Li Tra
Li Tra
Li Tra
What’s New
F2023
January 2023
r9766
Welcome to PowerFactory 2023!
DIgSILENT constantly updates and improves PowerFactory, not only to ensure that it remains compat-
ible with current standards but also to reflect new technologies and innovations in network analysis. In
PowerFactory 2023, we have many different developments and new features to share with you, some
targeted at specific user groups, others of benefit to all.
In our Distribution Network Tools module, we now have a dedicated Low Voltage Load Flow Calculation
function (see section 1.10.1), which provides a sophisticated modelling option using coincidence curves
to take the stochastic consumption behaviour of low voltage loads into account.
Other functions such as Arc-Flash Analysis and Optimal Power Flow have been enhanced with a
number of new features. See section 1, where all the main calculation developments are described.
Alongside the calculation developments, we have implemented a new reporting concept, primarily
aimed at allowing reports to be generated as PDF documents. This has allowed us to replace some of
the existing ASCII output with PDF reports, but the development goes much further than that, offering
an inbuilt PDF Viewer, support of various export formats, and the possibility of custom report design.
The new development is introduced in section 1.14.1.
Modelica simulation models can now be developed via block diagrams, using the graphical editor.
This feature is described in section 1.3.2. And whether DSL macros or Modelica models are used,
the process of model development has been made easier with the introduction of a library browser
presented in the graphical window, for the selection of these macros and models. Details can be found
in section 1.3.1.
Another important development for simulations is the option for automatic pre-compilation of user-
defined DSL models at initialisation. This feature, which will be delivered in a service pack rather
than at the initial version release, will result in significant improvements in simulation times. See section
1.3.4 for details.
Most users will see some differences in the user interface and data handling in this version. The context
menus have been redesigned and the new-object dialog has been simplified. In the background,
we have improved user interface responsiveness and implemented full support for Unicode (UTF-8) text
encoding, which now allows texts to contain any characters from any of the world’s writing systems.
All these changes, together with enhancements relating to database management, are described in
section 3.
The main developments related to power equipment models are detailed in section 4. New models,
such as the Fault Current Limiter and Static Synchronous Series Compensation model are listed first,
followed by details of many enhancements to existing models.
The last few sections of the document cover interfaces and scripting, and a reminder that from Power-
Factory 2023 onwards, PowerFactory is only offered as a 64-bit version.
We hope you enjoy reading about all the new features and we wish you continued success with
PowerFactory 2023.
Contents
1 Analysis Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.5.1 DC Arc-Flash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.14 Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.1 Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2 Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.4 Unicode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.6 Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.3 Transformers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.5 Controllers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5.1 Python . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
7 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
7.2 Licences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
1 Analysis Functions
The thermal rating object (IntThrating) has been enhanced to allow users to enter data for continuous
and short-term thermal ratings that vary depending on ambient temperature, wind speed or solar
irradiance. The dependency is configured by the user by specifying break-points on the configuration
page, and then the data is entered in tabular format. In the element that uses the rating object, the
relevant Meteorological Station (ElmMeteostat) is referenced. This provides the temperature, wind
speed or solar irradiance data from which the applicable thermal rating is determined. In the example
in Figure 1.1, a line rating is dependent on the ambient temperature measured by a nearby weather
station.
Furthermore, it is possible for the settings in the Meteorological Station to be modified by a time
characteristics, to provide typical values according to the time of day or year.
In addition to this enhancement, it is now possible to use thermal rating objects with lines that contain
line sections. See section 4.2.1 for details.
The Load Flow reports, previously produced using the “Output of Results” command, as ASCII reports
in the Output Window, are now generated as PDF documents. See section 1.14.1 for details.
In PowerFactory 2023, both DC short-circuit calculation methods, IEC 61660 (DC) and ANSI/IEEE
946, support overcurrent protection devices, for users of the Time-Overcurrent Protection module. The
tripping times are calculated based on steady-state short-circuit currents. Users are now able to model
protection devices in DC networks and verify the coordination using time-overcurrent plots and short-
circuit calculations.
In PowerFactory 2021, the Complete Method short-circuit calculation was updated to reflect the changes
in G74 Issue 2. One feature not included at that time is now delivered: there is now an enhancement to
take into account the automatic triggering of bypass for elements such as series capacitors (see Figure
1.2). Details can be found in the relevant technical references.
This methodology is in line with the recommendations in section 7.6.4 of G74 and results in a more
accurate modelling of series capacitors and TCSCs for close-up faults. The same methodology is
extended to other elements:
• Static Synchronous Series Compensation: The option to activate a bypass circuit based on a
current threshold can be enabled. For a detailed description of this new model see section 4.1.2.
• Fault Current Limiter: The trigger unit changes the resistance to a maximum value based on a
current threshold. For a detailed description of this new model see section 4.1.1.
• Rectifier and HVDC-LCC: For close-up faults the DC valves are blocked. This behaviour can be
enabled with an undervoltage threshold defined within the corresponding element.
There are no additional options in the short-circuit command; the enhancement to the algorithm is
simply activated if any relevant element in the network has its specific option selected.
The drawing toolbar for dynamic model block diagrams now presents a library browser for DSL macros
and Modelica models. This library browser displays all DSL macros / Modelica models from the global
DIgSILENT Library and any user-defined DSL macros / Modelica models from the project library’s
Dynamic Models folder. Sub-folders and their contents are displayed in a tree structure, as shown
below in Figure 1.3.
