Chapter 39

Distribution Network Tools

39.1 Introduction

The chapter presents the PowerFactory tools for assessment and optimisation of distribution networks.
The areas of analysis are highlighted in Figure 39.1.1 Each section of this chapter introduces the tool,
presenting a general description, the objective function, the optimisation procedure, and the command
dialogs.

Figure 39.1.1: How to access the Distribution Network tools

39.2 Voltage Sag

The Voltage Sag Table Assessment (ComVsag) can be used to assess the expected frequency and
severity of voltage sags within a network during an operating period, and determine the expected
number of equipment trips due to deep sags. The PowerFactory Voltage Sag tool calculates a short-
circuit at the selected load points within the system and uses the failure data of the system components
to determine the voltage sag probabilities.

Voltage sag analysis is similar to probabilistic reliability analysis, in that it uses fault statistics to describe

DIgSILENT PowerFactory 2018, User Manual 863

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

the frequency of faults, and then use these statistics to weight the results of each event and to calculate
the overall effects of failures. However, reliability analysis looks for sustained interruptions as one aspect
of quality of supply, whereas voltage sag analysis calculates the voltage drop during the fault until the
protection system has disconnected the defective component.

This section describes the calculation options, how to perform a Voltage Sag Table Assessment, and
how to view the results.

39.2.1 Calculation Options

[Link] Basic Options Page

Load selection

Reference to the set of load points. A load point can be defined by a busbar, terminal, or load.

Short-circuit command

Displays the short-circuit command that is used. The options for the short-circuit type will be changed
during the voltage sag calculation, depending on the Advanced Options specified in the ComVsag
dialog. However, other settings can be inspected or changed by clicking on the Edit button ( ).

Results

Reference to the results file that is used for storage of results.

Exposed area limit

This defines the minimum remaining voltage for the voltage sag calculation to continue calculating short-
circuits at busbars which are further away from the selected load points. If short-circuits at all busbars
(at a certain distance away from all load points) result in voltages at the load points being higher than
this limit, then no further short-circuit will be analysed.

[Link] Advanced Options Page

TheAdvanced Options page shows the various short-circuit types that can be analysed by the voltage
sag assessment command. All components for which a failure model has been defined use the same
short-circuit frequency. The relative frequency for each type of short-circuit is entered uniformly for all
components.

39.2.2 How to Perform a Voltage Sag Table Assessment

A voltage sag table assessment is performed in two phases:

1. A results file with remaining voltages and short-circuit impedances is created by executing the
ComVsag command. This can be done by selecting one or more nodes, right-clicking and
executing the Calculate. . . → Voltage sag table. . . option, or by initiating the command directly
from the main toolbar by clicking on the Voltage Sag Table Assessment icon ( ).
2. A voltage sag plot is created by selecting one or more of the nodes for which the ComVsag
command was executed, right-clicking and selecting the option Show → Voltage Sag Plot. . .

Alternatively,
• The Load selection in the ComVsag dialog can be completed manually with a set of objects. A
load point is defined by a terminal, a busbar, or by a single-connection element (a load, motor,

864 DIgSILENT PowerFactory 2018, User Manual

 39.2. VOLTAGE SAG

generator, etc.). These kinds of elements can be multi-selected from the single-line diagram or
Data Manager. Once selected, right-click on them and select Define. . . → General Set from the
context-sensitive menu. This set can then be selected as the Load selection.
• A voltage sag plot can be created from the Insert Plot dialog ( ) manually, and the load points
can then be selected from the list of analysed load points.
If several objects are selected which are all connected to the same busbar, then that busbar will be
added only once to the set of load points.

The Load selection parameter in the voltage sag assessment command should be set to use the
SetSelect which has the Used for: Voltage sag table flag set. However, any other selection can be
assigned to the Load selection.

The voltage sag analysis simulates various faults at the selected busbars. The calculation starts with
the selected load points, and proceeds to neighbouring busbars until the remaining voltage at all load
points does not drop below the defined Exposed area limit. The remaining voltages and the short-circuit
impedances for all load points are written to the results file specified by the Results parameter.

After all relevant busbars have been analysed, the sag table assessment continues by analysing short-
circuits at the midpoint of all lines and cables that are connected between the relevant busbars. Again,
the remaining voltages and short-circuit impedances for all load points are written to the results file.

After the complete exposed area has been analysed in this way, the results file contains the values for
Z_F1, Z_F2, Z_F0, Z_S1, Z_S2, Z_S0 and ura, uia, urb, uib, urc, uic for the two ends of all relevant
lines and cables and at their midpoints.

To reduce computation time, the written impedances are interpolated between the ends of a line and the
middle with a second-order polynomial. Then, the remaining voltages and various source impedances
are estimated. These estimated impedances are also interpolated between the ends and the midpoint.
The interpolated impedances are then used to estimate the remaining voltages between the ends and
the midpoints of the lines or cables. This quadratic interpolation gives a good approximation for longer
lines, as well as parallel lines.

39.2.3 Voltage Sag Table Assessment Results

The voltage sag tables are not calculated until a voltage sag plot is constructed. Upon reading the
remaining voltages, short-circuit frequencies and short-circuit impedances from the results file, a voltage
sag table is constructed for each selected load point.

Because there is no single definition of a voltage sag, the plot offers a selection of sag definitions:
• Minimum of Line-Neutral Voltages.
• Minimum of Line-Line Voltages.
• Minimum of Line-Line and Line-Neutral Voltage.
• Positive Sequence Voltage.
Secondly, the x-variable against which the sag frequency will be shown has to be selected. Possible
x-variables are:

• Remaining Voltage.
• Nom. Voltage at Shc-Busbar.
• Fault Clearing Time.
• Short-Circuit Type.
Additionally, the x-variable can be sub-divided according to a split-variable (parameter name: Split Bars
in). Possible split variables are:

DIgSILENT PowerFactory 2018, User Manual 865

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

• no split.
• any of the possible x-variables.

The same parameter cannot be selected for the x-variable and the split-variable.

An example of the resulting voltage sag plot, is shown in Figure 39.2.1.

Figure 39.2.1: Example Voltage Sag Plot

The voltage sag plot always shows the annual frequency of occurrence on the y-axis.

The example plot shows a bar for each load point for each x-variable, which is the Remaining Voltage.
All three loads can be seen to suffer either deep sags (remaining voltage less than 0.4 p.u.), or shallow
sags, although the values at 0.8 p.u. are also significant. Each bar is subdivided to the nominal voltage
at SHC-Busbar. The shallow sags are caused by the low voltage network, as well as the deep sags.
The high voltage network seems to cause moderate voltage sags. This is caused by the fact that the
low voltage networks in this example are radially operated and the higher voltage networks are meshed.
More detailed information about a specific value in the voltage sag plot can be viewed in the balloon
help that appears when placing the mouse over a bar or part of a bar (without clicking).

The voltage sag plot dialog has a Report button which outputs the voltage sag plot data to the output
window. A table for each selected load point will be written in accordance to the selected Voltage Sag
definition, x-Variable and Split Bars in selection.

39.3 Voltage Profile Optimisation

The Voltage Profile Optimisation (VPO) command (ComVoltplan) is used to optimise distribution trans-
former taps over the expected range of network load and generation conditions. It can be selected from
the Distribution Network Tools, as shown in Figure 39.1.1.

The VPO calculation considers two scenarios:

• A maximum demand/minimum generation scenario, or “Consumption Case”.

• A minimum demand/maximum generation scenario, or “Production Case”.

