Frequently Asked Question
Press "Ctrl + F" to find the keyword of your interest.
If you wish to have a direct link access to the video timestamps, please follow these instructions.
Found this video helpful? Why not take the whole HIL Specialist course? A Certificate is waiting for you for free at HIL Academy.
Would you or your organization benefit from having these videos narrated in your native language? Contact us and let us know if you wish to contribute.
TRANSCRIPT
Hello and welcome to the Thermal modelingapplications lesson.
This lesson will review applications and usecases for thermal modeling within Typhoon
and also provide insight into methods andtechniques for coupling the electrical and
thermal behaviors of systems.
You will also learn about the different approachesavailable to create thermal models inside
of Typhoon.
First, it's important to understand the motivationbehind developing thermal models.
Historically, industry manufacturers havefabricated power electronics modules which
were intended to be passively cooled usingheat sinks.
In purely passively cooled applications, thereis no control algorithm to monitor and control
the cooling process.
Heat sink sizing does not require real-timecalculation.
However, active cooling has become very commonin power electronics and electric motor applications.
Active thermal management is either implementedthrough a dedicated fan control unit or it
can be integrated into the main controller.
These cases often benefit from real-time simulation.
Dynamic instantaneous temperature responseduring some transients in load or switching
patterns can also be interesting to simulatefor the purpose of controller thermal protection
testing.
For example, some MOSFETs can be destroyedat exactly 150 C in intervals on the order
of milliseconds.
If thermal management is important in yourapplication, having the ability to C-HIL test
it adds value to your HIL device and extendsyour testing range to the system level.
You can achieve C-HIL testing of your system'sthermal performance by running coupled electrical
and thermal models in parallel, in real-time.
Let's explore coupled electrical and thermalmodels using an example model.
This is a Typhoon HIL example model of a variablespeed drive with IGBT losses calculation and
a thermal model.
Since our main focus is the thermal model,lets focus on the Heatsink thermal model and
the Inverter subsystems.
The Heatsink thermal model subsystem has twoinputs - power losses from the inverter switches
and ambient temperature.
The output of this component is an array ofcase temperatures for the different semiconductor
devices.
The ambient temperature can be changed inreal-time from SCADA, while power losses from
the switches and diodes are calculated inreal time in the inverter.
Let's enter this subsystem.
In the center of this subsystem, there isa Thermal network component called "Heatsink
model."
This component takes the vector values ofpower losses from the converter and ambient
temperatures.
In the power losses vector, each value representsthe power losses from one switch or diode.
Using this arrangement, you can observe temperatureand power losses per switch or per diode.
The temperature values computed in the Heatsinkmodel are fed back into the inverter.
You can see how the heatsink thermal modelconnects to the thermal model of the inverter
by looking inside of the Inverter subsystem.
In the inverter subsystem, you can see a Three-phaseinverter component, with an additional input
for case temperatures.
These case temperature values are informedfrom the heatsink model.
The inverter component has additional outputsfor power losses and temperatures at the junctions
of each switch and diode.
Since all signals here are configured in vectorform, you can access and monitor values at
the diode or switch level.
The power losses output provides both switchinglosses and conduction losses.
It is important to note that when temperaturecalculation is enabled in any converter, the
model computes junction temperatures by applyingthe same principle as the Thermal network
component.
This means that understanding how the thermalnetwork component function will also give
you insight into how converter thermal lossesare modeled.
If you want to know more about how to couplethermal and electrical models, please refer
to the how-to guide which is linked in thematerials tab.
If you open the mask of the thermal networkcomponent, you can see that the first property
is the number of thermal networks that willbe computed by this component.
Since there are six switches and six diodesin the three-phase inverter component, you
need 12 thermal networks.
This also corresponds to the vector lengthfor the input and output ports of the component.
The next property is the network type whichcan be specified as Cauer or Foster.
An understanding of these network types isimportant when determining which network to
use in different situations.
The Thermal network component represents ageneral thermal model which can be parametrized
with either Cauer thermal model or Fosterthermal model data.
A thermal model can be used to represent heattransfer through multiple layers of semiconductor,
or heat transfer through heatsinks, thermistorsand so on.
