LV4MV: A CONCEPT FOR OPTIMAL POWER FLOW MANAGEMENT IN DISTRIBUTION GRIDS, USING DER FLEXIBILITY



Similar documents
Operating the power system while enabling more renewables : which solutions?

Il Progetto INTEGRIS. Risultati e nuove prospettive per lo sviluppo delle Smart Grid October 2nd 2013 Brescia

SOFTWARE FOR THE OPTIMAL ALLOCATION OF EV CHARGERS INTO THE POWER DISTRIBUTION GRID

VOLTAGE CONTROL IN DISTRIBUTION SYSTEMS AS A LIMITATION OF THE HOSTING CAPACITY FOR DISTRIBUTED ENERGY RESOURCES

Energy Storage Systems in Grid4EU Demo4 and in Other Enel Distribuzione Projects Speaker

Venteea A Smart Grid demonstrator for the DRES integration on the MV distribution networks.

ADMS(Advanced Distribution Management System ) in Smart Grid

ENEL plans for storage introduction in Italian distribution network

Preparing for Distributed Energy Resources

Research of Smart Distribution Network Big Data Model

MANAGEMENT OF ELECTRIC VEHICLES CHARGING PROCESSES IN A DSO CONTROL CENTER

De Jaeger, Emmanuel ; Du Bois, Arnaud ; Martin, Benoît. Document type : Communication à un colloque (Conference Paper)

Executive summary. by Michel Clemence, Renzo Coccioni, and Alain Glatigny

APPLICATION NOTE. Increasing PV Hosting Capacity on LV Secondary Circuits with the Gridco System empower TM Solution

Advanced Data Management and Analytics for Automated Demand Response (ADR) based on NoSQL

PV GRID. Efficient Grid Integration of European PV: Project Results. Jörg Mayer, Managing Director German Solar Industry Association (BSW-Solar)

Modeling of PV Based Distributed Generator Systems with Diverse Load Patterns

General Validation Test Program for Wind Power Plants Connected to the Hydro-Québec Transmission System

RENEWABLE ENERGY INFEED R Herman, M Malengret and CT Gaunt University of Cape Town, South Africa

What Matters for Successful Integration of Distributed Generation

European Distribution System Operators for Smart Grids. Flexibility: The role of DSOs in tomorrow s electricity market

Overview and Comparison of Demand Response Programs in North American Electricity Markets

European Distribution System Operators for Smart Grids. Position paper on Electric Vehicles Charging Infrastructure

Enel s experience on smart grids

3-PHASE LOW VOLTAGE NETWORK LOAD BALANCER : A COST EFFECTIVE SOLUTION TO LINE VOLTAGE VARIATIONS

How To Integrate Renewable Energy With Smart Grids

Investor day. November 17, Energy business Michel Crochon Executive Vice President

D 6.4 and D7.4 Draft topics of EEGI Implementation Plan Revision: Definitive

ESB Networks Response. ERGEG Consultation. Voltage Quality Regulation in Europe

A Network Reference Model for Distribution Regulation and Planning: the Spanish Case

A Reactive Tabu Search for Service Restoration in Electric Power Distribution Systems

Next Generation Distribution Management Systems (DMS) and Distributed Energy Resource Management Systems (DERMS)

SmartGrids SRA Summary of Priorities for SmartGrids Research Topics

Demand Response Management System ABB Smart Grid solution for demand response programs, distributed energy management and commercial operations

Course notes for EE394V Restructured Electricity Markets: Locational Marginal Pricing

Study to Determine the Limit of Integrating Intermittent Renewable (wind and solar) Resources onto Pakistan's National Grid

GENe Software Suite. GENe-at-a-glance. GE Energy Digital Energy

Executive summary 2. 1 Introduction 4

Advanced Network Management Systems: A Risk-Based AC OPF Approach

ICT Architecture for an Integrated Distribution Network Monitoring

ECCO International, Inc. 268 Bush Street, Suite 3633 San Francisco, CA 94104

Phase Balancing of Distribution Systems Using a Heuristic Search Approach

How can the Future Internet enable Smart Energy?

