Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “convergence”

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 1,621 records · Page 90

Learning-Accelerated ADMM for Distributed DC Optimal Power Flow

We propose a novel data-driven method to accelerate the convergence of Alternating Direction Method of Multipliers (ADMM) for solving distributed DC optimal power flow (DC-OPF) where lines are shared between independent network partitions. Using previous observations of ADMM trajectories for a given system under varying load, the method trains a recurrent neural network (RNN) to predict the converged values of dual and consensus variables. Given a new realization of system load, a small number of initial ADMM iterations is taken as input to infer the converged values and directly inject them into the iteration. We empirically demonstrate that the online injection of these values into the ADMM iteration accelerates convergence by a significant factor for partitioned 14-, 118- and 2848-bus test systems under differing load scenarios. The proposed method has several advantages: it maintains the security of private decision variables inherent in consensus ADMM; inference is fast and so may be used in online settings; RNN-generated predictions can dramatically improve time to convergence but, by construction, can never result in infeasible ADMM subproblems; it can be easily integrated into existing software implementations. While we focus on the ADMM formulation of distributed DC-OPF in this paper, the ideas presented are naturally extended to other distributed optimization problems.

alternating direction method of multipliers↗

Scalable Simulation of Pressure Gradient-Driven Transport of Rarefied Gases in Complex Permeable Media Using Lattice Boltzmann Method

Accurate representations of slip and transitional flow regimes present a challenge in the simulation of rarefied gas flow in confined systems with complex geometries. In these regimes, continuum-based formulations may not capture the physics correctly. This work considers a regularized multi-relaxation time lattice Boltzmann (LB) method with mixed Maxwellian diffusive and halfway bounce-back wall boundary treatments to capture flow at high Kn. The simulation results are validated against atomistic simulation results from the literature. We examine the convergence behavior of LB for confined systems as a function of inlet and outlet treatments, complexity of the geometry, and magnitude of pressure gradient and show that convergence is sensitive to all three. The inlet and outlet boundary treatments considered in this work include periodic, pressure, and a generalized periodic boundary condition. Compared to periodic and pressure treatments, simulations of complex domains using a generalized boundary treatment conserve mass but require more iterations to converge. Convergence behavior in complex domains improves at higher magnitudes of pressure gradient across the computational domain, and lowering the porosity deteriorates the convergence behavior for complex domains.

42 ENGINEERING↗

Seismicity and the subduction process

There is considerable variation between subduction zones in the largest characteristic earthquake within each zone. Assuming that coupling between downgoing and upper plates is directly related to characteristic earthquake size, tests for correlations between variation in coupling and other physical features of subduction zones are conducted: the lateral extent and penetration depth of Benioff zones, age of subducting lithosphere, convergence rate, and back-arc spreading. Using linear multivariate regression, coupling is correlated with two variables: convergence rate and lithosphere age. Secondary correlations within the data set are penetration depth versus lithosphere age, and lateral extent versus convergence rate. Taken together, the observed correlations suggest a simple qualitative model where convergence rate and lithosphere age determine the horizontal and sinking rates, respectively, of slabs: these parameters influence the seismic coupling in the subduction zone. In the limit of a fast sinking rate and slow convergence rate, back-arc spreading occurs and thereby appears to be a passive process.

Ruff, L.↗

Dilution Jet Mixing Program, phase 1

The effect of jet to mainstream density ratio, flow area convergence as encounted in transition sections, and nonuniform mainstream profile upstream of dilution orifices on the mixing of a row of jets with a confined cross flow was quantified. It is found that: (1) jet spreading rate in transverse direction is increased with increasing J, H/D and with decreasing S/D; (2) the density ratio has only a second order effect on the jet mixing characteristics for a constant momentum ratio; (3) the temperature distributions in the jet mixing region are strongly influenced by the undisturbed mainstream profile; (4) flow area convergence enhances mixing in radial and transverse directions. An asymmetric convergent duct with flat wall injection has the same jet mixing characteristics as a symmetric convergent duct. An asymmetric convergent duct with slant wall injection has a faster jet spreading rate in the transverse direction.

