Welcome Guest! To enable all features please Login or Register.

Notification

Icon
Error

4 Pages123>»
Options
Go to last post Go to first unread
Offline RFreund  
#1 Posted : 13 April 2014 18:18:03(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
I'm sure there are many others like me who are attracted to SMath because of how it creates easy to follow calculations for engineers. Not to mention its free, open source (edit - not open source), has a great community and is easy to learn the basics. However what I would like to do is try to understand the best way to create interconnected calculations. As engineers (I will speak regarding structural engineering but I'm sure it can apply to other fields as well) we start with determining loads then choosing preliminary sizes, analyzing the structure to determine forces then designing the members. What I would like to do is to be able to create a continuous calculation where my loads are linked with analysis and design. The problem I run into when creating these calculations is size, speed and formating. I am curious to hear how others handle this and below I will describe what I would believe to be ideal.

I think essentially what I have in mind is similar to creating a program using code. However my code is the calculations and the end program is the structure design.
So here is an outline:

    [1]Snippets

First pretty much all calculations would be add to the calulcation page via snippets. Meaning that you start with your Loading section and you choose snippet -> dead load, snow load, wind load, etc. Then for design you choose snippet -> steel beam, concrete beam or whatever you need. All will have the ability to reference loads defined above.

    [2]Loads

Loads will need to use many different data tables. These could be kept as spreadsheets and imported to the document where they will then become Matrices. See here for further discussion.

    [3]Design / Analysis

For most hand calculations the design and analysis go hand in hand. The problem I see is that it takes considerable space to create a continuous beam analysis calculation then you must also include the design of the beam. Now you could use Insert -> Area regions to hide parts but then formating sometimes becomes a pain and there are various bugs (such as inserting snippets with area regions and deleting area regions). What would be advantageous is to be able to send input (i.e. beam length and loading) to a program which runs as a beam analysis program and then sends back the desired output (i.e. reactions, moment, shear, etc). However this program would be called for each beam instance without overwriting the output of the calculations above.



I'm curious to know how others are doing this or any tips that you care to provide.

Thanks in advance!!!!

**Edit - SMath is not Opensource, not sure what I was thinking there (thanks Ioan for the catch) and Thanks Mr. Andrey IVASHOV for the program.

Edited by user 21 September 2014 20:20:32(UTC)  | Reason: Title Spelling errors

Offline uncblue  
#2 Posted : 14 April 2014 20:58:48(UTC)
uncblue


Rank: Member

Groups: Registered
Joined: 28/08/2013(UTC)
Posts: 18
United States

Was thanked: 2 time(s) in 2 post(s)
Hey fellow structural engineers. When I first starting using SMath, I was trying to implement similar ideas with read/write to excel back and forth. Due to speed issues, I gave up and now create inputs for loads,reactions, etc. This was something I created a while back before I gave up and feel free to use any helpful info from it. I turned off autocalculate and removed company header. Although the software isn't able to be as automatic as I would like, I'm still very thankful to the developers and contributors for their hard work.
Edit: File was last saved with SMath Studio 0.97.5189.

Edited by user 14 April 2014 21:06:27(UTC)  | Reason: Not specified

File Attachment(s):
test.zip (233kb) downloaded 245 time(s).
Offline mkraska  
#3 Posted : 15 April 2014 00:03:47(UTC)
mkraska


Rank: Advanced Member

Groups: Registered
Joined: 15/04/2012(UTC)
Posts: 1,495
Germany

Was thanked: 796 time(s) in 509 post(s)
You might consider setting optimization numeric for data read operations. In your case this cuts by half the execution time of the whole sheet.

File Attachment(s):
DataExchange.zip (382kb) downloaded 97 time(s).
mkraska attached the following image(s):
performance.PNG
Martin Kraska

Pre-configured portable distribution of SMath Studio: https://en.smath.info/wi...th%20with%20Plugins.ashx
thanks 1 user thanked mkraska for this useful post.
on 15/04/2014(UTC)
Offline RFreund  
#4 Posted : 15 April 2014 17:02:25(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
Thanks guys. I appreciate the input thus far.

