Implement an FIR filter to calculate radiation forces #1071

salhus · 2023-06-07T14:24:55Z

This follows from #585
This PR implements an FIR filter to calculate the radiation forces. The user needs to set a flag simu.FIR = 1 to enable this implementation.
This PR uses automatic code generation to FIR filter blocks for each mode (both diagonal and off-diagonal terms) and works with body-2-body interactions enabled.
I have modified a line of code in the RM3 example that has the simu.FIR flag for easier review.

The PR still needs documentation.

remove sv_FIR flag

akeeste

Nice work @salhus!

There's a couple remaining items to fine tune this capability:

Update Line 3 in the FIR Filter Matrix mask to include body-to-body interactions. You might need to use body.total instead of length(body) because the highest level body block already defines body as body(1), body(2), etc
Add a short line of documentation in the comment at WECSim_Lib_Body_Elements/Rigid Body/Hydrodynamic Body/Wave Radiation Forces Calculation/SS CI and Constant-Damping-CoeVariant Subsystem to state what flag will turn on the FIR option, e.g. "If another wave type and simu.FIR == 1, use the FIR filter calculation"
Add advanced feature documentation
Add a WEC-Sim Application that compares the various radiation damping calculations. This will ensure this functionality is tested and would be a nice way to highlight the differences to users.

akeeste · 2023-06-30T16:36:19Z

examples/RM3/userDefinedFunctions.m

@@ -1,25 +1,25 @@
-%Example of user input MATLAB file for post processing


Revert to original version upon completion

akeeste · 2023-06-30T16:37:47Z

examples/RM3/wecSimInputFile.m

@@ -1,20 +1,21 @@
 %% Simulation Data
 simu = simulationClass();               % Initialize Simulation Class


Revert to original version upon completion

salhus · 2023-08-09T23:48:45Z

Hi Adam @akeeste ,
I added the rate transition blocks, and the sampling time to simu.cicDt.
I set the number of row blocks to 6 and wrapped a try statement when adding additional instances, so working on the block is easier. The row adding cannot be linked to body, since each body block will have their own index, and it will mess up the rest of the logic downstream. At this point I am deciding against enabling GBM.
Also, I would stick with length(body) instead of blody.total as body.total only counts hydrodynamic bodies.

In terms of development, I think this is done. The tests are passing.
I am only left with documentation.

cheers,
Sal

akeeste · 2023-08-11T16:18:47Z

Hi @salhus

Thanks for those updates. I added a few more changes that I described above into PR salhus/WEC-Sim#11. Please review and merge that PR into your branch. I updated the rate transition blocks to specify use of simu.cicDt or else they will inherit the same time step as the rest of the simulation. I also updated the counting of DOFs and bodies in the library block so that the FIR setting can now be used with:

nonhydro bodies
B2B
GBM

Note that non-hydro bodies can't be used with B2B if they're included in the h5 file, but that is a larger issue already present in WEC-Sim. I also don't think this will work with B2B and GBM, but that is a niche case that can't be tested easily and I'm unsure if GBM+B2B is currently supported or not.

Updates for FIR filter

salhus · 2023-08-15T17:57:51Z

@akeeste
Hi Adam,
Thanks for those additions ! I merged your PR. Thanks for the changes to the class. And it seems to work for body.dof, since body.total would have not worked as intended.

So just to confirm with you the work is done, and I am moving on the documentation now.

Cheers,
Sal

nathanmtom · 2023-09-21T23:56:17Z

@salhus Thanks for developing this feature as it provides an alternative to the primary CIC option and highlights how we can leverage automatic generation to auto build.

After completing my initial review, I believe you have responded to @akeeste requests expect for documentation and developing an applications to compare the three different approaches to the convolution integral calculation. We also should be adding a test case to ensure this feature is not broken in the future. I do need to run the WEC-Sim Applications Test with this branch to see if there are any further bugs to fix.

In terms of next steps the priority is documentation. Secondary is the applications and test case which can be added as a project board item and addressed in the next fiscal year. I'll be reviewing the N Waves PR next and if I have time I will try to assist with the applications and test case development.

salhus · 2023-09-22T00:01:06Z

@nathanmtom Thanks for reviewing the PR. I think I should be able to wrap-up the documentation by Monday. I have a PR on the applications repository for N-Wave PR and will add another PR that corresponds to this FIR PR tonight.

nathanmtom · 2023-09-27T05:12:20Z

docs/user/advanced_features.rst

+	+------------------------------------------------+-------------------------------------------+
+	|   *Radiation Force Calculation Approach*       | *Time (s)*  	                             |
+	+------------------------------------------------+-------------------------------------------+
+	|   No Convolution                               | 3.699 s    	                             |


@salhus Minor comment: can we call No Convolution "Constant Coefficients" as this is what the block is called in the Wave Radiation Forces Calculation Block.

Second when I run the simulation I get slightly different times, so I wanted to request we report normalized times. So No Convolution would be 1 and all other times would be referenced to that. I think there will be confusion if these numbers are not reproduced by users unfortunately there is a lot of factors impacting the final values.

Updated the docs.

done

nathanmtom · 2023-09-27T05:14:33Z

docs/user/advanced_features.rst

 .. _user-advanced-features-time-step:

+
+FIR Filters


@salhus Minor comment can we spell out FIR and say "Finite Impulse Response (FIR)". Also on line 218 below write out fully the first time and then use FIR for the rest.

