NorthWestern Energy Group, Inc. · NASDAQ:NWE
NorthWestern Energy — Ascend Analytics — Technical Presentation (June 2017)
71 slides from NorthWestern Energy's technical presentation, Jun 15, 2017
Slide 1: PowerSimm for Applications of Resource Valuation
PowerSimm for Applications of Resource Valuation Presented by: Sean Burrows, PhD Alankar Sharma/Kristina Wagner [email protected] 303.415.1400 June 15, 2017
Slide 2: Agenda
Agenda Defining Model Objectives for Resource Evaluation Principal Model Drivers for Asset Valuation Validation Results Evaluation of Flexible Generation Technologies PowerSimm Configurations & Operations 2 ascend analytics
Slide 3: Key Questions Addressed
Key Questions Addressed • Resource Requirement—Addressed in April ETAC 1) How does PowerSimm model the need for flexible generation? 2) How does PowerSimm model the need for capacity? • Resource Valuation—Today’s ETAC 1) How does PowerSimm validate inputs of market conditions for modeling? a) Foreca
Slide 4: Agenda
Agenda Defining Model Objectives for Resource Evaluation Principal Model Drivers for Asset Valuation Validation Results Evaluation of Flexible Generation Technologies PowerSimm Configurations & Operations 4 ascend analytics
Slide 5: PowerSimm Compared to Other Solutions
PowerSimm Compared to Other Solutions Capability Competitors Ascend Detailed sub hourly dispatch optimization Flexible requirements as function of renewables X Integrated ability to calculate peak resource (MW) X Simulated weather, load, renewables, price X Substantiate “meaningful uncertainty” X Ma
Slide 6: Keys to Future Supply Portfolios
Keys to Future Supply Portfolios • Need to manage cost uncertainty and not position uncertainty Recognize new exposure is to price volatility not cost of energy Declining prices limit exposure of average open positions Increased volatility in prices increase exposure to future cost uncertainty • Nee
Slide 7: Resource Valuation: Objective Functions
Resource Valuation: Objective Functions Min Expected Value of Total Cost E[NPV of Total Costs] = E ∗ ∗ 1 • Fixed Costs follow revenue requirements Depreciation, Amortization, Current Taxes, Deferred Taxes, Insurance, Property Taxes, On-going capital improvements, return on equity and debt • Variable
Slide 8: Calculation of Loss of Load Probability
Calculation of Loss of Load Probability Weather drives load and Simulated renewables Simulated generation renewable generation generation & load outages Hydro Gen + = Capability to serve peak load Simulated Weather Renewables Generation outages PowerSimm captures the Note: Competing models to PowerS
Slide 9: Resource Adequacy
Resource Adequacy PowerSimm’s simulation framework captures the structural relation of weather with renewables and load over a wide range of future conditions MW Short @ 0% Reserve Margin 120 100 80 MW Short 60 40 20 0 60 Hours Short by Year Hours (Load > Capacity) 50 MEAN 40 P5 30 P95 20 10 0 9 asc
Slide 10: Minutely Dispatch and Regulation
Minutely Dispatch and Regulation 2020 Baseline + 200 MW Wind + 100 MW Solar • Solar 1 MW ~ 0.2 MW flexible resources • Wind 1 MW ~ 0.6 MW of flexible resources • Changes in gradient of net load • Base load 3 ∆’s in gradient • Net load 7 ∆’s in gradient 10 ascend analytics
Slide 11: Total Flexibility Requirements in NWE – Minimizing Fast Ramp
Total Flexibility Requirements in NWE – Minimizing Fast Ramp Flexible requirement = 15 minute ramp (Inc) + 1 minute ramp (Regulation) Illustration of Flexible Capacity Requirement Regulation Flexible Year Scenario Inc Capacity Capacity Requirements 2016 Base 168 28 196 2017 Base 169 28 197 2018 Base
Slide 12: Agenda
Agenda Defining Model Objectives for Resource Evaluation Principal Model Drivers for Asset Valuation Validation Results Evaluation of Flexible Generation Technologies PowerSimm Configurations & Operations 12 ascend analytics
Slide 13: Implications of Weather During Delivery Conditions
Implications of Weather During Delivery Conditions 13 ascend analytics
Slide 14: Simulation Engine: Integrating Physical and Financial Uncertainty
