Home

STM User Manual

1. 000 33300 20 3140 2 2009 06 26 00 00 00 000 2 00 000 38500 000 000 000 0 000 33300 21 3141 2 2009 06 26 00 00 00 000 2 00 000 38500 0 00 000 000 0 O00 333 00 22 3142 2 2009 06 26 00 00 00 000 2 Filename 260609 z seve Joo oos 500 000 ooo ooo o ooo 33200 23 3143 2 2009 06 26 00 00 00 000 2 Save as type Report fies r xl Cancel 00 000 38500 0 00 000 000 0 000 33300 24 514 2 2009 05 26 00 00 00 000 24 Ps on oo Do 400 000 38500 000 000 000 0 O00 33300 i25 345 2 2009 05 26 00 00 00 000 25 00 25800 0 00 0 00 9000 90 00 91 00 237 00 23600 0 00 000 170 00 170 00 0 00 40400 0 00 38500 000 0 00 000 0 000 33300 26 3146 2 2009 06 26 00 00 00 000 26 00 25800 0 00 0 00 9000 90 00 91 00 23700 23600 000 000 170 00 17000 0 00 40400 000 38500 000 0 00 000 0 000 33300 2 347 2 2009 05 26 00 00 00 000 27 00 25800 0 00 0 00 90 00 90 00 91 00 23700 23600 000 000 170 00 170 00 0 00 40400 000 38500 000 000 000 0 000 33300 28 3148 2 2009 05 26 00 00 00 000 28 00 25800 000 0 00 9000 90 00 91 00 23700 23600 000 000 170 00 000 0 00 40400 000 38500 000 0 00 000 0 000 33300 29 3149 2 2009 05 26 00 00 00 000 29 00 25800 000 0 00 9000 000 91 00 23700 23600 000 000 170 00 000 0 00 40400 000 38500 000 0 00 000 000 33300 30 3150 2 2009 05 26 00 00 00 000 30 00 25800 0 00 000 000 000 91 00 23700 23600 000 0 00 170 00 0 00 0 00 40400 000 38500 000 0 00 000 0 000 33300 31 3151 2 2009 06 26 00 00 00 000 31 00 25800 0
2. This assumption is based on the observation that SEMO has historically dispatched Peat plant at or close to their availabilities gt Validation is applied so that no Minimum On Time or Minimum Off Times parameters are breached within historical availability data 11 5 Pumped Storage and Hydro gt The hydro Data Query returns the aggregate half hourly MSQ data for all hydro plants over the last 7 28 days gt The relevant average half hourly MSQ for periods 1 to 48 for both weekdays and weekends respectively 1s calculated These half hourly values are the forecast hydro MSQs over the 7 day period selecting weekday or weekend where appropriate gt The pumped storage Data Query returns the historical values for the last 28 days gt The pumped storage starting condition parameter is reset by reference to the previous days pumped storage ending condition gt The relevant average pumped storage parameters for both weekdays and weekends respectively are calculated These values are the forecast 35 pumped storage parameters over the 7 day period selecting weekday or weekend where appropriate 11 6 IC Data gt gt The Interconnector Data Query returns all relevant half hourly import and export bids for the latest available bid day for all Interconnector parties Bids with no import or export capacity or no incremental quantity are eliminated A single consolidated PQ input file is constructe
3. 00 000 000 000 91 00 23700 23600 000 000 170 00 0 00 0 00 40400 000 38500 000 0 00 000 Q 000 33300 32 3152 2 2009 06 26 00 00 00 000 32 00 25800 0 00 0 00 0 00 000 91 00 23700 23600 000 0 00 170 00 0 00 0 00 40400 000 38500 000 0 00 000 0 000 33300 33 3153 2 2009 06 26 00 00 00 000 3300 25800 0 00 000 000 000 91 00 23700 23600 000 0 00 170 00 0 00 0 00 40400 000 38500 000 0 00 000 0 000 33300 4 354 2 2009 06 26 00 00 00 000 34 00 25800 0 00 0 00 0 00 O00 91 00 23700 23600 000 000 170 00 0 00 0 00 40400 000 38500 000 000 000 0 000 33300 Query executed successfully 127 0 0 1 SQLEXPRESS 1965 30 ATM ELPlVAdministator 72 Lingo 00 00 09 0 rows gt Alternatively the on screen grid view can be selected by clicking the top left hand corner of the grid frame This will highlight all data retrieved by the query and this can be copied using a right mouse click The data can then be pasted into a number of applications e g Microsoft Excel 9 2 SQL Queries gt The two main queries which are used to extract SMP Uplift and MSQ results from the SQL database directly are respectively Use Lingo SELECT FROM Uplift WHERE Scenario ID 1 ORDER BY RunDate ASC Time ASC and Use Lingo SELECT FROM MSQ WHERE Scenario ID 1 ORDER BY RunDate ASC Time ASC gt These queries can be pasted into the white upper pane in SQL and then executed by pressing the Execute button red exclamation mark on the sub
4. Files Energy Link STI STE exe Path to Uplift Files Energy Link STH Uplitt exe 2 g Auto Correct MOFFT MONT amp MSG Errors QE Cancel 11 gt If AutoCorrect was not selected then any errors will be left and will just be reported in the Verification Errors Tab If it is selected the corrections will be made and the Scenario will be created as if no errors had occurred 6 3 2 Real Time Scenario Creation It is possible to use Scenario Manager in an enhanced form to generate datasets and scenarios for real time forecast purposes This functionality and methodology is described in detail in Chapters 10 12 and includes the detailed calculations of how the fuel prices are calculated 6 3 3 Constrained Scenario Creation It is possible to use Scenario Manager and STM in an enhanced form to generate and run System Constrained Scenarios This functionality and methodology is described in detail in Chapter 13 6 3 4 BackCast Scenario Creation It is possible to use Scenario Manager and STM in an enhanced form to generate System Constrained Scenarios This functionality and methodology is described in detail in Chapter 14 6 4 Running Scenarios Forecast and Real Time Select Scenario then Manage from the main menu or press the toolbar Manage button T below the main menu The Scenarios which are available to be solved are listed and the user selects The available actions with be sh
5. Scenarios Help amp j ob G e Exit Settings CSVs Backecast Realtime Forecast Manage About New Scenario mx Create New Scenario Scenario Title Enter a descriptive scenario name Start Date 01 01 2010 EndDate 01 02 2010 a Number of Contigous Days 1 Allocate Processors Cores 1 Demand Extension 12 Scenario Case Data Sets Availability Pleaseselectyourdataset Please select your clata set B BackCastq1 10 Demand Data BackCast0910 Test Data 48 cols Bi Test with ddmmyyyy format Fuel Prices Please select your data set a Hydro Data Please select your data set z Inter Connector Please select your data set mt Plant Data Please select your data set 4 Pumped Storage Please select your data set a Backcast A Realtime Forecast Constraints Verification Errors Verify amp Save 5 Cancel Select an Availability Data Set from those in the list The user can select the number of processors to be used for the run which must be a number at least one less than the total number or processors including networked processors available for the run The Number of Contiguous Days is the number days from 1 to7 to be included in each optimisation run selecting 2 or more will slow the solve process significantly For unconstrained runs under the T amp SC is always one day However the system operator may look beyond one day say to two or ev
