Search NASA⌕ Search

SEARCH · Search NASA

Results for “Multi-zone model”

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 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
NASA NTRS1988

The multi-zone calculation of turbomachinery flows. II - The multi-zone calculation of the turbulent, two-specie flow through the SSME HPFTP first and second stage cavities

A multi-zone Navier-Stokes methodology to calculate the two-specie flow through the first and second stage cooling cavities of the Space Shuttle Main Engine (SSME) high pressure fuel turbopump (HPFTP) is developed. A simplified two-component fluid formulation is used to model the interaction of coolant and hot gas. Johnston's secant approximation is used to define an appropriate near wall velocity for use in a three-dimensional law of the wall. The basic Navier-Stokes algorithm used is a finite-volume, predictor-corrector algorithm which uses a pressure correction technique. A multi-zone method is used to partition each cavity into easily handled subdomains. The results show that coolant flow is pumped up the turbine wheel for both cavities, creating a region of large temperature gradients on the turbine shank.

Williams, M.↗
DOE OSTI2020

An encoder–decoder LSTM-based EMPC framework applied to a building HVAC system

Numerous studies have demonstrated the benefit of economic model predictive control (EMPC) applied to building heating, ventilation, and air conditioning (HVAC) systems. However, the construction and training of predictive models for building HVAC systems are widely recognized as a key technological barrier preventing large-scale adoption of EMPC for buildings. In this work, an encoder–decoder long short-term memory-based EMPC framework is developed. The key advantage of the approach is that a model may be automatically generated from a list of inputs and outputs. From the definition of inputs and outputs, the constructed model may be trained and automatically embedded into the EMPC framework for real-time estimation and control. The overall end-to-end EMPC framework from model training to on-line estimation and control are described. To this end, the encoder–decoder model provides a natural framework for state estimation (encoder), which is required to provide an initial condition for the predictive model of EMPC (decoder). Closed-loop simulations using EnergyPlus are performed to demonstrate the approach. The simulated closed-loop system consists of a building zone from a multi-zone building, which is served by an air handling unit-variable air volume HVAC system. For the HVAC example considered, the trained encoder–decoder model can predict the indoor air temperature and HVAC sensible cooling rate of a building zone over a two-day horizon with high accuracy. Overall, we find that considering a time-of-use electric rate structure, the EMPC, which manipulates the zone temperature setpoint, can reduce the HVAC power consumption cost relative to keeping the zone temperature setpoint at its maximum value (i.e., minimum energy approach).

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2022

Resilient cooling through geothermal district energy system

Decarbonization and resilience to heat waves have recently become high priorities for building and district energy systems. Geothermal coupled district heating and cooling systems that operate a water loop near ground temperature gain increasing adoption to support decarbonization. In these systems, vapor-compression machines, distributed in the energy transfer stations, lift the temperature up or down to the needs of the particular building. In principle, these systems can provide low-power, free cooling from the geothermal bore field during heat waves when electricity is often scarce. However, the performance of such a resilience operation mode and its implication on the energy system configuration and the sizing of the bore field and HVAC equipment is not yet understood. Consequently, we are assessing their resilience, power use and design implications under a scenario of a heat wave on five working days during which chillers are switched off to reduce electrical consumption. Our analysis is based on high-fidelity, coupled dynamic models of district energy, building-side HVAC and actual control logic, with whole building energy simulation used to assess thermal conditions in a 2004 vintage multi-zone office building in Chicago, IL. The results show that relying only on waterside economizer cooling, the indoor thermal conditions can be maintained in a tolerable range for the majority of the building zones with half the electrical energy compared to standard chiller operation. Thermal comfort in the hottest zones can be further improved by oversizing the cooling coil. However, the waterside economizer has significant implications on the system configuration and sizing: The geothermal bore field needs to be sized about 30% larger than the upper limit of the range observed for conventional geothermal systems. Nevertheless, if a central chiller plant is added, the bore field can be downsized to the typical design range. The latter configuration still allows compressor-less cooling during the heat wave with peak power reduced by 60% compared to the standard design and chiller operation.

