Person:

Mavris, Dimitri N.

Permanent Link

https://hdl.handle.net/1853/71501

Associated Organization(s)

Organizational Unit

Daniel Guggenheim School of Aerospace Engineering

Organizational Unit

Aerospace Systems Design Laboratory (ASDL)

Full item page

Publication Search Results

Now showing 1 - 10 of 86

A Method for Modeling System-Driven Uncertainty during Probabilistic Part Life Analyses

(Georgia Institute of Technology, 2004-11) Wallace, Jon Michael ; Volovoi, Vitali V. ; Mavris, Dimitri N.

Probabilistic part life analyses of turbine components have typically been conducted in an ad-hoc fashion with respect to the influence of the system. While this approach greatly simplifes the analysis, signifcant errors and misleading results are possible. However, directly modeling the system analyses in a fully probabilistic and integrated fashion can be prohibitive in terms of the infrastructure required. An effcient approach to characterizing and quantifying the system-driven input for probabilistic part life assessments is proposed. The approach is demonstrated for a turbine blade operating in a medium size commercial transport jet. The results of this demonstra- tion illustrate how the component parameters and failure mechanisms can be qualitatively identifed and the complex probabilistic input modeled as driven by the system behavior.
A Methodology for Assessing Business Models of Future Air Transportation in the Atlanta Regional Transportation System

(Georgia Institute of Technology, 2004-09) Lim, Choon Giap ; Lewe, Jung-Ho ; DeLaurentis, Daniel A. ; Mavris, Dimitri N.

A methodology employing physics-based and economics-based tools in conjunction with probabilistic treatment is developed to study Personal Air Vehicle business model. In the context of the paper, a business model is a mathematical representation of a service provider business operation. Vehicle concepts and hypothesized metrics such as mobility freedom and 'value of time'are embedded in the methodology. Market behavior of the complex transportation environment is captured as part of the equation through Agent-based Modeling and Monte Carlo Simulation techniques. This simulation platform for the transportation environment facilitates the case study of the Atlanta Regional Transportation System. The establishment of this model lays the foundation for creating a robust and adaptive design methodology that allows experts in fields other than aerospace engineering to contribute their expertise towards the realization of this very diverse and dynamic future air transportation system.
Use of Probability of Success as an Independent Variable for Decision-Making

(Georgia Institute of Technology, 2004-09) Frits, Andrew P. ; Mavris, Dimitri N.

Early phases of design are characterized by risk and uncertainty. Appropriate accounting for this uncertainty is an important requirement for any designer. This work suggests collapsing risk and uncertainty into a single metric called the probability of success, which accounts for the probability of a given design simultaneously meeting all of the design requirements. Optimal or lowest cost designs can then be found for various levels of probability of success. These designs can be compared to each other, creating a trade-off between the cost of a design and its risk. These risk versus cost figures can be generated before a decision-maker commits to the design. Thus, the decision-maker will have all the information regarding the cost and risk of potential designs before making any design decisions. The decision-maker can thus treat the probability of success, or risk, as an independent variable, choosing the level of risk that he or she finds acceptable based upon the cost of the system, with the corresponding
Inclusion of Tactical Considerations for System-of-Systems Optimization of Torpedoes

(Georgia Institute of Technology, 2004-08) Frits, Andrew P. ; Weston, Neil R. ; Mavris, Dimitri N.

In the current torpedo design process, torpedoes are often designed independently from the tactics with which they are employed. This serial design process, of first developing tactics, then designing the torpedo, then re-developing tactics leads to torpedo designs that are sub-optimal when viewed from the greater system-of-systems perspective. This paper looks at the effects that tactics have on the design of torpedoes. It proposes a new paradigm, of simultaneous tactics development and torpedo design, and looks at the implications of various tactics on the optimal design of torpedo systems.
Benefits of Non-Dimensionalization in Creation of Designs of Experiments for Sizing Torpedo Systems

(Georgia Institute of Technology, 2004-08) Frits, Andrew P. ; Reynolds, Kristen ; Weston, Neil R. ; Mavris, Dimitri N.

Non-dimensionalization is useful at many stages in the conceptual design process. One area of usefulness is in the creation and execution of Design of Experiments. A Design of Experiments that is run with dimensional quantities can often have a large number of failed or infeasible cases or require frustratingly small ranges on the design variables in order to execute cleanly. However, with the use of non-dimensional parameters in the Design of Experiments, the dimensional values being used in the analysis tool automatically scale themselves so that appropriate magnitudes of each parameter are always being used. This automatic scaling greatly increases the stability of Design of Experiments when non-dimensional parameters are used, limiting the number of failed cases. This paper explores potential non-dimensional parameters for use in the conceptual design of torpedo systems. The paper shows that traditional non-dimensional parameters used in propulsor design, such as advance ratio and thrust coefficient, also work well as torpedo design parameters. A short example is given where the performance of a Design of Experiments for a torpedo system is improved via the use of non-dimensional parameters.
System Reliability Assessment using Covariate Theory

(Georgia Institute of Technology, 2004-01) Wallace, Jon Michael ; Mavris, Dimitri N. ; Schrage, Daniel P.

