SEARCH · Search NASA

Results for “optimal scheduling”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 343 records · Page 19

All sources NASA NTRS DOE OSTI

DOE OSTI2024

RAP: Resource-aware Automated GPU Sharing for Multi-GPU Recommendation Model Training and Input Preprocessing

Ensuring high-quality recommendations for newly onboarded users requires the continuous retraining of Deep Learning Recommendation Models (DLRMs) with freshly generated data. To serve the online DLRM retraining, existing solutions use hundreds of CPU computing nodes designated for input preprocessing, causing significant power consumption that surpasses even the power usage of GPU trainers. To this end, we propose RAP, an end-to-end DLRM training framework that supports Resource-aware Automated GPU sharing for DLRM input Preprocessing and Training. The core idea of RAP is to accurately capture the remaining GPU computing resources during DLRM training for input preprocessing, achieving superior training efficiency without requiring additional resources. Specifically, RAP utilizes a co-running cost model to efficiently assess the costs of various input preprocessing operations, and it implements a resource-aware horizontal fusion technique that adaptively merges smaller kernels according to GPU availability, circumventing any interference with DLRM training. In addition, RAP leverages a heuristic searching algorithm that jointly optimizes both the input preprocessing graph mapping and the co-running schedule to maximize the end-to-end DLRM training throughput. The comprehensive evaluation shows that RAP achieves 78.3× speedup on average over CPU-based DLRM input preprocessing frameworks. In addition, the end-to-end training throughput of RAP is only 2.04% lower than the ideal case, which has no input preprocessing overhead.

Wang, Zheng↗

DOE OSTI2020

Advanced Tri-lab Software Environment (ATSE)

The Advanced Tri-lab Software Environment (ATSE) is an effort to build an open, modular, extensible, community-engaged, and vendor-adaptable software ecosystem that enables the prototyping of new technologies for improving the ASC computing environment. The initial target for ATSE is to accelerate the maturity of the Arm ecosystem for supporting ASC computing and the high-performance computing community more broadly. ATSE provides an integrated and optimized software stack that includes: 1) Application development environment and libraries including compilers, math libraries, tools, MPI, and OpenMP, 2) Low-level system software including optimized Linux, network stack, file systems, containers, and virtual machines, 3) Job scheduling and management including workload manager, application launcher, and user tools, and 4) System administration and management tools supporting booting, monitoring, and operating system image management. SAND2020-12377 M Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

Younge, Andrew J.↗

DOE OSTI2021

Hamilton: Flexible, Open Source $10 Wireless Sensor System for Energy Efficient Building Operation

Sensors for improving building performance are rapidly populating the market, driven in part by the drive to reduce greenhouse gas emissions resulting from energy production as well as improve the interior environment for healthy and more productive spaces. UC Berkeley has led wireless sensor development over the past 25 years (e.g., Telos mote), with the Hamilton (named after Alexander Hamilton on the US $10 bill) as the most recent. The Hamilton sensor was designed as a low-cost high-performance sensor that is modular and interoperable. The objective of the Hamilton project was to create, evaluate and establish the technological foundations for secure and easy to deploy building energy efficiency applications utilizing pervasive, low-cost wireless sensors integrated with traditional Building Management Systems (BMS), consumer-sector building components, and powerful data analytics. The project included iterative hardware design, incorporating a high-performance database (BTrDb, http://btrdb.io/), creating and iterating the development of secure data middleware (BOSSwave, WAVE/WAVEMQ), working with and pushing the development of an open-source tiny operating system RiotOS, and implementing and improving protocols such as Thread/OpenThread and TCP/IP. The hardware benefited from careful design to drive down the cost; the design included a System-on-a-Chip (SoC), chip antenna, single crystal and five passive components. Careful design of the operating system created a low-power design to enable a long life with small batteries. The hardware included several sensors: temperature, radiant temperature, relative humidity, magnetometer, accelerometer, and light, with an optional occupancy (Passive InfraRed) sensor. The project was the basis of several applications, both internal to the research team and other researchers and professionals at other institutions. Several applications used the sensor hardware as the basis for other complex devices. Other applications used the sensors to improve building performance through interoperating with the building Heating Ventilation and Air-Conditioning (HVAC) system, such as using occupancy and/or distributed temperature sensing to reduce HVAC zone energy while still providing thermal comfort and to reduce peak loads in small commercial buildings. We demonstrated cloud-based energy analytics, implemented a schedule and a Model Predictive Controller in a small commercial building to optimize HVAC energy, occupancy and electricity price. Initial integration of these technological innovations was performed through the creation of execution containers containing the WAVE agent and various driver, proxy, or building system function logic. The research added to the understanding of efficient sensor hardware, secure middleware, time-series data management (high performance database), efficient communication protocols, and interoperating with applications and building systems. The project showed the technical effectiveness and economic feasibility of creating a low-cost, modular, and easy-to-deploy sensor. Through conversations with multiple end users, the research team discovered that many customers wanted data management and services in addition to the sensors. HamiltonIOT developed packages of sensors, border router, and data services to provide a seamless “plug-and-play” sensor deployment. Some customers were willing to pay for higher quality sensors (such as light); some customers wanted a robust enclosure (waterproof).

