U.S. Variable Generation Integration Study Modeling ologies and Applications Kevin Porter ter Associates, Inc. China Renewable Energy Society Renewable Energy Grid Integration Committee Beijing, China March 13, 2015 General Description of Models Elements of a Variable Generation Integration Study Where Does the Data for a Variable Generation Integration Study in the U.S. Come From Mesoscale Modeling A Few Recent Trends in Variable Generation Integration Studies in the U.S. Outline General Description of Models Used in U.S. Related to Electric Power Capacity Expansion Planning Models Long term capacity expansion for generation and, sometimes, limited representation of transmission Simplified dispatch and operations algorithms, though still sometimes 8760 hours Used for broad picture high- penetration renewable energy studies, and utility resource planning not used for detailed, operational RE integration studies Examples ReEDs, Switch, Strategist, CREAM-EDO General Description of Models Used in U.S. Related to Electric Power Production Cost Simulation Models Used for optimization of dispatch and operations Used for detailed, operational RE integration detailed representation of plants, flexibility, etc. with hourly or even sub-hourly representation Not used for capacity expansion Examples GE MAPS, Plexos, GridView, Promod General Description of Models Used in U.S. Related to Electric Power, cont. Load Flow Models Ensuring reliability of transmission system to meet N-1 conditions Snapshot of time, like hour of peak demand or minimum net load Example GE PSLF Development of multiple models driven, at least in part, by desire to limit computational time and cost. With advances in computing power, expect to see these models used together, such as capacity expansion models and production cost models. See NREL report at http//www.nrel.gov/docs/fy14osti/ 61185.pdf for more ination. Elements of a Variable Generation Integration Study Mesoscale models for projected wind and solar generation and forecasts Statistical analysis Load flow modeling for transmission Production cost modeling Determination of capacity value Emissions, wear and tear, and cost impacts of increased cycling of fossil fuel plants Dynamic transient studies Frequency and stability analysis http//www.ieawind.org/index_page_postings/ 100313/RP201620Wind20Integration 20Studies_Approved20091213.pdf Where Does the Data for a Variable Generation Integration Study in the U.S. Come From Government sources EIA, FERC, EPA Non-disclosure agreements Transmission planning process open to the public. Scenarios and assumptions are debated and agreed to. Stakeholders can request transmission providers to per specific transmission studies during “study windows” Wind and solar resource databases are posted on the NREL web site Data availability and study transparency ensures robustness of VG integration studies and increases the chance of industry and public acceptance of study results Qualities of high quality wind and solar data sets High spatial resolution e.g., 2 km High time resolution e.g, 5 minutes Validation against data from met towers, actual wind plants or clusters of regional aggregation of wind plants Forecasts at multiple time levels 1, 4, 6 and 24 hour resolutions Forecasts trained with actual VG production data and calibrated with real forecast errors High resolution VG data sets will better represent geographic diversity of VG and a more precise representation of variability and uncertainty Costs can range from 10K for a 15 km dataset to 1 million for a 2 km dataset covering three years. 1 /15 1 /2 Mesoscale Modeling Memo and presentation on the availability, quality and cost of wind resource data, prepared by Justin Sharp of Sharply Focused LLC, is available from the Energy Foundation of China and explains these issues in more detail Evolution of Integration Studies Can We Do This Utility-by-utility focus Development of mesoscale model data Higher penetration of wind and, increasingly, solar Regional studies e.g., Western Wind and Solar Integration Study, Eastern Wind Integration and Transmission Study A Few Recent Trends in Variable Generation Integration Studies in the U.S. Higher penetration levels 40 wind in Minnesota study 40 Effects of cycling Phase 3 of the Western Wind and Solar Integration Study found that 33 renewables increases cycling costs between 35 million and 157 million in the West, as compared to 7 billion in fuel cost savings from higher levels of VG 33 3500 1 5700 70 Five scenarios, all 33 VG but varying levels of wind and solar 5 33 Production cost modeling the primary tool generation commitment and dispatch, emissions, costs and transmission path flows were the outputs A Few Recent Trends in Variable Generation Integration Studies in the U.S., cont. Frequency Response and Transient Stability Concern is displacement of conventional generation by VG will result in deterioration of frequency response and transient stability GE/NREL study basically found that frequency response and transient stability can be maintained with high levels of VG with transmission upgrades, and further improved with synchronous condenser conversions and frequency response controls on VG GE/NREL Measured effect on grid with different levels of VG and the loss of two nuclear plants 2,750 MW; the loss of three nuclear plants; the loss of all solar DG; and a three-phase fault at a 500 kV transmission line 2750 500 APPENDIX Base system plus projected transmission additions Perance of transmission grid in steady state peak load, N-1 conditions, and contingency conditions N-1 Measure whether transmission is overloaded at particular locations and at what times Add transmission lines, transmission upgrades, and transers additions or replacements Load Flow Analysis More recent integration studies simulate both Day-Ahead Security Constrained Unit Commitment and Real-Time Constrained Economic Dispatch “Interleaved” modeling, with one simulation feeding into another Captures forecast uncertainties between day-ahead and real-time Include current generating units, known and projected capacity additions and retirements, additional renewables needed for meeting scenario or government-required targets Load from actual year e.g., 2006, then scaled by projected load growth to future year e.g., 2020. 2006 2020 Model outputs by scenario annual energy production by type of generation, renewable energy curtailment, risk of reserve violations and unserved load hourly dispatch, emissions, fuel consumption, production costs. VG curtailment Available VG capacity exceeds transmission capacity, reflecting N-1 or transmission outage N-1 Aggregate VG is more than what can be accommodated, even if other generation at minimum and exports are maximized Hourly results will identify “periods of interest” for more intensive study on a sub-hourly basis Production Cost Models More cycling of fossil-fueled generation can contribute to thermal and pressure stresses, resulting in component failures, higher O 4-hour ahead for committing combined cycle and gas steam units, with 4-hour ahead wind and solar forecasts; and 5- minute dispatch 4 4 5 Western Wind and Solar Integration Study Phase II, cont. Concern is displacement of conventional generation by VG will result in deterioration of frequency response and transient stability GE/NREL study basically found that frequency response and transient stability can be maintained with high levels of VG with transmission upgrades, and further improved with synchronous condenser conversions and frequency response controls on VG GE/NREL Data and Assumptions Wind and solar capacity and production data by location Light load scenarios with base and high renewables, and summer peak demand scenarios with base and high renewables Power flow and transient stability simulation tool GE’s PSLF GE PSLF Transmission planning database from the Western Electricity Coordinating Council WECC WECC Western Wind and Solar Integration Study, Phase III GE/NREL Measured effect on grid with different levels of VG and the loss of two nuclear plants 2,750 MW; the loss of three nuclear plants; the loss of all solar DG; and a three-phase fault at a 500 kV transmission line 2750 500 Simulated grid operations with active power control and inertial control on wind plants and utility-scale solar plants, and with the addition of energy storage Western Wind and Solar Integration Study, Phase III cont.