INCREASING THE PENETRATION OF RENEWABLE ENERGY SOURCES · 6 increase synthesis document list of figures figure 3.1: main increase results, their interdependencies and timing 15 figure

Increasing the penetration of renewable energy sources in the distribution grid

SYNTHESIS DOCUMENT January 2017

INCREASE was funded by the 7th Framework Programme of the EC

www.project-increase.eu

http://www.project-increase.eu/

Increasing the penetration of renewable energy sources in the distribution grid by developing control strategies and using ancillary services

SYNTHESIS DOCUMENT

January 2017

AUTHORS

Andreas Tuerk (JR), Claudia Fruhmann (JR), Bart Meersman (UGent) in cooperation with all project partners.

PROJECT

This project has received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement no. 608998.

More information about the project: www.project-increase.eu

ACKNOWLEDGEMENT & DISCLAIMER

Neither the European Commission nor any person acting on behalf of the Commission is responsible for the use which might be made of the following information. The views expressed in this publication are the sole responsibility of the authors and do not necessarily reflect the views of the European Commission.


CONSORTIUM

Aristotle University of

Thessaloniki

Eandis

Elektro Gorenjska

Elia

Joanneum Research

Korona

Liander

Mastervolt

Ghent University

Energie Netze Steiermark

Technische Universiteit Eindhoven

University of Ljubljana

ILPRA

5 INCREASE SYNTHESIS DOCUMENT

Contents

1 EXECUTIVE SUMMARY 9

2 INTRODUCTION AND PROJECT DETAILS 11

2.1 INTRODUCTION AND PROJECT AIMS 12

2.2 PROJECT DETAILS 12

3 OVERVIEW OF MAIN RESULTS 14

4 INCREASE TECHNICAL SOLUTIONS 17

4.1 INCREASE LOCAL CONTROL 18

4.2 INCREASE OVERLAYING CONTROL 20

4.3 INCREASE SCHEDULING CONTROL 24

4.4 THE INCREASE SIMULATION PLATFORM 28

5 INCREASE LAB TEST AND FIELD TRIALS 32

5.1 FIELD TRIALS: GENERAL OVERVIEW AND KEY FACTS 33

5.2 THE BELGIAN FIELD TRIAL 34

5.3 THE SLOVENIAN FIELD TRIAL 35

5.4 DUTCH FIELD TRIAL 39

5.5 AUSTRIAN FIELD TRIAL 41

6 BUSINESS CASES AND NEEDED FRAMEWORK CONDITIONS 45

6.1 STAKEHOLDERS, SERVICES AND BUSINESS MODELS 47

6.2 INCREASE BUSINESS MODEL FOR EVALUATION 47

6.3 INCREASE EVALUATION SCENARIOS AND SENSITIVITY ANALYSIS 49

6.4 THE NEEDED REGULATORY FRAMEWORK 51

7 IMPACT CREATION AND EXPLOITATION 53

7.1 OUTREACH AND DISSEMINATION 54

7.2 EXPLOITATION 55


List of figures

FIGURE 3.1: MAIN INCREASE RESULTS, THEIR INTERDEPENDENCIES AND TIMING 15

FIGURE 4.1: DROOP CURVE OF THE VBD CONTROL 19

FIGURE 4.2: PROBABILITY OF HAVING A VOLTAGE AT A NODE IN 3 DIFFERENT SCENARIOS 20

FIGURE 4.3: LOW VOLTAGE TEST FEEDER AND ICT INFRASTRUCTURE (N=7,M=5) 21

FIGURE 4.4: POWER INJECTION BY INVERTERS FOR DIFFERENT CONTROLS – SNAP SHOT 22

FIGURE 4.5: POWER INJECTION BY INVERTERS – TIME SERIES SIMULATION 22

FIGURE 4.6: THE COORDINATION OF OLTC AND DRES 23

FIGURE 4.7: GENERAL OVERVIEW OF SCHEDULING CONTROL LAYERS 24

FIGURE 4.8: LOAD PROFILE WITH ADDED OPERATION MODES OF DR UNITS (HEAT PUMPS) 25

FIGURE 4.9: GREEN (PV) ENERGY INFEED DIFFERENCES BETWEEN ECONOMIC AND ENERGY OPTIMIZATION IN SUMMER – SIMULATIONS BY SCENARIOS (PV PENETRATION LEVELS) 27

FIGURE 4.10: THE INCREASE SIMULATION PLATFORM 29

FIGURE 5.1: SUBSTATION WITH OLTC (LEFT) AND CABLE CABINETS WITH MEASUREMENTS AND COMMUNICATIONS (RIGHT) 34

FIGURE 5.2: BASIC SET-UP OF THE SLOVENIAN FIELD TRIAL 36

FIGURE 5.3: STATISTICAL ANALYSIS OF AVERAGE PHASE VOLTAGES 37

FIGURE 5.4: OPERATION OF EDC AND FPS 39

FIGURE 5.5: CONGESTION MANAGEMENT - TOTAL LOADING OF TRANSFORMER 40

FIGURE 5.6: CONGESTION MANAGEMENT – RELATIVE CURTAILMENT PER INVERTER 40

FIGURE 5.7: BASIC SET-UP OF THE AUSTRIAN FIELD TRIAL 42

FIGURE 5.8: INITIAL (LEFT) AND FINAL (RIGHT) SITE CONFIGURATION 42

FIGURE 5.9: REDUCTION OF THE NEUTRAL CURRENT IN CASE OF UNBALANCE CONTROL ACTIVATION 43

FIGURE 5.10: IMPACT OF THE UNBALANCE CONTROL ON MAXIMAL PHASE VOLTAGE DIFFERENCES AND NEUTRAL CURRENT 43

FIGURE 6.1: THE PROCESS OF EVALUATION OF THE INCREASE SOLUTIONS 47

FIGURE 6.2: THE CHOSEN AGGREGATOR BUSINESS MODEL FOR EVALUATION 48

FIGURE 6.3: ECONOMIC OPTIMIZATION WITH 10,000 AND 1,000 UNITS, FOR DIFFERENT PENETRATION SCENARIOS FOR AN EU GRID 50

FIGURE 7.1: INCREASE SUMMER SCHOOL IN LJUBLJANA 54

FIGURE 7.2: TECHNICAL WORKSHOP IN THESSALONIKI IN SEPTEMBER 2015 54


List of tables

TABLE 4.1: TYPES OF TRAFFIC LIGHT SYSTEMS 26

TABLE 4.2: DR DEVELOPMENT SCENARIOS 26

TABLE 6.1: NUMBER OF EMPLOYED PERSONNEL FOR DIFFERENT SCENARIOS AND POOL SIZE, WITH AMOUNT OF AGGREGATED UNITS 49

TABLE 6.2: SENSITIVITY ANALYSIS PARAMETERS 49

TABLE 7.1: OVERVIEW OF EXPLOITABLE KNOWLEDGE 55


Abbreviations

AS ANCILLARY SE RVICES PQ POWE R QUALITY

CM CONGE STION MANAGEMENT SC SCHEDULING CONTROL

DG DISTRIBUTED GENERATION SCA SCHEDULING CONTROL AGE NT

DMS DISTRIBUTION MANAGE MENT SYSTE M

TLS TRAFFIC L IGHT SYSTEM

DR DEMAN D RE SPON SE TSO TRANSMISSION SYSTE M OPE RATOR

DRE S DISTRIBUTED RENE WABLE ENERGY SOURCE S

UPS UNINTE RRUPTABLE POWER SUPPLY

DSO DISTRIBUTION SYSTE M OPE RATOR

VAT VALUE ANALYSIS TOOL

EDC EMULATE D DROOP CON TROL VBD VOLTAGE -BASED DROOP

FCR FRE QUE NCY CONTAINME NT RESE RVE

FEP FLE XIBLE ENE RGY PRODUCTS

FPS FAIR POWER SHARING

GUI GRAPHICAL USE R IN TERFACE

IL INTE RRUPTIBLE LOADS

JADE JAVA AGEN T DE VELOPE D ENVIRONMENT

KER KE Y EXPLOITABLE RE SULTS

KPI KE Y PE RFORMAN CE IN DICATOR

LDC L INE DROP COMPENSATION

LC LOCAL CON TROL

LV LOW VOLTAGE

MAS MULTI -AGE NT SYSTEM

MV MEDIUM VOLTAGE

OC OVE RLAYING CONTROL

OLTC ON-LOAD TAP CHARGER


1 EXECUTIVE SUMMARY

1 x

Executive summaryx


The significant rise in distributed renewable energy sources (DRES) has placed an enormous burden on the secure operation of the electrical grid. The massive increase of the intermittent DRES in low (LV) and medium (MV) voltage networks has led to a bidirectional power flow, which raises the urgent need for new operational and control strategies in order to maintain the ability of the system operator to provide the consumers a reliable supply of electricity at an acceptable power quality level and cost.

