Add sensitivities on intensity #231

Djazouli · 2021-02-26T16:51:09Z

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)

Djazouli · 2021-02-26T16:52:17Z

src/main/java/com/powsybl/openloadflow/ac/equations/AcEquationSystem.java

+
+        if (i2 != null) {
+            Equation i =  equationSystem.createEquation(bus2.getNum(), EquationType.BUS_I).addTerm(i2);
+            i.setActive(false); // equation is set to inactive so it does not appear in jacobian


This is certainly not the proper way to do this

Djazouli · 2021-02-26T16:54:17Z

src/test/java/com/powsybl/openloadflow/sensi/ac/AcSensitivityAnalysisTest.java

+        assertEquals(5, result.getSensitivityValues().size());
+        // todo: Remove the *100 from assert. Currently the
+        // todo: loadflow from OLF and Hades provides different results
+        assertEquals(766.4654d, getFunctionReference(result, "l23"), LoadFlowAssert.DELTA_I * 100);


Those *100 are needed because of a difference I have between Hades loadflow and Openloadflow loadflow. But maybe this is juste my hades2 config that is wrong ?

Djazouli · 2021-02-26T16:55:08Z

src/test/java/com/powsybl/openloadflow/util/LoadFlowAssert.java

@@ -23,6 +23,7 @@
    public static final double DELTA_ANGLE = 1E-6d;
    public static final double DELTA_V = 1E-3d;
    public static final double DELTA_POWER = 1E-3d;
+    public static final double DELTA_I = 1E-2d;


10e-3 would be better, but as I said before, I have some trouble mapping the Hades2 loadflow with the OpenLoadFlow loadflow to check my results, even on a very small network (4 buses, 1 transfo)

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

annetill · 2021-03-04T10:52:43Z

...a/com/powsybl/openloadflow/ac/equations/ClosedBranchSide1IntensityMagnitudeEquationTerm.java

+        double p1 = r1 * v1 * (g1 * r1 * v1 + y * r1 * v1 * sinKsi - y * R2 * v2 * sinTheta);
+        double q1 = r1 * v1 * (-b1 * r1 * v1 + y * r1 * v1 * cosKsi - y * R2 * v2 * cosTheta);
+
+        i1 = Math.hypot(p1, q1) / (Math.sqrt(3.) * v1 / 1000);


@Hadrien-Godard @Djazouli I think we don't have to compute the intensity from p and q. I have done that before as it was a post-processing. Can we compute the intensity through the Ohm law ?

Yes we can, but we will end up with the same formula since Ohm law is used to obtain above expression of p and q from v and phi.
The difficulty here is that Ohm law expresses a difference of complex tensions in term of a complex intensity, or the intensity that is measured (here i1) is in fact the magnitude ofthe complex intensity. Therefore the Math.hypot is needed to express this magnitude.

Yes but it would be simpler to calculate re(i) and img(i) and then mod(i) that calculating p and q and then dividing by voltage for nothing

Sure it would be more efficient.

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

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

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

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

…l-open-loadflow into intensity_from_ac

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

annetill · 2021-03-11T09:16:08Z

src/main/java/com/powsybl/openloadflow/ac/equations/AbstractBranchAcFlowEquationTerm.java

@@ -19,6 +19,8 @@
 */
 abstract class AbstractBranchAcFlowEquationTerm extends AbstractNamedEquationTerm {

+    public static final double NORMALIZATION_FACTOR = 1000d / Math.sqrt(3d);


Maybe you could add a reference to CURRENT in the name.

annetill · 2021-03-11T09:31:33Z

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

        }
    }

-    private List<SensitivityValue> getPostContingencySensitivityValues(List<LfSensitivityFactor<ClosedBranchSide1ActiveFlowEquationTerm>> lfFactors,
+    protected void setReferenceActivePowerFlows(List<LfSensitivityFactor<? extends AbstractClosedBranchAcFlowEquationTerm>> factors) {


You have to change the method name, and go back to one of your first choices setReferenceFunctions maybe?

annetill

Some small changes but okay!

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

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

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

annetill · 2021-03-15T15:52:18Z

It seems that we have performance issues with that PR. Can we add a parameter to compute intensity only for sensitivity analysis involving factors that need too ? For load flow computation, we will have to compute intensity only after a NR convergence.

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

annetill · 2021-03-16T15:39:11Z

src/main/java/com/powsybl/openloadflow/util/currentBranchesManager/AllBranchesManager.java

+/**
+ * @author Gael Macherel <gael.macherel at artelys.com>
+ */
+public class AllBranchesManager implements CurrentBranchesManager {


Maybe AllCurrentBranchesManager?

annetill · 2021-03-16T15:39:46Z

src/main/java/com/powsybl/openloadflow/util/currentBranchesManager/SpecificBranchesManager.java

+/**
+ * @author Gael Macherel <gael.macherel at artelys.com>
+ */
+public class SpecificBranchesManager implements CurrentBranchesManager {


Maybe SpecificCurrentBranchesManager?

