Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Computational Complexity”

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 613 records · Page 34

A source-attractor approach to network detection of radiation sources

Radiation source detection using a network of detectors is an active field of research for homeland security and defense applications. We propose Source-attractor Radiation Detection (SRD) method to aggregate measurements from a network of detectors for radiation source detection. SRD method models a potential radiation source as a magnet -like attractor that pulls in pre-computed virtual points from the detector locations. A detection decision is made if a sufficient level of attraction, quantified by the increase in the clustering of the shifted virtual points, is observed. Compared with traditional methods, SRD has the following advantages: i) it does not require an accurate estimate of the source location from limited and noise-corrupted sensor readings, unlike the localizationbased methods, and ii) its virtual point shifting and clustering calculation involve simple arithmetic operations based on the number of detectors, avoiding the high computational complexity of grid-based likelihood estimation methods. We evaluate its detection performance using canonical datasets from Domestic Nuclear Detection Office s (DNDO) Intelligence Radiation Sensors Systems (IRSS) tests. SRD achieves both lower false alarm rate and false negative rate compared to three existing algorithms for network source detection.

Wu, Qishi↗

8760-Based Method for Representing Variable Generation Capacity Value in Capacity Expansion Models

Capacity expansion models (CEMs) are widely used to evaluate the least-cost portfolio of electricity generators, transmission, and storage needed to reliably serve load over many years or decades. CEMs can be computationally complex and are often forced to estimate key parameters using simplified methods to achieve acceptable solve times or for other reasons. In this paper, we discuss one of these parameters -- capacity value (CV). We first provide a high-level motivation for and overview of CV. We next describe existing modeling simplifications and an alternate approach for estimating CV that utilizes hourly '8760' data of load and VG resources. We then apply this 8760 method to an established CEM, the National Renewable Energy Laboratory's (NREL's) Regional Energy Deployment System (ReEDS) model (Eurek et al. 2016). While this alternative approach for CV is not itself novel, it contributes to the broader CEM community by (1) demonstrating how a simplified 8760 hourly method, which can be easily implemented in other power sector models when data is available, more accurately captures CV trends than a statistical method within the ReEDS CEM, and (2) providing a flexible modeling framework from which other 8760-based system elements (e.g., demand response, storage, and transmission) can be added to further capture important dynamic interactions, such as curtailment.

14 SOLAR ENERGY↗

Two-level Clustering-based Target Detection Through Sensor Deployment and Data Fusion

Target detection is one fundamental problem in many sensor network-based applications, which is typically tackled in two separate stages for sensor deployment and data fusion. We propose an integrated solution, referred to as 2C-SSEM,which combines 2-level clustering-based sensor deployment and source strength estimate map-based data fusion for the detection of a single static or moving target. 2C-SSEM conducts the first level of clustering to determine a sensor deployment scheme and the second level of clustering to divide the deployed sensors into several subsets. For each sensor, the source strength is estimated at each grid point of the entire region based on a signal attenuation model, and for each subset of sensors, the target location is estimated using a distribution map-based statistical analysis method. A final detection decision is made by thresholding the clustering degree of the target location estimates computed by all subsets of sensors. Compared with traditional grid-based target detection methods, 2C-SSEM significantly reduces the computation complexity and improves the detection performance through an integrated optimization strategy. Extensive simulation results show the performance superiority of the proposed solution over several well-known methods for target detection.

Rao, Nageswara↗

Model-free Control of Building HVAC Systems to Accommodate Solar Photovoltaic Energy

With the increasing demand for energy in day-today life, the need for integrating renewable energy is growing. In this paper, we take a step in addressing the issue by employing solar photovoltaic (PV) energy to maintain indoor temperatures in buildings. The objective we aim to meet is to use as much of the generated PV energy as possible while maintaining the indoor comfort temperatures. We set this objective in order to reduce wastage of renewable energy and cut down the need for storage devices. We formulate a thermal model of the building HVAC system and employ the model-free control (MFC) technique to achieve the defined objective. Although recently introduced, MFC has shown promising results in various domains and it is more appealing due to its light computational complexity. Furthermore, we analyze the relationship between the generated PV energy and the total energy consumption of the buildings. The results show that the aforementioned relationship is linear in nature as the number of buildings increases.

