A problem about SAWS material

[TOKYOTech]

Hello everyone，

I am recently doing some research about wood structure and trying to use the SAWSMaterial to simulate the nail joints between shell wall and wood frame.
When I test the material's hysteresis properties with a simple model, a problem occurred as the following picture shows.
https://farm2.staticflickr.com/1510/249 ... _z.jpg[IMG]
The force between the two parallel unloading and reloading line didn't follow the backbone curve while it should have done. I don't this happens because of the material code itself or my model.
Please help me with that.
And the following is my code to test the material， the picture above is made by the displacement of node2.

wipe
puts "system"
model basic -ndm 3 -ndf 6
puts "node"
node 1 0 0 0
node 2 0 0 0
fix 1 1 1 1 1 1 1
puts "material"
#K0-K5 follow the Fig 2.17 on page 39 of paper02 units N,M,Pa
set F0 669.48
set F1 142
set K1 352.07e3
set K2 28.76e3
set K3 31.85e3
set K4 785e3
set K5 28e3
set DU 15.5e-3
set alpha 0.8
set beta 1.1
#R0-R4 follow the fig of opensees SAWS guide
set S0 $K1
set R1 [expr $K2/$K1]
set R2 [expr -$K3/$K1]
set R3 [expr $K4/$K1]
set R4 [expr $K5/$K1]
uniaxialMaterial SAWS 1 $F0 $F1 $DU $S0 $R1 $R2 $R3 $R4 $alpha $beta
uniaxialMaterial Elastic 2 1.0e10

puts "element"
element twoNodeLink 1 1 2 -mat 1 1 2 2 2 2 -dir 1 2 3 4 5 6 -orient 1 0 0 0 1 0
#element multipleShearSpring 2 1 2 32 -mat 1 -lim 0.0001 -orient 0 0 1 0 1 0

puts "recorder"
recorder Node -file node2.out -time -node 2 -dof 1 2 disp

puts "pattern"
pattern Plain 1 Linear {
load 2 1 0 0 0 0 0
}

puts "analysis"
constraints Plain
numberer Plain
system FullGeneral
test EnergyIncr 1.0e-12 10 2
algorithm ModifiedNewton
analysis Static
integrator DisplacementControl 2 1 0.0002
analyze 25
integrator DisplacementControl 2 1 -0.0002
analyze 50
integrator DisplacementControl 2 1 0.0002
analyze 75
integrator DisplacementControl 2 1 -0.0002
analyze 100
integrator DisplacementControl 2 1 0.0002
analyze 125
integrator DisplacementControl 2 1 -0.0002
analyze 150
integrator DisplacementControl 2 1 0.0002
analyze 175
integrator DisplacementControl 2 1 -0.0002
analyze 200
integrator DisplacementControl 2 1 0.0002
analyze 225
integrator DisplacementControl 2 1 -0.0002
analyze 250
integrator DisplacementControl 2 1 0.0002
analyze 260
integrator DisplacementControl 2 1 -0.0002
analyze 270
integrator DisplacementControl 2 1 0.0002
analyze 280
integrator DisplacementControl 2 1 -0.0002
analyze 280

[YuZhang]

Hi TOKYOTech,

I also had this problem in my SAWS models. What's more, when I ran the dynamic analyses, they gave me different results each time I ran the codes. I have sent an email with my codes to Frank to report this problem. Unfortunately I haven't received any replies yet. I believe there are some inherent problems in the source codes for the SAWS materials. Hoping someone could help us out.

[emjac]

Hi,

Just to be clear, the problem is that after approximately disp = 0.025 the material seems to stay on the R4*S0 stiffness instead of moving to the R2*S0 stiffness?

