Search NASA⌕ Search

SEARCH · Search NASA

Results for “Optimal control design”

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 127 records · Page 7

The Microgravity Isolation Mount: A Linearized State-Space Model a la Newton and Kane

Vibration acceleration levels on large space platforms exceed the requirements of many space experiments. The Microgravity Vibration Isolation Mount (MIM) was built by the Canadian Space Agency to attenuate these disturbances to acceptable levels, and has been operational on the Russian Space Station Mir since May 1996. It has demonstrated good isolation performance and has supported several materials science experiments. The MIM uses Lorentz (voice-coil) magnetic actuators to levitate and isolate payloads at the individual experiment/sub-experiment (versus rack) level. Payload acceleration, relative position, and relative orientation (Euler-parameter) measurements are fed to a state-space controller. The controller, in turn, determines the actuator currents needed for effective experiment isolation. This paper presents the development of an algebraic, state-space model of the MIM, in a form suitable for optimal controller design. The equations are first derived using Newton's Second Law directly; then a second derivation (i.e., validation) of the same equations is provided, using Kane's approach.

Hampton, R. David↗

Synthesis Methods for Robust Passification and Control

The research effort under this cooperative agreement has been essentially the continuation of the work from previous grants. The ongoing work has primarily focused on developing passivity-based control techniques for Linear Time-Invariant (LTI) systems. During this period, there has been a significant progress made in the area of passivity-based control of LTI systems and some preliminary results have also been obtained for nonlinear systems, as well. The prior work has addressed optimal control design for inherently passive as well as non- passive linear systems. For exploiting the robustness characteristics of passivity-based controllers the passification methodology was developed for LTI systems that are not inherently passive. Various methods of passification were first proposed in and further developed. The robustness of passification was addressed for multi-input multi-output (MIMO) systems for certain classes of uncertainties using frequency-domain methods. For MIMO systems, a state-space approach using Linear Matrix Inequality (LMI)-based formulation was presented, for passification of non-passive LTI systems. An LMI-based robust passification technique was presented for systems with redundant actuators and sensors. The redundancy in actuators and sensors was used effectively for robust passification using the LMI formulation. The passification was designed to be robust to an interval-type uncertainties in system parameters. The passification techniques were used to design a robust controller for Benchmark Active Control Technology wing under parametric uncertainties. The results on passive nonlinear systems, however, are very limited to date. Our recent work in this area was presented, wherein some stability results were obtained for passive nonlinear systems that are affine in control.

Kelkar, Atul G.↗

Glovebox Integrated Microgravity Isolation Technology (g-LIMIT): A Linearized State-Space Model

Vibration acceleration levels on large space platforms exceed the requirements of many space experiments. The Glovebox Integrated Microgravity Isolation Technology (g-LIMIT) is being built by the NASA Marshall Space Flight Center to attenuate these disturbances to acceptable levels. G-LIMIT uses Lorentz (voice-coil) magnetic actuators to levitate and isolate payloads at the individual experiment/sub-experiment (versus rack) level. Payload acceleration, relative position, and relative orientation measurements are fed to a state-space controller. The controller, in turn, determines the actuator Currents needed for effective experiment isolation. This paper presents the development of an algebraic, state-space model of g-LIMIT, in a form suitable for optimal controller design. The equations are first derived using Newton's Second Law directly, then simplified to a linear form for the purpose of controller design.

Hampton, R. David↗

Flight Control Research at NASA Ames

An overview of the flight control work in the Intelligent Systems Division at NASA Ames is presented. The highlight focuses on efforts surrounding performance-adaptive aeroelastic wing shaping for aircraft with flexible wings. Topics covered include aeroservoelastic modeling capabilities, online drag-optimizing control designs, gust and maneuver load alleviation techniques, and related wind tunnel demonstrations.

Hashemi, Kelley↗

Optimal Regulator Algorithms for the Control of Linear Systems

Control-theory design package, called Optimal Regulator Algorithms for Control of Linear Systems (ORACLS), aids in design of controllers and optimal filters for systems modeled by linear, time invariant differential and difference equations. ORACLS is particularly attractive rigorous tool for dealing with multi-input and multi-output dynamic systems in both continuous and discrete forms.

