement a comprehensive inspection plan to determine the structural integrity of all Reinforced Carbon-Carbon (RCC) system components. This inspection plan should take advantage of advanced non-destructive inspection technology. This paper describes a non-intrusive technology with a micro-flying robot to continuously monitor inside a space vehicle for any stress related fissures, cracks and foreign material embedded in walls, tubes etc.

Web-based toxic gas dispersion model for shuttle launch operations

During the launch of the Space Shuttle vehicle, the burning of liquid hydrogen fuel with liquid oxygen at extreme high temperatures inside the three space shuttle main engines, and the burning of the solid propellant mixture of ammonium perchlorate oxidizer, aluminum fuel, iron oxide catalyst, polymer binder, and epoxy curing agent in the two solid rocket boosters result in the formation of a large cloud of hot, buoyant toxic exhaust gases near the ground level which subsequently rises and entrains into ambient air until the temperature and density of the cloud reaches an approximate equilibrium with ambient conditions. In this paper, toxic gas dispersion for various gases are simulated over the web for varying environmental conditions which is provided by rawinsonde data. The model simulates chemical concentration at ground level up to 10 miles (1 KM grids) in downrange up to an hour after launch. The ambient concentration of the gas dispersion and the deposition of toxic particles are used as inputs for a human health risk assessment model. The advantage of the present model is the accessibility and dissemination of model results to other NASA centers over the web. The model can be remotely operated and various scenarios can be analyzed.

Human health risk assessment simulations in a distributed environment for Shuttle launch

During the launch of a rocket under prevailing weather conditions, commanders at Cape Canaveral Air Force station evaluate the possibility of whether wind blown toxic emissions might reach civilian and military personnel in the near by area. In our model, we focused mainly on Hydrogen chloride (HCL), Nitrogen oxides (NOx) and Nitric acid (HNO3), which are non-carcinogenic chemicals as per United States Environmental Protection Agency (USEPA) classification. We have used the hazard quotient model to estimate the number of people at risk. It is based on the number of people with exposure above a reference exposure level that is unlikely to cause adverse health effects. The risk to the exposed population is calculated by multiplying the individual risk and the number in exposed population. The risk values are compared against the acceptable risk values and GO or NO-go situation is decided based on risk values for the Shuttle launch. The entire model is simulated over the web and different scenarios can be generated which allows management to choose an optimum decision.

Prediction of aerodynamic coefficients for wind tunnel data using a genetic algorithm optimized neural network

A fast, reliable way of predicting aerodynamic coefficients is produced using a neural network optimized by a genetic algorithm. Basic aerodynamic coefficients (e.g. lift, drag, pitching moment) are modelled as functions of angle of attack and Mach number. The neural network is first trained on a relatively rich set of data from wind tunnel tests or numerical simulations to learn an overall model. Most of the aerodynamic parameters can be well-fitted using polynomial functions. A new set of data, which can be relatively sparse, is then supplied to the network to produce a new model consistent with the previous model and the new data. Because the new model interpolates realistically between the sparse test data points, it is suitable for use in piloted simulations. The genetic algorithm is used to choose a neural network architecture to give best results, avoiding over- and under-fitting of the test data.

Ground and Range Operations for a Heavy-Lift Vehicle: Preliminary Thoughts

ABSTRACT 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. INTRODUCTION Simulation modeling is one of the most important areas for exploration. The NASA Office of Chief Technologist (OCT) [6] has stated that “Simulation focuses on the design, planning, and operational challenges of NASA’s distributed, long-lived mission systems.” We agree that a model represents the features of a system from a dimensional or multidimensional viewpoint. On the other hand, simulation is the execution of a model which has the possibility (if the model is able to capture appropriately the features up to certain level of fidelity) to represent its behavior. In addition, OCT states [6] that “Through the combination of the two, we can make better decisions and communicate those decisions early enough in the design and development process that changes are easy and quick, as opposed to during production when they are extremely costly and practically impossible.” There are several principles with complex systems such as “emergent behavior” which can be discovered with simulation. Simulation modeling has some interesting benefits and features: 1. It helps to understand complex problems from different viewpoints: We have to understand the system and its structure, goals and objectives. We have to view complex problems from different dimensions. It is usual a multi-disciplinary effort. 2. Basic theory can be combined with experiments and expert opinions: A simulation model can fused data and information from first-principle models, empirical, and expert opinions. 3. Ontologies will be very important to increase agility and interoperability: This will support the development of knowledge discovery mechanisms and the potential automation of generating simulation models. 4. A map of models is important: A map of “models” (analytical and empirical) from the different points in the life-cycle of a system is an important endeavor in order to determine gaps.