15 GEOTHERMAL ENERGY↗
DOE OSTI2021

What are the Impacts on the HVAC System when it Provides Frequency Regulation? – A Comprehensive Case Study with a Multi-Zone Variable Air Volume (VAV) System

Heating, ventilation, and air-conditioning (HVAC) systems in commercial buildings have proven to be a great asset for providing secondary frequency regulation (FR) service to the power grid. However, how the provision of FR service affects the HVAC system Quality of Service (QOS) is still largely unknown. In particular, there exists a controversy regarding the energy efficiency impact. This paper investigates how a HVAC system performance and its associated controls are impacted during and after the FR service. For the assessment, we developed a dynamic Modelica-based medium office building model as a virtual testbed with the Air Handling Unit (AHU) fans providing the FR service. Different regulation capacities (demand levels) and standard test signals under various load profiles in the U.S. are considered. The results show that FR has little or no impact on the HVAC operation and occupant thermal comfort while the building is providing the service if an appropriate regulation capacity is pre-determined. However, if the regulation capacity is overestimated, the quality of the FR service provided by the HVAC system and the HVAC energy efficiency will be affected due to the equipment operating constraints. In addition, the AHU supply air temperature fluctuates, and the controller cannot maintain the desired setpoint as the FR magnitude is increased. After the FR service time, the system takes a certain period to return to normal operation. The recovery period extends with the increase in the FR magnitude. For this case study, there exists a control conflict between the FR control and the existing HVAC control (from ASHRAE Guideline 36-2018) during high load conditions, causing the fan speed to saturate. The fan power consumption could increase significantly; up to 77% increase compared to the baseline was observed. This case study demonstrates that the HVAC system controls need to be enhanced and adapted to minimize the interaction between the FR controller and the indoor environmental controls in order to preserve the HVAC system QOS while (during and after) providing satisfactory FR service.

Lu, Xing↗
DOE OSTI2021

Guideline 36 Savings Calculator v0.1

This software package contains a simple calculator that can be used to estimate savings from implementing a retrofit of HVAC controls to ASHRAE Guideline 36 in a multi-zone office building. Implemented in Excel, the calculator has a user-interface front-end to configure parameters of the case study, such as location, building size, and existing control strategies, and a back-end containing results from a large number of pre-run annual energy simulations. The software package also contains the simulation models and associated pre- and post-processing scripts to generate these results.

Blum, David↗
DOE OSTI2023

End-Use Savings Shapes Measure Documentation: Variable Refrigerant Flow with Heat Recovery and Dedicated Outdoor Air System

Building on the successfully completed effort to calibrate and validate the U.S. Department of Energy's ResStock™ and ComStock™ models over the past 3 years, the objective of this work is to produce national data sets that empower analysts working for federal, state, utility, city, and manufacturer stakeholders to answer a broad range of analysis questions. The goal of this work is to develop energy efficiency, electrification, and demand flexibility end-use load shapes (electricity, gas, propane, or fuel oil) that cover a majority of the high-impact, market-ready (or nearly market-ready) measures. "Measures" refers to energy efficiency variables that can be applied to buildings during modeling. An end-use savings shape is the difference in energy consumption between a baseline building and a building with an energy efficiency, electrification, or demand flexibility measure applied. It results in a time-series profile that is broken down by end use and fuel (electricity or on-site gas, propane, or fuel oil use) at each time step. ComStock is a highly granular, bottom-up model that uses multiple data sources, statistical sampling methods, and advanced building energy simulations to estimate the annual subhourly energy consumption of the commercial building stock across the United States. The baseline model intends to represent the U.S. commercial building stock as it existed in 2018. The methodology and results of the baseline model are discussed in the final technical report of the End-Use Load Profiles project. This documentation focuses on a single heating, ventilation, and air-conditioning (HVAC) end-use savings shape measure - a variable refrigerant flow with heat recovery (VRF HR) heating and cooling system coupled with a dedicated outdoor air system (DOAS) for ventilation. This measure replaces existing multi-zone variable air volume (VAV) systems or single-zone rooftop units (RTU) with a VRF HR system coupled with a DOAS that includes an energy/heat recovery ventilator (E/HRV). The measure covers 53% of exisiting building stock's floor area and is not applicable to HVAC system types using district heating or cooling or buildings/spaces that include high-ventilation spaces such as kitchens where the amount of exhaust air is large. A DOAS with E/HRV is used to provide required outdoor ventilation air to spaces since ventilation air is generally not supplied by a VRF HR system. An exhaust air energy recovery ventilator (ERV ) with sensible and latent heat exchange is added to humid climate zones while a heat recovery ventilator (HRV ) with sensible only exchange is added to drier climate zones. The ERV is modeled as a fixed membrane plate counterflow heat exchanger, while the HRV is modeled as a sensible-only fixed aluminum plate counterflow heat exchanger. Both systems include a bypass (for temperature control and economizer lockout) and minimum exhaust temperature control for frost prevention.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
NASA NTRS1992