@fmk : I looked at the code in SAWSMaterial.cpp. In LPATH5 and LPATH7 (when loading is negative or positive on the pinched curve) there is a correction noted in the code that changed DINT4 from DINT4 = (-FU + FI + (R2*S0*DU))/(S0*(R4 - R2)); to DINT4 = (-FU - FI + (R2*S0*DU))/(S0*(R4 - R2));. As far as I can tell, the displacement is checked against DINT4 to see if we should be on pinched part of the curve or on the post-capping stiffness (R2*S0). When I reverse the correction and run the mode given by TOKYOTech above I get the following hysteresis : https://goo.gl/photos/R3dovKNP5eHkbEHu9

So this seems to correct the problem, however I am not experienced enough to test further implications of this change... It would be nice to know why McDonald changed it in the first place.

@YuZhang : I have been usings SAWS for some Non-linear Time History Analysis but I am not experiencing the problem you describe (results are the same from one analysis to the next).

[YuZhang]

@emjac
Thanks for your info. I was doing IDA to collapse at that time, and I used maximum inter-story drift ratios as the indicator. I got the same results in one folder, and different results if just copy the codes and ran them in another folder. This sounds very unusual but it did happen.

[TOKYOTech]

@YuZhang @emjac
Thank you very much for your help. Recently I am too busy to check this web. Actually I find the mistake in the source code of SAWS just after I post this page. But I don't know how to introduce a new material into OpenSees =_= !. So I couldn't test it. Just like emjac said, the DINT 4 in the code is not right in both LPATH 5 and LPATH 7. The DINT 4 means the maximum displacement in this material, if the displacement is larger than DINT4, the material will fail and return a warning with stress equal to 0. I checked the original paper " Cyclic Analysis of Wood Shear Walls" by Bryan Folz and Andre Filiatrault. However, the paper doesn't define how to calculate the maximum displacement. The SAWS code defines the maximum displacement as the following picture shows. Using simple geometric knowledge we could calculate the DINT4 = (-FU + FI + (R2*S0*DU))/(S0*(R4 - R2)) (line 672) and DINT4 = (FU + FI + (R2*S0*DU))/(S0*(R4 - R2)) (line 784). So The code really has something wrong. @fmk
https://farm2.staticflickr.com/1474/263 ... 9bcf_k.jpg[/img]

[TOKYOTech]

@fmk @emjac Could you teach me the way to modify the code of SAWS and utilize it in the OpenSees program? I tried to use the *.dll method to build a MSAWSMaterial after changing the code of SAWSMaterial.h and SAWSMaterial.cpp. I changed all the "SAWSMaterial" to "MSAWSMaterial" in related files and fixed the DINT4 error. However, it returens" WARNING could not create uniaxialMaterial MSAWSMaterial ". I used the latest developer folder and followed the 32 procedures. My os is 64bit so I changed the build platform to x64. The example of ElasticPPcpp worked well. But when it came to MSAWSMaterial, the .dll file was generated but it didn't work with my .tcl file. Now it is killing me... Thank you very much for your time!

[emjac]

Sorry I only build on Linux so I won't be any help on Windows, but a suggestion... if you can build correctly with the unmodified SAWSMaterial then instead of renaming things just comment out that line in SAWSMaterial and build. You won't have access to both but at least you'll have the one you want. Good luck!

[TOKYOTech]

@emjac The problem is I could not build the opensees with the source code, tons of errors just pop out from nowhere. That is why I chose the dll method. But it failed as well. Anyway thank you very much.

[AndongZhan]

Hello,i am a beginner of opensees,and my research is about timber structual.Recently,i want to simulate glulam frame-shear wall but have some difficulty.Unfortunately,nobody around me use this software.So,can i get your e-mail and consult some question.Thanks a lot!

[TOKYOTech]

AndongZhan wrote:
> Hello,i am a beginner of opensees,and my research is about timber
> structual.Recently,i want to simulate glulam frame-shear wall but have some
> difficulty.Unfortunately,nobody around me use this software.So,can i get
> your e-mail and consult some question.Thanks a lot!
wudi198911@gmail.com