Optimal control of distributed parameter systems using multilevel techniques
Optimal control of distributed parameter systems using multilevel techniques
SEARCH · Engineering Papers
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.
Optimal control of distributed parameter systems using multilevel techniques
This study compares two statistical modeling approaches to construct summary dose-response curves for low boom community noise surveys. NASA field survey data were used that consist of multiple responses from a survey participant. These data require an approach that accounts for the correlation among repeated annoyance observations from the same participant. A multilevel model accounts for the correlation by allowing estimated parameters to vary with each survey participant. On the other hand, a population average model utilizes generalized estimating equations and accounts for the correlation via a user-specified within-subject correlation structure. A visual comparison of the dose-response curves for these two methods reveals similar results. When comparing specific points along the summary curves, the multilevel model yields more precise confidence bounds than the population average model. The similarity between the summary curves derived from each model lends validity to both approaches for approximating a population representative summary curve.
Evolution has conserved “economic” systems that perform many functions, faster or better, with less. For example, three to five leukocyte types protect from thousands of pathogens. In order to achieve so much with so little, biological systems combine their limited elements, creating complex structures. Yet, the prevalent research paradigm is reductionist. Focusing on infectious diseases, reductionist and non-reductionist views are here described. Furthermore, the literature indicates that reductionism is associated with information loss and errors, while non-reductionist operations can extract more information from the same data. When designed to capture one-to-many/many-to-one interactions—including the use of arrows that connect pairs of consecutive observations—non-reductionist (spatial–temporal) constructs eliminate data variability from all dimensions, except along one line, while arrows describe the directionality of temporal changes that occur along the line. To validate the patterns detected by non-reductionist operations, reductionist procedures are needed. Integrated (non-reductionist and reductionist) methods can (i) distinguish data subsets that differ immunologically and statistically; (ii) differentiate false-negative from -positive errors; (iii) discriminate disease stages; (iv) capture in vivo, multilevel interactions that consider the patient, the microbe, and antibiotic-mediated responses; and (v) assess dynamics. Integrated methods provide repeatable and biologically interpretable information.
The goal of this project is to model and control a novel unidirectional cascaded multilevel bridgeless rectifier as an active front end in medium and high voltage applications. This topology has many advantages over a conventional cascaded H-bridge rectifier, such as lower implementation cost, higher reliability, and greater flexibility with similar power quality.The steady-state mathematical model is used to develop a method for the voltage balancing of dc cells. Power factor analysis is discussed to achieve unity power factor using fully controlled hbridge cells. Power loss, efficiency, and cost comparison studies between the traditional cascaded H-Bridge converter and the proposed bridgeless converter demonstrate the advantages. A novel control strategy is proposed to achieve dc voltage balancing, fast and robust grid synchronization and power factor correction under unbalanced load conditions. Simulation and experimental results validate the models and control method.
A novel discharging control strategy is proposed and implemented for submodules in modular multilevel converters (MMCs). By sequentially controlling the arm currents of a MMC to the arm resistors, all submodule capacitors are discharged safely and rapidly without adding additional devices and circuits, resulting in improved performance and reduced cost. Both open- and closed-loop control methods are evaluated for effectiveness of the new control strategy. Simulation and experimental results are provided to validate the new control strategy.
The traditional single-level checkpointing method suffers from significant overhead on large-scale platforms. Hence, multilevel checkpointing protocols have been studied extensively in recent years. The multilevel checkpoint approach allows different levels of checkpoints to be set (each with different checkpoint overheads and recovery abilities), in order to further improve the fault tolerance performance of extreme-scale HPC applications. How to optimize the checkpoint intervals for each level, however, is an extremely difficult problem. In this paper, we construct an easy-to-use two-level checkpoint model. Checkpoint level 1 deals with errors with low checkpoint/recovery overheads such as transient memory errors, while checkpoint level 2 deals with hardware crashes such as node failures. Compared with previous optimization work, our new optimal checkpoint solution offers two improvements: (1) it is an online solution without requiring knowledge of the job length in advance, and (2) it shows that periodic patterns are optimal and determines the best pattern. We evaluate the proposed solution and compare it with the most up-to-date related approaches on an extreme-scale simulation testbed constructed based on a real HPC application execution. Simulation results show that our proposed solution outperforms other optimized solutions and can improve the performance significantly in some cases. Specifically, with the new solution the wall-clock time can be reduced by up to 25.3% over that of other state-of-the-art approaches. Lastly, a brute-force comparison with all possible patterns shows that our solution is always within 1% of the best pattern in the experiments.