7. Reactive energy compensation

Modeling Simulation Technology Research for Distribution Network Planning *

Calculation of Voltage Sag Indices for Distribution Networks

SURVEY OF HARMONIC DISTORTION IN LV AND MV NETWORKS: RESULTS AND CORRECTIVE STRATEGIES

23/11/2015 TASK 4 REPORT - TECHNICAL ANALYSIS. Use of Task 4 analyses. Introduction and approach - Task 4 objective

A Fuzzy Based Solution for Improving Power Quality in Electric Railway Networks

How To Grow A Global Power Company

Lead Beneficiary: INESC Porto

Operational Security Network Code

PRACTICAL DATA MINING IN A LARGE UTILITY COMPANY

VOLATILITY AND DEVIATION OF DISTRIBUTED SOLAR

A Novel QoS Framework Based on Admission Control and Self-Adaptive Bandwidth Reconfiguration

Impact of electric vehicles on the IEEE 34 node distribution infrastructure

2Azerbaijan Shahid Madani University. This paper is extracted from the M.Sc. Thesis

EURELECTRIC presentation. EURELECTRIC contribution to a reference model for European capacity markets. DG COMP workshop 30 June 2015

Development of a Conceptual Reference Model for Micro Energy Grid

Agenda do Mini-Curso. Sérgio Yoshio Fujii. Ethan Boardman.

ARTIFICIAL NEURAL NETWORKS FOR ADAPTIVE MANAGEMENT TRAFFIC LIGHT OBJECTS AT THE INTERSECTION

White Paper. Convergence of Information and Operation Technologies (IT & OT) to Build a Successful Smart Grid

PROGRAM. Giga Sponsors. Supported by. Media Partner. Organizer

Content. Research highlights and demonstrations

Grid Edge Control Extracting Value from the Distribution System

Synchronized real time data: a new foundation for the Electric Power Grid.

itesla Project Innovative Tools for Electrical System Security within Large Areas

Optimal Power Flow Analysis of Energy Storage for Congestion Relief, Emissions Reduction, and Cost Savings

Bogdan Vesovic Siemens Smart Grid Solutions, Minneapolis, USA

PSS E. High-Performance Transmission Planning Application for the Power Industry. Answers for energy.

Realization of control center HMIs by using IEC and CIM data bases for communication and data handling

PHASOR MEASUREMENT UNIT (PMU) AKANKSHA PACHPINDE

ALSTOM Energy Management Business. Challenges related to Smart Energy Eco Systems

Chapter 5. System security and ancillary services

METER DATA MANAGEMENT FOR THE SMARTER GRID AND FUTURE ELECTRONIC ENERGY MARKETPLACES

Overcurrent Protection for the IEEE 34 Node Radial Test Feeder

C I R E D 23 rd International Conference on Electricity Distribution Lyon, June 2015

Coordinated and Optimized Control of Distributed Generation Integration

An Active Packet can be classified as

Context: significant penetration of DG = increased risks for system security

Institute of Power Engineering

Power-monitoring unit PowerLogic System. Technical data sheet. Power Meter Series 800

DESIGN FOR MONITORING AND OPTIMIZATION OF POWER DEMAND FOR WIRELESSLY COMMUNICATING ELECTRIC LOADS

AMI and DA Convergence: Enabling Energy Savings through Voltage Conservation

Digital Metering: a key enabling factor to foster RES development

EWE's approach to double grid connection capacity by using DERflexibility

Voltage Control on LV Distribution Network: Local Regulation Strategies for DG Exploitation

Experimental Verification of Advanced Voltage Control for Penetration of PV in Distribution System with IT Sectionalizing Switches

ALSTOM Grid s solution

SIMULATING HYBRID ENERGY GRIDS IN SMART CITIES FOCUS ON ELECTRIC ENERGY SYSTEMS

Field experience and best practices in managing MW scale Li-ion energy storage systems coupled to large wind and solar plants

