Increasing Real-Time and Day-Ahead Market and Planning Efficiency Through Improved Software

 

Tuesday, July 8, 2025
9:15 AM	Introduction Monica Ferrera, Federal Energy Regulatory Commission (Washington, D.C.)
9:30 AM	Session T1   Copula-Based Uncertainty Modeling for Grid Operation Mingguo Hong, ISO New England (Holyoke, MA) Slava Maslennikov, ISO New England (Holyoke, MA) Xiaochuan Luo, ISO New England (Holyoke, MA) Tongxin Zheng, ISO New England (Holyoke, MA)   Enhancing SCUC Scalability through Redundant Constraint Filtering Sushant Varghese, New York ISO (Rensselaer, NY) Shubo Zhang, New York ISO (Rensselaer, NY) Matthew Musto, New York ISO (Rensselaer, NY) Kanchan Upadhyay, New York ISO (Rensselaer, NY)   2024 Long-Term System Assessment Results for the ERCOT Grid Pengwei Du, ERCOT (Austin, TX)
11:00 AM	Break
11:15 AM	Session T2   Pricing Compounded Congestion from Multiple Contingencies in California ISO’s Markets Haifeng Liu, California ISO (Folsom, CA) Fan Zhang, California ISO (Folsom, CA) Guillermo Bautista Alderete, California ISO (Folsom, CA) Khaled Abdul-Rahman, California ISO (Folsom, CA)   The MRI-based Capacity Accreditation: Interpretation and Properties Feng Zhao, ISO New England (Holyoke, MA) Tongxin Zheng, ISO New England (Holyoke, MA) Dane Schiro, ISO New England (Holyoke, MA) Xiaochu Wang, ISO New England (Holyoke, MA)   Revenue Sufficiency with Scarcity Pricing and Capacity Market: the Role of Coordinated Operating Reserve and Capacity Demand Curves Dongwei Zhao, Argonne National Laboratory (Westmont, IL) Zhi Zhou, Argonne National Laboratory (Westmont, IL) Jonghwan Kwon, Argonne National Laboratory (Westmont, IL) Todd Lovin, Argonne National Laboratory (Westmont, IL)
12:45 PM	Lunch
2:00 PM	Session T3 Interconnection Life Cycle Intelligence and Automation Brian Fitzsimons, GridUnity (Summit, NJ) Casey Cathy, Southwest Power Pool (Little Rock, AR) Deb Le Vine, California ISO (Folsom, CA)   Accelerating the Interconnection Process with Advanced Software and Automation Andy Witmeier, Midcontinent ISO (Carmel, IN) David Bromberg, Pearl Street Technologies (Pittsburgh, PA)   Developments of Cluster-Based Generator Interconnection Study Procedures in TVA and Automation Tool in TARA Luke Ellis, Tennessee Valley Authority (Hixson, TN) Seungwon An, PowerGEM (Niskayuna, NY)
3:30 PM	Break
4:00 PM	Session T4-A   Bridging the Gaps: Co-Optimizing Energy Systems for a Resilient Future Rob Homer, Energy Exemplar (Salt Lake City, UT)   Dynamic Reserve Requirements Chen-Hao Tsai, Midcontinent ISO (Carmel, IN) Arezou Ghesmati, Midcontinent ISO (Carmel, IN) Bing Huang, Midcontinent ISO (Carmel, IN)   Modernizing Grid Risk Management with AI Advances and Cloud Technology  Arezou Ghesmati, Midcontinent ISO (Carmel, IN) Long Zhao, Midcontinent ISO (Carmel, IN) Concong Wang, Midcontinent ISO (Carmel, IN) Evan Sattler, Midcontinent ISO (Carmel, IN)
4:00 PM	Session T4-B   Modeling Of Internal Controllable HVDC Lines in Energy Market Operations Hossein Lotfi, New York ISO (Rensselaer, NY) Bo Yuan, Cornell University (Ithaca, NY) Muhammad Marwali, New York ISO (Rensselaer, NY)   Optimizing DC Tie Utilization in Southwest Power Pool’s Expanded RTO Seth Mayfield, Southwest Power Pool (Little Rock, AR) Caroline Chapman, Southwest Power Pool (Little Rock, AR)   How Software can be Used to Automate Much of Today's Transmission Planning Workload Andrew Martin, Nira Energy (Denver, CO)
5:30 PM	Adjourn

 

 

Wednesday, July 9, 2025
9:00 AM	Session W1-A   Tapestry’s AI-enabled Interconnection Automation Software: Speeding Grid Interconnections and Enabling Dynamic Planning and Operations Dr. Cat Wong, Tapestry (Mountain View, CA) AJ Lambert, Interconnection Planning Projects Manager, PJM Interconnection (Pottstown, PA) Solving Mathematical Programs with Embedded Surrogate Models using GAMSPy Adam Christensen, GAMS Development Corporation (Fairfax, VA) Hamdi Burak Usul, GAMS Development Corporation (Fairfax, VA) Muhammet Soyturk, GAMS Development Corporation (Fairfax, VA) Steven Dirkse, GAMS Development Corporation (Fairfax, VA) Michael Bussieck, GAMS Development Corporation (Fairfax, VA)   Building Open-source AI Solutions and Foundation Models for Market and Planning Efficiency Alexandre Parisot, Linux Foundation Energy (Menlo Park, CA) François Mirallès, Hydro-Québec (Varennes, Canada) Jonas Weiss, IBM Research (Rueschlikon, Switzerland) Thomas Brunschwiler, IBM Research (Rueschlikon, Switzerland) Hendrik Hamann, IBM Research (Yorktown Heights, NY)
9:00 AM	Session W1-B   Prioritizing Interconnection Applications via High-Fidelity Stochastic Capacity Expansion Planning Elizabeth Glista, Lawrence Livermore National Laboratory (Livermore, CA) Jean-Paul Watson, Lawrence Livermore National Laboratory (Livermore, CA)   Managing Variability with High Fidelity Capacity Expansion Russ Philbrick, Polaris Systems Optimization (Shoreline, WA)   PTDF Powerflow Representation for Accelerating Large-scale Stochastic Nodal Capacity Expansion Planning Tomas Valencia Zuluaga, Lawrence Livermore National Laboratory (Livermore, CA) Amelia Musselman, Lawrence Livermore National Laboratory (Livermore, CA) Jean-Paul Watson, Lawrence Livermore National Laboratory (Livermore, CA) Shmuel Oren, University of California at Berkeley (Berkeley, CA)
10:30 AM	Break
10:45 AM	Session W2-A   Generative AI Software for High-Fidelity Household Load Profiles: Innovations in Resource Adequacy Modelling and Planning Efficiency Gareth Jones, Octopus Energy Group (London, United Kingdom) Gus Chadney, Octopus Energy Group (Marlow, United Kingdom) Sheng Chai, Octopus Energy Group (London, United Kingdom) Amber Woodward, Octopus Energy Group (London, United Kingdom)   Machine Learning Enhanced Formulation Tightening of Energy Storage Resource Constraints in Unit Commitment Farhan Hyder, Rochester Institute of Technology (Rochester, NY) Uyen Nhi Quang, Rochester Institute of Technology (Rochester, NY) Bing Yan, Rochester Institute of Technology (Rochester, NY)
10:45 AM	Session W2-B   A Novel Partitioning Algorithm for Constructing Diverse Solution Sets for the Unit Commitment Problem Ignacio Aravena, Lawrence Livermore National Laboratory (Livermore, CA) Jisun Lee, University of California at Berkeley (Berkeley, CA) Jean-Paul Watson, Lawrence Livermore National Laboratory (Livermore, CA) Alper Atamturk, University of California at Berkeley (Berkeley, CA)   Next-Generation Market Approaches to Support System Stability and Reliability Allison Campbell, Pacific Northwest National Laboratory (Portland, OR) Matthew Cornachione, Pacific Northwest National Laboratory (Richland, WA) Jesse Holzer, Pacific Northwest National Laboratory (Richland, WA) Molly Rose Kelly-Gorham, Pacific Northwest National Laboratory (Richland, WA) Liping Li, Pacific Northwest National Laboratory (Richland, WA) Ki Yeob Lee, Pacific Northwest National Laboratory (Richland, WA) Eran Schweitzer, Pacific Northwest National Laboratory (Richland, WA)
11:45 AM	Break
12:00 PM	Panel W2-A Panel Discussion on AI Integration in Power Markets
12:45 PM	Lunch
2:00 PM	Session W3-A   Advancing Market Efficiency and Reliability: Forecasting Emissions and Asset Risk in Electricity Systems Andres F. Ramirez, Lehigh University (Bethlehem, PA) Alberto J. Lamadrid, Lehigh University (Bethlehem, PA)   Factor-Based Portfolio Optimization for Risk Management in Electricity Markets with Renewable Energy Audun Botterud, Massachusetts Institute of Technology (Cambridge, MA) Yusu Liu, Massachusetts Institute of Technology (Cambridge, MA) Arnab Sur, Lehigh University (Bethlehem, PA) Alberto J. Lamadrid, Lehigh University (Bethlehem, PA)   On the Significance of High-Fidelity Resource Adequacy Assessment: The Case of New York ISO Aleksandr Rudkevich, Newton Energy Group LLC (Newton, MA) F. Selin Yanikara, Newton Energy Group LLC (Newton, MA) Russ Philbrick, Polaris Systems Optimization (Shoreline, WA)
2:00 PM	Session W3-B   Effective Congestion Mitigation with Transmission Topology Optimization at Alliant Energy and ATC – Case Studies and Practical Lessons Learned Mitchell Myhre, Alliant Energy (Madison, WI) Kristie Erickson, ATC (Waukesha, WI) Pablo A. Ruiz, NewGrid, Inc. (Somerville, MA) Paola Caro, NewGrid, Inc. (Somerville, MA) German Lorenzon, NewGrid, Inc. (Somerville, MA)   RiskVIEW: An Interactive Software Solution for Implementing the FERC Order No. 881 for Power Transmission Networks Ali Bidram, Grid Modernization Solutions LLC (Salt Lake City, UT) Jairo Giraldo Trujillo, Grid Modernization Solutions LLC (Salt Lake City, UT) Alex Farley, Grid Modernization Solutions LLC (Salt Lake City, UT) Masood Parvania, Grid Modernization Solutions LLC (Salt Lake City, UT)   Market-based Incentives for Adoption of Grid-Enhancing Technologies Mostafa Ardakani, University of Utah (Salt Lake City, UT)
3:30 PM	Break
3:45PM	Session W4-A   Market-based Storage Dispatch and Electricity Reliability Siva Visvesvaran, Cornell University (Ithaca, NY) Jacob Mays, Cornell University (Ithaca, NY)   Unlocking the Full Value of Energy Storage with Dynamic Optimization Software Michael Baker, Tyba (Oakland, CA)   Locational Energy Storage Bid Bounds for Facilitating Social Welfare Convergence Ning Qi, Columbia University (New York, NY) Bolun Xu, Columbia University (New York, NY)   Battery Operations in Electricity Markets: Strategic Behavior and Distortions Jerry Anunrojwong, Columbia University (New York, NY) Santiago R. Balseiro, Columbia University (New York, NY) Omar Besbes, Columbia University (New York, NY) Bolun Xu, Columbia University (New York, NY)
3:45PM	Session W4-B   Modeling Extreme Heat Wave and Wildfire Impacts on Power Reliability Juliette Franzman, Lawrence Livermore National Laboratory (Livermore, CA) Hannah Burroughs, Lawrence Livermore National Laboratory (Livermore, CA) Jhi-Young Joo, Lawrence Livermore National Laboratory (Livermore, CA) Andrew Mastin, Lawrence Livermore National Laboratory (Livermore, CA) Christabella Annalicia, Lawrence Livermore National Laboratory (Livermore, CA) Jean-Paul Watson, Lawrence Livermore National Laboratory (Livermore, CA)   Weathering the Firestorm: Mitigating Electric Grid Ignited Wildfires Brian J. Pierre, Sandia National Laboratories (Albuquerque, NM)   Multistage Generator Outage Simulation for Probabilistic Valuation of Operational Response to Extreme Events Luke Lavin, National Renewable Energy Laboratory (Golden, CO) Jose Daniel Lara, National Renewable Energy Laboratory (Golden, CO) Matthew Bossart, National Renewable Energy Laboratory (Golden, CO) David Palchak, National Renewable Energy Laboratory (Golden, CO)   Prognostics-Driven Operations & Maintenance to Enhanced Grid Reliability Feng Qiu, Argonne National Laboratory (Lemont, IL) Shijia Zhao, Argonne National Laboratory (Lemont, IL) Murat Yildirim Wayne State University (Detroit, MI) Zhaoyu Wang, Iowa State University (Ames, IA) Joydeep Mitra, Michigan State University (East Lansing, MI)
5:45 PM	Adjourn

 

 