866 DIgSILENT PowerFactory 2018, User Manual

 39.3. VOLTAGE PROFILE OPTIMISATION

It requires that loads be represented as medium voltage (MV) loads (ElmLodmv ). MV load elements
include transformer and LV network parameters, as illustrated in Figure 39.3.1. To show terminal
colouring based on maximum/minimum LV grid voltages, select View → Diagram Colouring from the
main menu (or select the Diagram Colouring icon). Under 3. Other select Results → Voltages / Loading.
Click on Colour Settings, go to the second page of the Voltages / Loading page, and select Consider
LV grid voltages for colouring. In the example below, the minimum voltage is below the lower limit and
the maximum voltage is above the upper limit (the limits set in the colouring options) and the terminal
therefore shows two colours.

The load and generation scaling factors used in the tap optimisation calculation override the values
specified on the “Load Generation Scaling” tab of the Load Flow Calculation; but the Load Flow Calcu-
lation settings remain unchanged.

Figure 39.3.1: MV load example

Note: The transformer tap changer is represented on the LV side of the MV load.

39.3.1 Optimisation Procedure

The optimisation procedure can be summarised as follows:

1. If Distribution Transformer Tap Limits are specified by the user, the tap range of transformers will
be limited within Min. allowed tap position and Max. allowed tap position. This is illustrated in
Figure 39.3.2, where a transformer with seven tap positions is limited to taps “-1” to “2”, which
limits the transformer voltage rise to 7 % and voltage drop to -5 %. The height of each bar is
determined by the voltage rise and voltage drop across the transformer in the production and
consumption cases, respectively.

DIgSILENT PowerFactory 2018, User Manual 867

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

Figure 39.3.2: Distribution transformer tap limits

2. The Upper tap limit and Lower tap limit are calculated based on settings that will keep the range of
expected LV grid voltages within the Upper voltage limit and Lower voltage limit. This is illustrated
in Figure 39.3.3, where the limits are set to between 0.92 p.u. and 1.10 p.u. In cases where only
Production case or Consumption case is set, only the corresponding voltages within the LV grid
will be considered.
3. Both tap positions “0” and “1” would be acceptable, and maintain transformer voltage drop and
LV grid voltages within acceptable limits. The optimisation routine selects the optimal tap position
based on the objective function defined in the command. Figure 39.3.3 shows an example for the
objective function Maximisation of generation. The lower tap limit (position “0” in Figure 39.3.3) is
selected in order to minimise the voltage rise.

868 DIgSILENT PowerFactory 2018, User Manual

 39.3. VOLTAGE PROFILE OPTIMISATION

Figure 39.3.3: Voltage limits for LV grids with the objective function Maximisation of generation selected

The possible scenarios for optimisation are summarised as follows:

1. There is a single tap position that will satisfy both LV grid lower and upper voltage limits -> this
tap is selected.
2. There are multiple tap positions that will satisfy both LV grid lower and upper voltage limits -> the
selected objective function defines which tap position is selected.
3. There is no tap position at all, that will satisfy either the lower nor upper voltage limits of the LV grid
-> PowerFactory will try to secure the LV grid violations of either the minimum (for Maximisation
of consumption) or maximum (for Maximisation of generation) voltage limit.
4. There are tap positions that will satisfy the LV grid upper voltage limit, but all of them violate the
lower voltage limit -> the highest tap position that will not violate the upper voltage limit is selected.
5. There are tap positions that will satisfy the LV grid lower voltage limit, but all of them violate the
upper voltage limit -> the lowest tap position that will not violate the lower voltage limit is selected.
Note that Distribution Transformer Tap Limits, if specified on the Advanced Options page, take prece-
dence over the Upper voltage limit and Lower voltage limit specified on the Basic Options page. This
means that if distribution tap limits are considered, a tap range is first determined which respects these
drop/rise limits over the MV/LV transformer (HV to LV side of the MV/LV transformer). Afterwards, an
optimal tap position which obeys the voltage limits (in the LV feeder) for the selected Calculated cases
is sought within this range.

DIgSILENT PowerFactory 2018, User Manual 869

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

39.3.2 Basic Options Page

• Calculation mode
– Optimisation: the tap of the distribution transformers will be set to the optimal position within
the given limits.
– Verification: the tap position of the distribution transformers will remain as-is. The algorithm
checks whether this setting exceeds the given limits. This could be used to verify whether
the given tap positions in a network are still valid after changes within the LV grid.
• Calculation cases
– Consumption- and production case simultaneously: the consumption- and production
case will be calculated and shown in the results. In addition, a tap position which conforms
in both cases is selected.
– Consumption case only: the consumption case will be calculated with the Voltage limits for
LV grids. Hence, the tap position will be optimised for the highest possible voltage within the
limit. A typical application would be a LV grid with high power consumption and no generation
units.
– Production case only: the production case will be calculated with the Voltage limits for LV
grids. Hence, the tap position will be optimised for the lowest possible voltage within the limit.
A typical application would be a LV grid with a high proportion of photovoltaics.
• Objective function (only available for Calculation method: Optimisation)
– Maximisation of generation: if multiple tap positions of the distribution transformer meet the
given limits for the consumption- and/or production case, the result with the lowest voltage
level will be used.
– Maximisation of consumption: if multiple tap positions of the distribution transformer meet
the given limits for the consumption- and/or production case, the result with the highest
voltage level will be used.
• Voltage limits for LV grids
– Upper voltage limit: upper limit that the LV grid must not exceed (e.g. 1.1 p.u.).
– Lower voltage limit: lower limit that the LV grid must not fall below (e.g. 0.9 p.u.).
• Consumption case (not available for Production case only )
– Load scaling factor: percentage load scaling for the calculation of the consumption case
(e.g. 100 %).
– Generation scaling factor: percentage generation scaling for the calculation of the con-
sumption case (e.g. 0 %).
• Production case (not available for Consumption case only )
– Load scaling factor: percentage load scaling for the calculation of the production case (e.g.
25 %).
– Generation scaling factor: percentage generation scaling for the calculation of the produc-
tion case (e.g. 100 %).
• Load Flow calculation: a reference to the Load Flow command used by the optimisation algo-
rithm. A copy is made of the command meaning that any changes made do not affect the settings
of the original Load Flow command.

39.3.3 Output Page

In cases where Calculation method “Consumption- and production case simultaneously” is used, the
following option is available:

• Shown results
– Consumption case: the results for the consumption case are shown.
– Production case: the results for the production case are shown.

870 DIgSILENT PowerFactory 2018, User Manual

 39.3. VOLTAGE PROFILE OPTIMISATION

39.3.4 Advanced Options Page

Distribution Transformer Tap Limits

Transformer Maximum Allowed Voltage Rise and Maximum Allowed Voltage Drop can be optionally
specified. These limits restrict the feasible range of taps in the optimisation procedure.

39.3.5 Results of Voltage Profile Optimisation

The result of the Voltage Profile Optimisation can be shown as a tabular or ASCII report, or as a voltage
profile plot.

Tabular and ASCII report

The tabular or ASCII reports, which show the recommended tap settings, including details of MV loads
with critical voltage drop or rise can be accessed after the Voltage Profile Optimisation has been calcu-
lated. This is done by clicking on “Reports Voltage Profile Optimisation” ( ). An example of the Optimal
Transformer Tap Positions section of the report is shown below in Figure 39.3.4 (results consistent
with Figure 39.3.1 and the discussion in Section 39.3.1). In the case where only the production or
consumption case are calculated, only the corresponding results will be available in the last columns
(voltages).

Figure 39.3.4: Voltage profile results

The recommended tap settings are also available on the Flexible Data page of MV loads under the
Voltage Profile Optimisation calculation parameter “c:nntap”. To update the network model with the
recommended tap settings, the user may either manually adjust MV load tap positions, or click the
Update Database icon on the main toolbar ( ), and update the case with the calculated distribution
transformer taps.

