-
Acceleration of Power System Dynamic Simulations using a Deep Equilibrium Layer and Neural ODE Surrogate
Authors:
Matthew Bossart,
Jose Daniel Lara,
Ciaran Roberts,
Rodrigo Henriquez-Auba,
Duncan Callaway,
Bri-Mathias Hodge
Abstract:
The dominant paradigm for power system dynamic simulation is to build system-level simulations by combining physics-based models of individual components. The sheer size of the system along with the rapid integration of inverter-based resources exacerbates the computational burden of running time domain simulations. In this paper, we propose a data-driven surrogate model based on implicit machine…
▽ More
The dominant paradigm for power system dynamic simulation is to build system-level simulations by combining physics-based models of individual components. The sheer size of the system along with the rapid integration of inverter-based resources exacerbates the computational burden of running time domain simulations. In this paper, we propose a data-driven surrogate model based on implicit machine learning -- specifically deep equilibrium layers and neural ordinary differential equations -- to learn a reduced order model of a portion of the full underlying system. The data-driven surrogate achieves similar accuracy and reduction in simulation time compared to a physics-based surrogate, without the constraint of requiring detailed knowledge of the underlying dynamic models. This work also establishes key requirements needed to integrate the surrogate into existing simulation workflows; the proposed surrogate is initialized to a steady state operating point that matches the power flow solution by design.
△ Less
Submitted 10 May, 2024;
originally announced May 2024.
-
Synchronization in Modern Heterogeneous Power Networks with Inverter-Based Resources
Authors:
Mark Walth,
Amir Sajadi,
Bri-Mathias Hodge
Abstract:
Synchronized operation of generators in a power network is paramount to the stability and reliability of energy delivery. In this paper, we address the synchronization problem in a heterogeneous power grid, consisting of both synchronous generators and inverter-based generators, and derive the necessary and sufficient condition guaranteeing the existence of a unique locally stable synchronized mod…
▽ More
Synchronized operation of generators in a power network is paramount to the stability and reliability of energy delivery. In this paper, we address the synchronization problem in a heterogeneous power grid, consisting of both synchronous generators and inverter-based generators, and derive the necessary and sufficient condition guaranteeing the existence of a unique locally stable synchronized mode. Further, our results have implications for utilizing grid-following inverters versus grid-forming inverters in order to enhance network stability. This work is of particular importance as power grids around the world are undergoing a transition from being predominantly composed of synchronous generators towards a grid consisting of a preponderance of inverter based generators.
△ Less
Submitted 9 April, 2024;
originally announced April 2024.
-
Plane Wave Dynamic Model of Electric Power Networks with High Shares of Inverter-Based Resources
Authors:
Amirhossein Sajadi,
Bri-Mathias Hodge
Abstract:
Contemporary theories and models for electric power system stability are predicated on a widely held assumption that the mechanical inertia of the rotating mass of synchronous generators provides the sole contribution to stable and synchronized operation of this class of complex networks on subsecond timescales. Here we formulate the electromagnetic momentum of the field around the transmission li…
▽ More
Contemporary theories and models for electric power system stability are predicated on a widely held assumption that the mechanical inertia of the rotating mass of synchronous generators provides the sole contribution to stable and synchronized operation of this class of complex networks on subsecond timescales. Here we formulate the electromagnetic momentum of the field around the transmission lines that transports energy and present evidence from a real-world bulk power network that demonstrates its physical significance. We show the classical stability model for power networks that overlooks this property, known as the "swing equation", may become inadequate to analyze systems with high shares of inverter-based resources, commonly known as "low-inertia power systems". Subsequently, we introduce a plane wave dynamic model, consistent with the structural properties of emerging power systems with up to 100% inverter-based resources, which identifies the concept of inertia in power grids as a time-varying component. We leverage our theory to discuss a number of open questions in the electric power industry. Most notably, we postulate that the changing nature of power networks with a preponderance of variable renewable energy power plants could strengthen power network stability in the future; a vision which is irreconcilable with the conventional theories.
△ Less
Submitted 29 January, 2024;
originally announced January 2024.
