Energy Yield Reconciliation in Monthly O&M Reports Claire Puttock, Lee Cameron & Alex Clerc April 15, 2016 EWEA Technology Workshop, Bilbao
Contents Introduction to variance explanation Motivation Key resources How is variance explained? Example results Variance explanation for young sites Summary
Introduction Automated analysis which gives a breakdown of difference ( variance ) between budget and actual production Variance broken down into three sources, assumed to be independent Variance(%) = Actual Production MWh Budget Production(MWh) Budget Production (MWh) Performance Ratio How well did the plant perform w.r.t. expectations? Energy Resource How favourable was the weather w.r.t. expectations? Budget Error How realistic is the budget?
Production Motivation Improved communication between owner and operator Enhanced understanding of asset as well as a portfolio 250 200 150 Identify opportunities to increase yield 100 Acknowledge possible budget uncertainty / inaccuracy 0 50 Budget Actual
Jan-06 Jun-06 Nov-06 Apr-07 Sep-07 Feb-08 Jul-08 Dec-08 May-09 Oct-09 Mar-10 Aug-10 Jan-11 Jun-11 Nov-11 Apr-12 Sep-12 Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Wind speed (m/s) Key Resources Reference Data Describes energy resource MERRA, RES internal WRF simulations 14 13 12 11 10 9 8 7 6 5 4 SMART Site Management, Analysis and Reporting Tool Suite of RES internal asset management tools Supports ad-hoc analysis e.g. in Python Month
Production (MWh) SMART Stores operational and budget data in a consistent format Automatically calculates performance losses 14000 12000 10000 8000 6000 4000 2000 Automatic data quality checks Actual Production Budget Production Reporting tool for asset managers Overview of portfolio statistics for asset owners
Variance(%) = Actual Production MWh Budget Production(MWh) Budget Production (MWh) Performance Ratio Budget Error Energy Resource
Variance Due to Performance Ratio Performance Ratio % = Net Production (MWh) Net Production (MWh) + Performance losses (MWh) Performance losses calculated using SMART SMART allows you to break down these losses into any form (e.g. Contractual downtime, grid etc..) Compare actual losses to budget losses to calculate variance due to PR
Variance(%) = Actual Production MWh Budget Production(MWh) Budget Production (MWh) Performance Ratio Budget Error Energy Resource
Jan-06 Jun-06 Nov-06 Apr-07 Sep-07 Feb-08 Jul-08 Dec-08 May-09 Oct-09 Mar-10 Aug-10 Jan-11 Jun-11 Nov-11 Apr-12 Sep-12 Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Wind speed (m/s) Yield Variance Due to Energy Resource Ideal yield* using all historic data compared to predicted ideal yield of analysis period Default assumption: linear relationship between monthly ideal yield and reference wind speed as per industry standard post construction analysis Can also use energetic reference data to reduce scatter 160 14 13 140 12 120 11 10 100 9 80 8 60 7 6 40 5 20 4 0 0 5 10 15 Month Reference Wind Speed *Yield corrected for losses, aka Gross Yield
Variance(%) = Actual Production MWh Budget Production(MWh) Budget Production (MWh) Performance Ratio Budget Error Energy Resource
Production (MWh) Variance Due to Budget Error Post construction analysis used to reforecast budget Possible results different annual budget, refined monthly profile Difference in owner s budget to reforecast budget contributes to production variance 8000 7500 7000 Unrealistic monthly budget profile 6500 6000 5500 5000 4500 4000 Jan-15 Feb-15 Mar-15 Apr-15 May-15 Jun-15 Jul-15 Aug-15 Sep-15 Oct-15 Nov-15 Dec-15 Budget Production (MWh) RES Best LT Estimate (MWh)
Remaining Variance Remaining variance = Variance (Variance due to PR + Variance due to Energy Resource + Variance due to Budget Error) Remaining (%) Variance = Actual Budget Budget Indicator of how successfully analysis has explained variance Performance Ratio (%) Budget Error (%) May be high due to missing or inaccurate SCADA data, inconsistent reference data, physical changes to site Energy Resource (%) Normally possible to reconcile variance within 1% over a year
