See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/261419600

Real-time simulation of More-Electric Aircraft power systems

Conference Paper · September 2013

DOI: 10.1109/EPE.2013.6632008

CITATIONS READS

9 578

3 authors, including:

Christian Dufour Jean Mahseredjian

Opal-RT Technologies Polytechnique Montréal
117 PUBLICATIONS   2,663 CITATIONS    199 PUBLICATIONS   2,733 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Power System Test Cases for EMT-type Simulation Studies View project

Multi-rate decoupling techniques for real-time simulation application View project

All content following this page was uploaded by Christian Dufour on 25 May 2016.

The user has requested enhancement of the downloaded file.

 Paper presented at the EPE'13 ECCE Europe conference, September 3-5, 2013, Lille, France

Real-time Simulation of More-Electric Aircraft Power Systems

Leonardo Montealegre Lobo Christian Dufour

Instituto Costarricense de Electricidad Opal-RT Technologies
Department of Research and Development 1751 Richardson, suite 2525
Apdo. 10032-1000, San Jose, Costa Rica Montreal, PQ, H3K 1G6, Canada
[email protected] [email protected]

Jean Mahseredjian
École Polytechnique de Montréal
Department of Electrical Engineering
Montreal, PQ, H3T 1J4, Canada
[email protected]

Acknowledgements
This paper has been realized in the framework of projects supported and funded by Bombardier
Aerospace, OPAL-RT Technologies, Pratt & Whitney Canada and by the Consortium for Research
and Innovation in Aerospace in Quebec (CRIAQ).

Keywords
More-Electric-Aircraft, Power Management, Real time simulation, Test bench, certification.

Abstract
The More-Electric Aircraft (MEA) requires thorough testing procedures to meet extreme security
requirements. Real-time simulation technologies have been traditionally part of this validation and
testing process for conventional aircraft designs. The MEA concept is now pushing HIL technologies
toward aircraft electric system real-time simulation. This paper discusses the real-time simulation
requirements for MEA, analyses an aircraft electrical power system, identifies real-time modeling
bottlenecks and validation needs. The case of Bombardier Global Express aircraft is used for this
purpose.

Introduction
Traditional aircrafts use pneumatic, hydraulic, mechanical and electrical energy sources to supply their
systems, which normally imply increases in fuel consumption, low power efficiency and heavier
aircraft weight. Therefore, a new concept in aircraft industry (the more-electric concept) has tried to
considerably reduce the impact of these trends, by increasing the implementation of electrical
components in aircrafts for generating, distributing and utilizing electrical power. Thus, the more-
electric aircraft (MEA) concept impacts significantly on aircraft electrical power system design, due to
the fact that many functions conventionally managed by hydraulic, pneumatic and mechanical power
are replaced by electric systems in order to reduce size and weight, and improve fuel efficiency [1].

In contrast, the increased usage of electrical power intensifies the power demands on the electrical
system, placing new constraints on its dynamic performances and on power quality, so that new power
system architectures must be designed, extensively analyzed, tested, validated and certified before
implementation in an actual aircraft.

In this context, mathematical modeling and simulation tools constitute an efficient approach for
predicting operational behavior, correcting design errors, eliminating prototyping steps and reducing
component and overall testing cycles. This is essential to study the impact on costs and electric design
modifications. Furthermore, efforts have been made to develop studies on power distribution systems
Paper presented at the EPE'13 ECCE Europe conference, September 3-5, 2013, Lille, France

at fixed frequency [2] [3], variable frequency [4], and power distribution at high DC voltage [5].
Regardless of the existence of all these studies, many approximations are made, either by necessity or
by lack of data. Consequently, research must be pursued in order to study the stability of new
distribution architectures with more precision, as the presence of power electronics on aircrafts is in
constant growth [6]. This is part of the model-based design paradigm in used for many years in the
aerospace industry [7].