Thursday, July 10, 2025
9:15 AM	Session H1-A   Probabilistic Grid Reliability Analysis with Energy Storage Systems: An Open-Source Tool for Assessing the Resource Adequacy of Power Systems Atri Bera, Sandia National Laboratories (Albuquerque, NM)   GridCal - Open-source for Modern Power Systems Josep Fanals i Batllori, eRoots Analytics SL (Barcelona, Spain) Santiago Peñate-Vera, eRoots Analytics (Santa Brígida, Spain)   Multiscale Stochastic Day- and Week-ahead Scheduling of a Large Renewable-dominated Power System – The Brazilian Case Mario V. Pereira, PSR (Rio de Janeiro, Brazil) Joaquim Dias Garcia, PSR (Rio de Janeiro, Brazil) Thiago Cesar, PSR (Rio de Janeiro, Brazil) Julio Alberto, PSR (Rio de Janeiro, Brazil) Luiz Carlos Costa, PSR (Rio de Janeiro, Brazil) Guilherme Bodin, PSR (Rio de Janeiro, Brazil) Andre Dias, PSR (Rio de Janeiro, Brazil) Raphael Chabar, PSR (Rio de Janeiro, Brazil)
9:15 AM	Session H1-B   A Comparison of Energy Market Pricing Methods Richard O'Neill (Silver Spring, MD)   Evaluating the Benefits of Combining Short-term and Long-term Flexible Ramping Products in Real-Time Electricity Markets Aravind Retna Kumar, Pennsylvania State University (University Park, PA) Anthony Giacomoni, Manager, PJM Interconnection (Audubon, PA) Shailesh Wasti, Pennsylvania State University (University Park, PA) Mort Webster, Pennsylvania State University (University Park, PA)   European Merging Function: Merging Grid Models to Enhance Coordination of Interregional Flows and Reliability with Open-source Operational Computation Modules Gladys Leо́n Surо́s, Artelys (Paris, France) Damien Jeandemange, Artelys (Paris, France) Maja Markovic, Maja Markovic PR Digital Energy (Belgrade, Serbia) Tengixang Ren, Artelys (Montreal, Canada) Alexis Godefroy, Artelys (Lyon, France) Nicolas Omont, Vice President, Artelys (Paris, France)
10:45 AM	Break
11:00 AM	Session H2-A   Impacts of Distributed Energy Resource Aggregations in Transmission Systems Matthew Cornachione, Pacific Northwest National Laboratory (Richland, WA) Brent Eldridge, Pacific Northwest National Laboratory (Richland, WA) Jesse Holzer, Pacific Northwest National Laboratory (Richland, WA) Eran Schweitzer, Pacific Northwest National Laboratory (Richland, WA) Liping Li, Pacific Northwest National Laboratory (Richland, WA)   Smart Grids as Coupled Physical and Economic Systems Leigh Tesfatsion, Iowa State University (Ames, IA)   Coordinated Transmission and Distribution Networks: A Bi-Level Framework for Resilience, Economic Efficiency, and Sustainability Moses Amoasi Acquah, University of Connecticut (Storrs, CT) Zongjie Wang, University of Connecticut (Storrs, CT)
11:00 AM	Session H2-B   Wholesale Electricity Markets with Multiple Virtual Power Plants Andrew Liu, Purdue University (West Lafayette, IN) Jun He, Purdue University (West Lafayette, IN)   Synthesizing Grid Data with Cyber Resilience and Privacy Guarantees Vladimir Dvorkin, University of Michigan (Ann Arbor, MI) Shengyang Wu, University of Michigan (Ann Arbor, MI)   PowerSAS.jl: A High-Performance Transmission Planning Reliability Analysis Tool in Julia Wei Gao, Argonne National Laboratory (Lemont, IL) Feng Qiu, Argonne National Laboratory (Lemont, IL) Shijia Zhao, Argonne National Laboratory (Lemont, IL)
12:00 PM	Adjourn

 

 

 

Session T1 (Tuesday, July 8, 9:30 AM)

---

Copula-Based Uncertainty Modeling for Grid Operation

Dr. Mingguo Hong, Principal Analyst, ISO New England (Holyoke, MA)

Dr. Slava Maslennikov, Technical Manager, ISO New England (Holyoke, MA)

Dr. Xiaochuan Luo, Manager, Power System Technology, ISO New England (Holyoke, MA)

Dr. Tongxin Zheng, Chief Technologist, ISO New England (Holyoke, MA)

With the increasing penetration of renewable, power system operation faces significant reliability and economic challenges.  One primary issue lies in the uncertainty of power generation of the renewable resources due to their weather dependency.  While the accurate forecasts of renewable generation and loads continue to be essential, such forecasts are probabilistic in nature and the uncertainty grows as the operational time horizon expands into the future.  At the same time, spatiotemporal correlations among renewable generations should also be considered in the forecast to better assess the aggregated impact on power system security.  This presentation discusses the recent research at the ISO New England for quantifying the forecast uncertainties using Copulas where both resource and region-level operational scenarios can be created for assessing power system security risks in operation planning studies.  The copula-based uncertainty model captures the inherent spatiotemporal correlations among the multivariate forecast timeseries and helps to develop probabilistic power system studies.

---

Enhancing SCUC Scalability through Redundant Constraint Filtering

Dr. Sushant Varghese, Market Solutions Engineer, New York ISO (Rensselaer, NY)

Dr. Shubo Zhang, Senior Market Solution Engineer, New York ISO (Rensselaer, NY)

Matthew Musto, Principal, Market Solutions Engineering, New York ISO (Rensselaer, NY)

Kanchan Upadhyay, Senior Energy Market Engineer, New York ISO (Rensselaer, NY)

The New York Independent System Operator (New York ISO) administers electricity markets in the state of New York.  A key component of its market clearing engine is the Security-Constrained Unit Commitment (SCUC) model.  However, stringent security constraints can lead to computational slowdowns and feasibility challenges when solving large-scale optimization problems.  These issues can complicate contingency modeling, selection, and monitoring, ultimately limiting the scalability of model accuracy.  In this presentation, we propose a two-stage constraint screening method to identify redundant constraints generated by Network Security Analysis.  Stage 1 employs a computationally efficient, bound-based approach that verifies whether individual constraints would bind under conservative input bounds.  Stage 2 builds on this by utilizing a comprehensive method that integrates both bound and constraint information; this stage leverages linear programming to account for interactions among a predefined set of constraints, thereby identifying an umbrella set of critical constraints.  The screening method is tested on the New York ISO market clearing engine under both static and dynamic reserve requirements.  Results show that removing redundant constraints in the day-ahead market significantly reduces SCUC solution times.  This improvement highlights the potential for enhanced scalability, particularly as the New York ISO accommodates increasing renewable generation, manages uncertain import availability, and implements nodal based, dynamically calculated reserve requirements.

---

2024 Long-Term System Assessment Results for the ERCOT Grid

Dr. Pengwei Du, Supervisor - Economic Analysis & Long Term Planning Studies, Electric Reliability Council Of Texas (Austin, TX)

The bulk transmission network within ERCOT consists of 60-kilovolt (kV) and higher transmission lines and associated equipment.  In planning, for both additions and upgrades to this infrastructure, ERCOT conducts a variety of forward-looking reviews to help ensure continued system reliability and efficiency.  ERCOT’s planning process covers several time horizons to identify and endorse new transmission investments.  The near-term needs are assessed in the six-year planning horizon through the development of the Regional Transmission Plan (RTP).  The Long-Term System Assessment (LTSA) provides an evaluation of the potential needs of ERCOT’s extra-high voltage (345-kV) system in the 10- to 15-year planning horizon.  One key highlight in 2024 is the unprecedented load growth that moved the ERCOT region into a new era of planning.  Both the resource mix and transmission build out needed to support the new era of planning must be carefully evaluated, and the 2024 LTSA intends to provide some insights on these new challenges which the ERCOT region is facing.  The LTSA guides the six-year planning process by providing a longer-term view of system reliability and economic needs.  ERCOT also studies different scenarios in its long-term planning process to account for the inherent uncertainty of planning the system beyond six years.  The goal of using various scenarios in the LTSA is to identify upgrades that are robust across a range of scenarios or more economical than the upgrades that would be determined considering only near-term needs.

---

Session T2 (Tuesday, July 8, 11:15 AM)

---

Pricing Compounded Congestion from Multiple Contingencies in California ISO’s Markets

Dr. Haifeng Liu, Manager, Market Validation, California ISO (Folsom, CA)

Dr. Fan Zhang, Senior Advisor of Power System Technology Development Group, California ISO (Folsom, CA)

Dr. Guillermo Bautista Alderete, Director of Marketing Analysis and Forecasting, California ISO (Folsom, CA)

Dr. Khaled Abdul-Rahman, Chief Information & Technology Officer, California ISO (Folsom, CA)

The reliable operation of the California ISO system requires the management and enforcement of transmission security constraints.  Credible contingencies are identified and managed through California ISO’s markets.  As part of the congestion management process, the markets will redispatch resources to maintain contingency limits and price congestion.  Pricing locations in the system will reflect the marginal cost of congestion based on the effectiveness of relieving congestion.  Under tight system conditions with limited flexibility to redispatch resources, the market may dispatch non-economic bids, even if they are uneconomical, or relax transmission constraints, including contingencies.  Based on practical experience, the California ISO system may encounter scenarios where multiple contingencies are binding for the same protected element.  Typically, in real-time operations, the subset of resources available to manage congestion across multiple contingencies for the same protected element is limited.  As a result, pricing multiple contingencies may not lead to additional congestion relief and may only increase the congestion price.  To address this, California ISO has developed a mathematical formulation that identifies and prices only the most limiting contingency from the set of contingencies being relaxed, mitigating the compounded price effect.  In this presentation, California ISO will explain the mathematical formulation for pricing multiple contingencies and will present actual market cases illustrating the outcomes when using this formulation.

---

The MRI-based Capacity Accreditation: Interpretation and Properties

Dr. Feng Zhao, Manager, ISO New England, Inc. (Holyoke, MA)

Dr. Tongxin Zheng, Chief Technologist, ISO New England (Holyoke, MA)

Dr. Dane Schiro, Principal Analyst, ISO New England (Holyoke, MA)

Dr. Xiaochu Wang, Analyst, ISO New England (Holyoke, MA)

Capacity accreditation reform has been a focus in several regions of the US.  The main objective of these reforms is to align a resource’s accredited capacity with its contribution to reliability.  ISO New England has been developing the marginal reliability impact (MRI)-based capacity accreditation method, a form of marginal accreditation.  This talk will discuss its interpretation and some important properties.  The method will also be compared with ICAP, UCAP, average and marginal Effective Load Carrying Capability (ELCC) methods.

---

Revenue Sufficiency with Scarcity Pricing and Capacity Market: The Role of Coordinated Operating Reserve and Capacity Demand Curves

Dr. Dongwei Zhao, Energy System Engineer, Argonne National Laboratory (Westmont, IL)

Dr. Zhi Zhou, Principal Computational Scientist, Argonne National Laboratory (Westmont, IL)

Dr. Jonghwan Kwon, Principal Energy Systems Engineer, Argonne National Laboratory (Westmont, IL)

Dr. Todd Lovin, Electricity Markets Team Lead, Argonne National Laboratory (Westmont, IL)

Missing money is a long-standing issue in the wholesale electricity market.  This study systematically evaluates the revenue sufficiency of various generation resources, considering the roles of scarcity pricing and capacity markets.  We develop a coordinated simulation framework that integrates energy and capacity markets while dynamically updating operating reserve demand curves (ORDCs) and capacity market demand curves (CDCs) to reflect evolving market conditions.  In the energy market, we characterize ORDCs based on system reliability conditions and clear daily energy markets.  In the capacity market, we construct demand and supply curves by accounting for generation units' energy market profits and investment costs, and then clear the market accordingly.  To capture market dynamics and inform market design, we conduct a comprehensive sensitivity analysis focusing on revenue evaluation.  Specifically, we analyze the impact of varying levels of variable renewable energy (VRE) penetration, changes in the energy price cap, ORDC price cap, update frequency of ORDCs and CDCs, and different reference technologies used in CDC design.  Through coordinated simulations, we assess how ORDCs and CDCs are shaped by system operational conditions and market design factors.  By considering energy and capacity market interactions, we illustrate how energy market revenues influence capacity market demand curves, supply curves, and overall revenues.  Furthermore, we provide insights into how market design choices and system conditions affect the revenue sufficiency of different generation technologies.

---

Session T3 (Tuesday, July 8, 2:00 PM)

---

Interconnection Life Cycle Intelligence and Automation

Brian Fitzsimons, CEO, GridUnity Inc. (Summit, NJ)

Casey Cathy, Vice President, Engineering, Southwest Power Pool (Little Rock, AR)
Deb Le Vine, Director, Infrastructure Contracts &Management, California ISO (Folsom, CA)

The generation interconnection serves as the front door to operational efficiency.  An effective interconnection process gets generation on-line to meet increasing demand and resource adequacy needs.  GridUnity is providing a common platform across transmission providers (ISO/RTO/non-RTO TP’s) and their transmission owner members to accelerate the interconnection study and commissioning process – reducing delays, increasing collaboration and accountability, and resolving interconnection bottlenecks.  The GridUnity platform has created the environment for a powerful network effect to be harnessed across all stakeholders and ultimately support the transformation of interconnection and planning.  Working with leading grid operators like the California Independent System Operator, Southwest Power Pool, their transmission owner members as well as non-RTO transmission providers like Southern Company, the opportunities to increase transparency, identify responsible parties, and track progress throughout the interconnection life cycle are significant.  GridUnity will discuss several new use cases (a. California ISO, first ready first served rating system implementation approved 09/30/2024 and its measurable success, b. ISO/TO network effect benefits across regions c. advanced intelligence usage) of the deployed solution and identify the key data required to be shared through the Interconnection request, study and commissioning process.  These use cases will address implementation challenges as well as the identified benefits from the platform’s intelligence capabilities introduced in Q1 2025.  We will also cover the new automated grid data validation and communication capabilities enabled through the network effect.  The multi-party collaboration is instrumental in meeting the study timeframes established in FERC Order 2023 and continuing the generator and the network upgrade construction progress.  A high-level overview of the technical approach, the importance of working towards standardized agreements amongst stakeholders, and the challenges of coordinating a large set of project participants with differing approaches will be explored.  The presentation will also discuss ongoing work and new opportunities for using AI and ML tools to further validate intake data and analyze historical data to identify best practices in the industry and continuous improvement methods to inform all stakeholders.  The ability to safely streamline the interconnection process using emerging technology tools, communication mechanisms, and intelligent automation is achievable through the industry embracing innovative solutions with proven results.  This is critical at this time to meet the growing demands on the grid and the growing opportunities in electrification.  GridUnity looks forward to participating in the technical conference and sharing our perspective.

