Skip to content
BY-NC-ND 4.0 license Open Access Published by De Gruyter Open Access March 8, 2018

Thermal constitutive matrix applied to asynchronous electrical machine using the cell method

Pablo Ignacio González Domínguez EMAIL logo , José Miguel Monzón-Verona , Leopoldo Simón Rodríguez and Adrián de Pablo Sánchez
From the journal Open Physics

Abstract

This work demonstrates the equivalence of two constitutive equations. One is used in Fourier’s law of the heat conduction equation, the other in electric conduction equation; both are based on the numerical Cell Method, using the Finite Formulation (FF-CM). A 3-D pure heat conduction model is proposed. The temperatures are in steady state and there are no internal heat sources. The obtained results are compared with an equivalent model developed using the Finite Elements Method (FEM). The particular case of 2-D was also studied. The errors produced are not significant at less than 0.2%. The number of nodes is the number of the unknowns and equations to resolve. There is no significant gain in precision with increasing density of the mesh.

1 Introduction

The Heat Transmission Equation in FF-CM (consult [1, 3, 5]) is expressed in Eq. (1) as:

GTMλGτ+MCpdτdt=W(1)

Two constitutive matrices are observed: [Mλ] and [MCp]. The constitutive matrix [Mλ] belongs to Fourier’s law of the heat conduction equation. The constitutive matrix [MCp] belongs to the Heat transmission equation with change of temperature.

Fourier’s law of the heat conduction equation applied to stator-airgap-rotor is:

GTMλGτ=D~qλ(2)

The heat sources, corresponding to Joule effects and magnetic hysteresis and eddy currents, are represented as the vector [W].

The heat produced by mechanical friction is not studied in this work.

The changes in the internal energy, dependent on the component materials of the asynchronous machines, are represented in the following expression:

MCpdτdt(3)

We establish an analogy between thermal conduction and electrical conduction as is indicated in Eq. (4).

GTMλGτ=D~qλGTMσGv=D~I(4)

The Fourier heat transmission equation is treated in 2-D and 3-D. The cartesian absolute coordinates of the cell are converted to local standardized coordinates. The origin is one of the four vertices of the tetrahedron used as cell [1,2,3]. The dual variables are projected to the primal using geometrical methods. Planar symmetries and axisymmetric symmetries are used. Quadratic interpolations are applied to obtain distributions of temperature [1,2,3,4, 5].

The main parameters used in this work are the following:

Assuming the analogy indicated in Eq. (4) gives:

MλMσ(5)

SymbolNameunit
D, Primal and dual discrete divergence operators.-
dτdtTemperature variation.°C·s–1
G, Primal and dual discrete gradient operators.-
IElectric current.A
[Mλ]Constitutive matrix in Fourier’s Law of Heat ConductionW·K–1
[MCp]Constitutive matrix in heat transition in transitory state.J·K–1
[Mσ]Constitutive matrix in electrical conduction.Ω–1
[qλ]Heat flow crossing a surface.W
S, Primal and dual surfaces.m2
VTTetrahedron volume.m3
V, Primal and dual volumes.m3
[v]Electrical potential matrix.V
[W]Heat generated during a process.W
λkVolumetric thermal conductivity.W·K–1·m–1
σElectrical conductivity.Ω–1·m–1

Using the expression proposed by [6], the matrices exposed in Eq. (5) are topological equivalent:

λ1VtS~iS~jσ1VtS~iS~j(6)

Where matrix [S] is defined according to (8), following Specogna and F. Trevisan [6],

[S]6×6=S~0S~0S~0S~5S~2S~2S~5S~0S~5S~5(7)

is the dot product of the vectors that define the dual faces of the cell and (VT) is the volume of the tetrahedron, see Figure 1.

Figure 1 Dual faces vectors of a tetrahedrical cell with baricentric
Figure 1

Dual faces vectors of a tetrahedrical cell with baricentric

The matrix [Mσ] must be calculated for the electro-thermal phenomena occurring in the asynchronous machine. So why do the calculations twice? Several numerical experiments were designed to confirm this assumption.

The obtained results were contrasted with the results obtained from the same model developed with Finite Element Methods (FEM). The particular caseof 2-D was also studied.

2 Numerical experiments

2.1 Numerical experiments description

In the numerical experiments, pure conduction of heat is assumed, there are no inner heat sources, and temperatures are stationary, as shown in Eq. (8).

MλGτ=qλ(8)

Two boundary conditions, of type Dirichlet, were established. Two temperatures, τout = 30 °C and τin = 80 °C were fixed on two faces of a tube, which represents the rotor or stator of an electric machine.

These faces are axial sides to thetube. The other top face and down face of the tube were considered as perfect heat insulators.

The temperatures are considered in steady state. There are no internal heat sources. Various structural materials with different volumetric thermal conductivity coefficients were reviewed. The volumetric thermal conductivity is λ1 = 50 Wm–1K–1 for the inner tube and λ2 = 193 Wm–1K–1 for the outer one. As there is no heat source inside the domain, the heat flows in a radial direction.

Temperatures are measured in a parametric cut line as shown in Figure 3.

