IFAQ – 04.06.

2018

Description, Modelling and Simulation of a

Benchmark Test System for Converter Dominated
Grids (Part II)

98% - RES Benchmark on Basis of Fast Voltage

Source controlled Converters (EMT & RMS)
Prepared for

DIgSILENT PowerFactory Users

 DIgSILENT GmbH
Heinrich-Hertz-Strasse 9
D-72810 Gomaringen
Tel.: +49 7072 9168 - 0
Fax: +49 7072 9168- 88
http://www.digsilent.de
e-mail: mail@digsilent.de

Contact:

Martin Schmieg
+49-7072-9168-10
m.schmieg@digsilent.de

Rev. Date Scope Author

0.1 29.05.2018 Initial Implementation M.Schmieg
0.9 29.05.2018 Model and Simulation Review S.Weigel
1.0 04.06.2018 Model and Presentation Review A.Constantin

EMT/RMS VSCC-Benchmark 2
 Table of Contents

Table of Contents

1 Introduction ......................................................................................................................................... 4

2 The PowerFactory 98%-RES Model .................................................................................................... 5

3 EMT- and RMS- Simulation Cases ........................................................................................................ 7

3.1 Case 1: 3phase Fault, 150ms at Busbar “A” ............................................................................................. 8

3.2 Case 2: Trip of 800 MW Load at busbar “A” ............................................................................................ 11

3.3 Case 3: Connection of a 250 MW load at busbar “A” ................................................................................ 13

3.4 Conclusions ......................................................................................................................................... 14

4 References ......................................................................................................................................... 15

EMT/RMS VSCC-Benchmark 3
 1 Introduction

1 Introduction
This paper presents the implementation in PowerFactory of a fast voltage controlled based method for converter
dominated power systems, as proposed in /2/ and its behavior within a benchmark system, as proposed in /1/.
Its principal function and overall stability are demonstrated for a representative 9 bus system with a 98% share
of renewables, full-scale converter based WT generation (Type 4). Nonetheless, the proposed converter
regulation strategy would be applicable for other power sources such as VSC-based HVDC, PV- or storage
systems.

The proposed converter control method takes into account that the further development of actually operated
power systems will not allow a fundamental change in control- and operation philosophy as e.g. proposed by /3/.
In other words, the traditional synchronous machine based control of voltage and frequency for balancing
reactive and active power demand as such has to be maintained, however with a steadily increasing percentage
of converter based generation such as wind power (WT), PV generation, battery storage and other PQ-controlling
devices. As mentioned already in /1/, this will require converter based systems to be also operated in voltage
source and frequency balancing control modes (VSCC). The critical challenge here is the frequency control as the
share of inertia based, synchronous generation may fade out in future to a relative low margin. As a
consequence, load balancing has to be implemented in a superfast way aiming in the limitation of frequency
excursions within acceptable operational limits.

Of course, a number of issues are still open and need further investigations. Among those are:

1. The proposed fast voltage control of converter based generation (here, the 98%-RES case) is showing a
very stable behavior. The remaining synchronous generation is “swinging” around the quite rigid RES-
generation. The damping of the machine oscillation can be tuned via PSS as it is done in classical
systems. However, it is unclear, if the dynamic interactions between thousands of converters in case of
very large interconnected systems (such as the ENTSO-E system) can be easily and robustly tuned
resulting in an unconstrained and unconditioned damped behavior.

2. The comparison of the EMT- and RMS- based simulations for identical cases and models are showing an
excellent match for all relevant quantities. However, it is still to be investigated if large, geographically
extended systems with considerable travelling wave effects will still behave in the same way. We might
possibly end-up in a situation where stability analysis in dominant converter based, geographically
extended systems will further need EMT-simulations to validate classical stability results.

3. The stability of low inertia systems will not allow any more the application of classical, feeder based
frequency dependent load shedding mechanisms. The time left to decide if a relay should operate is too
short as the system frequency will have already reached the next frequency stages when the first stage
has decided to operate. As a consequence, there might be the need to operate sufficient active power
reserve as required for a certain “design” contingency. However, in case of more severe faults, the
system might inevitable collapse or possibly split into several regions or sub-grids. Instead on relaying
on frequency depending load reduction, a voltage reduction schemes included in the fast converter
control might help preventing system collapse situations.