-
Towards Robust and Scalable Dispatch Modeling of Long-Duration Energy Storage
Authors:
Omar J. Guerra,
Sourabh Dalvi,
Amogh A. Thatte,
Brady Cowiestoll,
Jennie Jorgenson,
Bri-Mathias Hodge
Abstract:
Existing modeling approaches for long-duration energy storage (LDES) are often based either on an oversimplified representation of power system operations or limited representation of storage technologies, e.g., evaluation of only a single application. This manuscript presents an overview of the challenges of modeling LDES technologies, as well as a discussion regarding the capabilities and limita…
▽ More
Existing modeling approaches for long-duration energy storage (LDES) are often based either on an oversimplified representation of power system operations or limited representation of storage technologies, e.g., evaluation of only a single application. This manuscript presents an overview of the challenges of modeling LDES technologies, as well as a discussion regarding the capabilities and limitations of existing approaches. We used two test power systems with high shares of both solar photovoltaics- and wind (70% - 90% annual variable renewable energy shares) to assess LDES dispatch approaches. Our results estimate that better dispatch modeling of LDES could increase the associated operational value by 4% - 14% and increase the standard capacity credit by 14% - 34%. Thus, a better LDES dispatch could represent significant cost saving opportunities for electric utilities and system operators. In addition, existing LDES dispatch modeling approaches were tested in terms of both improved system value (e.g., based on production cost and standard capacity credit) and scalability (e.g., based on central processing unit time and peak memory usage). Both copper plate and nodal representations of the power system were considered. Although the end volume target dispatch approach, i.e., based on mid-term scheduling, showed promising performance in terms of both improved system value and scalability, there is a need for robust and scalable dispatch approaches for LDES in transmission-constrained electric grids. Moreover, more research is required to better understand the optimal operation of LDES considering extreme climate/weather events, reliability applications, and power system operational uncertainties.
△ Less
Submitted 29 January, 2024;
originally announced January 2024.
-
STint: Self-supervised Temporal Interpolation for Geospatial Data
Authors:
Nidhin Harilal,
Bri-Mathias Hodge,
Aneesh Subramanian,
Claire Monteleoni
Abstract:
Supervised and unsupervised techniques have demonstrated the potential for temporal interpolation of video data. Nevertheless, most prevailing temporal interpolation techniques hinge on optical flow, which encodes the motion of pixels between video frames. On the other hand, geospatial data exhibits lower temporal resolution while encompassing a spectrum of movements and deformations that challeng…
▽ More
Supervised and unsupervised techniques have demonstrated the potential for temporal interpolation of video data. Nevertheless, most prevailing temporal interpolation techniques hinge on optical flow, which encodes the motion of pixels between video frames. On the other hand, geospatial data exhibits lower temporal resolution while encompassing a spectrum of movements and deformations that challenge several assumptions inherent to optical flow. In this work, we propose an unsupervised temporal interpolation technique, which does not rely on ground truth data or require any motion information like optical flow, thus offering a promising alternative for better generalization across geospatial domains. Specifically, we introduce a self-supervised technique of dual cycle consistency. Our proposed technique incorporates multiple cycle consistency losses, which result from interpolating two frames between consecutive input frames through a series of stages. This dual cycle consistent constraint causes the model to produce intermediate frames in a self-supervised manner. To the best of our knowledge, this is the first attempt at unsupervised temporal interpolation without the explicit use of optical flow. Our experimental evaluations across diverse geospatial datasets show that STint significantly outperforms existing state-of-the-art methods for unsupervised temporal interpolation.
△ Less
Submitted 31 August, 2023;
originally announced September 2023.
-
Parameterized Linear Power Flow for High Fidelity Voltage Solutions in Distribution Systems
Authors:
Marija Marković,
Bri-Mathias Hodge
Abstract:
This paper introduces a new model for highly accurate distribution voltage solutions, coined as a parameterized linear power flow model. The proffered model is grounded on a physical model of linear power flow equations, and uses learning-aided parameterization to increase the fidelity of voltage solutions over a wide range of operating points. To this end, the closed-form analytic solution of the…
▽ More
This paper introduces a new model for highly accurate distribution voltage solutions, coined as a parameterized linear power flow model. The proffered model is grounded on a physical model of linear power flow equations, and uses learning-aided parameterization to increase the fidelity of voltage solutions over a wide range of operating points. To this end, the closed-form analytic solution of the parameterization approach is obtained via a Gaussian Process using a deliberately small input sample and without the need for recomputation. The resulting "self-adjusting" parameter is system-specific and controls how accurate the proposed power flow equations are according to loading conditions. Under a certain value of the resulting parameter, the proposed model can fully recover the linearized formulation of a specialized branch flow model for radial distribution systems, the so-called simplified DistFlow model. Numerical examples are provided to illustrate the effectiveness of the proposed model as well as the improvement in solution accuracy for voltage magnitudes over the simplified DistFlow model and several other linear power flow models, at multiple loading levels. Simulations were carried out on six small- and medium-sized test systems.