We present a topology-based method for mesh-partitioning in three-dimensional discrete fracture network (DFN) simulations that takes advantage of the intrinsic multi-level nature of a DFN. DFN models are used to simulate flow and transport through low-permeability fractured media in the subsurface by explicitly representing fractures as discrete entities. The governing equations for flow and transport are numerically integrated on computational meshes generated on the interconnected fracture networks. Modern high-fidelity DFN simulations require high-performance computing on multiple processors where performance and scalability depends partially on obtaining a high-quality partition of the mesh to balance work-loads and minimize communication across all processors. The discrete structure of a DFN naturally lends itself to various graph representations, which can be thought of as coarse-scale representations of the computational mesh. Using this concept, we develop two applications of the multilevel graph partitioning algorithm to partition the mesh of a DFN. In the first, we project a partition of the graph based on the DFN topology onto the mesh of the DFN and in the second, this DFN-based projection is used as the initial condition for further partitioning refinement of the mesh. We compare the performance of these methods with standard multi-level graph partitioning using graph-based metrics (cut, imbalance, partitioning time), computational-based metrics (FLOPS, iterations, solver time), and total run time. The DFN-based and the mesh-based partitioning methods are comparable in terms of the graph-based metrics, but the time required to obtain the partition is several orders of magnitude faster using the DFN-based partitions. The computation-based metrics show comparable performance between both methods so, in combination, the DFN-based partitions are several orders of magnitude faster than the mesh-based partition. Furthermore, the method which uses the DFN-partition solution as the initial condition of the mesh partition provided cut and imbalance values that were close to the mesh-based partition but in a fraction of the time. In turn, this hybrid method outperformed both of the other methods in terms of the total run time.
A multispectral, multiresolution (MSMR) method is developed for analyzing scenes of overlapping cloud layers. The MSMR method is applied to data from the NOAA 11 advanced very high resolution radiometer (AVHRR) and the high-resolution infrared radiometer sounder (HIRS-2). The data are from a nighttime oceanic scene in which a semitransparent cirrus veil overlays a large-scale stratus cloud. Low-cloud and clear-sky radiances are determined using a spatial coherence technique. Middle to upper level cloud pressures and radiances are estimated from HIRS-2 15 micrometer CO2 band radiometric data. The MSMR method improves the interpretation of a nighttime, oceanic scene containing thin cirrus over a large-scale stratiform cloud. If, for example, the same scene is analyzed using only the AVHRR 10.8 micrometer channel, the accompanying retrieved cloud heights are found to be between the cirrus and stratus cloud heights and are incorrectly identified as midlevel altostratus clouds. Theoretical radiative transfer model results for both water droplet spheres and randomly oriented hexagonal ice crystals are compared to observed AVHRR brightness temperature differences (BTD) between the 3.7- and 10.8 micrometer channels (BTD(sup 34)) and between the 10.8- and 12- micrometer channels (BTD(sup 45)) to distinguish among the effects of cloud optical depth, particle size, and phase for both single-layer clouds and overlapping two-layer clouds. Theoretical BTD calculations are used to estimate the range of effective particle sizes for eac h cloud layer. The data for the cirrus in the case study region near Bermuda are consistent with theoretical results for relatively small randomly oriented hexagonal ice crystals. The observed BTD(sup 34) and BTD(sup 45) values are lower for the cirrus above a lower-level cloud than for single-level cirrus with no underlying cloud. In certain cases the BTD analysis provides a way to distinguish between clouds composed of supercooled water droplets rather than ice particles. Analysis of nighttime data permits determination of stratus infrared optical depths smaller than 4.
Emerging quantum processors provide an opportunity to explore new approaches for solving traditional problems in the post Moore’s law supercomputing era. However, the limited number of qubits makes it infeasible to tackle massive real-world datasets directly in the near future, leading to new challenges in utilizing these quantum processors for practical purposes. Furthermore, hybrid quantum-classical algorithms that leverage both quantum and classical types of devices are considered as one of the main strategies to apply quantum computing to large-scale problems. In this article, we advocate the use of multilevel frameworks for combinatorial optimization as a promising general paradigm for designing hybrid quantum-classical algorithms. To demonstrate this approach, we apply this method to two well-known combinatorial optimization problems, namely, the Graph Partitioning Problem, and the Community Detection Problem. We develop hybrid multilevel solvers with quantum local search on D-Wave’s quantum annealer and IBM’s gate-model based quantum processor. We carry out experiments on graphs that are orders of magnitude larger than the current quantum hardware size, and we observe results comparable to state-of-the-art solvers in terms of quality of the solution.