Voltage profile diagram

To display a plot of the resultant profile for one feeder for the consumption case, production case or both,
select the Voltage Profile Plot icon ( ). Figure 39.3.5 shows an example plot for the consumption case
only:

DIgSILENT PowerFactory 2018, User Manual 871

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

Figure 39.3.5: Voltage profile plot for the consumption case

39.4 Tie Open Point Optimisation

The function of the Tie Open Point Optimisation (TOPO) (ComTieopt) is to optimise a radial system
of connected feeders by determining the best location for network open points. An open point can be
moved by the TOPO tool by opening and closing switches on the networks to be optimised.

This chapter is separated into three sub-sections. Firstly, the steps to access the TOPO tool are
described. Next, the background and function of the TOPO tool is presented and finally the procedure
for running a Tie Open Point Optimisation is described. The Tie Open Point Optimisation Command
can accessed as shown in Figure 39.1.1

39.4.1 Tie Open Point Optimisation Background

The function of the Tie Open Point Optimisation (TOPO) tool is best explained using an example.
Consider the network illustrated in Figure 39.4.1

872 DIgSILENT PowerFactory 2018, User Manual

 39.4. TIE OPEN POINT OPTIMISATION

Figure 39.4.1: Example network for Tie Open Point Optimisation

The network consists of three feeders, one from each of the three “stations”. Each feeder begins at a
“station” and ends at one of the two illustrated open points. The two open points in this network are not
necessarily the optimum open points. For example, it might be more economic (i.e. less network losses
and / or less impact of outages) to shift these open points by closing the open switches and opening
two switches in different positions on the feeders. The purpose of the TOPO tool is determine these
optimum open points automatically. Additionally, the TOPO tool can automatically consider network
voltage and thermal constraints - for instance it might be economic to shift an open point in terms of
reducing systems losses, however doing so might cause a cable to overload.

39.4.2 How to run a Tie Open Point Optimisation

This section describes the procedure for running a Tie Open Point Optimisation (TOPO) calculation.
The steps are summarised below, and discussed in more detail in the following sections:
• How to Create Feeders.

• How to configure the Tie Open Point Optimisation Command.

• How to configure constraints for the Tie Open Point Optimisation.
• How to configure the Advanced Options.
• How to configure Reliability Options.

How to Create Feeders

The TOPO tool requires that feeders are defined for the section of the network that you wish to opti-
mise. Additionally, the TOPO tool only works on radial feeders - mesh systems cannot be optimised

DIgSILENT PowerFactory 2018, User Manual 873

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

automatically. Furthermore, it is recommended that the target feeders for optimisation do not have any
overloaded components or voltage violations in the base case.

To define a feeder, right-click the cubicle at the beginning of the feeder and select the Define → Feeder.
Alternatively, for fast creation of multiple feeders right-click the bus the feeder/s are connected to and
select the option Define → Feeder. More information on feeders and feeder creation can be found in
Chapter 15: Grouping Objects, Section 15.5.

How to configure the Tie Open Point Optimisation Command

After a set of feeders has been defined, open the TOPO tool and configure the basic options:

1. Click the Change Toolbox icon ( ) and select Distribution Network Tools.

2. Open the dialog for the Tie Open Point Optimisation tool ( ).
3. Use the selection control for Feeding Points to select previously defined feeder/s, or a feeder
“Set”. If the Select option is chosen and multiple feeders are selected, a “Set” of feeders will
automatically be created within the active study case. By default the set will be named ’Tie Open
Point Optim. - Feeder Set’.

Note: It is generally recommended to define all feeders in the network as Feeders, and to conduct
a TOPO calculation for ’All Feeders’.

4. Select the desired Objective Function to minimise losses and / or reliability indices. If Optimisation
of Reliability Indices or Cost Optimisation (Losses + Reliability) is selected, complete the required
fields on the Reliability page, see (How to configure Reliability Options).
5. “Method” can be selected, where the optimisation explores the meshes iteratively, uses a stochas-
tic optimisation according to section 39.8 (Simulated Annealing or Genetic Algorithm).

6. “Balanced, positive sequence” or “Unbalanced” network representation can be selected. The

Load-flow command referenced below these radio buttons is automatically adjusted to the correct
calculation method based on this selection.
7. Optional: You can inspect and alter the settings of the load-flow command that is used for deter-
mining the losses and identifying the constraints of the system by clicking the blue selection arrow
next to load-flow command.
8. Optional: Change the “Saving of solution” option. The two options are as follows:
• Change Existing Network (Operation Scenario). This is the default option. The TOPO tool
modifies the base network model. Note that if a variation is active, the changes will be
implemented in the variation.
• Record to Operation Scenario. If you choose this option a selection control appears and you
can choose an existing operation scenario to save the results of the Optimisation procedure
to. Alternatively, you can leave the selection empty and PowerFactory automatically activates
a new Operation Scenario called “Tie Open Point Optimisation Results”. Any changes made
to the network as a result of the optimisation procedure are stored within this operation
scenario. You can revert to the original network by disabling the scenario.
9. Optional: Disable the “Report” flag. This control, enabled by default, allows you to turn off the
automatic printing of an ASCII report to the output window.
10. Optional: Select the Before Optimisation and After Optimisation results objects.
How to configure constraints for the Tie Open Point Optimisation

It is optional whether you choose to consider thermal and voltage constraints for the Tie Open Point
Optimisation. If you wish to consider constraints follow these steps:
1. Open the Tie Open Point Optimisation dialog and select the Constraints page.

874 DIgSILENT PowerFactory 2018, User Manual

 39.4. TIE OPEN POINT OPTIMISATION

2. Optional: Choose to enable or disable the option Consider Thermal Constraints. If enabled,
the TOPO tool will automatically consider thermal constraints in the network. Therefore, if an
optimal point were to cause an thermal overload on any system component, then this would not
be considered as a valid open point for reconfiguration of the system. There are two more options
for thermal constraints:
• Global constraint for all components. This is the default option. If enabled you must enter a
maximum thermal loading percentage in the Max. thermal loading of components field. Note
this option overrides the individual component thermal limits.
• Individual constraint per component. Select this option to automatically consider each com-
ponent’s unique thermal rating. Note, the thermal rating for each component is determined
by the field Max Loading within the Tie Open Point Optimisation page of each component.
3. Optional: Choose to enable or disable the option Consider Voltage Constraints. If this option is
enabled then each terminal in the system is checked against the Lower and Upper limit of allowed
voltage. If a particular open point causes a voltage violation, then such an open point cannot be
considered as “optimal”. There are two options for configuring the upper and lower voltage limits:
• Global constraints for all terminals (absolute value). If you choose this option then you must
enter an upper and lower voltage limit in the two corresponding fields within this dialog box.
• Individual constraint per terminal. If you choose this option, then each terminal has a unique
voltage limit which is assigned on the Tie Open Point Optimisation page of each terminal
(note that this excludes Substation internal nodes).
4. Optional: Choose to enable or disable the option Consider Voltage Drop / Rise. If this option is
enabled then each feeder in the system is checked against the Maximum Voltage Drop / Rise. If
a particular open point causes a voltage violation, then such an open point cannot be considered
as “optimal”. There are two options for configuring the maximum voltage drop / rise limits:
• Global constraints for all feeders (percent). If you choose this option then you must enter the
Maximum Voltage Drop and Maximum Voltage Rise in the two corresponding fields within
this dialog box.
• Individual constraint per feeder. If you choose this option, then each feeder has a unique
voltage drop / rise limit which is assigned on the Tie Open Point Optimisation page of each
feeder.
5. Choose the ignore all constraints for. . . option. You can use these options to optionally ignore
constraints where the nominal voltage is above or below user-defined thresholds entered here.
This can be useful for example to exclude all LV systems (say less than 1 kV) from the constraints
identification process as it may be acceptable to have these systems outside the “normal” range.
How to configure the Advanced Options