---

Accelerating the Interconnection Process with Advanced Software and Automation

Andy Witmeier, Director of Resource Utilization, Midcontinent ISO (Carmel, IN)

Dr. David Bromberg, Co-founder and Chief Executive Officer, Pearl Street Technologies (Pittsburgh, PA)

Midcontinent ISO’s generator interconnection queue has seen a marked increase in requests over the past several cycles, with net capacity approaching or exceeding peak load.  The large volume of projects has presented challenges in expedient processing of the queue.  Midcontinent ISO strives to remain a leader in innovation within the generator interconnection process, with a primary goal of being able to complete the generation interconnection process in a 1-year timeline.  Process improvements and automation will be a large part of being successful in achieving this goal and supporting the Transmission Evolution aspect of Midcontinent ISO’s Reliability Imperative.  To that end, Midcontinent ISO has deployed Pearl Street Technologies’ advanced SUGAR software to automate the Phase 1 study of its interconnection process.  This software is designed to perform power flow analysis, identify network upgrades (NUs), allocate costs using Midcontinent ISO’s Transmission Cost Estimation Guide and Transmission Owner upgrade inputs, and validate constraint mitigation.  In a benchmark exercise against Midcontinent ISO’s DPP 2021 Phase 1 study, Pearl Street’s software came within 1% of total network upgrade cost with a 99%+ reduction in time – cutting down 600+ days to less than 10 days total and just 7 hours of computation time.  Midcontinent ISO utilized SUGAR to complete its DPP 2022 Cycle Phase 1 study, which had been in progress for roughly two years, enabling results to be posted for review within one day of inputs being provided.  This presentation will outline technical details of the SUGAR deployment, including Midcontinent ISO’s effort to outline a transparent process for network upgrade selection based on sound engineering principles, flexibility for Transmission Owners to provide alternative upgrades and refined cost estimates, and Pearl Street’s cloud architecture.

---

Developments of Cluster-Based Generator Interconnection Study Procedures in TVA and Automation Tool in TARA

Luke Ellis, Transmission Planning Engineer, Tennessee Valley Authority (Hixson, TN)

Dr. Seungwon An, Principal Consultant, PowerGEM (Niskayuna, NY)

In collaboration with PowerGEM, TVA has developed a comprehensive cluster-based generator interconnection study procedure and a new automation tool to comply with FERC Order 2023.  This presentation offers an in-depth overview of the new TVA GD study procedures implemented within the PowerGEM TARA software.  The development began with evaluating the Generation Deliverability (GD) methodology used by several US RTOs, including PJM, Midcontinent ISO, and California ISO, tailored to TVA's operational conditions across different seasons such as summer, spring, and shoulder periods.  We discuss the limitations of traditional ERIS comparative analysis methodologies developed over the years and introduce a new refined ERIS methodology with significant benefits.  These benefits include streamlined study case development (all queue generators offline), no need to model network upgrades associated with queue generators, and simpler retool studies with queue withdrawals.  Central to this process is TARA’s advanced flowgate screening technology, which effectively identifies potential thermal violations while considering all possible dispatch scenarios, preventing the oversight of potential N-1 transmission constraints.  We also review TVA-specific GD criteria based on TVA’s seasonal operation practices.  For instance, TVA has benchmarked its Spring ERIS studies against ERIS comparative study methodologies, resulting in a refined flowgate screening-based ERIS study process.  This innovative approach incorporates TVA’s historical generation additions and the TVA pool-wide average generation outage rate, setting a limit on the total dispatch capabilities of new queue generators and mitigating the risk of unrealistic assumptions of large reductions in the dispatch of TVA’s existing generation fleets.

---

Session T4-A (Tuesday, July 8, 4:00 PM)

---

Bridging the Gaps: Co-Optimizing Energy Systems for a Resilient Future

Rob Homer, Senior Product Manager, Energy Exemplar (Salt Lake City, UT)

As the energy landscape evolves, the integration of electricity and gas systems have become a critical challenge for planners, policymakers, and market operators.  With increasing electrification, aggressive regional decarbonization targets, and shifting energy market dynamics, traditional siloed approaches to planning are no longer sufficient.  Regions across North America –from New England’s constrained grid to Texas’ variable renewable generation –illustrate the urgent need for a co-optimized approach that considers the interdependencies between power and gas.  This presentation will explore the risks of neglecting these linkages, drawing on real-world examples where energy supply security, affordability, and emissions goals have been jeopardized by fragmented planning.  It will examine a look into results from a co-optimized ISONE power and gas system.  Results from this study should reshape the way industry leaders approach integrated resource planning, policy analysis, and energy investments.  By leveraging a whole-system view, energy decision-makers can better anticipate supply constraints, optimize resource allocation, and create more resilient energy markets.  Join us for a thought-provoking discussion on how sector coupling and co-optimization are shaping the future of energy reliability.

---

Dynamic Reserve Requirements

Dr. Chen-Hao Tsai, Advisor II Operations Risk Assessment, Midcontinent ISO (Carmel, IN)

Dr. Arezou Ghesmati, Senior AI/ML Scientist, Midcontinent ISO (Carmel, IN)

Dr. Bing Huang, Research Engineer, Midcontinent ISO (Carmel, IN)

The shifting energy landscape is driving Midcontinent ISO to proactively adapt to growing uncertainty and heightened reliability risks.  With the rapid expansion of variable generation resources, Midcontinent ISO has observed increasing frequency and intensity of uncertainty across multiple timeframes –from multi-day forecasts to real-time system ramps.  To address these challenges and enhance system flexibility, Midcontinent ISO has initiated a series of enhancements to its reserve products, including Short-Term Reserve, Ramp Capability, and Regulation Reserve.  These improvements generally follow a two-phase approach.  The first phase focuses on refining uncertainty quantification and establishing corresponding reserve requirements to capture long-term trends and seasonal variations.  The second phase introduces a dynamic process that enables Midcontinent ISO to proactively increase reserve requirements when elevated uncertainty is anticipated.  Since January 2024, Midcontinent ISO has implemented a dynamic short-term reserve and next-day uncertainty forecasting process powered by advanced AI-based models.  This new approach has successfully enabled Midcontinent ISO to increase short-term reserve requirements in the Day-Ahead market based on forecasted uncertainty – proving especially effective during recent winter storms Heather (January 2024) and Enzo (January 2025).  Midcontinent ISO’s next area of focus is enhancing Ramp Capability and Regulation Reserve.  In the initial phase, Midcontinent ISO is conducting a cost-benefit analysis of various Ramp Capability Requirement levels using market simulations, which will inform future requirement recommendations.  A real-time market simulation tool is currently under development to assess the benefits and costs associated with different uncertainty-based reserve requirements.  These efforts aim to provide recommendations for replacing the existing static uncertainty component in Midcontinent ISO’s current Ramp Capability Up requirement.  To support both Ramp and regulation requirements, Midcontinent ISO will utilize advanced AI-based pattern recognition techniques to identify recurring trends, sudden shifts, and temporal correlations within net uncertainty.  This approach will aid analysts in detecting and understanding unusual system behavior, providing a solid foundation for developing and refining net uncertainty forecast models to dynamically configure Ramp and regulation needs.

---

Modernizing Grid Risk Management with AI Advances and Cloud Technology 

Dr. Arezou Ghesmati, Senior Data Scientist, Midcontinent ISO (Carmel, IN)

Dr. Long Zhao, Principal Operations Risk Assessment Advisor, Midcontinent ISO (Carmel, IN)

Dr. Concong Wang, Director Markets and Grid Research, Midcontinent ISO (Carmel, IN)

Evan Sattler, Senior Data Scientist, Midcontinent ISO (Carmel, IN)

The Operations Uncertainty Platform is a multi-year, multi-million-dollar effort funded in part by a $3 million U.S. Department of Energy grant.  It serves as a centralized solution for scenario-based risk assessment, providing the flexibility and scalability needed to manage increasing variability and uncertainty in large-scale electric grid operations.  Key features of the platform include: (1) a centralized data repository housing critical risk components such as weather, renewable, and load forecast scenarios and actuals, along with fuel supply and infrastructure risks; (2) AI/ML-powered predictive analytics for generating operational scenarios, quantifying uncertainties, predicting risks, and informing dynamic reserve and unit commitment decisions; (3) advanced visualizations that deliver actionable insights; and (4) APIs to integrate with downstream tools and inform public stakeholders.  A core focus of the Operations Uncertainty Platform is the integration of AI/ML-powered forecasting models to enhance the accuracy of renewable generation and load forecasts, while also identifying operational risks driven by net uncertainty.  These models support dynamic reserve adjustments and unit commitment decisions to strengthen grid reliability.  As the grid shifts toward higher levels of renewable energy, managing variability has become increasingly vital.  To meet this need, Midcontinent ISO’s data science team is developing advanced statistical and AI-driven systems that intelligently select or blend multiple vendor forecasts, ensuring the most accurate and actionable insights for reliable grid operations.  To that end, multiple AI-driven approaches have been explored, including real-time decision-making algorithms designed to switch between different renewable forecast vendors based on performance metrics.  These models have been rigorously tested across several months – from early 2024 through 2025 – demonstrating superior accuracy and reduced bias compared to forecasts from any single vendor.  In parallel, the team has developed machine learning-based ensemble models that utilize outputs from all available forecast providers to generate an internally optimized renewable forecast.  Evaluation of both the intelligent vendor selection system and the ensemble forecasting model against individual vendor outputs has shown significant improvements in forecast accuracy and operational efficiency, highlighting their potential to enhance Midcontinent ISO’s forecasting capabilities in support of grid reliability. 

---

Session T4-B (Tuesday, July 8, 4:00 PM)

---

Modeling of Internal Controllable HVDC Lines in Energy Market Operations

Dr. Hossein Lotfi, Senior Market Solutions Engineer, New York ISO (Rensselaer, NY)

Bo Yuan, Graduate Researcher, Cornell University (Ithaca, NY)

Dr. Muhammad Marwali, Energy Market Engineering Manager, New York ISO (Rensselaer, NY)

New York State’s Climate Leadership and Community Protection Act (CLCPA) aims to deliver 70% of New York’s energy from renewable resources by 2030 and 100% emission-free electricity by 2040.  However, transmission constraints limit clean-energy technologies in upstate NY from serving southeast NY, where most of the electricity demand is located.  The internal controllable line (ICL) was proposed to alleviate transmission constraints and to deliver clean power from upstate NY to NYC.  It is a nominal 1,300 MW underground HVDC line between Fraser substation in upstate NY and Rainey substation in NYC.  Since both ends of this HVDC line are inside the New York control area, the New York Independent System Operator, Inc. (New York ISO) will have full control of the asset to schedule power.  It is distinct from HVDC interconnectors between the New York ISO and its neighboring ISOs/RTOs, which link two control areas and are not scheduled solely by the New York ISO.  Therefore, the HVDC line is referred to as an ICL.  This presentation proposes a generalized locational marginal price (LMP)-based energy market model for ICLs.  A comparison between ICL, wheel, and PAR models is provided and the advantages of ICL are highlighted in the presentation.  A key advantage of the proposed model is that the ICL operator can bid competitively in the energy market in both directions of power flow to maximize its profit.  In addition to bidding competitively in both directions, transmission losses are modeled and the power flow on the ICL is optimized to leverage the LMP difference and the cost of the ICL transmission losses.

---

Optimizing DC Tie Utilization in Southwest Power Pool’s Expanded RTO

Seth Mayfield, Manager of Market Design, Southwest Power Pool (Little Rock, AR)

Caroline Chapman, Environmental Markets Director, Southwest Power Pool (Little Rock, AR)

Southwest Power Pool (SPP), a Regional Transmission Organization (RTO), is expanding its footprint into the Western Interconnection, becoming the first RTO to operate a single market across both the Eastern and Western Interconnections of the U.S. power grid.  This unprecedented expansion will introduce a unique operational challenge: managing energy flows across two asynchronous Balancing Authority Areas (BAAs) using three Direct Current (DC) Ties.  Historically, DC Tie flows have been scheduled through hourly tags, limiting the ability to dynamically respond to real-time system conditions and market price signals.  Post-expansion, SPP will implement an optimized approach to DC Tie scheduling, leveraging advanced economic dispatch and optimization algorithms to determine both the direction and magnitude of flow on a Day-Ahead and Real-Time basis.  By integrating DC Tie operations into its market structure, SPP will dynamically adjust flows as frequently as every five minutes, ensuring more efficient utilization of transmission capacity and improved market efficiency.  This presentation will explore the software and algorithmic innovations behind this optimization, including: • Economic Dispatch Enhancements: Leveraging price differentials and grid constraints to optimize DC Tie flows. • Real-Time Adjustments: Transitioning from static hourly schedules to rapid 5-minute adjustments. • Grid Stability Considerations: Balancing economic efficiency with reliability requirements across two asynchronous systems.  By optimizing DC Tie utilization, SPP aims to facilitate the lowest-cost delivery of power across its expanded footprint, improving market efficiency, integrating more renewable energy, and enhancing overall grid resilience.  This session will provide insights into the software challenges, data-driven methodologies, and real-world implementation strategies that will enable this transformation.