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2022

Options for Subscale Maturation of Advanced Reactor Technologies Testing for Nuclear Thermal Propulsion

Several options could be implemented to establish an irradiation testing capability suitable for investigation of the performance of multiple nuclear thermal propulsion fuel elements at prototypic conditions. The prototypic conditions of interest are based on the current needs of the National Aeronautics and Space Administration’s Space Nuclear Power Program. The results of such testing are also intended to reduce the risks currently seen for any future subscale or full-scale ground testing of an engine-reactor system. The optimal solution is dependent upon several factors such as performance, cost, availability, schedule, technology readiness level, and plans for future testing in the SNP Program. Three options, based on different combinations of these factors, are considered in this report.

33 ADVANCED PROPULSION SYSTEMS↗

DOE OSTI2022

Use Probabilistic Forecasts in Reliable and Economic Electricity Market Scheduling and Operations

The value of probabilistic forecasts in electricity market operations is being increasingly recognized lately, however, the use is still limited. This study demonstrates two cases of using probabilistic forecasts in scheduling and operations of the California electricity market. We first develop a data-driven method to give weather-informed estimates of ramping reserves based on short-term probabilistic solar irradiance forecasts. By using forecasts across multiple sites, our proposed method presents advantages over the real-world baseline in terms of system reliability and economics. Our second case uses probabilistic hydro power forecasts in the scheduling of a hydro power plant under economic and environmental constraints. By optimizing the bids into both the day-ahead and real-time markets, our results demonstrate considerable economic benefits.

13 HYDRO ENERGY↗

DOE OSTI2021

Improving Performance of Tape Restore Request Scheduling in the Storage System dCache

Given the anticipated increase in the amount of scientific data, it is widely accepted that primarily disk based storage will become prohibitively expensive. Tape based storage, on the other hand, provides a viable and affordable solution for the ever increasing demand for storage space. Coupled with a disk caching layer that temporarily holds a small fraction of the total data volume to allow for low latency access, it turns tape based systems into active archival storage (write once, read many) that imposes additional demands on data flow optimization compared to traditional backup setups (write once, read never). In order to preserve the lifetime of tapes and minimize the inherently higher access latency, different tape usage strategies are being evaluated. As an important disk storage system for scientific data that transparently handles tape access, dCache is making efforts to evaluate its recall optimization potential and is introducing a proof-of-concept, high-level stage request scheduling component within its SRM implementation.

Morschel, Lea↗

DOE OSTI2023

Extreme-scale stochastic optimization and simulation via learning-enhanced decomposition and parallelization (Final Technical Report)

Stochastic optimization and simulation models ubiquitously arise in designing and operating complex service/engineering systems. They can be extreme in scale due to high-dimensional data and decisions, and can also involve decisions made sequentially in response to newly revealed data, both causing significant computational challenge. The objective of this research is to explore a unified framework that integrates machine learning with discrete optimization and risk-averse modeling, to improve the efficiency of decomposition paradigms for stochastic optimization and simulations at extreme scale. The models we consider represent a broad class of complex decision-making problems, where 0-1 or continuous decisions are made before and/or after knowing multiple sources of uncertainties that could be correlated. We will employ machine learning methods to dynamically decide and prioritize computational procedures, including cut generation, branching, and bounding of the optimal objective. Furthermore, the research will shed new lights on the traditional decomposition algorithms for extreme-scale computing. Deliverables of the research include new modeling and computational methods for advancing the state-of-the-art research in optimization and simulation, bringing many relevant risk-averse, data-driven optimization problems in practice within the range of tractability. Examples include distributed computing server scheduling and sensor deployment for monitoring critical infrastructures. Success in this effort will enable progress in solving multiple extreme-scale problems in the complex system design and operations arising from DoE missions in energy, environment, and national security.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2024