Analysis of surface roughness generation in aircraft ice accretion

Patterns of roughness evolution have been studied analysis of high magnification video observations of accreting ice surfaces provided by the NASA Lewis Research Center. Three distinct patterns of surface roughness generation have been identified within the parametric regions studied. They include: Rime, Multi-Zone Glaze, and Uniform Glaze. Under most icing conditions, a brief period of transient rime ice growth was observed caused by heat conduction into the body. The resulting thin rime layer explains previously observed insensitivity of some ice accretions to substrate insensitivity of some ice accretions to substrate surface chemistry and may provide justification for simplifying assumptions in ice accretion sailing and modeling effects.

Hansman, R. J., Jr.↗
DOE OSTI2021

Grid-Interactive Multi-Zone Building Control Using Reinforcement Learning with Global-Local Policy Search

In this paper, we develop a grid-interactive multi-zone building controller based on a deep reinforcement learning (RL) approach. The controller is designed to facilitate building operation during normal conditions and demand response events, while ensuring occupants comfort and energy efficiency. We leverage a continuous action space RL formulation, and devise a two-stage global-local RL training framework. In the first stage, a global fast policy search is performed using a gradient-free RL algorithm. In the second stage, a local fine-tuning is conducted using a policy gradient method. In contrast to the state-of-the-art model predictive control (MPC) approach, the proposed RL controller does not require complex computation during real-time operation and can adapt to nonlinear building models. We illustrate the controller performance numerically using a five-zone commercial building.

30 DIRECT ENERGY CONVERSION↗
DOE OSTI2021

Grid-Interactive Multi-Zone Building Control Using Reinforcement Learning with Global-Local Policy Search: Preprint

In this paper, we develop a grid-interactive multi-zone building controller based on a deep reinforcement learning (RL) approach. The controller is designed to facilitate building operation during normal conditions and demand response events, while ensuring occupants comfort and energy efficiency. We leverage a continuous action space RL formulation, and devise a two-stage global-local RL training framework. In the first stage, a global fast policy search is performed using a gradient-free RL algorithm. In the second stage, a local fine-tuning is conducted using a policy gradient method. In contrast to the state-of-the-art model predictive control (MPC) approach, the proposed RL controller does not require complex computation during real-time operation and can adapt to non-linear building models. We illustrate the controller performance numerically using a five-zone commercial building.

30 DIRECT ENERGY CONVERSION↗
DOE OSTI2024

Insights into the broadband emission of the TeV blazar Mrk 501 during the first X-ray polarization measurements

