Compute reference active power flows for sensitivity analysis #206

Djazouli · 2021-01-26T11:36:42Z

Signed-off-by: Gael Macherel gael.macherel@artelys.com

Please check if the PR fulfills these requirements (please use '[x]' to check the checkboxes, or submit the PR and then click the checkboxes)

The commit message follows our guidelines
Tests for the changes have been added (for bug fixes / features)
[?] Docs have been added / updated (for bug fixes / features)

Does this PR already have an issue describing the problem ? If so, link to this issue using '#XXX' and skip the rest
#205

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

…s within the main component Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

src/main/java/com/powsybl/openloadflow/network/LfNetwork.java

src/main/java/com/powsybl/openloadflow/network/util/ActivePowerDistribution.java

src/main/java/com/powsybl/openloadflow/sensi/DcSensitivityAnalysis.java

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

src/main/java/com/powsybl/openloadflow/sensi/DcSensitivityAnalysis.java

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Signed-off-by: Anne Tilloy <anne.tilloy@rte-france.com>

…alization when having successive contingencies Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Signed-off-by: Anne Tilloy <anne.tilloy@rte-france.com>

geofjamg · 2021-02-02T08:19:57Z

Because in DC approximation, the power factor does not exist for example. And we have not implemented the slack distribution on conform load (and we will not in the short term) in sensitivity analysis.

But this could be solved by a pre-check on sensi DC side and keep the utility class generic

src/main/java/com/powsybl/openloadflow/network/util/ActivePowerDistribution.java

src/main/java/com/powsybl/openloadflow/network/util/GenerationActionPowerDistributionStep.java

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

src/main/java/com/powsybl/openloadflow/network/util/GenerationActionPowerDistributionStep.java

src/main/java/com/powsybl/openloadflow/sa/OpenSecurityAnalysis.java

annetill · 2021-02-02T16:56:37Z

src/main/java/com/powsybl/openloadflow/sensi/DcSensitivityAnalysis.java

-        DcLoadFlowParameters dcLoadFlowParameters = new DcLoadFlowParameters(lfParametersExt.getSlackBusSelector(), matrixFactory,
-                true, lfParametersExt.isDcUseTransformerRatio(), lfParameters.isDistributedSlack(),  lfParameters.getBalanceType());
-        DcLoadFlowResult dcLoadFlowResult = new DcLoadFlowEngine(lfNetworks, dcLoadFlowParameters).run();
+    private GraphDecrementalConnectivity<LfBus> createConnectivity(LfNetwork lfNetwork) {


@floriand-e2r Can you move this function to a network util class ?

I've put it in LfNetwork directly (similarly to createZeroImpedanceSubGraph())

annetill · 2021-02-02T16:57:43Z

src/main/java/com/powsybl/openloadflow/sensi/DcSensitivityAnalysis.java

+        Map<LfBus, BusState> busStates = new HashMap<>();
+        if (lfParameters.isDistributedSlack()) {
+            double mismatch;
+            if (connectivity != null) {


@floriand-e2r we have to think about making this "common".

src/main/java/com/powsybl/openloadflow/sensi/DcSensitivityAnalysis.java

annetill · 2021-02-02T17:00:33Z

src/main/java/com/powsybl/openloadflow/sensi/DcSensitivityAnalysis.java

-        participatingElements = step.getParticipatingElements(lfNetwork);
-        ParticipatingElement.normalizeParticipationFactors(participatingElements, "bus");
-        return participatingElements;
+        return getParticipatingElements(lfNetwork, loadFlowParameters, element -> true);


I am not sure to get the aim of this function...

It allows you to get the normalized participating elements

Indeed I get it, it is to filter. Now, you have on bus a boolean on bus isDisabled. Do you think that it is possible to use that attribute to deal with connectivity? Then it will update the participating elements.

It is possible to use the isDisabled attribute, but it less efficient.
To use the isDisabled parameter, I think we would need to:

Save all the buses states

Go through all the buses and disconnect the one that needs to be disconnected

Go through all the buses to get the non-disconnected participating elements

Restore all the buses states

With the current filtering in the DC getParticipatingElements, we are just doing:

Go through all the buses to get participating elements and get the one in the right connected component

I was thinking about save the state of the bus that initial contains a participating element. Then, you disabled those that are not in the main connected component of this connectivity. Then, you can use the classical getParticipatingElements. Then you restore the buses that initially contain participating element. It is just to ensure coherence in the code, but it is not a bit deal.

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Signed-off-by: Anne Tilloy <anne.tilloy@rte-france.com>

Signed-off-by: Florian Dupuy <florian.dupuy@rte-france.com>

flo-dup

I first thought that you needed a change in GraphDecrementalConnectivity to simplify your connectivity code, but in fact you can use getSmallComponents() instead of getComponentNumber() almost everywhere, if not everywhere. There is two places where I'm not so sure about that:

areVariableAndFunctionDisconnected(): do you really need to know if variable and function are in two distinct "small" connected components?
getElementsToReconnect(): I don't think we can avoid using getComponentNumber(), do you agree on that?

