Towards the Implementation of a Digital Twin

for Induction Motors

Adamou Amadou Adamou(B) and Chakib Alaoui

INSA Euro-Méditerranéen, EUROMED University, Fes, Morocco

[email protected], [email protected]

Abstract. Energy Efficiency (EE) is one of the main pillars of Industry 4.0. The
road map of this new industrial revolution has imposed a certain degree of accuracy
for conditions monitoring of industrial machines, where Induction Motors (IMs)
represent the most important machines in terms of energy consumption (42.2% of
the global energy demand). Achieving an accurate EE monitoring for IMs imposes
a good Stray Load Loss (SLL) estimation by taking into consideration the In-situ
aspect. In this paper, the five-dimension digital twin (DT) model is used to propose
a new approach to integrate the EE for DT. The proposed method is the first step
toward DT implementation for IM. Three state-of-the-art SLL models for IM are
compared and the most accurate expression to integrate into the EE model for the
proposed DT is then selected. The selected model takes into consideration all SLL
sources in IM represented in the Electric Equivalent Circuit (EEC) by the stator
and the rotor leakage inductances and the magnetizing inductance.

Keywords: Industry 4.0 · Digital Twin · Energy efficiency · Induction motor ·

Stray load losses

1 Introduction
One of the most important challenges in today’s environment and Industry 4.0 is energy
efficiency with low cost, where the energy management systems (EMS) allow to control
and reduce the energy consumption of industrial systems without affecting their oper-
ational performance. According to the International Energy Agency (IEA) [1], 53% of
electrical energy consumption goes to Electric Drives (ED) or Electric Motor-Driven
Systems (EMDS). Among the EMDS, 77% of the electrical energy consumption is due
to medium-sized motors, where 80% are Induction Motors of different types [2]. The use
of a large quantity of IMs makes them a good candidate for energy efficiency improve-
ments. The medium-sized IMs of 84% of energy efficiency consume 10,500 TWh per
year, so if a mere one percent (1%) of EE is improved during the operation, 105 TWh
could be saved [3]. The efficiency of IMs has emerged as having the biggest potential in
energy saving [4]. Improving the efficiency of IMs cannot be achieved without accurate
estimation of all losses of said motors, and in particular, the SLL. Beyond the design
stage, in Industry 4.0, the operation condition of IM implies an accurate determination
of efficiency, hence stray load losses.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022

S. Motahhir and B. Bossoufi (Eds.): ICDTA 2022, LNNS 454, pp. 513–523, 2022.
https://doi.org/10.1007/978-3-031-01942-5_51
514 A. A. Adamou and C. Alaoui

The SLL has been the subject of various discussions, since 1912 with E. M. Olin [5]
who suggested accounting for the stray load losses empirically relating no-load to load
losses. Until today, the discussion on these losses is still on-going because it is confusing
origins [6]. According to [5], the efficiency estimation methods of IM as the input-output
method are precise but difficult to perform to obtain an accurate result. While the loss
segregation method is easy to perform accurately for the conventional losses, the overall
result, taking into account only these conventional losses (see Eq. 4) is less precise. The
difference between the results of these two methods was quickly associated with the
stray losses. The harmonics losses, also called high frequency stray no-load and stray
load-losses, are related most closely to the loading operation of a machine, hence they are
called stray load losses or load loss or unknown losses [5, 8]. Despite all disagreement,
there is a consensus that the stray load losses are defined as the difference between the
total power losses of the machine on load and the losses determined by the segregation
method which are the conventional losses (see Eq. 1) [5–7].
Digital twin (DT) is one of the most capable enabling technologies for realizing
smart manufacturing and Industry 4.0. They are characterized by the seamless integration
between the cyber and physical spaces. The importance of DTs is increasingly recognized
by research institutions and the industrial sector [10]. The condition monitoring using DT
depends on the study goals such as design, production, manufacturing, and maintenance
[11]. The stray loss models are suggested and the most accurate is then selected as a
candidate to improve the DT implementation of IM. Among the energy-saving methods,
Energy-based predictive maintenance is one of the intuitive approaches that we can use.
This is clearly shown in [12], where the authors propose a new approach to predict and
classified the Squirrel Cage Induction Motor failure based on only energy consumption
data using machine learning techniques. The work is based on the energy consumption
data (current, voltage, and time) which correlated with faulty SCIM motors vibration
data regarding torque and speed.