Technically, INCREASE has focused on how to manage renewable energy sources in LV and MV networks and to provide ancillary services (AS) (towards distribution system operators (DSOs), but also transmissions system operators (TSOs)), in particular voltage control and the provision of reserve. A simulation platform has been developed enabling an efficient development of the INCREASE solutions. It also proved to be an efficient tool for DSOs to study the impact new solutions have on the congestion challenges related to the penetration of renewables as well as to investigate the influence of DRES on their distribution network. Starting from the technical solutions, INCREASE studied the regulatory framework, grid code structure and ancillary market mechanisms, and proposed adjustments to the economic and regulatory framework to facilitate successful provisioning of AS that are necessary for the operation of the electricity grid, including flexible market products.

INCREASE proposed a three level approach for maximizing the integration of renewables. The first level only uses local parameters (voltage at the point of connection, exchanged power) for the control. This is a fast control that mitigates the voltage unbalance (at the LV network) and uses P/V droops to achieve soft curtailment to solve the overvoltage problem. The first level control ensures the reliability and stability of the system. The second level control results in an optimal system and aims to minimize the loss of renewable energy as well as ensuring a fair distribution of the possible curtailed energy. This second level control is achieved by a multi-agent aggregator concept and consists of fair power sharing, the coordination of On-Load Tap Charger (OLTC) control and PV inverters to solve (current and voltage) congestion. In order for the DSO to evolve from congestion manager to capacity manager, a service layer is developed. This service layer is the third level in the INCREASE approach. It solves a multi-objective optimization problem by combining and extending optimization strategies and results in flexible energy products to provide AS with them. The INCREASE solutions were validated (i) by lab tests, as well as (ii) in four field trials in the real-life operational distribution network of Energienetze Steiermark in Austria, of Eandis in Belgium, of Elektro Gorenjska in Slovenia, and of Liander in the Netherlands. Different parts of the INCREASE solutions were successfully tested in the different field trials. Based on the developed control strategies business models were developed, that may help to source flexibility from demand response (DR) and PV. With the help of economic evaluations possible revenues for different actors (e.g. aggregators) could be quantified. In addition the needed framework conditions (economic, regulatory) to enable the proposed business case were investigated.


2 INTRODUCTION AND PROJECT DETAILS

Introduction and project detailsx

2 x


2.1 INTRODUCTION AND PROJECT AIMS

To meet the European 20-20-20 targets, the share of renewable energy needs to be 20% of the electric energy demand in 2020. This ambitious objective can only be met if the number of distributed renewable energy sources (DRES) at the low voltage (LV) electric distribution grid will significantly increase and large wind or solar plant farms will be installed at the medium voltage (MV) level. The significant rise in distributed renewable energy sources is placing an enormous burden on the secure operation of the electrical grid, impacting as well the transmission (TSOs) as the distribution system operators (DSOs).

Five scientific and eight industrial partners were participating in the ambitious research project INCREASE ("Increasing the penetration of renewable energy sources in the distribution grid by developing control strategies and using ancillary services") which aimed to solve these problems with innovative three-phase grid-connected inverters and new operational and control strategies in order to maintain the ability of the system to provide the consumers with reliable supply of electricity at an acceptable power quality level.

The aim of the project was to develop control strategies, test them and assess options to provide ancillary services (AS) to markets. In addition to technical solutions focussing on integration and intelligent control of DRES, the INCREASE project also developed accompanying control strategies and investigated their value provided to various stakeholders. For this purpose, a Multi-Agent System (MAS) control strategy was devised with a hierarchical structure, featuring Local Control (LC), Overlaying Control (OC) and Scheduling Control (SC). INCREASE investigated the regulatory framework, grid code structure and ancillary market mechanisms, and proposed adjustments to facilitate successful provisioning of AS necessary for the electricity grid operation, including flexible market products. It aimed to enable DRES and loads to go beyond just exchanging power with the grid which enables the DSOs to evolve from congestion to capacity manager. This facilitates higher DRES penetration at reduced cost. The simulation platform developed in the scope of INCREASE enables the validation of the proposed solutions and provides a tool for the DSOs to investigate the DRES influence in their network. Not only by lab test at TU/e and at Lemcko (Ghent University), but also real field trials in Austria, Slovenia and the Netherlands validated the INCREASE solutions. Based on the technical solutions, economic assessments were made and business models were developed including the needed framework conditions (economic, regulatory), to enable their implementation.

2.2 PROJECT DETAILS

Duration and budget

INCREASE was initiated in September 2013. The project duration was 40 months and led by Ghent University. The total EU contribution was about 3.2 million.


More project details at: http://www.project-increase.eu/

The INCREASE consortium

INCREASE is oriented at delivering tools and solutions to DSOs. To ensure that these tools and solutions are well accepted and fit the needs of this target groups as much as possible, four DSOs from different regions in Europe were partner in the project, apart from several research institutes and industrial partners.

Austria Netherlands

Energie Netze Steiermark www.e-netze.at

JOANNEUM RESEARCH www.joanneum.at

Belgium

Eandis www.eandis.be

Elia www.elia.be

Ghent University www.ugent.be

Greece

Aristotle University of Thessaloniki www.auth.gr

ILPRA (Greece) http://ilpra.gr

Liander www.liander.nl

MASTERVOLT www.mastervolt.com

Technische Univesiteit Eindhoven www.tue.nl

Slovenia

Elektro Gorenjska www.elektro-gorenjska.si

KORONA www.korona.si

University of Ljubljana www.fe.uni-lj.si


http://www.e-netze.at/

http://www.joanneum.at/

http://www.eandis.be/

http://www.auth.gr/

http://www.liander.nl/

http://www.mastervolt.com/

http://www.elektro-gorenjska.si/


3 OVERVIEW OF MAIN RESULTS

Overview of main resultsx

2 x


INCREASE developed algorithms, tools and hardware, some of them resulting in exploitable products. This chapter aims to give a brief overview of the results, their interdependencies and the timeline of their creation.

The results of the project, shown in Figure 3.1, can be grouped into algorithms (in blue), hardware and software tools (in green), and in support analytical framework (in orange). The control algorithms (Local, Overlaying and Scheduling Algorithm) have been designed, developed, tested and demonstrated in the project in order to control the local PV generation in connection with the demand response units. Their activation and operation depends on the hardware solution, the advanced 3-phase inverter, and their testing and to some extent demonstration on the Integrated simulation platform, both developed within the INCREASE project and described below.

In order to enable the simulations, we have designed technical scenarios as a part of the support analytical framework. They rested on the Framework Matrix, which extracted all the relevant ambient variables of the energy markets and the electric networks in an explicit way and enabled their comparison and control in the simulations. The results of the simulations were used to analyze the quality of the proposed business models using the multicriteria Value Analysis Tool, described below.

Figure 3.1: Main INCREASE results, their interdependencies and timing

Overlaying Control

Algorithm

Local Control

Algorithm

Scheduling Control

Algorithm

1 6 18 12 24 30 36

Integrated simulation platform

3 phase inverter

Technical Scenarios Business Models Frame-work

Matrix VAT


INCREASE Inverter

The INCREASE inverters are advanced LV inverters aimed at small-scale distributed generation, ideal for rooftop PV plants. They are equipped with advanced INCREASE control algorithms, implementing smart control strategies that tackle the issues of voltage unbalance mitigation and voltage-based droop control. The inverters which are currently on the market do not have the ability to provide voltage unbalance mitigation services nor have a P/V droop capability. These features enable the inverter to remain connected to the grid for a longer period of time compared to the existing systems which results in a higher production of renewable energy. The INCREASE inverters could also be used in battery systems (large (> 50kWh) and small-scale (< 50 kWh)) as well in UPS systems.

Simulation Platform

The INCREASE simulation platform is a software tool for the analysis of smart grids. It performs simultaneous simulations of the power grid, the distributed intelligent controls and the communication system, using multiple layers. Its flexible, modular open source architecture allows the easy integration of models of any novel smart grid component, as well as of existing software packages of other vendors.

Value analysis tool

The INCREASE Value Analysis Tool (VAT) was created within the INCREASE project and calculates the aspects of control strategies necessary for conducting a comprehensive socio-economic assessment. By providing quantitative information for financial, environmental, energy generation- and energy security aspects of various INCREASE solutions, the VAT is designed to serve as a decision support tool for policy makers, potential investors in renewable technologies and for market actors. The cornerstone of the methodology is the “Framework Matrix” that comprises the salient context variables as a series of parameters – the key performance indicators (KPIs), which describe the different facets of the context of the INCREASE solutions in which the market actors operate (e.g. advantages and disadvantages in terms of costs, involved actors, size, efficiency, business performance, and regulatory boundaries). The VAT has different parameters of possible business models as input such as the revenue sharing between actors and can calculate the profits for different actors under different scenarios. Based on the results of economic calculation with the VAT, different possible business models for aggregators were derived in INCREASE for aggregating flexibility from demand response and from PV. With help of sensitivity analyses, the needed framework conditions were assessed to enable the proposed business models.


4 INCREASE TECHNICAL SOLUTIONS

INCREASE technical solutionsx

4 x


The problem

The significant rise in DRES has placed an enormous burden on the secure operation of the electrical grid. The massive increase of the intermittent DRES in LV and MV networks has led to a bidirectional power flow, which raises the urgent need for new operational and control strategies in order to maintain the ability of the system operator to provide the consumers a reliable supply of electricity at an acceptable power quality level and cost.