The efficient solution of sparse, linear systems that arise through the discretization of partial differential equations remains a key challenge for a range of high performance scientific simulations. One approach for reducing data movement and improving performance is by exposing and exploiting structure in a problem through the use of robust structured multilevel solvers. By choosing coarsening that preserves the structure of the problem, these methods maintain efficient structured computation and communication throughout the multigrid hierarchy. However, when coarsening is not permitted to be dependent on the operator, anisotropy must be addressed by the smoother — producing error compatible for coarse-grid correction with structured coarsening. Here, the components required in a scalable parallel structured solver are described with a focus on memory and communication efficiency of robust smoothers. While the implementation of communication and memory reduction techniques in smoothers integrated in a complete 3D solver present a significant engineering challenge, a novel approach is proposed that addresses these challenges systematically through a change to the solver’s execution model. Enabled by user-level threading paired with a set of data and communication abstractions, this approach permits seamless aggregation of communication in plane smoothers — directly reusing code for a 2D distributed multilevel cycle. Results show an effective reduction in communication costs for coarse-grid problems, and result in a speedup of 8.7x in smoothing routines shown in Fig. 12 using this approach. This produces a significant improvement to strong scalability while maintaining favorable weak scaling behavior. Finally, a parallel scaling study using a series of refined meshes is included that demonstrates the effectiveness of this approach in an application of interest.
We present a mechanically robust, ultraelastic foam with controlled multiscale architectures and tunable mechanical/conductive performance is fabricated via 3D printing. Hierarchical porosity, including both macro- and microscaled pores, are produced by the combination of direct ink writing (DIW), acid etching, and phase inversion. The thixotropic inks in DIW are formulated by a simple one-pot process to disperse duo nanoparticles (nanoclay and silica nanoparticles) in a polyurethane suspension. The resulting lightweight foam exhibits tailorable mechanical strength, unprecedented elasticity (standing over 1000 compression cycles), and remarkable robustness (rapidly and fully recover after a load more than 20 000 times of its own weight). Surface coating of carbon nanotubes yields a conductive elastic foam that can be used as piezoresistivity sensor with high sensitivity. For the first time, this strategy achieves 3D printing of elastic foam with controlled multilevel 3D structures and mechanical/conductive properties. In conclusion, the facile ink preparation method can be utilized to fabricate foams of various materials with desirable performance via 3D printing.
Stabilization of high-order constant systems by multilevel control is proposed in the framework of the decomposition-aggregation method for stability analysis of large-scale systems. An 11-order linear model of a spinning flexible spacecraft is stabilized by the method using both the passive and the active stabilization devices. The method can take advantage of the special structural features of the model to provide the best estimate of the stability region for an important parameter which couples the wobble and the spin motion of the spacecraft.
We discuss the recently introduced multilevel algorithm for the steady-state solution of Markov chains. The method is based on an aggregation principle which is well established in the literature and features a multiplicative coarse-level correction. Recursive application of the aggregation principle, which uses an operator-dependent coarsening, yields a multi-level method which has been shown experimentally to give results significantly faster than the typical methods currently in use. When cast as a multigrid-like method, the algorithm is seen to be a Galerkin-Full Approximation Scheme with a solution-dependent prolongation operator. Special properties of this prolongation lead to the cancellation of the computationally intensive terms of the coarse-level equations.