Real-time simulation tools are a robust and practical way to face these challenges during MEA’s
concept implementation. In addition, real-time simulation tools offer the computational speed
advantage and allow studying a very large number of operational scenarios in less time. Moreover,
such tools can provide real-time synchronized simulations and allow interfacing with external physical
devices (hardware-in-the-loop), which can be used to validate physical controllers, improve designs
and even develop models.

This paper presents an initial benchmark for the real-time simulation and analysis of the Bombardier
Global Express (GLEX) aircraft electrical power system. The tools considered are Simulink [8] (for an
off-line first implementation model) and the OPAL-RT [9] simulator for real-time simulations. In
addition, some validations on the off-line model are performed using EMTP-RV [10]. It is shown that
these tools allow the development of advanced models and testing the aircraft system in a wide range
of scenarios. The real-time simulation of the developed power system is particularly complex.
Consequently, a new real-time simulation method called SSN [11] (State-Space Nodal method) is used
and tested successfully for this class of simulation problem.

Testing Standards
In the frame of many research programs around North America and Europe, one of the main drivers
for MEA’s design, integration and performance has been the testing standards. The aim of these
standards is to further understand the requirements and explore the challenges for characterizing new
aircraft electric systems in terms of total power and power quality.

For the scope of this paper, the military standard MIL-STD704F [12] is used to assess power quality in
terms of DC bus voltage deviation under both normal and transient behavior. MIL-STD704F
establishes the requirements and characteristics of aircraft electric power provided at the input
terminals of electric utilization equipment [12].

Moreover, MIL-STAD704F provides a range of acceptable voltage values under both steady-state and
transient behavior. In addition, the military standard MIL-E-7016 [13] is used to perform an electrical
load analysis on the GLEX’s load buses. This standard addresses the methods and analysis of electric
loads and source capacity on military aircraft [13].

Although both standards were prepared specifically for military assignments, their use for commercial
applications has received a widespread acceptance over the years.

Furthermore, other avionics test standards, for both military and commercial purposes, are performed
along with [12] and [13]. For instance, system behavior must comply with Airbus directives such as
ABD01.1.8 [14]. In other cases, the increasing number of power electronic loads would comprise a
high number of AC-DC rectifier stages, which would have to comply with the Airbus’ AC quality
standard (ABD100.1.8.1) [15].
 Paper presented at the EPE'13 ECCE Europe conference, September 3-5, 2013, Lille, France

Modeling of GLEX Aircraft Electrical Power System

The GLEX electric system is separated into AC and DC systems, as shown in Fig. 1. The AC system
uses variable frequency generators (VFG) to supply 115/200 VAC, variable frequency (324-596 Hz)
3-phase AC power. For the purpose of this paper, the VFGs are modeled using ideal voltage sources.
However, this will be upgraded to include electromechanical dynamics and contributions to
electromagnetic transients.

The generator outputs are supplied to the AC Power Centre (ACPC), which in turn distributes the
power to the aircraft subsystems. The ACPC is a sort of switching control center for automatic AC
supply reconfiguration in case of failure and it is based on truth tables representing GLEX
reconfiguration logic [16]. It is responsible for sending binary signals to the external switching time
control for the ACPC switches, in order to change ON/OFF state positions. The amount of logical
controls is vast and implementing all of them is not the goal of this paper. Only basic architecture
reconfiguration according to VFG’s and TRU’s status is executed in the GLEX electric system.

Fig. 1: One-line diagram of aircraft electric power system

Within the AC systems, AC cable impedance parameters are created, scripted in Simulink and then
validated in EMTP-RV. Zero, positive and negative sequence impedances are given according to [17].
A correction factor is applied to use the inductive reactance given in [17] for the entire range of
operations frequencies. A sequence impedance matrix is formed for each wire size from 12 AWG to
00 AWG. Phase domain cable impedance parameters are found using the Clarke transformation matrix
and finally a coupled-RL model (depending on frequency operation, length and wire size) is fully
implemented in GLEX’s Simulink model. Moreover, both passive and dynamic AC loads are
considered in the model.
Paper presented at the EPE'13 ECCE Europe conference, September 3-5, 2013, Lille, France