Aims.We present the first multiwavelength study of Mrk 501 that contains simultaneous very-high-energy (VHE)γ-ray observations and X-ray polarization measurements from the Imaging X-ray Polarimetry Explorer (IXPE). Methods.We used radio-to-VHE data from a multiwavelength campaign carried out between March 1, 2022, and July 19, 2022 (MJD 59639 to MJD 59779). The observations were performed by MAGIC,Fermi-LAT,NuSTAR,Swift(XRT and UVOT), and several other instruments that cover the optical and radio bands to complement the IXPE pointings. We characterized the dynamics of the broadband emission around the X-ray polarization measurements through its multiband fractional variability and correlations, and compared changes observed in the polarization degree to changes seen in the broadband emission using a multi-zone leptonic scenario. Results.During the IXPE pointings, the VHE state is close to the average behavior, with a 0.2–1 TeV flux of 20%–50% of the emission of the Crab Nebula. Additionally, it shows low variability and a hint of correlation between VHEγ-rays and X-rays. Despite the average VHE activity, an extreme X-ray behavior is measured for the first two IXPE pointings, taken in March 2022 (MJD 59646 to 59648 and MJD 59665 to 59667), with a synchrotron peak frequency > 1 keV. For the third IXPE pointing, in July 2022 (MJD 59769 to 59772), the synchrotron peak shifts toward lower energies and the optical/X-ray polarization degrees drop. All three IXPE epochs show an atypically low Compton dominance in theγ-rays. The X-ray polarization is systematically higher than at lower energies, suggesting an energy stratification of the jet. While during the IXPE epochs the polarization angles in the X-ray, optical, and radio bands align well, we find a clear discrepancy in the optical and radio polarization angles in the middle of the campaign. Such results further support the hypothesis of an energy-stratified jet. We modeled broadband spectra taken simultaneous to the IXPE pointings, assuming a compact zone that dominates in the X-rays and the VHE band, and an extended zone stretching farther downstream in the jet that dominates the emission at lower energies.NuSTARdata allow us to precisely constrain the synchrotron peak and therefore the underlying electron distribution. The change between the different states observed in the three IXPE pointings can be explained by a change in the magnetization and/or the emission region size, which directly connects the shift in the synchrotron peak to lower energies with the drop in the polarization degree.

Astronomy & Astrophysics↗
DOE OSTI2023

A bi-level data-driven framework for fault-detection and diagnosis of HVAC systems

Long-term operation of heating, ventilation, and air conditioning (HVAC) systems will eventually lead to a range of HVAC system failures, resulting in excessive energy consumption and maintenance costs. Here, to avoid HVAC malfunctioning, fault detection diagnostic (FDD) is utilized as a common practice. Machine learning methods have lately received considerable interest for FDD analysis of HVAC systems due to their high detection accuracy. Meanwhile, HVAC malfunctions are regarded as rare occurrences, hence normal operating data samples are much more accessible than data samples in faulty and malfunctioning conditions. The dominating frequency of normal operation in HVAC datasets has also led to heavily biased classification algorithms within the literature. Moreover, the focus of previous literature has been on increasing the accuracy of the models which leads to a high number of false positives (misleading alarms) in the system. In order to enhance the performance of diagnostic procedures and fill the mentioned gaps, this study proposes a novel data-driven framework. A bi-level machine learning framework is developed for diagnosing faults in air handling units (AHUs) and rooftop units (RTUs) based on principal component analysis (PCA), time series anomaly detection, and random forest (RF). It is shown that PCA can reduce the dataset dimension with one principal component accounting for 95% of data variance. Also, the random forest could classify the faults with 89% precision for single-zone AHU, 85% precision for RTU, and 79% for multi-zone AHU. By proposing this framework, three persistent challenges are addressed: (I) minimizing false positives; (II) accounting for data imbalance; and (III) normal condition monitoring of equipment.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2023

Evaluation of thermostat location for multizone commercial building performance

In multi-zone buildings, it is often found that a single shared thermostat controls more than one conditioned zones. Although these shared zones are supposed to have similar thermal needs (e.g., cooling and heating load), in reality, they are not mainly due to different orientations, sizes of windows, occupancy, space types, etc. This can cause unnecessary energy waste or thermal discomfort for the occupants. How to quantify this impact in multizone buildings remains a research gap. Therefore, this study aims to evaluate the impact of different sensor (i.e., thermostat) locations for multizone commercial buildings through a comprehensive modeling study. Here, two different scenarios for the sensor locations were selected to evaluate the impact in terms of energy and thermal comfort. The scenario (1) is that one to five sensors distributed among the five zones, but the sensor readings from selected zones will be used for no-sensor zones, which is no-mean sensor scenario. The scenario (2) is that one to five sensors distributed among the five zones, but the average temperature from the shared zones will be used for each of the shared zones, which is a mean sensor scenario. The uncertainty analysis was performed for different sensor location scenarios.(a)The major findings from an energy perspective, for scenario (1), the differences of cooling energy go as high as 17% more or 12% less, compared with the baseline. For heating energy consumption, the discrepancies go as high as 51% more or 52% less, compared with baseline. For site energy consumption, the discrepancies go as high as 3.2% more or 3.2% less, compared with baseline. For fan energy consumption, the discrepancies go as high as 3.2% more or as low as 1.0% less, compared with baseline. For scenario (2), the discrepancies of cooling energy go as high as 3% more, or 0.5% less, compared with the baseline. For heating energy consumption, the discrepancies, are go as high as 10.1% more or as low as 3.0% less, compared with baseline. For site energy consumption, the discrepancies go as high as 1.3% more or 0.3% less, compared with baseline. For fan energy consumption, the discrepancies go as high as 3.1% more or as low as 1.0% less, compared with baseline.(b) In terms of the indoor thermal comfort, for the no-mean-sensor scenarios, the discrepancies of unmet hours for cooling mode can be as high as 1,200 h, compared with the baseline. The discrepancies of unmet hours for heating mode can be as high as 740 h, compared with the baseline. For the mean-sensor scenarios, the discrepancies of unmet hours for cooling mode can be as high as 750 h, compared with the baseline. The discrepancies of unmet hours for heating mode can be as high as 50 h, compared with the baseline.