Transcription:

LV4MV: A CONCEPT FOR OPTIMAL POWER FLOW MANAGEMENT IN DISTRIBUTION GRIDS, USING DER FLEXIBILITY Emmanuelle VANET Gaspard LEBEL Raphaël CAIRE G2Elab France G2Elab France G2Elab - France emmanuelle.vanet@g2elab.greble-inp.fr gaspard.lebel@g2elab.greble-inp.fr raphael.caire@g2elab.greble-inp.fr Nouredine HADJSAID Stéphane BEDIOU Alain GLATIGNY G2Elab France Schneider Electric France Schneider Electric - France uredine.hadjsaid@g2elab.greble-inp.fr stephane.bediou@schneider-electric.com alain.glatigny@schneider-electric.com ABSTRACT This paper introduces a concept for optimal power flow management in distribution grids, using DER flexibility. The LV4MV concept leads on a release of the admissible MV voltage range, in order to increase the number of suitable DER available dispatches at the LV level, willing at the end to reduce the constraint management costs of both MV and LV levels and to encourage the penetration of DER down to the lowest level of the grid. It aims to help the DSO to benefit of the available LV flexibility to solve voltage constraints on the both MV and LV levels. The method to identify the MV admissible voltage range for des where only LV network is connected is presented, taking into account the downstream LV flexibilities, if available. Test results concerning a LV unbalanced distribution system are then displayed, for different scenarii of LV flexibility. Finally, developed at the European scale within DREAM FP7 project, LV4MV implementation requirements and architecture are discussed. INTRODUCTION One step toward LV network automation The large penetration of Decentralized Generation (DG) among the distribution grids engenders, among other impacts, a modification of the voltage profiles on both the Low and Medium Voltage levels (LV & MV levels) [1]. To cope with these resulting voltage deviations, the Distribution System Operators (DSO) can either use network reinforcement, or voltage regulation through management of local active and reactive powers of the Distributed Energy Resources (DER), including DG, storage and controllable loads connected to both MV and LV levels. While the automation of the distribution grids focused mainly on the MV grids [2] [4], this new challenge of voltage deviation should encourage the DSOs to think at a partial automation of the LV grids. Short-term local flexibility offers With the growth of automation in distribution grids, networks are becoming more and more active [5]-[6], meaning that active and reactive powers management is possible via controllable sources and loads in order to solve constraints in the grid. This is possible through the activation of end users flexibility at both demand and supply side. On an individual level, flexibility is the modification of generation injection and/or consumption patterns in reaction to an external signal (price signal or activation) in order to provide a service within the energy system. The parameters used to characterize flexibility include the amount of power modulation, the duration, the rate of change, the response time and the location [7]. The day-ahead and intraday processes lead to an optimal plan of energy balance and to the kwledge of available flexibility offers which can be used for constraint management. Thanks to a decentralized approach, after the clearing of the both wholesale and balancing markets, distribution networks remaining flexibility services could allow the DSO to tackle network constraints up to the time of operation, maintaining reliability and quality of service and encouraging the increase of the rate of penetration of DER until the lowest levels of the grid. The LV4MV, a concept for optimal power flow management in distribution grids using DER flexibility, has been thought as a pre-operational service for DSOs, permitting operational planning. Proposed solution The LV4MV concept leads on a release of the admissible MV voltage range, to increase the number of suitable DER available dispatches at the LV level, willing at the end, to reduce the constraint management costs of the both MV and LV levels. LV4MV aims to help the DSO to benefit of the LV flexibility available on a given LV network to cope with voltage deviations appearing on both MV and LV levels. According to EN 50160, it is recommended that, under rmal conditions, the variations range of the r.m.s. magnitude at the supply terminals for LV and MV levels should be Un ±10% for 95% of a week, where Un is the minal voltage [8]. In practice the r.m.s. value could be determined over a fixed interval of 20 milliseconds and the basic measurement could be made by determining the average of these values over a period of 10 minutes. While EN 50160 gives general limits for public supply networks, various European countries commit to additional rules governing supply conditions. CIRED 2015 1/5

