Add Infrastructure for Geomechanical Analysis #5801

bska · 2024-12-16T11:08:05Z

In particular, this lays the foundation for adding new well connections as a result of fracturing processes.

bska · 2024-12-16T11:08:15Z

jenkins build this please

bska · 2024-12-16T11:47:27Z

jenkins build this please

bska · 2024-12-17T08:58:19Z

jenkins build this please

bska · 2024-12-18T10:59:46Z

jenkins build this please

bska · 2024-12-19T08:39:16Z

jenkins build this please

bska · 2024-12-19T10:17:18Z

jenkins build this please

bska · 2024-12-19T11:03:26Z

jenkins build this please

bska · 2024-12-19T15:08:53Z

jenkins build this please

bska · 2024-12-20T08:28:16Z

jenkins build this please

bska · 2024-12-20T09:29:40Z

jenkins build this please

bska · 2024-12-20T11:50:24Z

jenkins build this please

bska · 2024-12-20T12:06:06Z

jenkins build this please

akva2

A lot of back and forth. I gave up reviewing as I realized this is not structured as expected for a PR against a production branch. Maybe you want to address some of these, maybe not. I'll post them at least.

akva2 · 2024-12-20T12:08:24Z

opm/simulators/linalg/PropertyTree.hpp

+    boost::property_tree::ptree* getBoostParamPtr() const{
+      //just to be able to use full boost interface
+      return tree_.get();
+    }


this commit doesn't seem ready for inclusion in main line. if we need access to the whole api in simulator code the whole approach of encapsulation using this class falls on its rear end and we should consider removing the encapsulation.

I agree that this breaks encapsulation. I need a way to put values back into the tree, and especially into a subtree, and this was the quickest way to go about doing that. In particular, I didn't see a way of combining get_child() with put<>() in a way that would make the change persistent. I am open to suggestions.

akva2@0106bc9

should in principle do it, although it's has some rather nasty not-so-const-correct aspects.

akva2@0106bc9

should in principle do it, although it's has some rather nasty not-so-const-correct aspects.

Okay, thanks. I'll pick this back up after the holidays, then.

akva2 · 2024-12-20T12:11:21Z

opm/simulators/flow/FlowProblem.hpp

+
+
+protected:
+    ActionHandler<Scalar> actionHandler_;


if we are to introduce this member here, don't we need to move the initialization from the Blackoil child class also? and what happens with compositional, which afaik was the reason it was put in the Blackoil class?

akva2 · 2024-12-20T12:12:03Z

opm/simulators/wells/WellInterfaceGeneric.cpp

@@ -676,6 +676,36 @@ void WellInterfaceGeneric<Scalar>::resetWellOperability()
    this->operability_status_.resetOperability();
 }

+template<class Scalar>
+void WellInterfaceGeneric<Scalar>::addPerforations(const std::vector<std::tuple<int,double,double>>& perfs){


loong line. we should have a type alias for that novel sized type. braces.

akva2 · 2024-12-20T12:12:11Z

opm/simulators/wells/WellInterfaceGeneric.cpp

+void WellInterfaceGeneric<Scalar>::addPerforations(const std::vector<std::tuple<int,double,double>>& perfs){
+        int newperf = perfs.size();
+
+        for(auto& perf : perfs){


akva2 · 2024-12-20T12:12:17Z

opm/simulators/wells/WellInterfaceGeneric.cpp

+        for(auto& perf : perfs){
+            const auto [cell, WI, depth] = perf;
+            auto it = std::find(well_cells_.begin(), well_cells_.end(), cell);
+            if(it != well_cells_.end()){


akva2 · 2024-12-20T12:17:48Z

opm/simulators/flow/FlowGenericVanguard.cpp

@@ -424,6 +424,10 @@ void FlowGenericVanguard::registerParameters_()
    Parameters::Register<Parameters::OwnerCellsFirst>
        ("Order cells owned by rank before ghost/overlap cells.");
 #if HAVE_MPI
+    Parameters::Register<Parameters::AddCorners>
+        ("Add corners to partion.");


akva2 · 2024-12-20T12:18:45Z

opm/simulators/flow/FlowProblemBlackoil.hpp

    using FlowProblemType = FlowProblem<TypeTag>;
-
+    private:


we want the blank line.

akva2 · 2024-12-20T12:19:03Z

opm/simulators/flow/FlowProblemBlackoil.hpp

-
+        this->endStepApplyAction();
+    }
+    void endStepApplyAction()    


missing newlines.

akva2 · 2024-12-20T12:20:25Z

opm/simulators/flow/OutputBlackoilModule.hpp

@@ -882,6 +882,42 @@ class OutputBlackOilModule : public GenericOutputBlackoilModule<GetPropType<Type
                        else if (FluidSystem::phaseIsActive(waterPhaseIdx))
                            val.second = getValue(fs.temperature(waterPhaseIdx));
                    }
+                    else if (key.first == "BSTRSSXX")
+                        {


formatting is.. not as wanted.

akva2 · 2024-12-20T12:22:31Z

CMakeLists_files.cmake

@@ -574,9 +574,40 @@ list (APPEND TEST_DATA_FILES
 # originally generated with the command:
 # find opm -name '*.h*' -a ! -name '*-pch.hpp' -printf '\t%p\n' | sort
 list (APPEND PUBLIC_HEADER_FILES
+  flow/flow_blackoil.hpp


these (mostly) contain prototypes for functions that are never installed, ie not in the library. if we need need these in downstreams there is more work to be done, this achieves nothing.

this achieves nothing.

