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 everyone and welcome to our lesson on electrical circuit partitioning between cores.
You will acquire knowledge about all the relevant components,
and understand how and where to use each of them.Let's start with why we do electrical circuit
partitioning. The motivation is to enable parallel computing of complex
power electronics and power system models that usually contain a large number of switches.
The basic partitioning algorithm is based on the detection of unconnected circuits.
In this diagram, there are 3 separate unconnected circuits that will end up in three separate cores.
This is because when Schematic Editor detects unconnected circuits,
it assigns them automatically to separate cores. There is still the possibility to override this
and group some parts of separate circuits in the same core, which will be explained later.
Components used for Electrical circuit partitioning elements can be divided into
three main groups: Ideal transformer based, Transmission line model-based core couplings,
and Core marker components.
IT coupling components are ideal transformers with a transfer ratio of one. They are
used to partition the complete emulated power electronics circuit into multiple sub-circuits.
Consequently, each sub-circuit is assigned and executed on a separate processing core.
In addition, this ideal transformer block introduces a time delay of one simulation step
between the coupling s sides, which is negligible for most practical systems. They are also
asymmetrical which is emphasized by the red and green sides of the component icon. The red side of
a coupling component is the current source side, while the green side is the voltage source side.
IT coupling components can be single phase and multi-phase.
You can find out more details about each type in the documentation link in the Materials tab.
Depending on the position, IT core coupling elements can introduce instabilities in the
circuit. In order to verify the stability of IT core coupling components a coupling stability
analysis tool is available in Schematic Editor. You can enable this by clicking on Model, Model
Settings, Circuit Solver settings, and then check the Enable coupling stability analysis checkbox.
When coupling components are added to the circuit some topological conflicts can occur, as we
covered in lesson 3.2.1 on topological conflicts. These conflicts are solved by adding snubber
circuits in parallel with the coupling's current source and/or in series with the coupling's
voltage source. Snubbers are embedded in coupling components. We will cover this in more detail in
lesson 3.3.2.1.1 on Snubber parametrization.Now let's switch to TLM core couplings.
TLM, or Transmission Line Model, core coupling components are based on transmission-line links.
Similar to IT coupling components, they are used to partition the complete emulated power
electronics circuit into multiple sub-circuits. A coupling component can be either capacitive
or inductive. Core coupling elements can be single, three, four, or five phases.
You can find out more about each type in the documentation links in the Materials tab.
The main advantage of TLM couplings, compared to the ideal transformer-based coupling components,
is that TLM couplings are symmetrical components. Both sides of the TLM couplings
are voltage sources behind a resistance. Because of this property, TLM coupling rotation is not
important and they will not introduce any topological conflict in the circuit. The main
disadvantage is that they add some additional inductance or capacitance to the circuit.
However, it is better to replace an existing inductor/capacitor with a
TLM coupling in order to get better results. When it comes to TLM coupling components the
bilinear discretization method is recommended since TLM coupling is based on this approach.
But, if the TLM inductance or capacitance is relatively small, trapezoidal discretization
can be used. TLM coupling is in general more robust than Ideal Transformer coupling.
TLM coupling in combination with a bilinear or trapezoidal discretization method guaranties
stability in most practical use cases, so there is no need for an additional stability analysis.
There are five Component Properties available to parameterize the components:
Coupling Type, input values for L or C, Embedded inductors/capacitors, TLM / Embedded components
ratio, and Ratio. The coupling type can be either inductive or capacitive. According to
the selected type, a parameter for the Inductor value or the Capacitor value is available.
Based on these inputs, the TLM coupling behaves as an additional inductor/capacitor in the circuit.
If Embedded inductors/capacitors is enabled, embedded inductors/capacitors
are added to both sides of the coupling.If embedded components are added,
it is necessary to specify how the impedance is divided between the TLM coupling and embedded
components. That can be done using the TLM / Embedded components ratio and Ratio properties.
If the 'Automatic' option is selected, the ratio is determined by the discretization method.
For instance, if the bilinear discretization method is chosen,
the coupling element will use around 20% of the impedance and rest will be divided equally
between the TLM coupling embedded components. With the trapezoidal discretization method,
less than 5% of the impedance is used by the coupling element, while the rest is equally
divided between TLM embedded components. If the 'Manual' option is selected, the
ratio can be explicitly set to meet the model's requirements. It must be a value between 0 and 1.