2 Digital Twin of Induction Motor – Towards the Implementation

of Stray Losses
2.1 Definition and Description

Several studies have defined the concept of a digital twin in different ways, although
sometimes they are incomplete with a lack of information [11, 13]. In [14] DT is defined
as “an integrated multi-physics, multi-scale, and probabilistic simulation of a complex
product and uses the best available physical models, sensor updates, etc., to mirror the
life of its corresponding twin”. It is the most common definition of DT and it is similar to
the definition dictated in [15]. In [10] the simulation aspect of the digital twin is defined
as follows.
 Towards the Implementation of a Digital Twin for Induction Motors 515

Maintenance
Pronosc Health Management (PHM)
Lifecycle opmizaon
Process/logisc/producon
Design

0 10 20 30 40 50 60 70 80

Fig. 1. Digital Twin applications [11]

The Digital Twin refers to a description of a component, product, or system by a set

of well-aligned executable models with the following characteristics.

a. DT is the linked collection of the relevant digital artifacts including engineering

data, operation data, and behavior descriptions via several simulation models. The
simulation models making up the DT are specific for their intended use and apply
suitable fidelity for the problem to be solved.
b. DT grows along with the real system along the whole life cycle and integrates the
currently available knowledge about it.

DT presents many advantages to the modern industry, such as reduced production

downtime; reduced breakdowns; cost savings; improved productivity; elimination of
ambiguity in maintenance tasks; expended equipment life; improved customer service
and reputation; reduced energy waste; and improved facility security [11]. These advan-
tages may have different impacts on the various sectors, so the application domain of
the DT is categorized into five groups (see Fig. 1) where maintenance is the dominant
application of the DT.

2.2 The Proposed EE Model-Based Digital Twin

Choosing the right technologies and tools to implement DT is “la bête noire” of many
researchers and professionals in engineering. Fortunately, the 5-dimension digital twin
model allows for the implementation of this technology based on 5 keys [16, 17]. This
model is constituted of Physical entities, Virtual model, DT data, Services in DT, and
Connection. Based on this model, the proposed EE-based DT can be formulated as:

MDT = f (HP, SP, DD, Ss, Com) (1)

where HP is the hardware part, SP is the software part, DD is the data, Ss are the services
and Com is the communication part.

Hardware Part: The physical motors installed in the industry and the smart sensors
constitute the hard part of the physical system. At first, the EEC parameters are identified
w.r.t the used laboratory motor using the algorithm established in [18]. These parameters
are then integrated into the EE model to reduce the number of unknown parameters. The
remaining parameters are then found using a minimization algorithm.
516 A. A. Adamou and C. Alaoui

Software Part: It is the virtual environment (Digital Twin) that reproduces the defined
behavior of the physical system. The EE model is implemented and run by the soft part
as an RT simulation model. The Hard part information, the local and support services
are the entities that supply the soft part as inputs, and the results of the RT simulation
are shown by the HMI.
Services: The management of the DT ecosystem needs services to operate, these are
local and external (support) services from the partners. The local services are carried
out in the internal system by the company’s collaborators, these can be the applications
simulation and condition monitoring for EE, PHM, predictive maintenance, etc.
Data: This is the essential part of DT. It is the DT data server that is composed of
heterogeneous, multi-sources, and multi-scale data from the previous three entities. The
hard part provides static and dynamic data from the EEC to measure voltage, current,
and temperature.
Communication Part: The information exchange in the DT system is managed by
the communication part with help of the IoT technologies. It illustrates these iterations
(Fig. 2).

Services

Support services Local services

Softwares, AI EE Monitoring
algorithms, etc. PHM
Predictive maintenance

Physical In-Service IMs Virtual model

DT data server

HMI

Legend:
: Smart sensor networks
: Wireless communication
: Connection between the entities ( , , , )
Fig. 2. The proposed DT system is based the on EE model.

3 Stray Losses Model of IM

The stray load losses are the only losses in induction motor where researchers are not
unanimous on their characteristic aspects. In [19], Jimoh et al. detail 8 methods to eval-
uate the SLL, which are classified into 2 categories, full-load tests, and light-load tests.
 Towards the Implementation of a Digital Twin for Induction Motors 517

Among methods proposed in the previous paper, the Full-Load Tests are inaccurate and
non-economical. The Input-Output, Pump Back and Calorimetric methods are imprac-
ticable for a high efficient machine. However, the Light-Load Test methods are simple,
economical, and indirect, based on too many assumptions, which have generated consid-
erable controversy [19]. As mentioned previously, the origin of SLL can be divided into
three sources: Leakage flux, space, and the time harmonics in the Air-gap flux density.
According to the literature, there are 4 ways to evaluate the SLL which are, Measure-
ment, Analytical/Numerical methods, Equivalent Circuit, and Empirical relations [3].
Among these methods, EEC, Analytical and Empirical relations are suitable for In-Situ
EE monitoring. Equations (2) to (4) show these losses.