---

How Software can be Used to Automate Much of Today's Transmission Planning Workload

Andrew Martin, Co-founder and Head of Transmission, Nira Energy (Denver, CO)

Nira is a software company founded on the premise that much of today’s transmission planning workload should be automated.  Neary 100 of the largest developers in the country are using our software to get study results on the order of hours instead of weeks or months.  The interconnection study process follows a handful of steps: 1. A set of initial assumptions are established (e.g. applicant projects) 2. Models are built from those assumptions 3. The models are run through power flow software (e.g. PowerGEM’s TARA) 4. Results are reviewed by engineers 5. Transmission solutions are developed 6. Costs for solutions are allocated to projects.  Running a study requires orchestrating the flow of data from one step to the next, and there are often hundreds of manual iterations made between each step.  This means that most of a transmission planner's time is spent manually handling data in between each step – work that code can ideally handle.  This makes the process prone to human error and challenging to replicate.  Nira has automated all of these steps on top of industry-standard power flow solvers like TARA.  Existing solvers like TARA work very well for their specific purposes like model building and power flow analysis.  We believe that rebuilding the solver is not an efficient path to interconnection study process improvements.  Solvers such as TARA have been fine tuned for the past 20+ years, and replacing them does not address the core problem.  ISOs and software companies should think of existing solvers as infrastructure to build on top of.  Nira does this by layering process automation on top of solvers like TARA and as a result, we have been able to dramatically improve the speed at which we can run interconnection studies nationwide For some examples of the benefits that this can deliver, this has helped:  • Reduce study timelines down from months to hours: Our customers are viewing results for their projects immediately, rather than waiting months. • Increase repeatability of the process: We have helped both ISO’s and developers identify over $500M of cost allocation errors. • Fill the labor shortage gap:  Three engineers on our team can facilitate analysis for several hundred projects at a time.  Doing this manually within the same timeframe would take 30 engineers.  In summary, automation should facilitate all steps of the study process.  It will by no means replace engineers or be fully automated, but we have room to over 10x the speed and volume of engineering workflows while improving consistency and accuracy.

 

Session W1-A (Wednesday, July 9, 9:00 AM)

---

Tapestry’s AI-enabled Interconnection Automation Software: Speeding Grid Interconnections and Enabling Dynamic Planning and Operations

Dr. Cat Wong, Head of Technical Operations, Tapestry (Mountain View, CA)

AJ Lambert, Interconnection Planning Projects Manager, PJM Interconnection (Pottstown, PA)

Tapestry is at the forefront of developing and deploying artificial intelligence (AI) driven software solutions designed to revolutionize the electric grid interconnection process and enhance planning efficiency.  The increasing penetration of energy sources and the growing complexity of modern power systems necessitate a paradigm shift in how we approach grid planning and operation.  Traditional interconnection processes are often time-consuming, resource-intensive, and prone to bottlenecks, hindering the rapid integration of energy resources and growing customers.  Tapestry’s innovative software addresses these challenges by automating and optimizing key aspects of the interconnection workflow and planning analysis.  Our software solutions aim to significantly improve the efficiency of the interconnection process, a critical component of the integration process into holistic planning analysis.  By leveraging AI and machine learning algorithms, we are developing tools that can automate the analysis of vast datasets, including grid topology and generation profile.  This automation reduces the reliance on manual processes, minimizing human error and accelerating the study timeline.   In addition to automating interconnection studies and power flow analyses, Tapestry is developing software that automates other critical parts of the interconnection and planning process.  We’ve worked with Coordinador Electrico Nacional, Chile’s national system operator, to make their grid planning process faster and more efficient.  With Tapestry’s Grid Planning Tool, planners can run simulations 86% faster than legacy tools with 30 scenarios simultaneously instead of one at a time.  These improvements bring unprecedented speed and granularity to the grid planning process.  On the interconnection front, this includes streamlining the review of technical documents, organizing the massive data and managing the grid models.  By automating these tasks, we reduce administrative overhead and accelerate the overall interconnection timeline.  More broadly, Tapestry software is on the path to providing grid planners, operators and market participants the ability to plan, interconnect, and operate the grid all in one place.  In summary, Tapestry's AI-driven software solutions are designed to address the pressing challenges facing the electric grid, including the need to accelerate renewable energy integration and enhance grid resilience.  By automating and optimizing key processes, improving model accuracy, and enabling faster decision-making, our software is paving the way for a more efficient, reliable, and sustainable energy future.  Our focus on leveraging AI and machine learning to improve the interconnection process and enhance market efficiency positions Tapestry as a leader in the development of innovative solutions for the modern grid.

---

Solving Mathematical Programs with Embedded Surrogate Models using GAMSPy

Dr. Adam Christensen, Senior Technical Staff, GAMS Development Corporation (Fairfax VA)

Hamdi Burak Usul, GAMS Development Corporation (Fairfax, VA)

Muhammet Soyturk, Software Engineer, GAMS Development Corporation (Fairfax, VA)

Dr. Steven Dirkse, President, GAMS Development Corporation (Fairfax, VA)

Dr. Michael Bussieck, CEO, GAMS Development Corporation (Fairfax, VA)

Recent advances in machine learning (ML) and artificial intelligence (AI) have commoditized the creation of detailed models used to describe complex systems by leveraging tools such as PyTorch, Scikit-Learn, and TensorFlow.  These surrogate models simplify non-linear systems so that they can be incorporated as constraints within traditional optimization frameworks.  Embedding such models into algebraic modeling languages (AMLs), like GAMS and AMPL, has posed significant challenges because these systems were purpose-built for symbolic (sparse) algebra and are not easily integrated with the rapidly evolving world of third-party software packages.  The rise of Python as a dominant language for data science and ML has spurred a paradigm shift, motivating the development of tools that bridge the gap between classical algebraic modeling and modern computational techniques.  In this work we highlight GAMSPy –  a native Python AML that combines the mathematical transparency and scalability of traditional AMLs with the versatility of Python’s expansive ecosystem.  GAMSPy is engineered to maintain the syntactical advantages of handwritten algebra while accommodating the computational requirements of modern ML applications.  It achieves this through extensive operator overloading and set-driven constructs, which allow users to write algebraic expressions that closely mirror their handwritten counterparts.  GAMS “classic” – the performance engine used to generate models and interface with solvers – excels in creating indexed algebra.  However, ML/AI methods predominantly rely on matrix operations (which are dense data structures by definition).  We introduce new features and constructs into GAMSPy that enable essential ML operations such as matrix multiplication, transpositions, and norms.  In this talk, we showcase the use of these new tools to embed a trained neural network (NN)  model.  The NN was trained with Pytorch to describe a complicated energy system that was then used as a constraint set within a traditional optimization framework.  This framework allows the systems engineer to optimize the operation of the overall plant while taking into account the details of energy conversion.  This workflow is illustrative of other similar modeling efforts/research to incorporate ML models of complex weather patterns, market behaviors, etc. a variety of optimization models useful to a wide range of disciplines.  We delve into future prospects, show how GAMSPy's approach differs from existing alternatives and discuss innovative methods where mathematical modeling intersects with machine learning.  In summary, GAMSPy represents a significant convergence of classical algebraic modeling and modern computational methodologies.  Its design emphasizes computational efficiency, syntactic clarity, and scalability, thereby providing a robust platform for researchers and practitioners.  By uniting the rigor of AMLs with the dynamic capabilities of Python and ML frameworks, GAMSPy not only resolves longstanding integration challenges but also paves the way for innovative applications at the intersection of mathematical modeling and data science.

---

Building Open Source AI Solutions and Foundation Models for Market and Planning Efficiency

Dr. Alexandre Parisot, Director of Ecosystem AI and Energy Systems, Linux Foundation Energy (Menlo Park, CA)

Dr. François Mirallès, Researcher, Hydro-Québec (Varennes, Canada)

Dr. Jonas Weiss, Senior Research Scientist, IBM Research (Rueschlikon, Switzerland)

Dr. Thomas Brunschwiler, Principal Research Scientist, IBM Research (Rueschlikon, Switzerland)

Dr. Hendrik Hamann, Chief Science Officer for Climate and Sustainability, IBM Research (Yorktown Heights, NY)

This presentation will highlight recent advances in open source Artificial Intelligence (AI) projects to address increasing challenges of the power grid.  The projects are currently developed through interdisciplinary collaboration among utilities, system operators, researchers, and vendors.  Challenges they address range from grid forecasting, planning, and real-time operations, aiming at providing modular, reusable tools.  Artificial intelligence (AI) has emerged as a powerful technology, with great potential to improve market efficiency and reliability in power systems.  More recently the concept of foundation models, AI models that are pre-trained on large data-corpuses in a self-supervised way – meaning no data annotation is required – has even further increased AI capabilities.  For the power grid, this constitutes a superb opportunity to ingest readily available extensive power systems data into highly expressive grid foundation models.  Once trained, their generalization capability can serve many different “downstream tasks” by re-training (fine-tuning) them on small samples of application-specific data and deploying them.  This ambition is being pursued through open source collaboration in the context of the GridFM project hosted by LF Energy.  With increasing advances of current grid foundation models, we already see great potential on power-flow related, but also optimal power flow related applications.  By adding time-series, missing value state-reconstruction can gradually be expanded into state-forecasting, grid-dynamics, electromechanical transients or quasi static time series simulations for expansion planning, congestion management, and stability analysis.  We will also add additional modalities like protection- and control actions, market-data, weather and climate, telecom, IT, news-feeds or social media.  By further guiding the model training by penalizing violations of constraints, operational practices or policies or missing market targets, GridFM’s usability will easily span from power flow related problems, state estimation, control, trading and markets, load-, generation- and other forecasting tasks all the way to decision support.  This constitutes a truly scalable industry solution.  To reach this goal, we aim at sharing data, models and model-weights as openly as possible.  For this we will have to understand and properly address regulatory and (cyber)security/privacy concerns of the data-owners.  Distributed or federated learning for privacy preservation and extensively testing and challenging pre-trained models will be among many techniques used.  Also, many stakeholders will be able to share the development cost of this endeavor and jointly develop equitable and sustainable business-models.  Another step into the same direction is to learn from, collaborate with, build on top and mutually complement existing open source activities with similar and related goals.  These include Grid2op, a flexible simulation platform for power grid operations, OpenSTEF, delivering automated pipelines for short-term load forecasting using machine learning techniques, or OpenSynth, supporting the creation of synthetic smart meter data, e.g. required for training.  These projects also contribute to shared infrastructure for AI readiness – such as open benchmarks, explainability frameworks, and approaches to data governance – supporting the development of trustworthy and auditable AI tools.  As interest grows in applying AI to critical grid operations, open collaboration is accelerating the deployment of efficient and secure solutions across the power system.

---

Session W1-B (Wednesday, July 9, 9:00 AM)

---

Prioritizing Interconnection Applications via High-Fidelity Stochastic Capacity Expansion Planning

Dr. Elizabeth Glista, Power System Researcher, Lawrence Livermore National Laboratory (Livermore, CA)

Dr. Jean-Paul Watson, Senior Research Scientist, Lawrence Livermore National Laboratory (Livermore, CA)

Constructing a more resilient power grid is of growing importance due to a variety of concerns including booming demand for electric power due to data center growth.  However, the slow and uncertain regulatory process of grid interconnection limits the ability of new generation, storage, transmission, and large-load projects to come online in a transmission system.  While the amount of generation and storage capacity in U.S. interconnection queues exceeds 2,000 GW, the median time for a project to complete all the required studies and reach commercial operation is 5 years (based on 2023 data from LBNL).  Congestion in the interconnection queue is a major contributor to this time delay.  Additionally, many of the active projects in the existing interconnection queue may not contribute significantly to system-level goals like resource adequacy (RA).  Motivated by this, our work leverages an existing nodal capacity expansion planning (CEP) model to prioritize active projects in the interconnection queue.  We consider a two-stage stochastic CEP problem with uncertainty in generator availability and load based on high-fidelity weather predictions with high spatial and temporal resolutions, as obtained from a Regionally Refined Mesh (RRM) of the E3SM earth systems model.  Using this CEP framework, we demonstrate how active projects in the interconnection queue can be prioritized based on RA metrics like expected unserved energy (EUE) and how we can use CEP to understand the contribution of new projects to grid resilience under stressed conditions.  We solve this optimization problem using the parallelizable decomposition algorithm Progressive Hedging on Lawrence Livermore's High Performance Computing (HPC) platform.  To demonstrate the efficacy of our model, we consider the current set of active projects in California ISO's interconnection queue on a CEP model over a large-scale, synthetic-but-realistic test case of California with around 8,000 buses and a year of representative future weather data.  We show that our work can be used to prioritize existing projects in the interconnection queue, via an optimization-based and fair open-source modeling approach.

---

Managing Variability with High Fidelity Capacity Expansion

Dr. Russ Philbrick, President, Polaris Systems Optimization (Shoreline, WA)

Capacity expansion modeling is difficult because there are many choices and these need to be evaluated over time frames that are often decades long.  By necessity, we use simplified models that can be solved in a modest amount of time, and with a reasonable level of fidelity.  However, even with simplified representations of time using representative days or weeks, most models lack important operational constraints on resource flexibility and lack system constraints on powerflow.  Among other issues, this has led to the use of zonal model because nodal powerflow modeling is viewed as too hard.  However, the use of zonal models leads to inaccurate and even harder planning processes.  Nevertheless, high-fidelity models can be solved using more-efficient methods to represent time in capacity expansion models.  This presentation will present the benefits of "implicit chronology" based on the wrapper-constraint formulation.  These methods can adequately represent system variability and resource time coupling, particularly for storage, while avoiding many of the simplifications traditionally assumed as necessary.

---

PTDF Powerflow Representation for Accelerating Large-scale Stochastic Nodal Capacity Expansion Planning

Tomas Valencia Zuluaga, Postdoctoral Researcher, Lawrence Livermore National Laboratory (Livermore, CA)

Amelia Musselman, Systems Analyst, Lawrence Livermore National Laboratory (Livermore, CA)

Dr. Jean-Paul Watson, Senior Research Scientist, Lawrence Livermore National Laboratory (Livermore, CA)

Dr. Shmuel Oren, Professor, University of California at Berkeley (Berkeley, CA)

Ensuring the reliability and resilience of the modern power grid requires models that handle the inherent uncertainty of generation availability and electricity consumption.  These models must have sufficiently high spatial and temporal resolution to adequately capture weather variability and provide actionable siting and sizing decisions.  A stochastic nodal capacity expansion planning (CEP) model can satisfy these requirements, but has challenges of computational scalability.  We employ a power transfer distribution factors (PTDF) representation of power flow constraints within a stochastic nodal generation, transmission, and storage CEP model to improve scalability.  We implement a method to address the changes in system topology that occur during the optimization when this type of model is used.  The method proposed consistently outperforms the more commonly used b-theta formulation of linear DC power flow in terms of the resulting optimal cost and computational solution time when applied to realistic test systems based on California and South Carolina.