Figure 1.3: Library browsers for DSL macros and Modelica models
For quick access, it is possible to search for particular macros / models using the element filter of the
drawing toolbar. With a right-click, it is possible to show the contents of a selected macro / model. DSL
macros / Modelica models from the drawing toolbar can be inserted in block diagrams of DSL / Modelica
model types using drag & drop or a select & click action, as shown in Figure 1.4 below.
See also section 3.6.3, which describes a number of changes made to the Dynamic Models library.
Figure 1.4: The drawing toolbar for dynamic model block diagrams, showing the filtered list and dragging
macro into block diagram
PowerFactory 2023 gives model developers a new dimension of dynamic modelling capability by sup-
porting the graphical definition of clocked Modelica models within a Modelica standard compliant devel-
opment environment. This allows hierarchically structured Modelica models to be graphically defined in
an intuitive manner.
Complex models can now be defined on the basis of a number of basic blocks already supplied in the
PowerFactory library. Alternatively, individual blocks can be defined, based on a similar process as
already existing for DSL modelling (using DSL macros). A simple example of a graphically developed
Modelica Model is shown in Figure 1.5 below.
Figure 1.5: Example of a graphically defined digital controller of a power inverter using Modelica
diagrams
Each Modelica Model supports an “Icon” view (as illustrated in Figure 1.6 below), which allows user-
defined graphical representations of the specific model whenever it is used as a hierarachically subor-
dinated block. The available drawing tools allow users to draw and customise the graphical appearance
of the block.
Graphically defined models support hierarchically structured parameter sets depending on the internal
structure of the model. Consequently, tree-like parameterisation views, as shown in Figure 1.7 below,
allow for consistent grouping of model parameters, thus avoiding cluttered parameter lists for complex
models.
PowerFactory 2023 comes with a new Modelica Library containing basic building blocks specifically
tailored for controls of power system equipment. The supplied library (as shown on the right-hand side
of Figure 1.3) is actively maintained, supported and enhanced by DIgSILENT.
“Pack” function
The Modelica Type object (TypMdl) supports now a subblock “Pack” function, similar to that of DSL
models. With this function, any Modelica model that uses subblocks can pack these subblocks within
a dedicated folder located inside the object. This allows simplified export and sharing of the Modelica
models across project boundaries or outside the PowerFactory database.
For Modelica clocked models intended to execute at a fixed rate defined by the solver’s simulation time
step, PowerFactory 2023 introduces a new Sampling option for Modelica models named “According
to integration step size” (Modelica Model (ElmMdl) dialog, Clocked Model page). When selected,
the model’s sampling period will be automatically set at runtime, depending on the step size settings
available in the “Calculation of initial conditions” command.
This option is highly useful for models dedicated to post-processing tasks (identification of minima and
maxima of a large set of variables at each simulation time step, RMS voltage computation according
to various algorithms for later use in a technical report, etc.) or for replicating the behaviour of legacy
simulation environments based purely on procedural models. The option is available for FMI based
models as well.
The PowerFactory 2023 installation provides an option to include the freely available MinGW C-Compiler
(the option is initially set by default). This feature allows users to compile dynamic models directly from
PowerFactory without the need to install any other external software. Detailed licence information for
MinGW itself, the MinGW runtime and additional libraries can be found in the installation directory of
PowerFactory, subfolder MinGW → licences.
PowerFactory 2023 is capable of automating the process of model compilation, enabling users to
benefit from improved simulation performance without the need of manually compiling individual models.
NOTE: This feature will not be available in the initial release of PowerFactory 2023 but will be delivered
in a subsequent service pack.
The new “Automatic compilation” option for DSL models will be included in the “Calculation of initial
conditions” command (Solver Options page, Models tab). With this function, the initialisation command
can automatically check for compilation and compile all currently non-compiled DSL models of level 5 or
higher. The functionality is effective only at run-time, so users can still apply modifications to the existing
DSL models, even if PowerFactory has already generated a compiled code for them. PowerFactory
automatically monitors the state of the DSL models as well, meaning that any subsequent simulation
runs on an unchanged DSL model will not trigger a re-compilation. Conversely, if any modification has
been applied to a DSL model then a re-compilation is triggered for that specific model.
A number of Simulation Scan modules are available for RMS simulations, allowing variables to be
monitored during the simulation and events triggered accordingly. Now, in PowerFactory 2023, the
variable scan module (ScnVar ) is also available for EMT simulations.
The Protection Coordination Assistant was extended in PowerFactory 2022 to consider overcurrent
devices. The report command now also has the capability to create time-overcurrent plots showing
coordination results for overcurrent devices such as relays and fuses. Figure 1.8 shows calculated
settings for two overcurrent relays. As the plot legend indicates, the solid curves are settings calculated
by the Protection Coordination Assistant.
It is now also possible to show existing and calculated settings of the same relay within the time-
overcurrent plot. This is also shown in Figure 1.8. The relay R_AS5 is shown twice in the plot: once with
the calculated settings (solid curve) and once with the existing settings (dashed curve). This enables
the user to compare calculated settings with existing ones graphically.
Figure 1.8: Time-overcurrent plot with coordination results and existing settings
1.5.1 DC Arc-Flash
The Arc-Flash calculation has been extended to support DC arc-flash analysis. The functionality is
implemented for the NFPA 70E and the DGUV 203-077 methods. It is possible to determine the arc
duration based on fixed or relay tripping times (see also section 1.2.1). The NFPA 70E method supports
both simplified and detailed arc models, while for the DGUV 203-077 a worst-case calculation can be
enabled according to the recommendation.
Both DC arc-flash calculation methods are fully integrated into the arc-flash analysis, i.e. the handling
is the same as for the AC arc-flash. Tabular reports and diagram colouring are also both available for
DC arc-flash calculations.