Business Intelligence Modeling in Launch Operations

The future of business intelligence in space exploration will focus on the intelligent system-of-systems real-time enterprise. In present business intelligence, a number of technologies that are most relevant to space exploration are experiencing the greatest change. Emerging patterns of set of processes rather than organizational units leading to end-to-end automation is becoming a major objective of enterprise information technology. The cost element is a leading factor of future exploration systems. This technology project is to advance an integrated Planning and Management Simulation Model for evaluation of risks, costs, and reliability of launch systems from Earth to Orbit for Space Exploration. The approach builds on research done in the NASA ARC/KSC developed Virtual Test Bed (VTB) to integrate architectural, operations process, and mission simulations for the purpose of evaluating enterprise level strategies to reduce cost, improve systems operability, and reduce mission risks. The objectives are to understand the interdependency of architecture and process on recurring launch cost of operations, provide management a tool for assessing systems safety and dependability versus cost, and leverage lessons learned and empirical models from Shuttle and International Space Station to validate models applied to Exploration. The systems-of-systems concept is built to balance the conflicting objectives of safety, reliability, and process strategy in order to achieve long term sustainability. A planning and analysis test bed is needed for evaluation of enterprise level options and strategies for transit and launch systems as well as surface and orbital systems. This environment can also support agency simulation based acquisition process objectives. The technology development approach is based on the collaborative effort set forth in the VTBs integrating operations, process models, systems and environment models, and cost models as a comprehensive disciplined enterprise analysis environment. Significant emphasis is being placed on adapting root cause from existing Shuttle operations to exploration. Technical challenges include cost model validation, integration of parametric models with discrete event process and systems simulations, and large-scale simulation integration. The enterprise architecture is required for coherent integration of systems models. It will also require a plan for evolution over the life of the program. The proposed technology will produce long-term benefits in support of the NASA objectives for simulation based acquisition, will improve the ability to assess architectural options verses safety/risk for future exploration systems, and will facilitate incorporation of operability as a systems design consideration, reducing overall life cycle cost for future systems.

NASA System-Level Design, Analysis and Simulation Tools Research on NextGen

ABSTRACT A review of the research accomplished in 2009 in the System-Level Design, Analysis and Simulation Tools (SLDAST) of the NASAs Airspace Systems Program is presented. This research thrust focuses on the integrated system-level assessment of component level innovations, concepts and technologies of the Next Generation Air Traffic System (NextGen) under research in the ASP program to enable the development of revolutionary improvements and modernization of the National Airspace System. The review includes the accomplishments on baseline research and the advancements on design studies and system-level assessment, including the cluster analysis as an annualization standard of the air traffic in the U.S. National Airspace, and the ACES-Air MIDAS integration for human-in-the-loop analyzes within the NAS air traffic simulation. INTRODUCTION As the air traffic demand increases to levels that are not sustainable by the air traffic management system, NASA has been mandated by Congress to support the development and implementation of the Next Generation Air Transportation System (NextGen) of the U.S.A. National Air Space (NAS). This national effort is led by the Joint Planning and Development Office (JPDO) “ to retain U.S. leadership in global aviation, expand capacity, ensure safety, protect the environment, ensure our national defense, and secure the nations” [1]. Although the estimated air traffic demand is subject to forecasts under present economic uncertainties, the benefits of NextGen are important as the latest 2011 report of the Federal Aviation Administration entitled “FAAs NextGen Implementation Plan” [2] states: “NextGen is a comprehensive overhaul of our National Airspace System to make air travel more convenient and dependable, while ensuring your flight is as safe, secure and hassle free as possible. … Our latest estimates, which are sensitive to traffic and fuel price forecasts, indicate that by 2018, NextGen will reduce total delays (in flight and on the ground) by about 35 percent compared with what would happen if we did nothing. That delay reduction will provide, through 2018,