The DC power system supply consists of four Transformer Rectifier Units (TRUs) for normal power
distribution and two NiCad batteries (nickel-cadmium). Each TRU produces unregulated 28 VDC,
rated at 150 A [16]. Lack of information prevents modeling the TRU as in [18] so that a simplified
model similar to [18] is created in Simulink, using a 12-pulse power converter. For DC cable
modeling, a block is created, edited and scripted in Simulink to generate DC cable impedance based
on [17]. However, more sophisticated models could be developed to account for higher frequency
transients. For instance, it might be needed to account for temperature and altitude effects using
dynamic functions. In addition, DC loads are assumed to be constant impedance loads, drawing a
specific current at 28 VDC, while NiCad batteries are modeled using simple ideal DC sources.

Real-time Modeling of GLEX Aircraft Electric Power System

In order to use a Simulink model in real-time with OPAL-RT’s software and hardware, some
modeling concepts must be followed when building the application. One of the most important
concepts in real-time simulation using OPAL-RT is grouping the model into parallel subsystems (each
subsystem is assigned to a single CPU), using various parallelizations, so that the real-time model can
be computed on multiple cores for achieving real-time performance. In addition, Fig. 1 presents the
GLEX electric system grouped into parallel subsystems or RT-LAB cores.

One of the main challenges of MEA real-time simulation resides in the very large number of switches
in the circuits that are closely coupled to each other due to the lumped nature of MEA circuits. The
ARTEMiS-SSN solver provides a very good solution to this challenge.

Real-time Modeling Using the ARTEMiS SSN Solver

The real-time simulation of the GLEX aircraft is done with ARTEMiS State-Space Nodal (SSN)
solver. The SSN solver creates virtual partitions of the model, described by state-space equations, and
combines them using a nodal admittance method [11]. The partitions are created so that the number of
switches within each of them is kept low enough to enable pre-computation. Then, the common
simultaneous solution of all partitions is found by the nodal method, using on-line LU factorization.

Fig. 2: The SSN algorithm with parallelization capability on multiple cores

Paper presented at the EPE'13 ECCE Europe conference, September 3-5, 2013, Lille, France

This allows for the simulation of large circuits without any added artificial delays and with a large
number of switches, provided that these switches are grouped in different partitions (or ‘SSN groups’).
One particularity of the SNN algorithm and its implementation into the OPAL-RT simulator is that
SSN partitions can be computed in parallel in different cores of the simulator to further accelerate the
simulation. The SSN algorithm is depicted in Fig. 2.

As mentioned before, one of the main challenges in the real-time simulation of the electric circuit of
an aircraft is the very high number of topologically connected switches or breakers. In the case of the
GLEX model, there are a total of 109 switches. These switches are grouped into 25 SSN partitions
combined with a nodal admittance matrix of rank 48, which is 80% sparse.

The model, grouped into 25 SSN partitions, is executed in real-time on the RT-LAB platform (v.10.6)
at a time step of 39 µs. The RT-LAB target was a 3.33GHz 12-core PC running under RT-Linux. To
achieve this performance, the SSN simulation used 4 cores in parallel and made tri-diagonal
reordering of the nodes to take advantage of the sparsity of the nodal matrix, by using a non-sparse
method. Further gain in performance is expected when a fully sparse method will be available in SSN
[20]. Some tests, involving steady-state and transient behavior, are performed on the GLEX’s electric
system model in order to validate real-time performance. Description of the tests and results are shown
in the next section.

Test Results for the Global Express Aircraft

Test 1: VFG1 failure with ACPC response
Fig. 3 summarizes the events in this test, in which ACPC detects a failure in VFG1 at 200 ms and
reconfigures to transfer energy from VFG4 to AC BUS1 as indicated in [19]. After 100 ms, VFG1
returns in operation and ACPC restores the power supply between VFG1 and AC BUS1, as AC BUS1
power supply is no longer provided by VGF4. It is noticed that ACPC transfers and restores the power
supply between VFG1 and AC BUS1 within the 50 ms delay for switching operation, as required in
[16]. Due to this delay, VFG4 supplies energy to AC BUS1 at 250 ms and VFG1 restores power
supply at 350 ms.