6. User select the number of instances EE ELP Day Ahead Model v1 0 0 4 for x64 m Ha x CCE Failed Pending Best Objective 2 151779E 007 Bound Objective 6 496529E 006 Total Iterations 69 81 Elapsed Time 00 41 Energy Link Stephen Mooney Scenario Selection LINE v Instame Un Solved Total Solved Failed EmorConole pest otiective Bound Objective Total Iterations Objective Tolerance Energy Link Stephen Mooney 17 gt The Advance to Next Day button can be used at any point Time to relative to move the solver on the next day whilst saving the current best results for the advanced day This function is designed for time pressured short term runs where the User can decide is an acceptable solution tolerance has been achieved before manually advancing to the next day gt The date of the current run day is shown in the bottom left hand corner with the elapsed time shown at bottom right The panel on the right also shows the current value of both the Best and Bound Objectives the tolerance percentage of the current Best Objective and the total number of iterations solutions analysed so far The panel under the selection buttons shows the current number of days solved pending and failed and also the number of days manually advanced kB ELP Day Ahead Model v1 0 0 4 for x64 Scenario Selection Scenario Instance Count Run Typ
7. applied are the key determinates of required run time Very powerful machines can get near final schedules after 30 40 seconds but may take several minutes to prove they have indeed found the optimal solution 22 fo tee M Solved Total Solved Failed Advanced Pending 1 1 1 0 0 pom Ss 0 Error Console Elapsed Time Advance to Next Day System Default Properties Max Memory Object IP Tolerance Time to Relative 120 v Time Limit 0 Iteration Limit Energy Link Stephen Mooney Finally some upgrades to STM may be provided with more than one solver type to emulate different types of solution types which may be used by the MSP Software If this option is enabled solver selection can be made using the third tab as illustrated below a ee ee enario Select Solved 0 J 0 Total iterations 0 ees Elapsed Time me ff Advance to Next Day ls IID 1 23 8 Background information on Uplift gt The Uplift calculation is run using a separate module This can only be used when all days in a scenario have been successfully completed by STM This is because costs and Shadow Prices must be available for the calculation to be performed The application is stated by double clicking on the Uplift exe icon The application shows all Scenarios Titles available to be calculated For each scenario
8. bids GO CO2 RolGas DBASGB NBP Moffat Agency Gas CV NTS Comm SO to Exit T Exchange Rates PoundperEuro CommodityRol Carbon tonne Bid CO2 in bids Gas CO2 CapacityRol NiGas GASGB_NBP Moffat_Agency Gas_CV NTS_Comm SO_to_Exit 10 Exchange_Rates_PoundperEuro CommodityNi Exchange Rates PoundperEuro 1 DO Carbon tonne Bid CO2 in bids Gas CO2 4CapacityNl HFO HFOARA HFO Trans RolyExchange Rates USDperEuro HFOExciseyHFO C 1000 Carbon tonne Bid CO2 in bids HFO CO2 NI Coal Coal amp RA Exchange Rates USDperEuro Coal TransNl Exchange Rates PoundperEuroy Coal Cv 10D0 Carbon tonne Bid CO2 in bids Coal CO2 NI Distillate GOARA GO TransNl Exchange Rates USDperEuroVGO C 1000 Carbon tonne Bid CO2 in bids GO CO2 40 12 Constraints Feature in Scenario Creation gt When creating a Scenario where the Constraints Module is installed the User can then select the dispatch constraints to be applied by selecting the Constraints tab at the bottom of the pane Constraint can be individually selected constraints as illustrated below for eg Edenderry These individual constraints are classified under the four categories of Plant Specific Location Specific Other Supply Side and Other Demand Side Once a constraint has been selected a chain appears next to it in the application as seen for Edenderry Scenario Manager alpha v0 0 1 11 3 File New Scenarios Help EH 2 e Exit Settings CSVs Backcast Realtime For
9. end and so is not available for runs until the last day of the month Therefore demand Data Queries runs toward the month end on the last but one day of a month will not have any data past D 4 This defaults to the annual forecast adjusted to the D 4 forecast A wind output MWh forecast 1s required over the D 7 period in order that the schedule demand can be calculated from the above demand forecast This 1s derived as follows Standing data for transmission connected and unconnected wind generation capacities has been set up for 2010 These capacities will need occasional review by the User The wind Data Queries return the latest D 2 rolling wind forecast demand A scaling factor is calculated based on the last day in the D 2 forecast and this is applied to the D 7 forecast to get an adjusted wind output profile for the rest of the 7 day forecast period Persistence is assumed thereafter if the D 7 is still short of the 7 day target 3l gt If available third party wind forecast up to 6 days ahead can be included in the above calculations If this data is available it will be mapped to the D 2 wind forecasts and rolled forward on the calculated correlation gt Standing data for embedded generation output MWh has been set up for 2010 These output assumptions will need occasional review by the User gt The embedded element of the wind output will be calculated as will the scheduled demand to be met by d