Tabular report
The existing Arc-Flash report, shown in Figure 1.9, has been enhanced and become more intuitive with
respect to reproducing an arc-flash situation. For a particular arc-flash scenario, the newly introduced
options in the report allow the user to re-calculate a short-circuit or a short-circuit trace with the consid-
ered arc resistance at the desired fault location. This enhancement also enables the users to observe
the corresponding arcing current seen by a particular relay or a set of relays, in time-overcurrent plots.
Figure 1.9: Enhanced Arc-Flash tabular report with additional options to reproduce arc-flash situations
In addition to the enhanced report, the existing result variables list for Arc-Flash Analysis is extended by
adding new set of monitoring variables for different calculation methods. For example, it is now possible
to monitor the arc resistance at the fault locations. For the method DGUV 203-077, the variables for
current limiting factor i.e. 𝑘B and relative arc power i.e. 𝑘P are added.
Report document
The reporting command, which can be used to create arc-flash labels directly from PowerFactory, now
creates a PDF document, as shown in Figure 1.10 below. In addition, it has an option to allow user-
defined report or label templates in order to standardise the result export. Please see section 1.14 for
further information about the new reporting concept.
Several developments are introduced in the Harmonic Load Flow and the Frequency Sweep calculation
commands.
Consideration of contingencies
A major enhancement is the implementation of Contingency Analysis into the Harmonic Load Flow and
the Frequency Sweep calculation. It allows the user to consider all available fault cases (n-1, n-2, n-k)
in the analysis, with the results for each contingency being stored in a separate result file. In addition
to the new calculation feature, several new reports and a plot wizard for an automated creation of plots
have been introduced to allow easy analysis of results. See section 1.6.2 and 1.6.3 for details.
Frequency Range
The Harmonic Load Flow command now offers the option to execute the calculation for a defined
frequency range.
New commands have been introduced for the Harmonic Load Flow and Frequency Sweep calculation,
for loading results from existing result files into the memory. This allows the user to display calculation
results of a specific order without executing the command again, or likewise the results of a specific
contingency.
The former general Harmonics Calculation command (ComHmc) has been removed. Only the separate
Harmonic Load Flow (ComHldf ) and Frequency Sweep (ComFsweep) commands are available from
now on.
The results handling for the Harmonic Load Flow and the Frequency Sweep calculations has been
reengineered. Instead of one result object for all result parameters, separate sub-files are created as
required, which means that for the Harmonic Load Flow there will be one for order-specific results and
one for summary results. For both Harmonic Load Flow and the Frequency Sweep, if contingencies are
considered, there will be one sub-file for each contingency plus the base case.
The selection of harmonic distortion limits for the Harmonic Load Flow has been improved by an
additional input for the THD limit. A new object allows the user to define harmonic distortion limits
for different voltage levels. The new limit definition is automatically used when creating new tabular
reports or plots, as well as for recording results when considering contingencies.
New additional tabular reports for calculations with and without the consideration of contingencies have
been introduced.
For the Harmonic Load Flow calculation, several new reports are introduced, to give an overview of the
harmonic distortion in the network, along with critical contingencies. The following new tabular reports
are available, and Figure 1.11 shows an example:
Figure 1.11: Tabular report showing the highest total harmonic distortion (THD) for each terminal
For the Frequency Sweep calculation, several new reports are introduced to give an overview of the
resonances in the network. The impact of all considered contingencies is also analysed. The following
new tabular reports are available:
• All resonances: reports all occurring resonances over the analysed frequency range for each
recorded terminal and contingency.
• Resonances in a given range: resonances for each recorded terminal and contingency can be
filtered according to the frequency range of interest.
• Resonances near existing harmonic injections: takes existing harmonic injections from sources in
the network into account and lists all resonances which are close to the existing injections.
For both the Harmonic Load Flow and the Frequency Sweep calculation, new commands for an auto-
mated creation of plots are introduced. The commands are able to plot results of a certain variable for a
user-defined selection of terminals and contingencies. This allows the user to create as many plots as
are needed for terminals and contingencies of interest.
The Harmonic Load Flow Plot command allows the creation of harmonic distortion (HD) plots.
The Frequency Sweep Plot command is able to create frequency curves for the impedance and its
phase angle. In addition, an R-X plot can be created. Figure 1.12 shows the Frequency Sweep Plot
command along with its settings, and Figure 1.13 shows the resultant plots.
Figure 1.13: Frequency Sweep plot created by the new Frequency Sweep Plot command
Due to the increasing penetration of renewable energies and charging stations in distribution networks,
PowerFactory 2023 offers the newest guidelines for connection request assessments, to meet the
growing challenge of integrating new plants. For low-voltage networks, VDE-AR-N 4100 is introduced,
which is specifically designed for loads such as charging stations. The newest version of the D-A-CH-CZ
is now also available in the current state of its release.
D-A-CH-CZ Edition 3
The newest version of the D-A-CH-CZ has been published and will also be available in this PowerFactory
version. The new version is designed to strive for the best possible harmonisation in between the D-
A-CH-CZ and VDE and IEC guidelines. Edition 3 implements the need to assess storage units as well
as increasingly widespread plants with power electronics. Additionally, updated calculation methods
are introduced, in particularly for harmonics and asymmetries, for a more up-to-date assessment. In
the current publication, the guideline covers the updated assessments for connection requests of loads,
generation and storage units in low-voltage and medium-voltage networks.