Fig. 3: Sequence of events in terms of energy supply for each VFG

The SSN results at a simulation step around 50 µs show a very good match when compared to
simulation at a very low time step of 1 µs. The results, depicted in Fig. 4, show a slight voltage drop
with increased time step (0.05 VDC drop from 28 VDC at a time step of 50 µs). These values are
within the limits established in Table IV of the military standard MIL-STD704F [12], which indicates
that a DC voltage of a system operating at nominal 28 VDC must be between 22 VDC et 29 VDC.
Paper presented at the EPE'13 ECCE Europe conference, September 3-5, 2013, Lille, France

30

28

26

24
Voltage (V DC)

22

20

18

16
SPS 1us
SPS 50us
SSN 50us
14
0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Time (s)

Fig. 4: DC ESS Bus for real-time and off-line simulations

In addition, Fig. 5 shows current simulation results for all VFG buses when VFG1 fails at 200 ms and
restores operation at 350 ms. It is observed that VFG4 supplies more current when VFG1 is off. The
magnitude and duration of the overcurrent in VFG4, as well as the magnitude of currents at new
steady-state operation, meet the requirements established in [12]. Furthermore, steady-state voltages of
all VFG buses before the failure and after the recovery are within the established in Table II of the
military standard MIL-STD704F [12], which indicates that an AC voltage of a variable frequency
system operating at nominal 115 VAC must be between 108 VAC and 118 VAC. Finally, transient
behavior of all VFG buses behaves according to the Envelope of Normal AC Voltage Transient, as in
Figure 3 of the military standard MIL-STD704F [12].
VFG1 Bus Phase A
Phase B
Current (A AC)

50
Phase C
0
-50

0.2 0.22 0.24 0.26 0.28 0.3 0.32 0.34 0.36

VFG2 Bus
100
Current (AAC)

-100
0.25 0.255 0.26 0.265 0.27 0.275 0.28 0.285 0.29 0.295 0.3
VFG3 Bus
Current (A AC)

100

-100
0.25 0.255 0.26 0.265 0.27 0.275 0.28 0.285 0.29 0.295 0.3
VFG4 Bus
Current (A AC)

100

-100
0.25 0.26 0.27 0.28 0.29 0.3
Time (s)

Fig. 5: Zoom in the current simulation results from VFG buses when VFG1 fails
Paper presented at the EPE'13 ECCE Europe conference, September 3-5, 2013, Lille, France

Test 2: VFG4 failure with ACPC response

Finally, it is important to analyze TRU behavior before and after a switching event. In this test, VFG4
fails at 200 ms of simulation, so that VFG1 supplies energy to AC BUS4 as indicated in [16]. After
100 ms, VFG4 returns in operation and restores the power supply with AC BUS4. Again, there is a 50
ms delay between switching operations. Furthermore, the voltage drop is detected by the controls
implemented in Simulink for triggering the closure of the solid state power controllers (SSPC) in 2 ms,
in order to supply battery power to the DC bus. As observed in Fig. 6, the minimum voltage sag
reaches 14.31 VDC in less than 2 ms, while the maximum overvoltage reaches 43.06 VDC during 2.95
ms. Both voltage levels are within the requirements in section 5.3.2.1 of the military standard MIL-
STD704F [12].
BATT Bus

45
43.06 VDC
VFG 1 supplies
40 AC Bus 4

35 VFG 4 on
SSPC supplies
29.4 VDC 29.11 VDC (28.01 VDC)
25.08 VDC

30
Voltage (VDC)

25
VFG 4 failure
(off)
SSPC supplies
20
25.08 VDC

15
18.7 VDC

10
14.31 VDC

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Time (s)

Fig. 6: TRU voltage transient behavior during VFG4 failure

Test 3: VFG1 and VFG4 failure with a miss operation occurring while switching
In this test the phase A of the VFG4 switch remains connected while VFG4 fails at 200 ms.
In addition, the phase B and phase C of the VFG1 switch remains connected while VFG1 fails at the
same time as VFG4. The idea of this test is to simulate a possible problem when a switch does not
operate properly.