Until w in Europe, based on a fit-and-forget approach, the DSO currently guaranties voltage margins in MV level in order to ensure the operation in LV level. For example in France, the voltage at each MV network connection point should be within the range of Un ±5% [9]. This might lead to suboptimal operation of the system network, adding operational constraints for the DSO. The LV4MV concept has been designed to release these constraints, taking advantages of active distribution networks. CONCEPT Identification of the MV admissible voltage range interval The LV4MV concept aims to dissociate the MV and the LV levels via the management of the LV DER available dispatches. The objective of LV4MV is to determine the constraint-dependent admissible voltage range for each MV de where only LV network is connected, considering the downstream LV flexibilities, if available. Let [, ;, ] the MV admissible voltage range interval at the de of the considered MV distribution network, assuring that all LV des fed by this connection point are respecting the LV admissible voltage constraints. At a given time interval, for each MV de where only LV network is connected and MV consumer, it is possible to determine several MV admissible voltage range intervals, depending on the state of activation of downstream LV flexibilities. Hence, a table including the different voltage range intervals and their corresponding price of LV flexibilities activation can be constructed for each considered MV de. Therefore, instead considering a fix MV voltage range for all the MV des, it enables to increase the degree of freedom of those where only LV networks are connected: at these des, the voltage at the MV/LV transformer might be outside of the MV admissible voltage range commonly used in network planning, while all LV des fed by this connection point are respecting the LV admissible voltage constraints. Optimization of the MV grid operation with more precise voltage constraints Thanks to this increase of degree of freedom of the MV des, it will be easier to optimize the overall MV grid operation. This global ecomic and technical optimization should take into account the OLTC position of the HV/MV transformer but also the cost of activation of MV and LV flexibilities. The acquired precision on the voltage range constraints will help the DSO to plan and operate the network in a more accurate way and to realize benefits on flexibilities activation. ALGORITHM In the present article, the voltage variations range are respecting grid codes: the MV end users voltage variations range is set at Uc ±5% and the LV voltage variations range is set at [Un -10%; Un +10%]. Figure 1 presents the flowchart of the initialization step of the LV4MV, performed for each MV de where only LV network is connected, and for each dispatch of downstream LV flexibilities. Vref is the current standard voltage at the secondary part of the MV/LV transformer under rmal conditions. It is stored in the variable V0 for the initialization of the algorithm. After performing a first unbalanced three-phased loadflow on the downstream LV network, three situations might appear: case 1 corresponds to a situation where voltage at each LV de is included in the LV voltage variations range, whereas cases 2 and 3 correspond respectively to a LV under-voltage situation and to a LV over-voltage situation. Figure 1- Initialization flow-chart of LV4MV The next step consists in the computation of the MV admissible voltage range interval. The algorithm corresponding to case 1 is presented in figure 2. Cases 2 and 3 are t presented here due to high similarities. The algorithm consists in computing the admissible voltage limits and at the secondary part of the MV/LV transformer, assuring that all downstream LV connection point are respecting the LV admissible voltage range. In order to determine the voltage, the first step is the computation of the minimal voltage deviation between the highest voltage of the LV network and the admissible LV maximal voltage. The second step consists in increasing the voltage at the secondary part of the MV/LV transformer of value and to test thanks to a unbalanced three-phased load flow computation, if all the LV connected points are still respecting the LV admissible voltage range. This procedure is repeated until a detection of LV voltage deviation or a critical threshold of. The determination of is roughly based on the same sequence. CIRED 2015 2/5