VDE-AR-N 4100
With the introduction of the VDE-AR-N 4100, not only generation units but also loads can be connected
and assessed in low-voltage networks. As the number of charging station installations in low-voltage
networks increases, this new feature will help with the management of such requests.
The Connection Request Assessment reports, previously produced as ASCII reports in the Output
Window, are now generated as PDF documents. See section 1.14.1 for details.
The Optimal Power Flow function has been extended and the graphical user interface revised. The
additional level of detail for several models allows a wider range of applications for OPF, including the
optimisation of reactive redispatch measures to complement the Unit Commitment functionality. The
support of more sophisticated solvers also leads to faster computation times.
The cost components for generators now include redispatch costs in addition to the operating costs, in
order to penalise deviation from the non-optimised operating point. These costs are supported in the AC
and DC OPF, and also in the contingency constrained DC OPF, and their use enables the optimisation
of post-fault redispatch actions. Costs for the deviation of reactive power can be used in the AC OPF to
weight the individual power plants when maximising the reactive power reserve.
Parallel transformers can now be grouped via an external tap controller and optimised as one control.
The combination of several transformers guarantees coherent behaviour in the optimisation.
Penalty costs have been introduced for several classes of network elements, to enable the implemen-
tation of a deployment sequence and to prevent excessive use of certain measures. As well as being
used by the “Minimisation of costs” objective function, the penalty costs can be used to weight the
control variable deviations for other objective functions.
The IPOPT (Interior Point OPTimizer) solver has been introduced as an additional solver for AC OPF.
A well-known non-commercial solver for large-scale non-linear optimisation problems, IPOPT provides
stable and robust solutions and can be used as an alternative to the standard inbuilt solver. In the DC
OPF, the user can now include commercial solvers such as GUROBI or IBM CPLEX to achieve faster
solving times. The CBC Solver is now the standard solver for DC OPF.
The save to memory and save to file functionality implemented in the Load Flow calculation is now also
supported in AC OPF. This is advantageous if several OPF calculations are to be performed on the
same network model with different configurations or investigation times. Reusing the previous results
can provide a better starting point than a flat start or using load flow results, so improving convergence
and reducing the number of iterations needed, leading to faster calculation times and better solution
stability.
An option to force discrete active power operating points for generators has been introduced. This can
be used, for example, for renewable generation that can only be instructed to certain operating levels.
Unscheduled flows, such as loop flows, may cause unexpectedly high loading of network elements.
Therefore, Transmission System Operators (TSOs) need to assess the impact of unscheduled flows in
order to maintain the integrity and security of their system. With the results of this assessment, they can
then reduce the amount of capacity available for trading, to ensure that they will be able to cope with
the unscheduled flows.
PowerFactory 2023 has been enhanced by a new functionality called “Flow Decomposition”, available
in the Transmission Network Tools module, to calculate these flows.
Using Flow Decomposition, the user can trace the active power flows within meshed networks, in order
to identify loop, transit and import/export flows. The flow types are classified according to the zonal
allocation of source, sink and network element. A loop flow is for example an active power flow on a line
in zone A that starts and ends in zone B. For the zonal allocation, grids, zones or areas defined within
the PowerFactory network model can be used.
The Flow Decomposition calculation can be carried out for one point in time, using a load flow, or over
a period of time, using a Quasi-Dynamic Simulation. The basic options are shown in Figure 1.16.
In addition to the commonly known loop flow calculation, PowerFactory can also distinguish cycle flows
caused by HVDC links and phase-shifting transformers (PSTs), giving a better understanding of the
cause of the loop flow. The report specially developed to present the results of the Flow Decomposition
provides a structured overview of all analysed network components and how the active power flow is
split up into loop, internal, import, export, transit, PST and HVDC cycle flow, as shown in Figure 1.17.
Figure 1.17: Report for Flow Decomposition for two lines and two points in time
Furthermore, the new command also offers a so-called “Power Exchange Analysis”, which makes it
possible to analyse individual flows more in detail. The results of the calculation can be presented in
a dedicated tabular report showing the source and sink for each flow type and component. Depending
on the level of detail needed, the user can display the source and sink either as grouping element (e.g.
zone or area) or as the actual generation/load node. Figure 1.18 shows an example report for the Power
Exchange Analysis. In this example case, the loop flows are caused by load and generators in the zone
“Northeast” and “Southwest”, whereas most of the transit flow is from “Southwest” to “Northwest”.
Figure 1.18: Report for Power Exchange Analysis for a specific line
The composition of electrical loads in low-voltage grids is undergoing rapid change. New types of
electrical consumers such as electric vehicles and heat pumps are connecting to the grid and lead to
a greater utilisation of the existing power system. Existing grids have to be re-evaluated taking these
new consumers into account and weak areas reinforced so as to ensure future network resilience. The
specific power consumption of each consumer is not known. Therefore, it is necessary to evaluate
the network utilisation using load flow calculations based on minimal data. This is done using load
categorisation.
With all these requirements in mind, the new Low Voltage Load Flow Calculation has been developed
as part of the Distribution Network Analysis toolbox, as shown in Figure 1.19. This tool enables the easy
calculation of realistic maximum loadings of lines and transformers, and minimum voltages at all nodes
in a network. In essence, the electrical loads only need to be categorised in order to find bottlenecks in
the grid whilst not overestimating the criticality. The Low Voltage Load Flow Calculation can be used in
both radial and meshed grids and is based on the use of coincidence curves.
Figure 1.19: New function for the Low Voltage Load Flow in the Distribution Network Analysis toolbox
Coincidence curves g(n) take the stochastic consumption behaviour of low voltage loads into account.