Simulation Engine: Integrating Physical and Financial Uncertainty Ascend’s PowerSimm is able to capture meaningful uncertainty, through: The flow chart • Producing realistic simulations of physical system and market conditions outlines • Capturing the core structural relationship between weather, lo
Slide 15: PowerSimm Mirrors NWE's Weather-Load Relationship
PowerSimm Mirrors NWE’s Weather-Load Relationship Weather is a key driver during delivery conditions Space cooling rapidly drives up energy usage Space heating drives a slower load response 15 ascend analytics
Slide 16: PowerSimm Simulates Realistic Price-Load Relationships
PowerSimm Simulates Realistic Price-Load Relationships • Red plus signs are historical data • Blue circles are simulated data • A high degree of overlap shows the model is effectively capturing the underlying dynamics 16 ascend analytics
Slide 17: Renewables vs Temperature
Renewables vs Temperature Solar Generation (MWh) Wind generation has a structural relationship with temperature differences but is weakly correlated with average temperatures Wind Generation (MWh) Average Temp F Solar generation exhibits an upward trend as temperature increases Average Temp F 17 asc
Slide 18: Seasonal Energy Production of NorthWestern Hydros
Seasonal Energy Production of NorthWestern Hydros Generation peaks with spring snow melt in April & May Generation MWh Hydro assets are largely run-of-river, but have pondage to provide capacity and some flexible generation 18 ascend analytics
Slide 19: Fundamental Determinants of Volatility in Power Prices
Fundamental Determinants of Volatility in Power Prices 19 ascend analytics
Slide 20: Market Potential for Capacity: Energy Flows throughout the WECC
Market Potential for Capacity Energy Flows throughout the WECC • PowerSimm performs fundamental modeling of the WECC region inclusive of transmission congestion and losses determined through an integrated generation dispatch and DC optimal power flow model. • Major interties support seasonal flow of
Slide 21: Price Volatility in CAISO over time
Price Volatility in CAISO over time 22% 7% • Peaks and Valleys in Renewables (as % of load) correspond to peaks & valleys in pricing volatility. 60% 42% • Increase in wind/solarincrease volatility in LMPs. 100% 50% 21 ascend analytics
Slide 22: Over-generation due to increasing renewables is already here
Over-generation due to increasing renewables is already here EIM Real Time Price Map. March 27, 2017 at 11:00 AM • kkkk Energy Imbalance Market in the Northwest 22 ascend analytics
Slide 23: California needs to export excess solar to the WECC immediately
California needs to export excess solar to the WECC immediately • CA is ahead of expectations in terms of renewable export due to 5 GW of rooftop solar plus utility scale solar. • March 2017 – 80 GWh of renewables were curtailed due to over-supply. • The Western Energy Imbalance Market is expected t
Slide 24: Changing Market Dynamics as a Function of Renewables
Changing Market Dynamics as a Function of Renewables Hour 24 ascend analytics
Slide 25: Observed and Forecasted Increases in Price Volatility through Time WECC
Observed and Forecasted Increases in Price Volatility through Time WECC 10% 7% Mid-C Volatility of DA LMP 25% 11% 25 ascend analytics
Slide 26: WECC Market Analysis: Renewables and Market Volatility
WECC Market Analysis: Renewables and Market Volatility Recorded Forecast Recorded Forecast • Market price volatility is expected to increase as renewable generation increases • Greater premium on highly flexible capacity • By 2027 the growth in price volatility diminishes • Volatility peaks during t
Slide 27: Forward Prices and Volatilities
Forward Prices and Volatilities 27 ascend analytics
Slide 28: Forward Fuel Prices
Forward Fuel Prices Prices peak in winter months (January) Source: Canadian Gas Price Reporter 28 ascend analytics
Slide 29: Long-term Implied Heat Rates (Annual)
Long-term Implied Heat Rates (Annual) Implied Heat Rates - Annual 14 Market quotes for forwards till 2020; 12 liquid forward curves post-2020 10 Implied Heat Rate (MMBTU/MWh) 8 6 4 2 0 Heavy Load HRs Light Load HRs • With the increasing penetration of zero variable cost renewable generation in WECC,