Results Variance Explanation for 1 financial year Results collated and exported automatically via RES reporting tool Typically takes analyst about one day to complete analysis for 20 sites Site A Production Variance to Budget Variance Due to Performance Ratio Variance Due to Wind Variance due to Budget Error Remaining Variance -5% -6% 1% -1% 1% Production is 5% below budget PR loss budget is 9% Actual losses are 15%: - 2.5% Converter - 2.5% Curtailment - 2% Noise - 2% Yaw system - 2% Gearbox - 1% High ambient temperature - 3% Other (Generator/Grid/Scheduled Maintenance )
Production (MWh) Results Variance Explanation for 1 financial year Results collated and exported automatically via RES reporting tool Typically takes analyst about one day to complete analysis for 20 sites Site B Production Variance to Budget Variance Due to Performance Ratio Variance Due to Wind Variance due to Budget Error Remaining Variance -3% 7% 0% -10% 0% 4500 Production is 3% below budget 4000 3500 3000 Budget is 10% too optimistic 2500 2000 Budget Production (MWh) RES Best LT Estimate (MWh)
Results Variance Explanation for 1 financial year Results collated and exported automatically via RES reporting tool Typically takes analyst about one day to complete analysis for 20 sites Site C Production Variance to Budget Variance Due to Performance Ratio Variance Due to Wind Variance due to Budget Error Remaining Variance 0% -10% 7% 4% -1% PR loss budget is 4% Actual losses are 14%: - 7% Curtailment - 5% Blade repairs - 1% HV site repairs - 1% Hub/pitch system PR losses compensated by windy year and pessimistic budget
Pre Construction Variance Explanation I Post Construction variance explanation works well with >1 year of operational data Can we compare a site s operational yield to pre construction expectations? Proposal find relationship between monthly reference wind speed and pre construction modelled yield Historic measured wind speed Measure, correlate, predict Time series site wind climate Yield modelling tool Time series modelled yield Reference wind speed
Yield Pre Construction Variance Explanation II Plot time series modelled yield against reference wind speed data Compare real yield to modelled to determine variance due to wind Budget error is not possible to calculate as requires more data 250 200 y = 22,5x - 67,5 150 100 50 0 4 5 6 7 8 9 10 11 12 13 14 Reference Wind Speed (MERRA node X)
Yield Pre Construction Variance Explanation II Plot time series modelled yield against reference wind speed data Compare real yield to modelled to determine variance due to wind Budget error is not possible to calculate as requires more data Time Series Modelled Data Real Data 250 y = 22,5x - 67,5 200 150 100 50 0 4 5 6 7 8 9 10 11 12 13 14 Reference Wind Speed (MERRA node X)
Yield Pre Construction Variance Explanation II Plot time series modelled yield against reference wind speed data Compare real yield to modelled to determine variance due to wind Budget error is not possible to calculate as requires more data Time Series Modelled Data Real Data 250 y = 22,5x - 67,5 200 150 100 50 0 4 5 6 7 8 9 10 11 12 13 14 Reference Wind Speed (MERRA node X)
Yield Pre Construction Variance Explanation II Plot time series modelled yield against reference wind speed data Compare real yield to modelled to determine variance due to wind Budget error is not possible to calculate as requires more data Time Series Modelled Data Real Data 250 y = 22,5x - 67,5 200 150 100 50 0 4 5 6 7 8 9 10 11 12 13 14 Reference Wind Speed (MERRA node X)
Yield Pre Construction Variance Explanation II Plot time series modelled yield against reference wind speed data Compare real yield to modelled to determine variance due to wind Budget error is not possible to calculate as requires more data Time Series Modelled Data Real Data 300 250 200 150 100 y = 22,5x - 67,5 These numbers could be delivered in a pre-construction report! y = 23,4x - 96,5 50 Derived from production data 0 4 6 8 10 12 14 16 Reference Wind Speed (MERRA node X)
Production Summary Variance explanation used to give breakdown of reasons why actual production differs from budget 250 200 Variance broken down into three categories assumed to be independent: 150 100 - Performance Ratio losses 50 - Energy Resource variation - Budget error 0 Budget Actual For young sites, real yield could be compared to modelled yield derived from pre construction estimates to calculate variance due to wind Site Variance Variance Due to Performance Ratio Variance Due to Wind Variance Remaining due to Variance Budget Error A -5% -5% 1% -2% 1% B -3% 7% 0% -10% 0% C 0% -10% 7% 4% -1%
Any Questions?