The options in the Advanced page can generally be left on default values. The options are described
as follows:

• Switches to be optimised. These options configure the switches / elements considered by the
optimisation procedure.
– All switches. All switches will participate in the optimisation.
– Selected switches. Only the selected switch types will participate in the optimisation. For
example, if “Circuit-Breaker” and “Load-Breaker-Switch” are ticked, then both circuit breakers
and load breakers will be considered by the optimisation. The switch type is defined on the
switch element “Basic Data” page. Similar to Switch type, only the selected control types will
participate in the optimisation. The control type is defined in switch element “Reliability” page
in the “Sectionalising” field. Switches are considered in the optimisation only when its switch
type AND the control type satisfies the selected settings.
– Assume each edge element is switchable. If selected, lines that do not have a switch can
also be switchable (either out of service or in service).
• Maximum number of outer loops. This option controls the maximum number of outer loops which
is the total number of times the optimisation procedure will be repeated when searching for an
optimal solution.

DIgSILENT PowerFactory 2018, User Manual 875

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

• Maximum change in system losses. This option determines the threshold above which a change
in open point is considered. If the reduction in losses is below this threshold, the iteration will stop.
• Constraint Priority options can be selected for the relevant constraints. For example, consider the
following scenario:
– The TOPO calculation is to consider Global Thermal constraints, with the Max. thermal
loading of components set to 100 %, and Global Voltage Constraints with a Lower limit of
0.90 p.u.
– The constraint priorities for loading constraint is set to 1, and for voltage lower limit is set to
3.
– In the current configuration, a line is loaded to 102 % of rating.
– Shifting the open point causes the voltage at a terminal on an adjacent feeder to decrease
5 % below 0.90 p.u. (i.e. 0.855 p.u).
– As a result of the priorities, the thermal loading deviation will be “penalised” to a greater
extent than the voltage deviation, and the open point will change, despite the resultant voltage
deviation.

How to configure Reliability Options

If Optimisation of Reliability Indices is selected, the user may select between optimisation of SAIFI or
EPNS indices on the Reliability page. Where:

• SAIFI (System Average Interruption Frequency Index) in units of [1/C/a], indicates how often the
average customer experiences a sustained interruption in one year. Note that the number of
customers at each load should be defined on the Reliability page.
• EPNS (Expected Power Not Supplied) is in units of [MW]. Multiplying EPNS by the study duration
gives the expected energy not supplied.
Contingency definitions can be optionally considered for Busbar / terminals, Lines / Cables, and Trans-
formers.

If Cost Optimisation (Losses + Reliability) is selected, Costs for Losses and Interruption costs per
customer should be defined, as these are used in the Objective Function calculation to determine
the network configuration that optimises both Losses and Reliability.

39.5 Backbone Calculation

This section describes the Backbone Calculation command (ComBbone) dialogs and presents an
example calculation. To run a Backbone Calculation, either:

• Select the Backbone Calculation icon under Distribution Network Tools as shown in Figure 39.1.1.

• From the Data Manager select and then right-click previously defined feeders and click Calculate
→ Backbone Calculation. . . .
• From the main menu, select Calculation → Distribution Network Tools→ Backbone Calculation.

The Backbone Calculation is used to determine the main paths between adjacent feeders connected via
open points, that may serve to restore lost load in case of failures inside a feeder. The command creates
objects in the Network Data folder (ElmBbone) with the Backbones constituent network elements. This
simplifies visualisation of the main path(s) between feeder(s), particularly in large distribution networks
where the main paths may not be apparent from the single line diagram.

Backbone objects are created for all feeders or a user-defined set of feeders based on path load, cross-
section, network structure, or scoring method criteria. The command can optionally consider existing
remote controlled switches at open points, and the availability of connections to alternative transformers
or substations when creating Backbones.

876 DIgSILENT PowerFactory 2018, User Manual

 39.5. BACKBONE CALCULATION

From the Backbone dialog, the Backbone contents (elements) can be viewed, marked in the graphic,
and checked (see example in Section 39.5.4). The Check Backbone button is used to verify that the
backbone object still defines a valid inter-feeder path matching its calculated parameters.

39.5.1 Basic Options Page

Generate backbones

Specify all feeders or a user-defined set of feeder/s for the Backbone Calculation.

Calculation based on:

Note: For all calculation methods, feeder is supposed to be operated radially must be selected on the
Basic Options page of the relevant Feeder/s.

• Path load: Backbones are determined based on the MVA load on the paths between adjacent
feeders.
– (Optional) specify the max. number of backbones per feeder.
– Optionally select to Report results to the output window, including details of backbone open
points.
– Pointer to load-flow command (note for balanced calculations only).
• Cross section: Backbones are determined based on the cross-section of lines/cables connecting
adjacent feeders.
– (Optional) specify the max. number of backbones per feeder.
– Choose to determine backbone using either the mean cross section of lines in the path or
the minimum cross section in path.
– Optionally select to Report results to the output window, including details of backbone open
points, and minimum and mean cross-section.
• Network structure: Backbones are determined based on the network structure. If none of the
options are selected, Backbones are calculated for all feasible inter-feeder paths.
– (Optional) create backbones only if path leads to different substation.
– (Optional) create backbones only if path leads to different HV/MV-transformer.
– (Optional) create backbones only if tie open point is remote-controlled (as specified on the
Reliability page of each switch).
– Optionally select to Report results to the output window, including details of backbone open
points.
• Scoring method: Backbones are determined using a scoring algorithm based on the restoration
ability of the adjacent feeder. Scoring method settings are entered on the Scoring Settings page.
– (Optional) specify the max. number of backbones per feeder.
– Optionally select to Report results to the output window, including details of backbone open
points, and loading/voltages of limiting elements.
– Pointer to load-flow command (note for balanced AC calculation only).

39.5.2 Scoring Settings Page

If scoring method is selected on the Basic Options page, enter scoring settings on the Scoring Settings
page. Backbones are determined based on the restoration ability of every inter-feeder path using
Topology, Loading violation, and Voltage violation criteria.

For each criteria satisfied, the path receives the entered number of points. The path with the greatest
number of points is the “winning” path.

DIgSILENT PowerFactory 2018, User Manual 877

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

Topology scoring

Define scoring settings for Topology scoring criteria:

• Path leads to different substation.
• Path leads to different HV/MV-transformer.

• Tie open point is remote controlled.

• Greater than a specified number of remote-controlled switches on path. A path to another Feeder
receives the entered number of points if more (closed) remote-controlled switches than the entered
number are on the path of the Backbone contained in the initial feeder.

Loading violation scoring

Assign Points for loading violations based on individual loading constraints or global loading constraints.
If no element is overloaded, the calculation assigns the specified number of points. If global loading
constraints is selected, then Max. Loading should also be defined.

Define scoring settings for Loading violation scoring criteria:

• Restoring transformer (restoration mode). Consider a path from initial “feeder A” to “feeder B”.
“Feeder A” is de-energised and the connection to “feeder B” via the tie open point is closed.
A load flow is calculated in this so-called restoring mode and the entered number of points is
assigned if the supplying HV/MV-transformer is not overloaded.