They depend on the number of customers n of a specific load type connected to the branch being
investigated. The more consumers connected to a branch of the network, the lower the power of each
individual consumer becomes, statistically. This can be seen in the small example in Figure 1.20. It
should be noted that Kirchhoff’s current law (the sum of all currents is zero at each node) is not valid
in these calculations, as the coincidence factor can be different at each branch connected to the same
node.
The existing Low Voltage Loads (ElmLodlv and ElmLodlvp) and their associated Low Voltage Load Type
(TypLodlv ) have been extended to be usable in the new Low Voltage Load Flow Calculation. As shown
in Figure 1.21, a Coincidence Definition can now be linked to the LV load type or the LV load itself.
In addition to the standard coincidence curve functions, it is possible for users to define their own
curves, using the “Table” input mode. This is useful, for example, for new load categories such as
electric vehicles and heat pumps, where the first n instances are not expected to reduce power due to
their timely correlation and duration of power consumption.
The power consumed by each load, considered individually at each branch in the network, is determined
by the coincidence value gathered from the corresponding curve and scaled with the maximum base
power, which by default is taken from the coincidence curve, too (see settings in Figure 1.21). However,
to improve the calculation results, the known annual energy demand of metering points and the peak
load values can be applied to some individual or all loads in a grid, as shown in Figure 1.22.
In order to be able to quickly run through different scenarios, it is possible to adjust all loads in the
network that share a load category, by using a global scaling object that is linked in the calculation
command. This can simply be done via absolute scaling of the power. However, the process can be
refined using a stochastic scaling, in which the percentage of consumers to be taken into account for
the coincidence is also specified. This makes it possible, for example, to vary the penetration of the grid
with electric vehicles and thus to carry out prognosis studies. In addition to scaling, coupling between
different load categories can be defined in the calculation command with the coupling object, to define
which load categories are to be taken into account for the coincidence and which are not.
In Fig. 1.23, the result of a Low Voltage Load Flow Calculation can be seen. The result parameters are
stored in the same variables as for the normal Load Flow Analysis, therefore the same user-defined
routines and automations can be applied.
Using the Low Voltage Load Flow Calculation, it is now possible to obtain precise results with minimal
data. Since network branches with fewer network connections tend to be rated more critically, voltage
problems can be predicted more easily. At the same time, the loading at the beginning of feeders is not
overestimated.
The Power Park Energy Analysis reports, previously produced as ASCII reports in the Output Window,
are now generated as PDF documents. See section 1.14.1 for details.
The Techno Economical Analysis reports, previously produced as ASCII reports in the Output Window,
are now generated as PDF documents. See section 1.14.1 for details.
The Reliability Assessment reports, previously produced as ASCII reports in the Output Window, are
now generated as PDF documents. See section 1.14.1 for details. The Recovery Scheme report for
Optimal Power Restoration is also now generated as a PDF.
Neural Network Training now supports two different neural network architectures: “Graph Convolutional
Networks” and “Fully Connected Networks”. Graph Convolutional Networks can improve the quality of
the approximation, especially for large power grids and a limited number of training samples. The user
can select the neural network architecture in the Neural Network Training command (see Figure 1.24).
Figure 1.24: New option to select different neural network architectures for Neural Network Training
1.14 Reports
A new reporting concept has been introduced, enabling users to create PDF reports in PowerFactory.
In an ongoing process, the PDF reports will replace the older ASCII reports, which generate output in
the Output Window.
The PDF report documents are compiled from report templates, which are self-contained, project-
independent database objects that encapsulate both data acquisition and document design. While Pow-
erFactory ’s DIgSILENT Library already contains a variety of ready-to-use report templates for different
calculation commands, users can also make their own report templates tailored to their specific needs,
by creating new report templates from scratch or copying and modifying existing report templates.
The Report Generation command makes it easy to prepare professional-looking reports, with options
for customising front pages, headers and footers etc. The user has the flexibility to select from a range
of reports applicable to the calculation just executed, and generate either separate PDF reports or a
single, combined report, as shown in Figure 1.25 below. For calculations which produce result files,
reports can be generated even if the calculation results have been reset.
The use of filters (see Figure 1.26) allows the user to ensure that the report focuses on areas of interest.
Where reports offer different options, these are entered as parameters to the command, on the Param-
eters page.
On the Export page, the user can choose to view the reports inside PowerFactory or export them to a
given location.
When reports are viewed using the new PDF viewer in PowerFactory, element names in the reports
have hyperlinks to actual elements in the active project.
Customised reports
As well as making use of the existing report templates, users can design modified versions of these, or
indeed design their own reports, using the Report Designer, as shown below in Figure 1.27.
Although the new reports are by default generated as PDF documents, users will find that as an
advanced option it is also possible to generate reports in other formats. The following formats are
supported:
• PDF
• MS PowerPoint
• MS Excel
• MS Word
• HTML
• JSON
• Text
2.1 Diagrams
PowerFactory 2023 includes a new symbol editor, meaning that the symbols of network elements in a
single line diagram can now be edited graphically, as shown in Figure 2.1.
Alternatively, the symbol specifications can be defined in the Geometry page on the symbol (IntSym)
using the Modelica language.
The drawing toolbar for dynamic model block diagrams now has a library browser for DSL macros and
Modelica models. Details can be found in section 1.3.1
2.2 Plots
There has been an update to the options for the user to choose line styles, not only for plots but also for
branch elements in a single-line diagram. Previously the same list options was offered in both cases,
but this has now been rationalised, removing options that are not relevant to the context.
The previous “RAW DATA” option, intended for showing data points in plots, has been removed. In its
place, there is a Sample Marker option. This is used to highlight data points with a selected marker.