Historical evolution of engineering disciplines and the complexity of the MDO problem suggest that disciplinary autonomy is a desirable goal in formulating and solving MDO problems. We examine the notion of disciplinary autonomy and discuss the analytical properties of three approaches to formulating and solving MDO problems that achieve varying degrees of autonomy by distributing the problem along disciplinary lines. Two of the approaches-Optimization by Linear Decomposition and Collaborative Optimization-are based on bi-level optimization and reflect what we call a structural perspective. The third approach, Distributed Analysis Optimization, is a single-level approach that arises from what we call an algorithmic perspective. The main conclusion of the paper is that disciplinary autonomy may come at a price: in the bi-level approaches, the system-level constraints introduced to relax the interdisciplinary coupling and enable disciplinary autonomy can cause analytical and computational difficulties for optimization algorithms. The single-level alternative we discuss affords a more limited degree of autonomy than that of the bi-level approaches, but without the computational difficulties of the bi-level methods. Key Words: Autonomy, bi-level optimization, distributed optimization, multidisciplinary optimization, multilevel optimization, nonlinear programming, problem integration, system synthesis
The global dynamics of the ionized and neutral components in the environment of Io plays an important role in the interaction of Jupiter's corotating magnetospheric plasma with Io. The stationary simulation of this problem was done in the MHD and the electrodynamics approaches. One of the main significant results from the simplified two-fluid model simulations was a production of the structure of the double-peak in the magnetic field signature of the I0 flyby that could not be explained by standard MHD models. In this paper, we develop a method of kinetic ion simulation. This method employs the fluid description for electrons and neutrals whereas for ions multilevel, drift-kinetic and particle, approaches are used. We also take into account charge-exchange and photoionization processes. Our model provides much more accurate description for ion dynamics and allows us to take into account the realistic anisotropic ion distribution that cannot be done in fluid simulations. The first results of such simulation of the dynamics of ions in the Io's environment are discussed in this paper.
Recent work in MPACT, the primary deterministic neutron transport solver in CASL’s VERA, has focused on significantly improving the overall performance of the calculations and many components within the code have been improved. The paper focuses on the most recent development to improve the 2D/1D subplane method in preparation for a production-level capability. The key developments presented here focus on incorporating a multilevel-in-space CMFD solver that also collapses on the subplane bases, a dynamic subplane meshing capability that better preserves accuracy during control rod movements, and overall usability improvements to make using the subplane method seamless. Compared to the current default 2D/1D solver in MPACT, substantial improvements are demonstrated on two progression problems, including a standard quarter core model, where potential savings observed at ~2× in memory and ~5-6× in core-hour requirements.
Recent work in MPACT, the primary deterministic neutron transport solver in CASL’s VERA, has focused on significantly improving the overall performance of the calculations and many components within the code have been improved. The paper focuses on the most recent development to improve the 2D/1D subplane method in preparation for a production-level capability. The key developments presented here focus on incorporating a multilevel-in-space CMFD solver that also collapses on the subplane bases, a dynamic subplane meshing capability that better preserves accuracy during control rod movements, and overall usability improvements to make using the subplane method seamless. Compared to the current default 2D/1D solver in MPACT, substantial improvements are demonstrated on two progression problems, including a standard quarter core model, where potential savings observed at ~2× in memory and ~5-6× in core-hour requirements.
Cardiovascular pathologies, such as a brain aneurysm, are affected by the global blood circulation as well as by the local microrheology. Hence, developing computational models for such cases requires the coupling of disparate spatial and temporal scales often governed by diverse mathematical descriptions, e.g., by partial differential equations (continuum) and ordinary differential equations for discrete particles (atomistic). However, interfacing atomistic-based with continuum-based domain discretizations is a challenging problem that requires both mathematical and computational advances. We present here a hybrid methodology that enabled us to perform the first multiscale simulations of platelet depositions on the wall of a brain aneurysm. The large scale flow features in the intracranial network are accurately resolved by using the high-order spectral element Navier–Stokes solver NεκTαr. The blood rheology inside the aneurysm is modeled using a coarse-grained stochastic molecular dynamics approach (the dissipative particle dynamics method) implemented in the parallel code LAMMPS. The continuum and atomistic domains overlap with interface conditions provided by effective forces computed adaptively to ensure continuity of states across the interface boundary. A two-way interaction is allowed with the time-evolving boundary of the (deposited) platelet clusters tracked by an immersed boundary method. The corresponding heterogeneous solvers (NεκTαr and LAMMPS) are linked together by a computational multilevel message passing interface that facilitates modularity and high parallel efficiency. Results of multiscale simulations of clot formation inside the aneurysm in a patient-specific arterial tree are presented. We also discuss the computational challenges involved and present scalability results of our coupled solver on up to 300 K computer processors. Validation of such coupled atomistic-continuum models is a main open issue that has to be addressed in future work.