The computation of the applied offset voltage corresponding to the deviation between the nearest voltage of the LV network and the LV voltage limit permits to gain CPU time doing less load flow computations than when performing iterations based on small step increases. Case 1 Vadm,min =V0 Compute V min Vmin > s1 Start V min = V min /2 V 0 = V 0 - V min /2 V max = V max /2 Perform a triphased load flow Current congestions in the feeders Vi >VLVmin End Vadm,min =V0 Compute V max Vmax > s2 V 0 = V 0 + V max Perform a triphased load flow Current congestions in the feeders Vi >VLVmax Figure 2 LV4MV flowchart corresponding to case 1 SIMULATION RESULTS Test Network The LV4MV is applied at a random MV de, which would be only connected to a real three-phase unbalanced 12-des LV grid, composed of one feeder. No MV end user is connected to this de. Figure 3 shows the considered LV distribution system. Table 1 presents the lines characteristics of the considered LV network. Figure 3 - Considered unbalanced 12-des LV network End Table 1 - Mutual and Self impedances of the considered network (mω) Z(L1) Z(L2) Z(L3) Z(L4) Z(L5) Z(L6) Z(L7) Z(L8) Z(L9) Z(L10) Z(L11) 0,7 + 1,2 i 0,1 + 1,0 i 0,1 + 1,0 i 0,1 + 1,0 i 0,1 + 1,0 i 0,7 + 1,2 i 0,1 + 1,0 i 0,1 + 1,0 i 0,1 + 1,0 i 0,1 + 1,0 i 0,7 + 1,2 i 0,1 + 1,0 i 0,1 + 1,0 i 0,1 + 1,0 i 0,1 + 1,0 i 0,9 + 1,2 i 10,5 + 18,6 i 1,2 + 16,7 i 1,2 + 16,7 i 1,2 + 16,7 i 1,2 + 16,7 i 10,5 + 18,6 i 1,2 + 16,7 i 1,2 + 16,7 i 1,2 + 16,7 i 1,2 + 16,7 i 10,5 + 18,6 i 1,2 + 16,7 i 1,2 + 16,7 i 1,2 + 16,7 i 1,2 + 16,7 i 15,0 + 18,9 i 42,8 + 75,5 i 4,7 + 67,7 i 4,7 + 67,7 i 4,7 + 67,7 i 4,7 + 67,7 i 42,8 + 75,5 i 4,7 + 67,7 i 4,7 + 67,7 i 4,7 + 67,7 i 4,7 + 67,7 i 42,8 + 75,5 i 4,7 + 67,7 i 4,7 + 67,7 i 4,7 + 67,7 i 4,7 + 67,7 i 60,9 + 76,9 i 21,7 + 38,3 i 2,4 + 34,4 i 2,4 + 34,4 i 2,4 + 34,4 i 2,4 + 34,4 i 21,7 + 38,3 i 2,4 + 34,4 i 2,4 + 34,4 i 2,4 + 34,4 i 2,4 + 34,4 i 21,7 + 38,3 i 2,4 + 34,4 i 2,4 + 34,4 i 2,4 + 34,4 i 2,4 + 34,4 i 30,9 + 39,1 i 1,3 + 2,3 i 0,1 + 2,1 i 0,1 + 2,1 i 0,1 + 2,1 i 0,1 + 2,1 i 1,3 + 2,3 i 0,1 + 2,1 i 0,1 + 2,1 