Moddelling Thermal Behaviour
Moddelling Thermal Behaviour
Abstract – In electric power products development, thermal aspects are becoming very
important due to the fact that they guarantee the correct operation of product. The specified
temperature limits cannot be exceeded. Thus, knowledge of temperature behavior is very
important factor in order to predict reliability and performance in its working environment. In
this paper some bases of Thermal Network Method (TNM) modeling in application to transient
rating calculations of typical GIS compartment arrangements are presented. Thermal network
composed of heat sources, thermal resistances and capacitances can be easily represented and
calculated by means of ATP-EMTP program (also OrCAD Capture environment).
1 Introduction
Thermal analysis is a very important issue in electrical devices and apparatus design especially
for high current ratings. According to the international IEC 62271-203 standard [2], a
temperature rise type test has to be performed, in order to ensure that the allowed temperature
limits are not exceeded. To optimize the number of needed design tests, it is important to predict
the temperature rise with a simulation tool. Different approaches are available to thermal
simulation, selection of which is always a matter of balance between the modeling complexity,
accuracy, and computational cost [7]. Among them, the thermal network method (TNM) [3]
has extremely low computational cost at decent accuracy. The basic idea of TNM is to describe
a thermal problem by an equivalent circuit network and solve the network. If one uses the
OrCAD software a thermal library is available (created and built at ABB [7]), that provides the
needed circuit element symbols and the corresponding thermal models for, e.g. power loss,
thermal conduction, convective heat transfer, radiation, and more. Alternative solution is to use
the ATP-EMTP program to simulate the thermal behavior [5] by setting up an equivalent
electrical circuit.
This paper addresses on thermal behavior of a typical ELK-3 GIS compartment using OrCAD
Capture environment and ATP-EMTP program. Modeling and simulation results were built and
compared. These analyses are a fundament for more advanced ATP-EMTP application in
Thermal Network Modeling.
where heat flow and temperature correspond to current flow and voltage, respectively. Similar
difference ΔQ, thermal resistance R and the transferred thermal output P is presented by
to the Ohms law for electric streaming fields, the following relation results between temperature
formula (1)
∆Θ=P⋅R (2-1)
where: voltage (U) and current (I) corresponds to temperature (Q) and heat transfer rate (P),
sources in thermal network mean power losses and are equivalent to current sources in electrical
circuit [5], [8].
Thermal conduction, radiation and convection are represented by thermal resistors.
· resistance of conductors,
There are four mechanisms of changing electric energy into heat:
· magnetic hysteresis of
materials in magnetic
· eddy currents,
field,
· selected phenomena in
dielectric materials.
There are also three ways of heat
· conduction,
transfer in electric devices:
· radiation,
· convection.
It is commonly known, that differential equations for heat transfer are similar to equations,
which describe electric current flow. This fact enables simplification of thermal simulations by
using thermal network method. Main aspect of this method is replacement of real geometrical
elements and thermal phenomena by electrical elements such as resistors, current sources and
capacitances.
The network model for a thermal problem could be created and solved using a circuit simulator.
The full set of thermal – electrical system’s analogies is presented in Fig. 1 [4]. The heat sources
of a switchgear device are the temperature dependent ohmic losses of current carrying
conductors, contacts, and connections. Heat is conducted along the current path and dissipated
to the gas and enclosure through convection and radiation. The heated enclosure is cooled down
through convection and radiation to the ambient. Using TNM, the power loss of a conductor is
calculated by its geometrical dimensions and material properties. Heat dissipations via radiation
and convection are described by thermal resistances, which are calculated based on the surface
areas, emissivity numbers and the installation conditions. An iterative solver is necessary due
to the dependency of the conductor resistivity, the convective heat transfer coefficient and
radiation coefficient on the temperature. Fig. 2 shows an example of a thermal network model
for two pieces of conductors connected by a screw joint within a cubical, as a rudiment
illustration of a switchgear model [3].
A
2570
A-A
f 160
f 820
Fig. 3: Typical dimension of GIS ELK-3 type compartment
Nominal parameters of the GIS ELK-3 for which the solution is intended are indicated. Typical
· frequency: 50 Hz;
· SF6 absolute pressure: 5 bar (SF6 pressure influence on the thermal convection
Fig. 5: Cross-section of analyzed case and equivalent thermal network (model 1); main
conductor with surrounding air pressure, R1 - thermal resistance for radiation, R2 – thermal
resistance for convection to air, C1 – conductor thermal capacitance (15.06 kJ/K)
Fig. 6: Cross-section of analyzed case and equivalent thermal network (model 2); main
conductor in enclosure; R1 - thermal resistance for radiation, R2 – thermal resistance for
convection to SF6, R3 - thermal resistance for convection from SF6 to enclosure, R4 – thermal
resistance for radiation to air, R5 – thermal resistance for convection to air, C1 – conductor
thermal capacitance (15.06 kJ/K), C2 – SF6 thermal capacitance (0.92 kJ/K), C3 – enclosure
thermal capacitance (157.47 kJ/K)
The OrCAD capture and ATP-EMTP simulation results for steady state final temperature were
compared in Table 2. Drawings, diagrams and simulation results are presented in Figures 7÷12.