Telsang, Bhagyashri↗

Learning Deep Neural Network Representations for Koopman Operators of Nonlinear Dynamical Systems

The Koopman operator has recently garnered much attention for its value in dynamical systems analysis and data-driven model discovery. However, its application has been hindered by the computational complexity of extended dynamic mode decomposition; this requires a combinatorially large basis set to adequately describe many nonlinear systems of interest, e.g. cyber-physical infrastructure systems, biological networks, social systems, and fluid dynamics. Often the dictionaries generated for these problems are manually curated, requiring domain-specific knowledge and painstaking tuning. In this paper we introduce a deep learning framework for learning Koopman operators of nonlinear dynamical systems. We show that this novel method automatically selects efficient deep dictionaries, outperforming state-of-the-art methods. We benchmark this method on partially observed nonlinear systems, including the glycolytic oscillator and show it is able to predict quantitatively 100 steps into the future, using only a single timepoint, and qualitative oscillatory behavior 400 steps into the future.

koopman operators, machine learning, deep learning↗

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↗

Reduced Order Model to Predict Dispersion of Flammable Refrigerant into a Space

As the HVAC&R industry mobilizes to deploy more low-GWP refrigerants, relevant standards are being continually reviewed and updated. Those include the general safety standards ISO 5149 and ASHRAE 15, and the equipment standards IEC and UL. The standards systematically set the allowable maximum amount of refrigerant that should be used in different equipment types and different applications. To do so, they rely on predictions of how a leaked refrigerant mass will disperse into a space. Dispersion characteristics, such as total flammable volume and its residence time, determine the risk associated with the presence of the flammable refrigerant. The standards have included provisions for the use of flammable refrigerants for approximately two decades. They relied on limited analytical analyses and test cases in their development. Dispersion of a refrigerant into a space is complex. Computational fluid dynamics (CFD) are the most accurate in predicting a given problem. However, CFD is computationally expensive and requires specialized expertise and resources and is not suitable for use by standards development working group as prediction tool. This paper presents the development of a reduced order model (ROM) that predicts the key dispersion characteristics relevant to the dispersion of a leaked refrigerant into a space for any combination of input variables. The inputs are the refrigerant release height, the total released refrigerant mass and its release flow rate, the refrigerant molecular weight, the ventilation flow rate, the floor area and height of the space, recirculation air flow rate, and the tightness of the space. The outputs are histograms of volume fraction of the room in prescribed concentration bins and the total mass of the refrigerant in each bin normalized by the total refrigerant charge at 13 prescribed simulation time stamps between 1 and 900 seconds. The ROM is constructed from a set of CFD simulations with carefully chosen combinations of input parameters. The selection if done using a multidimensional sparse grid which is a generalization of the classical tensor approach but offers additional flexibility and thus can be more carefully tuned towards a specific model. The tuning is done to improve the accuracy, measured in the difference between the output values of the ROM and the CFD model, while minimizing the computational cost, measured in number of CFD simulations which is orders of magnitude more expensive than the processing the training data.

Edwards, Dean↗

A Novel Multi-Agent Deep Reinforcement Learning-enabled Distributed Power Allocation Scheme for mmWave Cellular Networks

We consider the power allocation problem over shared spectrum for millimeter-Wave (mmWave) cellular downlink. Existing approaches usually find sub-optimal solutions by solving a non-convex optimization which leads to scalability issues due to centralized control. Therefore, distributed and adaptive approaches are desirable. Recently, model-free Deep Reinforcement Learning (DRL) has achieved success in such wireless resource management tasks. By modeling the radio environment as a Markov Decision Process (MDP) with the base stations (BSs) being the agents, power allocation can be automated at the agent level with comparable throughput performance to conventional centralized schemes. The multi-agent setting presents new challenges as the radio environment is impacted by the joint actions of the agents and is no longer stationary from any individual agent’s perspective. Existing literature bypasses this non-stationarity violation by ignoring it which may cause performance degradation. To tackle this issue, we propose a distributed continuous power allocation scheme based on a modified version of multi-agent Deep Deterministic Policy Gradient (MADDPG) that is tailored for the distributed multiple-agent setting. The proposed scheme employs a centralized-training distributed-execution framework where Q-functions are trained over subsets of BSs while each BS determines its transmit power based only on its own local observation. It admits constant per-BS communication and computation complexity and is thus scalable to large networks. Numerical evaluation shows that the proposed scheme adapts well to a wide range of interference conditions and can achieve comparable or better performance than several state-of-the-art non-learning approaches.