42 ENGINEERING↗
DOE OSTI2020

Smart Ventilation for Advanced California Homes

This project investigated smart ventilation approaches to minimize energy use for providing indoor air quality (IAQ) in high performance new California homes. Evaluation criteria included annual ventilation-related energy, peak energy and time-of-use savings, and the indoor air quality relative to a minimally code-compliant ventilation system. The simulations used CONTAM’s air flow and contaminant transport model, combined with the EnergyPlus building loads model. House types representing the default California Energy Code compliance homes were investigated for four California climate zones, covering a wide range of climate types. Both single and multi-zone smart ventilation controls were investigated. Contaminant sources included contaminants emitted continuously and varying with time, temperature and relative humidity, episodic emissions from occupant activities and outdoor particles. Single-zone ventilation controls that varied ventilation depending on outdoor temperatures were able to consistently save half of ventilation-related energy without compromising long-term IAQ. Ventilation strategies that tracked occupancy were less successful, because this work included generic contaminants with constant background emission rates. Energy performance for occupancy controls improved with a one-hour pre-occupancy flush out strategy. The addition of zoning ventilation controls did not offer significant IAQ to energy improvements compared to non-zonal versions of the same ventilation system type. The best controls had HVAC energy savings of 10-20%, with individual cases reaching up to 40% savings. However, these savings cannot be achieved without worsening personal exposures for at least one contaminant. A metric is needed to assess the competing changes in exposure to different contaminants in order to determine the net-health impacts of a control strategy. Controls that directly sensed contaminants and controlled them to acceptable levels showed that the California OEHHA limit for formaldehyde completely dominates system performance, with homes not able to meet the limit even with continuous operation of a fan sized to twice the current code minimum.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2022

Sensor Impact Evaluation at Different Sensor Locations in a Multi-zone Office Building

FY 2022 Q1 deliverables include the development of an emulator that can evaluate the sensor impacts at different sensor locations (i.e., thermostat locations) in a multizone office building. This report presents the detailed procedure of developing the emulator and using the emulator for preliminary sensor impact analysis. In designing new multizone buildings or retrofitting existing buildings, the room thermostat locations or subzoning design has been often determined by best practices without considering the effects of this design in terms of energy or thermal comfort. In subzoning design, the total number of thermostats is usually smaller than the total number of rooms. As a result, one thermostat in one room often controls the indoor temperature of several other adjacent rooms. For example, five zones might share one thermostat located in one of the zones. Because the demands for thermal load in different zones could be different for the multizone buildings, this subzoning design can compromise control performance and waste building energy consumption. Furthermore, for multizone buildings, subzoning could introduce thermal discomfort for zones. This issue has not be thoroughly investigated in simulation/field studies, and the US Department of Energy’s Oak Ridge National Laboratory explored the impacts of subzoning design through modeling and experimental study.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
NASA NTRS1993