PStray = PTotalloss − PConventional (2)

PTotalloss = PElectrical − PMechanical (3)

PConventional = PJs + PJr + PIron + PMec (4)

3.1 Model-Based on EEC

Several models have been proposed to evaluate the SLL based on the equivalent circuit,
the first model with a modified EEC including SLL resistance was proposed by Lamine
et al. in [25]. The approach used can be divided into two categories: Some evaluate the
SLL based on the stator and rotor leakages inductances, and the magnetizing inductance
in the equivalent circuit which represent the influence of slots on the magnetic field [3,
26], Others evaluate the SLL by adding resistance as an SLL resistance [9, 27] (Fig. 3).

Fig. 3. T-model equivalent circuit of IM [3]

SLL Model Based on Leakage and Magnetizing Inductances. Rajendra et al. [3]
proposed in their paper a new SL model for grid-connected IM which is judged accurate.
This model can also be used for the design stage. The model is based on 11 sources (Table
I in [3]) of SL which are represented by three inductances as stator leakage inductance
(Lls ), rotor leakage inductance (Llr ), and the magnetizing Inductance (Lm ). Find by using
blocked rotor test and no-load test. The model is developed, as follows.
518 A. A. Adamou and C. Alaoui

Model Formulation. Depending on the leakage flux, the SLL is directly proportional to
the LIS and LIr , and it’s also inversely proportional to Lm [26]. Based on these obser-
vations, eight (8) models were proposed in [26] and six (6) models in [3] to evaluate
the SLL. Each study selects an accurate model based on certain criteria. In the first
paper [26], standard deviation (S.D) of the errors and correlation coefficients (r) are
calculated as KPIs to identify the most appropriate model (Eq. 5). In the 2nd paper [3],
Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) are used
to select the suitable model (Eq. 6) for their relative capability of SLL representation.
The model is based on the result reported by many papers that explained that the SLL
is proportional to the output power (1% to 4% of the output power according to [28]),
hence SLL = KSL Pout .
 m  n
LlS Llr
KSL1 = ( ) ( ) .pf (5)
Lm Lm
LlS m Llr n
KSL2 = ( ) ( ) (6)
p.Lm p.Lm
So the SLL models from the equations above are F-1: KSL1 Pout and F-2: KSL2 Pout .
Accurate expressions of inductances have been expressed in [29] taking into con-
sideration the leakage and the harmonics sources of SLL. The inductances expressions
as reported in [3] are calculated using the procedure laid in IEC 60034-2-1 and the test
data given by Toshiba Mitsubishi-Electric Industrial Systems Corporation (TMEIC).

SLL Model Based on Additional Resistances. These models involve additional resis-
tances in the EEC to represent the SLL. In [27], Nasir proposed a new EEC including
the Stray load and no-load losses by adding additional resistances called stray loss resis-
tances, in parallel with the stator and rotor leakage reactances. In the usual EEC, the iron
core resistances (stator and rotor) are not represented, the proposed circuit [27], includes
these resistances to model the SLL in the stator and rotor circuit. These are expressed in
(7) and (8).

XlS2 .RSiC XlS2 .RS XlS2 .RS

RSSSL = + ≈ (7)
R2SiC + XlS2 R2S + XlS2 R2S + XlS2
S 4 .XlS2 S 2 .XlS2 S 2 .XlS2
RRSSL = + ≈ (8)
RSiC Rr Rr

Hence the losses are calculated as: Stator: PSSL = 3.RSSSL .IS2 ; Rotor: PSSL =
3.RRSSL .Ir2 .
However, in [9], Boglietti et al. proposed an EEC including the SLL due to the effects
of the spatial harmonic in IM. Theses model is not convenient for our search because it
considers only the additional losses due to the rotor cage currents induced by the air gap
spatial harmonics. As performed in [21] by the sensitivity analysis, it is imperative to
have the input power, the load torque, and the input current measured with high accuracy
to guarantee a good estimation for the SLL.
 Towards the Implementation of a Digital Twin for Induction Motors 519