Ioan - your first post makes some very good points. You're right that SMath should not replace programming languages but I realize that I think a program that could emulate the "rules" of a programing language in a simplified program would be excellent for engineering calculation purposes. Really when you put together calculations for a project you are creating a small program. Now I'm not really talking about new low/mid/high rise structures which we use FEM or other full structure packages. I'm talking about small projects which probably make up 90% of structural engineering work. I think most engineers try to use spreadsheets in this manner and try to link them as best as possible. I suppose you could use Excel and VBA to create something similar to what I'm getting at but you lose so much transparency (not to mention beauty) of the calculation.

There is a software TEDDS by CSCWorld (see here) that is sort of similar to the scheme in which I discuss but there are MANY limitations. They separate the calculation page into sections which calculate only when asked.

Anyway it's good to hear that I'm on the right track. I'll share what I come up with.

Thanks again!!
thanks 1 user thanked RFreund for this useful post.
on 15/04/2014(UTC)
Offline uncblue  
#5 Posted : 15 April 2014 20:24:53(UTC)
uncblue


Rank: Member

Groups: Registered
Joined: 28/08/2013(UTC)
Posts: 18
United States

Was thanked: 2 time(s) in 2 post(s)
RFreund,

I've only seen an output from another company for CSC TEDDS. It was a concrete beam analysis and anchor rod analysis. I don't see why this couldn't be done in SMath. I'm someone who has created spreadsheets, Mathcad, and SMath (superior to Mathcad in my opinion) sheets without paying an arm and a leg to a company like CSC just to make it look "pretty".

Forgive me, but I'm not sure I understand what you are trying to accomplish as far as design checks. So when you select the material, do you want the same SMath sheet to perform the necessary design checks based on whatever material you choose? Or one SMath sheet for steel checks, another for concrete, etc, but each sheet will be similar in approach as far as organization of loads and import/export data?
I'm willing to help if I can understand the goal. I realize what you are trying to accomplish isn't the easiest thing to explain through text on a forum.

ioan92 and mkraska,
Thanks for the tips. The tips reduced the load/calculate times tremendously and now may be useful to me.

Edit:
So I just came across the snippets feature. I never knew this even existed sadly.

Edited by user 15 April 2014 20:55:30(UTC)  | Reason: Not specified

thanks 1 user thanked uncblue for this useful post.
on 15/04/2014(UTC)
Offline RFreund  
#6 Posted : 16 April 2014 05:32:44(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
UNC - There are subtle differences in TEDDS but it is very similar to the snippets concept (except they have many predefined calculations) but there is no good way to store data or to even use matrices.

As far as my goal -> I'm not really sure. I suppose ultimately the goal is to link all the calcs together as a first step. So the overall picture -

  • Define your loads - say snow, wind, EQ, etc. And all these are different snippets that you pull into the calculation.

  • Next say you have a steel roof beam. So you define some tributary width and reference your loads you previously defined. Now you need a beam analysis calculation to find your shear and moments. This could be a large program if you have multiple point loads and varying loads and uniform loads.

  • Now I want to design the beam that I analyzed above. So I insert a steel design snippet. Usually design snippets are of smaller size and run time length, but they may require large data tables

  • Now I need to analyze another beam. Well this means pulling in the snippet for me analysis again which is a large calculation. Too bad I can't just have this be an "off page" calculation (I've also previously called this an 'engine page' where I send my input and get a result

  • again I design the beam

  • Now I design the steel column by inserting my column design snippet and loads are based on the beam or from the original design loads

  • Now design the foundation by inserting spread foundation design snippet which again references the original loads or column reaction


So in the end, most of this is doable. The problems are that a beam analysis program can be large/slow and also how to effectively use what I call "data tables" without slowing the calculation down.

Then if we are really asking for more we get into formating: See Table of Contents and Program Format

Also the point of this is to discover any other tips that I've been missing out on i.e.
Quote:

So I just came across the snippets feature. I never knew this even existed sadly.

Offline RFreund  
#7 Posted : 14 July 2014 04:41:43(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
I think I am starting to see a possible way to accomplish what I'm looking to do.
I am starting to define more functions, large functions. These are then saved as snippets.

So for example you have a wind load function. The wind pressure is based on the tributary and location of the member. Therefore I have a snipet which is a function for wind load which I load at the beginning of the page. Then for each element I can send the tributary are and location to the function and retrieve the pressure for each member.
I can see this working for the beam analysis program as well.