---

Session W-2-A (Wednesday, July 9, 10:45 AM)

---

Generative AI Software for High-Fidelity Household Load Profiles: Innovations in Resource Adequacy Modelling and Planning Efficiency

Gareth Jones, Chief Operating Officer, Octopus Energy Group (London, United Kingdom)

Gus Chadney, Data Lead, Octopus Energy Group (Marlow, United Kingdom)

Sheng Chai, Senior Data Scientist, Octopus Energy Group (London, United Kingdom)

Amber Woodward, International Public Affairs Manager, Octopus Energy Group (London, United Kingdom)

Our energy system is undergoing immense and rapid change.  Generation is decentralizing; power flows are increasingly bidirectional; we’re relying more on weather-dependent renewables.  Meanwhile, consumption is spiking with electrification; behaviors are fast evolving with low carbon technology (LCT) adoption, smart tariffs and changing weather patterns.  Despite these changes, decision-makers are constrained by outdated and aggregated energy data, which hinders accurate demand profiling, real-time planning, and long-term resource adequacy assessments.  To enhance market efficiency and reliability of the bulk power system, grid operators, policymakers, and researchers require access to granular, high-fidelity demand data.  The digitalization of energy presents a growing opportunity to maximize data use across all services and markets – particularly in resource adequacy frameworks, real-time and day-ahead planning.  However, Advanced Metering Infrastructure (AMI) data and other data sources come with privacy, proprietary, and cybersecurity risks that prevent the wide sharing and dissemination of data for modelling purposes.  Furthermore, the absence of labelled attributes, such as Low Carbon Technologies, limits the ability to represent load diversity in resource adequacy modelling.  Current approaches rely on computationally intensive, physics-based models to estimate demand, yet these lack scalability and adaptability to emerging consumption patterns.  Synthetic smart meter data offers a breakthrough solution, drawing on pioneering approaches in healthcare, finance and technology sectors.  By using generative AI techniques to generate high-fidelity synthetic datasets and taking an open-source approach, synthetic data allows policymakers, regulators, grid operators to improve real-time and day-ahead market forecasting, enhance grid reliability, and better model the impacts of distributed energy resources (DERs) and storage – without compromising consumer privacy.  Faraday, built by Centre for Net Zero, is a Generative-AI model to produce synthetic, household-level AMI data.  The Variational Auto-Encoder (VAE)-based model is trained on over 1.8-billion-meter readings over one year from the largest energy supplier in the UK.  Faraday produces household-level synthetic load profiles consisting of half-hourly kWh consumption.  Profiles can be modified based on defined user-specified inputs such as days of week, months of year, LCT ownership, Energy Performance Certificate (EPC) rating, property type, and tariff type.  The model is currently being tested with AMI data provided by a US state government agency to ensure its applicability across different regulatory and market environments.  In collaboration with Linux Foundation Energy, Centre for Net Zero has established an international data community, Open Synth, to facilitate the sharing of synthetic AMI data and algorithms to scale quickly, particularly in areas with limited data access.  Supported by research and industry leaders – including Delft Technical University, Alliander, HydroQuebec, and the University of Oxford – OpenSynth is an open-source one-stop-shop for synthetic smart meter data, including: (i) Model repository to host algorithms for generating synthetic data, on Github; (ii) Data repository to host synthetic data sets, on Hugging Face; (iii) A community for quality assurance, in particular to agree on common evaluation frameworks.  Further models and datasets are anticipated to be added throughout 2025, including an expansion from household-level to Commercial & Industrial data, and other synthetic datasets of relevance for Linux Foundation Energy’s project – GridFM.  By harnessing synthetic data, electric power system operators, government, research and academia can address fundamental challenges in resource adequacy assessments, market planning, and support the development of novel AI-based approaches to ensure the reliability and efficiency of the power system.

---

Machine Learning Enhanced Formulation Tightening of Energy Storage Resource Constraints in Unit Commitment

Farhan Hyder, Ph.D. Candidate, Rochester Institute of Technology (Rochester, NY)

Uyen Nhi Quang, Student, Rochester Institute of Technology (Rochester, NY)

Dr. Bing Yan, Assistant Professor, Rochester Institute of Technology (Rochester, NY)

Recent years have seen an increased penetration of utility-scale energy storage resources (ESRs) into the power grid.  To incorporate ESRs into the unit commitment (UC) problem, binary variables are required to prevent simultaneous charging and discharging.  This further increases the problem's complexity, prompting computationally efficient modeling of ESRs.  A systematic formulation tightening approach based on constraint-to-vertex conversion, developed in our previous work, has shown effectiveness in tightening formulations of conventional generators.  However, it requires extensive manual analysis and expert knowledge during the parameterization step (expressing numerical coefficients as combinations of generator parameters).  In this paper, the systematic formulation tightening approach is enhanced via machine learning for ESRs.  To address the challenge caused by the manual analysis for parameterization, a novel machine learning model is developed by taking ESR parameters as inputs and numerical coefficients of the tight constraints as outputs to identify the mathematical relationship between them.  Furthermore, the entire tightening process is automated for flexibility and applicability.  Results based on the IEEE 118-bus system show a significant reduction in solving time (up to 45%) while maintaining the solution quality of UC with ESRs, demonstrating the effectiveness of the approach.  The approach is general and has great potential for tightening complicated MBLP problems in power systems and beyond.

---

Session W2-B (Wednesday, July 9, 10:45 AM)

---

A Novel Partitioning Algorithm for Constructing Diverse Solution Sets for the Unit Commitment Problem

Dr. Ignacio Aravena, Operations Research Engineer, Lawrence Livermore National Laboratory (Livermore, CA)

Jisun Lee, Ph.D. Candidate, University of California at Berkeley (Berkeley, CA)

Dr. Jean-Paul Watson, Senior Research Scientist, Lawrence Livermore National Laboratory (Livermore, CA)

Dr. Alper Atamturk, Professor and Chair, University of California at Berkeley (Berkeley, CA)

A known challenge in solving unit commitment problems is the large number of acceptable solutions that lie within any practical optimality tolerance.  During operations, existing software arbitrarily selects one among the set of acceptable solutions.  Whereas these acceptable solutions are similar in total cost (or welfare), they can lead to radically different outcomes in other aspects, such as prices, transfers, and reliability.  A number of studies in the literature have relied on enumeration techniques (no-good cuts) to produce and investigate the variations among different acceptable solutions.  These techniques, however, struggle to obtain diverse subsets of acceptable solutions, requiring long computations to achieve the level of diversity that allows the study of differences among them.  In this talk, we present a novel enumeration approach, designed to generate diverse subsets of acceptable solutions for the unit commitment problem, aiming at achieving the maximum diversity within a set wall clock time limit.  Instead of selecting and removing points from the acceptable solution set one-by-one, our approach removes them in pairs, selecting the most distant solutions at each iteration.  This two-solution removal step is followed by a careful (non-intersecting) partition of the reminder of solution space.  The two-solution removal is then recursively carried out for each remaining component of the partition, a process that we organize using a queue, and later parallelize using high-performance computing.  We discuss (integer programming) strength properties of our approach in comparison with alternatives and discuss details of its implementation.  Finally, we present a numerical comparison of our approach against approaches from the literature, showing how it is able to produce more diverse sets of solutions at a similar or smaller computational cost.

---

Next-Generation Market Approaches to Support System Stability and Reliability

Dr. Allison Campbell, Power Systems Data Scientist, Pacific Northwest National Laboratory (Portland, OR)

Dr. Matthew Cornachione, Senior Scientist, Pacific Northwest National Laboratory (Richland, WA)

Dr. Jesse Holzer, Senior Scientist, Pacific Northwest National Laboratory (Richland, WA)

Molly Rose Kelly-Gorham, Ph.D. Candidate, Pacific Northwest National Laboratory (Richland, WA)

Dr. Liping Li, Electrical Engineer, Pacific Northwest National Laboratory (Richland, WA)

Dr. Ki Yeob Lee, Postdoctoral Researcher, Pacific Northwest National Laboratory (Richland, WA)

Dr. Eran Schweitzer, Power Systems Engineer, Pacific Northwest National Laboratory (Portland, WA)

Traditional bulk electricity markets do not adequately account for the variability and uncertainty of renewable energy (RE) resources, which are already deployed at high penetration in some U.S. markets.  The increased dependence on renewable resources has led to market volatility, where increasingly frequent negative price bids from non-fuel resources impact investment recovery for firm capacity, and system disturbances, where limited visibility into inverter setpoints strain system stability (California ISO, 2020).  Bulk electricity markets have historically been constructed to minimize the cost of fuel required to operate generators while maintaining system reliability through constraints on assets.  While in principle this valuation approach incentivizes the dispatch of least-cost resources, in practice system operators often override their market solutions because RE resources are susceptible to dramatic ramps in production (variability) and forecast errors (uncertainty).  This results in energy payments to firm capacity exceeding market value (California ISO 2023) and unused generation from RE which had been rate-based by customers.  Most of these resources are sited in rural, transmission-constrained regions, where variability can lead to system overload, resulting in voltage drops and regional blackouts.  Uncertainty compounds this, as the need to balance demand with supply requires setting aside enough dispatchable generation to absorb forecast errors.  These grid reliability and stability issues are already being seen in areas with high penetration of renewable energy (NERC 2022).  Given the prevalence of RE resources already participating in the U.S. power generation mix, new market formulations and valuation strategies are needed to ensure reliability, resiliency, security and affordability.  This talk describes recent work developed at PNNL to introduce new value stream approaches into the security constrained unit commitment (SCUC) objective formulation to incentivize reliable and affordable power delivery.  The new approaches described in this talk include: (1) valuation of uncertainty from all units (RE and thermal) to standardize resource predictability, (2) new market products for inertia and damping to compensate all resources for the stability they provide to the system, and (3) risk-endogenous reserves where units are commensurably awarded for their performance and can thereby improve their market-participant reliability.  These new approaches are expected to improve system stability by valuing and minimizing the risk inherent to variable and uncertain generation.  Finally, this talk will provide guidance on metrics to adequately demonstrate that proposed model formulations achieve the desired outcome in the bulk power system.

---

Panel W2-A (Wednesday, July 9, 12:00 PM)

---

Panel Discussion on AI Integration in Power Markets

Discussion with several panelists on current and potential future integration of AI into power markets.  

Panelists:

Dr. Adam Christensen (Senior Technical Staff, GAMS Development Corporation)

Brian Fitzsimons (CEO, GridUnity)

Andy Witmeier (Director of Resource Utilization, Midcontinent ISO)

Moderator:

Dr. Emma Nicholson (Federal Energy Regulatory Commission)

---

Session W3-A (Wednesday, July 9, 2:00 PM)

---

Advancing Market Efficiency and Reliability: Forecasting Emissions and Asset Risk in Electricity Systems

Dr. Andres F. Ramirez, Postdoctoral Scientist, Lehigh University (Bethlehem, PA)

Dr. Alberto J. Lamadrid, Associate Professor, Lehigh University (Bethlehem, PA)

As power systems evolve to incorporate higher levels of renewable energy, understanding emissions forecasting and generator risk is critical to ensuring market efficiency and grid reliability.  This work integrates advanced forecasting techniques with risk assessment methodologies to enhance real-time and day-ahead market operations.  We leverage historical emissions data from the U.S. Environmental Protection Agency (EPA) and optimization tools from the Advanced Research Projects Agency-Energy (ARPA-E) PERFORM project to project 2030 pollutant emissions scenarios in the New York Independent System Operator (New York ISO) region.  Additionally, we introduce a machine learning framework that quantifies generator risk by training models on Security Constrained Economic Dispatch (SCED) and Security Constrained Unit Commitment (SCUC) simulations.  This operational risk score provides a data-driven measure of the likelihood that a generator will cause imbalances due to variability and uncertainty.  By combining emissions forecasting with risk-aware optimization, this research contributes to improving market efficiency, informing policy decisions, and supporting the integration of clean energy resources into the bulk power system.

---

Factor-Based Portfolio Optimization for Risk Management in Electricity Markets with Renewable Energy

Dr. Audun Botterud, Principal Research Scientist, Massachusetts Institute of Technology (Cambridge, MA)

Dr. Yusu Liu, Research Scientist, Massachusetts Institute of Technology (Cambridge, MA)

Dr. Arnab Sur, Research Scientist, Lehigh University (Bethlehem, PA)

Dr. Alberto J. Lamadrid, Associate Professor, Lehigh University (Bethlehem, PA)

The large-scale integration of variable renewable energy (VRE) sources to decarbonize the power sector increases the need for appropriate risk management in electricity market operations.  We propose a novel factor-based portfolio optimization approach to manage VRE risks.  The method applies Principal Component Analysis to decompose forecasting errors, while accounting for their geographical correlations.  In turn, we impose the resulting factor loadings and corresponding risk quotients for individual VRE assets on the day-ahead electricity scheduling and market clearing formulation.  In a case study based on data from the New York Independent System Operator (ISO) system, we illustrate how the factor-based portfolio construction reduces system risk exposure through lower real-time shortfalls from VRE resources.  We also simulate resulting day-ahead scheduling and real-time dispatch.  The results show a substantial reduction in average real-time prices, with over 85% of the locations in the system seeing reductions in prices.  Moreover, simulations also indicated substantial reductions in expected system cost and expected unserved demand.