10. processed to create short term projections of the likely variations to this data based on a variety of algorithms and methodologies These are set out in more detail below The processing also includes forward market prices for fuels to be provided by the User These short term projections of inputs assumptions are provided in the required formats for direct use in Scenario Manager STM to produce short term price and often more importantly plant dispatch forecasts 10 2 Data Requirements 10 2 1 SEMO Data Requirements This upgrade to Scenario Manager requires a direct link to a database of SEMO published data Whilst in theory any data source could be ulitised the real time forecast upgrade is based around Energy Link s PDD database Further information on this database 1s available from Energy Link on request 10 2 2 Fuel Price Forecasts The upgrade includes an additional fuel input price form through which the User inputs raw fuel price and FX forecasts for up to seven days and the historical fuel prices and FX data for the base date of the observed plant 29 bidding behavior These historical prices are required to establish a basis point for calibration of changes to offer prices over the forecast period Fuel units are in commonly traded values i e p therm tonne etc 10 2 3 Constraints Assumptions To accurately model plant dispatch requires additional plant constraints to be applied to the in the Real Time foreca
11. 0 01 60 01 12 01 99 00 1 40 0 90 2 2010 05 31 30 31 300 31 700 31 60 31 12 31 99 00 1 40 0 90 a 3 2010 05 30 30 30 300 30 700 30 60 30 12 30 99 00 1 40 0 90 4 2010 05 29 30 29 300 29 700 29 60 29 12 29 99 00 1 40 0 90 5 2010 05 28 30 28 300 28 700 28 60 28 12 28 99 00 1 40 0 90 6 2010 04 25 30 25 300 25 700 25 60 25 12 25 33 00 1 40 0 90 5 Backcast C Realtime Forecast Constraints Verification Errors 4j Update Fuel fe Use Fuel Values Verify amp Save 2 Cancel Copvright c 2009 2010 Energy Link Partnership Ltd 62 The last previously input raw prices are pulled onto screen The User updates gas oil carbon prices and exchange rates by either typing the values or pasting them from another application e g Microsoft Excel When the prices have been updated the user commits them to the database by pressing the Update Fuel button The user then selects the Use Fuel Values button Scenario and the datasets are created First the raw fuel prices are converted to MWh including the relevant carbon and transportation excise costs for each fuel type and the fuel input dataset created The other 6 datasets follow If there are no data problems the Verify and Save button becomes highlighted and the User can press this button and has completed the scenario creation process returning to Manage Scenarios by clicking ie to run it 31 10 4 Error Handling gt If the Verify and Save button is grayed out the user sho
12. 00 25800 0 00 0 00 0 00 40 00 91 00 23700 23600 000 000 0 00 4000 0 00 40400 000 38500 000 0 00 000 0 000 33300 4 3124 2 2008 06 26 00 00 00 000 4 poemes Ge 000 38500 000 000 000 0 000 33300 5 315 2 2009 05 26 00 00 00 000 5 00 0 00 38500 000 000 000 0 000 33300 amp 12 2 2009 0626 00000000 6 59 ht Ca Documents and Settings O x a B to 00 0 00 38500 000 000 O00 0 00 33300 Z 27 2 2009 06 26 00 00 00 000 7 Administrator 00 0 00 38500 000 000 000 0 000 33300 8 3128 2 2009 05 26 00 00 00 000 8 3 All Users 00 000 38500 000 000 000 0 000 33300 g 3129 2 2009 06 26 00 00 00 000 9 Desktop Default User 00 000 38500 000 000 000 0 000 33300 10 3130 2 2009 06 26 00 00 00 000 1 00 000 38500 000 000 000 0 000 33300 EB 3131 2 2009 06 26 00 00 00 000 1 LJ 00 000 38500 000 000 000 0 000 33300 12 3132 2 2009 06 26 00 00 00 000 1 my projects 00 0 00 385 00 000 000 000 0 O00 33300 EE 20 3 2 2009 05 26 00 00 00 000 1 00 000 38500 000 000 000 0 000 33300 M 3134 2 2009 06 26 00 00 00 000 1 00 0 00 38500 000 0 00 000 0 000 33300 15 5135 2 2009 06 26 00 00 00 000 1 00 0 00 38500 000 000 000 0 000 33300 ae 3136 2 2009 06 26 00 00 00 000 1 My Computer 00 0 00 38500 000 000 000 0 000 33300 NEM 3137 2 2009 06 26 00 00 00 000 1 00 000 38500 000 000 000 0 000 33300 18 3138 2 2009 06 26 00 00 00 000 1 00 000 38500 000 000 000 0 000 33300 ag 3038 2 2009 06 26 00 00 00 000 1 00 000 38500 000 000 000 0
13. 9 2010 Energy Link Partnership Ltd G9 gt The User then inputs a Scenario Title which should be regular letters and spaces If this Scenario Title is not unique then a message box requests another scenario name be selected gt User the start and end dates for the run from the drop down calendars which defines the number of days to be run Once selected the menu will populate with datasets that exist for the selected date range only There may be a slight delay while this information is derived Scenario Manager alpha v0 0 1 11 lt a File New Scenarios Help Aj OP CG Le Exit Settings CSVs Backeast Realtime Forecast Manage About New Scenario mx Create New Scenario Scenario Title Enter a descriptive scenario name Start Date 01 01 2000 End Date 01 02 2000 Janu 2010 Number of Contigous Days 1 Allocate Processors Cores 1 CET n Sun Mon Tue Wed Thu Fri Sat Scenario Case Data Sets 27 29 30 31 L 2 3 4 5 6 7 8 9 Availability 10 LL 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Demand Data i 31 L 4 5 6 Fuel Prices Y Hydro Data E Inter Connector E Plant Data Pumped Storage m Backcast A Realtime B Forecast Constraints Verification Errors Copyright c 2009 2010 Energy Link Partnership Ltd Verify amp Save 35 Cancel G Scenario Manager alpha v0 0 1 11 s File New
14. CONFIDENTIAL STM and Scenario Manager User Application Manual Version 1 0 May 2012 The Intellectual Property Rights and copyrights of this document belong to Energy Link Partnership Limited The contents shall not be reproduced copied or passed to any third party without the express written permission of Energy Link Partnership Limited A SS y um SN t D W AINN Z pr ge Dec CS 7 CONFIDENTIAL 4 X ff LS D N 4 a WP dd N X CONTENTS Introduction Model Variants Summary of Model Environment Data Input Files 4 1 Introduction to Data Inputs 4 2 Availability 4 3 Demand 4 4 Hydro 4 5 Interconnector 4 5 1 GB Price Model 4 5 2 Interconnector Template 4 6 Plant Data 4 7 Pumped Storage A A 5 5 Error Bookmark not defined Error Bookmark not defined Error Bookmark not defined Error Bookmark not defined Error Bookmark not defined Error Bookmark not defined Error Bookmark not defined Error Bookmark not defined Error Bookmark not defined 4 8 Fuel Error Bookmark not defined 4 8 1 Fuel Template Error Bookmark not defined Application Operation 6 5 1 Scenario Manager 6 5 1 1 Input Files Library 7 5 1 2 Scenario Creation 8 5 2 STM Error Bookmark not defined S22 Basic STM Operations 15 3 2 2 Advanced STM Functions 19 5 3 Uplift 24 5 4 Results Extraction 26 5 4 1 Scenario Manager Results Extraction Error Bookmark not d