i 0,1 + 2,1 i 0,1 + 2,1 i 1,3 + 2,3 i 0,1 + 2,1 i 0,1 + 2,1 i 0,1 + 2,1 i 0,1 + 2,1 i 1,9 + 2,4 i 43,5 + 76,7 i 4,8 + 68,7 i 4,8 + 68,7 i 4,8 + 68,7 i 4,8 + 68,7 i 43,5 + 76,7 i 4,8 + 68,7 i 4,8 + 68,7 i 4,8 + 68,7 i 4,8 + 68,7 i 43,5 + 76,7 i 4,8 + 68,7 i 4,8 + 68,7 i 4,8 + 68,7 i 4,8 + 68,7 i 61,8 + 78,1 i 18,1 + 16,9 i 1,0 + 14,6 i 1,0 + 14,6 i 1,0 + 14,6 i 1,0 + 14,6 i 18,1 + 16,9 i 1,0 + 14,6 i 1,0 + 14,6 i 1,0 + 14,6 i 1,0 + 14,6 i 18,1 + 16,9 i 1,0 + 14,6 i 1,0 + 14,6 i 1,0 + 14,6 i 1,0 + 14,6 i 13,1 + 16,6 i 36,9 + 65,0 i 4,1 + 58,3 i 4,1 + 58,3 i 4,1 + 58,3 i 4,1 + 58,3 i 36,9 + 65,0 i 4,1 + 58,3 i 4,1 + 58,3 i 4,1 + 58,3 i 4,1 + 58,3 i 36,9 + 65,0 i 4,1 + 58,3 i 4,1 + 58,3 i 4,1 + 58,3 i 4,1 + 58,3 i 52,5 + 66,3 i 73,0 + 128,7 i 8,1 + 115,3 i 8,1 + 115,3 i 8,1 + 115,3 i 8,1 + 115,3 i 73,0 + 128,7 i 8,1 + 115,3 i 8,1 + 115,3 i 8,1 + 115,3 i 8,1 + 115,3 i 73,0 + 128,7 i 8,1 + 115,3 i 8,1 + 115,3 i 8,1 + 115,3 i 8,1 + 115,3 i 103,8 + 131,2 i 26,5 + 46,7 i 2,9 + 41,9 i 2,9 + 41,9 i 2,9 + 41,9 i 2,9 + 41,9 i 26,5 + 46,7 i 2,9 + 41,9 i 2,9 + 41,9 i 2,9 + 41,9 i 2,9 + 41,9 i 26,5 + 46,7 i 2,9 + 41,9 i 2,9 + 41,9 i 2,9 + 41,9 i 2,9 + 41,9 i 37,7 + 47,6 i 26,0 + 45,8 i 2,9 + 41,0 i 2,9 + 41,0 i 2,9 + 41,0 i 2,9 + 41,0 i 26,0 + 45,8 i 2,9 + 41,0 i 2,9 + 41,0 i 2,9 + 41,0 i 2,9 + 41,0 i 26,0 + 45,8 i 2,9 + 41,0 i 2,9 + 41,0 i 2,9 + 41,0 i 2,9 + 41,0 i 36,9 + 46,6 i Load modelling The static characteristics of the load can be classified into constant power, constant current and constant impedance load, depending on the power relation to the voltage [10]. Considering the network data, the constant power characteristic is applied for load modelling. Table 2 Phases repartition of residential customer profile characteristic at the different des Node Phase 1 Phase 2 Phase 3 4 1 1 1 5 5 4 1 6 2 2 2 8 3 3 3 10 1 0 0 12 3 3 4 In order to simulate different power patterns, the French residential customer profile characteristic for a standard weather condition of the 3rd Thursday of January [11]- [12] has been applied several times on the different CIRED 2015 3/5

