SEARCH · Search NASA
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.
The startup of systems composed of power electronic systems is becoming more complex. As the systems grow, standard pre-configuration sequencing methods will struggle. This work proposes a linear programming optimization method to couple power electronic systems in a series of startup commands. This formulation is demonstrated in a real-time implementation on hardware.
This paper considers the suboptimal stochastic control of linear discrete-time dynamical systems with unknown or stochastically varying parameters. The suboptimal scheme is based upon the use of the open-loop-feedback-optimal method. The state and parameter estimates are generated by an extended Kalman filter algorithm. Numerical results for first-order systems are presented.
We present robust trajectory optimization techniques using a sweeping gradient method for ordinary differential equations with events (SGM) and linear covariance analysis (LinCov). SGM is a method for computing the gradient of trajectory analyses defined by performance indices over initial value problems with events with respect to static parameters. LinCov is an analytic technique for predicting stochastic behavior of dynamical systems. By combining SGM and LinCov, it is possible use efficient, off-the-shelf, gradient-based optimizers to solve robust optimal trajectory design problems. We describe the individual methods and some details on how they can be combined. Then we apply the combined techniques to a variety of orbital trajectory design problems to demonstrate its use, including minimum fuel transfer and mid-course correction burn scheduling.
Linear systems simultaneous optimal stochastic control and observation strategies, assuming quadratic observation cost
Gust analysis is added to a linearized frequency-domain method in FUN3D, a NASA computational fluid dynamics solver. The method linearizes about a nonlinear static equilibrium condition and is therefore appropriate for problems with small perturbations such as transonic stochastic gust analysis. In addition to the gust analysis, adjoint-based sensitivities of stochastic gust constraints are implemented for multidisciplinary design optimization. The linearized frequency-domain gust model and adjoint-based sensitivities are described and verified. The method is applied to an optimization for mass minimization of the AGARD 445.6 wing subject to a stochastic gust constraint limiting the displacement of the wing tip.
Equations are derived for the angles of general multivariable root loci and linear quadratic optimal root loci, including angles of departure and approach. The generalized eigenvalue problem is used to compute angles of approach. Equations are also derived to find the sensitivity of closed loop eigenvalue and the directional derivatives of closed loop eigenvectors. An equivalence class of quadratic weights that produce the same asymptotic eigenstructure is defined, a canonical element is defined, and an algorithm to find it is given. The behavior of the optimal root locus in the nonasymptotic region is shown to be different for quadratic weights with the same asymptotic properties. An algorithm is presented that can be used to select a feedback gain matrix for the linear state feedback problem which produces a specified asymptotic eigenstructure. Another algorithm is given to compute the asymptotic eigenstructure properties inherent in a given set of quadratic weights. Finally, it is shown that optimal root loci for nongeneric problems can be approximated by generic ones in the nonasymptotic region.
An optimization system is described that consists of the finite-element analysis program known as Engineering Analysis Language (EAL), the CONMIN optimization program, and a piecewise linear analysis. The optimization analysis system (OPTEAL) is used to minimize the mass of a transport wing subject to stress constraints for two static load conditions. The optimization procedure and the optimization analysis system (OPTEAL) are described and the optiomized results with fully stressed design results compared.
Optimal power flow (OPF) problems, which dispatch power targets to controllable generating units across a network, must generally account for non-convex constraints on power flow. Furthermore, adapting those problems so as to make them solvable with convex optimization techniques is an area of much academic and operational interest. In this paper, we present a method for solving OPF as a quadratic program by iteratively refining and re-initializing a linearized model of power flow based on the outputs of an associated nonlinear solver. The linear model on which we demonstrate this method is an adapted version of an approximation designed for use with unbalanced distribution networks. As an important benefit, the model allows for the explicit inclusion of nodal voltage phasor values in both the OPF problem's objective and its constraints, which opens the door to the idea of phasor-based control (PBC) design. We show in simulations on the IEEE 13-node test feeder that our method quickly converges to a set of phasor targets that are sufficiently precise for use in operations at the distribution level.