The Cauer model has physical meaning; everylayer in the model represents one material
layer in the physical system.
Comparatively, the Foster model is a purelymathematical representation of the system.
The parameters of this model can be obtainedfrom real measurements because model parameters
are time constants and thermal resistances.
However, this model does not have physicalmeaning.
An in-depth overview of the fundamental conceptsand operational principles of these thermal
network models can be found in the paper linkedin the materials tab.
Let's examine both thermal models separately.
Here you can see a representation of the Cauermodel using elements from the electrical domain.
This example contains four layers with eachlayer consisting of a resistor and capacitor.
This network could represent, for instance,a semiconductor which includes a silicon chip,
solder, substrate, solder, and base, suchas the semiconductor structure shown here.
Temperatures marked T1 to T5 would then representthe temperatures of layers between the mentioned
parts.
Since the layers of this model have physicalmeaning, the number of layers will depend
on factors such as the type of element beingmodeled and whether there is a thermistor
or a heatsink.
On the right of the screen, you can see themathematical equations that describe the Cauer
thermal model.
The first equation describes the relationshipbetween thermal capacity, designated as Cthermal,
and thermal time constant and resistance,designated as t and Rthermal respectively.
The second equation describes the thermalenergy transferred between layers.
In the numerator of this equation, T12 representsthe temperature difference between points
1 and 2.
The last equation describes the thermal energytransferred through the thermal capacitance.
The derivative in this expression describesthe change in temperature of point 1.
Returning to the model, you can see how theCauer model is parametrized.
The thermal resistance and thermal capacitanceparameters are entered as lists.
The lengths of these lists reflect the numberof layers of the thermal model.
From this, you can see that the Cauer modeldescribed here contains two layers.
The resistance and capacitance of the firstlayer are described by the first elements
of the thermal resistance and thermal capacitancelists.
The second layer is similarly described bythe second elements.
You can repeat this for as many layers asyou wish to include in your model.
The second network type is the Foster thermalnetwork.
In contrast to the Cauer network, the individualRC elements here do not represent material
layers.
The nodes within the network do not have anyphysical meaning.
Coefficients for this network can easily beextracted from a measured cooling curve which
is often provided in datasheets.
This model is parametrized with time constantsand thermal resistances.
On the right, you can see how these valuesinform the thermal capacitance.
Keep in mind that nodes internal to the networkdon't have physical meaning.
The inclusion of multiple layers is used toincrease fidelity of the temperature transient
described at the ends of the network.
Returning to the model again, you can seehow the Foster thermal model is parametrized.
In this example model, the Three-Phase Invertercomponent uses a Foster thermal network.
You can see this network description on theLosses tab of the component parameters when
Temperature calculation is enabled.
This component has its thermal network typeset to Foster.
You can also see that the switches and diodesare parametrized separately for this converter.
Since this is a Foster thermal network, thermalresistances and thermal time constants are
used for parametrization.
Data for Foster thermal models can often becollected from the datasheet.
Let's take a look at the datasheet of an IGBTintegrated power module provided in the materials
tab.
In this datasheet, thermal impedance informationis presented on the last page - page 9.
On the left side of the page, you can seea plot of transient Thermal impedance vs time.
On the right of the page, you can see themathematical expression that is used to model
the transient thermal impedance.
To match this thermal response in your system,you should apply the parameters in the table
below the mathematical expression to yoursystem.
For example, to create the IGBT thermal model,you should populate the Thermal resistance
with a list containing the three resistancevalues from the first row.
Similarly, the thermal time constant shouldbe populated with a list containing the time
constant values from the second row.
You also have the option to import switchand diode parameters from an xml file.
It is important to note that these files alsocontain parameters needed for temperature
calculation meaning parameters of thermalmodels.
Because of this, if you import an xml filefor a specific switch or diode, the thermal
model properties of the converter will alsobe populated.
This concludes the less on thermal models.
If you are interested learning more aboutthe Thermal network component, please refer
to the component documentation linked in thematerials tab.
Keep in mind that if you want to create yourown custom thermal model, you can always do
so using the signal processing toolbox.
Thank you for your attention.