△ Less
Submitted 9 September, 2022;
originally announced September 2022.
-
Droop-e: Exponential Droop as a Function of Power Output for Grid-Forming Inverters with Autonomous Power Sharing
Authors:
Rick Wallace Kenyon,
Amirhossein Sajadi,
Bri-Mathias Hodge
Abstract:
This paper presents the novel Droop-e grid-forming inverter control strategy, which establishes an active power-frequency relationship based on an exponential function of the inverter power dispatch. The advantages of this control strategy include an increased utilization of available headroom, mitigated system frequency dynamics, and a natural limiting behavior, all of which are directly compared…
▽ More
This paper presents the novel Droop-e grid-forming inverter control strategy, which establishes an active power-frequency relationship based on an exponential function of the inverter power dispatch. The advantages of this control strategy include an increased utilization of available headroom, mitigated system frequency dynamics, and a natural limiting behavior, all of which are directly compared to the hitherto standard static droop approach. First, the small signal stability of the Droop-e control is assessed on a 3-bus system and found stable across all possible inverter power dispatches. Then, time-domain simulations show improved frequency dynamics at lower power dispatches, and a limiting behavior at higher dispatches. Finally, a novel secondary control scheme is introduced that achieves power sharing following the primary Droop-e response to load perturbations, which is shown to be effective in time-domain simulations of the 3- and 9-bus systems; comparative simulations with a static 5% droop yields unacceptable frequency deviations, highlighting the superiority of the Droop-e control.
△ Less
Submitted 7 July, 2022;
originally announced July 2022.
-
Hierarchical Control of Utility-Scale Solar PV Plants for Mitigation of Generation Variability and Ancillary Service Provision
Authors:
Simon Julien,
Amirhossein Sajadi,
Bri-Mathias Hodge
Abstract:
Renewable energy technologies including solar and wind inevitably play a leading role in meeting the growing demand for a decarbonized and clean power grid. However, these technologies are highly dependent of meteorological conditions of power plant site and the challenge remains on how to cope with their short-term and momentarily variability. This paper presents a hierarchical control system to…
▽ More
Renewable energy technologies including solar and wind inevitably play a leading role in meeting the growing demand for a decarbonized and clean power grid. However, these technologies are highly dependent of meteorological conditions of power plant site and the challenge remains on how to cope with their short-term and momentarily variability. This paper presents a hierarchical control system to provide ancillary services from a solar PV power plant to the grid without the need for additional non-solar resources. With coordinated management of each inverter in the system, the control system commands the power plant to proactively curtail a fraction of its instantaneous maximum power potential, which gives the plant enough headroom to ramp up or down power production from the overall power plant, for a service such as regulation reserve, even under changing cloud cover conditions. A case study from a site in Hawaii with one-second resolution solar irradiance data is used to verify the efficacy of the proposed control system. The algorithm is subsequently compared with an alternative control technology from the literature, the grou** control algorithm; the results show that the proposed hierarchical control system is over 10 times more effective in reducing generator mileage to support power fluctuations from solar PV power plants.
△ Less
Submitted 30 June, 2021;
originally announced July 2021.