Fig. 7 zooms on the moment where the fault occurs. Since two phases of the VFG4 are OFF, ACSC
sends a fault signal to the ACPC in order to command VFG1 to supply AC BUS 4. Therefore, AC Bus
1 shows an overcurrent when failure occurs. In addition, a transient in AC BUS 4 is observed at 250
ms when ACPC tries to restore the power supply.

Regarding the DC buses, DC ESS Bus shows a 7.7% voltage drop through failure, since there is an
overcurrent in AC Bus 1. In this case, SSPC does not work during failure since DCSCC does not see
any fault signal because there is only one phase OFF. Moreover, DC BATT Bus does work on battery
back-up through failure, since two phases are OFF and DCSCC interprets this situation as a failure.
Both DC voltages reach steady-state values after transients and meet the requirements in section
5.3.2.1 of the military standard MIL-STD704F [12]. Fig. 8 shows the DC voltages from DC ESS Bus
and BATT Bus under the conditions stated above.
Paper presented at the EPE'13 ECCE Europe conference, September 3-5, 2013, Lille, France

AC BUS 1
100

50

Current (A AC) 0

-50

-100
0.2 0.22 0.24 0.26 0.28 0.3 0.32 0.34 0.36 0.38 0.4

AC BUS 4 Phase A
Phase B
100 Phase C

50
Current (A AC)

-50

-100

0.2 0.22 0.24 0.26 0.28 0.3 0.32 0.34 0.36 0.38 0.4
Time (s)

Fig. 7: Zoom in on the current simulation results from AC BUS 1 and AC BUS 4
DC ESS Bus

40
25.85 VDC

28.01 VDC 28.01 VDC

Voltage (VDC)

30

20 Voltage drop

10

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
BATT Bus

40
Voltage (VDC)

30

28.01 VDC 28.01 VDC

20 25.08 VDC

On battery supply
10 SSPC

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Time (s)

Fig. 8: Voltage simulation results from DC ESS Bus and BATT Bus

It is noticed that the analyses presented in this section meet the intent of MIL-E-7016F [13] as well.
Furthermore, as shown above, to develop as accurate a model as possible, the power system military
standard MIL-STD-704f [12] is used to meet basic steady-state and transient output requirements.

Other validations are made on the model performance during failure, involving energy transfers in
accordance with ACPC truth table logic. The power quality of the simulation results is evaluated
according to [12] and [13]. Results show that voltage ripples, as well as AC and DC responses are
within the standards.
 Paper presented at the EPE'13 ECCE Europe conference, September 3-5, 2013, Lille, France

Conclusion
This paper presented an initial attempt to model GLEX electric systems in steady state and transient
behaviors for real-time simulation. It demonstrated the capability to study system performance under
given operating conditions and potential benefits of conducting real-time simulations for MEA electric
systems. This testing methodology can be used to study various aircraft power system architectures
and test their performance within practical economical and reliability constraints.

The paper also demonstrated that the SSN solver is efficient enough to enable real-time simulation of
MEA with a very high number of switches present in the model. Further work is ongoing with the
objective of decreasing the time step of the real-time simulations, increasing the number of switches
usable in real-time simulation and performing faster HIL tests, as required by complex MEA’s electric
systems.

New concepts or designs are constantly emerging and must be extensively analyzed, tested, validated
and certified before its implementation in an actual aircraft. One long term objective of this work is to
contribute to the creation of a model-based certification process for MEA, for research on simulation
methods for aircraft power systems.