15. MSQ results from Completed STM runs can be viewed by selecting a scenario and pressing the MSQ report button This data can be placed on the clipboard by clicking the top left hand corner and right clicking and selecting copy Additional reports may be provided in future releases SQL can be used to extract other results if required see SQL reporting section Scenario Manager alpha v0 0 1 11 xn File New Scenarios Help Currently Available Scenario List o HN E E Manage BackCast qi 10 Z Solve Uplift CPM MSQ Failed Reset Del Output Preview mx Preview Pane Run Date Time MSQ 7 L 2010 03 30 48 302 00 2 2010 03 30 47 330 00 3 2010 03 30 46 330 00 4 2010 03 30 45 330 00 5 2010 03 30 44 330 00 6 2010 03 30 43 330 00 7 2010 03 30 42 330 00 8 2010 03 30 41 330 00 19 2010 03 30 40 330 00 10 2010 03 30 39 330 00 11 2010 03 30 38 330 00 12 2010 03 30 37 383 00 13 2010 03 30 36 383 00 14 2010 03 30 35 383 00 15 2010 03 30 34 383 00 16 2010 03 30 33 383 00 Fr Copyright c 2009 2010 Energy Link Partnership Ltd G9 14 7 Background information on STM gt STM is the core application which determines the plant Market Schedule Quantities MSQ and the associated Shadow Prices The application can be started separately from Scenario Manager and can be used to execute runs which have already been set up using Scenario Manager or manually in SQL The security dongle m
16. ant availability 1s reset to zero 11 3 Fuel Assumptions gt The fuel Data Query returns the exchange rates published by SEMO for application to the historical fuel price data gt The core calculations and inputs to be applied to the forecast fuel prices are set out in Appendix 3 gt Range validation is applied to the calculated prices e g to check that p therm has been input rather than therm 11 4 Plant Data Assumptions gt The plant data queries return the latest historical values for Conventional Plants and peat plants These are derived from Technical and Commercial Bids Exchange Rates Fuel Types and Fuel Prices and Plant Location and are combined with latest Bid Correlation Factors gt The bid correlation components of the Data Query read in the last 7 days bid data and fuel price data and calculate the correlation relationships for each fuel plant bid parameter utilising the latest historical P Qs exchange rate corrected data gt The Fuel Base prices are automatically calculated and inserted following completion of the fuel inputs above gt The parameter MSQ DQ which related to spinning reserve levels of each plant is input manually calculated and input automatically 34 gt The data is checked to ensure it has monotonically increasing Qs and Ps Ps of 999 are inserted wherever Qx Qx 1 to improve solution speed gt Peat plant bids are amended to ensure they are must run
17. d for both imports and exports This procedure sorts the data into 5 PQ import pairs and 5 PQ export pairs These prices are then scaled for each day and hence period within day by gas price correlation factors plus a constant value These gas correlation factors vary by Summer Winter and weekday and weekend This data is then validated e g to check that both Interconnector bids and offers are available on the base date for the plant data and fuel prices to ensure that gas correlation factors can be derived 11 7 Demand Data gt The demand Data Queries return the 2 day rolling forecast D 2 the 4 day rolling forecast D 4 and the month ahead forecast M 1 and the annual forecast A scaling factor relative to the month ahead forecast is calculated based on the last day in the D 2 forecast and this is applied to the D 7 forecast 36 to get a adjusted profile for the rest of the 7 day forecast period Persistence is assumed thereafter if the D 7 is still short of the 7 day target A scaling factor is calculated based on the last day in the D 4 forecast relative to the month ahead forecast and this is applied to the M 1 forecast to get an adjusted profile for the rest of the month Finally a combined forecast is built up from the scaled D 2 D 4 and M 1 forecasts with the shorter term forecasts having precedence over the longer term forecasts The M 1 forecast is published at the month
18. e Unconstrained Solved Total Solved Failed Advanced Pending 0 0 0 0 25 Error Console Best Objective 1 083328E 007 Bound Objective 6 633921E 006 Total Iterations 92773 Objective Tolerance 38 76 Elapsed Time 01 11 Energy Link Stephen Mooney Solving 03 11 2009 Please wait Online Energy Link Library Models IID 1 18 7 2 Advanced STM Functions gt More advanced functions are selected using Scenario from the main menu FE ELP Day Ahead Model v1 0 0 4 for x64 Instance Count Run Type v Unconstained Solved Total Solved Failed Advanced Pending 1 1 1 0 0 Error Console Best Objective Bound Objective Total Iterations 0 Objective Tolerance 0 Elapsed Time Energy Link Stephen Mooney Select a Scenario to Solve form those available in the list Online Energy Link Library Models IID 1 gt The run results from an entire scenario can be deleted by selecting Scenario then Reset Scenario The selected scenario run results will be deleted from the SQL database though the input files and scenario itself are unaffected The Scenario can be rerun from either Scenario Manager of from STM as above 19 FEl ELP Day Ahead Model v1 0 0 4 for x64 Scenario Selection Scenario Instance Count Run Type 1710 vw 1 Instance Unconstrained Solved Total Solved Failed Advanced Pending 1 1 1 0 0 Error Console Best Objec
19. e distinct scenarios could be run using only these inputs During loading all inputs are validated for formats and completeness Invalid inputs files are not loaded and error messages report the reason for failure 6 3 Scenario Creation The following describes the scenario creation process using Scenario Manager The User starts the application by double clicking the SMT icon which opens a splash screen and the various activities e g loading ODBC drivers connection to the various Databases etc can be seen in the top RHS When this is complete Scenario Manager opens From the main menu the User selects Forecast or RealTime 6 3 1 Forecast Scenario Creation Selecting Forecast or clicking on the Icon loads the view shown below 8 Scenario Manager alpha v0 0 1 11 lt File New Scenarios Help amp j G Y Exit Settings CSVs Backeast Realtime Forecast Manage About New Scenario mx Create New Scenario 4 Scenario Title Enter a descriptive scenario name Start Date 01 01 2000 End Date 01 02 2000 Number of Contigous Days 1 Allocate Processors Cores 1 Demand Extension 12 Scenario Case Data Sets Availability Y Demand Data x Fuel Prices ed Hydro Data a Inter Connector Plant Data Y Pumped Storage x Backcast A Realtime E Forecast Constraints Verification Errors Verify amp Save 5 Cancel Copyright c 200