EMT/RMS VSCC-Benchmark 4
 2 The PowerFactory 98%-RES Model

4. The benchmark presented in this paper and respective PowerFactory project is based on a hyper-fast
performance of the Type 4 WT-control. Except for the PLL, there is no delay assumed between the
measured voltage at the WT-busbar and the grid-injected current. Consequently, any delay of data
acquisition, signal filtering, etc. are neglected. Here, further analysis is required to determine the
maximum acceptable signal- and control processing delays which do not jeopardize the effectiveness of
the proposed control scheme. Of course, the direct feedthrough of voltage and current quantities require
specific algorithmic capabilities of the simulation software.

5. A further serious aspect to be considered are the torsional interactions of the very fast converter
controls with turbine shaft dynamics of large power plants but also small diesel units designed with
flexibles clutches.

So far, we consider the presented 98%-RES project as a first step allowing to further investigate the above
mentioned issues aiming in supporting a solid and stable transition from synchronous machine dominated power
systems towards stable and robust developments with very high RES share.

Unfortunately, grid code requirements imposed by utilities in the past years have not adequately supported a
global view on system stability issues. Even the latest German grid codes VDE-AR-N 4105/4110/4120 and 4130
do not consider basic stability issues by setting-up sustainable requirements for the control characteristics of
converter based generation.

2 The PowerFactory 98%-RES Model

The benchmark model has been generally described in /1/ covering the grid definitions as well as the
synchronous generator and WT type 4 converter models and parameters. In the course of the preparation of the
project, two levels of converter modelling has been tested. Initially, the DC-intermediate circuit has been detailed
on basis of dedicated DC components along with a more detailed PWM-converter setup.

However, it was found that the response of the model version where the DC- intermediate circuit has been
included in a static generator control setup along with a simplified static generator concept, will basically produce
comparable results. Consequently, the finally presented model is based on a static generator model
(Elm.GenStat) with an implicit DC-intermediate circuit being part of the converter control.

As indicated in Fig. 1 the dq-quantities are generated via the “AC voltage” and “AC current measurement” blocks
along with the PLL output signals in the block “Park Transformation”. The idq, udq and further calculated pq-signals
are fed along with the measured PLL-frequency, the remote bus voltage (remote bus control) and the simulated
voltage of the DC-intermediate circuit to the main control block “Udc_Q Controller”.

As described in /1/ section 3.6 the “Udc_Q Controller” together with the “Damping Terms” and the “Modulation
Limitation” include the following main functions:

1. Reactive power control consisting of the superimposed remote bus voltage control and fast local voltage
control with damping terms and transient current limiting control logic. The fast local voltage control is
only taking the converter reactance into account, neglecting the converter active losses. As a
consequence, the model would not initialize correctly if the active power losses are considered for the

EMT/RMS VSCC-Benchmark 5
 2 The PowerFactory 98%-RES Model

static generator (Elm.GenStat). However, those losses are neglectable compared to the rating of the
device.

2. The active power control path includes the presently in Europe (ENTSO-E) applied frequency dependent
power reduction or power increase in case of frequency excursion outside the 50±0.2 Hz band, a
voltage dependent power reduction as well as the damping terms along with the transient current
limiting logic.

The active power side is insofar simplified that the active power source (wind, solar irradiation, battery
storage, etc.) is neglected. It is simply assumed that an active power resource is available at any time within
the rating of the unit.

Figure 1: Overview of the proposed Converter Model Frame.

All details of the various model components can be studied in the project “98RES_TestGrid_v1.pfd” made
available by DIgSILENT via the FAQ access “https://www.digsilent.de/en/faq-powerfactory.html” (no login
required).

The synchronous generator model incl. prime mover (IEEEG1), AVR (EMAC1T) and a specifically designed PSS is
tuned to provide sufficiently fast active power response along with a fast and damped voltage support.

All applied plant models are identical for the prepared EMT- and RMS-simulations. There are no variations and no
specific operation scenarios.