Why am I trying to make those changes? Because in implementations of decremental connectivity the getComponentNumber needs some additional time to compute the map giving the number from the vertex.

flo-dup · 2021-02-03T12:01:28Z

src/main/java/com/powsybl/openloadflow/sensi/DcSensitivityAnalysis.java

+            // set buses injections and transformers to 0 outside the main connected component
+            int mainComponentNumber = connectivity.getComponentNumber(network.getSlackBus());
+            List<Integer> columnsToSetToZero = network.getBuses().stream()
+                    .filter(lfBus -> connectivity.getComponentNumber(lfBus) != mainComponentNumber)


To get the buses which are not in the main component you should use directly connectivity.getSmallComponents()

Unfortunately, I think getSmallComponents() returns the smallest components in size, but it has nothing to do with the slack bus (the slack could be in a small component). What I want here are the buses that are not in the same component than the slack bus.

Indeed, but couldn't we make the assumption that the slack bus is in the biggest component in size? I thought you could, but well I might be mistaken - calling @annetill and @geofjamg to the rescue!

flo-dup · 2021-02-03T12:03:46Z

src/main/java/com/powsybl/openloadflow/sensi/DcSensitivityAnalysis.java

                lfFactors.forEach(factor -> factor.setPredefinedResult(null));
-                cutConnectivity(lfNetwork, connectivity.getConnectivity(), breakingConnectivityCandidates);
+                cutConnectivity(lfNetwork, connectivity, breakingConnectivityCandidates);
+                int mainComponent = connectivity.getComponentNumber(lfNetwork.getSlackBus());


Always equal to 0

See previous comment, I am not sure about this

flo-dup · 2021-02-03T12:13:23Z

src/main/java/com/powsybl/openloadflow/sensi/DcSensitivityAnalysis.java

+            // check if the factor function and variable are in different connected components
+            if (factor.areVariableAndFunctionDisconnected(connectivity)) {
+                factor.setPredefinedResult(0d);
+            } else if (!factor.isConnectedToComponent(mainComponent, connectivity)) {


You always call isConnectedToComponent with mainComponent, so in fact you could rename it to isConnectedToMaintComponent and use getSmallComponents() instead of getComponentNumber()

See previous conversation

Signed-off-by: Anne Tilloy <anne.tilloy@rte-france.com>

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

…or get branch Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

…not need to set alpha 1 to 0 Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

sonarqubecloud · 2021-02-08T09:40:54Z

Kudos, SonarCloud Quality Gate passed!

0 Bugs
0 Vulnerabilities
0 Security Hotspots
3 Code Smells

98.6% Coverage
0.0% Duplication

Djazouli linked an issue Jan 26, 2021 that may be closed by this pull request

Recompute function references when doing sensitivity analysis with contingencies #205

Closed

jonenst force-pushed the master branch from 9542f6d to a371a95 Compare January 26, 2021 15:51

compute function references from within sensitivity analysis

2d924e3

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Djazouli force-pushed the recompute_function_ref branch from 438e157 to 2d924e3 Compare January 26, 2021 16:11

Djazouli added 8 commits January 26, 2021 17:29

do not use streams to get element from Collection

0bdc3d9

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

assert impossible nullpointerexception to pass sonar

841e8d5

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

cache the result of base sensitivity for each contingency

0774a43

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

remove code smells

a001db4

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

refactor the factor state recomputation if distributer or glsk

bce98c0

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

remove dead code now that we are only computing sensitivity for value…

de686c5

…s within the main component Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

do some refactoring to make the code more understandable

e44039a

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Merge branch 'master' into recompute_function_ref

d86c431

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Djazouli requested review from annetill and geofjamg January 28, 2021 13:06

annetill requested changes Jan 28, 2021

View reviewed changes

Djazouli added 3 commits January 28, 2021 16:55

Factorize setAlphaFlow and setAlphaSensi

0d5b0ae

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

add memorization of previous value within step constructors

4c7436b

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

some renaming

58d8312

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

annetill requested changes Jan 29, 2021

View reviewed changes

Djazouli and others added 4 commits January 29, 2021 16:03

remove branch tripping from this PR

f9f313d

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Add getLfBus to ParticipatingElement

3ae3176

Signed-off-by: Anne Tilloy <anne.tilloy@rte-france.com>

remove caching from participating elements, as it leads to wrong norm…

3ebde47

…alization when having successive contingencies Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

fix small bug with multiples contingencies on the same element

c114a33

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Djazouli force-pushed the recompute_function_ref branch from fd67e6b to c114a33 Compare February 1, 2021 18:39

annetill added 2 commits February 2, 2021 07:56

Move method to ActivePowerDistribution.

ced6252

Signed-off-by: Anne Tilloy <anne.tilloy@rte-france.com>

Finish merge.