This paper presents an empirical study of some nonexhaustive approaches to optimizing preferences within the context of constraint-based, mixed-initiative planning for mission operations. This work is motivated by the experience of deploying and operating the MAPGEN (Mixed-initiative Activity Plan GENerator) system for the Mars Exploration Rover Mission. Responsiveness to the user is one of the important requirements for MAPGEN, hence, the additional computation time needed to optimize preferences must be kept within reasonabble bounds. This was the primary motivation for studying non-exhaustive optimization approaches. The specific goals of rhe empirical study are to assess the impact on solution quality of two greedy heuristics used in MAPGEN and to assess the improvement gained by applying a linear programming optimization technique to the final solution.
Motivated by an energy efficient building application, we want to optimize a quadratic cost functional subject to the Boussinesq approximation of the Navier-Stokes equations and to bilateral state and control constraints. Since the computation of such an optimal solution is numerically costly, we design an efficient strategy to compute a sub-optimal (but applicationally acceptable) solution with significantly reduced computational effort. We employ an economic Model Predictive Control (MPC) strategy to obtain a feedback control. The MPC sub-problems are based on a linear-quadratic optimal control problem subjected to mixed control and state constraints and a convection-diffusion equation, reduced with proper orthogonal decomposition. Finally, to solve each sub-problem, we apply a primal-dual active set strategy. The method can be fully parallelized, which enables the solution of large problems with real-world parameters.
In this paper, robust optimization is performed on trajectory correction maneuvers during the lunar lander return phase of an Artemis mission, treating the trajectory from one hour after low lunar orbit departure to arrival in the vicinity of the lunar Gateway as a relative motion problem. To enable rapid stochastic optimization techniques requiring many candidate trajectories, SPICE kernel relative motion as implemented by the Quadratic Interpolated State Transition (QIST) system is used as the underlying dynamics propagation. The optimization is performed with a genetic optimizer using linear covariance (LinCov) software in a simplified operational context, taking into account the availability of navigation sensors with varying measurement models, ranges, and accuracies. No numerical integration is used, since the relative motion around Gateway is fully characterized with the a priori computation of the QIST coefficients. Maneuver placements are computed to optimize the minimum 3σ delta-v of the trajectory, the position dispersion at a target point, and a convex combination of these two metrics. An order of magnitude runtime improvement is provided over legacy methods with less than 10% error introduced. All QIST results are shown to be in-family with legacy methods. The tradespace for optimal delta-v design is found to range from 77.0 to 93.9 m/s, while the range of optimal dispersion is between 1.4 and 11.7 km.
Continuous linear programming problems with time delays in constraints, noting construction of solutions to dual problem from solution to primal problem
This paper is concerned with the application of optimal estimation and control concepts to the ground alignment of inertial platforms subjected to random sway motion. A seventh-order Kalman filter has been designed for estimating the platform misalignments and sway motion. Using the linear quadratic optimal regulator techniques, feedback controller gains have been designed to drive the initial misalignments to a small value. A reconfiguration of the filter controller scheme has been proposed and verified by Monte Carlo simulation for improving the speed of alignment without changing the alignment accuracy.
Linear back-drive differentials have been proposed as alternatives to conventional gear differentials for applications in which there is only limited rotational motion (e.g., oscillation). The finite nature of the rotation makes it possible to optimize a linear back-drive differential in ways that would not be possible for gear differentials or other differentials that are required to be capable of unlimited rotation. As a result, relative to gear differentials, linear back-drive differentials could be more compact and less massive, could contain fewer complex parts, and could be less sensitive to variations in the viscosities of lubricants. Linear back-drive differentials would operate according to established principles of power ball screws and linear-motion drives, but would utilize these principles in an innovative way. One major characteristic of such mechanisms that would be exploited in linear back-drive differentials is the possibility of designing them to drive or back-drive with similar efficiency and energy input: in other words, such a mechanism can be designed so that a rotating screw can drive a nut linearly or the linear motion of the nut can cause the screw to rotate. A linear back-drive differential (see figure) would include two collinear shafts connected to two parts that are intended to engage in limited opposing rotations. The linear back-drive differential would also include a nut that would be free to translate along its axis but not to rotate. The inner surface of the nut would be right-hand threaded at one end and left-hand threaded at the opposite end to engage corresponding right- and left-handed threads on the shafts. A rotation and torque introduced into the system via one shaft would drive the nut in linear motion. The nut, in turn, would back-drive the other shaft, creating a reaction torque. Balls would reduce friction, making it possible for the shaft/nut coupling on each side to operate with 90 percent efficiency.