20. ecast Manage About New Scenario mx Create New Scenario Scenario Title Real Time 2010 05 28 for 4 Days 15 59 24 Start Date 3 05 2010 End Date 01 06 2010 Number of Contigous Days 1 Allocate Processors Cores 1 Demand Extension 12 Default Constraint Sets Unconstrainedl 4 Plant Specific West Offaly C d Historical Last Month West UlTaly Constrainec All excluding PS B32 Kilroot Tarbert and Spinning Reserve Edenderry Constrained All excluding PS B32 Kilroot and Tarbert Lough Rea Constrained All excluding Pump Storage Limits All available constraints Kilroot Constrained MSQ DQ Only Ballylumford Unit B32 E 7 Monevpoint Constraint 6 4 Location Specific La Ballvlumford Site Maximum Output C rstrained Moyle Import Limits o Moyle Export Limits Tarbert Dav Time Constraint Tarbert Night Time Constraint Aghada Marina Day Time Constraint Aghada Marina Night Time Constraint Aghada Marina 80Mw Night Time Constraint Aghada Marina 20MW Night Time Constraint Tarbert L2S5MW Day Time Constraint Tarbert 125MW Night Time Constraint 4 Other Supply Side On Time Dav Backcast O Realtime 3 Forecast Constraints Verification Errors Update Fuel Use Fuel Values A Verify amp Save St Cancel 6 6 Copyright c 2009 2010 Energy Link Partnership Ltd 6 e 9 e gt Alternatively the User can then select a predefined constraint set from the drop down list based on pre
21. efined 5 4 2 Results Extraction from SQL 26 5 4 3 SQL Queries 24 Appendix 1 Real Time Forecast Functionality 20 6 1 Introduction to Real Time Forecasts 20 6 2 Data Requirements 20 6 2 1 SEMO Data Requirements 20 6 2 2 Fuel Price Forecasts 20 6 2 3 Constraints Assumptions 30 6 3 Real Time Forecast Creation 30 Appendix 2 Real Time Forecast Algorithms and Methodologies 33 ii CONFIDENTIAL 7 1 Introduction to Algorithms and Methodology 7 2 Availability Assumptions 1 2 1 122 7 2 3 7 24 1 25 Fuel Assumptions Plant Data Assumptions Pumped Storage and Hydro IC Data Demand Data Appendix 3 Fuel Price Calculations 8 1 Fuel Inputs 8 2 Transportation and Excise Inputs 8 3 Exchange Rates and Carbon Bid Inputs 8 4 Conversion Factors 8 5 Calculations SS N 3 quo t WP if Se g AN cash CL E T 1 Introduction Energy Link Partnership Limited Energy Link has developed STM an SEM System Marginal Price and Capacity Payment Mechanism CPM price forecast application The objective of this application is to simulate the operation of the Two Stage relaxed integer unit commitment approach adopted by the MSP Software Details of the precise workings of this engine are unknown Energy Link updates and refines the solution process within STM as SEMO market data becomes available The basic STM model can be upgraded to include additional features such as d
22. en three days so as to optimise total physical dispatch costs over this period e g including shutdown costs Whether this longer term optimisation is occurring can be inferred from analysis of actual published physical dispatch data Regardless of the number of days selected only the 10 results for the first day are saved to the SQL database the remaining days being discarded The Demand Extension Period is the extra time in half hours from 6 to18 to added to each optimisation run which for unconstrained runs under the T amp SC is always 12 half hours This is again to reflect that the System Operator may be considering different total optimisation periods within the constrained run The default value 1s 12 half hours The User then selects the pre loaded Case Data for each of the seven input categories from the drop down menus Finally once a valid set of Case Data has been selected the User can press the Validate and Save button and final validation will be undertaken on the compatibility of the selected datasets For example is it only after the plant and availability data cases have been selected that Minimum Stable Generation Minimum On Time and Minimum Off Time validation routines can be applied The action taken on error depends on the AutoCorrect MOFFT MONT and MSG Errors are selected via the toolbar Setting button TE SM Default Properties ODBC Name sM_OCT dl CSV Output Path to STA n
23. ermination of the service as per the terms and conditions of the Model Agreement Scenario Manager is written primarily in opensource QT and C utilizing Boost maths libraries Source code is complied with the Intel C compiler professional version 11 0 1 54 All inputs and outputs relating to the applications are stored in an SQL database Data can be loaded or extracted to from the SQL database using Scenario Manager or for experienced SQL users directly to SQL itself from other databases or csv files 4 Data Input Files Before using the applications it is important that the User is familiar with the seven main application input files These inputs are stored as permanent entries until deleted to the SQL database and can be used for any number of scenarios All key inputs are in comma delimited csv format and are described in full detail in the attachments Data Input Files accompanying this manual and outline the validation and checking that is available through Scenario Manager 5 Application Operation There are three components to the overall price forecasting application which can be used individually Scenario Manager STM and the Uplift module In practise the majority of the functionality of the three applications is available through use of Scenario Manager itself Each of the three applications is described in more detail below 6 Scenario Manager The objective of Scenario Manager is to simplify the creation of