CFD Parametric Study of Consortium Impeller

Current design of high performance turbopumps for rocket engines requires effective and robust analytical tools to provide design impact in a productive manner. The main goal of this study is to develop a robust and effective computational fluid dynamics (CFD) pump model for general turbopump design and analysis applications. A Finite Difference Navier-Stokes flow solver, FDNS, which includes the extended k-epsilon turbulence model and appropriate moving interface boundary conditions, was developed to analyze turbulent flows in turbomachinery devices. A second-order central difference scheme plus adaptive dissipation terms was employed in the FDNS code, along with a predictor plus multi-corrector pressure-based solution procedure. The multi-zone, multi-block capability allows the FDNS code to efficiently solve flow fields with complicated geometry. The FDNS code has been benchmarked by analyzing the pump consortium inducer, and it provided satisfactory results. In the present study, a CFD parametric study of the pump consortium impeller was conducted using the FDNS code. The pump consortium impeller, with partial blades, is a new design concept of the advanced rocket engines. The parametric study was to analyze the baseline design of the consortium impeller and its modification which utilizes TANDEM blades. In the present study, the TANDEM blade configuration of the consortium impeller considers cut full blades for about one quarter chord length from the leading edge and clocks the leading edge portion with an angle of 7.5 or 22.5 degrees. The purpose of the present study is to investigate the effect and trend of the TANDEM blade modification and provide the result as a design guideline. A 3-D flow analysis, with a 103 x 23 x 30 mesh grid system and with the inlet flow conditions measured by Rocketdyne, was performed for the baseline consortium impeller. The numerical result shows that the mass flow rate splits through various blade passages are relatively uniform. Due to the complexity of blade geometries, the TANDEM blade configurations were analyzed with the multi-zone grid structure. Both the 7.5 deg- and the 22.5 deg-clocking TANDEM blade cases utilized a 80K mesh system. The numerical result of two TANDEM blade modifications indicates the efficiency and the head are worse than those of the baseline case due to larger flow distortion. The gap between the TANDEM blade and the full blade allows the flow passes through and heavily loads the pressure side of the partial blade such that flow reversal occurs near the suction side of the splitter. The flow split at the exit of impeller blades is very non-uniform for TANDEM blade cases, and this will greatly induce the side load on the diffuser. Therefore, the TANDEM blade modification in the present CFD analysis does not improve the performance of the consortium impeller.

Cheng, Gary C.↗
NASA NTRS1993

Advanced Multi-Phase Flow CFD Model Development for Solid Rocket Motor Flowfield Analysis

It is known that the simulations of solid rocket motor internal flow field with AL-based propellants require complex multi-phase turbulent flow model. The objective of this study is to develop an advanced particulate multi-phase flow model which includes the effects of particle dynamics, chemical reaction and hot gas flow turbulence. The inclusion of particle agglomeration, particle/gas reaction and mass transfer, particle collision, coalescence and breakup mechanisms in modeling the particle dynamics will allow the proposed model to realistically simulate the flowfield inside a solid rocket motor. The Finite Difference Navier-Stokes numerical code FDNS is used to simulate the steady-state multi-phase particulate flow field for a 3-zone 2-D axisymmetric ASRM model and a 6-zone 3-D ASRM model at launch conditions. The 2-D model includes aft-end cavity and submerged nozzle. The 3-D model represents the whole ASRM geometry, including additional grain port area in the gas cavity and two inhibitors. FDNS is a pressure based finite difference Navier-Stokes flow solver with time-accurate adaptive second-order upwind schemes, standard and extended k-epsilon models with compressibility corrections, multi zone body-fitted formulations, and turbulence particle interaction model. Eulerian/Lagrangian multi-phase solution method is applied for multi-zone mesh. To simulate the chemical reaction, penalty function corrected efficient finite-rate chemistry integration method is used in FDNS. For the AL particle combustion rate, the Hermsen correlation is employed. To simulate the turbulent dispersion of particles, the Gaussian probability distribution with standard deviation equal to (2k/3)(exp 1/2) is used for the random turbulent velocity components. The computational results reveal that the flow field near the juncture of aft-end cavity and the submerged nozzle is very complex. The effects of the turbulent particles affect the flow field significantly and provide better prediction of the ASRM performance. The multi-phase flow analysis using the FDNS code in the present research can be used as a design tool for solid rocket motor applications.