• On backbone of restoring feeder (normal mode). Consider a path from initial “feeder A” to “feeder
B”. A load flow is calculated (in so-called normal mode) and the entered number of points is
assigned if no element on the potential backbone path contained in “feeder B”, the restoring feeder
is overloaded in the base case.
• On complete backbone (restoration mode). Consider a path from initial “feeder A” to “feeder B”.
“Feeder A” is de-energised and the connection to “feeder B” via the tie open point is closed. A load
flow is calculated in this so-called restoring mode and the entered number of points is assigned if
no element on the potential backbone path is overloaded.
• In complete feeder (restoration mode). Consider a path from initial “feeder A” to “feeder B”. “Feeder
A” is de-energised and the connection to “feeder B” via the tie open point is closed. A load flow
is calculated in this so-called restoring mode and the entered number of points is assigned if
no element in the complete resulting feeder is overloaded (not only on the backbone as for the
previous option).
Voltage violation scoring

Define scoring settings for voltage violation criteria based on individual voltage drop/rise constraints or
global voltage drop/rise constraints. If global voltage drop/rise constraints is selected, then Max. drop
and Max. rise should also be defined. If no voltage limits are violated, the calculation assigns the
specified number of points.

• On backbone of restoring feeder (normal mode). Consider a path from initial “feeder A” to “feeder
B”. A load flow is calculated (in so-called normal mode) and the entered number of points is
assigned if no terminal on the potential backbone path contained in “feeder B” violates its voltage
drop constraint and voltage rise constraint.

• On complete backbone (restoration mode). Consider a path from initial “feeder A” to “feeder B”.
“Feeder A” is de-energised and the connection to “feeder B” via the tie open point is closed. A load
flow is calculated in this so-called restoring mode and the entered number of points is assigned if
no terminal on the potential backbone path violates its voltage drop and rise constraint.

• In complete feeder (restoration mode). Consider a path from initial “feeder A” to “feeder B”. “Feeder
A” is de-energised and the connection to “feeder B” via the tie open point is closed. A load flow
is calculated in this so-called restoring mode and the entered number of points is assigned if no

878 DIgSILENT PowerFactory 2018, User Manual

 39.5. BACKBONE CALCULATION

terminal in the complete resulting feeder violates its voltage drop and rise constraint (not only on
the backbone as for the previous option).

39.5.3 Tracing Backbones

When a Backbone is calculated, it always contains a connection to another Feeder via a tie open point.
In the worst case of an outage close to the feeding point of the initial feeder, the initial feeder is de-
energised by opening its feeding switch and restored by the second Feeder via the tie open point. These
restoration steps can be simulated for an existing Backbone using the Backbone trace functionality. The
trace buttons are located beside the ComBbone command, and can also be accessed via the main
menu Calculation → Distribution Network Tools→ Start trace. . . .

39.5.4 Example Backbone Calculation

Consider a case where there are two parallel feeders with multiple open-points. A Backbone calculation
is conducted based on a criteria of minimum cross section in path, and with the Max. number of
backbones per feeder set to “1”. Backbone objects are created within the Network Data folder.

To highlight Backbones, from the main menu select View → Diagram Colouring (or select the Diagram
Colouring icon). Under 3. Other select Topology Feeders. Click on Colour Settings, and on the Feeders
page select Highlight backbones.

Figure 39.5.1 shows the result, where the path through “Open Point 2” is highlighted as a result of the
cross section of conductors in this path. Refer to Section 39.5.3 for details of how to trace the Backbone
restoration steps.

Figure 39.5.1: Example Backbone Calculation

DIgSILENT PowerFactory 2018, User Manual 879

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

39.6 Optimal Capacitor Placement

Optimal Capacitor Placement (OCP) is an automatic algorithm that minimises the cost of losses and
voltage constraints (optional) in a distribution network by proposing the installation of new capacitors
at terminals along the selected feeder/s. The optimal size and type of capacitor is selected from a
list of available capacitors entered by the user. The algorithm also considers the annual cost of such
capacitors and only proposes new capacitors for installation when the reduction of energy loss and
voltage constraint costs exceeds the annual cost of the capacitor (investment, maintenance, insurance
etc).

To access the OCP tool, select the OCP toolbar from the toolbar selection window as illustrated in
Figure 39.6.1.

Figure 39.6.1: Optimal Capacitor Placement Tools

The buttons in the OCP toolbar are as follows:

• The main Optimal Capacitor Placement command is started with the Calculate Optimal Capacitor
Placement icon ( ). The command and the various user-defined options are described in detail
in Sections 47.3.1 to 43.3.3.

• After a successful optimisation, the list of nodes (terminals) where capacitors are proposed for
installation can be accessed by selecting the Show nodes with New Capacitors icon ( ).

• Following a successful OCP, the list of proposed capacitors can be accessed with the Show New
Capacitors icon ( ).

• The Remove previous solution icon ( ) deletes the results (removes all placed capacitors) from
a previous OCP routine.
• To list all results from the OCP in a ASCII text report printed to the output window use the Output
Calculation Analysis icon ( ). The report also displays the original system losses and voltage
constraint costs and such costs after the installation of the proposed capacitors.

880 DIgSILENT PowerFactory 2018, User Manual

 39.6. OPTIMAL CAPACITOR PLACEMENT

39.6.1 OCP Objective Function

The OCP optimisation algorithm minimises the total annual network cost. This is the sum of the cost of
grid losses, the cost of installed capacitors, and optionally the fictitious penalty cost of voltage violations:

𝑚
∑︁ 𝑛
∑︁
𝑇 𝑜𝑡𝑎𝑙𝐶𝑜𝑠𝑡𝑠 = 𝐶𝐿𝑜𝑠𝑠𝑒𝑠 + (𝐶𝐶𝑎𝑝𝑖 ) + (𝐶𝑉 𝑜𝑙𝑡𝑉 𝑖𝑜𝑙𝑖 ) (39.1)
𝑖=1 𝑖=1

Where:

• 𝐶𝐿𝑜𝑠𝑠𝑒𝑠 is the annual cost of grid losses (i.e. including the grid losses, not only the feeder/s
for which the optimal capacitor placement is performed). Essentially, this is the 𝐼 2 𝑅 loss of all
elements in the network.
• 𝐶𝐶𝑎𝑝𝑖 is the annual cost of a capacitor (investment, maintenance, insurance), as entered by the
user in the list of possible capacitors. m is the total number of installed capacitors.
• 𝐶𝑉 𝑜𝑙𝑡𝑉 𝑖𝑜𝑙𝑖 corresponds to a fictitious cost used to penalise a bus (terminal) voltage violation. 𝑛
is the total number of feeder terminals with voltage violations.
Note that if the OCP is not able to reduce the Total Costs by installation of a capacitor/s, the following
message will be reported:

Costs can not be reduced with the given “Available Capacitors”

Evaluating the Voltage Violation Cost

As there is no ’real’ cost for a voltage violation, if the user wants to consider voltage violations as part
of the OCP algorithm, they must assign a ’fictitious’ cost for such violations. The voltage violation
cost is calculated based on the user specified voltage limits and penalty factors. The voltage limits
are defined in the ’Basic Options’ tab of the OCP command dialog (’vmin’ and ’vmax’ parameters, see
Section 47.3.1: Basic Options Page). The penalty factors are defined in the ’Advanced Options’ tab of
the same command (’weight’ and ’weight2’ fields, see Section 43.3.3: Advanced Options Page). The
penalty values are applied for voltages inside the admissible voltage band (parameter ’weight’: Penalty
Factor 1) and for voltages outside the admissible band (parameter ’weight2’: Penalty Factor 2).