---

On the Significance of High-Fidelity Resource Adequacy Assessment:  The Case of New York ISO

Dr. Aleksandr Rudkevich, CEO, Newton Energy Group LLC (Newton, MA)

Dr. F. Selin Yanikara, Energy Research Analyst, Newton Energy Group LLC (Newton, MA)

Russ Philbrick, CEO and CTO, Polaris Systems Optimization (Shoreline, WA)

The presentation contrasts the differences in assessing resource adequacy of constrained power systems using traditional methods relying on stylized system representations and on high-fidelity models incorporating the full network representation and operational details.  Both methodologies are implemented within the same modeling platform and datasets *) thus assuring full comparability of results.  We outline the key computational steps for assessing power system reliability using chronological Monte Carlo replications of the MIP based SCUC/SCED scheduling and nodal price formation under multiple weather scenarios.  We analyzed traditional system-wide adequacy indicators such as Loss of Load Hours (LOLH) and Expected Unserved Energy (EUE) as well as the nodal reliability metrics based on the statistical analysis of dual variables reported by the optimization engine.  The system-wide and locational results obtained from the high-fidelity model are compared to those produced by Monte Carlo replications with a simplified zonal model of the kind presently used to assess resource adequacy of real systems in forming reserve margin and local installed capacity requirements.  Using the New York ISO system and published methodology, we demonstrate the feasibility of the high-fidelity analysis and highlight deficiencies in the existing zonal modeling approaches. *) ENELYTIX-PSO; https://enelytix.com

---

Session W3-B (Wednesday, July 9, 2:00 PM)

---

Effective Congestion Mitigation with Transmission Topology Optimization at Alliant Energy and ATC – Case Studies and Practical Lessons Learned

Mitchell Myhre, Senior Manager of Business Planning and Regulatory Strategy, Alliant Energy (Madison, WI)

Kristie Erickson, Reliability and Coordination Manager, ATC (Waukesha, WI)

Pablo A. Ruiz, CEO, NewGrid, Inc. (Somerville, MA)

Paola Caro, Principal Engineer, NewGrid, Inc. (Somerville, MA)

German Lorenzon, Senior Engineer, NewGrid, Inc. (Somerville, MA)

Timely and effective management of transmission system limitations has become critical to support the ongoing energy transition as well as economic development.  Additional transmission capability is needed quickly to support increased deployment of variable energy resources, which often are located away from demand centers, the electrification of traditionally fossil energy uses, and other load growth, such as data centers.  Transmission congestion can become acute in many regions, and has imposed over $20 billion of added costs on US transmission customers in 2022.  In also impacts grid reliability and resilience throughout the country.  To cost-effectively manage transmission constraints will require improved congestion management practices and technologies to better utilize the existing grid, especially as new transmission facilities are planned and constructed.  Transmission topology optimization identifies grid reconfiguration options to quickly re-route power flow around bottlenecks, increasing transfer capability across congested areas.  These reconfiguration solutions reduce the market redispatch and generation curtailments otherwise necessary to manage congestion and, by doing so, provide operational, efficiency, reliability, and resilience benefits.  In collaboration with NewGrid and ATC, Alliant Energy has been able to advance regionally beneficial and reliable reconfiguration opportunities to mitigate major bottlenecks in the area, relieving transmission limitations during both transmission and generation outages.  Working with the Midcontinent ISO, ATC and other Transmission Operators in the region evaluate and leverage these opportunities.  Alliant customers have reduced their congestion costs, avoided curtailments of wind and solar renewable plants, and facilitated the commissioning of new resources.  This presentation will review the Alliant Energy and ATC experience with the use of topology optimization, including case study results and practical lessons learned for the successful management and use of reconfiguration opportunities.

---

Power Purchase Agreements for Secured 24/7 Energy Delivery – A Comprehensive Review of Literature and Current Practices

Zhi Zhou, Principal Computational Scientist, Argone National Laboratory  (Lemont, IL)

Dongwei Zhao, Energy Systems Engineer, Argonne National Laboratory (Lemont, IL)

Daniel Boff, Economist, Pacific Northwest National Laboratory  (Richland, WA)

Alisha Fernandez, Systems Engineer, Pacific Northwest National Laboratory (Richland, WA)

As a critical complementary mechanism to pool markets, Power Purchase Agreements (PPAs) are becoming increasingly preferred among large energy consumers, such as data centers, to secure cost-effective energy. Traditionally, PPAs operate on a pay-as-produced basis, balancing supply and demand annually. However, the focus is shifting toward matching supply and demand on an hourly basis to fully meet energy needs. This shift requires the integration of flexible energy resources, such as hydropower and energy storage, to complement variable renewable energy sources like wind and solar, forming the foundation for 24/7 energy PPAs. This work provides a comprehensive review of emerging market trends, current practices, and state-of-the-art design methodologies for 24/7 PPAs, with a special focus on the role of hydropower. We further analyze potential resource mixes, including hydropower, wind, and solar from the supplier’s perspective to ensure continuous and reliable generation. From the buyer’s perspective, we examine diverse load profiles and their temporal variations, which influence procurement strategies and risk exposure. The 24/7 PPA design involves significant pricing and delivery risk due to market and generation uncertainties. We review the sources of these risks and the tools for risk management, highlighting how financial options and flexible resources such as hydropower can mitigate these risks. Ultimately, this study presents pathways for future PPA design, offering strategic insights into how integrated resource planning, financial structuring, and risk mitigation can lead to more secure and cost-effective 24/7 energy agreements.

 

---

Market-based Incentives for Adoption of Grid-Enhancing Technologies

Dr. Mostafa Ardakani, Associate Professor, University of Utah (Salt Lake City, UT)

The US power grid is under stress, hindering its economic efficiency and reliability.  Numerous studies show a need for substantial investment to increase the transmission network’s capacity.  Some of this need can be met through grid-enhancing technologies (GETs) that enable more efficient utilization of the existing grid.  Compared to building new transmission lines, GETs are significantly cheaper and can be deployed much faster.  Unfortunately, current compensation mechanisms do not provide sufficient incentives for adoption of GETs or their efficient operation.  This presentation explores two potential solutions through allocation of incremental financial transmission rights and marginal value pricing.  Both methods rely on market-based signals to compensate GET owners.  The methods are revenue adequate, align the incentives with social welfare improvements, and transfer the investment risks from ratepayers to the investors.  Preliminary results show that the developed market-based incentives generate payments that are much larger than those of a regulated rate of return, for efficient GET projects.

---

Session W4-A (Wednesday, July 9, 3:45 PM)

---

Market-based Storage Dispatch and Electricity Reliability

Siva Visvesvaran, Ph.D. Candidate, Cornell University (Ithaca, NY)

Dr. Jacob Mays, Assistant Professor, Cornell University (Ithaca, NY)

A primary aim of electricity market design is to align the private incentives of market participants with the regulatory goal of reliability at least cost.  While idealized markets achieve this alignment through the formation of efficient spot prices, concerns about market power and price volatility have led many market operators to suppress prices below the efficient level.  Because storage resources offer into the market based on private estimates of opportunity costs, a failure to produce full-strength prices leads to lower offers and premature dispatch leading up to potential scarcity events, threatening reliability.  This study develops a parsimonious model to describe the impact of energy market offer caps, capacity market non-performance penalties, and market participant risk attitudes on battery dispatch decisions.  Through a series of simulations incorporating bids from both risk-neutral and risk-averse battery operators, we demonstrate the reliability risk created by current capacity market configurations and explore potential solutions.  We argue that current capacity markets will be increasingly inadequate to the task of ensuring reliability as liberalized markets add storage, necessitating reforms to strengthen spot price formation.

---

Unlocking the Full Value of Energy Storage with Dynamic Optimization Software

Michael Baker, Co-founder & Chief Executive Officer, Tyba (Oakland, CA)

As energy storage assets proliferate across the U.S. power grid, their unique operational characteristics and flexibility present opportunities for grid stability and project profitability.  Many current operating models fall short in capturing the dynamism of these assets.  This results in suboptimal dispatch decisions, inefficient resource utilization, and ultimately lower returns for asset operators.  This session will present an advanced software and optimization methodology that significantly improves storage operations.  The approach begins with machine learning-driven price forecasting to configure and submit optimal day-ahead bids, accounting for expected market conditions and project specifications.  Once bids are cleared, the software continuously re-forecasts prices and re-optimizes the operating plan in real time, recalculating multi-leg price-quantity (PQ) bids before each interval to automatically ensure the battery is maximizing performance.  This real-time adaptability allows operators to capture price volatility and respond dynamically to changing grid conditions.  We will showcase real-world applications where this software has delivered significant performance improvements, including up to 40% higher revenue compared to median-performing assets in competitive markets such as ERCOT and California ISO.  Through collaboration with leading independent power producers and real-time integration with market systems, our approach has consistently demonstrated increased market efficiency and optimized asset utilization.

---

Locational Energy Storage Bid Bounds for Facilitating Social Welfare Convergence

Dr. Ning Qi, Research Scientist, Columbia University (New York, NY)

Dr. Bolun Xu, Assistant Professor, Columbia University (New York, NY)

Current market practices primarily rely on storage participants submitting strategic bids.  However, evidence from California ISO suggests that storage participants tend to overly withhold their availability, which will compromise market efficiency.  To address this issue, we have developed a novel regulatory software that generates bid bounds serving as offer caps for energy storage in electricity markets to help reduce system costs while mitigating potential market power exercises.  We derive the bid bounds based on a tractable multi-period economic dispatch chance-constrained formulation that systematically incorporates the uncertainty and risk preference of the system operator.  The bid bounds can seamlessly integrated into current market clearing software and enables operators to remain neutral, fostering competition among strategic storage participants, while capping offers to prevent excessive withholding that could compromise system efficiency.  The key analytical results verify that the bounds effectively cap storage bids across all uncertainty scenarios with a guaranteed confidence level.  We show that bid bonds decrease as the state of charge increases but rise with greater net load uncertainty and risk preference.  Agent-based numerical simulations based on the 8-zone ISO New England test system verify our theoretical findings and show the proposed approach can reliably reduce system cost and regulate storage profit, especially mitigating extreme withholding cases that also improve storage profits.  These benefits scale up with increased storage economic withholding and storage.

---

Battery Operations in Electricity Markets: Strategic Behavior and Distortions

Jerry Anunrojwong, Ph.D. Candidate, Columbia University (New York, NY)

Dr. Santiago R. Balseiro, Associate Professor, Columbia University (New York, NY)

Dr. Omar Besbes, Professor, Columbia University (New York, NY)

Dr. Bolun Xu, Assistant Professor, Columbia University (New York, NY)

Electric power systems are undergoing a major transformation as they integrate intermittent renewable energy sources, and batteries to smooth out variations in renewable energy production.  As privately-owned batteries grow from their role as marginal “price-takers” to significant players in the market, a natural question arises: How do batteries operate in electricity markets, and how does the strategic behavior of decentralized batteries distort decisions compared to centralized batteries?  We propose an analytically tractable model that captures salient features of the highly complex electricity market.  We derive in closed form the resulting battery behavior and generation cost in three operating regimes: (i) no battery, (ii) centralized battery, and (ii) decentralized profit-maximizing battery.  We establish that a decentralized battery distorts its discharge decisions in three ways.  First, there is quantity withholding, i.e., discharging less than centrally optimal.  Second, there is a shift in participation from day-ahead to real-time, i.e., postponing some of its discharge from day-ahead to real-time.  Third, there is reduction in real-time responsiveness, or discharging less in response to smoothing real-time demand than centrally optimal.  We quantify each of the three forms of distortions in terms of market fundamentals.  We also quantify the impact of the battery market power on total system cost via the Price of Anarchy metric, and prove that it is always between $9/8$ and $4/3$.  That is, incentive misalignment always exists, but it is bounded even in the worst case.  We calibrate our model to real data from Los Angeles and Houston.  Lastly, we show that competition is very effective at reducing distortions, but many market power mitigation mechanisms backfire, and lead to higher total cost.  The work provides stakeholders with a framework to understand and detect market power from batteries.  It also shows that the potential loss from battery market power is relatively small compared to the cost reduction achievable from having enough battery capacity in the system.  Therefore, independent system operators in rapidly changing markets might want to prioritize market entry of batteries and only shift to market power mitigation once the market is more mature.

---

Session W4-B (Wednesday, July 9, 3:45 PM)

---

Modeling Extreme Heat Wave and Wildfire Impacts on Power Reliability

Juliette Franzman, Data Scientist, Lawrence Livermore National Laboratory (Livermore, CA)

Hannah Burroughs, Staff Engineer, Lawrence Livermore National Laboratory (Livermore, CA)

Dr. Jhi-Young Joo, Grid Data Analytics Area Lead, Lawrence Livermore National Laboratory (Livermore, CA)

Dr. Andrew Mastin, Operations Research Scientist, Lawrence Livermore National Laboratory (Livermore, CA)

Christabella Annalicia, Academic Graduate Appointee, Lawrence Livermore National Laboratory (Livermore, CA)

Jean-Paul Watson, Senior Research Scientist, Lawrence Livermore National Laboratory (Livermore, CA)

Heat waves can have significant human and economic impacts including strain on the power grid.  In addition, if wildfires happen concurrently with heat waves, they can further strain the power grid due to requirements to de-energize power lines near active fires or electrical infrastructure damaged within wildfire perimeters.  As the probability and severity of concurrent heat waves and wildfires increases due to global climate change it is insufficient to use historical data to understand the potential impacts of these extreme events.  We develop a methodology to study the power reliability impacts of future concurrent extreme heat and wildfires.  First, we identify plausible future heat wave events using temperature data from a high-resolution global climate model.  The temperature data set is then used to predict increase in electrical load.  In parallel, we identify future high risk wildfire areas using global climate model projections.  To determine the power grid impact of plausible future wildfire events we use a bi-level optimization method.  The objective is to find the set of power grid components which, if damaged by wildfire, will lead to the largest loss of load, with an understanding of how the grid operator will adjust the system in the event of outages.  Finally, we use the hourly electrical load data and wildfire contingency set (electrical components disable during fire) to study the reliability impacts on the power grid (using a steady state AC power flow) from these concurrent events.