The INCREASE solutions

The technical solutions in INCREASE focused on the development of multi-level control strategies. INCREASE developed local, overlaying and scheduling control strategies. In addition a simulation platform was created to simulate control strategies in specific networks.

4.1 INCREASE LOCAL CONTROL The main objective of the INCREASE LC strategies are the continuous mitigation of overvoltage and voltage unbalance in LV distribution networks. The INCREASE LC strategies thus improve the power quality of the distribution grid. The control strategies were implemented in the inverter that connects solar panels to the grid. The inverter measures the voltage and power at his terminals and based on this reacts. There is no need for communication with external parties, the inverter works independently.

INCREASE local control consists of 2 strategies:

The voltage-based droop control

This strategy lowers the produced PV power in case the grid voltage rises so that disconnection of the solar installation is avoided.

The voltage unbalance mitigation control

The strategy distributes the produced PV power in the different phases so that less loaded phases receive more power than the higher loaded phases. This strategy can only be applied in three-phase inverters. One of the extra advantages of this control strategy is that the grid is also supported during voltage dips caused by grid faults.

In INCREASE, we proposed an active power (P/V) control, namely the voltage-based droop (VBD) control. The basic principle behind this control strategy is depicted in Figure 4.2. The inverter injects the available power into the grid until a certain voltage threshold is reached (e.g. vg,up = 248V). Then, the injected power is decreased compared to the available power. This is achieved by forcing the inverter to deviate from his maximum power point.


Figure 4.1: Droop curve of the VBD control

The voltage unbalance mitigation strategy is implemented in three-phase inverters. This control distributes the solar power between the three phases such that lower loaded phases receive more power compared to higher loaded phases.The voltage unbalance mitigation strategy is achieved by giving the inverter a resistive behavior towards the zero-sequence and negative-sequence voltage component. This results in the inverter injecting more current in the phase with lower voltage and higher current in the phase with a higher voltage.

Figure 4.3 depicts the probability of having a certain voltage at a node in a LV network in four different cases:

Classical on/off control – inverter switches off when 1.1pu is reached

3ph-DPC: INCREASE local control strategy

3ph-Sym: 3-phase connection of renewables with VBD control

Q-control: voltage control by means of reactive power control

INCREASE local control always results in the lowest grid voltage which means that there is more room to integrate more renewables in the LV distribution grid. The proposed solution does not use communication which is very beneficial for the implementation of the solution.


Figure 4.2: Probability of having a voltage at a node in 3 different scenarios

We can thus conclude that:

The INCREASE local control results in the most efficient use of the LV distribution grid

The INCREASE local control solves the on-off oscillation problem and leads to a better power quality

The INCREASE local control ensures a stable and reliable operation of the distribution grid

4.2 INCREASE OVERLAYING CONTROL The OC of INCREASE aims to mitigate over-voltage and over-loading problems in LV networks while providing an equal opportunity for all the customers. It has been developed as a complementary control mechanism to the local control for providing coordination means with other network controllers like PV inverters and the On-Load Tap Changer (OLTC).

The INCREASE OC consists of three main control modules with their dedicated control tasks:

The Fair Power Sharing (FPS) module aims to mitigate over-voltage problems by curtailing active power equally among PV inverters;

The Congestion Management (CM) module aims to relieve the congestion occurring at the MV/LV transformer;


The Coordination control module for OLTC and PV inverters aims to maximize renewable energy production.

The ICT-aided implementation of OC

In order to realize the aforementioned control modules in real-life, the integration of a cyber domain and a physical domain is required. Therefore an interface platform called middleware layer was developed in INCREASE. In addition, an agent-aggregator architecture was developed using a middleware later, which laid the foundation for the development of OC algorithms. It consists of two main software entities including an aggregator/grid agent and a DRES agent to deploy data acquisition and control functions in various environments, like Matlab/Simulink and Java run-time.

FAIR POWER SHARING (FPS)

The LC mitigates the overvoltage problem however it leads to unequal/unfair power curtailment among the inverters. Therefore, OC complements the LC and the FPS algorithm of OC redistributes the curtailed power uniformly/fairly among the inverters. The control algorithm was implemented using the inverters of DRES.

ACL messages

MV LV

250 kVA20/0.4 kV

1 3 nUtility grid

House load

House load

2

TCP/IP communication

, ,i i iinj g netP V P

R jX+ R jX+

House load

Simulink

JADE

1 2 M,i

ref OCP,i

ref OCP ,i

ref OCP

1 2 M

4

Inv. 1 Inv. 2 Inv. m

Figure 4.3: Low voltage test feeder and ICT infrastructure (n=7,m=5)

The network simulated in Simulink/SimPowerSystems consists of a MV/LV transformer and a LV feeder. Five PV inverters are connected to the LV feeder as shown in Figure 4.3. It can be noted that the middleware later is implemented using Java Agent Development Environment (JADE). Three different controls were tested using the system shown in Figure 4.3 and the corresponding results are shown in Figure 4.4.


Figure 4.4: Power injection by inverters for different controls – snap shot

Figure 4.4 shows that ON/OFF control and local control curtail power in inverter 4 and 5 whereas other inverters do not curtail any power. In contrast, FPS algorithm of OC leads to equal/fair power curtailment. In short, FPS shows more benefits gained in the scenarios having very high penetration of solar PV. In addition, the time series simulation is conducted and the result is shown in Figure 4.5, which shows that OC leads to equal power injection repeatedly.

Figure 4.5: Power injection by inverters – time series simulation

CONGESTION MANAGEMENT

Congestion generally refers to a situation when the demand or generation at a certain point in the distribution network exceeds the transmission capabilities. The developed CM solution of INCREASE is based on the similar principle of fair power curtailment dealing with an excessive reversed flow due to PV production.

ON/OFF control Local control Overlaying Control

Pow

er in

ject

ion

(kW

)

0

1

2

3

4

Inv. 1 Inv. 2 Inv. 3 Inv. 4 Inv. 5


Distribution transformers are more prone to congestion than the cables. In the Netherlands, 87% of the MV/LV transformers will be overloaded in 2040 compared to 34% of cables

The amount of required curtailment depends on the local loads at any given time

Higher local loads reduce the reverse power flow across the transformer thus the congestion does not occur

A numerical simulation was conducted to verify the congestion management and it could be observed that OC reduces the reverse power flow.

COORDINATION FOR OLTC & PV INVERTERS

The development of the agent-aggregator platform enables the possibility for INCREASE’s OC to coordinate OLTC with DRES inverters in real-time. Based on the remote measurements at critical locations in the downstream area of the LV grid, the coordination control algorithm determines tap position of OLTC such that the voltages in the LV network remains within the predefined voltage limits.

The coordinated control is verified using a numerical simulation and the results are shown in Figure 4.6. It shows that the maximum (Bus 4) and minimum (Bus 6) voltages in the simulated LV network is regulated by coordinating tap position and DRES power generation.

Figure 4.6: The coordination of OLTC and DRES

0 500 1000 15000.9

1

1.1

Vol

tage

(p.u

.)

Transformer LV side Bus 4 Bus 6 Voltage limits

0 500 1000 1500-40

-20

0

20

Pow

er in

ject

ions

(kW

)

Pinj of inv. 1 Pinj of inv. 2 Load 1 Load 2

0 500 1000 1500

-5

0

5

Tap

posi

tion

Time (s)


4.3 INCREASE SCHEDULING CONTROL As the top hierarchical layer of a MAS control structure, SC plans the operation of demand response (DR) units according to various optimization criteria. The whole process of scheduling and grid operation integrates inputs like market price, advanced forecasting of demand, DRES generation, network topology and power quality (PQ) boundaries.

Figure 4.7: General overview of scheduling control layers

This way, DRES units are able to generate maximum energy within the boundaries set by the objectives of LC and OC while the scheduling control agent (SCA) schedules the DR unit based on the energy or economic optimization according to their optimizations criteria, e.g. maximum green energy infeed and maximum profit.

The DR units used for simulations were based on heat pumps parameters and had three operation modes: normal level of operation PDR/2, where the heat pump is running on its medium installed power, full power mode PDR MAX and full stop PDR 0. In the simulations this load is added on top of the various household load profiles generated from the real measurements of a demo site. This load profile creation together with heat pump operation modes is presented in Figure 4.8.


PDR/2

PDR MAX

PDR 0

Figure 4.8: Load profile with added operation modes of DR units (heat pumps)

Economic optimization of DR units

The economic optimization schedule is based on the input data of wholesale electricity market prices and internal costs of DR units. Deployment schedules for each DR unit are made considering the limits of individual DR unit parameter values. It maximizes the profit of DR schedule through maximizing the income of DR unit’s flexibility on wholesale day-ahead market and minimizing the DR unit’s adjustment cost. By increasing the consumption of DR units to PDR MAX in the hours with lower energy prices and reducing it to PDR 0 during the high price periods, the SCA realizes profit/savings with its flexible energy portfolio.