-
Synchronization in Electric Power Networks with Inherent Heterogeneity Up to 100% Inverter-Based Renewable Generation
Authors:
Amirhossein Sajadi,
Rick Wallace Kenyon,
Bri-Mathias Hodge
Abstract:
The synchronized operation of power generators is the foundation of electric power network stability and a key to the prevention of undesired power outages and blackouts. Here, we derive the conditions that guarantee synchronization in power networks with inherent generator heterogeneity when subjected to small perturbations, and perform a parametric sensitivity analysis to understand synchronizat…
▽ More
The synchronized operation of power generators is the foundation of electric power network stability and a key to the prevention of undesired power outages and blackouts. Here, we derive the conditions that guarantee synchronization in power networks with inherent generator heterogeneity when subjected to small perturbations, and perform a parametric sensitivity analysis to understand synchronization with varied types of generators. As inverter-based resources, which are the primary interfacing technology for many renewable sources of energy, have supplanted synchronous generators in ever growing numbers, the center of attention on associated integration challenges have resided primarily on the role of declining system inertia. Our results instead highlight the critical role of generator dam** in achieving a stable state of synchronization. Additionally, we report the feasibility of operating interconnected electric grids with up to 100% power contribution from inverter-based renewable generation technologies. Our study has important implications as it sets the basis for the development of advanced control architectures and grid optimization methods that ensure synchronization and further pave the path towards the decarbonization of the electric power sector.
△ Less
Submitted 5 May, 2022; v1 submitted 8 April, 2021;
originally announced April 2021.
-
Interactive Power to Frequency Dynamics Between Grid-Forming Inverters and Synchronous Generators in Power Electronics-Dominated Power Systems
Authors:
Rick Wallace Kenyon,
Amirhossein Sajadi,
Matt Bossart,
Andy Hoke,
Bri-Mathias Hodge
Abstract:
With increased attention on grid-forming inverters as a power system stabilizing device during periods of high shares of inverter-based resource operations, there is a present need for a transparent and rigorous investigation of the inverted and direct power to frequency control capabilities, and associated impacts, of these devices on hybrid systems. Here, analysis of the frequency dynamics of th…
▽ More
With increased attention on grid-forming inverters as a power system stabilizing device during periods of high shares of inverter-based resource operations, there is a present need for a transparent and rigorous investigation of the inverted and direct power to frequency control capabilities, and associated impacts, of these devices on hybrid systems. Here, analysis of the frequency dynamics of the droop controlled grid-forming inverter and the synchronous generator illuminates the inverted active power-frequency relationship and the frequency response order reduction, forming the basis for novel, non-linear frequency droop control approaches. Device-level electromagnetic transient domain simulations corroborate the order-reduction findings, establish that a properly designed DC-side system has a negligible dynamical impact on active power transfer and will not impede frequency regulation, and confirm the frequency response improvement with non-linear control. Simulations of the 9- and 39-bus test systems validate the order reduction and associated decoupling of the nadir and rate of change of frequency in large networks. The primary system oscillatory mode confirms the correlation between high shares of grid-forming inverters and increased mode frequency and dam**; a sharp decrease in dam** is observed at shares above 80%, whether by grid-forming device quantity in large networks or rating variations in a small test system. Finally, simulation results on the Hawaiian island of Maui show a trend towards a first order frequency response with a grid-forming inverter, further corroborating the analytic findings and network impacts.
△ Less
Submitted 8 October, 2022; v1 submitted 24 February, 2021;
originally announced February 2021.
-
Voltage Estimation in Low-Voltage Distribution Grids with Distributed Energy Resources
Authors:
Marija Marković,
Amirhossein Sajadi,
Anthony Florita,
Robert Cruickshank III,
Bri-Mathias Hodge
Abstract:
The present distribution grids generally have limited sensing capabilities and are therefore characterized by low observability. Improved observability is a prerequisite for increasing the hosting capacity of distributed energy resources such as solar photovoltaics (PV) in distribution grids. In this context, this paper presents learning-aided low-voltage estimation using untapped but readily avai…
▽ More
The present distribution grids generally have limited sensing capabilities and are therefore characterized by low observability. Improved observability is a prerequisite for increasing the hosting capacity of distributed energy resources such as solar photovoltaics (PV) in distribution grids. In this context, this paper presents learning-aided low-voltage estimation using untapped but readily available and widely distributed sensors from cable television (CATV) networks. The CATV sensors offer timely local voltage magnitude sensing with 5-minute resolution and can provide an order of magnitude more data on the state of a distribution system than currently deployed utility sensors. The proposed solution incorporates voltage readings from neighboring CATV sensors, taking into account spatio-temporal aspects of the observations, and estimates single-phase voltage magnitudes at all non-monitored buses using random forest. The effectiveness of the proposed approach was demonstrated using a 1572-bus feeder from the SMART-DS data set for two case studies - passive distribution feeder (without PV) and active distribution feeder (with PV). The analysis was conducted on simulated data, and the results show voltage estimates with a high degree of accuracy, even at extremely low percentages of observable nodes.