Liaw, Paul↗
NASA NTRS2000

Use Computer-Aided Tools to Parallelize Large CFD Applications

Porting applications to high performance parallel computers is always a challenging task. It is time consuming and costly. With rapid progressing in hardware architectures and increasing complexity of real applications in recent years, the problem becomes even more sever. Today, scalability and high performance are mostly involving handwritten parallel programs using message-passing libraries (e.g. MPI). However, this process is very difficult and often error-prone. The recent reemergence of shared memory parallel (SMP) architectures, such as the cache coherent Non-Uniform Memory Access (ccNUMA) architecture used in the SGI Origin 2000, show good prospects for scaling beyond hundreds of processors. Programming on an SMP is simplified by working in a globally accessible address space. The user can supply compiler directives, such as OpenMP, to parallelize the code. As an industry standard for portable implementation of parallel programs for SMPs, OpenMP is a set of compiler directives and callable runtime library routines that extend Fortran, C and C++ to express shared memory parallelism. It promises an incremental path for parallel conversion of existing software, as well as scalability and performance for a complete rewrite or an entirely new development. Perhaps the main disadvantage of programming with directives is that inserted directives may not necessarily enhance performance. In the worst cases, it can create erroneous results. While vendors have provided tools to perform error-checking and profiling, automation in directive insertion is very limited and often failed on large programs, primarily due to the lack of a thorough enough data dependence analysis. To overcome the deficiency, we have developed a toolkit, CAPO, to automatically insert OpenMP directives in Fortran programs and apply certain degrees of optimization. CAPO is aimed at taking advantage of detailed inter-procedural dependence analysis provided by CAPTools, developed by the University of Greenwich, to reduce potential errors made by users. Earlier tests on NAS Benchmarks and ARC3D have demonstrated good success of this tool. In this study, we have applied CAPO to parallelize three large applications in the area of computational fluid dynamics (CFD): OVERFLOW, TLNS3D and INS3D. These codes are widely used for solving Navier-Stokes equations with complicated boundary conditions and turbulence model in multiple zones. Each one comprises of from 50K to 1,00k lines of FORTRAN77. As an example, CAPO took 77 hours to complete the data dependence analysis of OVERFLOW on a workstation (SGI, 175MHz, R10K processor). A fair amount of effort was spent on correcting false dependencies due to lack of necessary knowledge during the analysis. Even so, CAPO provides an easy way for user to interact with the parallelization process. The OpenMP version was generated within a day after the analysis was completed. Due to sequential algorithms involved, code sections in TLNS3D and INS3D need to be restructured by hand to produce more efficient parallel codes. An included figure shows preliminary test results of the generated OVERFLOW with several test cases in single zone. The MPI data points for the small test case were taken from a handcoded MPI version. As we can see, CAPO's version has achieved 18 fold speed up on 32 nodes of the SGI O2K. For the small test case, it outperformed the MPI version. These results are very encouraging, but further work is needed. For example, although CAPO attempts to place directives on the outer- most parallel loops in an interprocedural framework, it does not insert directives based on the best manual strategy. In particular, it lacks the support of parallelization at the multi-zone level. Future work will emphasize on the development of methodology to work in a multi-zone level and with a hybrid approach. Development of tools to perform more complicated code transformation is also needed.

Jin, H.↗
NASA NTRS2003

Accelerating Large Data Analysis By Exploiting Regularities

We present techniques for discovering and exploiting regularity in large curvilinear data sets. The data can be based on a single mesh or a mesh composed of multiple submeshes (also known as zones). Multi-zone data are typical to Computational Fluid Dynamics (CFD) simulations. Regularities include axis-aligned rectilinear and cylindrical meshes as well as cases where one zone is equivalent to a rigid-body transformation of another. Our algorithms can also discover rigid-body motion of meshes in time-series data. Next, we describe a data model where we can utilize the results from the discovery process in order to accelerate large data visualizations. Where possible, we replace general curvilinear zones with rectilinear or cylindrical zones. In rigid-body motion cases we replace a time-series of meshes with a transformed mesh object where a reference mesh is dynamically transformed based on a given time value in order to satisfy geometry requests, on demand. The data model enables us to make these substitutions and dynamic transformations transparently with respect to the visualization algorithms. We present results with large data sets where we combine our mesh replacement and transformation techniques with out-of-core paging in order to achieve significant speed-ups in analysis.

Moran, Patrick J.↗