Slide 30: Long-term Implied Heat Rates (Monthly)
Long-term Implied Heat Rates (Monthly) Implied HRs - Monthly Late summer (August-September) peaks in implied heat rates slowly decline 18 over time due to increasing solar and wind generation. By 2025, with market maturity, the peaks in late summer and late winter (February) normalize to 16 the same
Slide 31: Forward Price Volatility Inputs
Forward Price Volatility Inputs Forward Price Volatility Inputs • Forward price 50% volatilities are used to accurately model 45% correlations between 40% monthly gas and power prices 35% Forward Volatilities 30% • Monthly forward volatilities are higher 25% closer to delivery 20% date of the forwar
Slide 32: Spot Price Volatility Inputs
Spot Price Volatility Inputs • Spot price volatilities account for increasing variability in intra-month power prices • Spot price volatilities for power are calculated as = (Standard Deviation of prices / Mean of prices) on monthly time steps • Volatility forecasts are developed based on relationsh
Slide 33: Demonstrate PowerSimm UI – Forward Curve Editor, Spot Price Editor
Demonstrate PowerSimm UI – Forward Curve Editor, Forward Curve Constraints Editor (Corr, Vol), Spot Price Editor 33 ascend analytics
Slide 34: Raw Spot Price Simulations
Raw Spot Price Simulations • Raw spot price data shows simulations consistent with market dynamics • Higher spot volatilities signified by increasingly larger price fluctuations in the future • Frequency of prices being >$100 increase over time consistent with observations in WECC region • Negative
Slide 35: Long-Run Equilibrium Evolution to High Renewable Penetration Rates
Long-Run Equilibrium Evolution to High Renewable Penetration Rates 35 ascend analytics
Slide 36: New Thermal Units – Capacity Factors
New Thermal Units – Capacity Factors • With decreasing power price forward marks, the monthly capacity factors for each generation type decreases. • CCs suffer marginally more than ICEs with more volatile power prices and incurred startup costs Thermal Assets - Capacity Factors 100% 90% 80% 70% 60%
Slide 37: Long-run Equilibrium: Thermal Generation
Long-run Equilibrium: Thermal Generation • Planning traditionally assumes long-run equilibrium conditions are satisfied through wholesale price conditions supporting new unit entry of a CC or CT. • Due to the lowering of implied heat rates and increasing volatility caused by renewables, Montana’s ma
Slide 38: ColStrip – Capacity Factors
ColStrip – Capacity Factors • ColStrip without must run component shows declining capacity factors with longterm price conditions • If modeled as must run, the must run component incurs annual losses owing to power prices being unfavorable (negative power prices) to keep the unit on ColStrip Capacit
Slide 39: ColStrip – Net Revenue Returns ($/kW year)
ColStrip – Net Revenue Returns ($/kW year) • With cheap coal prices, ColStrip plants will continue to generate revenues with increasing volatile prices • Revenues gained from higher positive prices will be greater than losses incurred from operating during negative price hours due to inflexibility •
Slide 40: Agenda
Agenda Defining Model Objectives for Resource Evaluation Principal Model Drivers for Asset Valuation Validation Results Evaluation of Flexible Generation Technologies PowerSimm Configurations & Operations 40 ascend analytics
Slide 41: Forward Price Simulations
Forward Price Simulations • PowerSimm simulates monthly forward curves for power for both on-peak (heavy) and off-peak (light) periods • PowerSimm simulations closely represent volatility conditions with higher price spreads closer to today’s transaction date and lower spreads further out For peak p
Slide 42: Forward Price Simulations
Forward Price Simulations Figures represent 5 simulated forward curve price paths for Heavy and Light load hours 42 ascend analytics
Slide 43: Load Validation
Load Validation Load Simulation Validation Results • Demonstrate a robust model to simulate future load • Capture uncertainty in load growth with extreme weather scenarios validated prior to a load simulation • Validated monthly load values • Validated hourly load values for hours 1-24 Load Simulati