Energy optimization of DR units

The goal of energy based optimization is to fully support DRES production in the network and enable maximum injection of DRES produced green energy. DR schedule is designed to minimize the difference between local consumption and production in the LV grid. With a decrease of the energy difference in the network, power flows are minimized and voltage fluctuations are lowered. Due to a lower voltage level in the network DRES inverter control curtails less active power and greater DRES energy injection is achieved in the network. By energy optimization a profit loss for the aggregator is expected.

The energy difference between the values Wload,k and WPV,k is reduced by activation of the DR units to increase or decrease their consumption in order to achieve energy balance.

Traffic light system

The Scheduling Control solution as well as the entire MAS control concept in INCREASE have been designed in such a way that they enable DSOs to always have control over the grid. For this purpose, a Traffic Light System (TLS) concept has been introduced that gives DSOs the ultimate control over DR unit schedules. The main purpose of TLS is to assure that the schedules of DR units, generated by the aggregator, do not have negative impacts on network situations. TLS is applied to the schedules on day-ahead and shorter time-frames (i.e. intraday and balancing) due to very low accuracy of week-ahead forecasts on the distribution network.


For the implementation of the TLS, DSOs are equipped with an additional TLS module to check for the possible effects of the initial DR schedule on the network, in order to detect any PQ violations in the system. In INCREASE, the PQ violations typically include over/under voltages and current congestions in the feeder under consideration. The TLS module’s main task is to determine whether the SC-supplied schedule of each individual DR unit causes any PQ violations. The TLS is fed with information from the power flow based on the joint load schedule of DR units and inflexible load. The different TLS function logic is shown in Table 4.1 below.

Table 4.1: Types of Traffic Light Systems

Action taken Does the joint load schedule causes PQ violations?

Direction: Does DR schedule help reduce PQ violations?

TLS Type No Yes No Yes

Simple Accept DR schedule

Reject DR schedule

/ /

Advanced Accept DR schedule

Check direction Reject DR schedule Accept DR schedule

Intelligent Use 15-min forecasting to optimally schedule DR units.

Results of the simulations and conclusion

The simulations were performed on the Elektro Gorenjska demo site Suha grid with the use of OpenDSS and Matlab for calculation of power flows and technical analysis for different PV penetration scenarios which are presented in Table 4.2.

Table 4.2: DR development scenarios

Scenario Number of PV

installed Number of DR

units Installed DR units power

[kW] DR integration [%]

1 6 5 44 7.14

2 9 10 95 14.29

3 15 15 139 21.43

4 21 20 183 28.57

5 24 25 224 35.71

6 30 30 264 42.86


The results of the key project parameter, the PV energy infeed, are shown for different penetration levels at both, economic and energy, scheduling optimizations in Figure 4.9.

Figure 4.9: Green (PV) energy infeed differences between economic and energy optimization in summer – simulations by scenarios (PV penetration levels)

Scheduling the flexibility of the DR units with economic optimization yields a higher profit for aggregators on the market. In scenarios with higher levels of DR implementation, economically scheduled units can cause additional PQ violations in the network due to unfavourable nature of the schedules towards the grid situation. Causing the PQ violation would result in penalties for aggregators due to unsuccessful activation of the units. It also negatively impacts the green energy infeed level of the RES, due to curtailment of the units.

Energy based optimization is a more grid oriented form of DR unit scheduling. It focuses on supporting the production of the RES in the network as opposed to more market or price oriented schedules of economic optimization. As a result, aggregators profit on the market achieved with DR flexibilities is lower, but production of RES in the network is increased and more energy is sold on the market. The situation in the grid is more stable and less PQ violations occur. RES units are achieving higher production rates due to less curtailment of the output.

With utilization of advanced forms of TLS, used for evaluation of DR schedule validity, production of RES can be further increased and better operation of the DR units can be achieved.

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1 2 3 4 5 6

Ener

gy [M

Wh/

wee

k]

Scenario

Economic opt Energy opt.


4.4 THE INCREASE SIMULATION PLATFORM

The problem

Most of the commercially available software platforms for power system analysis allow the use of customisable configurations covering almost all electrical aspects in generation, transmission and distribution, including steady-state calculations, power quality optimization and protection coordination. However, a major drawback is that they usually have a closed-form architecture, not allowing the easy and efficient integration of user developed models, especially regarding control systems. Software tools based on time-domain solutions simulate the power system network by solving differential equations that describe all power system components. Although these simulation tools provide the ability to integrate distributed generation (DG) control schemes in a very straightforward manner, they cannot be used for the simulation of extended networks, due to the prohibitive execution times. Software platforms based on phasor solutions use algebraic equations for the simulation of power systems operating in steady-state condition. As a result, power flow and harmonic analysis calculations can be readily performed with small execution times, even in cases of extended networks. The major drawback of such models is that their core calculation routines may become cumbersome with the incorporation of DG droop controls or of other additional control schemes.

The INCREASE solution

The INCREASE simulation platform comprises the benefits of both software categories and allowing the efficient integration of several DRES control strategies.

Over the past decades, research has focused on developing software tools to analyse, plan, optimize and simulate electrical networks. Although novel control schemes have been proposed for interfacing DGs to the grid and mitigating these issues, such controlled inverters need to be also efficiently incorporated in the simulation software packages.

The new simulation tool has the following basic characteristics:

Employs phasor-domain solutions, resulting in short execution times even in cases of extended distribution networks

Offers a graphical user interface (GUI) for the convenient input and configuration of the system under study

Can allow the efficient incorporation of any DG control scheme

Can be interlinked with other software platforms and tools to form a generic co-simulation platform, capable to simulate modern power system networks from both power control and communication point of view.


The INCREASE simulation platform is presented in Figure 4.10. The developed software comprises different open-source tool components and their mutual interconnections.

Figure 4.10: The INCREASE simulation platform

More specifically, the INCREASE simulation platform includes:

The Core engine, which is the base of the simulation platform. The core is developed in MATLAB and implements the interconnections between the different components of the INCREASE simulation platform.

The Draw tool, which is a graphical pre-processor with design capabilities to allow the user-friendly input and configuration of the distribution or transmission network under investigation. All elements work in a drag and drop environment by simply putting them in the design area, configuring their required data, and making the appropriate connections.

The OpenDSS software, which is a phasor-domain grid simulator, capable of performing unbalanced power flow calculations. OpenDSS provides all the advantages of an open-source software, while it also offers highly accurate results, remarkable numerical performance, and vast communication abilities with external software tools.

The JADE software, which is the tool integrating any intelligent multi-agent based control system and the corresponding communication in the INCREASE platform for the implementation of high level control strategies.

CORE


The OMNeT++ simulator, which is an open source LAN simulator, used for the analysis and the evaluation of the communication infrastructure of the examined smart grid.

The INCREASE simulation platform provides the user with a co-simulation tool that can be used to investigate the influence of DRESs on their distribution system. In general, the simulation platform offers the following major features:

Detailed analysis of MV/LV electrical power grids, including all potential DRESs, DGs, loads and control systems. Due to the required analysis, a quasi-dynamic solution is adopted.

Incorporation of an adaptive MAS taking into account multi-objective control algorithms as well as the communication among the individual agents.

Implementation of a multi-layer control strategy for the secure and optimal operation of active distribution grids. The distinct control strategies are coordinated by employing a user defined timeslot concept.

Ability to simulate balanced as well as unbalanced distribution networks with high accuracy.

A near-zero mismatch between the results obtained from the INCREASE simulation platform and most of the available commercial time-domain software platforms.

Provides reduced order equivalent models for both passive and active parts of extended distribution grids.

Can incorporate the most high-end control schemes in DRESs for the effective overvoltage and voltage unbalance mitigation.

Offers the ability to perform time-series simulation with reduced execution times.

The INCREASE simulation platform allows the integration of load and generation forecasting algorithms for short- and medium-term provision of reserve, focusing on the power loss reduction, maximizing active power injection and the optimal performance of DRES. The platform offers the ability to simulate several control schemes for the locally controlled inverter-interfaced DRESs. Τhe P(V) droop control is considered as the default control strategy. Other control schemes such as the Q(V) and the cosφ(P) control strategy can be readily incorporated into the simulation platform. It employs a discrete LAN simulator of communication networks to evaluate the communication performance and the vulnerability of the MAS control system. The LAN simulator is mainly used to analyse possible contingencies of the communication infrastructure on the operation of the MAS control system. However, it can also be used to investigate alternative options on the design of the necessary infrastructure and to examine the communication system vulnerability and its impact on the control system performance.


Simple GUI for the power system design with user-friendly post-processing tools for reporting and plotting results.

Built-in import and export features for most common data and calculation formats.

Flexible platform based on open-source software with modular architecture structure to readily integrate future packages, features and functions.

The above features and advantages make the INCREASE simulation platform a competitive simulation program among other commercial and open-source software packages.