References
[1] R. Newman, "The more electric engine concept," Society of Automotive Engineers, 2004.
[2] A. Eid, H. El-Kishky, M. Abdel-Salam, and M.T. El-Mohandes, "On Power Quality of Variable-Speed
Constant-Frequency Aircraft Electric Power Systems," IEEE Transactions on Power Delivery, vol. 25, no. 1,
pp. 55-65, January 2010.
[3] A. Eid, H. El-Kishky, M. Abdel-Salam, and T. El-Mohandes, "Power quality investigations of VSCF aircraft
electric power systems," in 2010 42nd Southeastern Symposium on System Theory (SSST), Tyler, TX, USA,
2010, pp. 171-176.
[4] T. Jomier, "Final public MOET Technical Report," European Commission 6th Framework Programme,
Toulouse, France, Technical Report D0.02.3, 2009.
[5] NASA. (2010, November) Dryden Flight Research Center. [Online].
http://www.nasa.gov/centers/dryden/home/index.html
[6] M.A. Lemaire, L. Montealegre Lobo, J. Mahseredjian, F. Sirois, and C. Lavoie, "Simulation of aircraft
electrical power systems: tools and challenges," in Canadian Aeronautics and Space Institute AERO 2011,
Montreal, Canada, 2011.
[7] S. Cavalcanti, and M. Papini, "Preliminary Model Matching of the EMBRAER 170 Jet," Journal of Aircraft,
vol. 41, no. 4, pp. 703-710, 2004.
[8] Simulink, "User's Guide," The Math Works Inc., Version 7.4 November 2012. [Online].
http://www.mathworks.com/products/simulink/
[9] C. Dufour, L-A. Grégoire, and J. Bélanger, "Solvers for Real-Time Simulation of Bipolar Thyristor-Based
HVDC and 180-cell HVDC Modular Multilevel Converter for System Inteconnection and Distributed Energy
Integration," in 2011 CIGRE Conference Proceedings, Recife, Brazil, April, 2011.
[10] Jean Mahseredjian, S. Dennetière, L. Dubé, B. Khodabakhchian, and L. Gérin-Lajoie, "On a new approach
for the simulation of transients in power systems," Electric Power Systems Research, vol. 77, no. 11, pp.
1514-1520, September 2007.
[11] C. Dufour, J. Mahseredjian, and J. Belanger, "A Combined State-Space Nodal Method for the Simulation of
Power System Transients," IEEE Transactions on Power Delivery, vol. 26, no. 2, pp. 928-935, April 2011.
[12] Department of Defense, "Aircraft Electric Power Characteristics," United State Department of Defense,
Washington, USA, Military Standard MIL-STD-704f, 2004.
[13] Department of Defense, "Analysis of Electric Load and Power Source Capacity," United State Department
of Defense, Washington, USA, Military Standard MIL-E-7016, 1976.
[14] J. Faucher, Aeroconseil Aircraft Eng, “Simulation Study of new Aircraft Electrical Power Network
performances”, MOET Forum, Barcelona, Sept 2009, http://www.eurtd.com/moet/PDF/B.pdf
 Paper presented at the EPE'13 ECCE Europe conference, September 3-5, 2013, Lille, France

[15] J.C. Swierczek, F. Moller, C. Saudemont, R. Meuret, and B. Robyns, "Power Management of a
Regenerative Local HVDC Aircraft Network Using Supercapacitors," in Proceedings of the 15th
International Power Electronics and Motion Control Conference, EPE-PEMC 2012 ECCE Europe, Novi
Sad, Serbia, 2012, pp. LS4a.5
[16] Bombardier Aerospace, "ATA 24, Technical Training Guide," Montreal, Canada, Technical Report 2003.
[17] S.J. Exner, "Impedance Data for 400-Cycle Aircraft Distribution Systems" AIEE, Technical Paper 1952.
[18] K.W.E. Cheng, "Comparative study of AC/DC converters for more electric aircraft," in Proceedings of the
1998 7th International Conference on Power Electronics and Variable Speed Drives, London, UK, 1998, pp.
229-304.
[19] F. Tinney and W.S. Meyer, "Solution of Large Sparse Systems by Ordered Triangular Factorization," IEEE
Transactions on Automatic Control, vol. 18, no. 4, pp. 333 - 346, August 1973.

View publication stats