Context menus are used throughout PowerFactory. Right-clicking on any object gives access to the
many options available to the user, from editing and changing the object to using it as an input to a
calculation. In this latest version, these context menus have been reviewed and updated. Users will
also notice that some of the commonly used actions are now highlighted with familiar icons, making it
easier to find the option you want.
When a user wishes to insert a new object into a PowerFactory project, by using either the context
menu or the new-object icon , the dialog that is presented is now a simple search dialog, as can be
seen in Figure 3.2 below.
The user can search for the required object using a text string, and the dialog will filter the list accord-
ingly. The filtering is based on searching for the string not only in object descriptions (e.g. 2-Winding
Transformer), but also in the class names (e.g. ElmTr2, ComLdf ).
The responsiveness of the user interface (UI) has been greatly improved. The application window now
keeps updating even during long calculations, resulting in better user experience.
3.4 Unicode
With this version, PowerFactory now completely supports Unicode (UTF-8) for text encoding. Any text
written in Latin or other character sets, such as Cyrillic, Arabic or Chinese can be displayed and handled
correctly.
For most users, this change will be transparent, but it offers greater flexibility and will also benefit
customers who need to work with projects that were created using different writing systems.
When users wish to export data from a Network Model Manager or Data Manager, it is usual to simply
copy the data, or use Spread Sheet Format → Copy (with column headers), and then paste. Now
(and already in PowerFactory 2022 SP4), the colouring is kept when the data is pasted into an Excel
spreadsheet. When copying from a tabular report, the bold and italic formatting is also preserved.
Some CIM object attributes have units associated with them, such as km for line length. These units
are now included in the column headers when the objects are viewed in a Data Manager.
3.6 Database
Whenever a user installs a new version of PowerFactory (as opposed to a service pack), the user’s
database needs to be migrated in order to work with this version. This means that, in addition to
updating libraries and configuration data etc., each project needs to be upgraded so that new object
attributes are integrated into the project, and revised dialogs with new and changed options are made
available. This can result in a long down-time for users who work with large multi-user databases.
In PowerFactory 2023, the migration process has been changed, so that only a minimal structural
migration is initially required. All other objects are migrated when they are first needed. Unused objects
or historic object records remain unmigrated without limiting working with PowerFactory. An explicit
time-consuming “project migration” of all object records is no longer required.
This will greatly reduce the downtime of a database when migrating to a new version. In most cases,
the database can be used immediately by all users once the structural migration has completed.
A new tool has been developed to assist users who wish to transfer their PowerFactory multi-user
databases from one database management system to another. The export-import process can be used
for any combination of database systems, namely to/from SQL Server, Oracle and PostgreSQL.
The tool is made available through an additional corporate licence and will work with all supported
versions of PowerFactory.
The advantages of using the tool over approaches such as project export and import are the ease of
use and the fact that all object relations and references within the database are maintained. This latter
point is especially important for customers who make use of PowerFactory ’s archiving mechanism, as
it allows for restoration of projects that were archived before the database was transferred.
The Dynamic Models library within the global DIgSILENT Library has been restructured and updated.
The models are now sorted according to application (e.g. synchronous generator power plants, inverter-
based resources/storage, HVDC, FACTS).
A new set of dynamic models has been added for CGMES Profile 3.1 (related to IEC 61970 – Part 457:
Common Information Model (CIM) for Dynamics Profile, Ed. 2.0). PSS/E compatible dynamic models
have been extended by a new set of models related to PSS/E version 35.2.
For synchronous generator power plants, a new composite model frame is introduced, which is suitable
for dynamic models corresponding to CGMES (Profile 2.4.15, Profile 3.1), IEEE (IEEE Std 421.5, IEEE
PES-TR1), PSS/E (version 35.2), as well as for DIgSILENT dynamic models. This model frame super-
sedes former model frames for synchronous generator power plants that were individually designed for
each set of dynamic models. With the new model frame, it is possible to combine dynamic models from
different model sets, as far as the combination is feasible given the required signal connections of the
individual models. Please note that for data exchange in CGMES or PSS/E data format, only dynamic
models of the corresponding model sets are to be used.
The relay library has been extended to include the follow new models:
• ABB REX 640
• ABB REF 620
• ABB REM 620
In this section, all new models or elements are described. In section 4.2 onwards, readers will find
details of improved modelling capabilities that have been delivered as enhancements to existing models.
In PowerFactory 2023 a new Fault Current Limiter (FCL) model (ElmFcl) has been implemented.
The model represents a resistive type of superconducting fault current limiter (SFCL). During normal
operation the FCL is in superconducting state. When triggered, the resistance increases in the quench-
ing period, consequently limiting the current (see Figure 4.1). Its main purpose is the reduction of
mechanical stress on electrical equipment resulting from fault events.
Different trigger options are available depending on the analysis. In the Complete Method short-circuit
calculation, a current threshold triggers the FCL (see section 1.2.2) while in EMT simulation the FCL
can be additionally triggered with a current slope threshold. It is also possible to trigger the FCL via
external signals and even apply completely different resistance characteristics using dynamic models.
For any other analysis function the FCL is in superconducting state, i.e. steady state resistance value.
The model is available as 3-phase and single phase versions for AC networks and can also be used in
DC networks.
A model of a Static Synchronous Series Compensator (SSSC, or S3C) has been added. An SSSC is
a FACTS device (Flexible AC Transmission System) used for series compensation within AC networks.