4 Discussion
4.1 Comparison of Three States of the Art SLL Model
The stray loss model proposed in the previous papers by Rajendra et al. [26] and [3]
include the leakage flux and the magnetizing reactance’s, the difference is due on the
pole pair which appear in the model (3), and the power factor which appear in the
model (2). It’s important to note that the exponent’s m and n are calculated by regression
analysis with the model and the SLL test data. At a steady-state, the efficiency can
be expressed as η = (ratedTorque)×(ratedspeed
3UI ×pf
)
, we know that the synchronous speed is
inversely proportional to the pole pair (ns = 120f s
p ) which imply that the pole pair (p) is
linked to the pf . Hence, we can conclude that the expressions (Eqs. (5) and (6)) above
have the same parameter dependent (Table 1).

Table 1. Comparison of different SLL models.

SLL model Include leakage flux Include space Include time-harmonic

harmonic
KSL1 Pout YES YES YES
KSL2 Pout YES YES YES
3.(RSSSL .IS2 + YES NO NO
RRSSL .Ir2 )

However, the model proposed by Nasir in [27] doesn’t include the harmonics due to
the Air Gap flux, the resistances which modeled the SLL represent only the leakage flux
caused by the stator and rotor windings.
A comparison is proposed to select the best model between the models from model
F-1 or F-2. Matlab is used to draw the SLL measured for 14 motors based on IEC-
60034-2 standard, and the errors of the calculated SLL from the Eqs. (5 and 6) w.r.t
the measured SLL (see Fig. 4). Standard deviation (SD) of the errors in calculated SLL
from F-1 and F-2 w.r.t measured SLL are evaluated to be respectively 19.45446424 and
9.57042857. The most accurate model is the one with the lowest SD, hence the selected
model is F-2.

100.00 4

50.00 2

0.00 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14

SLL (measured) % error F-1 % error F-2

Fig. 4. SLL errors analysis with F-1 and F-2 (data provided from [3, 26])
520 A. A. Adamou and C. Alaoui

4.2 Towards the EE Monitoring Based on DT

To achieve the proposed DT implementation, a first step of the process is illustrated (see
Fig. 5). On a lab scale, the multi-physics model (including all the stator and rotor losses)
outputs are used to run the 3D model which highlights the output magnitudes (Torque
and speed). A virtual sensor is used to measure the output power and then combines with
the input power to measure the EE of the machine. The way to measure virtually the
internal temperature and the vibration from different measured points can be envisaged
using the IM thermal model and the Finite Element Analysis (FEA) methods.

L1
Multiphysic-based EE modelling
L2 EE model = f(V, I, T)
of IM (including SLL)
L3

Voltage, Current
Temperature TCP/IP
RT SIMULATOR

Workstation Oscilloscope

Legend:
: EE model based on multiphysics analysis
RT SIMULATOR : Digital RT Simulator from OPAL-RT Technologies
: Connection between entities
Fig. 5. The first step process towards the proposed DT model (lab test bench)

S. Bouzid et al. describe in their paper the development of RT DT of a wound rotor IM

based on the finite element method. In this work [31], the RT outputs of electromagnetic
quantities otherwise difficult to evaluate are computed using DT.

5 Conclusion

Stray load losses implementation in Induction Motor Digital Twin application is high-
lighted in this paper. Beyond the SLL measurement methods and models presented, the
best model is selected for DT-based EE implementation.
This study is the first step towards an accurate EE model establishment for DT in
Industry 4.0. Hence the next works can be directed towards the establishment of the
so-called EE model, then, based on these models, the energy efficiency improvement
methods during the operation can be defined. From the mentioned works, energy-based
predictive maintenance is one of the intuitive methods to improve the EE during the
operation. And a co-simulation with virtual Reality to the development of the proposed
 Towards the Implementation of a Digital Twin for Induction Motors 521

DT can also improve the system as user-friendly as shown in [32]. In terms of the per-
spective, it is important to note that the historical EE data of the machine during operation
allows to development of very good methods for energy consumption forecasting using
machine learning (ML) methods. According to [33], among the ML algorithms such as
Linear Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), Artificial
Neural Networks (ANN), and Recurrent Neural Networks (RNN) linear regression meth-
ods were compared, the linear regression shown the most accurate result for industrial
energy consumption forecasting.