Srinivasan, R.↗

Application of improved numerical schemes

Two approaches which accelerate the solution of the steady state Navier-Stokes equations are discussed. The SIMPLER algorithm, a revised version of SIMPLE, provides a more accurate pressure field for each iteration through the momentum equations, thereby speeding convergence. PISO (Pressure Implicit Split Operator), performs a secondary correction of the velocity and pressure fields (after the typical pressure correction) which enhances convergence. Both schemes account for terms neglected in the SIMPLE approach, but do so in slightly different ways. Two dimensional driven cavity flow and flow over a step were calculated to examine the effect of geometry on the performance of these schemes. Computations were carried out on a series of progressively finer grids. The effect of relaxation number on convergence rate was analyzed, using results from SIMPLE as criteria for performance correlation. Results show: (1) the improved schemes promoted convergence by up to 60% for the driven cavity and 40% for flow over a step; (2) for the driven cavity problem, the efficiency of PISO and SIMPLER increased as the number of nodes increased; and (3) to ensure faster convergence, higher relaxation numbers must be applied.

Neely, G. H.↗

Operator induced multigrid algorithms using semirefinement

A variant of multigrid, based on zebra relaxation, and a new family of restriction/prolongation operators is described. Using zebra relaxation in combination with an operator-induced prolongation leads to fast convergence, since the coarse grid can correct all error components. The resulting algorithms are not only fast, but are also robust, in the sense that the convergence rate is insensitive to the mesh aspect ratio. This is true even though line relaxation is performed in only one direction. Multigrid becomes a direct method if an operator-induced prolongation is used, together with the induced coarse grid operators. Unfortunately, this approach leads to stencils which double in size on each coarser grid. The use of an implicit three point restriction can be used to factor these large stencils, in order to retain the usual five or nine point stencils, while still achieving fast convergence. This algorithm achieves a V-cycle convergence rate of 0.03 on Poisson's equation, using 1.5 zebra sweeps per level, while the convergence rate improves to 0.003 if optimal nine point stencils are used. Numerical results for two and three dimensional model problems are presented, together with a two level analysis explaining these results.

Decker, Naomi↗

Operator induced multigrid algorithms using semirefinement

A variant of multigrid, based on zebra relaxation, and a new family of restriction/prolongation operators is described. Using zebra relaxation in combination with an operator-induced prolongation leads to fast convergence, since the coarse grid can correct all error components. The resulting algorithms are not only fast, but are also robust, in the sense that the convergence rate is insensitive to the mesh aspect ratio. This is true even though line relaxation is performed in only one direction. Multigrid becomes a direct method if an operator-induced prolongation is used, together with the induced coarse grid operators. Unfortunately, this approach leads to stencils which double in size on each coarser grid. The use of an implicit three point restriction can be used to factor these large stencils, in order to retain the usual five or nine point stencils, while still achieving fast convergence. This algorithm achieves a V-cycle convergence rate of 0.03 on Poisson's equation, using 1.5 zebra sweeps per level, while the convergence rate improves to 0.003 if optimal nine point stencils are used. Numerical results for two- and three-dimensional model problems are presented, together with a two level analysis explaining these results.

Decker, Naomi Henderson↗

A parallel algorithm for the eigenvalues and eigenvectors for a general complex matrix

A new parallel Jacobi-like algorithm is developed for computing the eigenvalues of a general complex matrix. Most parallel methods for this parallel typically display only linear convergence. Sequential norm-reducing algorithms also exit and they display quadratic convergence in most cases. The new algorithm is a parallel form of the norm-reducing algorithm due to Eberlein. It is proven that the asymptotic convergence rate of this algorithm is quadratic. Numerical experiments are presented which demonstrate the quadratic convergence of the algorithm and certain situations where the convergence is slow are also identified. The algorithm promises to be very competitive on a variety of parallel architectures.

Shroff, Gautam↗

Use of the surface wind field as a predictor of thunderstorms and cloud-to-ground lightning at Kennedy Space Center

The convergence-lightning nowcasting method that began to be operationally tested and evaluated by USAF personnel at the KSC in 1986 uses surface convergence, especially the areal average, to identify the potential for new thunderstorm growth. When this areal method fails, normalized cell convergence, which is defined as the integral of cellular convergence with respect to its area, can be used as a complementary technique in association with the areal one. A detailed examination is conducted for one day, and the relationship of normalized cell convergence to lightning flash counts is examined in a 68-day ensemble.