That's mostly true. They're used directly in Main.hpp so we either need the headers, or the include statements in Main.hpp must go away.

yes, i am aware of that but Main.hpp is also not structured to be installed, it's only meant to end up in simulators, which again are not meant to be downstream accessible.

my point is that if we really want to prepare the code for downstreams we have a lot of things that need to be addressed, not just in the code but also in the build system., e.g. how we handle gpu stuff.

i'm not really opposing this change, it makes thing no worse than it already was and helps a little so...

yes, i am aware of that but Main.hpp is also not structured to be installed, it's only meant to end up in simulators, which again are not meant to be downstream accessible.

True. I guess it slipped through in #2532 (8c2ba0b) then.

my point is that if we really want to prepare the code for downstreams we have a lot of things that need to be addressed, not just in the code but also in the build system., e.g. how we handle gpu stuff.

I think we may getting to a point where a careful review of the full OPM build system is in order anyway.

bska · 2024-12-20T14:41:01Z

Resetting this back to draft state pending resolution of some of the concerns raised during review.

…ctures on wells is growing

- added numOverlap to flowgeneric

Not all 'problem's have a geomechanical model, so limit block level stress collection to those that do. While here, also install more headers to enable downstream/out-of-tree users of the main simulator classes. Finally, reduce scope of some objects and be (more) 'const' correct.

bska force-pushed the geomech branch from dcf8ea8 to 982b9cb Compare December 16, 2024 11:46

bska force-pushed the geomech branch 2 times, most recently from 1a50042 to eaae892 Compare December 17, 2024 08:53

bska force-pushed the geomech branch from eaae892 to c60cb18 Compare December 17, 2024 11:12

bska mentioned this pull request Dec 18, 2024

Report Connection Level Fracturing Statistics to I/O Layer #5811

Draft

bska force-pushed the geomech branch 3 times, most recently from f53428c to 6a6c47f Compare December 19, 2024 08:31

bska force-pushed the geomech branch from 6a6c47f to 1524a32 Compare December 19, 2024 10:14

bska force-pushed the geomech branch from 1524a32 to cfc4085 Compare December 19, 2024 11:02

bska force-pushed the geomech branch 2 times, most recently from 4a7f0cf to b34550d Compare December 19, 2024 14:22

bska force-pushed the geomech branch from b34550d to e155c71 Compare December 20, 2024 08:27

bska force-pushed the geomech branch from e155c71 to 8a9f983 Compare December 20, 2024 09:29

bska force-pushed the geomech branch from 8a9f983 to 7c48c28 Compare December 20, 2024 11:49

bska force-pushed the geomech branch from 7c48c28 to 7504499 Compare December 20, 2024 12:05

akva2 reviewed Dec 20, 2024

View reviewed changes

bska marked this pull request as draft December 20, 2024 14:41

hnil and others added 12 commits December 20, 2024 15:47

added instantiation of bool in propertytree

45dc902

Added utitlites for addin new wellIndices and modify old: used if fra…

b977d6e

…ctures on wells is growing

adde hack in propertytree

8ba1238

added more output

15e8b4a

fixes for to get geomech at 2025.04

1055037

Rearranged to make it possible to do easier output of matrices

748e8a2

Added more asserts for not setting wrong interior size

08c3c69

Added possibilities to set addCorners and numOverlap

f5764ab

-made more flexible structure for endStep

4ee3ee8

- added numOverlap to flowgeneric

-add blockoutput of stress

473a16f

Remove Unused Object

1282dce

bska force-pushed the geomech branch from 7504499 to 1282dce Compare December 20, 2024 14:47

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add Infrastructure for Geomechanical Analysis #5801

Add Infrastructure for Geomechanical Analysis #5801

bska commented Dec 16, 2024

bska commented Dec 16, 2024

bska commented Dec 16, 2024

bska commented Dec 17, 2024

bska commented Dec 18, 2024

bska commented Dec 19, 2024

bska commented Dec 19, 2024

bska commented Dec 19, 2024

bska commented Dec 19, 2024

bska commented Dec 20, 2024

bska commented Dec 20, 2024

bska commented Dec 20, 2024

bska commented Dec 20, 2024

akva2 left a comment

akva2 Dec 20, 2024

bska Dec 20, 2024

akva2 Dec 20, 2024

bska Dec 20, 2024

akva2 Dec 20, 2024

akva2 Dec 20, 2024

akva2 Dec 20, 2024

akva2 Dec 20, 2024

akva2 Dec 20, 2024

akva2 Dec 20, 2024

akva2 Dec 20, 2024

akva2 Dec 20, 2024

akva2 Dec 20, 2024

bska Dec 20, 2024

akva2 Dec 20, 2024

bska Dec 20, 2024

bska Dec 20, 2024

bska commented Dec 20, 2024

Add Infrastructure for Geomechanical Analysis #5801

Are you sure you want to change the base?

Add Infrastructure for Geomechanical Analysis #5801

Conversation

bska commented Dec 16, 2024

bska commented Dec 16, 2024

bska commented Dec 16, 2024

bska commented Dec 17, 2024

bska commented Dec 18, 2024

bska commented Dec 19, 2024

bska commented Dec 19, 2024

bska commented Dec 19, 2024

bska commented Dec 19, 2024

bska commented Dec 20, 2024

bska commented Dec 20, 2024

bska commented Dec 20, 2024

bska commented Dec 20, 2024

akva2 left a 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

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

Choose a reason for hiding this comment

bska commented Dec 20, 2024