Fig. 7: Thermal network diagram for model (1) – OrCAD program (acc.to Fig. 5)
Pipe radiation
to air
MODEL
hradiati
Pipe temp
V T
I I
MODEL
heatsrc
U(0)
+
Pipe convection
to air
M
Recalculate current [A] MODEL
into joule heat source [W] hconv
Fig. 8: Thermal network diagram for model (1) – ATP-EMTP program (acc.to Fig. 5, Fig. 7)
a) b)
80 160
70 140
120
60
temperature [C]
tem perature [C ]
100
50
80
40
60
30
40
20 20
0 50 100 150 200 250 0 50 100 150 200 250
time [min] time [min]
Fig. 9: ATP-EMTP simulation dynamic results for conductor in air (without enclosure
element) for rated currents 6300 A (a) and 4000 (b)
Fig. 10: Thermal network diagram for model (2) – OrCAD program (acc.to Fig. 6)
V T V T
I I
MODEL
heatsrc
U(0)
+
T V T T
U(0)
U(0)
+
Fig. 11: Thermal network diagram for model (2) – ATP-EMTP program (according to Fig. 6,
Fig. 10)
b) b)
60 120
conductor conductor
enclosure enclosure
100
50
te m p e ra tu re [C ]
te m p e r a tu r e [C ]
80
40
60
30
40
20 20
0 50 100 150 200 250 0 50 100 150 200 250
time [min] time [min]
Fig. 12: ATP-EMTP simulation result for conductor, SF6 pressured gas and enclosure
element; for rated currents 6300 A (a) and 4000 a (b)
Table 2: Simulation results – temperatures on conductor comparison
Rated
OrCAD Capture ATP-EMTP
current Setup
[A] [°C] [°C]
Model (1) 29.5 31.4
1250
Model (2) 27.0 29.0
Model (1) 45.1 46.8
2500
Model (2) 37.3 39.1
Model (1) 75.0 73.7
4000
Model (2) 56.7 57.5
Model (1) 149.5 149.7
6300
Model (2) 101.8 104.4
To validate the ATP-EMTP model several currents (untypical) levels were used as a heat
source. Results presented in Table 2 indicate the ATP-EMTP models correctly estimate the
final temperature of the whole compartment. OrCAD software is a program dedicated to the
electrical calculation – using thermal library (created and built at ABB [7-8]) can be dedicated
to the thermal network calculations. ATP-EMTP program has allowed us to achieve the
approximate value based on the initial assumptions.
The difference between OrCAD and ATP-EMTP comes from the fact that the OrCAD utilizes
more sophisticated thermal model to calculate the temperature.
5 Conclusions
Heat phenomena in working electrical apparatuses are the condition which defines their current
limits. Therefore using the computational techniques allows engineers to analyze thermal aspect
of systems in the early design stage and is a must for modern designers. Such analysis helps in
the estimation of the required cooling or alternation of the construction to meet thermal
requirements. Therefore the thermal network concept offers approach which is simple and
intuitive, especially for electric engineers. And with high speed of computations and good
precision it is worth to be considered.
The thermal networks of two models of ELK-3 GIS busduct have been built. Calculations were
made for two approaches: GIS busduct without enclosure – surrounding air pressure (model 1)
and GIS busduct with enclosure – SF6 pressure inside (model 2). All calculations were made
for simplified GIS busduct design and were simulated for operating currents (also untypical
values): 1250, 2500, 4000 and 6300 A. Both models returns relatively close solutions for the
compartment temperature. Additionally if one utilizes the ATP-EMTP software time constants
for temperature increases can be read. ATP-EMTP model overestimate the thermal value by
the several Celsius degrees. This is due to fact that ATP-EMTP utilizes only fundamental
thermal dependencies. Further expansion is possible. Important fact is that ATP-EMTP allows
for fast and easy assessment if the thermal conditions are meet or are close to their limits.
· For typical atmospheric pressure without enclosure temperature on conductor was equal
Following conclusions can be listed, based on simulation results:
· If GIS busduct with enclosure and SF6 gas pressure inside – temperatures are below
approx. 150°C (not applicable according to IEC 62271-203 [2]);