2 - Simulation des grands systèmes, projet PEGASE

11/06/2013
Auteurs :
OAI : oai:www.see.asso.fr:4404:4406
DOI :

Résumé

2 - Simulation des grands systèmes,  projet PEGASE

Métriques

28
5
1.62 Mo
 application/pdf
bitcache://73c8de301576b04aa4c3847d17e4f00276242bf2

Licence

Creative Commons Aucune (Tous droits réservés)

Sponsors

image1.jpeg
<resource  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
                xmlns="http://datacite.org/schema/kernel-4"
                xsi:schemaLocation="http://datacite.org/schema/kernel-4 http://schema.datacite.org/meta/kernel-4/metadata.xsd">
        <identifier identifierType="DOI">10.23723/4404/4406</identifier><creators><creator><creatorName>Bertrand Haut</creatorName></creator></creators><titles>
            <title>2 - Simulation des grands systèmes,  projet PEGASE </title></titles>
        <publisher>SEE</publisher>
        <publicationYear>2013</publicationYear>
        <resourceType resourceTypeGeneral="Text">Text</resourceType><dates>
	    <date dateType="Created">Sat 15 Jun 2013</date>
	    <date dateType="Updated">Mon 25 Jul 2016</date>
            <date dateType="Submitted">Fri 17 Aug 2018</date>
	</dates>
        <alternateIdentifiers>
	    <alternateIdentifier alternateIdentifierType="bitstream">73c8de301576b04aa4c3847d17e4f00276242bf2</alternateIdentifier>
	</alternateIdentifiers>
        <formats>
	    <format>application/pdf</format>
	</formats>
	<version>28838</version>
        <descriptions>
            <description descriptionType="Abstract"></description>
        </descriptions>
    </resource>
.