Makes sense.

done.

nathanmtom · 2023-09-27T05:18:09Z

Thanks for updating the documentation @salhus. I have some minor requests on the documentation that I would appreciate you incorporating. Also if you can respond to @akeeste comments about reverting two files back to the original we should be ready for merging by end of this week. I'll respond with feedback on the applications case in that PR.

editorial changes

salhus · 2023-09-27T21:01:31Z

@nathanmtom ,
Thanks Nathan.
I updated the docs and restored the RM3 example to its original state.

nathanmtom · 2023-09-27T23:47:08Z

Thanks @salhus for addressing my comments for this PR. I accept your latest changes and all tests are passing except for the windows-latest which has been identified in issue #1128.

salhus · 2023-09-27T23:50:54Z

@nathanmtom
Thanks Nathan for patiently going through it all!

salhus added 8 commits June 6, 2023 16:03

row-working

344b4a3

Update WECSim_Lib_Body_Elements.slx

86b2444

Update WECSim_Lib_Body_Elements.slx

75febaa

update body block and start variant

4e96b05

not_working_but_have_variants_setup

1e21afd

almost_there

a74b034

b2b0works

8af479c

b2b works

7e3deb8

salhus requested a review from akeeste June 7, 2023 14:25

salhus marked this pull request as draft June 7, 2023 14:25

akeeste self-assigned this Jun 7, 2023

salhus requested review from dav-og, dforbush2, jleonqu, jtgrasb, kmruehl and nathanmtom June 7, 2023 14:26

akeeste added the Feature new feature request label Jun 7, 2023

salhus added 3 commits June 7, 2023 08:34

Update stopWecSim.m

fdfe58a

remove sv_FIR flag

Delete block_params.m

3567545

Update RM3.slx

bc9e6eb

kmruehl removed request for dav-og, dforbush2, jleonqu, jtgrasb, kmruehl and nathanmtom June 14, 2023 14:21

akeeste mentioned this pull request Jun 14, 2023

Review implementation of convolution integral in radiation problem #585

Closed

akeeste requested changes Jun 30, 2023

View reviewed changes

Update WECSim_Lib_Body_Elements.slx

ac9bb32

salhus added 4 commits August 9, 2023 17:06

in-line comments inside the block

50a514c

replace library

534e1f2

minor bug fixes

7959a7e

Saving file in 2020b format

2ed40b4

akeeste added 2 commits August 11, 2023 10:43

fix body.b2bDOF definition

41bfb78

update library dof and body counting, lib to R2020b

d132158

Merge pull request #11 from akeeste/Fr-FIR-filter

4f26763

Updates for FIR filter

kmruehl requested a review from nathanmtom September 20, 2023 14:43

kmruehl assigned nathanmtom and unassigned akeeste Sep 20, 2023

salhus marked this pull request as ready for review September 21, 2023 04:51

salhus mentioned this pull request Sep 22, 2023

Example demonstrating the FIR Filter method to Calculate the Radiation Force WEC-Sim/WEC-Sim_Applications#43

Merged

Update advanced_features.rst

96da582

nathanmtom reviewed Sep 27, 2023

View reviewed changes

salhus added 2 commits September 27, 2023 14:57

Update advanced_features.rst

86f20f3

editorial changes

restore RM3 example to original state

b714a77

salhus added 2 commits September 27, 2023 15:45

Update advanced_features.rst

64ad13a

Update advanced_features.rst

e102211

nathanmtom merged commit 9cb6724 into WEC-Sim:dev Sep 27, 2023

salhus deleted the Fr-FIR-filter branch October 4, 2023 18:36

		@@ -1,25 +1,25 @@
		%Example of user input MATLAB file for post processing

		@@ -1,20 +1,21 @@
		%% Simulation Data
		simu = simulationClass(); % Initialize Simulation Class

Implement an FIR filter to calculate radiation forces #1071

Implement an FIR filter to calculate radiation forces #1071

Uh oh!

Conversation

salhus commented Jun 7, 2023

Uh oh!

akeeste left a comment • edited by nathanmtom Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

akeeste Jun 30, 2023

Choose a reason for hiding this comment

Uh oh!

akeeste Jun 30, 2023

Choose a reason for hiding this comment

Uh oh!

salhus commented Aug 9, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

akeeste commented Aug 11, 2023

Uh oh!

salhus commented Aug 15, 2023

Uh oh!

nathanmtom commented Sep 21, 2023

Uh oh!

salhus commented Sep 22, 2023

Uh oh!

nathanmtom Sep 27, 2023

Choose a reason for hiding this comment

Uh oh!

salhus Sep 27, 2023

Choose a reason for hiding this comment

Uh oh!

nathanmtom Sep 27, 2023

Choose a reason for hiding this comment

Uh oh!

salhus Sep 27, 2023

Choose a reason for hiding this comment

Uh oh!

nathanmtom commented Sep 27, 2023

Uh oh!

salhus commented Sep 27, 2023

Uh oh!

nathanmtom commented Sep 27, 2023

Uh oh!

salhus commented Sep 27, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

akeeste left a comment •

edited by nathanmtom

Loading

salhus commented Aug 9, 2023 •

edited

Loading