It contains a power electronic converter which is series connected to the network, either directly or via
a booster transformer. With an SSSC, the power flow (or current flow respectively) in a meshed AC
network can be controlled very flexibly.
The SSSC is modelled either as a series connected voltage source, or as a two-level converter con-
nected via a series-connected booster transformer. The first modelling approach represents the effect
of an SSSC: a controlled series-injected voltage. It can be used as a generic model of an SSSC,
or to represent a particular type of SSSC that contains a converter directly connected with the grid.
This first approach avoids the need to enter detailed equipment data. The second approach allows
a detailed representation of an SSSC that contains a two-level converter connected to the grid via a
booster transformer. In this case, the type data of the booster transformer and the basic data of the
converter are required.
• Current magnitude
• Effective reactance (with option of additional current droop)
• Voltage injection
• Current droop
The voltage source model supports both symmetrical and phase-wise control.
For power quality assessments (Harmonic Load Flow calculation, impedance Frequency Sweep), the
SSSC is represented by a Thevenin equivalent, with the options to enter a harmonic voltage spectrum
and the frequency-dependent characteristic of the impedance (i.e. of inductance and resistance).
For RMS/EMT simulations, input signals for dynamic control of the SSSC are available. An option to
link a detailed submodel to the SSSC network element is available for use in EMT simulations.
A new Duplex Reactor model (ElmDuplex) has been introduced in PowerFactory 2023. This element is
also known as a Duplex Current Limiting Reactor. The Duplex reactor comprises two coils with negative
magnetic coupling. During normal operation, when the current flows from A towards terminals B and C,
the current flowing in one coil counteracts the voltage drop in the other coil. This effect produces a total
reactance lower than the coil reactance. During a short-circuit, for example at B terminal, the current
through that coil sees the whole reactance of the coil. The characteristic of the duplex reactor of having
a low reactance (and hence lower voltage drops) during normal conditions and a high reactance during
short-circuit conditions makes it attractive for some applications. A typical application of the duplex
reactor, between a generator and two distribution feeders, is shown in Figure 4.3.
A new PLL Phase/Frequency Measurement (StaPll) model has been added, which represents a more
enhanced model compared to the existing Phase Measurement Device PLL (ElmPhi__pll).
The new model provides an easy-to-use and robust phase and frequency measurement element for
RMS and EMT simulations.
In addition to the previously available Synchronous Reference Frame PLL model (SRF-PLL), a Double
Second-Order Generalised Integrator PLL (DSOGI-PLL) can now be used for EMT simulations. This
allows for robust positive sequence tracking capabilities even if subjected to negative sequence or har-
monic components. Additionally, a Notch filter for oscillations at double nominal angular frequency has
been introduced to the SRF-PLL for EMT simulations, to suppress the impact of a negative sequence
voltage component.
With PowerFactory 2023 it is possible to select a thermal rating object (IntThrating) for line sections
(ElmLnesec). If the thermal rating object is dependent on the wind speed, ambient temperature or solar
radiation (see Section 1.1.1), the corresponding meteorological station (ElmMeteostat) is used from the
parent line element (ElmLne).
PowerFactory 2023 supports a zero sequence mutual coupling impedance to couple parallel lines. The
coupling of lines is done via a “Zero sequence mutual impedance” object (ElmZ0mut), without the need
of geometric line tower modelling. The zero sequence mutual impedance is defined for a coupling length
associated with individual start and end positions of the coupled lines, see Fig. 4.4. This modelling
approach is especially used for compatibility with external data formats, for example for CIM/CGMES
import/export.
Figure 4.4: Coupling of parallel lines via zero sequence mutual coupling impedances
4.3 Transformers
In PowerFactory 2023, the existing booster transformer model (ElmTrb) has been enhanced and now
supports a single-phase model with four connections. This model makes it easier, for example, to model
a single-phase railway booster transformer, compared with previous PowerFactory versions. Saturation
of the single-phase booster transformer can be considered in load flow calculation, RMS and EMT
simulations (as was already the case for the 3-phase booster transformer).
In addition, it is now possible to define a frequency dependency of leakage and magnetising impedances
of single-phase and three-phase booster transformers for harmonic studies. The ratio of winding eddy
current losses to copper losses has also been included, which is used for K-factor, Factor-K, FHL and
skin effect assessment.
For short-circuit calculation using the Complete Method, the following device models have been en-
hanced with an option to consider a voltage deadband for the short-circuit current contribution:
• Static generator (ElmGenstat)
• PV system (ElmPvsys)
It expands the calculation of their short-circuit current contribution |∆𝐼1 | to the following:
⎧
⎨0
⎪ if ∆𝑢1 ≤ 0
|∆𝐼 1 | = 𝐾 · ∆𝑢1 · 𝐼N if 𝐾 · ∆𝑢1 < 𝑖max
⎪
𝑖max · 𝐼N if 𝐾 · ∆𝑢1 ≥ 𝑖max
⎩
where:
• ∆𝑢1 is the voltage deviation in p.u. considered for the short-circuit current contribution
• 𝑈1,ldf is the pre-fault voltage in kV
• 𝑈1,rot is the remaining voltage in kV caused by contributions from rotating machines
• 𝑢DB is the voltage deadband in p.u. based on the nominal voltage.
For the doubly fed asynchronous generator and dynamic voltage support short-circuit models, the
voltage deadband is considered for the transient and steady-state short-circuit current contribution.
For the full size converter short-circuit model, it applies to the subtransient contribution as well.