24. etailed capacity payment revenues dispatch forecasting back testing and real time short term forecasting To assist easy of use of STM Energy Link has developed a parallel application Scenario Manager The objectives of this application are to facilitate loading and validation of inputs to STM creation and launching of runs and extraction of results Finally an Uplift application is provided which calculates Uplift and hence System Marginal Price SMP as a post process after each STM scenario run This manual relates to STM version v1 0 0 8 x64 Scenario Manager version alpha V 0 0 0 11 and Uplift Calculator v1 0 0 4 x64 2 Model Variants As noted above STM and Scenario Manager can be specified with a number of upgrades dependent on the user requirements The additional features of these upgrades are set out in Appendices 1 and 2 Abbreviation for SEM Trading and Settlement Code Model 3 Summary of Model Environment STM is written primarily in C and runs on the NET Framework v 3 5 or above running Windows It utilises the new Lingo 12 0 API a third party add in from Lindo systems inc to provide a Mixed Integer Programme solving capability The licenses required to operate these have been purchased by Energy Link and to ensure compliance with these the STM application is secured with a security dongle which must be inserted into a free USB port on the host PC This security dongle must be returned with the t
25. ethodologies applied to the returns from these Data Queries in producing forecast data sets 11 2 Availability Assumptions gt The availability Data Query establishes the maximum of the MSQ and actual operation MW from the published Ex Ante Schedule for each plant The query also returns the actual Ex_Post_Indicative schedule availability for each plant and takes the maximum of the ex ante derived an ex post values as the basis of the availability forecast This 1s to ensure that if a plant which was actually unavailable ex post 1s forecast to be available ex ante then the forecast availability takes precedence gt Persistence is applied to this base availability for period of forecast The relevant persistence value is derived from the last half hour the entire last 24 hour period of data and projected forward This is compared with the each plant rating MW capacity contained in the Rating Table gt In addition to the above SEMO publishes a monthly Outage Return file in PDF format The User can manually update the application with this data or other data and this data will automatically override the persistence values calculated above gt The availability forecast for a specific generator can be directly loaded to PDD which will then be used as part of the forecast dataset 33 gt Validation is applied so that no half hourly availabilities are greater than zero but lower than the MSG In this case the relev
26. ispatchable plant in the dispatch schedule 11 8 Fuel Price Calculations The fuel price calculations are derived from the following input conversion and calculation components 11 9 Fuel Inputs GASGB NBP p therm HFOARA tonne GOARA tonne CoalARA tonne User Input User Input User Input User Input Carbon tonne tonne User Input 11 10 Transportation and Excise Inputs NTS Comm SO to Exit Moffat Agency CapacityRol CapacityNI p kWh 0 0181 p therm 0 MWh 0 MWh 0 38 CommodityRol MWh 0 284 CommodityNI kWh 0 0005352 HFO Trans Rol tonne 15 HFOExcise tonne 1 306 GO Trans Rol tonne 33 GO ExciseRol tonne 56 07 GO TransNI tonne 26 Coal TransRol USD tonne 0 8 Coal TransRol Euro tonne 0 9 Coal TransNI tonne 8 6 11 11 Exchange Rates and Carbon Bid Inputs Exchange Rates Exchange Rates CO2 in Bid COz2 in bids bids 11 12 Conversion Factors CO2 Coal CO2 tonnes MWh CO2 Gas CO2 tonnes MWh CO2 HFO CO2 tonnes MWh CO2 GO CO2 tonnes MWh Coal CV Gas OV HFO CV GO OV KWh tonne KWh tonne KWh tonne KWh tonne 39 11 13 Calculations FuelPrices Formula Rol Coal Coal amp RA Coal TransRol USD Exchange Rates USDperEuro Coal TransRol Eura Coal Cv 1000 Carbon tonne Bid CO2 in bids Coal CO2 RolGasoil GOARA GO Trans RolyExchange Rates USDperEuro GO ExciseRol GO Cv 10D00 Carbon tonne Bid CO2 in
27. lts saved to the SQL database before the entire run completes 9 1 Results Extraction from SQL Results can be extracted directly from SQL using a small library of SQL queries These must be edited to reference the correct Scenario Title as defined by a single numerical value The results for any scenario can be saved out from SQL as a text or RPT format file for easy analysis in spreadsheet applications These formats are easily converted to CSV or XLS formats This form of extraction is illustrated below LAS Microsoft SQL Server Management Studio Express File Edit View Query Tools Window Community Help Bnew Query Liy G7 il at 3 BB Be gu Tri E A lid 5 ae se LAS 2 ty 8 3 26 _ 127 0 0 1 MSQ Sorted sql view dbo ScenarioList Summary X SELECT FROM MSQ WHERE Scenario ID 42 ORDER BY RunDate ASC Time ASc Results FE Messages ib Scenario ID RunDate Time AD1 aps ati Ar2 ata B10 ex 832 ecri ecr2 eoc4a Bocs Bocs caso careloet ent an a2 cz HN2 HNC Kkcaa 1 32012 2009 06 26 00 00 00 000 1 00 25800 0 00 000 000 59 50 91 00 237 00 23600 000 000 000 0 00 0 00 40400 000 38500 000 0 00 000 0 000 33300 0 00 EN 3122 2 2009 06 26 00 00 00 000 200 25800 0 00 0 00 000 40 00 91 00 23700 23600 000 000 0 00 4000 0 00 40400 000 38500 0 00 0 00 000 0 000 33300 EN 3123 2 2009 06 26 00 00 00 000 3
28. menu The queries can then be saved under User chosen names using the File then Save as buttons on the main menu gt The queries have to be edited to include the numerical value which represents the Scenario Title in the SQL database This edit is made at Scenario _ID 1 in the query replacing 1 with the appropriate value This numerical value can be obtained by first navigating to the Lingo database in SQL and selecting Tables The Scenario table should then be opened by right clicking 27 E Microsoft SQL Server Management Studio Express File Edit View Query Designer Tools Window Community Help X U Tables CARLO MOBYSQLEXPRESS Databases ingo Tables 35 Item s Name Schema Created Availability dbo 30 07 2009 CapacityPeriodPayments dbo 30 07 2009 CapacityPeriodPaymentsum dbo 30 07 2009 Costs dbo 30 07 2009 CPGP Data dbo 30 07 2009 CPM dbo 30 07 2009 CPM Data dbo 30 07 2009 C Demand Calcs dbo 30 07 2009 Demand Data dbo 30 07 2009 E FlagZero dbo 30 07 2009 FuelPrices dbo 30 07 2009 Fuels dbo 30 07 2009 El Hydro Data dbo 30 07 2009 E HydroPS dbo 30 07 2009 cJIC Data dbo 30 07 2009 LossOfLoad dbo 30 07 2009 Margin Data dbo 30 07 2009 E MSQ dbo 30 07 2009 E NettingGeneration dbo 30 07 2009 PlantData dbo 30 07 2009 E PSHICOutput dbo 30 07 2009 Scenarios dbo 30 07 2009 ScriptKeys dbo 30 07 2009 Script