References
1. 4E_2017_Annual_Report_100518.pdf. https://www.iea-4e.org/wp-content/uploads/public
ations/2018/05/4E_2017_Annual_Report_100518.pdf
2. Waide, P., Brunner, C.U.: Energy-Efficiency Policy Opportunities for Electric Motor-Driven
Systems. OCDE, Paris (2011). https://doi.org/10.1787/5kgg52gb9gjd-en
3. Kumar, R., Kumar, P., Kanekawa, T., Oishi, K.: Stray loss model for induction motors with
using equivalent circuit parameters. IEEE Trans. Energy Convers. 35, 1036–1045 (2020).
https://doi.org/10.1109/TEC.2020.2964616
4. Mzungu, H.M., Sebitosi, A.B., Khan, M.A.: Comparison of standards for determining losses
and efficiency of three-phase induction motors. In: 2007 IEEE Power Engineering Society
Conference and Exposition in Africa - PowerAfrica, Johannesburg, South Africa, pp. 1–6.
IEEE (2007). https://doi.org/10.1109/PESAFR.2007.4498051
5. Olin, E.M.: Determination of power efficiency of rotating electric machines: summation of
losses versus input-output tests. Proc. Am. Inst. Electr. Eng. 31, 1287–1310 (1912). https://
doi.org/10.1109/PAIEE.1912.6660632
6. Wang, J., et al.: Stray load loss calculation for induction motor by combination of general
airgap field modulation theory and 2D FEA. IEEE Trans. Energy Convers. 36, 2524–2533
(2021). https://doi.org/10.1109/TEC.2021.3052365
7. International Standards for the Induction Motor Efficiency Evaluation A Critical Analysis of
the Stray-Load Loss Determination.pdf
8. W. Novotny, D.: High-frequency losses in induction motors. University of Wisconsin,
Department of Electrical and Computer Engineering, Wisconsin, USA (2021)
9. Boglietti, A., Cavagnino, A., Ferraris, L., Lazzari, M.: Induction motor equivalent circuit
including the stray load losses in the machine power balance. IEEE Trans. Energy Convers.
23, 796–803 (2008). https://doi.org/10.1109/TEC.2008.921467
10. Boschert, S., Rosen, R.: Digital twin—the simulation aspect. In: Hehenberger, P., Bradley, D.
(eds.) Mechatronic Futures, pp. 59–74. Springer, Cham (2016). https://doi.org/10.1007/978-
3-319-32156-1_5
11. Errandonea, I., Beltrán, S., Arrizabalaga, S.: Digital Twin for maintenance: a literature review.
Comput. Ind. 123, 103316 (2020). https://doi.org/10.1016/j.compind.2020.103316
12. Laayati, O., Bouzi, M., Chebak, A.: Smart energy management system: SCIM diagnosis
and failure classification and prediction using energy consumption data. In: Motahhir, S.,
Bossoufi, B. (eds.) ICDTA 2021. LNNS, vol. 211, pp. 1377–1386. Springer, Cham (2021).
https://doi.org/10.1007/978-3-030-73882-2_125
13. Grieves, M., Vickers, J.: Digital twin: mitigating unpredictable, undesirable emergent behav-
ior in complex systems. In: Kahlen, F.-J., Flumerfelt, S., Alves, A. (eds.) Transdisciplinary
Perspectives on Complex Systems, pp. 85–113. Springer, Cham (2017). https://doi.org/10.
1007/978-3-319-38756-7_4
522 A. A. Adamou and C. Alaoui