Armstrong, E. S.↗

Blade planform design optimization to enhance turbine wake control

Abstract This study considers optimizing the planform of wind turbine blades to ultimately enhance wind plant controls, namely, wake steering strategies. Adjoint‐enabled unsteady actuator line simulations are carried out to obtain gradients for optimization of several different performance objectives with respect to blade chord length at 10 locations along blade span. We demonstrate different blade design optimizations that can maximize time‐averaged lateral wake deflection, entrainment of kinetic energy, or total power of multiple turbines. Our optimized designs can produce a 4+% increase in wake deflection, a 4× increase in vertical kinetic energy entrainment, or a 3.6% increase in power when compared with the baseline case. While lateral wake deflection is only modestly sensitive to chord changes, we find that increasing the outboard chord length can dramatically increase kinetic energy entrainment, resulting in faster wake recovery and gains in net power. While this work develops only a few case studies emphasizing relative performance improvements and general trends, these results show the promise of a framework that combines mid‐fidelity computation with adjoint‐based optimization for control and design problems.

WIND ENERGY↗

Applied & Computational MathematicsChallenges for the Design and Control of Dynamic Energy Systems

The Energy Independence and Security Act of 2007 (EISA) was passed with the goal 'to move the United States toward greater energy independence and security.' Energy security and independence cannot be achieved unless the United States addresses the issue of energy consumption in the building sector and significantly reduces energy consumption in buildings. Commercial and residential buildings account for approximately 40% of the U.S. energy consumption and emit 50% of CO{sub 2} emissions in the U.S. which is more than twice the total energy consumption of the entire U.S. automobile and light truck fleet. A 50%-80% improvement in building energy efficiency in both new construction and in retrofitting existing buildings could significantly reduce U.S. energy consumption and mitigate climate change. Reaching these aggressive building efficiency goals will not happen without significant Federal investments in areas of computational and mathematical sciences. Applied and computational mathematics are required to enable the development of algorithms and tools to design, control and optimize energy efficient buildings. The challenge has been issued by the U.S. Secretary of Energy, Dr. Steven Chu (emphasis added): 'We need to do more transformational research at DOE including computer design tools for commercial and residential buildings that enable reductions in energy consumption of up to 80 percent with investments that will pay for themselves in less than 10 years.' On July 8-9, 2010 a team of technical experts from industry, government and academia were assembled in Arlington, Virginia to identify the challenges associated with developing and deploying newcomputational methodologies and tools thatwill address building energy efficiency. These experts concluded that investments in fundamental applied and computational mathematics will be required to build enabling technology that can be used to realize the target of 80% reductions in energy consumption. In addition the finding was that there are tools and technologies that can be assembled and deployed in the short term - the next 3-5 years - that can be used to significantly reduce the cost and time effective delivery of moderate energy savings in the U.S. building stock. Simulation tools, which are a core strength of current DOE computational research programs, provide only a part of the answer by providing a basis for simulation enabled design. New investments will be required within a broad dynamics and control research agenda which must focus on dynamics, control, optimization and simulation of multi-scale energy systems during design and operation. U.S. investments in high performance and high productivity computing (HP2C) should be leveraged and coupled with advances in dynamics and control to impact both the existing building stock through retrofits and also new construction. The essential R&D areas requiring investment are: (1) Characterizing the Dynamics of Multi-scale Energy Systems; (2) Control and Optimization Methodologies of Multi-scale Energy Systems Under Uncertainty; and (3) Multiscale Modeling and Simulation Enabled Design and Operation. The concept of using design and control specific computational tools is a new idea for the building industry. The potential payoffs in terms of accelerated design cycle times, performance optimization and optimal supervisory control to obtain and maintain energy savings are huge. Recent advances in computational power, computer science, and mathematical algorithms offer the foundations to address the control problems presented by the complex dynamics of whole building systems. The key areas for focus and associated metrics with targets for establishing competitiveness in energy efficient building design and operation are: (1) Scalability - Current methodology and tools can provide design guidance for very low energy buildings in weeks to months; what is needed is hours to days. A 50X improvement is needed. (2) Installation and commissioning - Current methodology and tools can target a three month window for commissioning of building subsystems; what is needed is one week. A 10X improvement is needed. (3) Quality - Current design tools can achieve 30% accuracy; what is needed to make design decisions is 5% with quantification of uncertainty. A 5X improvement is needed. These challenges cannot be overcome by raw computational power alone and require the development of new algorithms. Here algorithms mean much more than simulating the building physics but need to be inclusive of a much better understanding of the building and the control systems associated with the building and to capture the entire set of dynamics. The algorithmsmust represent computationally new mathematical approaches to modeling, simulation, optimization and control of large multi-scale dynamic systems and bringing these elements to bear on industry in simulation enabled design approaches.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