Application of Banking Scoring and Rating for Coherent Risk Measures in Electricity Systems ABSCORES

This project developed a framework for asset and system risk management that can be incorporated into current electricity system operations to improve economic efficiency and establish an Electric Assets Risk Bureau. We leveraged scoring and ratings from banking and financial institutions alongside current optimization methods in dispatching power systems to help system operators and electricity markets schedule resources. This approach is based on the observation that there are major discrepancies between the power scheduled by a system operator and the actual power generated/consumed. These discrepancies—exacerbated by unplanned contingencies (e.g., natural disasters)—are caused by multiple factors, including the different financial, environmental and risk preferences of power producers, consumers, and aggregators. We developed a framework that counteracts two failures in electricity system operations: imperfect information and missing markets for products. The technical approach included five tasks. Tasks 1 and 2 supported the development of risk scores at the asset level with historical data collected for this project. Tasks 3, 4, and 5 incorporated scoring into decision-making at the system level. The proposed effort achieved PERFORM's Program Objectives because the proposed outputs and algorithms do not exist in the electricity industry and are an innovative approach to managing risk. Since the acknowledged need to better assess and act upon risk profiles for grid assets has not been met by the industry, this project will also impact ARPA-E's Mission Areas, including improving energy efficiency and giving the U.S. a technological lead in advanced energy technologies.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

DOE OSTI2020

Locality-Aware Scheduling for Scalable Heterogeneous Environments

Heterogeneous computing promise boost performance of scientific applications by allowing massively parallel execution of computational tasks. However, manually managing extremely heterogeneous, multi-device systems is complicated and may result in sub-optimal performance. Specifically, data management is an extremely challenging problem on multi-device systems. In this work, we introduce two locality-aware schedulers for the Minos Computing Library (MCL), an asynchronous, task-based programming model and runtime for extremely heterogeneous systems. The first scheduler implements a pure locality-aware algorithm to maximize data reuse, though it might incur in ”hot-spots” that limit system utilization. The second scheduler mitigates this drawback by dynamically targeting between locality-awareness and system utilization based on the current workload and available computing devices. Our results show that locality-awareness greatly benefit applications that exhibit data reuse, providing up to 6.9x and 7.9x over the original MCL scheduler and equivalent OpenCL implementations, respectively. Moreover, our schedulers introduce negligible overhead compared with the original MCL scheduler and achieve similar performance for applications that don’t benefit from data locality.

Architecture, co-design, Task-based programming mo↗

DOE OSTI2022

MARS: Malleable Actor-Critic Reinforcement Learning Scheduler

In this paper, we introduce MARS, a new scheduling system for HPC-cloud infrastructures based on a cost-aware, flexible reinforcement learning approach, which serves as an intermediate layer for next generation HPC-cloud resource manager. MARS ensembles the pre-trained models from heuristic workloads and decides on the most cost-effective strategy for optimization. A whole workflow application would be split into several optimizable dependent sub-tasks, then based on the pre- defined resource management plan, a reward will be generated after executing a scheduled task. Lastly, MARS updates the Deep Neural Network (DNN) model based on the reward. MARS is designed to optimize the existing models through reinforcement mechanisms. MARS adapts to the dynamics of workflow applications, selects the most cost-effective scheduling solution among pre-built scheduling strategies (backfilling, SJF, etc.) and self- learning deep neural network model at run-time. We evaluate MARS with different real-world workflow traces. MARS can achieve 5%-60% increased performance compare to state-of-the- art approaches.