An existing inductor or capacitor can be replaced with a corresponding TLM coupling. In this case,
the TLM coupling inductance/capacitance needs to be the same as the inductance/capacitance of the
replaced element. If Embedded inductors/capacitors is enabled, the inductance/capacitance will be
divided between TLM and embedded components. In that case, the best results are obtained. It is
recommended to use TLM couplings in this way.TLM partitioning elements can also be single
phase or multi-phase. In the Materials tab, you can find more about each type.
As mentioned before, the basic partitioning algorithm is based on the detection of
unconnected circuits. Each unconnected circuit is executed on a separate core.
But there is component called a core marker which can connect separate cores into one core.
This component has only one parameter which is Marker ID. Independent circuits,
unconnected circuits, or circuits divided by a coupling component from the full schematic model
are emulated on the same FPGA solver core if they share a core marker with the same Marker ID.
Core markers are not mandatory in order to partition the model properly, but they can help
you to save on core resources. You can find more details about core markers in the Materials tab.
Now let s see how we can simulate some parts of the
circuit in the same core using the example model.
As you can see, we have three separate circuits which are all simulated on different FPGA cores.
Now we will add core markers in these two cores. For this,
let s make sure they share the same ID. Any name can be used for the ID, so let s use Id0.
Now if we compile the model again, you can see that in this case the model fits into two cores.
Each type of core coupling has its advantages and disadvantages.
The ideal transformer (IT) coupling is simpler in structure, but can present a challenge with
circuit placement and orientation. If not done optimally numerical instabilities can happen,
such as arithmetic overflow, shown by the AO flag. Because of that, it's very important to
know the orientation of the core coupling and if snubbers are needed to parametrize them correctly.
Determining the position for the coupling component is crucial in order to preserve
stability and accuracy of the simulation. In some cases, this step is trivial, while there
are use cases when it is more challenging. Here are some general rules for coupling placement:
IT Coupling elements should be placed in the model where the dynamics in signals are lower.
Usually, these spots are next to capacitors or next to inductors.
The red side of the coupling component represents the current source side of a coupling. This
side should be rotated towards a slow changing voltage, usually a voltage source or capacitor.
The green side of a coupling component represents the voltage source side of a coupling. This side
should be rotated towards a slow changing current, usually, a current source or an inductor.
These are general rules. It is not always possible to place coupling component in an ideal position.
Power electronic circuits are highly dynamic. Each switch permutation
can modify the topology in a way that a capacitive circuit becomes more inductive,
therefore a circuit snubber might be needed.IT coupling placement is explained in detail
in the link in the Materials tab. In this document, you will find more examples.
The transmission line model (TLM) core coupling is more complex in structure, but is much more
forgiving in terms of placement. There's no requirement for orientation or snubbers.
However, these benefits don't come for free. The discretization type must be changed, which many
times results in extra time slot utilization. More importantly, these couplings present a parasitic
capacitance. The value of the capacitance is a function of the inductance and the selected
simulation time step. The smaller the inductance set for the coupling, the bigger the parasitic
capacitance added to the circuit. The bigger the time step, the bigger the capacitance.
TLM coupling brings an additional inductor or capacitor in the circuit,
according to the selected coupling type. Because of that,
it is recommended to remove an existing inductor (or capacitor) from the circuit
and then place a corresponding TLM coupling component at that spot.
You can find more examples about TLM coupling placement in the link in the Materials tab.
Empirically, the IT core coupling has shown to be more applicable in power electronics while the TLM
core coupling finds its application better suited to microgrids and power systems. That being said,
they should not be understood as separate solutions for these two application domains.
In addition to the core coupling components that we ve already covered, there are some components
that have embedded core couplings as well. For example, if we check transformer components,
you can see that next to the General tab is a Coupling tab.
This shows that you are able to use embedded couplings inside of some other components.
Now let's look at how an embedded coupling can be used in the single phase two winding transformer.
There are two possible options for embedded coupling in the single phase two winding
transformer: Ideal Transformer based coupling and TLM coupling. If Embedded coupling is set
to Ideal Transformer, an Ideal Transformer based coupling will be placed between the two windings
of the transformer. If Embedded coupling is set to TLM, a secondary winding inductor (L2) will
be replaced with a TLM coupling component. The inductance will be divided between the coupling
and the embedded inductors hidden in the TLM.The TLM to embedded inductors ratio can be
determined by the compiler, but also it can be specified explicitly. If the
Automatic option is selected, the ratio will be determined by the discretization method.