phases at some des of the considered LV network (table 2). Assuming that the mean annual consumption per house in 2013 in France is 4113kWh [13], the active power withdraw value of a typical end user is computed for each half-hour. The power factor for residential customer has been set to 0.95. LV4MV results The LV4MV algorithm has been performed for different activations of LV flexibilities in the 12-des network, at different time of the day. The different cases of available flexibilities are listed below: Scenario 0. Initial case: without activating any flexibility Scenario A. 30% decrease of consumed power on phase 1 at de 5 Scenario B. 25% decrease of consumed power on phase 3 at de 12 Scenario C. 17% decrease of consumed power on each phase at de 8 Scenario D. 50% decrease of consumed power on phase 1 at de 10 Table 3 presents the results of the LV4MV algorithm applied for different scenarii of activation of LV flexibilities, at 3.00 pm and 7.30 pm respectively. Table 3 LV4MV results (3.00 pm, 7.30 pm) Time 3.00 pm 7.30 pm Scenario Vadm_min Vadm_max Vadm_min Vadm_max 0 0,94 1,1 0,95 1,1 A 0,94 1,1 0,95 1,1 B 0,93 1,1 0,95 1,1 C 0,94 1,1 0,95 1,1 D 0,94 1,1 0,95 1,1 A, B 0,93 1,1 0,95 1,1 A, C 0,94 1,1 0,95 1,1 A, D 0,94 1,1 0,95 1,1 B, C 0,93 1,1 0,95 1,1 B, D 0,93 1,1 0,94 1,1 C, D 0,94 1,1 0,95 1,1 A, B, C 0,93 1,1 0,95 1,1 A, C, D 0,94 1,1 0,95 1,1 B, C, D 0,93 1,1 0,94 1,1 A, B, C, D 0,93 1,1 0,94 1,1 Depending on the load of the network and the activated flexibilities, it is possible to find the MV admissible voltage range at each MV de, which guarantee that all downstream LV connection points are respecting LV voltage constraints. In some of these scenarii, table 3 shows that MV admissible voltage ranges are larger than the one commonly assumed for network planning. Others situations might lead to stricter MV voltage ranges, the benefit of this algorithm would stay to get more information for real-time network operation. Figure 4 shows the three-phases voltage profile along the downstream LV network at 3.00 pm, when scenarii A,B,C and D are applied, and with a reference voltage V0=0.95pu at the secondary part of the MV/LV transformer. Voltage 0.95 0.94 0.93 0.92 0.91 0.9 X: 1 Y: 0.95 2 4 6 8 10 12 Nodes Figure 4 Three-phases voltage profile along the LV network (3.00 pm, scenario A,B,C,D, V0=0.95pu) According to LV4MV computations, the voltage at the secondary part of the MV/LV transformer can be set down to 0.93pu, ensuring that LV voltage constraints will be respected for the whole downstream LV network. However for V0=0.92pu, LV voltage constraints are t respected anymore. Influence on voltage profile of a LV flexibility activation In term of power decrease need, an end user s flexibility activation corresponds either to a consumption reduction of a controllable load, or to a production increase assuming that some reserves have been pre-contracted. Mathematically, it can be represented as a single-phase consumed power reduction at one de. Figure 5 illustrates the scenarii A and D applied on phase 1 simultaneously at 7.30 pm, which is a 30% decrease of consumption at de 5, coupled with a 50% decrease of consumption at de 10. Voltage 1.01 1.005 1 0.995 0.99 0.985 0.98 0.975 0.97 0.965 0.96 LV minimum admissible voltage Voltage profile along the feeder Voltage profile along the feeder Without LV flexibility activation With LV flexibility activation at des 5 and 10 0.004 pu 2 4 6 8 10 12 Nodes Phase 1 Phase 2 Phase 3 0.0327 pu Figure 5 Voltage profile on phase 1 along the feeder at 7:30 pm with and without LV flexibilities activation Figure 5 shows that a LV flexibility activation could have a large impact on voltage profile. This impact on voltage profile is larger when the network is highly loaded and CIRED 2015 4/5