5 INCREASE LAB TEST AND FIELD TRIALS

INCREASE lab test and field trialsx

5 x


5.1 FIELD TRIALS: GENERAL OVERVIEW AND KEY FACTS

Belgian field trial

• Responsible DSO: EANDIS • Location: Koningshooikt, a rural region near

Mechelen • Pilot network: Cable type LV network with a

significant share of DRES on the LV level as well as on MV level. Main types of DRES are PV and combined heat and power (CHP).

Dutch field trial

• Responsible DSO: LIANDER • Location: Holiday park Bronsbergen • Pilot network: LV network of the first micro-

grid in the Netherlands with a high share of PV (installed on approx. 50% out of 210 cottages).

Slovenian field trial

• Responsible DSO: ELEKTRO GORENJSKA (EG) • Location: Village Suha near Kranj • Pilot network: Rural cable type LV network

with a high penetration of PV. The total installed capacity of the PV is 210 kW.

Control strategies for OLTC

Local control (over-voltage mitigation)

Coordinated control

Local control (unbalance mitigation)

Local control (over-voltage mitigation)

Fair power sharing

Congestion management

Control strategies for the OLTC

Austrian field trial

• Responsible DSO: ENERGIENETZE STEIERMARK (ENS)

• Location: Main administration building of ENS in Graz

• Pilot network: Roof-mounted part of a PV installation, which consists of several single-phase PV inverters with a total installed capacity of 47 kWp.


5.2 THE BELGIAN FIELD TRIAL

The pilot network selected for the Belgian field trial contained a large number of PV and heat loads, connected to LV level. Consequently, there are high voltage variations, which are in some cases close to the limits, specified by the power quality standard EN 50160. A large PV and CHP unit on MV level additionally contributes to high voltage variations in the LV network. To cope with these variations, EANDIS decided to investigate the possibilities of voltage control using OLTC, and to examine whether OLTC is a suitable techno-economic measure for increasing the load and DRES penetration in LV networks.

To set-up the field trial, the Belgian DSO EANDIS replaced a conventional distribution transformer with a 400 kVA unit, equipped with OLTC. To monitor the impact of OLTC, monitoring devices were deployed at the transformer station and at four key points, located at the end-points of LV cables. The communication infrastructure was also deployed, to provide real-time acquisition of the measurements and remote control of OLTC, using a centralised algorithm. Measurements were transmitted to the distribution management system (DMS) of EANDIS via a mobile network.

Figure 5.1: Substation with OLTC (left) and cable cabinets with measurements and communications (right)

EANDIS conducted extensive tests of OLTC performance for more than one year. Different OLTC control algorithms have been evaluated:

A basic OLTC control, where there is a fixed voltage set-point at the LV side of the transformer and the tap position is adjusted in order to obtain the given voltage set-point.

A basic control algorithm with line drop compensation (commonly known as LDC with ABC points), which operates in a similar way as the basic control algorithm, but here the voltage set-point depends on a power flow through the transformer.

A central OLTC algorithm (remote control from the control centre), which takes into account the measurements from critical points of the network (e.g. end-points of the long feeders) and selects the optimal tap position accordingly.


First two algorithms operate locally based on local parameters (voltage at LV bus of transformer), while the third operates on a central location (DMS in control centre), based on measurements from multiple points of the network. Therefore adequate communication and measurement infrastructure is required.

As can be concluded from the results for different demonstration cases basic local OLTC control improves the voltage only locally at the LV bus of transformer and shortest LV feeder. High voltage variations remain in range of the longest LV feeder, since basic local OLTC control does not consider voltage conditions of other points of the network. In case of LDC and centralised OLTC control, an overall improvement of voltage can be observed, since the voltage variations are significantly reduced, thus also at the end of the longest LV feeder.

A very similar performance of LDC and centralised algorithm can be observed, especially in considering the most critical point at the end of the longest LV feeder. Based on these conclusions, EANDIS decided that the most suitable control algorithm for future implementation would be LDC control, since there is no need to invest into additional communication and measurement infrastructure, as in the case of centralised algorithm. There is also a reduced risk in the case of communication failure and a need for back-up of the local OLTC control algorithm in case of communication failure is avoided.

The most valuable outcome of the Belgian field trial is the fact that OLTC can be a suitable techno-economic option for the integration of additional loads and DRES into the LV network. To identify an optimal OLTC control algorithmspecifics of the LV network must be considered, not only voltage variations, but also available communication and measurement infrastructure. If there is a need to invest into new measurements and communications, classical local OLTC control algorithms might be a better option than centralised algorithms.

5.3 THE SLOVENIAN FIELD TRIAL

The pilot network selected for the Slovenian field trial consisted of many PV units causing voltage rises on the one hand, and several long feeders without PV units, causing voltage drops on the other hand. Specifics of the pilot network provided a suitable testing environment for the evaluation of different INCREASE and alternative solutions for improvement of the voltage conditions in LV networks. Besides classical OLTC control algorithms, the Slovenian field trial focused on the evaluation of local control for mitigation of over-voltages and coordinated control of OLTC and local control. To set-up the trial, Elektro Gorenjska implemented network analysers for monitoring of the key nodes in the pilot network (substation, PV units, end-node of the longest feeder without PV units). Measurements were acquired within a 1-minute period and transmitted to the control centre of Elektro Gorenjska (PQ database) via their private WiMAX communication network.

To enable voltage control, a 400 kVA transformer, equipped with OLTC was deployed. A dedicated remote control algorithm for OLTC was developed by TU/e and deployed on a


separate computer. This algorithm uses measurements from multiple points of the network to determine new tap settings of OLTC.

To demonstrate functionalities of the INCREASE local control, a 22 kW PV unit, where the highest voltages appear, was selected. A controllable load was installed in parallel with the PV unit, as an introduction of a new controllable 3-phase inverter would lead to significant interactions with private installations of PV owners (large space requirements, reconnection of the PV panels, potential risks of new equipment). That way the same effect as in the case of a controllable inverter was achieved, however without interfering with private owned PV installations.

Figure 5.2: Basic set-up of the Slovenian field trial

To connect all devices used in the Slovenian field trial, a dedicated state of the art SCADA system (SIPRONIKA Unifusion server) was deployed, to enable communication between the systems in a standardised way (standard communication protocols such as OPC UA, DNP3.0). Furthermore, the SCADA system served for a basic visualisation of the network measurements. The complete set-up is shown Figure 5.2.

Following demonstration scenarios were evaluated in the scope of the Slovenian field trial:

SC1 Business as usual, no control measures applied

SC2 Basic local OLTC control, with a fixed voltage set-point

SC3 INCREASE centralised OLTC control, using average phase voltages

SC4 INCREASE centralised OLTC control, using phase voltages


SC5 INCREASE local control (using controllable load) only

SC6 Coordinated control: basic local OLTC control and INCREASE local control

SC7 Coordinated control: INCREASE centralised OLTC control (phase voltages) and INCREASE local control

Figure 5.3: Statistical analysis of average phase voltages

The results shown in Figure 5.3 were evaluated for the four key nodes: LV bus of transformer; PV Urh, 22 kW PV unit, where the highest voltages occur and controllable load is deployed; PV Bassol, 22 kW PV unit and Suha 56, node where the lowest voltages are experienced.

Scenario 1 reveals significant voltage variations in the pilot network. There are voltage rises at PV units and voltage drops at the end of the longest feeder without a PV unit. Similar as in the case of the Belgian field trial, the basic local OLTC control (scenario 2) improves voltage only locally, at the LV bus of transformer, where voltages are concentrated around a given set-point of 235 V. High voltage variations remain unchanged, especially in case of PV Urh.

Centralised OLTC control algorithms, developed within INCREASE (scenario 3 and 4) result in more variations at the LV bus of transformer, but improve the voltage at other critical nodes. Better results are obtained in the case of scenario 4, where the algorithm considers minimal and maximal phase voltages - voltage at the node with the highest voltage drops is slightly increased, while significant voltage variations at PV Urh and PV Bassol are reduced.

Scenario 5 was meant to evaluate the performance of the INCREASE local control for mitigation of over-voltages, without the support of OLTC. Due to local characteristics of such

Transformer PV Urh

PV Bassol Suha 56


control, there is no significant impact on other nodes than PV Urh, where controllable load was deployed. The load started to consume PV generation (imitation of PV curtailment), when the voltage of PV Urh exceeded a value of 236 V. Controllable load (INCREASE local control) contributed to less significant voltage rises at PV Urh.

In scenario 6, there was a coordinated control of a basic local OLTC control and INCREASE local control for mitigation of over-voltages. Similar as in the case of scenario 2 (basic local OLTC control), most significant voltage improvement is achieved on the LV bus of transformer. In addition, the voltage rises were reduced at PV Urh, due to PV curtailment using controllable load.

Compared to scenario 6, overall improvement of voltage was achieved in scenario 7, since both, the lowest and highest voltages, were considered. Lowest voltages are mainly tackled by centralised INCREASE OLTC control (using phase voltages), which has a priority to solve lowest voltages in the network by increasing the tap position. Highest voltages on the other hand are reduced with INCREASE local control, installed at PV Urh, with most significant voltage rises.