The train and line elements have been enhanced to support train simulations with trains connected to
lines with explicit representation of phase and neutral (1PH-N). This allows train simulations with net-
work models in which the current flow through catenary/contact wires (phase), rails (neutral conductor)
and earth should be distinguished. With line elements that consider the impedances of phase and
neutral explicitly, the voltage drops are also calculated correctly in cases in which the current through
neutral (rails) differs significantly from the current in the phase (catenary/contact wire). This is in contrast
to simple line representations (1PH) without explicit neutral, in which a lumped loop impedance of
catenary/contact wires and rails is entered.
For example, this is especially important for AT systems (2-phase power supply networks with autotrans-
formers), in which a large amount of the return current through rails commutates into the negative feeder
line at the next autotransformer (AT). While in the line section with the train, the current magnitudes in
contact wire and rails are (almost) identical, in the line sections behind the next AT the current in the
rails is significantly smaller than the current in the contact wire.
Figure 4.6: Train with 1PH-N connection in an AT system (the arrows show power flow direction)
The synchronous machine model (ElmSym) has now been enhanced to support three-phase permanent
magnet machine types. Various applications are now better supported, including directly coupled wind
turbines, flywheel energy storage systems, etc.
4.5 Controllers
The PV System element (ElmPvsys) can now be added to a power-frequency controller. For PV systems
where the model is set to “Solar Calculation”, the maximum active power will be the power of the PV
system based on actual solar calculation and the secondary controller will only control the active power
on these elements by reducing it.
A new option for the PWM converter, to consider a deadband for short-circuit current contribution, is
available. See section 4.4.1 for details.
The Pulse Generator element (ElmPulsegen) has been enhanced with two new features:
• The Phase-Shift PWM (PS-PWM) now supports an individual reference signal for each valve.
Unipolar switching is applied in this case. The individual reference signal per valve allows to
improve the capacitor voltage balancing, when a suitable control is implemented.
• The Phase-Disposition PWM (PD-PWM) supports now the sorting algorithm for capacitor voltage
balancing
Surge arresters are typically installed phase to earth at each terminal. However, surge arresters
between phases are required to address overvoltages between phases resulting from switching surges,
lightning surges, or from backflashover caused by a lightning to a pole.
The three-phase surge arrester model StaSua (only considered for EMT simulations) has been en-
hanced. Besides the default YN connection, it now also offers a delta-connection in which three surge
arresters are connected between the phases (a-b, b-c, c-a).
The calculation of the RMS phasors for EMT simulations in Voltage, Current or PQ Measurement
devices (StaVmea, StaImea, StaPqmea) now offers two additional methods for calculating the fre-
quency for the Discrete Fourier Transformation (DFT). In addition to the previously available options
to use the nominal frequency or a Fast Fourier Transformation (FFT), the frequency can now also be
obtained either through a phase angle derivative calculation or a Frequency Locked Loop (DSOGI-FLL)
calculation.
In addition to the measuring the phase-to-earth (PH-E) voltage, the Voltage Measurement (StaVmea)
now also provides the option to output the phase-to-phase (PH-PH) and phase-to-neutral (PH-N) volt-
ages as signals, which for example can be used for dynamic models in RMS and EMT simulations.
The existing ABB starting unit has been completely revised. Among other improvements, a new starting
model for REX 640 series has been added to the list of models supporting this unit. Also, the REL 670
model now offers the overcurrent starting method. For more information, please refer to the Technical
Reference document.
In addition, relay model REL 100 has been enhanced to operate the zones only when the minimum
operating current criterion is fulfilled.
The harmonic blocking function of the differential unit has been enhanced. It can now be used without
the additional measurements if the inputs are already phasor quantities. For more information, please
refer to the Technical Reference document.
5.1 Python
Python 3.11 has been released, and with PowerFactory 2023, Python 3.11 is supported and will be
selected as default Python version. Former versions 3.10, 3.9, 3.8 and 3.7 are also still supported and
can be selected within the configuration.
For more information about the changes of Python 3.11, please refer to the official Python release notes.
The PowerFactory CIM (CGMES) converter now supports the import and export of commonly-used
single- and two-phase elements. These elements are connected, in the PowerFactory model, to single-
and two-phase PowerFactory cubicles, represented by single- and two-phase CIM Terminals. This
development is available for imports/exports using CGMES 3.0.
7 Installation
PowerFactory 2022 was the last PowerFactory version to support 32-bit technology. From PowerFactory
2023 onwards, only a 64-bit version is available.
Discontinuing the support of a 32-bit version of PowerFactory is a natural step for us, as modern CPUs
are ending support for 32-bit architecture. Concentrating on 64-bit architecture enables us to take
advantage of newer technologies, increased performance, support for more memory and better stability.
If you are still using a 32-bit operating system, we strongly advise that you upgrade it to 64-bit, as this
will improve the security and performance of your system.
7.2 Licences
Using PowerFactory in a virtual (cloud) environment has always been possible, with the caveat that the
licence must still be provided via a physical device. Now, however, an alternative licence technology is
available, in the form of a cloud-based softkey.
These cloud-based softkey licences are available for Multi-User and Team-Edition licence servers, and
require a valid maintenance contract.
Existing licences can be exchanged for cloud-based softkeys, which are subject to an additional annual
cost. The use of cloud-based softkeys is not restricted to PowerFactory 2023; the licences can be used
with any supported version of PowerFactory.
DIgSILENT customers
DIgSILENT offices
Partners and representatives
DIgSILENT GmbH
Heinrich-Hertz-Straße 9
DIgSILENT GmbH is certified
72810 Gomaringen (Germany)
to the ISO 9001:2015 standard.
T: +49 7072 9168-0 More information is available at
[email protected] www.tuv-sud.com/ms-cert.