62676c8

Signed-off-by: Anne Tilloy <anne.tilloy@rte-france.com>

geofjamg reviewed Feb 2, 2021

View reviewed changes

src/main/java/com/powsybl/openloadflow/network/util/ActivePowerDistribution.java Outdated Show resolved Hide resolved

geofjamg reviewed Feb 2, 2021

View reviewed changes

src/main/java/com/powsybl/openloadflow/network/util/GenerationActionPowerDistributionStep.java Outdated Show resolved Hide resolved

Merge branch 'master' into recompute_function_ref

46bff86

Djazouli added 3 commits February 2, 2021 15:39

quick fix for function references with transformers

0e8691f

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

use bus states and add a test on reset

ba8668d

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

remove unused code

36b59aa

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

annetill reviewed Feb 2, 2021

View reviewed changes

src/main/java/com/powsybl/openloadflow/network/util/GenerationActionPowerDistributionStep.java Outdated Show resolved Hide resolved

annetill reviewed Feb 2, 2021

View reviewed changes

src/main/java/com/powsybl/openloadflow/sa/OpenSecurityAnalysis.java Show resolved Hide resolved

annetill reviewed Feb 2, 2021

View reviewed changes

src/main/java/com/powsybl/openloadflow/sensi/DcSensitivityAnalysis.java Outdated Show resolved Hide resolved

annetill reviewed Feb 2, 2021

View reviewed changes

Djazouli and others added 5 commits February 2, 2021 18:18

remove useless copy and default constructor

dbc8abf

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

manage balance type check in dc

c9ff234

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Small renaming.

8888188

Signed-off-by: Anne Tilloy <anne.tilloy@rte-france.com>

Create utility method in LfNetwork to create the DecrementalConnectivity

5826701

Signed-off-by: Florian Dupuy <florian.dupuy@rte-france.com>

Move BusStates create/restore methods in BusState

0610e6f

Signed-off-by: Florian Dupuy <florian.dupuy@rte-france.com>

flo-dup requested changes Feb 3, 2021

View reviewed changes

annetill and others added 4 commits February 3, 2021 15:15

Revert some renaming.

eb2c029

Signed-off-by: Anne Tilloy <anne.tilloy@rte-france.com>

Merge branch 'master' into recompute_function_ref

dfb5636

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

remove hacky patch for a1 in function references

e87f676

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

Merge branch 'master' into recompute_function_ref

e7a7bb2

annetill changed the title ~~[WIP] Compute function references from within sensitivity analysis~~ Compute reference active power flows for sensitivity analysis Feb 4, 2021

Djazouli added 3 commits February 5, 2021 12:23

Merge branch 'master' into recompute_function_ref

e48b2ca

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

merge fix

2b53ee1

Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

remove alpha1 from branch outside main component, and optimize a fact…

fae1027

…or get branch Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

flo-dup force-pushed the recompute_function_ref branch from 547ce9a to 2b53ee1 Compare February 5, 2021 16:38

add test combining transformers and contingencies to make sure we do …

646ba36

…not need to set alpha 1 to 0 Signed-off-by: Gael Macherel <gael.macherel@artelys.com>

annetill merged commit 0fae72a into master Feb 8, 2021

annetill deleted the recompute_function_ref branch February 8, 2021 09:45

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Compute reference active power flows for sensitivity analysis #206

Compute reference active power flows for sensitivity analysis #206

Djazouli commented Jan 26, 2021

geofjamg commented Feb 2, 2021

annetill Feb 2, 2021

flo-dup Feb 3, 2021

annetill Feb 2, 2021

annetill Feb 2, 2021

Djazouli Feb 2, 2021

annetill Feb 2, 2021 •

edited

Loading

Djazouli Feb 2, 2021

annetill Feb 2, 2021

flo-dup left a comment •

edited

Loading

flo-dup Feb 3, 2021

Djazouli Feb 3, 2021

flo-dup Feb 3, 2021

flo-dup Feb 3, 2021

Djazouli Feb 3, 2021

flo-dup Feb 3, 2021

Djazouli Feb 3, 2021

sonarqubecloud bot commented Feb 8, 2021

Compute reference active power flows for sensitivity analysis #206

Compute reference active power flows for sensitivity analysis #206

Conversation

Djazouli commented Jan 26, 2021

geofjamg commented Feb 2, 2021

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

annetill Feb 2, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

flo-dup left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

sonarqubecloud bot commented Feb 8, 2021

annetill Feb 2, 2021 •

edited

Loading

flo-dup left a comment •

edited

Loading