There are two possible situations for a terminal voltage and the calculation for the fictitious voltage
violation cost is slightly different for each situation. The two situations are explained as follows:
1. In situation one, the voltage 𝑈 of a terminal is within the allowed voltage band (between vmax and
vmin) but deviates from the nominal voltage of 1 p.u. The penalty cost is calculated as:

𝐶𝑉 𝑜𝑙𝑡𝑉 𝑖𝑜𝑙 = 𝑤1 · ∆𝑈 (39.2)

where:
∆𝑈 is the absolute deviation from the nominal voltage in p.u. (∆𝑈 = |𝑈 − 𝑈𝑛 |).
𝑤1 is the penalty factor (parameter ’weight’) inside the admissible voltage band in $/% from the
’Advanced Options’ tab.
2. For situation two, the voltage 𝑈 is outside the allowed voltage band (greater than vmax or less
than vmin) and the penalty cost is calculated as:
𝑈 > 𝑈𝑛 + ∆𝑈𝑚𝑎𝑥 , if voltage is higher than max. limit:
𝐶𝑉 𝑜𝑙𝑡𝑉 𝑖𝑜𝑙 = 𝑤2 · (∆𝑈 − ∆𝑈𝑚𝑎𝑥 ) + 𝑤1 · ∆𝑈
or
𝑈 < 𝑈𝑛 − ∆𝑈𝑚𝑖𝑛 , if voltage is lower than min. limit:
𝐶𝑉 𝑜𝑙𝑡𝑉 𝑖𝑜𝑙 = 𝑤2 · (∆𝑈 − ∆𝑈𝑚𝑖𝑛 ) + 𝑤1 · ∆𝑈
where

• ∆𝑈 is the absolute deviation from the nominal voltage 𝑈𝑛 in p.u.

DIgSILENT PowerFactory 2018, User Manual 881

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

• 𝑈𝑛 + ∆𝑈𝑚𝑎𝑥 is the higher voltage limit in p.u.

• 𝑈𝑛 − ∆𝑈𝑚𝑖𝑛 is the lower voltage limit in p.u.
• 𝑤1 is the penalty factor (parameter ’weight’) for voltage inside the admissible voltage band in
$/% from the ’Advanced Options’ tab.
• 𝑤2 is the penalty factor (parameter ’weight2’) for voltage outside the admissible voltage band
in $/% from the ’Advanced Options’ tab.
The algorithm can be summarised in as follows:
• If the voltages are inside the admissible band the penalty cost applied is equal to 𝑤1 · ∆𝑈
• If the voltages are outside the admissible band the penalty cost applied is equal to the penalty
inside the band (𝑤1 · ∆𝑈 ) plus the factor 𝑤2 · (∆𝑈 − ∆𝑈𝑙𝑖𝑚 , with ∆𝑈𝑙𝑖𝑚 being either the maximum
or the minimum limit value of the admissible band.
Figure 39.6.2 illustrates the concept of the voltage band violation cost.

Figure 39.6.2: Fictitious cost assigned by voltage band violations

39.6.2 OCP Optimisation Procedure

To find the optimal configuration of capacitors, PowerFactory applies the following steps:
• First a sensitivity analysis determines the ’best’ candidate terminal; This involves evaluating the
impact on the total cost (Losses + Voltage Violations) by connecting the largest available capacitor
from the user-defined list of capacitors to each target feeder terminal. At this stage the cost of the
largest capacitor is excluded.
• Terminals are ranked in descending order of total cost reduction. The terminal that provides the
largest cost reduction becomes the ’best’ candidate terminal for a ’new’ capacitor.
• The optimisation routine then evaluates the cost reduction at the candidate terminal using each
available capacitor from the user-defined list including the cost of each capacitor. The ’best’
capacitor is the one that reduces the cost the most when also considering the annual cost of that
capacitor.
• Repeat step one but any terminals that have previously been selected as candidates for capacitor
installation are not included in the ranking of candidate terminals. The algorithm stops when all

882 DIgSILENT PowerFactory 2018, User Manual

 39.6. OPTIMAL CAPACITOR PLACEMENT

terminals have had capacitors installed, or the installation of capacitors cannot reduce costs any
further.

Note: If Load Characteristics are considered, then the above algorithm will be completed for every
independent load state. See Section 39.6.5 for how the load states are determined.

39.6.3 Basic Options Page

Feeder

Here the target feeder for the optimum capacitor placement is selected. The feeder is a special
PowerFactory element that must be created by the user before it can be selected in this dialog (for
information about feeders refer to Chapter 15: Grouping Objects 15.5 (Feeders)).

Method

• Optimisation; This option calculates the optimal placement for capacitors using the methodology
described in Section 39.6.2. The output of the analysis is printed to the output window and any
new capacitors are connected to the target terminal/s if the ’Solution Action’ - ’Install capacitors’
is selected.
• Sensitivity Analysis; Performs the sensitivity analysis that ranks the candidate terminals according
to their impact on the total loss cost excluding the capacitor cost. The output is presented in the
output window. This option provides a quick indication of the most effective place for a single
capacitor. No capacitors are installed if this option is selected.

Network Representation

Here either a ’Balanced, positive sequence’ or a ’Unbalanced’ network representation can be selected.
The Load-flow command referenced below these radio buttons is automatically adjusted to the correct
calculation method based on this selection.

Constraints

Here the voltage constraint limits (upper and lower) can be entered, along with a limitation for the ’Total
Reactive Power of all Capacitors’ that can be added by the Optimal Capacitor Placement tool. The
total reactive power of all capacitors includes all existing capacitors along the feeder plus any more
capacitors proposed by the optimisation tool.

Note: The voltage constraints are meaningless if penalty factors for deviations outside of the nominal
range are not entered as discussed in detail in Section 39.6.1: OCP Objective Function.

Energy Costs

The energy cost ($/kWh) can be entered manually or taken from an External Grid. Note, if more than
one External Grid exists in the network, the algorithm takes the first External Grid by database ID. The
calculation of the cost of the network losses is as follows:

𝑇 𝐶 = 𝑀 𝐶 × 8760 × 𝐿 (39.3)

where:
𝑇 𝐶 is the total cost per annum in $;
𝑀 𝐶 is the energy cost of losses in $/kWh; and
𝐿 is the total losses in kW.

DIgSILENT PowerFactory 2018, User Manual 883

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

Note that if characteristics are applied to the loads and the analysis uses the option ’Consider Load
Characteristics’ (see Section 39.6.5), then the losses calculation becomes a summation over each time
state considered.

Note: The default energy cost units are $/kWh. However, this can be changed to Euro or Sterling (£) via
the project settings from the main menu bar. Edit → Project. . . Project Settings→ Input Variables
tab→ Currency Unit.

Solution Action
• Report only (do not modify network); The result of the optimisation is a report to the output window
only, no modifications are made to the network model.
• Install capacitors (modify network). If this option is chosen, the capacitors that the optimisation
proposes for the network will be automatically installed. However, note that the single line diagram
is not automatically updated, only the network model database. To draw the installed capacitors
in the SLD the option must be selected in the Advanced Options page (see Section 43.3.3). The
placed capacitors can be also visualised on the Voltage Profile Plot of the Feeder, see (Viewing
results on the Voltage Profile Plot) in Section 39.6.7.