Watson, Andrew I.↗

Approximation methods for stochastic petri nets

Stochastic Marked Graphs are a concurrent decision free formalism provided with a powerful synchronization mechanism generalizing conventional Fork Join Queueing Networks. In some particular cases the analysis of the throughput can be done analytically. Otherwise the analysis suffers from the classical state explosion problem. Embedded in the divide and conquer paradigm, approximation techniques are introduced for the analysis of stochastic marked graphs and Macroplace/Macrotransition-nets (MPMT-nets), a new subclass introduced herein. MPMT-nets are a subclass of Petri nets that allow limited choice, concurrency and sharing of resources. The modeling power of MPMT is much larger than that of marked graphs, e.g., MPMT-nets can model manufacturing flow lines with unreliable machines and dataflow graphs where choice and synchronization occur. The basic idea leads to the notion of a cut to split the original net system into two subnets. The cuts lead to two aggregated net systems where one of the subnets is reduced to a single transition. A further reduction leads to a basic skeleton. The generalization of the idea leads to multiple cuts, where single cuts can be applied recursively leading to a hierarchical decomposition. Based on the decomposition, a response time approximation technique for the performance analysis is introduced. Also, delay equivalence, which has previously been introduced in the context of marked graphs by Woodside et al., Marie's method and flow equivalent aggregation are applied to the aggregated net systems. The experimental results show that response time approximation converges quickly and shows reasonable accuracy in most cases. The convergence of Marie's method and flow equivalent aggregation are applied to the aggregated net systems. The experimental results show that response time approximation converges quickly and shows reasonable accuracy in most cases. The convergence of Marie's is slower, but the accuracy is generally better. Delay equivalence often fails to converge, while flow equivalent aggregation can lead to potentially bad results if a strong dependence of the mean completion time on the interarrival process exists.

Jungnitz, Hauke Joerg↗

The multigrid preconditioned conjugate gradient method

A multigrid preconditioned conjugate gradient method (MGCG method), which uses the multigrid method as a preconditioner of the PCG method, is proposed. The multigrid method has inherent high parallelism and improves convergence of long wavelength components, which is important in iterative methods. By using this method as a preconditioner of the PCG method, an efficient method with high parallelism and fast convergence is obtained. First, it is considered a necessary condition of the multigrid preconditioner in order to satisfy requirements of a preconditioner of the PCG method. Next numerical experiments show a behavior of the MGCG method and that the MGCG method is superior to both the ICCG method and the multigrid method in point of fast convergence and high parallelism. This fast convergence is understood in terms of the eigenvalue analysis of the preconditioned matrix. From this observation of the multigrid preconditioner, it is realized that the MGCG method converges in very few iterations and the multigrid preconditioner is a desirable preconditioner of the conjugate gradient method.

Tatebe, Osamu↗

Active Control of Wind Tunnel Noise

The need for an adaptive active control system was realized, since a wind tunnel is subjected to variations in air velocity, temperature, air turbulence, and some other factors such as nonlinearity. Among many adaptive algorithms, the Least Mean Squares (LMS) algorithm, which is the simplest one, has been used in an Active Noise Control (ANC) system by some researchers. However, Eriksson's results, Eriksson (1985), showed instability in the ANC system with an ER filter for random noise input. The Restricted Least Squares (RLS) algorithm, although computationally more complex than the LMS algorithm, has better convergence and stability properties. The ANC system in the present work was simulated by using an FIR filter with an RLS algorithm for different inputs and for a number of plant models. Simulation results for the ANC system with acoustic feedback showed better robustness when used with the RLS algorithm than with the LMS algorithm for all types of inputs. Overall attenuation in the frequency domain was better in the case of the RLS adaptive algorithm. Simulation results with a more realistic plant model and an RLS adaptive algorithm showed a slower convergence rate than the case with an acoustic plant as a delay plant. However, the attenuation properties were satisfactory for the simulated system with the modified plant. The effect of filter length on the rate of convergence and attenuation was studied. It was found that the rate of convergence decreases with increase in filter length, whereas the attenuation increases with increase in filter length. The final design of the ANC system was simulated and found to have a reasonable convergence rate and good attenuation properties for an input containing discrete frequencies and random noise.