14. Tao, F., Cheng, J., Qi, Q., Zhang, M., Zhang, H., Sui, F.: Digital twin-driven product design,
manufacturing and service with big data. Int. J. Adv. Manuf. Technol. 94(9–12), 3563–3576
(2017). https://doi.org/10.1007/s00170-017-0233-1
15. Lermer, M., Reich, C.: Creation of digital twins by combining fuzzy rules with artificial
neural networks. In: IECON 2019 - 45th Annual Conference of the IEEE Industrial Electron-
ics Society, Lisbon, Portugal, pp. 5849–5854. IEEE (2019). https://doi.org/10.1109/IECON.
2019.8926914
16. Mukherjee, V., Martinovski, T., Szucs, A., Westerlund, J., Belahcen, A.: Improved analytical
model of induction machine for digital twin application. In: 2020 International Conference
on Electrical Machines (ICEM), Gothenburg, Sweden, pp. 183–189. IEEE (2020). https://
doi.org/10.1109/ICEM49940.2020.9270916
17. Qi, Q., et al.: Enabling technologies and tools for digital twin. J. Manuf. Syst. 58, 3–21 (2021).
https://doi.org/10.1016/j.jmsy.2019.10.001
18. Pedra, J., Córcoles, F.: Estimation of induction motor double-cage model parameters from
manufacturer data. IEEE Trans. Energy Convers. 19, 8 (2004)
19. Jimoh, A.A., Findlay, R.D., Poloujadoff, M.: Stray losses in induction machines: part I,
definition, origin, and measurement. IEEE Trans. Power Appar. Syst. PAS 104, 1500–1505
(1985). https://doi.org/10.1109/TPAS.1985.319165
20. Cao, W.: Comparison of IEEE 112 and new IEC standard 60034-2-1. IEEE Trans. Energy
Convers. 24, 802–808 (2009). https://doi.org/10.1109/TEC.2009.2025321
21. Boglietti, A., Cavagnino, A., Lazzari, M., Pastorelli, M.: International standards for the induc-
tion motor efficiency evaluation: a critical analysis of the stray-load loss determination. In:
38th IAS Annual Meeting on Conference Record of the Industry Applications Conference,
2003, vol. 2, pp. 841–848 (2003). https://doi.org/10.1109/IAS.2003.1257626
22. Boglietli, A., Cavagnino, A., Lazzari, M., Pastorelli, A.: Induction motor efficiency measure-
ments in accordance to IEEE 112-B, IEC 34-2, and JEC 37 international standards. In: IEEE
International Electric Machines and Drives Conference, IEMDC 2003, Madison, WI, USA,
pp. 1599–1605. IEEE (2003). https://doi.org/10.1109/IEMDC.2003.1210664
23. Pillay, P., Al-Badri, M., Angers, P., Desai, C.: A new stray-load loss formula for small and
medium-sized induction motors. IEEE Trans. Energy Convers. 31, 1221–1227 (2016). https://
doi.org/10.1109/TEC.2016.2539959
24. Lu, B., Habetler, T.G., Harley, R.G.: A nonintrusive and in-service motor-efficiency estimation
method using air-gap torque with considerations of condition monitoring. IEEE Trans. Ind.
Appl. 44, 1666–1674 (2008). https://doi.org/10.1109/TIA.2008.2006297
25. Lamine, A., Levi, E.: Dynamic induction machine modeling considering the stray load
losses. In: 39th International Universities Power Engineering Conference, UPEC 2004, vol.
1, pp. 582–586 (2004)
26. Kumar, R., Kumar, P.: Modelling of stray-load loss for medium power induction motors. In:
IECON 2018 - 44th Annual Conference of the IEEE Industrial Electronics Society, pp. 571–
576 (2018). https://doi.org/10.1109/IECON.2018.8591650
27. Nasir, B.A.: Modeling of stray losses in the equivalent circuit of induction machines. In: AIP
Conference Proceedings, vol. 2307, p. 020006 (2020). https://doi.org/10.1063/5.0032902
28. Agamloh, E.B.: An evaluation of induction machine stray load loss from collated test results.
IEEE Trans. Ind. Appl. 46, 2311–2318 (2010). https://doi.org/10.1109/TIA.2010.2070474
29. Boldea, I., Nasar, S.A.: The Induction Machines Design Handbook, 2nd edn. CRC Press,
Boca Raton (2020)
30. Use of magnetic circuits in leakage reactance calculations. In: Introduction to AC Machine
Design, pp. 125–191. Wiley, Hoboken (2017). https://doi.org/10.1002/9781119352181.ch4
31. Bouzid, S., Viarouge, P., Cros, J.: Real-time digital twin of a wound rotor induction machine
based on finite element method. Energies 13, 5413 (2020). https://doi.org/10.3390/en1320
5413
 Towards the Implementation of a Digital Twin for Induction Motors 523

32. Havard, V., Jeanne, B., Lacomblez, M., Baudry, D.: Digital twin and virtual reality: a co-
simulation environment for design and assessment of industrial workstations. Prod. Manuf.
Res. 7, 472–489 (2019). https://doi.org/10.1080/21693277.2019.1660283
33. Bahij, M., Labbadi, M., Cherkaoui, M., Chatri, C., Lakrit, S.: A comparison study of machine
learning methods for energy consumption forecasting in industry. In: Motahhir, S., Bossoufi,
B. (eds.) ICDTA 2021. LNNS, vol. 211, pp. 165–175. Springer, Cham (2021). https://doi.org/
10.1007/978-3-030-73882-2_16