39.6.4 Available Capacitors Page

On this page, the user defines the available capacitors for the OCP command. One capacitor is entered
per row. To add a new capacitor, right-click within any cell and select the option ’Insert Rows’, ’Append
Rows’ or ’Append n Rows’. The following fields are mandatory for each row:

• Ignored; If this option is checked, then the capacitor specified in this row will be ignored by the
OCP command.
• Q per Step Mvar; Here the nominal reactive power of the capacitor in Mvar per step is specified.
• Switchable; If this option is enabled then the algorithm can use a capacitor with multiple steps.
• Max. Step; If the ’Switchable’ option is enabled, then this option specifies the maximum number of
steps available to the optimisation algorithm. The maximum available reactive power is therefore
Max. Step * Q per Step Mvar.
• Technology; Specifies whether the capacitor is Three-phase or Single-phase.
• Cost; Important. This is the total cost of the capacitor bank per annum. This is a critical parameter
for the OCP command as the capacitor will only be installed if the losses offset by its installation
are greater than the annual cost of the capacitor.

Note: It is theoretically possible to force the installation of a particular capacitor at an optimal location
on a feeder by defining a very low cost for the capacitor, and limiting the number of capacitors to
say, one.

Available Capacitors
• Allow use of each capacitor multiple times; This is the default option and it means that every
capacitor in the list can be used at more than one feeder terminal (multiple times).
• Use each capacitor only once; If this option is enabled then each capacitor can only be placed at
one terminal along the target feeder.

Treatment of 3-phase capacitors This option allows the specification of the ’technology’ type for 3-
phase capacitors. This option is only available when the ’Network Representation’ is set to ’Unbalanced’
in the Basic Options page.

884 DIgSILENT PowerFactory 2018, User Manual

 39.6. OPTIMAL CAPACITOR PLACEMENT

39.6.5 Load Characteristics Page

If load characteristics are to be considered by the optimisation algorithm, then the option ’Consider Load
Characteristics’ should be enabled on this page.

Load States

Two options are available:

1. ’Use existing Load States’; If this option is selected then the system load state that is active in
the system (the load state observed as a result of a single load-flow at the current point in time)
will be used as the load state for the optimisation algorithm. For example, if there is a 1 MW load
with a active characteristic that gives the current load value of 0.6 MW, then the load used for the
optimisation will be 0.6 MW, not 1 MW.
2. ’Create Load States’; If this option is selected then PowerFactory automatically discretises all load
characteristics into a number of ’states’ using a sophisticated algorithm. The algorithm iterates
through every hour of the selected time period to determine the number of unique operating load
states that exist. Every operating state is assigned a probability based on the number of times
that it occurs and this probability is used to determine the cost of losses for each state.

39.6.6 Advanced Options Page

Candidate Buses
• All terminals in feeder; If this option is selected, every terminal in the feeder is considered as a
possible candidate for a ’new’ capacitor.
• Percentage of terminals in feeder; Selecting this option and entering ’x’ percent for the parameter
means the optimisation algorithm will only consider ’x’ percent of the feeder terminals as targets
(candidates) for ’new’ capacitors. The ranking of terminals is according to the Sensitivity Analysis
as described in Section 39.6.2.
Max. Number of Iterations

This parameter determines the maximum number of iterations of the optimisation algorithm before it
automatically stops. As a maximum of one capacitor is placed per iteration, this can effectively limit the
total number of capacitors that can be placed by the optimisation routine.

Max. Execution Time

This parameter specifies the maximum time the optimisation routine can run before it is automatically
interrupted.

Penalty Factors for Voltage Deviation

• Factor for Deviation from 1 p.u (weight); This parameter is used to determine the total ’fictitious
cost’ for terminals deviating from 1 p.u. The cost is applied to each phase of the terminal. For
example, if a three phase terminal voltage is measured at 0.95 p.u for each phase and the ’fictitious
cost rate’ is $10,000/% then the total cost of this deviation is $150,000 (5% * $10,000/% * 3).

Note: If no penalty costs are to be applied within the admissible band, this factor should be set to zero.
If this value is greater than zero, the program will add costs to all terminals with voltage different
than 1.0 p.u.

• Additional Factor outside range [vmin, vmax] (weight2); This parameter can be used to apply an
additional weighting factor to the first deviation factor when the terminal voltage falls outside the
voltage limits defined on the ’Basic Options’ page. The factor is cumulative, so using the previous
example and a additional factor of 20,000/% with a vmin of 0.975, the fictitious cost becomes
$300,000 (5% * $10,000/% + 2.5% * $20,000/%) * 3.

DIgSILENT PowerFactory 2018, User Manual 885

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

Note: The values for the two voltage penalties ’weight’ and ’weight2’ should be carefully chosen be-
cause the target optimisation function is a sum of three objective functions (losses, capacitor cost
and voltage deviation cost). If the voltage weights are too high, the algorithm might not consider
the other two objectives. Likewise, if they are very low, the algorithm may not consider voltage
violations at all.

Print report after optimisation

The automatic printing of the optimisation results can be disabled by unchecking this option.

Draw the installed capacitors

This option draw the installed capacitors in the Single Line Diagram when checked.

39.6.7 Results

The last three OCP tool-bar buttons give access to the optimisation results.

Show Nodes with New Capacitors

When pressing the Show Nodes with New Capacitors icon ( ), after a successful optimisation is
complete, a list appears of all terminals where capacitors are proposed for installation.

Show New Capacitors

Pressing the Show New Capacitors icon ( ) shows a list of proposed new capacitors.

Output Calculation Analysis

This Output Calculation Analysis icon ( ) generates a report with the results of the sensitivity analysis
and the final optimisation procedure.

Viewing results on the Voltage Profile Plot

Following a successful optimisation, the ’new’ capacitors can be visualised on the voltage profile plot of
the feeder. To enable this, navigate to the voltage profile plot display after the optimisation and click the
rebuild button. An example of such a plot showing the placed capacitors is shown in Figure 39.6.3.

886 DIgSILENT PowerFactory 2018, User Manual

 39.7. PHASE BALANCE OPTIMISATION

Figure 39.6.3: Voltage profile plot showing the new capacitors after an Optimal Capacitor Optimisation.

Removing Capacitors Placed by the Optimal Capacitor Placement Routine

The capacitors placed by the OCP command can be removed at any time after the analysis has been
completed by using the Remove previous solution icon ( ). This button is like an ’Undo’ for the ’Optimal
Capacitor Placement’.

39.7 Phase Balance Optimisation

Distribution networks are generally designed to support asymmetric loads and feed in. This fact leads
to asymmetric load flows, which can be calculated with PowerFactory using the unbalanced Load Flow
Calculation. The asymmetric load flows result in higher loadings and losses in single phases and
transformer windings and are therefore not welcome. In networks with a high number of asymmetric
loads and/or elements with less than 3 phases, the Phase Balance Optimisation offers a possibility
to distribute the connected phases of asymmetric elements in a way to minimise the power unbalance
in the network. The feature works on radially operated feeders using one of two possible algorithms to
satisfy the chosen objective function.

The following sections describe the objective functions of the Phase Balance Optimisation, the imple-
mented algorithms and the possibilities regarding the solution and its output.

39.7.1 Objective functions

The optimisation algorithm may be executed for two different objective functions. The minimised quantity
is in both cases the power unbalance 𝑠. It is defined for branch elements as follows:

DIgSILENT PowerFactory 2018, User Manual 887

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

Let 𝑁 be the number of phases, and let

𝑁
1 ∑︁
𝑆ˆ = 𝑆𝑖
𝑁 𝑖=1

be the average complex power (at one end) of a branch element, where 𝑆𝑖 , 𝑖 = 1, . . . ,𝑁 are the complex
powers on phases 1, . . . ,𝑁 . Let
𝑁
1 ∑︁
𝑆¯ = |𝑆𝑖 |
𝑁 𝑖=1

be the average of the absolute values of the powers on the different phases. Then the power unbalance
factor 𝑠𝑏 for the branch element 𝑏 is defined as
¯ max {|𝑆𝑖 − 𝑆|}.
𝑠𝑏 := (1/𝑆) ˆ
𝑖=1,...,𝑁

The user can choose between the following two objective functions for the optimisation:
• Minimise average power unbalance: This function takes into account the power unbalance of
all 𝑀 branch elements, which are part of the analysed feeder. The average power unbalance is
defined as:
𝑀
1 ∑︁
𝑠¯ = 𝑠𝑏 .
𝑀
𝑏=1

• Minimise power unbalance at feeding point: This function permutes the connection of the
feeders elements to get a minimum power unbalance at the feeding point (branch element, where
feeder ’starts’), regardless the unbalance of the rest of the feeder elements.