Hollis, Patrick↗

An Algebraic Multigrid Solver for Navier-Stokes Problems in the Discrete Second-Order Approximation

An algebraic multigrid scheme is presented for solving the discrete Navier-Stokes equations to second-order accuracy using the defect-correction method. Solutions have been obtained for problems involving both structured and unstructured meshes, with the resolution and resolution grading controlled by global and local mesh refinements. The solver is efficient and robust to the extent that no underrelaxation of variables has been required to ensure convergence, but rates of convergence can be improved with small amounts of underrelaxation of the velocity-pressure coupling. Provided that the computational mesh can resolve the flow field, convergence characteristics are almost mesh independent. Rates of convergence actually improve with refinement, asymptotically approaching mesh independent values. For extremely coarse meshes where dispersive truncation errors would be expected to prevent convergence (or even induce divergence), solutions can still be obtained by using explicit underrelaxation in the iterative cycle.

Webster, R.↗

The Effects of Dissipation and Coarse Grid Resolution for Multigrid in Flow Problems

The objective of this paper is to investigate the effects of the numerical dissipation and the resolution of the solution on coarser grids for multigrid with the Euler equation approximations. The convergence is accomplished by multi-stage explicit time-stepping to steady state accelerated by FAS multigrid. A theoretical investigation is carried out for linear hyperbolic equations in one and two dimensions. The spectra reveals that for stability and hence robustness of spatial discretizations with a small amount of numerical dissipation the grid transfer operators have to be accurate enough and the smoother of low temporal accuracy. Numerical results give grid independent convergence in one dimension. For two-dimensional problems with a small amount of numerical dissipation, however, only a few grid levels contribute to an increased speed of convergence. This is explained by the small numerical dissipation leading to dispersion. Increasing the mesh density and hence making the problem over resolved increases the number of mesh levels contributing to an increased speed of convergence. If the steady state equations are elliptic, all grid levels contribute to the convergence regardless of the mesh density.

Eliasson, Peter↗

The Variability and Predictability of Precipitation over Southeast Asian

The large scale atmospheric moisture budget, including precipitation, moisture flux convergence, and evaporation is described over southeast Asian. ne results are from the Goddard Earth Observing System (GEOS-1) data assimilation system for the period March 1981 - February 1995. On average, in the late spring and early summer season, precipitation is largely balanced by the atmospheric moisture flux convergence over southeast Asian; evaporation is also an important contributor. The seasonal variability of precipitation is best related to variations in the moisture flux convergence. On the monthly time scales, both the precipitable water and precipitation anomalies are most closely related to moisture flux convergence anomalies and less closely related to evaporation anomalies. The impact of the low-level jet (LLJ) on the strong low-level moisture flux convergence will be discussed. An ensemble of GEOS-2 atmospheric General Circulation Model (GCM) seasonal forecasts and long-term simulation are analyzed to assess the controlling influences of boundary forcing and memory of initial conditions. We also assess the variability and potential predictability from these experiments.

Chang, Yehui↗

The Climate Signal in Regional Moisture Fluxes: A Comparison of Three Global Data Assimilation Products

This study assesses the quality of estimates of climate variability in moisture flux and convergence from three assimilated data sets: two are reanalysis products generated at the Goddard Data Assimilation Office (DAO) and the National Centers for Environmental Prediction/National Centers for Atmospheric Research (NCEPJNCAR), and the third consists of the operational analyses generated at the European Center for Medium Range Forecasts (ECMWF). The regions under study (the United States Great Plains, the Indian monsoon region, and Argentina east of the Andes) are characterized by frequent low level jets (LLJs) and other interannual low level wind variations tied to the large-scale flow. While the emphasis is on the reanalysis products, the comparison with the operational product is provided to help assess the improvements gained from a fixed analysis system. All three analyses capture the main moisture flux anomalies associated with selected extreme climate (drought and flood) events during the period 1985-93. The correspondence is strongest over the Great Plains and weakest over the Indian monsoon region reflecting differences in the observational coverage. For the reanalysis products, the uncertainties in the lower tropospheric winds is by far the dominant source of the discrepancies in the moisture flux anomalies in the middle latitude regions. Only in the Indian Monsoon region, where interannual variability in the low level winds is comparatively small, does the moisture bias play a substantial role. In contrast, the comparisons with the operational product show differences in moisture which are comparable torhe differences in the wind in all three regions. Compared with the fluxes, the anomalous moisture convergences show substantially larger differences among the three products. The best agreement occurs over the Great Plains region where all three products show vertically-integrated moisture convergence during the floods and divergence during the drought with differences in magnitude of about 25%. The reanalysis products, in particular, show good agreement in depicting the different roles of the mean flow and transients during the flood and drought periods. Differences between the three products in the other two regions exceed 100% reflecting differences in the low level jets and the large scale circulation patterns. The operational product tends to have locally larger amplitude convergence fields which average out in area-mean budgets: this appears to be at least in part due to errors in the surface pressure fields and aliasing from the higher resolution of the original ECMWF fields. On average, the reanalysis products show higher coherence with each other than with the operational product in the estimates of interannual variability. This result is less clear in the Indian monsoon region where differences in the input observations appears to be an important factor. The agreement in the anomalous convergence patterns is, however, still rather poor even over relatively data dense regions such as the United States Great Plains. These differences are attributed to deficiencies in the assimilating GCM's representations of the planetary boundary layer and orography, and a global observing system incapable of resolving the highly confined low level winds associated with the climate anomalies.