Baheri, Betis↗

DOE OSTI2023

Optimization of Energy Storage System Economics and Controls by Incorporating Battery Degradation Costs in REopt

The use of stationary electrochemical energy storage systems utilizing lithium-ion batteries has increased rapidly as the production scale and price for lithium-ion batteries has decreased. These energy storage systems are crucial for maintaining grid resiliency, especially for grids operating with high penetration of renewable energy generation assets or for with a variety of distributed energy generation and storage systems. One challenging factor for the development of battery energy storage systems is estimating the proper sizing, in terms of both power and energy, that minimizes total costs over the lifetime of the systems; this calculation is difficult in simple cases, where a battery is costed independently, but is extremely challenging when building loads and electrical generation by photovoltaic resources are also considered. REopt is a techoeconomic optimization tool developed by NREL to address these challenges. Previously, battery degradation has been priced by simply assuming a 10-year replacement schedule for battery systems. However, this does not account for varying degradation trends observed across real-world batteries, or allow for batteries to be operated in a degradation-aware manner that optimizes battery dispatch based on operating costs. This work incorporates a battery life model into REopt. This battery life model is simple, so that it may be solvable within the constrains of a mixed-integer linear optimization problem, but is fit to accelerated aging data recorded in the lab. To achieve the best possible accuracy for lifetime estimates given these constraints, parameters for the battery life model in REopt are estimated by fitting 20-year simulations of battery life after identifying state-space battery degradation model from accelerated aging data. Comparisons of battery life predicted in REopt and from the state-space battery degradation model to ensure validity of lifetime estimates made by REopt. Battery life and cost is optimized by controlling three decision to minimize system life cost: battery sizing, daily state-of-charge, and daily energy-throughput. The cost of battery degradation as a function of these control variables is then estimated assuming two possible maintenance strategies: replacement, where the entire battery system is replaced if cell reach an end-of-life capacity threshold; and augmentation, which establishes a fund to pay for continual purchase of new batteries to maintain the initial energy capacity of the system. These two strategies offer conservative (for replacement) and optimistic (for augmentation) bounds for total system cost. The degradation cost incurred by these strategies is then used to control battery dispatch decisions, operating the battery in a degradation-aware manner that maximizes battery lifetime while also providing energy when favorable. Because the mixed-integer linear program has perfect foresight of future energy needs, batteries with degradation costs are always operated using 'just-in-time' charging, which is unrealistic, as no energy is left in the storage system to perform other energy services or to serve as emergency back-up power. To combat this, an inequality constraint on the average annual state-of-charge is imposed, and the sensitivity of system cost to average stored energy, e.g., the cost of system resiliency, can be quantified. Analysis of results has several conclusions, for instance, oversizing of battery storage systems is not a cost burden when battery storage is an optimal solution, as any additional battery capacity can simply be utilized to avoid costs of purchasing energy from a utility.

battery↗

DOE OSTI2022

Design of Digital Twin Sensing Strategies Via Predictive Modeling and Interpretable Machine Learning

This work develops a methodology for sensor placement and dynamic sensor scheduling decisions for digital twins. The digital twin data assimilation is posed as a classification problem, and predictive models are used to train optimal classification trees that represent the map from observed data to estimated digital twin states. In addition to providing a rapid digital twin updating capability, the resulting classification trees yield an interpretable mathematical representation that can be queried to inform sensor placement and sensor scheduling decisions. The proposed approach is demonstrated for a structural digital twin of a 12 ft wingspan unmanned aerial vehicle. Offline, training data are generated by simulating scenarios using predictive reduced-order models of the vehicle in a range of structural states. Furthermore, these training data can be further augmented using experimental or other historical data. In operation, the trained classifier is applied to observational data from the physical vehicle, enabling rapid adaptation of the digital twin in response to changes in structural health. Within this context, we study the performance of the optimal tree classifiers and demonstrate how they enable explainable structural assessments from sparse sensor measurements and also inform optimal sensor placement.

47 OTHER INSTRUMENTATION↗

DOE OSTI2022

2022 FORCE Development Status Update

Integrated energy systems (IESs) are essential for decarbonizing electricity and industrial sectors and fully exploiting these systems requires sophisticated planning, scheduling, and dispatching tools to maximize their socio-economic benefits. The Framework for Optimization of Resources and Economics (FORCE) is a tool suite developed at Idaho National Laboratory and is dedicated to making robust analyses of IESs easier for analysts and researchers. The report summarizes the improvements and updates to FORCE during the most recent development cycle. Particularly, we demonstrate improvements to user accessibility, stochastic time-series analysis, and vertical integration with the software tools provided in FORCE.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2020