99 GENERAL AND MISCELLANEOUS↗

Level 1 transient model for a molybdenum-99 producing aqueous homogeneous reactor and its applicability to the tracy reactor

Babcock and Wilcox Technical Services Group (B and W) has identified aqueous homogeneous reactors (AHRs) as a technology well suited to produce the medical isotope molybdenum 99 (Mo-99). AHRs have never been specifically designed or built for this specialized purpose. However, AHRs have a proven history of being safe research reactors. In fact, in 1958, AHRs had 'a longer history of operation than any other type of research reactor using enriched fuel' and had 'experimentally demonstrated to be among the safest of all various type of research reactor now in use [1].' A 'Level 1' model representing B and W's proposed Medical Isotope Production System (MIPS) reactor has been developed. The Level 1 model couples a series of differential equations representing neutronics, temperature, and voiding. Neutronics are represented by point reactor kinetics while temperature and voiding terms are axially varying (one-dimensional). While this model was developed specifically for the MIPS reactor, its applicability to the Japanese TRACY reactor was assessed. The results from the Level 1 model were in good agreement with TRACY experimental data and found to be conservative over most of the time domains considered. The Level 1 model was used to study the MIPS reactor. An analysis showed the Level 1 model agreed well with a more complex computational model of the MIPS reactor (a FETCH model). Finally, a significant reactivity insertion was simulated with the Level 1 model to study the MIPS reactor's time-dependent response. (authors)

07 ISOTOPE AND RADIATION SOURCES↗

Extending Power of Nature from Binary Problems to Real-Valued Graph Learning in Real World

Nature performs complex computations constantly at clearly lower cost and higher performance than digital computers. It is crucial to understand how to harness the unique computational power of nature in Machine Learning (ML). In the past decade, besides the development of Neural Networks (NNs), the community has also relentlessly explored nature-powered ML paradigms. Although most of them are still predominantly theoretical, a new practical paradigm enabled by the recent advent of CMOS-compatible room-temperature nature-based computers has emerged. By harnessing the nature's power of entropy increase, this paradigm can solve binary learning problems delivering immense speedup and energy savings compared with NNs, while maintaining comparable accuracy. Regrettably, its values to the real world are highly constrained by its binary nature. A clear pathway to its extension to real-valued problems remains elusive. This paper aims to unleash this pathway by proposing a novel end-to-end Nature-Powered Graph Learning (NP-GL) framework. Specifically, through a three-dimensional co-design, NP-GL can leverage the nature's power of entropy increase to efficiently solve real-valued graph learning problems. Experimental results across 4 real-world applications with 6 datasets demonstrate that NP-GL delivers, on average, 6970X speedup and 10^5x energy consumption reduction with comparable or even higher accuracy than Graph Neural Networks (GNNs).

artificial intelligence↗

A Fast Dynamic Internal Predictive Power Scheduling Approach for Power Management in Microgrids: Preprint

This paper presents a Dynamic Internal Predictive Power Scheduling (DIPPS) approach for optimizing power management in microgrids, particularly focusing on external power exchanges among diverse prosumers. DIPPS utilizes a dynamic objective function with a time-varying binary parameter to control the timing of power transfers to the external grid, facilitated by efficient usage of energy storage for surplus renewable power. The microgrid power scheduling problem is modeled as a mixed-integer nonlinear programming (MINLP-PS) and subsequently transformed into a mixed-integer linear programming (MILPPS) optimization through McCormick's relaxation to reduce computational complexity. A predictive window window with 6 data points is solved at an average of 0.92s, a 97.6% improvement over the 38.27s required for the MINLP-PS formulation, implying the numerical feasibility of the DIPPS approach for real-time implementation. Finally, the approach is validated against a static objective using real-world load data across three case studies with different time-varying parameters, demonstrating the ability of DIPPS to optimize power exchanges and efficiently utilize distributed resources while shifting the external power transfers to specified time durations.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Sub-microsecond Transformers for Jet Tagging on FPGAs

We present the first sub-microsecond transformer implementation on an FPGA achieving competitive performance for state-of-the-art high-energy physics benchmarks. Transformers have shown exceptional performance on multiple tasks in modern machine learning applications, including jet tagging at the CERN Large Hadron Collider (LHC). However, their computational complexity prohibits use in real-time applications, such as the hardware trigger system of the collider experiments up until now. In this work, we demonstrate the first application of transformers for jet tagging on FPGAs, achieving $\mathcal{O}(100)$ nanosecond latency with superior performance compared to alternative baseline models. We leverage high-granularity quantization and distributed arithmetic optimization to fit the entire transformer model on a single FPGA, achieving the required throughput and latency. Furthermore, we add multi-head attention and linear attention support to hls4ml, making our work accessible to the broader fast machine learning community. This work advances the next-generation trigger systems for the High Luminosity LHC, enabling the use of transformers for real-time applications in high-energy physics and beyond.