An example square in 2-D, is shown in Figure 2, with sides of 200 cm and a superficial thermal conductivity coefficient equal to 1 WK–1. The rectangular perforations have assigned a temperature τin = 150 °C to their edges. The outer square edge has a temperature τout = 50 °C.The temperatures are considered in steady state.

Figure 2 (a) Square 2-D with three holes. Dimensions are in meters. (b) FF-CM model. (c) FEM model
Figure 2

(a) Square 2-D with three holes. Dimensions are in meters. (b) FF-CM model. (c) FEM model

Figure 3 Tube, as stator of an electric machine, with two different structural materials, also two different thermal conductivity coefficients
Figure 3

Tube, as stator of an electric machine, with two different structural materials, also two different thermal conductivity coefficients

Temperatures are measured in a horizontal parametric cut line at 1 meter from the base of the square, see Figure 2(a).

2.2 Results

Different meshes have been used in the model and the temperatures calculated with FF-CM. The meshing and the post-processing of the model in CM have been performed with Gmsh software [7]. The resulting temperatures are compared with the temperatures obtained with a 2-D axisymmetric FEM model using the software FEMM [8]. For a tube in 3-D, the results are showed in Figures 4 and 6.

Figure 4 Temperatures in a tube with two thermal conductivity
Figure 4

Temperatures in a tube with two thermal conductivity

For a square in 2-D, with one superficial thermal conductivity coefficient, the results are showed in Figure 5.

Figure 5 Temperatures in square 2-D with three holes
Figure 5

Temperatures in square 2-D with three holes

The difference between the normal distribution and the distribution obtained in the error histogram is due to the measurement method (see Figure 6). A parametric cutline is used in the FF-MC model and in the FEMM model, too. Although both parametric cut lines are made at the same distance, the meshes do not match exactly.

Figure 6 (a) 2-D axisymmetric FEM model using the software FEMM of a stator with two different structural materials, also two different thermal conductivity coefficients. (b) Errors histogram when are compared FF-CM model with FEM model
Figure 6

(a) 2-D axisymmetric FEM model using the software FEMM of a stator with two different structural materials, also two different thermal conductivity coefficients. (b) Errors histogram when are compared FF-CM model with FEM model

Unfortunately, the interpolations between the values of the nodes of the mesh and the values in the points of the cutline disagree. This is because the mesher of the GMSH is not the same as that of the FEMM.

3 Conclusions

The FF-CM is a strong method because we face a 3-D model in FF-CM, with a not very dense tetrahedral mesh, with a planar axisymmetric model calculated with FEM and very dense triangular mesh. The errors produced are not significant at less than 0.2%. The number of nodes is the number of unknowns and equations to resolve. There is no significant gain in precision with increasing density of the mesh. In the Figure 4, with 9,679 nodes and 13,136 nodes the errors are similar.

Acknowledgement

The authors thank the project TEC2014-60527-C2-1-R “SURF: Sistema Subacuático RFID para Acuicultura” from Ministerio de Economía y Competitividad (MINECO) for their contribution to the realization of this work.

References

[1] Bullo M., D’Ambrosio V., Dughiero F., Guarnieri M., Coupled electrical and thermal transient conduction problems with a quadratic interpolation cell method approach, Magnetics, IEEE Transactions on, 2006, 42(4), 1003-1006.10.1109/TMAG.2006.872471Search in Google Scholar

[2] Bullo M., D’Ambrosio V., Dughiero F., Guarnieri M., A 3D Cell Method Formulation for Coupled Electric and Thermal Problems, b). Miami, FL, USA, IEEE, 2006, 7-7.10.1109/CEFC-06.2006.1632800Search in Google Scholar

[3] Bullo M., D’Ambrosio V., Dughiero F., Guarnieri M., A 3-D Cell Method Formulation for Coupled Electric and Thermal Problems, Magnetics, IEEE Transactions on, 2007, 43(4), 1197-1200.10.1109/TMAG.2006.890955Search in Google Scholar

[4] Tonti E., Formulazione finita delle equazioni di campo: Il Metodo delle Celle, Atti del XIII Convegno Italiano di Meccanica Computazionale, Brescia, Italy, 2000.Search in Google Scholar

[5] Tonti E., A direct discrete formulation of field laws: The cell method, CMES-Computer Modeling in Engineering and Sciences, 2001, 2(2), 237-258.Search in Google Scholar

[6] Specogna R., Trevisan F., Discrete constitutive equations in A-Chi geometric eddy-current formulation, IEEE Trans. Magn, 2005, 41(4), 1259-1263.10.1109/TMAG.2005.844841Search in Google Scholar

[7] Geuzaine C., Remacle J. F., Gmsh: a three-dimensional finite element mesh generator with built-in pre-and post-processing facilities, International Journal for Numerical Methods in Engineering, 2009, 79(11), 1309-1331.10.1002/nme.2579Search in Google Scholar

[8] Meeker D., http://www.femm.info/wiki/HomePageSearch in Google Scholar

Received: 2017-11-01
Accepted: 2017-11-24
Published Online: 2018-03-08

© 2018 Pablo Ignacio González Domínguez et al., published by De Gruyter

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

Downloaded on 27.1.2023 from https://www.degruyter.com/document/doi/10.1515/phys-2018-0005/html
Scroll Up Arrow