PEGASE PROJECT Paris, June 11, 2013 2 Outline PEGASE general presentation Context Project description Project main achievements Focus on time domain simulation Objectives Algorithm improvements Prototypes demonstration Conclusions CONTEXT 4 Historical background up to year 2000 Power sector characteristics Utilities were most of the time national monopolies vertically integrated Interconnection of national networks used mainly for generation reserve sharing Consequences The flows on the interconnectors were normally low or zero. In case of major disturbances, the interconnections were intentionally tripped in order to avoid the extension of a possible blackout. The N-1 network security was managed at national level. Very few data exchange between countries was needed. 5 Today the paradigm of the ETN has changed Liberalization of the electricity market pushes to the use of the total available cross border transfer capacity Integration of a large amount of intermittent renewables Introduction of new ICT to improve the awareness of the system and its commandability Further extensions of the ETN considered Need for new solutionsNeed for new solutions PROJECT DESCRIPTION 7 PEGASE project objectives The PEGASE project addressed these issues and aimed at removing associated technical barriers By developing new tools to be capable of > Monitoring > Simulating > Optimizing The ETN as a whole. To support its real time control and its operational planning. Development of powerful simulation and optimization tools able to run the entire European Transmission Network Development of powerful simulation and optimization tools able to run the entire European Transmission Network 8 Power System Operation Paradigm Static & dynamic assessment Man Machine Interface Man Machine Interface Static & dynamic security assessment Real time operation Operational planning: day ahead to hour ahead Data acquisition & manager Individual Operators’ Perceptions State estimation Building anticipated state Single Country with limited neighbouring areas (external equivalents) 9 Power System Operation Paradigm Static & dynamic assessment Man Machine Interface Man Machine Interface Static & dynamic security assessment Real time operation Operational planning: day ahead to hour ahead Data acquisition & manager Individual Operators’ Perceptions State estimation Building anticipated state 10 The PEGASE consortium structure TE, Ulg, Elia TU/e RTE, CRSA-ECP, DIGITEO REE, AICIA NUCLEO REN FGH, UDE UoM DELING HEP TRANSELECTRICA ENERGOSETPROJECT SO-UPS TEIAS RTU LITGRID Consortium Nature 9 TSOs & Expert Companies 1 Manufacturer 11 Research Centers & Universities Project duration: 4 Years June 2008 – June 2012 Funding: EU: 8.6 M€ Total: 13.6 M€ PROJECT MAIN ACHIEVEMENTS 12 Smart Tools for Transmission Grid 1. New generation of groundbreaking algorithms 2. Prototypes demonstrated on Pan-European systems 3. Paving the way for future research and for industrialization Power System Operation SmartToolsforTransmissionGrid State Estimation Steady State Optimization Time Domain Simulation Dispatcher Training Simulator 13 Special focus on time domain simulation 1. New generation of groundbreaking algorithms 2. Prototypes demonstrated on Pan-European systems 3. Paving the way for future research and for industrialization Power System Operation SmartToolsforTransmissionGrid State Estimation Steady State Optimization Time Domain Simulation Dispatcher Training Simulator 14 Context Objective Develop new algorithms to > Simulate the time electro-mechanical evolution of very large power system > Decrease sufficiently simulation time for: - Full accuracy simulation for reference and offline studies - Dynamic Security Assessment by computing approximate trajectories - Real time simulation for DTS purpose PEGASE target Pre-Pegase status 15 min > 1 hour 15 sec / realtime 10 x realtime PEGASE target: develop and integrate these new algorithms in three prototypes PEGASE target: develop and integrate these new algorithms in three prototypes Context Increasing need for time domain simulation on Pan-European model > The system is operated closer to the limits > More protections and controllers > Poorly damped inter-area power oscillations can occur in large synchronous power system 15 ACHIEVEMENT 1: NEW GENERATION OF GROUND-BREAKING ALGORITHMS 16 Target problem Simulation of the electro-mechanical evolution of very large power system Very though mathematical problem Very large : +/- 140.000 variables for the European system Non-linear : park’s equations, loads,… Discontinuous : relay/switch in controllers Stiff Oscillating and badly damped Need of new advanced algorithms to reduce computation time 17 New generation of ground-breaking algorithms 18 Multi rate algorithm Many disturbances impact only a small part of the system Use of small stepsize for “fast variables” and large stepsize for “slow variable” Full accuracy prototype I Simplified prototype I DTS engine prototype fast variable: 26 steps slow variable: 5 steps single-rate multirate ||error|| 7.64 10-2 5.28 10-2 ||error||2 4.22 10-5 4.22 10-5 func evals 184 326 47 102 Simulation time 0.045 0.026 Multi rate computation for two components 19 Error control management Full accuracy prototype I Simplified prototype I DTS engine prototype The step size selection depends on the norm of a local error estimation an unsuitable norm can lead to the misdetection of some instabilities Development of a new norm suitable for very large system Approach: mixing L2 and L norm depending on the step size Lead to smaller step sizes allowing to detect instability In time domain simulation, a guarantee on the accuracy can be obtained only by adapting dynamically the stepsize length 20 Higher order methods Full accuracy prototype I Simplified prototype I DTS engine prototype An order 4 (Hammer-Hollingsworth) integration scheme investigated Allows to take larger steps but each step takes more time Suitable only when very high accuracy is targeted (tolerance less than 10-4) 21 Decomposition methods Full accuracy prototype I Simplified prototype I DTS engine prototype Coarse grain approach: parallelize the resolution of the system on different processors Technique: Cut the system into n subsystems Solve each system simultaneously on different processors Exchange values on the boundaries and iterate up to convergence > Boundary values assumed could have been erroneous. Mandatory to have good accuracy Benefit brought by parallelization limited by the additional iterations Speedup only if small number of iteration > Iteration number depends on many factors > Key factor: a preconditioner has to be introduced between two iterations Testcase 22 ACHIEVEMENT 2: PROTOTYPES DEMONSTRATED ON PAN-EUROPEAN SYSTEMS 23 Full European size testcase Static data : > anonymized and noised load flow model of UCTE (now ENTSO/e) EHV grid + Turkish EHV grid provided by TEIAS, which consists in real static data Dynamic data: > Fictitious data and controllers created to respect ENTSO-E statistics energy source Nodes 16578 Lines 14044 Transformers 9654 Generators 3240 Wind farms 707 HVDC 2 VSC, 1 LCC Total load 400 GW Nbr. of variables 140 000 All prototypes validated on a testcase representative of the full European size system Benchmarked on “standard” computer architecture 2 x Xeon X5690 @ 3.47GhZ CPU 48 GB RAM 24 Full accuracy prototype demonstrated on pan-European system Initial PEGASE target : 15 min Thermal cascade leading to system splitting simulated in full accuracy prototype CPU time : 12 min 35 s A 400 kV line is tripped at the time t =1s. As consequences nearby lines are overloaded and tripped by thermal protection. System splits at FR-ES border at time t=6.18s. Frequency in ES and PT part dips and when it reaches 49.5 Hz load shedding is activated at time t = 48 s. At the same time 3x600 MW generating units are shut down in FR area to restore power balance. After frequency restoration interconnection branches between ES and FR are closed causing successful resynchronisation at time t = 150s. 25 Full accuracy prototype demonstrated on pan-European system Unique combination of Performances on very large power systems > System splitting and resynchronization: 12 min 35 sec. > Voltage collapse: 4 min 42 sec Flexibility For model definition: user defined controllers through block diagrams For post-processing: access to all variables even without any pre-selection Accuracy Guarantee on the solution accuracy thanks to the variable step size strategy 26 DTS prototype demonstrated on pan-European system Performance improvements real time achieved Merging of the new investigated algorithms to achieve real-time Difference between simulation time and wall clock time seamless to operators 27 ACHIEVEMENT 3: PAVING THE WAY FOR FUTURE RESEARCH AND FOR INDUSTRIALIZATION 28 PAVING THE WAY FOR INDUSTRIALIZATION Performances reached are sufficient for time domain simulation of the whole Pan-European system Less than 15 minutes for a full accuracy simulation Around 15 seconds for a simplified simulation Real time simulation for DTS Prototypes validated and ready for industrialization Prototypes usable by other research projects Full accuracy prototype (based on EUROSTAG) to be used in iTesla project 29 Conclusions PEGASE was a major project addressing four topics in the field of very large power systems State estimation, steady state optimization, time domain simulation and Dispatcher Training Simulator Some key figures 150.000+ hours 30+ publications 7 prototypes Prototypes demonstrated on an European size testcase 30 How to remain updated with PEGASE results? www.fp7-pegase.eu : Public version of the deliverables posted QUESTIONS COMMENTS?