In the scope of the Slovenian field trial, all solutions were successfully validated, however a more significant impact of control strategies, especially of coordinated control, was expected before the trial. The deviation of achieved from expected impacts is a consequence of several challenges, faced during the implementation and operation phase:

Due to network configuration (high PV generation and long feeders without PV units), there were low and high voltages in different nodes at the same time, which hindered the operation of the OLTC (tap adjustment would improve voltage in some parts of the network, while voltage in the other parts of the network would be deteriorated)

Coordinated control of INCREASE centralised OLTC control using phase voltages and INCREASE local control (PV curtailment) is a possible solution for previous challenges, but extensive tests over a longer period with different parameters (settings) would be required

High voltage unbalances also hinder operation of OLTC control algorithms, especially in the case of algorithms based on average phase voltages (Scenario 3)

Bad weather conditions and practical implementation challenges, associated with integration of new equipment (controllable load) resulted in a limited amount of reliable results for last three scenarios

In conclusion, to obtain an optimal OLTC control algorithm and parameters, extensive tests of different algorithms and control settings of OLTC over a longer period (e.g. one year) are required. The same thing applies to INCREASE local control, where tests over different seasons are required to obtain optimal settings, i.e. voltage thresholds, where PV curtailment is initiated and voltage thresholds, where PV generation must be switched off.


In case that the LV network is experiencing low and high voltages at the same time, one solution (only OLTC) might not be enough to cope with these issues. Combinations of several measures (e.g. coordinated control) would be a better option in such cases.

5.4 DUTCH FIELD TRIAL Since the pilot network selected for the Dutch field trial contains many PV units, it was used to test INCREASE solutions which are meant to tackle over-voltages (local control, FPS) and congestion of distribution transformers.

The basic idea of the Dutch field trial was to replace some of the existing PV inverters with new ones, provided by MASTERVOLT. These new inverters have the capability to remotely adjust the active power limiting set-point. However, since new inverters did not incorporate INCREASE local control on a hardware level, the concept was imitated on a middleware layer using the cloud platform IntelliWeb. IntelliWeb enabled owners of the MASTERVOLT inverters to monitor the performance of their PV installation on a near real-time basis using a Wi-Fi network. To distinguish between physically integrated local control and solution, used in the Dutch field trial, a term emulated droop control (EDC) was used.

Modification of the IntelliWeb platform was done by TU/e in such a way that a two-way communication between inverter and control agent was enabled. Voltage and active power measurements were acquired with a 5s period. Using these measurements, the control agent provided active power limiting set-points to the inverter, based on selected control functionality and corresponding settings.

Figure 5.4: Operation of EDC and FPS

227

228

229

230

231

232

233

15:28 15:31 15:34 15:37 15:40 15:43 15:46 15:48

Volta

ge (V

)

Time

Upoc 49 Upoc 172

0

500

1000

1500

2000

2500

3000

15:28 15:31 15:34 15:37 15:40 15:43 15:46 15:48

Activ

e po

wer (

W)

Time

Pinj 49 Pinj 172 Pset 49 Pset 172

NO CONTROL EDC

EDC REDUCES VOLTAGE

FPS


Figure 5.4 shows some results of EDC and FPS operation on October 25th 2016. The measurements were acquired from inverters at 2 locations, Bronsbergen 49 and 172. The upper graph represents voltage measurements, while the lower graph represents active power measurements and active power limiting set-points.

Inverters are allowed to operate at 100% available PV injection for approx. 2 minutes, which corresponds to the initial procedure of EDC. After this initial procedure, EDC initiates PV curtailment, since the voltage of the inverter is above the 218.5 V threshold for activation of PV curtailment. A slight reduction of voltage can be observed after active power curtailment.

Active power measurements shown in Figure 5.4 indicate that after EDC was activated, the PV generation (curtailment) is not equally redistributed among the observed inverters. Around 15:39 h, FPS took over and provided new active power limiting set-points for both inverters in such a way that their PV output was equalised.

Figure 5.5: Congestion management - total loading of transformer

Figure 5.6: Congestion management – relative curtailment per inverter

-20

-15

-10

-5

0

5

10

11:05 11:10 11:15 11:20 11:25 11:30 11:35 11:40 11:45 11:50

Activ

e po

wer (

kW)

Time

Total loading Loading threshold

0,00

10,00

20,00

30,00

40,00

50,00

60,00

70,00

80,00

90,00

100,00

u1a u1b u3a u5b u6a 24a 28a 28b 46 49 111 116a 116b 125a 172 192

PV c

urta

ilmen

t (%

)

Inverter number

9-min

Control reacts


Congestion management of a distribution transformer was validated. Since there was no actual congestion of transformers in the pilot network, the algorithm was provided with such settings, that it reduced the reverse power flow, caused by PV. Congestion management was activated after there was a continuous reverse power flow for more than 9 minutes.

There were 16 inverters participating in the control action. The number of inverters activated during congestion management depends on the reliability of the Wi-Fi communication. Due to communication issues, the amount of controllable PV power within congestion management was limited.

After congestion management was activated, the reverse power flow reduced. But as in this specific case the amount of curtailed PV generation was very small, this reduction could also have been a consequence of changed solar irradiation or load conditions in the network. A proper response of congestion management was confirmed by inverters (Figure 5.6). However, in case of a more reliable communication, more controllable power would be available and the impact on the total loading of a transformer would be more significant.

Figure 5.6 shows relative values of the PV curtailment, thus in reference to PV generation before activation of congestion management. Results showed that PV curtailment is done in such a way that each inverter curtails an equal share of available PV generation.

Also in case of congestion management, a principle of fairness is considered, although a bit different as in the case of FPS, since the same relative value of PV injection must be achieved instead of the same absolute value.

5.5 AUSTRIAN FIELD TRIAL The pilot network of the Austrian field trial, consisting of different single-phase PV inverters, was chosen to validate and evaluate the concept of the INCREASE local control for mitigation of voltage unbalances (unbalance control in the following).

The basic idea of the Austrian field trial was to replace 2 existing single-phase 8 kVA PV inverters with 2 controllable three-phase 15 kVA inverters. In this case, remaining 4 single-phase inverters are reconnected to the same phase L1, to create voltage unbalances, which are then tackled by controllable inverters (unbalance control).

To set-up the trial, 2 controllable PV inverters were deployed, each with a corresponding controller, which was also used to acquire measurements at connection points of controllable inverters. A local computer was used to store the measurements from controllers and to provide remote access for settings of controllers and aggregation and processing of the measurements from controllers. In addition, a network analyser was deployed at the main bus of the 10th floor, to monitor the three-phase voltages and the total phase currents and the neutral current of the main line, supplying the 10th floor of the Energienetze Steiermark (ENS) building.


Figure 5.7: Basic set-up of the Austrian field trial

Figure 5.8: Initial (left) and final (right) site configuration

Two simple demonstration cases were shown in the network. First is the Control OFF scenario, the single-phase inverters created the unbalances, by injecting PV generation into the same phase L1, while the unbalance control of controllable inverters was switched OFF. The control ON scenario was the same, but here the unbalance control was switched ON, to tackle the unbalances, created by single-phase inverters.

The performance of the unbalance control was validated and evaluated using voltage and current measurements, acquired from the main bus and main line supplying the 10th floor of the ENS building. Figure 5.9 indicates high current unbalances, generated by unsymmetrical PV generation. The highest current was obtained in phase L1, since all single-phase PV inverters inject the power into phase L1. There was also a high neutral current, which further revealed high current and voltage unbalances. After the unbalance control was switched ON (right chart on Figure 5.9), the neutral current was reduced. This was not the case in the


control OFF scenario, which means that the unbalance control successfully reduced the (voltage) unbalances, created by unsymmetrical PV injection.

Figure 5.9: Reduction of the neutral current in case of unbalance control activation

Figure 5.10: Impact of the unbalance control on maximal phase voltage differences and neutral current

Proper response of the unbalance control is further confirmed in Figure 5.10, which shows results of correlation analysis. On the left chart, the correlation between the maximum absolute values of a phase voltage difference between two phases, indicating the voltage unbalance, and total apparent power has been analysed. The results indicate that in the control ON scenario, maximum voltage differences are reduced. As an example: in case of 15,000 VA, the maximum voltage difference decreases by 4 V.

On the right chart of Figure 5.10, a correlation between neutral current and total three phase current is noticeable. Also in that case, it is very clear that the unbalance control is proper operational. As an example: at a 3-phase current of 80 A, the neutral current is reduced by about 20 A.

The Austrian field trial confirmed that the unbalance control performs according to the concept, defined within INCREASE. As controllable inverters were introduced, voltage unbalances were reduced if the unbalance control was activated.

Using such controllable inverters could be a good solution for increasing PV penetration in networks facing voltage unbalances. The Austrian field trial shows also that PV inverters with unbalance control do not further deteriorate voltage unbalances, but improve the situation. A more significant improvement of the voltage unbalances may be achieved by increasing the share (total power capacity) of such controllable inverters.