when the flexibility is located at the end of a line. The activation of LV flexibilities will permit to increase the determined MV admissible voltage range intervals. The LV4MV is performed for each LV flexibilities offer and combinations, which are tested in merit order. Therefore, the logic is combinatorial and might t be optimal. A method based on the sensitivity analysis of local flexibilities for voltage regulation [14] completes this algorithm step, in order t to test every combination of available LV flexibilities but to sort them by ecomic and technical impacts. It will be then possible to compare the cost of LV flexibilities activation with the benefit in operation of the entire MV grid. IMPLEMENTATION Requirements Technically, LV4MV requires the processing of independent load-flow calculations for LV networks and then Optimal Power Flow (OPF) calculations for MV networks. It requires also the monitoring of the voltage values among the LV feeders thanks to Customer Energy Management System (CEMS) or Advanced Metering Infrastructures (AMI), the forecast of n-controllable loads and already trade DER programs and finally the gathering of DER bids, willing to support the grid for constraints management. Implementation architecture Developed within the FP7 DREAM project [15], this mechanism rests on a distributed mode of network operation, with an architecture where functionalities are distributed among the whole system network. The distribution of intelligence permits a repartition of computations among the network and avoids the reporting of all the data at a centralized level. This permits also to better consider all the available DER flexibilities down to the lowest level of the grid. Hence, assuming that RTU are installed in primary and secondary substations, the LV4MV algorithm could be embedded and performed at every secondary substation in order to facilitate the optimization and the operation of the entire considered network, completed in primary substations. Figure 6 presents the infrastructure to be deployed on targeted MV and LV flexible networks in order to benefit of the LV4MV concept. Figure 6 Infrastructure to be deployed on MV and LV flexible networks to benefit of the LV4MV concept CONCLUSION AND PERSPECTIVES A method to identify the MV admissible voltage range interval of des where only LV network is connected, taking into account the downstream LV flexibilities if available, has been presented and tested. Results shows that MV admissible voltage ranges might be larger or stricter than the one commonly assumed for network planning, depending on downstream LV flexibilities. The interaction between the LV4MV algorithm performed at each secondary substation and the optimization of the overall MV grid operation will be discussed in future work. This global ecomic and technical optimization will take into account the OLTC position of the HV/MV transformer but also the cost and performance of activation of MV and LV flexibilities, determined thanks to a method based on the voltage sensitivity analysis of local flexibilities. REFERENCES [1] T. Tran-Quoc, et Al., Technical impacts of small distributed generation units on LV networks, in IEEE Power Engineering Society General Meeting, 2003, 2003, vol. 4, p. -2464 Vol. 4. [2] R. E. Brown, Impact of Smart Grid on distribution system design, in 2008 IEEE Power and Energy Society General Meeting - Conversion and Delivery of Electrical Energy in the 21st Century, 2008, pp. 1 4. [3] A. S. Bouhouras et Al., Selective Automation Upgrade in Distribution Networks Towards a Smarter Grid, IEEE Trans. Smart Grid, vol. 1,. 3, pp. 278 285, 2010. [4] D. S. Popovic et Al., The optimal automation level of medium voltage distribution networks, Int. J. Electr. Power Energy Syst., vol. 33,. 3, pp. 430 438, 2011 [5] A. Korbik, S. D. J. Mcarthur, G. W. Ault, G. M. Burt, and J. R. Mcdonald, Enabling active distribution networks through decentralised automous network management, 18th International Conference and Exhibition on Electricity Distribution, 2005. CIRED 2005, 2005 [6] L. F. Ochoa, C. J. Dent, and G. P. Harrison, Distribution Network Capacity Assessment: Variable DG and Active Networks, IEEE Trans. Power Syst., vol. 25,. 1, pp. 87 95, 2010. [7] Eurelectric, Flexibility and aggregation: requirements for their interaction in the market, Technical report, 2014. [8] Eurelectric, Application to the European Standard EN50160 on Voltage characteristics of electricity supplied by public distribution systems, Jul. 1995. [9] ERDF, Description physique du réseau public, ERDF-NOI- RES_07E, 2008. [10] Load representation for dynamic performance analysis [of power systems], IEEE Trans. Power Syst., vol. 8,. 2, pp. 472 482, mai 1993. [11] ERDF, Profilage et Profils, 2013. [Online]. Available: http://www.erdf.fr/erdf_fournisseurs_electricite_responsables _Equilibre_Profils. [12] RTE, Règles relatives à la Programmation, au Mécanisme d Ajustement et au dispositif de Responsable d Equilibre, Section 2, chapitre F, 2013. [13] Commission de Régulation de l Energie, Observatoire des marchés de l électricité, du gaz et du CO2, 2eme trimestre 2013, 2013. [14] E. Vanet, R. Caire, and N. Hadjsaid, Sensitivity analysis of local flexibilities for voltage regulation in unbalanced LV distribution system, presented at the PowerTech, 2015 IEEE Eindhoven, in press. [15] DREAM project, DREAM reference object model and dictionary, Technical report D5.1, 2014. CIRED 2015 5/5