39.7.2 Methods

To achieve the minimisation of power unbalance, three different algorithms are available to choose from:
• Large loads and generators first

• Simulated annealing
• Genetic Algorithm
All three methods have their advantages. The Large loads and generators first method is easier to
understand and to configure, and leads in most cases to very good solutions. The Simulated annealing
and genetic algorithm methods are more theoretical regarding the configuration, but due to the random
approach can find solutions for networks in which the power balance is difficult to achieve.

Large loads and generators first:

This algorithm iterates over all loads and generators in order of their apparent power, starting with the
largest load or generator. For each load or generator, it will permute the connections of the load or
generator or their supplying branch elements. After calculating the objective function for all possible
connections, it will choose the best connection for this load or generator.

Settings:

The behaviour of the algorithm can be controlled with the following two settings:
The setting ’Disconnect loads and generators at beginning’ will disconnect all loads and generators
before the algorithm starts to iterate, and in each iteration, the actual load or generator will be connected
to the grid in the best way for this iteration step. If this setting is not chosen, the actual load or generator
will be just reconnected.
At each iteration, the algorithm will evaluate several modifications. The threshold in the frame ’Accep-
tance of change’ determines the minimal improvement required for such a modification. If a modification
does not lead to an improvement larger than this threshold, it is not applied to the solution.

888 DIgSILENT PowerFactory 2018, User Manual

 39.8. OPTIMISATION ALGORITHMS

Simulated annealing

This algorithm as well as the input parameters are described in detail in section 39.8.

39.7.3 Considered elements

The set of elements, whose connections are permuted during the optimisation can be parameterised
by the following settings:
• Allow phase permutation: defines, which types of elements are considered by the algorithm. At
least one box has to be checked.
• Elements with fixed phases: a selection of various elements may be chosen, which are excluded
from the optimisation.

39.7.4 Representation of solution

The optimised connections of the affected elements can be applied to the network by choosing one of
the two possible options:

The preferable option is to use ’Create new Variation’, where all connection changes will be stored into
a new variation. They can be undone by deactivating the newly created variation. This option is also
advantageous if the impact of the optimisation has to be analysed or if different settings have to be
compared between each other.

As alternative, the changes may be set directly in the network without creating a new variation. Select
this option only if surely intended!

After the execution of the optimisation the result boxes show the unbalance factors of power, current
(for branch elements) and voltage (nodes). By calculating another load flow, the result boxes are reset
to the normally shown variables.

39.7.5 Output

The Phase Balance Optimisation tool displays by default some information in the output window. The
internal and effective objective function value before and after the optimisation are printed (the internal
objective value may differ from the effective objective value due to an approximation made to achieve the
high performance). The number of modified elements, differentiated for the element types, are listed,
too. Additional information may be displayed by checking the following settings in the Output page of
the ComBalance-dialog:

• Output changed elements: lists after execution of the optimisation all elements, which were
changed including the affected phases.
• Report objective value after each iteration: displays the internal objective value for every
iteration in the output window.

39.8 Optimisation Algorithms

Genetic Algorithm and Simulated Annealing are well suited to solve the following optimisation problems.
• Objective function is not differentiable
• Only “discrete” states are allowed

DIgSILENT PowerFactory 2018, User Manual 889

CHAPTER 39. DISTRIBUTION NETWORK TOOLS

• Possibility of having lots of local minima

• large amount of solutions within the state space

39.8.1 Genetic Algorithm

To illustrate how the Genetic Algorithm improves the optimisation process, the procedure for solving the
Phase Balance Optimisation using this algorithm is shown below.

The starting point of this algorithm is the state space. This set contains all possible phase connection
permutations within the network as a sequence, the so-called genotypes. These permutations are
coded numerically within the genotypes.

Figure 39.8.1 illustrates one genotype used for Phase Balance Optimisation. In this example, the
numbers are the coded possibilities for phase permutations at the different cubicles within the network,
whereas each vertical line represents one cubicle.

Figure 39.8.1: Sample genotype for Phase Balance Optimisation simulation

Out of the set of possible genotypes, a user-defined number of states is drawn and defined to be the
population.

Starting from this population, the single genotypes are mutated and crossed over.

Mutation means the replacement of single code numbers, in this example phase permutations. A
possible mutation is shown in figure 39.8.2.

Figure 39.8.2: Sample genotype mutation for Phase Balance Optimisation simulation

Crossover stands for the exchange of sequences of code numbers (genes) between two genotypes.
These two steps are executed in every iteration.

In each iteration, the objective function is calculated for each genotype within the population. In this
example, the objective function would be the power unbalance. The genotype representing the minimum
of the objective function is stored globally and gives the resulting optimum at the end of the optimisation.

Settings: the algorithm stops if the ’Maximum number of iterations’ is reached or the objective value is
less than the defined value. If the latter is set to zero, the algorithm always stops after the maximum
number of iterations.
The population settings define how many genotypes will be considered.

890 DIgSILENT PowerFactory 2018, User Manual

 39.8. OPTIMISATION ALGORITHMS

The mutation rate defines the portion out of the population, for which a mutation will be executed.
The number of mutation points defines the number of mutations, executed within one genotype.
The number of crossover points defines the lengths of the sequence to be replaced during cross over
process.

39.8.2 Simulated Annealing

Simulated annealing is a stochastic optimisation method, which reconnects the grid randomly, and
during a cool down of the ’system’ will reach a good solution. During the execution of the algorithm, a
so called temperature 𝑇𝑛 is tracked, which reduces the longer the algorithm lasts.

At each iteration 𝑛 of the algorithm, first a new proposal for possible solutions is generated. These
solutions are then applied to the grid and the objective value 𝑣𝑝 for this proposal is calculated.

If the objective 𝑣𝑝 is better than the last objective value 𝑣𝑙 (meaning 𝑣𝑝 < 𝑣𝑙 ), the algorithm accepts the
proposal and will continue with the next iteration step.

If the objective 𝑣𝑝 is worse than the last objective value 𝑣𝑙 , with some probability 𝑝𝑛 ∈ (0, 1) the algorithm
still accepts the proposal. Note, that in this case 𝑣𝑝 > 𝑣𝑙 . The probability 𝑝𝑛 decays as the temperature
𝑇𝑛 decays:
𝑝𝑛 ∝ exp{−(𝑣𝑝 − 𝑣𝑙 )𝐼𝑛 },
where 𝐼𝑛 = 1/𝑇𝑛 is the inverse temperature.

Some stopping criteria can be given, to determine when the algorithm should stop iterating.

Settings: the algorithm stops if the ’Maximum number of iterations’ is reached or the objective value is
less than the defined value. If the latter is set to zero, the algorithm always stops after the maximum
number of iterations.

The two settings in the frame for the inverse temperature 𝐼𝑛 define how fast the variation between two
iterations decreases.

DIgSILENT PowerFactory 2018, User Manual 891