EMT/RMS VSCC-Benchmark 6
 3 EMT- and RMS- Simulation Cases

3 EMT- and RMS- Simulation Cases

Aiming in a demonstration and verification of the robustness of the proposed voltage source controlled converter
solution, a total of three fault scenarios have been prepared as follows:

Case 1: 3-phase fault with a duration of 150ms for both

EMT as well as RMS simulation setup. In addition,
a study case has been prepared to graphically
display selected EMT/RMS – variables within one
single graphics board

Case 2: Trip of a 800 MW load at busbar “A”

Case 3: Connect an additional 250 MW load at busbar “A”

When running the simulations, there are very few warning messages which can be totally ignored. The basic
integration formula settings are as follows:

RMS cases: adaptive step-sizes ranging from 5ms to 200ms, simulation error (grid) = 1 kVA
EMT cases: adaptive step sizes ranging from 0,01ms to 8ms, simulation error (grid) = 1 kVA

The simulation stop time is preferably set to 2.5s if the initial transients shall be viewed and to 5, 25 or even 50s
if the long term transients are of importance. The user will note that in case of considerable frequency excursions
due to the applied fault, noticeable frequency deviations can be observed between the EMT- and RMS cases.
Those deviations are caused by the neglection of the frequency dependency of grid component in case of the
RMS simulation model.

Note: The user may add or modify the prepared cases as per individual interest. However, it shall be
mentioned that the model has been tested and released for the three prepared cases only. Especially
when defining faults or events, that potentially require protection or additional control systems to
operate which are not included in the initial model, the prepared project may not work any longer (e.g.
when switching in a larger load which cannot be supplied by the WTGs and the SG).

EMT/RMS VSCC-Benchmark 7
 3 EMT- and RMS- Simulation Cases

3.1 Case 1: 3phase Fault, 150ms at Busbar “A”

Figure 2: RMS/EMT-quantities for the synchronous machine SG 2

Figure 3: RMS/EMT-quantities for Synchronous Machine SG2 (2,5s zoom)

EMT/RMS VSCC-Benchmark 8
 3 EMT- and RMS- Simulation Cases

Figure 4: EMT/RMS-quantities for grid objects

Figure 5: EMT/RMS-quantities for grid objects (2,5s zoom)

EMT/RMS VSCC-Benchmark 9
 3 EMT- and RMS- Simulation Cases

Figure 6: EMT/RMS-quantities for WTG1 (2,5s zoom)

Figure 7: EMT/RMS quantities for WTG1 (2,5s zoom); con’t

EMT/RMS VSCC-Benchmark 10
 3 EMT- and RMS- Simulation Cases

3.2 Case 2: Trip of 800 MW Load at busbar “A”

Figure 8: RMS-quantities for the synchronous machine SG2

Figure 9: RMS-quantities for grid objects

EMT/RMS VSCC-Benchmark 11
 3 EMT- and RMS- Simulation Cases

Figure 10: RMS-quantities for WTG1

EMT/RMS VSCC-Benchmark 12
 3 EMT- and RMS- Simulation Cases

3.3 Case 3: Connection of a 250 MW load at busbar “A”

Figure 11: RMS-quantities for the synchronous machine SG2

Figure 12: RMS-quantities for grid objects

EMT/RMS VSCC-Benchmark 13
 3 EMT- and RMS- Simulation Cases

Figure 13: RMS-quantities for WTG1

3.4 Conclusions
The benchmark project presented is demonstrating in a number of representative cases that the fast voltage
source controlling converter (VSCC) method can successfully handle power systems with an extra high share of
converter based generation. Compared to other discussed technologies such as the “Virtual Synchronous
Machine” method, the VSCC method is straight forward with a minimum of dynamic characteristics to be
considered. It simply balances active and reactive power via frequency and voltage, fast efficient and fully
compatible. It integrates seamless with actually operated power system technologies.

However, the reader is encouraged to ad his own control technique proposal and solution to the VSCC-method
presented, finally aiming in finding the best possible control technique for high RES share integration to existing
power systems.

EMT/RMS VSCC-Benchmark 14
 4 References

4 References
/1/ I. Erlich, A. Korai, “Description, Modelling and Simulation of a Benchmark System for Converter Dominated
Grdis (Part I)”, June 2018

/2/ I. Erlich, A. Korai et al., "New Control of Wind Turbines Ensuring Stable and Secure Operation Following
Islanding of Wind Farms," in IEEE Transactions on Energy Conversion, vol. 32, no. 3, pp. 1263-1271, Sept.
2017

/3/ H. Weber, P. Basker, A. Nayeemuddin, “Power System Control with Renewable Sources, Storage and Power
Electronic Converters”, 2018 IEEE International Conference on Industrial Technology (ICIT), 19.-22.2.2018,
Lyon, France

EMT/RMS VSCC-Benchmark 15