Practical lessons from the field trials

Planning and testing

The INCREASE field trials showed the need for a thorough planning and testing period, depending however on the specific technology. For equipment not being officially certified, FAT and SAT tests are mandatory in order to minimize expected risks. Any fault during exploitation is by default time consuming, usually complex and costly. Moreover equipment faults endanger safety of operations and project time schedules.

Communication systems

Communication seemed an issue in most cases. In particular, we suggest an internet platform designed for people to see energy demand and which enables two way communications. In case of using WiFi, as far as possible preparation for router adjustments have to be made. Using customers WiFi also raises question about privacy. Another issue related to communication is remote access.

Matching new with existing hardware

The field trials highlighted the need to compare the technical requirements of old and new systems as well as assess whether they are matching. It is also important to check the dimensions for replacement; users are using free space around the inverters, making the installation more difficult and time consuming. In the EG case, some of the panels were shadowed and communication antennas created noises in the measurement of controllers. Besides the utilization of new equipment, demo cases should also be based on well proven and reliable day-to-day technologies in order to minimize general demo risks. A thoughtful implementation of upgraded day-to-day systems enables later an easier implementation of positive demo outcome.

Interaction with customers

Some of the field trials underlined the importance of getting people / customers on board for the test to avoid problems at a later stage. This has proven to be intensive and time consuming. In particular, we find that it is of key relevance to have at least one meeting on the location of the field trial with all responsible partners where all details are discussed thoroughly. Demo customers, being directly or indirectly involved should be devoted maximum care. The sharing of information, clear project insight and fair play are mandatory. The quality of services should never be reduced and demo goals achieved in order to attract customers’ future interest and engagement. Necessary financial reimbursements should not be avoided


6 BUSINESS CASES AND NEEDED FRAMEWORK CONDITIONS

Business cases and needed framework conditionsx

6x


The problem

For many smart grid solutions there are no robust business cases available, this include also aggregation of flexibility and in many EU countries there are no suitable regulatory framework conditions for aggregation. In particular the implications of the INCREASE technical solutions and the control strategies on possible new business models have to be investigated.

The INCREASE solution

A Value Analysis Tool incorporating the systematic Value Analysis Methodology, based on development scenarios, was developed within the INCREASE project. It assessed the profitability but also effects on emission and energy security of different scenarios. All relevant stakeholders were identified, relations between them and remuneration options.

The chosen business model was applied to the technical results of evolution scenarios performed on two different networks, the Elektro Gorenjska network and the representative synthetic European grid (EG). A sensitivity analysis varied specific parameters in order to determine their impact on the business of the actors involved in the business model. Impacts of variations for aggregators’ software costs, costs of his personnel and availability fees, were investigated along with changes on electricity prices on the wholesale electricity market. The assessments were complemented by an analysis of the needed regulatory framework to enable INCREASE solutions.

To assess the value obtained by market actors, the Value Analysis Methodology was applied to the simulation results obtained using the INCREASE Simulation Platform.

It starts with the Technical Analysis where the MAS control strategy operation is simulated in a typical distribution network, using Evaluation Scenarios.

The simulated outcomes of the MAS control strategies provide operating schedules of DRES- and DR-units and thus the ‘physical value’ of the four key INCREASE AS.

The economic, environmental and operation security value provided by results of MAS Evaluation Scenarios are assessed with the Value Analysis Tool using AS Value Analysis Scenarios.

Appropriate criteria and metrics including key performance indicators (KPIs) have been built into the VAT to quantify the value created via INCREASE solutions.


Figure 6.1: The process of evaluation of the INCREASE solutions

6.1 STAKEHOLDERS, SERVICES AND BUSINESS MODELS

One of the first steps for evaluating INCREASE solutions was the identification of all relevant stakeholders, relations between them and remuneration options. Following stakeholders have been identified as relevant: DR + PV consumers, Aggregators, DSOs, National FiT aggregators, and remuneration options: FiT, Premium, Market.

In addition to the already established business and remuneration models new AS achieved under INCREASE solutions were assessed:

Basic Service (handling the PV units in case of PQ violations),

Scheduling Service (integrating DR unit schedules into operation compared to only VC+VUM),

Upgrade of inverters in case of VC and the implementation of new ones in case of VUM.

Balancing Service (the aggregators align its activities based on SC Economic Optimization - maximizing economic remuneration or SC Energy-based Optimization - maximizing PV electricity injected).

6.2 INCREASE BUSINESS MODEL FOR EVALUATION Based on the results of initial economic calculation for different business models which were derived from possible combinations of relations between stakeholders and key services presented in Figure 6.2, a new INCREASE business model was proposed and selected for the evaluation of INCREASE solutions. This model focuses on improving the aggregators’ conditions and profits, where PV owners are aggregated in the aggregators PV portfolio and their produced energy is sold on the wholesale market. The PV unit owners pay a contractual share of their market revenue to the aggregator for its service as a broker to provide them access to the wholesale electricity markets. This share represents the cost of service of the aggregator to give the PV units market access and covers all the market risks associated with the selling of PV energy. PV owner’s costs consist of upgrades of the inverter control which allows them to actively participate in voltage control schemes and to increase their energy production.

Parameters

Framework matrix

Technical

Market

Economic

Regulatory

Business case Value Analysis Tool Value of the INCREASE AS

Stakeholders Technical Analysis

of Evaluation Scenarios


In the chosen model, for the DR units in the aggregators’ portfolio, the aggregator has to cover the costs of acquiring DR unit flexibility capability CDRe. This includes installation, operation and maintenance of the Advanced Metering Infrastructure (AMI, ADR box and the required communication equipment) and the availability fee payment to the DR unit owners. In addition to DR costs, the aggregator covers his own cost CSCA, which represents his cost of software purchase and updates together with personnel and overhead costs. The revenue of the aggregator also includes the income RA-DR stems from successful sales of Flexible Energy Products (FEP) on the wholesale electricity markets. The FEP are composed of the DR units’ energy changes in a given time interval, assumed that they are sold at the same market price SM in both directions (increase and decrease of consumption).

Relative and absolute profitability of the aggregator was determined for both cases of operation: joint PV and DR unit portfolio and DR unit aggregation only. Initial set of results for value analysis showed, that implementation of availability fee paid by the aggregator to DR unit owners, set the DR unit in positive orientation of the profits, but the aggregators expenses exceeds his income from selling DR FEP on the wholesale market. Due to the fact that the difference between the high rate and the low rate of electricity price is relatively low, his income from selling DR energy products on wholesale is too low. This conforms to the market reality as the aggregator usually offers his services on the reserve or balancing market, where prices are higher. The chosen INCREASE business model for evaluation with associated costs for each stakeholder is presented in Figure 6.2.

Figure 6.2: The chosen aggregator business model for evaluation

DR unit

Energy Market

PV unit - Market

SCA Aggregator

Energy flow

Cash flow

Prod

uced

ener

gySh

are o

f pro

fits f

rom

sellin

g the

ener

gy on

the w

holes

ale m

arke

t

Energy sales of PV production and

FEP products on wholesale market

Profits from PV energy and FEP products

Flexible Energy Products

Share of profits from selling FEP on

the wholesale market, DR availability

fee

Software cost and updates,

Personnel cost,Overhead costs

ADR box installation DR communication

equipment cost

Inverter LC software upgrade cost


6.3 INCREASE EVALUATION SCENARIOS AND SENSITIVITY ANALYSIS

The chosen business model was applied to the technical results of evolution scenarios performed on two different networks, the Elektro Gorenjska network and the representative synthetic European grid. In order to evaluate the added value of the proposed INCREASE business model and solutions, INCREASE first had to estimate the costs for providing the services. The costs were assessed through market research backed up by stakeholder discussion.

Table 6.1: Number of employed personnel for different scenarios and pool size, with amount of aggregated units

Scenario / utilization

100 unit pool size 1000 unit pool size 10000 unit pool size

Number of personnel

DR units

PV units

Number of personnel

DR units

PV units

Number of personnel

DR units

PV units

1 1 20 40 1 200 400 3 2,000 4,000

2 1 40 80 1 400 800 3 4,000 8,000

3 1 60 120 2 600 1,200 3 6,000 12,000

4 1 80 160 2 800 1,600 4 8,000 16,000

5 1 100 200 2 1,000 2,000 4 10,000 20,000

A sensitivity analysis varied specific parameters in order to determine their impact on the business of the actors involved in the business model. Impacts of variations for aggregators’ software costs, costs of his personnel and availability fees, were investigated along with changes on electricity prices on the wholesale electricity market. Parameters, which were defined in the business model and which determine the profit sharing between the actors, were also investigated together with the size of the aggregation pool in aggregators business. Parameters and their variations are given in Table 6.2, where values in bold represent initial parameter values, which were used for calculations for variations of other parameters.