Laatu, Lauri [Imperial Coll., London]↗

Complexity analysis of pipeline mapping problems in distributed heterogeneous networks

Largescale scientific applications require using various system resources to execute complex computing pipelines in distributed networks to support collaborative research. System resources are typically shared in the Internet or over dedicated connections based on their location, availability, capability, and capacity. Optimizing the network performance of computing pipelines in such distributed environments is critical to the success of these applications. We consider two types of largescale distributed applications: (1) interactive applications where a single dataset is sequentially processed along a pipeline; and (2) streaming applications where a series of datasets continuously flow through a pipeline. The computing pipelines of these applications consist of a number of modules executed in a linear order in network environments with heterogeneous resources under different constraints. Our goal is to find an efficient mapping scheme that allocates the modules of a pipeline to network nodes for minimum endtoend delay or maximum frame rate. We formulate the pipeline mappings in distributed environments as optimization problems and categorize them into six classes with different optimization goals and mapping constraints: (1) Minimum Endtoend Delay with No Node Reuse (MEDNNR), (2) Minimum Endtoend Delay with Contiguous Node Reuse (MEDCNR), (3) Minimum Endtoend Delay with Arbitrary Node Reuse (MEDANR), (4) Maximum Frame Rate with No Node Reuse or Share (MFRNNRS), (5) Maximum Frame Rate with Contiguous Node Reuse and Share (MFRCNRS), and (6) Maximum Frame Rate with Arbitrary Node Reuse and Share (MFRANRS). Here, 'contiguous node reuse' means that multiple contiguous modules along the pipeline may run on the same node and 'arbitrary node reuse' imposes no restriction on node reuse. Note that in interactive applications, a node can be reused but its resource is not shared. We prove that MEDANR is polynomially solvable and the rest are NP-complete. MEDANR, where either contiguous or noncontiguous modules in the pipeline can be mapped onto the same node, is essentially the Maximum n-hop Shortest Path problem, and can be solved using a dynamic programming method. In MEDNNR and MFRNNRS, any network node can be used only once, which requires selecting the same number of nodes for onetoone onto mapping. We show its NP-completeness by reducing from the Hamiltonian Path problem. Node reuse is allowed in MEDCNR, MFRCNRS and MFRANRS, which are similar to the Maximum n-hop Shortest Path problem that considers resource sharing. We prove their NP-completeness by reducing from the Disjoint-Connecting-Path Problem and Widest path with the Linear Capacity Constraints problem, respectively.

99 GENERAL AND MISCELLANEOUS↗

GDSA framework, a computational framework for complex modeling problems in radioactive waste management

This paper details a computational framework to produce automated, graphical workflows, and how this framework can be deployed to support complex modeling problems like those in nuclear engineering. Key benefits of the framework include: automating previously manual workflows; intuitive construction and communication of workflows through a graphical interface; and automated file transfer and handling for workflows deployed across heterogeneous computing resources. This paper demonstrates the framework's application to probabilistic post-closure performance assessment of systems for deep geologic disposal of nuclear waste. However, the framework is a general capability that can help users running a variety of computational studies.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Computation of the Complex Probability Function

The complex probability function is important in many areas of physics and many techniques have been developed in an attempt to compute it for some z quickly and e ciently. Most prominent are the methods that use Gauss-Hermite quadrature, which uses the roots of the n th degree Hermite polynomial and corresponding weights to approximate the complex probability function. This document serves as an overview and discussion of the use, shortcomings, and potential improvements on the Gauss-Hermite quadrature for the complex probability function.

97 MATHEMATICS AND COMPUTING↗

Computational Fluid Dynamics (CFD) and Safety Factors: Computer modeling of complex processes needs old-fashioned experiments to stay in touch with reality

Computational fluid dynamics (CFD) is recognized as a powerful engineering tool. That is, CFD has advanced over the years to the point where it can now give us deep insight into the analysis of very complex processes. There is a danger, though, that an engineer can place too much confidence in a simulation. If a user is not careful, it is easy to believe that if you plug in the numbers, the answer comes out, and you are done. This assumption can lead to significant errors. As we discovered in the course of a study on behalf of the Department of Energy's Savannah River Site in South Carolina, CFD models fail to capture some of the large variations inherent in complex processes. These variations, or scatter, in experimental data emerge from physical tests and are inadequately captured or expressed by calculated mean values for a process. This anomaly between experiment and theory can lead to serious errors in engineering analysis and design unless a correction factor, or safety factor, is experimentally validated. For this study, blending times for the mixing of salt solutions in large storage tanks were the process of concern under investigation. This study focused on the blending processes needed to mix salt solutions to ensure homogeneity within waste tanks, where homogeneity is required to control radioactivity levels during subsequent processing. Two of the requirements for this task were to determine the minimum number of submerged, centrifugal pumps required to blend the salt mixtures in a full-scale tank in half a day or less, and to recommend reasonable blending times to achieve nearly homogeneous salt mixtures. A full-scale, low-flow pump with a total discharge flow rate of 500 to 800 gpm was recommended with two opposing 2.27-inch diameter nozzles. To make this recommendation, both experimental and CFD modeling were performed. Lab researchers found that, although CFD provided good estimates of an average blending time, experimental blending times varied significantly from the average.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