---

Weathering the Firestorm: Mitigating Electric Grid Ignited Wildfires

Dr. Brian J. Pierre, Manager, Electric Power Systems Research Department, Sandia National Laboratories (Albuquerque, NM)

The U.S. Department of Energy has a significant focus on Wildfire Electric Grid Resilience.  This presentation aims to discuss R&D efforts to mitigate the ignition and decrease consequences of major wildfires through new tools and improved information pre-wildfire, better planning for wildfires, early response during wildfires to increase safety and minimize damage, and accelerated recovery following wildfires to maximize energy availability.  Better wildfire modeling, monitoring, and planning tools allows significant reduction in consequences once a wildfire ignites, and optimal investments, protection schemes, control methods, and vegetation management can reduce the probabilities of grid ignited wildfires.  Responding early and effectively during wildfires is key to reducing wildfire consequences.  A better understanding of wildfire ignition and lightning ignition probabilities can improve wildfire response time.  Better grid component modeling and monitoring can reduce grid ignited wildfires and better uncertainty modeling and visualization can help decision makers understand when and if Public Safety Power Shutoffs (PSPS) are needed and when and if evacuations are needed.  Many utilities have created PSPS plans to mitigate wildfire ignition.  Significant R&D is focused on the reduction of the probability of grid-initiated wildfire ignition, and the reduction of consequence if wildfires occur, especially to critical infrastructure.  In addition, if we can help reduce wildfire ignition and impact, we can mitigate PSPS.  This presentation will highlight Sandia National Laboratories significant research portfolio in the Wildfire Electric Grid Resilience Program.

---

Multistage Generator Outage Simulation for Probabilistic Valuation of Operational Response to Extreme Events

Dr. Luke Lavin, Senior Research Engineer, National Renewable Energy Laboratory (Golden, CO)

Jose Daniel Lara, Senior Staff Researcher Power Grid Optimization, National Renewable Energy Laboratory (Golden, CO)

Matthew Bossart, Ph.D. Student, University of Colorado Boulder, (Boulder, CO)

David Palchak, Researcher, National Renewable Energy Laboratory (Golden, CO)

Resource adequacy modeling often assumes known, ex-ante distributions of generator outages, requiring translation to be used by power system operators.  We develop software more readily usable on operational decision timescales by integrating probabilistic draws of outages in multiple simulation stages through emulation and reserve deployment in real-time in NREL's open-source Sienna software suite.  This approach allows us to develop an insurance-like value-at-risk framework for valuing operational response to extreme events stressing power system reliability.  By leveraging empirical, NERC Generating Availability Data System-derived data on correlated generator outages, we also demonstrate methods with more direct applicability to system operators.

---

Prognostics-Driven Operations & Maintenance to Enhanced Grid Reliability

Dr. Feng Qiu, Group Leader, Argonne National Laboratory (Lemont, IL)

Dr. Shijia Zhao, Energy Systems Scientist, Argonne National Laboratory (Lemont, IL)

Dr. Murat Yildirim, Associate Professor, Wayne State University (Detroit, MI)

Dr. Zhaoyu Wang, Professor, Iowa State University (Ames, IA)

Dr. Joydeep Mitra, Professor, Michigan State University (East Lansing, MI)

Effective operations and maintenance (O&M) are crucial for enhancing the competitiveness and reliability of power system assets.  Traditional O&M strategies, such as periodic and diagnostics-based policies, often fail to address unit-specific degradation and forecast future failures, leading to inefficiencies and increased costs across various asset types, including photovoltaic (PV) systems, hydro generation, and other critical components.  This presentation introduces a prognostics-driven O&M approach designed to improve asset maintenance from reactive to proactive, with developed open-source tools on smart maintenance of solar inverters and prediction of hydro bearing degradations, as well as enhanced reliability modeling of solar inverters using semi-Markov approach.  The proposed technical framework leverages real-time sensor data and advanced prognostic models to accurately assess the state-of-health of diverse power system assets, predict their remaining life, and optimize maintenance schedules.  By integrating these predictions into stochastic programming, we develop large-scale O&M optimization models that enable fleet-wide, proactive decision-making.  This approach not only reduces maintenance costs and extends asset service life but also minimizes unplanned outages, thereby improving the overall reliability of the power system.  Key innovations and achievements include: (1) Continuous update of remaining life distributions (RLDs) using real-time sensor monitoring records with 89% accuracy in predicting inverter failures and domain-informed neural network to flag hydro bearing in degradation; (2) Dynamic maintenance model with fleet-level optimization to achieve 40+% saving in O&M cost using proposed prognostics-based maintenance; (3) Enhanced reliability modeling of solar inverters based on semi-Markov process, consisting of 5 transient and 1 absorbing states, with 95% accuracy to estimate practical survival function; (4) Open-source tools shared in GitHub (https://github.com/ANL-CEEESA/Fleet-Management-Tool-for-Solar-Inverters, https://github.com/ANL-CEEESA/Domain-informed-Models-for-Degradation-and-Prognostics-in-Hydro-Components ) co-developed and verified with industry partners.  The proposed framework empowers asset owners, power system operators, and maintenance service providers to enhance asset risk assessments and decision-making capabilities.  Additionally, utilities and independent power producers can leverage this framework to extend the lifetime of power system components and reduce operational interruptions.

---

 

Session H1-A (Thursday, July 10, 9:15 AM)

---

Probabilistic Grid Reliability Analysis with Energy Storage Systems: An Open-Source Tool for Assessing the Resource Adequacy of Power Systems

Dr. Atri Bera, Senior R&D Engineer, Sandia National Laboratories (Albuquerque, NM)

The Probabilistic Grid Reliability Analysis with Energy Storage Systems (ProGRESS) software is a Python-based open-source tool developed by Sandia National Laboratories (https://github.com/sandialabs/snl-progress) for assessing the resource adequacy of the evolving electric power grid integrated with energy storage systems (ESS).  This tool utilizes a Markov Chain Monte Carlo-based stochastic simulation engine to create diverse scenarios that test the limits of the modern power grid consisting of a high volume of ESSs and non-dispatchable resources.  State-of-the-art ESS models are incorporated within the Monte Carlo simulation engine.  The charge-discharge dynamics of ESSs, along with their evolving state-of-charge (SOC), are captured by the tool.  In addition, ESS failures and repair models are also built into the tool, allowing users to analyze the availability of their ESS devices when they are needed most.  ProGRESS also offers the capability of handling the uncertainty associated with non-dispatchable resources such as wind and solar, enabling the user to simulate different supply uncertainty scenarios based on weather conditions.  Users are able to build their own grid models, download and utilize historical weather data using APIs, analyze magnitude, duration, and frequency of expected future outages in terms of well-established reliability metrics including Loss of Load Probability (LOLP), Loss of Load Expectation (LOLE), and expected unserved energy (EUE).  ProGRESS allows users to make informed decisions and plan effectively considering uncertainties of the future electric grid.

---

GridCal – Open-source for Modern Power Systems

Josep Fanals i Batllori, Chief Executive Officer, eRoots Analytics SL (Barcelona, Spain)

Santiago Peñate-Vera, Chief Technology Officer, eRoots Analytics (Santa Brígida, Spain)

As power systems grow increasingly complex, driven by the integration of renewable energy, distributed generation, and data centers, the need for robust, accessible, and flexible simulation tools has never been greater.  GridCal, an open-source power systems simulation package mainly built by eRoots personnel, addresses this challenge by providing a comprehensive platform for modeling, analysis, and optimization of electrical grids.  This talk will introduce GridCal, highlighting its capabilities in load flow analysis, short-circuit calculations, contingency analysis, and optimal power flow, all within a user-friendly interface.  Designed for researchers, engineers, and educators, GridCal bridges the gap between academic research and industry practice, promoting collaboration and innovation in the energy sector.  Attendees will gain insights into how GridCal can be leveraged to tackle real-world power system challenges, from grid planning to operational decision-making, while contributing to the growing ecosystem of open-source tools for sustainable energy systems.  Join us to explore how GridCal is shaping the future of power systems analysis for a new generation of energy professionals.

---

Multiscale Stochastic Day- and Week-ahead Scheduling of a Large Renewable-dominated Power System – The Brazilian Case

Dr. Mario V. Pereira, Chief Innovation Officer, PSR (Rio de Janeiro, Brazil)

Dr. Joaquim Dias Garcia, Head, PSR (Rio de Janeiro, Brazil)

Thiago Cesar, Lead Specialist, PSR (Rio de Janeiro, Brazil)

Julio Alberto, Head, PSR (Rio de Janeiro, Brazil)

Dr. Luiz Carlos da Costa Jr., Technical Director, PSR (Rio de Janeiro, Brazil)

Guilherme Bodin, Head, PSR (Rio de Janeiro, Brazil)

Andre Dias, Head, PSR (Rio de Janeiro, Brazil)

Dr. Raphael Chabar, Executive Director, PSR (Rio de Janeiro, Brazil)

Brazil has one of the largest integrated power systems in the world, covering an area larger than the continental US or the European Union.  The energy mix is renewable dominated: 110 GW of hydro, 60 GW of solar, 35 GW of wind, 20 GW of biomass and 20 MW of thermal (mostly natural gas).  The HV grid has 15 thousand buses.  The system scheduling is carried out by the National System Operator, ONS.  The main challenge is to manage the very high variability of renewable production using the portfolio of hydro reservoirs, which range from multi-year to weekly storage capacity.  The optimal strategy is produced by a computational model based on the stochastic dual dynamic programming (SDDP) algorithm.  The SDDP model uses as input 1200 multivariate scenarios of inflows, wind, irradiation for each plant, plus temperature (which affects load).  The scenarios have weekly stages and hourly resolution in each stage and are produced by a stochastic model that captures the complex spatial and time dependencies of all inputs.  The SDDP optimal strategy produces a vector of “water values” (hydro opportunity costs) at the end of each week.  The STS – short-term scheduling (day- and week-ahead) – is formulated as the minimization of within-the-week costs plus end-of-week expected future costs.  The STS optimization problem represents all details of unit commitment, ramping constraints, transmission network constraints, multiple water use etc. and is solved by mixed integer programming and decomposition techniques.  One important limitation of the current STS model is the use of a single forecast of renewable production and load, that is, the STS is solved as a deterministic optimization problem, whereas the operation strategy is solved as a stochastic optimization problem.  This means that there are large mismatches between the day-ahead dispatch trajectory and the “true up” values in real time operation.  These mismatches are widening as new renewable capacity enters the system.  For this reason, ONS contracted PSR to design a new analytical tool for the STS & real time.  This presentation will describe the new tool, which uses an ensemble of forecasts produced by an AI-based model similar to GraphCast, and linear decision / affine rules to determine a stochastic short-term policy (SSP) instead of a single trajectory.  This SSP is sent to the real-time operation, which allows a tighter alignment with the true up results.  The new tool also includes the representation of grid-enhancing technologies such as Dynamic Line Rating, Smart Wires and new storage resources (batteries and pumped hydro) which will enter the system in the near future.

---

Session H1-B (Thursday, July 10, 9:15 AM)

---

A Comparison of Energy Market Pricing Methods

Dr. Richard O'Neill, Consultant (Silver Spring, MD)

We analyze the results of four pricing methods (LMP+, ELMP+, CHP and AIC pricing) theoretically, on various large available data sets, and on toy examples to compare their properties.  We will discuss the prices, make-whole payments and lost opportunity costs.  We measure the amount of uplift (lost opportunity cost (LOC) + make-whole payments).  Make-whole payments are not fully transparent and create inefficient peanut-buttering cost-allocation problems.  We show that LOCs need truthful information, are myopic, can cause infeasible solutions, and lower market efficiency.  We also show how allowing non-dispatched units to set prices creates difficulties.  Only AIC pricing fully eliminates make-whole payments.  For dynamic models, for example, day-ahead markets, we show how the pricing methods allocate avoidable fixed costs to the periods that caused them and whether the allocation is consistent with efficient pricing principles.  Since energy markets are, by design, highly granular (nodal, and hour or 5 minutes) compared to capacity markets that require more assumptions, administrative interventions and are more controversial.  Getting prices ‘right’ is much more difficult in less-granular capacity markets.  But increases in energy market prices results in lower capacity market prices or investment costs.  Finally, we compare each pricing method on market efficiency, transparency, make-whole payments, compatibility with ‘bilateral’ markets and the dispatch algorithm, investment signals, speed of computing, and compatibility with two-sided markets.

---

Evaluating the Benefits of Combining Short-term and Long-term Flexible Ramping Products in Real-Time Electricity Markets

Aravind Retna Kumar, Doctoral Candidate, Pennsylvania State University (University Park, PA)

Dr. Anthony Giacomoni, Manager, PJM Interconnection (Audubon, PA)

Shailesh Wasti, Ph.D. Student, Pennsylvania State University (University Park, PA)

Dr. Mort Webster, Professor, Pennsylvania State University (University Park, PA)

The need for greater system flexibility to manage increased net load volatility and forecast error has become a growing concern in energy market design.  Flexible Ramping Products (FRPs), system requirements for additional reserves to address net load variability distinct from contingency reserves, have been implemented in several markets as one possible avenue to increase system flexibility.  Typically, FRPs implemented in markets have relatively short delivery times, 10 or 15 minutes.  More recently, some markets have implemented, while others have proposed, a second FRP with a longer delivery time (one to two hours).  This is motivated by the assumption that a long-term product is needed to: a) address much longer, sustained steep ramps, which may become more frequent and accompanied by higher forecast errors with increased renewable generation, and b) induce commitments of additional resources that have longer startup times but lower marginal costs (e.g., combined-cycle units) than the traditional fast-start resources (e.g., combustion turbines).  Despite these developments, there has been relatively little rigorous analysis to evaluate whether the combination of short-term and long-term FRPs improves reliability or reduce cost relative to a market with only a short-term FRP.  In this work, we address the question: are there incremental benefits to implementing both a long-term (e.g., 60-minute) and a short-term (e.g., 10-minute) FRP within a Real-Time Market as compared with a market with only a short-term FRP?  Using a simple, illustrative power system, we systematically identify the range of possible system states for which the additional long-term FRP improves system reliability or reduces system cost.  Although on a smaller scale, our real-time market model captures several critical features of actual markets.  The model iterates between solving a multi-period unit commitment based on the current forecast, followed by a single-interval economic dispatch to meet actual realized net load.  This is repeated within a rolling horizon framework to emulate real-time market operations.  We compare simulated market outcomes across three market designs – no FRPs, a short-term FRP, and a combination of short-term and long-term FRPs – in terms of reliability and cost metrics.  This analysis is repeated over a search space by varying system attributes such as incremental costs between resource types, net load profiles, and real-time forecast errors in order to identify system states for which the addition of a long-term FRP provides benefits.  Across a wide range of system states, we find no cases in which a combined long-term and short-term FRP reduces the unserved energy relative to only a short-term FRP.  Additionally, the market design that combines short- and long-term FRPs results in operating costs that are equal to or higher than that of a system with a short-term FRP alone across most cases.  Only within a narrow range of state space does adding the long-term FRP reduce cost while achieving the same reliability.  Finally, we compare outcomes across these market designs for a detailed simulation of PJM’s energy markets and show that the same qualitative result holds: instances of improved reliability or reduced cost from adding a long-term FRP to a short-term FRP are rare across a range of historical load profiles and forecast errors.