A method is demonstrated that utilizes covariate theory to generate a multi-response component failure distribution as a function of pertinent operational parameters. Where traditional covariate theory uses actual measured life data, a modified approach is used herein to utilize life values generated by computer simulation models. The result is a simulation-based component life distribution function in terms of time and covariate parameters for each failure response. A multivariate joint probability covariate model is proposed by combining the covariate marginal failure distributions with the Nataf transformation approach. Evaluation of the joint probability model produced significant improvement in joint probability predictions as compared to the independent series event approach. The proposed methods are executed for a nominal aircraft engine system to demonstrate the assessment of multi-response system reliability driven by a dual mode turbine blade component failure scenario as a function of operational parameters.
Assessment of Uncertainty in Aerospace Propulsion System Design and Simulation

(Georgia Institute of Technology, 2003-12) Mavris, Dimitri N. ; Roth, Bryce Alexander

The subject of uncertainty analysis in complex systems design is a broad and bourgeoning field of study. This paper focuses on only three very specific areas of current propulsion research wherein uncertainty plays a pivotal role in the problem formulation. The first is probabilistic approaches to matching engine cycle models to test data. Engine cycle models must have a high confidence of representing the actual engine performance accurately. These models must be matched in the presence of measurement, manufacturing, and other sources of uncertainty. Moreover, the optimal model match tends to change with time such that the problem is stochastic in nature. Current efforts are focusing on using Bayesian statistics to enable a comprehensive (stochastic) treatment of the problem. The second research area of interest is probabilistic analysis methods for estimation of part life in life-limited gas turbine engines. There are many sources of uncertainty in estimating part life, including material properties, material cleanliness/flaw size, part loads, and usage profile. Moreover, life limited parts are subject to accumulated damage over time, and the damage accumulation rate is a strong function of vehicle mission profile and usage. Current efforts are therefore aimed at linking detailed part analysis (finite element and materials models) with higher-level system and mission-level parameters to enable rapid and accurate analysis with the least possible effort. Finally, the role of uncertainty in engine materials selection and insertion is discussed. The materials development process for critical turbine engine parts is very lengthy and subject to considerable uncertainty with regards to the optimal balance of materials properties required for a given application. This is an area of research that will benefit from the development of materials selection methods designed to yield robust materials applicable to the greatest possible number of engines.
Simulation-Based Parametric Reliability Modeling Using Covariate Theory

(Georgia Institute of Technology, 2003-11) Wallace, Jon Michael ; Mavris, Dimitri N.

Reliability analysis methods used for preliminary safety assessments of complex systems typically assume a predetermined and invariant set of input variable statistical properties. However, during the product development phase of the system and especially during in service operation, the characteristics of the random variables can themselves be subject to variation. Thus, the resulting failure probability distribution can vary greatly from early predictions. The objective of this paper is to explore a technique used to create a general parametric failure probability distribution as a function of key input variables. This technique is constructed around covariate theory which is the basis of the familiar Accelerated Life Testing and Proportional Hazards Modeling approaches. Where these approaches have traditionally been used with physical experiments, they are applied within this study to Monte Carlo simulation data generated using an available component modeling and simulation environment of a gas turbine airfoil limited by a single failure mode. Necessary modifications to the traditional form of the covariate approach are identified for application to controlled Monte Carlo simulation data. Implications to potential safety improvements early on in the product development phase are discussed.
Implementation of a Physics-Based Decision-Making Framework for Evaluation of the Multidisciplinary Aircraft Uncertainty

(Georgia Institute of Technology, 2003-09) Briceño, Simón Ignacio ; Mavris, Dimitri N.

In today's business climate, aerospace companies are more than ever in need of rational methods and techniques that provide insights as to the best strategies which may be pursued for increased profitability and risk mitigation. However, the use of subjective, anecdotal decision-making remains prevalent due to the absence of analytical methods capable of capturing and forecasting future needs. Negotiations between airframe and engine manufacturers could benefit greatly from a structured environment that facilitates efficient, rational, decision-making. Creation of such an environment can be developed through a parametric physics-based, stochastic formulation that uses Response Surface Equations as meta-models to expedite the process. This paper describes the implementation of such an approach in order to demonstrate the types of insights that might be gained as an engine manufacturer tries to forecast the effects of the associated airframe uncertainties (structural, aerodynamic, etc. design changes) on engine related characteristics for the design of a hypothetical regional business jet.
Estimation of Turbofan Engine Performance Model Accuracy and Confidence Bounds

(Georgia Institute of Technology, 2003-09) Roth, Bryce Alexander ; Mavris, Dimitri N. ; Doel, David L.

This paper explores the application of Inference and Bayesian Updating principles as a means to efficiently incorporate probabilistic data into the turbine engine status model matching process. This approach allows efficient estimation of nominal model match parameters from test data and also enables quantification of model accuracy and confidence bounds. The basic concepts are developed in detail and formulated into a status matching approach. This method is then applied to a simple surrogate matching problem using a cantilever beam matching exercise to illustrate the methods in a clear and easy-to-understand way. Typical results are presented and are directly analogous to status matching of a gas turbine engine cycle model.