An framework for robust flight control design using constrained optimization

An analytical framework is described for the design of feedback control systems to meet specified performance criteria in the presence of structured and unstructured uncertainty. Attention is focused upon the linear time invariant, single-input, single-output problem for the purposes of exposition. The framework provides for control of the degree of the stabilizing compensator or controller.

Palazoglu, A.↗

Computationally Efficient Control Co-Design Optimization Framework with Mixed-Fidelity Fluid and Structure Analysis: OpenTurbineCoDe (OTCD) - ARPA-E ATLANTIS Phase I Project (Final Technical Report)

The goal of the Department of Energy (DOE) Advanced Research Projects Agency-Energy (ARPA-E) Aerodynamic Turbines Lighter and Afloat with Nautical Technologies and Integrated Servo-control (ATLANTIS) Program is to develop new technologies for floating offshore wind turbines, or wind farms, using the discipline of control co-design (CCD.) In this context, our goal is to develop a computationally efficient optimization framework for design of floating offshore wind turbines. Our specific aim is to utilize high-fidelity structural, aerodynamic, aero-structural tools, and to derive control-oriented reduced- or low-order models directly from the high-fidelity tools. We proposed a mixed-fidelity modeling approach which means that we are also using low- and mid-fidelity tools when necessary. This research is conducted by a multidisciplinary team consisting of Rutgers University, University of Michigan, Brigham Young University, and the National Renewable Energy Laboratory (NREL). The computational framework, called OpenTurbineCoDe, is designed to integrate, where possible, traditional structural, aerodynamic, aeroelastic models (e.g., OpenFAST) and advanced control algorithms with higher fidelity simulation tools including Reynolds-averaged Navier–Stokes (RANS) solvers, and three-dimensional structural finite element solvers. All the high-fidelity tools used in this research provide numerically exact gradients to facilitate both efficient optimization and local linearization for control implementation.

17 WIND ENERGY↗

Data-Driven Optimization for the Design and Control of Large-Scale Systems (LDRD Final Report)

Engineering decisions are often formulated as optimization problems such as the optimal design or control of physical systems. In these applications, the resulting optimization problems are constrained by large-scale simulations involving systems of partial differential equations (PDEs), ordinary differential equations (ODEs), and differential algebraic equations (DAEs). In addition, critical components of these systems are fraught with uncertainty, including unverifiable modeling assumptions, unknown boundary and initial conditions, and uncertain coefficients. Typically, these components are estimated using noisy and incomplete data from a variety of sources (e.g., physical experiments). The lack of knowledge of the true underlying probabilistic characterization of model inputs motivates the need for optimal solutions that are robust to this uncertainty. In this report, we introduce a framework for handling "distributional" uncertainties in the context of simulation-based optimization. This includes a novel measure discretization technique that will lead to an adaptive optimization algorithm tailored to exploit the structures inherent to simulation- based optimization.

42 ENGINEERING↗

Control of a flexible space shuttle vehicle

The application of optimal control technology to the design of a controller for the space shuttle descent was studied using weighting matrices adjusted through iteration to reduce response peaks, which are defined as mean magnitude plus sigma values. The synthesis techniques, and optimal controller design are discussed along with simplified controller design. Quantitative characteristics of the optimal controller, effects of variation of sample rate, and simplified controller data are included.

Harvey, C. A.↗

A Mathematical Formulation of the SCOLE Control Problem. Part 2: Optimal Compensator Design

The study initiated in Part 1 of this report is concluded and optimal feedback control (compensator) design for stability augmentation is considered, following the mathematical formulation developed in Part 1. Co-located (rate) sensors and (force and moment) actuators are assumed, and allowing for both sensor and actuator noise, stabilization is formulated as a stochastic regulator problem. Specializing the general theory developed by the author, a complete, closed form solution (believed to be new with this report) is obtained, taking advantage of the fact that the inherent structural damping is light. In particular, it is possible to solve in closed form the associated infinite-dimensional steady-state Riccati equations. The SCOLE model involves associated partial differential equations in a single space variable, but the compensator design theory developed is far more general since it is given in the abstract wave equation formulation. The results thus hold for any multibody system so long as the basic model is linear.