Min, Wei↗

Uncertainties Associated with Theoretically Calculated N2-Broadened Half-Widths of H2O Lines

With different choices of the cut-offs used in theoretical calculations, we have carried out extensive numerical calculations of the N2-broadend Lorentzian half-widths of the H2O lines using the modified Robert-Bonamy formalism. Based on these results, we are able to thoroughly check for convergence. We find that, with the low-order cut-offs commonly used in the literature, one is able to obtain converged values only for lines with large half-widths. Conversely, for lines with small half-widths, much higher cut-offs are necessary to guarantee convergence. We also analyse the uncertainties associated with calculated half-widths, and these are correlated as above. In general, the smaller the half-widths, the poorer the convergence and the larger the uncertainty associated with them. For convenience, one can divide all H2O lines into three categories, large, intermediate, and small, according to their half-width values. One can use this division to judge whether the calculated half-widths are converged or not, based on the cut-offs used, and also to estimate how large their uncertainties are. We conclude that with the current Robert- Bonamy formalism, for lines in category lone can achieve the accuracy requirement set by HITRAN, whereas for lines in category 3, it 'is impossible to meet this goal.

Ma, Q.↗

Low-Thrust Trajectory Optimization with Simplified SQP Algorithm

The problem of low-thrust trajectory optimization in highly perturbed dynamics is a stressing case for many optimization tools. Highly nonlinear dynamics and continuous thrust are each, separately, non-trivial problems in the field of optimal control, and when combined, the problem is even more difficult. This paper de-scribes a fast, robust method to design a trajectory in the CRTBP (circular restricted three body problem), beginning with no or very little knowledge of the system. The approach is inspired by the SQP (sequential quadratic programming) algorithm, in which a general nonlinear programming problem is solved via a sequence of quadratic problems. A few key simplifications make the algorithm presented fast and robust to initial guess: a quadratic cost function, neglecting the line search step when the solution is known to be far away, judicious use of end-point constraints, and mesh refinement on multiple shooting with fixed-step integration.In comparison to the traditional approach of plugging the problem into a “black-box” NLP solver, the methods shown converge even when given no knowledge of the solution at all. It was found that the only piece of information that the user needs to provide is a rough guess for the time of flight, as the transfer time guess will dictate which set of local solutions the algorithm could converge on. This robustness to initial guess is a compelling feature, as three-body orbit transfers are challenging to design with intuition alone. Of course, if a high-quality initial guess is available, the methods shown are still valid.We have shown that endpoints can be efficiently constrained to lie on 3-body repeating orbits, and that time of flight can be optimized as well. When optimizing the endpoints, we must make a trade between converging quickly on sub-optimal endpoints or converging more slowly on end-points that are arbitrarily close to optimal. It is easy for the mission design engineer to adjust this trade based on the problem at hand.The biggest limitation to the algorithm at this point is that multi-revolution transfers (greater than 2 revolutions) do not work nearly as well. This restriction comes in because the relationship between node 1 and node N becomes increasingly nonlinear as the angular distance grows. Trans-fers with more than about 1.5 complete revolutions generally require the line search to improve convergence. Future work includes: Comparison of this algorithm with other established tools; improvements to how multiple-revolution transfers are handled; parallelization of the Jacobian computation; in-creased efficiency for the line search; and optimization of many more trajectories between a variety of 3-body orbits.

Parrish, Nathan L.↗