If the Manual option is selected, the ratio can be explicitly set to meet your requirements.
You can find the list of components that have embedded coupling options in the Materials tab.
Now let s look at a practical example of when it is recommended to use IT core
coupling and when it is better to use TLM coupling elements.
The following example will illustrate some of the common circuit partitioning patterns.
Let's enter the Example Explorer, choose grid-connected converter
and three-phase back-to-back converter.As you can see there are two converters
coupled through a large DC link capacitor. This circuit partition is necessary,
since the overall converter weight is 6, and if we check the device table,
we can see that the maximum converter weight for the chosen device configuration is 3 per core.
There are two suitable positions for coupling. Those positions are either
to the left side of the capacitor or the right side of capacitor,
since the capacitor should change voltage slowly. The red, current, side of the
coupling should be connected in parallel with the capacitor. Let's try both solutions,
we can figure out that in both cases a topological conflict appears. As you can see the issue appears
on the green side of the coupling when the inverter is in short-circuit mode.
During regular work, when it is not of interest that converter works in short-circuit mode,
this warning can be ignored. If we want to solve the warning,
we can enable a dynamic resistive snubber on the voltage source side of the coupling component.
Now let's switch to another example that demonstrates the use of IT core couplings as well.
In order to find this example, let s navigate to electrical drives and then induction machine
open loop in the Example Explorer.As you can see this example contains
a three- phase rectifier with a weight of 3 and a three-phase inverter with a weight of 3.
You can find the maximum converter weight in the device table for this device configuration;
here we can see the weight is 3 per core. That means that we have to separate this
model in two cores. Since IT couplings are better for Power Electronics applications,
we will use them to separate this circuit. The red side of the coupling
is placed towards the capacitor, as suggested previously in the lesson.
The next example demonstrates a typical use of TLM couplings. As we explained,
TLM couplings are mainly used in Power systems and microgrid applications. Let s look at two examples
of the use of TLM couplings. The first example presents a simple microgrid system consisting of
three battery components and the grid. In order to find this example, check Materials tab.
In this example, generic microgrid components are used. If we try to
compile this model without coupling, we can see that matrix memory is utilized over 100% .
Since this model should be coupled, the best place for the coupling placement is the RL section,
which has an embedded TLM coupling.
As you can see in this case it is not obvious where a good position for the IT coupling is.
If you follow the general rules mentioned above for IT coupling placement,
you can figure out that they do not apply to this example.
For instance, if we put IT coupling between the Grid and RL section,
the current side of the core coupling will be rotated towards the inductor in both cases,
which is not recommended since it can cause a topological conflict.
Here snubbers are needed and they must be parameterized,
which can be tricky because of issues with instability.
In cases where it is not so obvious where the best place for the IT core coupling component,
the TLM couplings are the recommended solution. By using TLM couplings,
this issue of coupling placement is avoided.The next example represents a power systems
application. Let s open the Power Systems IEEE33 bus system example in the Example Explorer.
As you can see this example is very complex, and it is not easy to find a
suitable position for IT coupling, since the coupling orientation is important.
In case you want to use IT couplings for this kind of application, you would need to add a snubber
which will be difficult to parametrize. Also, an IT coupling can cause some instability which would
be avoided by using TLM couplings. As you can see there are RL sections that represent a natural
place for TLM couplings. For these reasons, in this example, inductive TLM couplings are used.
Let's summarize what are the main advantages and disadvantages of each type of core coupling.
When it comes to symmetry, IT couplings are asymmetrical while TLM couplings are symmetrical.
Any discretization method can be applied for IT couplings, while bilinear or trapezoidal
discretization method is recommended for TLM couplings. TLM couplings are straightforward
to place, while IT couplings require some additional considerations for placement.
Snubbers are characteristic for IT couplings, while in TLM core couplings there is no need
for snubbers. TLM couplings inject parasitic impedance into the model,
while for IT couplings that is not the case. IT couplings are more suitable for power electronics
applications while TLM couplings are recommended for microgrid and power systems applications.
In this lesson we covered electrical circuit partitions, as well as the main groups of core
coupling components, their advantages, and disadvantages. In the following lessons,
we will focus more on Signal Processing Circuit Partitioning and IT snubber parametrization.
So, see you then. Thank you for watching!