Table 6.2: Sensitivity analysis parameters

Parameter Variation of default values

Energy price 100 %, 200 %, 300 %

Agg. share from PV sales 10 %, 20 %, 30 %

Agg. share from DR sales 75 %, 50 %, 25 %

Agg. SW cost [€] 100 000, 75 000, 50 000

Agg. DR pool size 100, 1000, 10000

Availability fee per DR unit 300 €, 600 €, 900 €


To apply INCREASE solutions within the wider, EU region, data on grid design and operation were gathered from several DSOs in central, southern and northern Europe. A representative synthetic grid was created based on a questionnaire about typical network settings and parameters in different European regions comprising the typical amount of feeders per transformer station, size of loads connected per feeder, and typical loading of the transformer and lines in the network. In the representative grid, the impact of INCREASE technologies’ implementation was analysed using a scenario approach. The sensitivity of those results to key drivers was subsequently analysed.

Key assumptions

For PV, the number of units reaches up to 20,000 in the biggest pool size for 10,000 DR units. 7.5 kW DR units were used in simulations, along with 20 kWp PV installations. The outcome of the control of flexible energy sources is highly dependent on the rules and boundary conditions within which the system operates. These rules encompass technical, economic, market and regulatory provisions and define the framework specific for each country in which the control solutions are deployed. The chosen cost structure is of high importance for the subsequent assessments. We assumed corresponding personnel or overhead cost and varying number of needed personnel dependent on the DR pool sizes, ranging from one to four. If an established company such as an energy trader starts with aggregation as part of a broad portfolio of business activities the costs may be lower.

Figure 6.3: Economic optimization with 10,000 and 1,000 units, for different penetration scenarios for an EU grid


The results of evaluation show that there are cases with positive revenues for aggregators implementing INCREASE solutions. The aggregator has a successful business if his costs are distributed among sufficient DR units in his pool. In case of an aggregator’s pool size of 10,000 aggregated DR units, costs can be spread across enough units to achieve a positive profit in operation. Scenarios with a lower level of integration of PV and DR are less profitable, which becomes problematic in case of a smaller aggregator’s pool size, where he becomes profitable only with increased energy prices or due to the inclusion of PV units in his business portfolio. For all of the pool sizes, a higher level of integration presents better cost distribution and more favourable conditions for his operations. Other factors, such as higher market prices as in the case of reserve markets, would promote profits for aggregators. Under chosen cost assumptions, for small pools, flexibility from DR is hardly profitable. DR units need to provide flexibility on reserve markets; the availability payments need to be low and the share of profits they get from the aggregator also needs to be low. For PV also smaller pools are profitable. This means for small pools the aggregators should mainly include PV in their aggregation.

Regarding cost assumptions, it was assumed that the aggregator starts-up his business. If established companies such as energy retailers that start to include aggregation in their business portfolio are considered, the costs may be much lower and also small DR pools may become profitable. Overall the EU grid is more profitable in comparison to the EG, even when assuming lower personal costs for Slovenia than for the EU average, which is due to a fact that the EG is very resilient and the impacts of new controls are smaller.

6.4 THE NEEDED REGULATORY FRAMEWORK

In several EU countries aggregators are entering the electrical energy or reserve markets. The market access for aggregators is improving in some EU countries, while others are still lagging behind. In several EU countries there is not even a legal framework for aggregation.

In Austria there is no DR participation in the spot markets. In 2014 however the technical prequalification in the reserve market of the minimum pool-size for tertiary reserve was decreased from 10 MW to 5 MW, making it easier for smaller aggregators to participate in the market. Currently, between 5 and 10 aggregators are active on the Austrian market, some of them providing tertiary reserve from industrial processes.

In Germany a first step towards demand side flexibility was made with the ordinance for interruptible loads (IL) introduced in 2012 and taking place in an organized market by regular tenders. Pursuant to the ordinance, large electricity consumers shed loads in cases of bottlenecks, thus stabilizing the grids. In return they receive a compensation that is passed on to electricity consumers. Germany is currently revising its regulatory framework with new provisions for the secondary and tertiary reserve to be adopted in 2016. It aims to allow additional actors providing flexibility from renewables. The access of demand side management providers and storage providers to the reserve markets was simplified. A daily


tender opened each working day for the secondary reserve will enable a better market access to actors with small renewable generation parks, or actors that provide flexibility from production processes. They could for example better forecast the capacities they could offer.

In Belgium in regard to the primary reserve (R1), the Belgian TSO Elia has developed a combination of asymmetric and symmetric Frequency Containment Reserve (FCR) products in order to allow demand side participation. Furthermore R1 requires 30 seconds for the entire volume to be delivered, unlike, for example, the Nordic markets that allow only 5 seconds to reach the entire volume. R3DP (tertiary reserve dynamic profile) is a product of Elia for balancing purposes on the distribution grid. Until 2015 this product was tendered yearly, but since 2016 there is also a monthly tendering, in order to follow the European trend to move towards short term sourcing to increase liquidity in the balancing market. Elia also has plans to move toward a weekly tendering.

Slovenia has among different products for tertiary reserve also one product that can be provided by dispersed production sources, and consumers who can provide demand response (Product DSM). The minimum offer size is 1 MW. The tender is annual, and due to the small size of the market there is currently no possibility to offer shorter-term tenders (e.g. weekly or daily). All together 134 MW are tendered with 15 MW reserved for DSM.

In Italy the wholesale market operators can act as demand aggregators (dispatching users). However, there are no independent demand response aggregators in Italy today. Regarding the balancing market, the regulatory framework for aggregated Demand Response participation is not yet in place.

Aggregation is not legal in Spain and there is only one scheme allowing demand response. For demand response at low level (residential or small communities) with aggregators, there is no regulation yet.

France needs a FCR of 572 MW. This is also symmetrical but the market is not completely open. Instead, each generator is required to provide capacity at a regulated price of €18 per MW per hour. Recently, the market opened up via bilateral agreements for DR participation (in 2014) and for traders (in 2015).

Often the regulatory frameworks are not supportive for demand side management or participation of distributed renewable generation. Important adjustments of market regulations as we observe in a few countries is the reduction of the minimum bid sizes to allow small renewable generations to participate in tenders, or shorter scheduling periods.

Regarding INCREASE solution mainly portfolio sizes above 1 MW, the minimum bid size in many reserve markets, led to business cases. However depending on cost assumptions also smaller amounts may lead to business cases and therefore the market provisions should not be prohibitive.


7 IMPACT CREATION AND EXPLOITATION

Impact creation and exploitationx

7 x


7.1 OUTREACH AND DISSEMINATION

Apart from close cooperation of the project partners, management meetings and technical meetings, INCREASE partners were very active in outreaching to a large range of actors. To disseminate the results of the project to the full extent, international summer schools for postgraduate students and young professionals were held. The main objectives were transfer of knowledge and latest research achievements as well as insights on lessons learned from practical experiences about the preliminary results but also include the view of EU countries not covered by INCREASE. During the project three separate summer schools were organized one in Ghent, one in Ljubljana and one in Eindhoven.

Figure 7.1: INCREASE summer school in Ljubljana

In particular also cooperation with other related EU funded project was sought, specifically with DREAM, EvolvDOS and IDEAL4grid. At several occasions common workshop sessions were organised. INCREASE organised several dedicated technical workshops with outreach to academia and industry.

Figure 7.2: Technical Workshop in Thessaloniki in September 2015


INCREASE was also present at major international conferences, such as MedPower or Barcelona Utility week in 2016.

7.2 EXPLOITATION

In this subsection we briefly provide an overview of the exploitable knowledge derived from the INCREASE project results and the methodology adopted to develop and describe detailed exploitation and dissemination pathways. Table 7.1 gives an overview of products and solutions that could possibly be exploited after the project’s end.

Table 7.1: Overview of exploitable knowledge

Exploitable Knowledge Description Area, field or sector Owner

INCREASE inverters with control algorithms

Smart control strategies for inverters for battery systems and UPS systems

Research community, Energy supplier, Electricity sector

UGent

OCSG software solution Plug in software application in grids

Operators, energy providers, electricity sector in general

TU/e

INCREASE simulation platform Software tool for the analysis of smart grids

Research community, DSOs, energy sector

AUTH

Scheduling control algorithm Publication of algorithm in academia

Research community UL

Value analysis Decision making solver for public bodies and private investors

Research community, Investors

JR

From a methodological point of view, the following steps and approaches were adopted to develop an exploitation and dissemination plan for INCREASE research activities:

Identification and definition of the exploitable knowledge;

Agreement on the common exploitation approach and selection of Key Exploitable Results (KERs) within the INCREASE project team

Workshop with INCREASE participants and consultants

Definition of specific project results relevant for commercialization;

Definition of special project results relevant for publication in scientific, peer-reviewed journals;

Definition of individual exploitation plan for each of the partner;


Communication and dissemination activities already launched for the exploitation of the results of the project.

By means of the CANVAS business model, target groups, customer segments, early adopters, unique selling points as well as key metrics were identified, the market as well as cost structures were analysed and exploitation channels were derived.

For a few products detailed exploitation plans were developed including the needed budget, and roadmaps towards commercialisation, for other products or results academic exploitation paths were defined. All INCREASE partners outlined their plans how to build on the INCREASE results.

Documents

INCREASING THE PENETRATION OF RENEWABLE ENERGY SOURCES · 6 increase synthesis document list of figures figure 3.1: main increase results, their interdependencies and timing 15 figure