△ Less
Submitted 11 November, 2020;
originally announced November 2020.
-
Computational Experiment Design for Operations Model Simulation
Authors:
Jose Daniel Lara,
Jonathan T. Lee,
Duncan Callaway,
Bri-Mathias Hodge
Abstract:
Computer simulations that demonstrate the valueof novel approaches are crucial to develo** more flexibleand robust power systems operations with high penetrations ofrenewable energy at multiple geographic and temporal scales.However, optimization-based simulations that depend on forecastdata often face challenges in evaluating performance, reproducingresults, and testing under realistic simulati…
▽ More
Computer simulations that demonstrate the valueof novel approaches are crucial to develo** more flexibleand robust power systems operations with high penetrations ofrenewable energy at multiple geographic and temporal scales.However, optimization-based simulations that depend on forecastdata often face challenges in evaluating performance, reproducingresults, and testing under realistic simulation scenarios. In thispaper, we develop scientific computing best-practices for thevalidation and reproduction of power systems operational models.We then employ two case studies to demonstrate the proposedvalidation and reproduction framework.
△ Less
Submitted 23 April, 2020;
originally announced April 2020.
-
An Unsupervised Clustering-Based Short-Term Solar Forecasting Methodology Using Multi-Model Machine Learning Blending
Authors:
Cong Feng,
Mingjian Cui,
Bri-Mathias Hodge,
Siyuan Lu,
Hendrik F. Hamann,
Jie Zhang
Abstract:
Solar forecasting accuracy is affected by weather conditions, and weather awareness forecasting models are expected to improve the performance. However, it may not be available and reliable to classify different forecasting tasks by using only meteorological weather categorization. In this paper, an unsupervised clustering-based (UC-based) solar forecasting methodology is developed for short-term…
▽ More
Solar forecasting accuracy is affected by weather conditions, and weather awareness forecasting models are expected to improve the performance. However, it may not be available and reliable to classify different forecasting tasks by using only meteorological weather categorization. In this paper, an unsupervised clustering-based (UC-based) solar forecasting methodology is developed for short-term (1-hour-ahead) global horizontal irradiance (GHI) forecasting. This methodology consists of three parts: GHI time series unsupervised clustering, pattern recognition, and UC-based forecasting. The daily GHI time series is first clustered by an Optimized Cross-validated ClUsteRing (OCCUR) method, which determines the optimal number of clusters and best clustering results. Then, support vector machine pattern recognition (SVM-PR) is adopted to recognize the category of a certain day using the first few hours' data in the forecasting stage. GHI forecasts are generated by the most suitable models in different clusters, which are built by a two-layer Machine learning based Multi-Model (M3) forecasting framework. The developed UC-based methodology is validated by using 1-year of data with six solar features. Numerical results show that (i) UC-based models outperform non-UC (all-in-one) models with the same M3 architecture by approximately 20%; (ii) M3-based models also outperform the single-algorithm machine learning (SAML) models by approximately 20%.
△ Less
Submitted 10 May, 2018;
originally announced May 2018.
-
Exact Ground State Energy of Hubbard Rings in the Atomic Limit
Authors:
W. B. Hodge,
N. A. W. Holzwarth,
W. C. Kerr
Abstract:
Using a straightforward extension of the analysis of Lieb and Wu, we derive a simple analytic form for the ground state energy of a one-dimensional Hubbard ring in the atomic limit. This result is valid for an \textit{arbitrary} number of lattice sites $L$ and electrons $N \leq L$. Furthermore, our analysis, including an application of the theory of stochastic matrices, provides insight into the…
▽ More
Using a straightforward extension of the analysis of Lieb and Wu, we derive a simple analytic form for the ground state energy of a one-dimensional Hubbard ring in the atomic limit. This result is valid for an \textit{arbitrary} number of lattice sites $L$ and electrons $N \leq L$. Furthermore, our analysis, including an application of the theory of stochastic matrices, provides insight into the degeneracy and spin properties of the ground states in the atomic limit. We give numerical results which illustrate how the atomic limit is approached.
△ Less
Submitted 6 October, 2009; v1 submitted 10 June, 2008;
originally announced June 2008.