Optimal assignment for the single-household shared autonomous vehicle problem

Autonomous vehicles have the potential to transform the way people are transported. While driverless technology may mean fewer vehicles are required to transport people to and from their daily activities, such changes may result in increased congestion or total miles traveled. In this study, we solve the single-household shared autonomous vehicle problem to identify cost-optimal routings of vehicles throughout the day. Such a tool will be useful for consumers seeking to minimize cost and for regulators seeking to understand and predict how people may behave in different scenarios. Here, we provide a thorough literature review and construct a mixed-integer linear program to minimize the daily travel cost of a household attending a given set of activities. Since solution time is a determinant for applicability of such a model, we present the model in a component-wise fashion. This approach allows us to understand which features most affect the problem complexity and solution time. We note that modeling carpooling is the feature that most increases time to find an optimal solution, and we therefore propose a novel modeling technique for carpooling two people. We illustrate the performance of our model by comparing it with other models from the literature and note that our model can solve significantly larger problem instances and in a time that is short enough to facilitate real-time scheduling. We also highlight the utility of our model for regulators, who can use it to analyze quickly produced optimal routes under different cost/tax scenarios.

33 ADVANCED PROPULSION SYSTEMS↗

DOE OSTI2021

High-dimensional Data-driven Energy optimization for Multi-Modal Transit Agencies (HD-EMMA) (Final Technical Report)

Public bus transit services in the U.S. are responsible for at least 19.7 million metric tons of CO 2 emission annually. Electric vehicles (EVs) can have a much lower environmental impact than comparable internal combustion engine vehicles (ICEVs), especially in urban areas. Unfortunately, EVs are also much more expensive than ICEVs. As a result, many public transit agencies can afford only mixed fleets of transit vehicles, consisting of EVs, hybrids (HEVs), and ICEVs. Transit agencies that operate such mixed fleets of vehicles face a challenging optimization problem: these agencies need to decide which vehicles are assigned to serving which transit trips. Since the advantage of EVs over ICEVs varies depending on the route and time of day (e.g., the benefit of EVs is higher in slower traffic with frequent stops and lower on highways), the assignment can have a significant effect on energy use and, hence, environmental impact. Through this project, we have developed reference data about energy collections and constructed a set of machine learning models that can accurately predict the energy consumption for the whole fleet at the level of each trip. We have used these models to develop a scheduling and assignment strategy that can rotate the different vehicle types across the transit agencies’ routes. The optimization algorithm ensures that the vehicles are matched to trips considering weather patterns, expected congestion, and road gradients to minimize the overall energy usage. We list the key observations from our project for other practitioners below. Details are available in the report, and the list of source code and our publications are included in the appendix. 1. We have demonstrated the feasibility of collecting, merging and analyzing large volumes of high-resolution real-world telemetry data from a mixed vehicle fleet. To mitigate the inherent noise of the recorded GPS points, the team developed an algorithm that filters data and maps the points onto a street. The algorithm considers previous and subsequent location measurements and different characteristics of nearby streets to determine how likely the vehicle travels on them. Then, the team segmented the time series into disjoint contiguous samples based on adjacent road segments and repeated the outlier detection and removal. For each data point, the team added features corresponding to elevation changes within the samples, weather features, such as temperature, and traffic data, such as speed ratio between actual speed and free-flow speed. 2. We have developed two forms of machine learning models that be used to understand and analyze the energy operations of a mixed vehicle transit fleet. The micro prediction model provides estimates of instantaneous energy prediction for all types of buses (diesel, hybrid, and electric). Such a model is important in evaluating the energy impacts of real-time bus operation strategies, but it is challenging due to diversified driving cycles of transit buses. The model can help the drivers understand the impact of their driving behaviors and short-term congestions. The macro prediction models estimate average energy consumption across the whole trip considering the features: distance traveled, various road-type features, elevation change, day of the week, time of day, various weather features (temperature, humidity, etc.), and traffic features (speed ratio and jam factor). 3. We have demonstrated that it is possible to transfer the machine learning models we have developed in this project to other teams and cities by using inductive transfer learning. We also showed that the performance of the macro energy prediction models can be improved using a multi-task learning approach where the learning parameters are shared between the models being developed for different vehicle types. The advantage of this approach is improved learning performance as the models can exploit common spatio-temporal and environmental characteristics. 4. Finally, we have developed trip and vehicle assignment and scheduling algorithms that use the energy prediction models and develop a trip to vehicle type (diesel, electric, hybrid) assignment for the whole operation to reduce overall emissions and cost. We have shown through simulations that the proposed algorithms can save $\$$ 48,910 in energy costs and 175 metric tons of CO 2 emission annually for CARTA.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2023