Slide 44: Temperature Validation
Temperature Validation PowerSimm simulates temperature at multiple sites BILLINGS MISSOULA PowerSimm simulates temperature at multiple sites 44 ascend analytics
Slide 45: Weather Simulations: WECC Regional Weather
Weather Simulations: WECC Regional Weather • In addition to simulating Northwestern’s weather driven customer load, PowerSimm also simulates weather and load conditions for WECC region • The structural relationship between WECC weather patterns and historical prices drive simulations of future price
Slide 46: Demonstrate PowerSimm UI – Main Market Model (WECC System Load)
Demonstrate PowerSimm UI – Main Market Model (WECC System Load) 46 ascend analytics
Slide 47: Agenda
Agenda Defining Model Objectives for Resource Evaluation Principal Model Drivers for Asset Valuation Validation Results Evaluation of Flexible Generation Technologies PowerSimm Configurations & Operations 47 ascend analytics
Slide 48: NWE System Optimization
NWE System Optimization Objective Function – Minimizing the cost to meet load min #0 − ( + 3$ ∗ .# + >- + ./ ?5 456578 97: System Level Constraints • Meet load: • #0 − ( + ∑456578 97: 3$ = Availability for energy and regulation • Maximum and minimum generation levels • Maintain sufficient flexibilit
Slide 49: Summary of Technologies and Valuation Approach
Summary of Technologies and Valuation Approach Technology Valuation Summary Standalone battery storage Batteries generate system capacity value when they free up thermal resources from the task of regulation. Combined cycle units can therefore move from the mid-point to operating at full capacity wi
Slide 50: Economics of Batteries
Economics of Batteries 50 ascend analytics
Slide 51: Three Uses of Batteries
Three Uses of Batteries Three Uses of Batteries Regulation Balancing out minute-to-minute fluctuations in load and generation Ramping Respond to sub-hourly and hourly ramps (e.g. evening ramp as solar generation tails off) Load shifting Store excess generation that would otherwise be dump energy and
Slide 52: Battery Benefits: Avoided Costs
Battery Benefits: Avoided Costs Battery benefits consist of two key components: 1.) The avoided fuel cost from batteries providing regulation services • Thermal units have a declining heat rate (i.e. operate more efficiently) as they run closer to full load capacity. When thermal units serve regulat
Slide 53: Battery Benefits: Capacity Value
Battery Benefits: Capacity Value 2.) Batteries create capacity by relieving thermal plants of regulation responsibilities and allowing them to instead commit energy during peak load conditions Capacity Value vs Battery Size Capacity value of the 10, 25, 50 and 100 MW ESS over time. 160 140 120 100 M
Slide 54: Production (Fuel) Cost Savings from Batteries
Production (Fuel) Cost Savings from Batteries • The value of the batteries is measured as the difference in net revenue with and without batteries. • Batteries primarily derive value from furnishing regulation, enabling thermal generation to operate more efficiently. Production Cost Savings vs. Batt
Slide 55: Economics of Energy-serving Batteries
Economics of Energy-serving Batteries Annual Battery Revenues $10,000 $9,000 $8,000 $7,000 $ per MWh $6,000 $5,000 $4,000 $3,000 $2,000 $1,000 $0 2017 2018 2019 2020 2021 Small Battery (300 MWh) Medium Battery (1000 MWh) Large Battery (5000 MWh) • In an economic dispatch scenario, batteries serving
Slide 56: Battery Charge and Discharge Relative to Power Price
Battery Charge and Discharge Relative to Power Price Medium Battery (1000 MWh) - Market Price vs Battery Net Position for Oct 22 and Oct 23 50 As price $40 spikes, battery $35 Battery Charge/Discharge (MWh) 30 discharges MID-C market Price ($) $30 10 $25 $20 -10 $15 As price Battery doesn’t cycle dr
Slide 57: Medium Battery State of Charge and Resource Interaction
Medium Battery (1000 MWh) State of Charge for Oct 22 and Oct 23 600 $40 $35 500 MID-C Market Price $30 400 $25 MWh 300 $20 Battery charges between $15 200 2:00am and 6:00am $10 100 $5 0 $0 0:00 6:00 12:00 18:00 0:00 6:00 12:00 18:00 Hour Battery State of Charge (MWh) MID-C Price Medium Battery (1000