Balakrishnan, A. V.↗

Optimal member damper controller design for large space structures

Consideration is given to the selection of velocity feedback gains for individual dampers for the members of a structurally controlled large flexible space structure. The problem is formulated as an optimal output feedback regulator problem, and necessary conditions are derived for minimizing a quadratic performance function. The diagonal nature of the gain matrix is taken into account, along with knowledge of noise covariances. It is pointed out that the method presented offers a systematic approach to the design of a class of controllers for enhancing structural damping, which have significant potential if used in conjunction with a reduced-order optimal controller for rigid-body modes and selected structural modes.

Joshi, S. M.↗

Practical optimal flight control system design for helicopter aircraft. Volume 2: Software user's guide

A method by which modern and classical control theory techniques may be integrated in a synergistic fashion and used in the design of practical flight control systems is presented. A general procedure is developed, and several illustrative examples are included. Emphasis is placed not only on the synthesis of the design, but on the assessment of the results as well. The first step is to establish the differences, distinguishing characteristics and connections between the modern and classical control theory approaches. Ultimately, this uncovers a relationship between bandwidth goals familiar in classical control and cost function weights in the equivalent optimal system. In order to obtain a practical optimal solution, it is also necessary to formulate the problem very carefully, and each choice of state, measurement and output variable must be judiciously considered. Once design goals are established and problem formulation completed, the control system is synthesized in a straightforward manner. Three steps are involved: filter-observer solution, regulator solution, and the combination of those two into the controller. Assessment of the controller permits and examination and expansion of the synthesis results.

Riedel, S. A.↗

Optimal Control within the Context of Multidisciplinary Design, Analysis, and Optimization

Multidisciplinary design, analysis and optimization involves modeling the interactions of complex systems across a variety of disciplines. The optimization of such systems can be a computationally expensive exercise with multiple levels of nested nonlinear solvers running under an optimizer.The application of optimal control in project development often involves performing trajectory optimization for fixed vehicle designs or parametric sweeps across some key vehicle properties.This information is then relayed to the subsystem design teams who update their designs and relay some bulk characteristics back to the trajectory optimization procedure.This iteration is then repeated until the design closes.However, with increasing interest in more tightly coupled systems, such as electric and hybrid-electric aircraft propulsion and boundary layer ingestion, this process is prone to ignore subtle coupling between vehicle subsystem designs and vehicle operation on a given mission.Integrating trajectory optimization into a tightly coupled multidisciplinary design procedure can be computationally prohibitive, depending on the complexity of the subsystem analyses and the optimal control technique applied.To address these issues a new optimal control software tool, Dymos, has been developed.Dymos is built upon NASA's OpenMDAO software and can leverage its capabilities to efficiently compute gradients for the optimization and optimize complex models in parallel on distributed memory systems.This report provides some explanation into the numerical methods employed in Dymos and provides several use cases that demonstrate its performance on traditional optimal control problems and improvements ino techniques have been used extensively in recent decades to solve a variety of optimal control problems, typically in the form of aerospace vehicle trajectory optimization.

pseudospectral↗

Controls-structures integrated design optimization with shape variations

The shape design variables have been introduced into the set of design variables of the Controls-Structure Integrated (CSI) Design of space-structures. The importance of the shape variations in improving the design (obtained with only control and sizing variables) has been aptly illustrated. Two different types of design variables that describe the shape variations of the structure have been introduced. In the first case, the nodal coordinates have been considered as design variables. This has the inherent difficulty of having too many design variables. This not only is time consuming but also memory intensive and may not yield a manufacturable shape to the structure. The second approach has been introduced to overcome this difficulty. The structure is allowed to vary in a particular pre defined pattern. The coefficients of these patterns are considered as the shape design variables. The eigenvalue and eigenvector sensitivity equations with respect to these coefficient design variables have been developed and are used to approximate the eigenvalues and eigenvectors in a perturbed design.

Koganti, Gopichand↗