Incorporating Operational Uncertainties into the Dispatch of an Integrated Solar and Storage System

The economic assessment of hybrid energy systems (HES) pairing battery energy storage systems (BESSs) and photovoltaics (PV) is highly important for advancing their deployment in power systems. This paper presents an innovative assessment framework, including an optimal control policy for dispatch under uncertainty and procedures for exploring control parameters that maximize economic benefits. The proposed dispatch policy consists of two steps using system forecast information. The first step is to determine whether a BESS will be used within an operational scheduling time frame based on the probability of events and their thresholds. Once the dispatch of BESS is triggered, a model predictive control (MPC) is carried out in the second step for scheduling using the expected value of system information. By exercising this policy with different thresholds, one can explore the trade-offs between short-term benefits and battery lifetime, and identify an optimal threshold that maximizes the total economic benefits within the battery lifetime. An evaluation study in a real-world HES project is presented to illustrate the proposed framework. Compared with traditional optimal dispatch algorithms, the proposed method can significantly improve the economic benefits of an HES scheduled under forecast uncertainties.

Ma, Xu↗

DOE OSTI2025

Deep Learning–Assisted Multiobjective Optimization of Geological CO 2 Storage Performance under Geomechanical Risks

In geological CO 2 storage, designing the optimal well control strategy for CO 2 injection to maximize CO 2 storage while minimizing the associated geomechanical risks is not trivial. This challenge arises due to pressure buildup, CO 2 plume migration, the highly nonlinear nature of geomechanical responses to rock-fluid interaction, and the high computational cost associated with coupled flow and geomechanics simulations. In this paper, we introduce a novel optimization framework to address these challenges. The optimization problem is formulated as follows: maximize total CO 2 storage while minimizing geomechanical risks by adjusting the injection schedules within bounded constraints. The geomechanical risks are primarily driven by injection-induced pressure build-up, which is characterized by ground displacement and the induced microseismicity. We used the Fourier neural operator (FNO)-based deep learning model to construct surrogate models, replacing the time-consuming coupled flow and geomechanics simulations for evaluating the aforementioned objective functions. The developed surrogate models have been incorporated into a multiobjective optimization framework through a genetic algorithm to reduce the computational burden. The proposed optimization framework reduces the computational cost from approximately 2,400 hours, when using objective function evaluations based on physics-based simulations, to around 20 minutes. A set of Pareto-optimal solutions of the proposed workflow yields nontrivial optimal decisions, reducing the microseismicity potential and the vertical displacement. This Pareto front highlights the optimal trade-offs between CO 2 storage amount, safety, and ground displacement, emphasizing the need for careful optimization and management of injection strategies to achieve a balanced outcome. The novelty of this work is twofold. First, we demonstrate the importance of incorporating the minimization of the geomechanical risks as objective functions into the CO 2 storage optimization workflow to mitigate the potential risk of induced microseismicity and ground displacement. Second, we leverage the FNO-based surrogate models to optimize a real-field CO 2 storage operation.

42 ENGINEERING↗

DOE OSTI2022

Volt-VAR Optimization in Distribution Networks Using Twin Delayed Deep Reinforcement Learning

Modern distribution grids are undergoing new challenges due to the stochastic nature of distributed energy resources (DERs). High penetration of DERs has a significant impact on Volt-VAR profile and system power losses. This work proposes a deep reinforcement learning (DRL)-based Volt-VAR optimization approach for improving voltage profile and reducing system power loss under high penetration of distributed energy resources, such as battery energy storage and solar photovoltaic units in distribution grids. The twin delayed deep deterministic policy gradient (TD3) method-based DRL agent is proposed to configure optimal set-points of reactive power outputs of fast responding smart inverters. The agent schedules the reactive power of inverters according to their physical capabilities, such as minimum allowed power factor, e.g., 0.9 leading/lagging. The reward function of the proposed DRL scheme is designed carefully to ensure a proper voltage profile of the grids with effective scheduling of reactive power outputs from inverters. The performance of the proposed model is verified on modified IEEE 34- and 123-bus systems and compared with base case with no reactive supply by inverters, and local droop Volt-VAR control approach. The results show that the proposed method performs better than the local droop control and deep deterministic policy gradient (DDPG)-based DRL method for reducing voltage fluctuation and minimizing power loss.

Hossain, Rakib↗