29. ow shows that the SQL database is linked to a local 1SQL instance However remote connections both over internal networks and public internet can be set by specifying a Server Address and associated port number Elapsed Time mm Advance to Next Day Energy Link Stephen Mooney Select a Scenario to Solve form those available in the ist 21 Solver settings can be selected which affect both the speed and accuracy of each daily run and hence the entire scenario The main settings are gt Max memory the amount of machine RAM allocated to the solution Reducing the memory below 20MB 1s likely to result in solver errors as some point gt IP Tolerance the minimum percentage difference between the best and bound objective before the solver moves on the next daily run or completes gt Time to Relative the minimum time seconds before the solver applies the selected IP Tolerance set above gt Time Limit the maximum run time seconds for each daily run Selection of 0 means that there 1s no time limit 1 e there 1s unlimited solution time until the required IP Tolerance is met Setting a positive solve time limit will prevent modest machines from hanging on difficult days gt Iteration Limit the maximum number of solutions which can be analysed before the solver moves on the next daily run or completes gt The IP Tolerances and the time before which they can be
30. own Highlighted on the toolbar All newly created Scenarios should offer the v Solve option Pressing the Solve button Es starts the STM module which appears in the foreground showing key run information e g running time solved etc For Constrained Scenarios no other options should be available as Uplift and CPM are not applicable All Unconstrained Scenarios should offer the 12 Uplift button and if of 1 or more calendar months duration the CPM button 5 Scenario Manager alpha v0 0 1 11 mn File New Scenarios Help Currently Available Scenario List e 8 EE Manage Please Select a Scenario 1 v Solve Uplift CPM MSQ Failed Reset Del Please Select a Scenario Output Preview BackCast q1 10 ex BackCast q4 09 v2 BackCast qi 10 Preview Pan v2 BackCast q4 09 BackCast 0910 BackCast 0910v2 Real Time 2010 05 18 for 4 Days 22 34 16 Copyright c 2009 2010 Energy Link Partnership Ltd 62 gt Unconstrained Scenarios which have been Solved will have the Uplift Button available and pressing it launches the Uplift application and starts calculating Uplift for the selected scenario gt Unconstrained Scenarios of calendar month duration which have been Solved and had Uplift calculated will have the CPM Button available and pressing this fires the CPM application in the background This typically takes 15 minutes per 12 month period 6 5 Scenario Manager Reporting 13 Summary
31. s dbo 30 07 2009 Cl States dbo 30 07 2009 E E StopZero dbo 30 07 2009 E Tasks dbo 30 07 2009 tmp Availability dbo 30 07 2009 tmp MSQ dbo 30 07 2009 za LE tmn Loft dha 20 07 2009 vi Ready gt The numerical ID number is displayed in the first column against the required Scenario Title E Microsoft SOL Server Management Studio Express File Edit View Query Designer Tools Window Community Help Li New Query a dg m ER ali E mj Change Type gt 1 sa SFE Table dbo Scenarios Summary X Title Starting Endina MaxCores Interval AvalabityC FuelPricesC PSCase ICCase DemandCase PlantDataC HydroCase ExtPeriods July2009 01 07 2009 31 07 2009 1 1 1 1 1 1 1 1 1 12 2010 01 01 2010 01 01 2011 1 1 2 2 1 2 2 2 2 12 NULL NULL NULL NULL NULL NULL NULL NULL NULL NULL NULL NULL NULL 9 9 9 o 9 9 lt K 4 1 of2 b bi b Cellis Read Only Ready E 28 10 Real Time Forecast Functionality 10 1 Introduction to Real Time Forecasts An upgrade to Scenario Manager can be provided which includes additional functionality to facilitate short term forecasts The objective of this upgrade is to create accurate input files for short term forecasts with as little manual input and expert decision making as possible utilising the most recent market data published by SEMO as the reference point This published data is
32. scenarios the running of uplift and CPM the loading and validation of inputs and the extraction of run outputs Scenario Manager can be used to Validate name and load input files Create scenario names and associated run ranges Select named input files for each scenario Launch scenario run which sparks STM Launch Uplift calculation following scenario run VV VV NV WV Extract Results 6 1 Loading Input Files Input files can most easily be loaded by selecting the toolbar button and dragging and dropping from another window into the white file input pane There can be a slight delay between dropping the files and them appearing in the input pane especially if the files are large There is also a browse button to allow navigation to specific directories File loading is illustrated below A Scenario Manager alpha v0 0 1 11 Ss File New Scenarios Help amp j E C n KD Fey Exit Settings CSVs Backcast Realtime Forecast Manage About Browse Import Cancel CSV File Importer Bx CSV File List File Data Type Rows State Case Description L P csv Plant 48 Sample PlantData File Copyright c 2009 2010 Energy Link Partnership Ltd t 6 2 Input Files Library Each CSV input which is loaded becomes part of a library with a unique name which can be used in multiple Scenarios For example there may be a high medium and low fuel price input though only one of each of the other types of input In total thre
33. st Under these forecast runs the reported MSQs for each plant are the forecast Dispatch Quantities DQs based on the forecast data set Under the Trading and Settlement Code T amp SC all runs used for setting SMP and hence for financial settlement are unconstrained 1 e they do not apply these additional plant constraints The User should not therefore use any price data derived from constrained plant dispatch forecasts in any financially based forecasts However the resulting DQs produced can be used to estimate constraint payment settlement under the T amp SC 10 3 Real Time Forecast Creation The following sets out the steps to be followed to create a real time forecast gt The User opens Scenario Manager and selects Real Time forecast The application moves to the Real Time Fuel inputs section illustrated below 30 Scenario Manager alpha v0 0 1 11 File New Scenarios Help a oe 7 Exit Settings CSVs Backeast Realtime Forecast Manage About New Scenario Bx Create New Scenario Scenario Title Real Time 2010 05 28 for 4 Days 15 59 24 Start Date 1 05 2010 End Date 01 06 2010 Number of Contigous Days 1 Allocate Processors Cores 1 Demand Extension 12 4 Raw Fuel Pricing Tansportation and Data Sets Gas HFO GO Coal Carbon Peat E Rate E Rate eh RunDate p therm tonne tonne fytonne tonne tonne to to L 2010 06 01 30 01 300 01 70