I'll try to post back when/if I make some headway
Offline RFreund  
#8 Posted : 24 July 2014 21:01:38(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
Ioan had previously volunteered some help for creating a continuous beam program, any interest in giving me some further direction on this?
I'd like to do it as a function where you could imput:
Length
Point load, location
Varying load, start and end location

if possible....

Thanks !
Offline RFreund  
#9 Posted : 25 July 2014 00:42:00(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
Sounds good, enjoy your Holiday!
Offline RFreund  
#10 Posted : 27 August 2014 16:14:43(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
Now that we are all rested up... Time to program a beam analysis function? Biggrin

I also wondered if there would be any benefit to instead create a beam analysis plugin rather than making it a function (maybe it is faster?).

Pretty good link here on creating a frame analysis program:
http://newtonexcelbach.w...ame-analysis-with-excel/
Offline RFreund  
#11 Posted : 28 August 2014 05:45:59(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
Yeah I think that sounds good.
I would like or atleast I think I would like to make the program so that it operates as a function. Reason being is that it will most likely be used multiple times in the calculation page. I'm just not sure how to make the input user friendly or even possible as a function. I think you would need some sort of matrix input.

If the output could show the shear, moment and deflection diagrams that would be good too (maybe only at discrete points or subdivide the beam into "X" number of nodes).
Offline RFreund  
#12 Posted : 29 August 2014 01:03:28(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
Was that suppose to link to something specific?
I did some quick searching but didn't turn up much.
Offline RFreund  
#13 Posted : 29 August 2014 16:53:22(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
Wow. That is impressive and will take some study.

I'd like to try and replicate that (I think) but turn it into a function if possible. As it would be needed many times throughout the calculation page.
Offline RFreund  
#14 Posted : 30 August 2014 06:02:49(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
Well I looked into Ber7's calculations. They make me think maybe I bit off more than I can chew (not sure if that is an expression in France). However there are a few problems. I'm looking for a program that is more flexible. I'm not so concerned with the output and graphics right now. I'd like to be able to input any number of loads and possible change the support conditions. Also I'd like to use units.

I think I need to start with a stiffness (matrix method) approach perhaps with a certain number of nodes along the beam.

Also I see alot of MathCAD files as MCDX as opposed to XMCD, do you know what the differences or why we can only open XMCD?
Offline Davide Carpi  
#15 Posted : 30 August 2014 11:20:21(UTC)
Davide Carpi


Rank: Advanced Member

Groups: Registered, Advanced Member
Joined: 13/01/2012(UTC)
Posts: 2,192
Man
Italy
Location: Italy

Was thanked: 1105 time(s) in 726 post(s)
Originally Posted by: RFreund Go to Quoted Post
Also I see alot of MathCAD files as MCDX as opposed to XMCD, do you know what the differences or why we can only open XMCD?

Googling it, seems that mcdx is the default (and only) way to save mathcad files with mathcad prime. Also, I've opened a couple of these files with a text editor, and is not an "open" xml scheme; seems a mix of xml code and proprietary format, thus I'm not sure that is possible to create a plugin to open it.

Edited by user 30 August 2014 11:24:31(UTC)  | Reason: Not specified

If you like my plugins please consider a donation to SMath Studio; for personal contributions to me: paypal.me/dcprojects
Offline RFreund  
#16 Posted : 30 August 2014 15:18:55(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
Thanks Davide.

Also attached are a couple of good spreadsheets I found for reference.
File Attachment(s):
BEAM ANALYSIS.xls (3,684kb) downloaded 190 time(s).
Steel JOIST.xls (1,950kb) downloaded 136 time(s).
Offline RFreund  
#17 Posted : 01 September 2014 04:15:05(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
Wow very impressive. This is almost exactly as I was describing. Clap
It does have these funny looking units that don't make much sense though. Crazy

Quote:
I have some reserves on your temptation to put many function in the same file. A real mess could happen with local and global variables.