Machine learning models are promising as surrogates in optimization when replacing difficult to solve equations or black-box type models. This work demonstrates the viability of linear model decision trees as piecewise-linear surrogates in decision-making problems. Linear model decision trees can be represented exactly in mixed-integer linear programming (MILP) and mixed-integer quadratic constrained programming (MIQCP) formulations. Furthermore, they can represent discontinuous functions, bringing advantages over neural networks in some cases. We present several formulations using transformations from Generalized Disjunctive Programming (GDP) formulations and modifications of MILP formulations for gradient boosted decision trees (GBDT). We then compare the computational performance of these different MILP and MIQCP representations in an optimization problem and illustrate their use on engineering applications. Importantly, we observe faster solution times for optimization problems with linear model decision tree surrogates when compared with GBDT surrogates using the Optimization and Machine Learning Toolkit (OMLT).
With the increasing penetration of intermittent renewable energy sources into the grid, there is a growing need for process systems-based strategies that integrate dispatchable and variable energy systems for supplying the demand while maintaining grid reliability. The proposed framework corresponds to a dynamic mixed-integer linear programming optimization approach that integrates coal-fired and natural gas-fired power plants, NaS batteries for energy storage, and solar/wind energy to supply the demand. This optimization approach considers an economic goal and constraints to provide power balance while maintaining the overall damage of the natural gas combined cycle (NGCC) power plant drum under a maximum stress as well as avoiding the overheating of the NGCC superheater and reheater. Renewable curtailment levels are also retained at minimum levels. Case studies are analyzed considering different loads and renewable penetration levels. The results show that the demand was met for all cases. Grid flexibility was mostly provided by the NGCC, while the batteries were used sparingly. In addition, considering a CO 2 equivalent analysis, the environmental performance was intrinsically connected to grid flexibility and the level of renewable penetration. Stress analysis results reinforced the necessity for an equipment health-related constraint.
Although monoclinic bismuth vanadate (BiVO 4 ) is a promising photoanode for solar water splitting, its practical use is hindered by imperfect photocurrent generation/collection, low photovoltage compared to the bandgap, and corrosion side reactions that limit durability. Here, we introduce a controlled-annealing sol–gel process for BiVO 4 thin-film photoanodes along with an optimized linear-sweep-voltammetry photodeposition of CoFeO x H y cocatalysts. The resulting BiVO 4 films annealed at 550 °C exhibited a photocurrent density of 4.1 mA/cm 2 at 1.23 V RHE under 1 sun AM 1.5G solar simulation and a low onset potential of 0.26 V RHE due to high majority carrier conductivity, a crystalline bulk with reduced defects as evidenced by x-ray photoelectron spectroscopy and photoluminescence lifetime analysis, and thus enhanced photocarrier collection. However, significant degradation in performance was found due to interfacial photocorrosion. To protect the surface and speed the oxygen-evolution reaction CoFeO x H y cocatalyst layers were deposited. By varying the number of consecutive sweeps and adjusting the applied bias range, an ultra-thin (~15 nm) CoFeO x H y cocatalyst layer was uniformly grown deposited over 30 cycles on the BiVO 4 surface. The resulting BiVO 4 /CoFeO x H y yielded 4.03 mA/cm 2 at 1.23 V RHE and onset potential of 0.24 V RHE , with stable operation (~15 % loss in photocurrent at 1.23 V RHE relative to ~60 % loss in the uncatalyzed control sample). These conformal CoFeO x H y catalytic layers function simultaneously to selectively collect photoexcited holes from the BiVO 4 , catalyze the water-oxidation reaction, and protect the BiVO 4 from photodegradation.
A nonlinear programming technique was developed for the automated design and optimization of autopilots for large flexible launch vehicles. This technique, which resulted in the COEBRA program, uses the iterative application of linear programming. The method deals directly with the three main requirements of booster autopilot design: to provide (1) good response to guidance commands; (2) response to external disturbances (e.g. wind) to minimize structural bending moment loads and trajectory dispersions; and (3) stability with specified tolerances on the vehicle and flight control system parameters. The method is applicable to very high order systems (30th and greater per flight condition). Examples are provided that demonstrate the successful application of the employed algorithm to the design of autopilots for both single and multiple flight conditions.