---

European Merging Function: Merging Grid Models to Enhance Coordination of Interregional Flows and Reliability with Open-source Operational Computation Modules

Dr. Gladys Leо́n Surо́s, Vice President, Artelys (Paris, France)

Damien Jeandemange, Senior Software Engineer, Artelys (Paris, France)

Maja Markovic, Owner, Maja Markovic PR Digital Energy (Belgrade, Serbia)

Tengixang Ren, Energy Consultant, Artelys (Montreal, Canada)

Alexis Godefroy, Power Systems Software Engineer, Artelys (Lyon, France)

Nicolas Omont, Vice President, Artelys (Paris, France)

The diversity and complementarity of power grids between regions of large power grids can be leveraged to make the best use of the existing infrastructures.  Enhanced interregional cooperation brings additional efficiency and reliability to the systems.  For example, in the USA, NERC’s 2024 ITCS sheds light on 35 GW of potential transfer capacity that could benefit to system reliability.  To materialize those benefits, without jeopardizing the security, interregional power flows should be correctly modeled and computed.  In Europe, the European Merging Function (EMF) is being developed with this purpose in mind.  It performs the merging and consolidation of Individual Grid Models (IGMs) provided in CIM format by each Transmission System Operator, into a unified Common Grid Model (CGM) that enables operational coordination to maximize exchanges while ensuring reliability.  An open source implementation module that scales the net positions of different control areas according to market data has been developed within LFEnergy PowSyBl Open Load Flow.  It is used in several implementations of the EMF, enabling the sharing of the development efforts.  The implementation of a similar function in the USA may contribute to ease inter-ISO coordination (the so-called “seam issues”), especially in regions facing important loop flows.

---

Session H2-A (Thursday, July 10, 11:00 AM)

---

Impacts of Distributed Energy Resource Aggregations in Transmission Systems

Dr. Matthew Cornachione, Electrical Engineer, Pacific Northwest National Laboratory (Richland, WA)

Brent Eldridge, Electric Grid Systems Optimization Researcher, Pacific Northwest National Laboratory (Richland, WA)

Jesse Holzer, Senior Scientist, Pacific Northwest National Laboratory (Richland, WA)

Dr. Eran Schweitzer, Power Systems Engineer, Pacific Northwest National Laboratory (Richland, WA)

Dr. Liping Li, Operation Research Scientist, Pacific Northwest National Laboratory (Richland, WA)

Following FERC Order 2222, more Distributed Energy Resource Aggregations (DERAs) are anticipated to interconnect to distribution systems in order to provide energy, ancillary services, and capacity to the bulk grid.  Detailed analysis of potential impacts on transmission system reliability are necessary because of the unique operational characteristics of DERAs and the potential interactions between and simplified aggregate resource assumptions, distribution system conditions, and market incentives.  This presentation focuses on three main aspects of the DERA simulation platform developed at PNNL: modeling assumptions, simulation implementation challenges, and an assessment of market design policies.  Results show potential impacts from assumed DERA bidding and dispatch behaviors.  We overview the simulation platform architecture and its capabilities for modeling large-scale DERA integration in realistically sized networks.  Lastly, we analyze various existing and proposed DERA market design policies, such as impacts from self-scheduling requirements, distribution factor bidding format, and PTDF-derived aggregation zones.  Impacts are measured both in terms of transmission reliability metrics as well as economic and financial impacts on market participants.

---

Smart Grids as Coupled Physical and Economic Systems

Dr. Leigh Tesfatsion, Professor Emerita and Courtesy Research Professor, Iowa State University (Ames, IA)

The concept of a smart grid has been applied to grid-supported electric power markets at both high-voltage transmission levels and lower-voltage distribution levels.  This presentation will focus on smart-grid concerns for grid-supported electric power systems that span both levels, hereafter referred to as Integrated Transmission and Distribution (ITD) systems.  The presentation will start by reviewing nine broadly-accepted goals for ITD systems.  It will then highlight several fundamental design issues that arise for ITD systems due to their strong dependence on coupled physical and economic processes.  This strong dependence seriously complicates the ability of researchers and stakeholders to bridge the gap between design conceptualization and design implementation.  The presentation will conclude by discussing a tiered Design Readiness Level (DRL) approach to the design of ITD systems, analogous to the Department of Energy's tiered Technology Readiness Level (TRL) approach to the development of new technologies.  As will be shown using illustrative examples, this tiered DRL approach is greatly aided by recent progress in the development of co-simulation platforms.

---

Coordinated Transmission and Distribution Networks: A Bi-Level Framework for Resilience, Economic Efficiency, and Sustainability

Dr. Moses Amoasi Acquah, Postdoctoral Research Associate, University of Connecticut (Storrs, CT)

Dr. Zongjie Wang, Assistant Professor, University of Connecticut (Storrs, CT)

Extreme weather events and the rapid integration of distributed energy resources (DERs) transform power system dynamics, necessitating advanced coordination between transmission and distribution (T&D) networks.  Traditional unidirectional power flow models, where distribution systems are treated as passive loads, are no longer sufficient due to the increasing complexity of DERs, demand response programs, and advanced grid management technologies.  This work introduces a bi-level optimization framework uniquely designed to model the intricate coordination of T&D networks incorporating islanding and network reconfiguration strategies while directly employing resilience metric load recovery resiliency index (LRRI) into the optimization process to quantify and enhance grid resilience, economic efficiency, and sustainability.  In the proposed framework, the upper-level optimization aims to minimize annual investment and operational expenses by determining the optimal sizing and strategic placement of DERs within the power grid.  Simultaneously, the lower-level optimization maximizes load restoration capability during extreme events to enhance grid resilience.  The methodology integrates network topology optimization to achieve a holistic assessment of trade-offs among economic objectives, resilience enhancement, and emission reductions.  Advanced load restoration techniques, including dynamic network reconfiguration and contingency analysis, are employed to ensure grid performance under disruptive conditions.  Simulation case studies on IEEE T&D benchmark systems under various failure scenarios have demonstrated the effectiveness and efficiency of the proposed bi-level optimization framework.  Results have shown significant improvements in grid resilience, cost-effectiveness, and carbon emission reductions.  By addressing coordination gaps between transmission system operators (TSOs) and distribution system operators (DSOs), this work provides a scalable solution for integrating DERs, optimizing demand response, and enhancing grid adaptability to climate change.  The proposed work supports the transition to a decarbonized, resilient energy future, addressing critical challenges in modern power system planning and operation.  Keywords: Grid Resilience, Bi-Level Optimization, T&D Coordination, Distributed Energy Resources, Load Restoration, Network Reconfiguration, Carbon Emissions, Extreme Weather Events.

---

Session H2-B (Thursday, July 10, 11:00 AM)

---

Wholesale Electricity Markets with Multiple Virtual Power Plants

Dr. Andrew Liu, Associate Professor, Purdue University (West Lafayette, IN)

Jun He, Ph.D. Student, Purdue University (West Lafayette, IN)

FERC Order 2222 mandates the participation of DERs in wholesale electricity markets, creating opportunities for VPPs to aggregate numerous small-scale resources.  Traditional VPPs typically use centralized command-and-control approaches, which often face resistance from prosumers reluctant to relinquish management of their assets.  To address this challenge, our study proposes a decentralized VPP framework in which prosumers maintain autonomy over their resource decisions.  Specifically, we investigate whether this decentralized decision-making leads to stable market outcomes – including convergent LMPs, reduced price volatility, flattened load curves, and mitigation of the “duck curves" – or if it results in chaotic and unpredictable market dynamics.  Our decentralized VPP framework employs a Mean-Field Control (MFC) approach in which aggregators centrally determine optimal operational policies for a large number of prosumers in response to LMP signals from wholesale markets.  Unlike traditional centralized methods, our MFC approach computes a representative control policy for an aggregated population of DERs, significantly reducing computational complexity.  Prosumers then execute these optimized policies autonomously, preserving local control of their assets.  Critically, under our proposed approach, RTOs/ISOs continue to perform essential market-clearing processes, including UC/OPF calculations.  These operations ensure grid reliability and maintain transmission feasibility.  Additionally, ISOs/RTOs retain the capability to send targeted event signals – such as renewable oversupply conditions or critical peak load events similar to traditional demand response programs -- to aggregators.  Our MFC-based VPP algorithms can integrate these signals into their policy optimization processes, enabling VPPs to adapt quickly and effectively to such market conditions.  To capture interactions among multiple VPPs within a wholesale market, our framework integrates Mean-Field Games (MFG) with the MFC approach.  While MFC optimizes policies within each VPP given LMP signals, MFG models strategic interactions among multiple aggregators whose collective actions dynamically influence market outcomes and equilibrium LMPs.  This combined MFC-MFG framework allows us to explicitly analyze whether decentralized policy execution by numerous VPPs leads to stable convergence of LMPs and delivers desired market benefits, or instead generates chaotic volatility.  Reinforcement learning (RL) methods, particularly Proximal Policy Optimization (PPO), efficiently solve this hybrid model, enabling VPPs to iteratively learn optimal strategies under uncertainty (in a noncooperative manner), adaptively responding to fluctuations in renewable generation, demand patterns, and market prices.  We conducted numerical simulations using realistic test systems based on the Oahu Island electricity grid.  Tested scenarios include significant renewable oversupply and critical peak load conditions during heatwaves.  Results demonstrate that our decentralized VPP model effectively reduces market volatility, prevents extreme LMP events, and improves market stability and resource adequacy.  Comparative analyses show notable improvements in market efficiency, lower cumulative costs, and significantly reduced price volatility compared to decentralized price-response methods that ignore multiagent interactions.  This research directly supports FERC’s objectives of improving market efficiency through innovative software solutions.  By offering a robust, scalable, and customer-centric decentralized VPP solution, we address critical operational and economic challenges facing market operators and participants.  Ultimately, our framework provides a practical path toward achieving FERC Order 2222’s goal of efficient DER integration in wholesale electricity markets.

---

Synthesizing Grid Data with Cyber Resilience and Privacy Guarantees

Dr. Vladimir Dvorkin, Assistant Professor, University of Michigan (Ann Arbor, MI)

Shengyang Wu, M.S. Student, University of Michigan (Ann Arbor, MI)

Differential privacy (DP) provides a principled approach to synthesizing data (e.g., loads) from real-world power systems while limiting the exposure of sensitive information.  However, adversaries may exploit synthetic data to calibrate cyberattacks on the source grids.  To control these risks, we propose new DP algorithms for synthesizing data that provide the source grids with both cyber resilience and privacy guarantees.  The algorithms incorporate both normal operation and attack optimization models to balance the fidelity of synthesized data and cyber resilience.  The resulting post-processing optimization is reformulated as a robust optimization problem, which is compatible with the exponential mechanism of DP to moderate its computational burden.

---

PowerSAS.jl: A High-Performance Transmission Planning Reliability Analysis Tool in Julia

Dr. Wei Gao, Postdoctoral Associate, Argonne National Laboratory (Lemont, IL)

Dr. Feng Qiu, Group Manager, Advanced Grid Modeling, Argonne National Laboratory (Lemont, IL)

Shijia Zhao, Energy Systems Scientist, Argonne National Laboratory (Lemont, IL)

Reliability analysis is essential in transmission planning, ensuring that power systems operate continuously and securely under normal and stressed conditions.  Emerging grid complexities, driven by renewable energy integration and extreme weather events, demand computational tools capable of rapid, robust, and scalable scenario analyses.  Traditional numerical methods, such as iterative Newton-Raphson techniques, face significant challenges, including convergence issues and computational inefficiencies, particularly in extensive contingency scenarios.  This work enhances the existing PowerSAS.m semi-analytical solution (SAS) based tool –developed at Argonne National Laboratory – by refactoring it with more features for power grid planning in Julia, a high-performance computing language specifically chosen for its numerical efficiency and advanced parallel processing capabilities.  The Julia-based PowerSAS.jl achieves substantial improvements in computational speed and scalability, effectively addressing the exponentially growing simulation demands in large-scale power system analyses (e.g., N-1 contingencies, transient stability, and extended-term cascading scenarios).  In addition, PowerSAS.jl is to be released as an open-source tool, thus the power systems community will gain direct access to robust and highly efficient computational methods, fostering collaboration, transparency, and widespread adoption.  Leveraging PowerSAS’s proven numerical robustness and Julia’s performance advantages, the enhanced tool directly supports proactive and comprehensive reliability assessments in transmission planning, significantly reducing simulation runtimes and enabling more accurate and extensive evaluations critical for real-time and day-ahead market efficiency and operational resilience.

---