Slide 58: Demonstrate Hourly Reports for Batteries
Demonstrate Hourly Reports for Batteries (scrollbar charts) 58 ascend analytics
Slide 59: Example Valuation Strategies by Technology
Example Valuation Strategies by Technology 59 ascend analytics
Slide 60: Stand-Alone Battery
Stand-Alone Battery Provides capacity and more efficient energy to the system by substituting for conventional power plants in providing for ancillary services The conventional power plants no longer have to back down from their most efficient operating point to hold capacity in reserve for flexibil
Slide 61: Pumped Hydro Storage
Pumped Hydro Storage Co-optimized operation between energy and ancillary services and provide capacity. Economics driven by the spread between low and high prices throughout the day. Perform some Hydro Net Position Discharge in morning regulation all day and evening ramps 100 80 60 40 20 MW 0 -20 -4
Slide 62: Internal Combustion Engine (ICE)
Internal Combustion Engine (ICE) Co-optimized operation between energy and ancillary services Net Position 100 MW plant does 100 50 MW of regulation 90 and 50 MW of energy 80 70 60 MW 50 40 30 20 10 0 Minutely Time-Scale Constraints for optimization: Ramp rate limit: 3$ − 3$ A ≤ !&! Power within cap
Slide 63: Agenda
Agenda Defining Model Objectives for Resource Evaluation Principal Model Drivers for Asset Valuation Validation Results Evaluation of Flexible Generation Technologies PowerSimm Configurations & Operations 63 ascend analytics
Slide 64: Ancillary Services Modeled
Ancillary Services Modeled Ancillary services modeled are specific to Northwestern Energy region; not tied to NERC standards Contingency Reserves: Resources that address outages of major units Spinning Online resources that can quickly address major outages; must respond within requirements 10 minut
Slide 65: Demonstrate PowerSimm UI - Portfolio, Generation, Ancillary Reserve Editor
Demonstrate PowerSimm UI - Portfolio, Generation (allowed ancillary contributions tab), Ancillary Reserve Editor, etc. 65 ascend analytics
Slide 66: Unit Behavior – Base Portfolio - Basin Creek and ColStrip
Unit Behavior – Base Portfolio - Basin Creek and ColStrip • Basin Creek CT unit primarily • ColStrip’s upper section (dispatchable, non-must-run) acts serves Non Spin and INC as a flexible generation resource serving regulation requirements services and spin, non spin as required • Must run portion
Slide 67: Unit Behavior – Base Portfolio - David Gates
Unit Behavior – Base Portfolio - David Gates • David Gates - fast ramping resource but expensive to run with high operating cost • With 6 total units modeled under David Gates plant, it can provide all types of ancillary services primarily providing regulation owing to high flexible capabilities 67
Slide 68: Unit Behavior – Generic ICE Portfolio – ICE vs David Gates
Unit Behavior – Generic ICE Portfolio – ICE vs David Gates • New ICE units are cheaper to operate compared to existing flexible David Gates unit • A new ICE introduced in NWE’s portfolio takes over providing regulation and freeing David gates to be available for Non Spin and INC incidences because i
Slide 69: Unit Behavior – Generic Battery Portfolio – Regulation Serving Battery
Unit Behavior – Generic Battery Portfolio – Regulation Serving Battery • Regulation serving batteries primarily serve flexible ramp requirements i.e. regulation • Batteries are cheaper to operate compared to thermal assets and are “always ON” causing batteries to primarily serve regulation requireme
Slide 70: Unit Behavior – Hydros
Unit Behavior – Hydros • Hydro units are modeled as energy only components and ancillary only components • Energy only hydro components are modeled on weather driven generation levels • Ancillary serving Hydro components are modeled as battery storage items providing corresponding ancillary service
Slide 71: Demonstrate Hourly Cube Reports for Ancillary Contributions
Demonstrate Hourly Cube Reports for Ancillary Contributions 71 ascend analytics
Ascend Analytics — Technical Presentation (June 2017) on 6ix