Yes very understandable. My reasoning is that I like to have the entire calculation "linked" together. Meaning that I like to define all the loads at the beginning of the calculation and then reference them through the calculation. So if I change the roof dead load at some later point. All my beam calculations below are 'automatically' updated for the new roof dead load. I do see there is chance to confuse local and global variables. I would like to ask though what is your definition of local and global variables? Is a local variable, a variable local to the function? Could you maybe provide an example of possible conflict between the two?

Quote:
My idea is to build a separated programs that are doing simple operation and to transfer informations using data files.
I used in my works such a solution with a reasonable success and I can assist you with some details.


I am very interested in this method. Could you provide an example if it is convenient?

Quote:
I haven't yet done a postprocess to obtain informations between the two ends of a element but it is enough simple to do.

Simple, ha, sure simple. Keep me posted I suppose.

Thanks again!!!
thanks 1 user thanked RFreund for this useful post.
on 01/09/2014(UTC)
Offline RFreund  
#18 Posted : 14 September 2014 15:51:08(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
OK, so after so time to digest and use I realize that maybe I was asking for the wrong thing and maybe jumped ahead of myself a bit too quickly. I think I would like to start with a simple beam calculator program. There is nothing wrong with Ioan's calculation and is still very impressive. However, there is a bit of effort requied to make sure you have input data correctly and for most beams the extra effort is not necessarily required. So in any case I suppose I should put forth some effort... Here goes..

So my plan here is to develop a simple beam calculator. First I am posting the calculation which is based on Walter D. Pilkey's Modern Formulas for Statics and Dynamics a Stress and Stain Approach pages 11 - 22 and also a spread by Alex Tomavich which I posted previously.

My next steps are to turn the worksheet into a function so it can be used through the same page but I have a few questions first.

1. What is the best way to determine maximum Moment, location of max moment, max deflection, location of max deflection? We know that derivative of set equal to zero will give us the answer but is there a way to perform this in SMath? My guess is most likely not due to the way I have set up the calculation. So my next choice is the set an input variable "number of points to divide beam" which will calculate the shear moment and deflection at these locations. Then I will need to figure out when the shear changes sign, etc.

After this I would like to graphics.
Then add more end conditions.
Then try to use the 3-Moment theory for a continuous beam.
File Attachment(s):
Simple Beam.sm (78kb) downloaded 130 time(s).
Offline mkraska  
#19 Posted : 15 September 2014 12:39:10(UTC)
mkraska


Rank: Advanced Member

Groups: Registered
Joined: 15/04/2012(UTC)
Posts: 1,495
Germany

Was thanked: 796 time(s) in 509 post(s)
I have made a simple FEA program for straight beams. Features and limitations:
- straight beams along the x-axis
- no longitudinal strain
- no shear strain
- elementwise constant bending stiffness
- exact solution for linear distributed load function given by left and right load values
- no hinges (could easily implemented by duplicate nodes and coupling elements)
- concentrated forces and moments at the nodes
- constraints for displacement w and slope phi can be given at any node and may have non-zero values
- postprocessing: support reactions, plots of displacement, slope, bending moment, shear force and applied distributed load, auto-scaling plots
- except plotting everything can be symbolic
- solution is Maxima-independent
- demonstrates standard FEA procedure: Element and node loops for system assembly, system reduction using constraints, solution of constrained system, expanding, postprocessing.
- model description by an element matrix E and a node matrix N

All this is more a concept demonstration than a streamlined application and for sure is buggy. Background on how the finite element was designed can be found here.

Edited by user 15 September 2014 16:18:38(UTC)  | Reason: Not specified

File Attachment(s):
Balkenelement FEM.sm (180kb) downloaded 198 time(s).
mkraska attached the following image(s):
fea-beam.PNG
Martin Kraska

Pre-configured portable distribution of SMath Studio: https://en.smath.info/wi...th%20with%20Plugins.ashx
thanks 2 users thanked mkraska for this useful post.
on 18/09/2014(UTC),  on 19/09/2014(UTC)
Offline RFreund  
#20 Posted : 15 September 2014 16:49:34(UTC)
RFreund


Rank: Advanced Member

Groups: Registered
Joined: 25/09/2013(UTC)
Posts: 285
United States
Location: IL

Was thanked: 18 time(s) in 16 post(s)
Very nice Martin I will take a look.

I hope to some day be half as smart as you guys.

Thanks again!
Users browsing this topic
4 Pages123>»
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.