34. the start and end dates total number of days in the run and numbers of days both solved and failed are displayed ELP Uplift Calculator v1 0 0 2 x64 File Scenaria List Title C Decog C Novog C Nov09 BT Con C Nov09 BT Con V2 C Nov09 BT Con V3 Starting 01 12 2008 03 11 2008 03 11 2008 03 11 2008 03 11 2009 Ending 31 12 2009 28 11 2009 28 11 2009 28 11 2003 28 11 2009 Copyright c 2009 Energy Link Partnership Ltd gt The scenario is selected by ticking the appropriate box and the calculation started by clicking Calculate ELP Uplift Calculator v1 0 0 2 x64 File Scenario List Title Deco9 M Novos C Nov09 BT Con Nov09 BT Con V2 E Nov09 BT Con V3 Starting 01 12 2009 03 11 2009 03 11 2009 03 11 2009 03 11 2009 Ending 31 12 2009 28 11 2009 28 11 2009 28 11 2009 28 11 2009 Calculate Copyright c 2009 Energy Link Partnership Ltd 24 gt The calculation for each day takes 1 2 seconds and the uplift results are written to the SQL database Once all days have been calculated the applications reports Done The application can then be exited by selecting File then Exit from the main menu 25 9 Results Extraction All output data for each run is stored securely in the SQL database During both STM and Uplift runs results are continuously written to the SQL database If required the User can view extract resu
35. tive 0 Bound Objective 0 Total Iterations 0 Objective Tolerance Confirm Reset 0 Are you sure you want to do this to 1710 This will delete any existing results and reset the Tasks List Elapsed Time 00 00 Advance to Next Day Energy Link Stephen Mooney Select a Scenario to Solve form those available in the list Online Energy Link Library Models IID 1 gt Alternatively where there are only a small number of failed days in a run these can be individually reset by selecting Scenario then Failed Days and confirming the delete Only the reset failed days will be rerun if the scenario is then reloaded and rerun In the event that particular days fail to Scenario Selection Scenario Instance Count Run Type 7 Dec09 w 1 Instance Unconstrained Solved Total Solved Failed Advanced Pending 29 29 2 0 0 Error Console Best Objective 0 Bound Objective 0 Total Iterations 0 Objective Tolerance 0 onfirm Reset Scenario contains 2 faled days Would you Ike to reset these Ce Elapsed Time 00 00 Energy Link Stephen Mooney Select a Scenario to Solve form those available in the list Onine Energy Link Library Models IID 1 20 gt The SQL database connections for the applications can be specified by selecting File then Settings 00 00 EnegyLnk Stephen Mooney Select a Scena ario to Solve orm thos avaiable nt wie wm gt The example bel
36. uld select the Dataset tab One of the 7 files listed will be missing and the data sources required to complete it should be double checked File New Scenarios Help amp j Qe Ej 4j T Ej psc Exit Settings CSVs Backcast Realtime Forecast Manage About New Scenario Create New Scenario Scenario Title Real Time 2010 05 28 for 4 Days 15 59 24 Start Date 2805 2010 End Date 01 06 2010 Number of Contigous Days 1 Allocate Processors Cores 1 z Demand Extension 12 Raw Fuel Pricing Tansportation and Data Sets Data Type Case Description Availability Availability for Real Time 2010 05 28 for 4 Days 15 59 24 Demand Demand for Real Time 2010 05 28 for 4 Days 15 59 24 sysop PNJ Fuel Prices Fuel Prices for Real Time 2010 05 28 for 4 Days 15 59 24 Hydro Hydro for Real Time 2010 05 28 for 4 Days 15 59 24 21V UK IC UK IC for Real Time 2010 05 28 for 4 Days 15 59 24 Plant Plant for Real Time 2010 05 28 for 4 Days 15 59 24 Pump Storage Pump Storage for Real Time 2010 05 28 for 4 Days 15 59 24 ss epa Reattime O Verification Errors A Update Fuel Use Fuel Values H Verify amp Save Copyright c 2009 2010 Energy Link Partnership Ltd 32 11 Real Time Forecast Algorithms and Methodologies 11 1 Introduction to Algorithms and Methodology Appendix 1 outlines how the Data Queries are performed when conducting Real Time Forecasts This section sets out in more detail m
37. ust be inserted in a free USB slot for STM to operate gt Solve times per day are approximately 2 minutes in full Mixed Integer Programming MIP mode though can be considerably longer dependant on the solution complexity A one year run typically takes 12 hours depending on machine speed A multi processor machine can execute concurrent operations and reduce overall solve times by a factor of 3 7 depending on the machine and how many instances are running For example one year on an 8 core PC would take 2 hours e g 365 days are split into 7 53 day runs undertaken concurrently with overlap All instances will be open on screen for monitoring gt Users will use STM predominately to select more advanced functions for runs These main advanced functions are described below 7 1 Basic STM Operations Onstarting up STM the screen below is displayed which will show the name of the first alphabetically scenario if any exist 15 Total Solve d Best Objective Total Iterations 0 a EN Elapsed Time 00 00 Energy Link Stephen Mooney gt To select a particular scenario the User selects the Scenario drop down button to display all available scenarios EE ELP Day Ahead Model v1 0 0 4 for x64 C CC 1710 i BI Failed 1 1 Eror Console BestObjective Bound Objective Objective Tolerance 0 Energy Link Stephen Mooney 16 gt As with Scenario Manager the
38. vious use and or preference Once the 41 constraint set is chosen the applicable individual constraints are automatically shown by a chainlink Constraints are applied to the scenario being created and persisted for the next Scenario forecast Real time keep separate preferences Periodic analysis of the actual application of constraints in the SEM schedules can be undertaken on request to better inform the application of this module 42

STM User Manual

Contents

Download Pdf Manuals

Related Search

Related Contents