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

geofjamg · 2021-03-17T07:09:17Z

src/main/java/com/powsybl/openloadflow/ac/equations/AcEquationSystemCreationParameters.java

@@ -17,14 +20,21 @@

    private final boolean forceA1Var;

+    private final BranchesCurrentManager branchesCurrentManager;


BranchesCurrentManager is probably too much for just a branch list filtering, why not just a list of branch id and empty (or null) means all branches?

geofjamg · 2021-03-17T07:10:12Z

src/main/java/com/powsybl/openloadflow/equations/EquationSystem.java

@@ -66,7 +66,8 @@ private void reIndex() {

            Set<Variable> variablesToFind = new HashSet<>();
            for (Equation equation : equations.values()) {
-                if (equation.isActive()) {
+                if (equation.isActive() && EquationUpdateType.DEFAULT.equals(equation.getUpdateType())) {


It is safe to compare enum with == in Java

geofjamg · 2021-03-17T07:10:18Z

src/main/java/com/powsybl/openloadflow/equations/EquationSystem.java

        for (Equation equation : equations.values()) {
-            equation.update(x);
+            if (updateType.equals(equation.getUpdateType())) {


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

sonarqubecloud · 2021-03-17T10:09:24Z

Kudos, SonarCloud Quality Gate passed!

0 Bugs
0 Vulnerabilities
0 Security Hotspots
4 Code Smells

91.2% Coverage
0.0% Duplication

Djazouli commented Feb 26, 2021

View reviewed changes

Base automatically changed from sensi_ac to master February 28, 2021 13:36

add intensity sensitivity

fd4b83f

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

Djazouli force-pushed the intensity_from_ac branch from a54456a to fd4b83f Compare March 1, 2021 14:23

annetill reviewed Mar 4, 2021

View reviewed changes

Djazouli and others added 15 commits March 5, 2021 17:55

Add coherence between DELTA_I and DELTA_POWER

d86716f

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

Merge branch 'master' into intensity_from_ac

891063f

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

merge fix

a0d651b

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

Merge branch 'master' into intensity_from_ac

52dc43d

merge fix

9f996a3

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

refactor closed branch intensity

26b0be0

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

Merge branch 'master' into intensity_from_ac

e00e0da

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

compute intensity on open branches without using p and q

e73a6f1

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

use i1 eval instead of old formula with p and q

3d02aa5

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

Add unit tests for LfSwitch.

304fc08

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

Fix unit test.

e4b1266

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

check that we cannot compute intensity sensi in DC

909e14d

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

First import of unit test for the new intensity magnitude equations.

cdc1de1

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

Merge branch 'intensity_from_ac' of https://github.com/powsybl/powsyb…

b70c001

…l-open-loadflow into intensity_from_ac

add test on open branch for intensity magnitude

835e20f

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

annetill reviewed Mar 11, 2021

View reviewed changes

Djazouli and others added 3 commits March 11, 2021 13:31

some renaming

c8a0f5a

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

Small fix.

6170b2d

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

Replace intensity with current

aff7569

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

annetill approved these changes Mar 12, 2021

View reviewed changes

Merge branch 'master' into intensity_from_ac

56f4852

annetill changed the title ~~add intensity sensitivity~~ Add sensitivit Mar 15, 2021

annetill changed the title ~~Add sensitivit~~ Add sensitivities on intensity Mar 15, 2021

Djazouli added 2 commits March 16, 2021 13:41

update intensity only after the convergence of newton raphson

c6ca639

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

create intensity equations only for the factors that requires intensity

bd43ab1

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

annetill reviewed Mar 16, 2021

View reviewed changes

rename BranchesCurrentManager

611800a

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

geofjamg requested changes Mar 17, 2021

View reviewed changes

changes requested

50077aa

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

geofjamg approved these changes Mar 17, 2021

View reviewed changes

Merge branch 'master' into intensity_from_ac

ba56bc6

annetill merged commit d8dad84 into master Mar 17, 2021

annetill deleted the intensity_from_ac branch March 17, 2021 10:13

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add sensitivities on intensity #231

Add sensitivities on intensity #231

Djazouli commented Feb 26, 2021

Djazouli Feb 26, 2021

Djazouli Feb 26, 2021

Djazouli Feb 26, 2021

annetill Mar 4, 2021

Hadrien-Godard Mar 4, 2021

geofjamg Mar 4, 2021

Hadrien-Godard Mar 4, 2021

annetill Mar 11, 2021

annetill Mar 11, 2021

annetill left a comment

annetill commented Mar 15, 2021

annetill Mar 16, 2021

annetill Mar 16, 2021

geofjamg Mar 17, 2021

geofjamg Mar 17, 2021

geofjamg Mar 17, 2021

sonarqubecloud bot commented Mar 17, 2021

		@@ -17,14 +20,21 @@

		private final boolean forceA1Var;

		private final BranchesCurrentManager branchesCurrentManager;

Add sensitivities on intensity #231

Add sensitivities on intensity #231

Conversation

Djazouli commented Feb 26, 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

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 left a comment

Choose a reason for hiding this comment

annetill commented Mar 15, 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

sonarqubecloud bot commented Mar 17, 2021