(function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start': new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0], j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src= 'https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f); })(window,document,'script','dataLayer','GTM-5M6SH59');
7 minutes read
5 December 2016

Connection rigidity impact on static design

7 minutes read

This is the third post about the rigidity of the connections in steel structures. Since we already have discussed what a hinge actually is, and how to calculate the connection rigidity, it is time to discuss how this all fits into the static design. It is also important to discuss the impact of connection rigidity on the outcomes from static design which is today’s topic.

Introduction

Knowing how to calculate connection rigidity is one thing, but today I want to show you how that rigidity impacts the outcomes from static calculations. I know for a fact, that connection rigidity is frown upon by many engineers (let’s face it, it is additional work, and there is no additional time!). I hope that this post will show how important this topic may be in your design.

To begin the discussion we should consider one simple idea:

If we are talking about rotational rigidity of the connection we expect the outcomes from the static design to be somewhere between those for a hinge and those for a rigid connection

If we are talking about translational rigidity of the connection the outcomes will be somewhere between those for a rigid connection and those with no connection at all!

If we would think about it for a second we can get to a point where we realize this is not only the issue for moment and force distribution in the model (which is what I want to discuss today). This will also have an impact on element or structure stability (and I will discuss it in the future)!

Connection rigidity-impact on moment distribution in the model

The simplest example is, of course, a simply supported beam. For this example, we are using an HEB 300 cross-section between 2 concrete walls. Since we have assumed in our design that the connections will be rigid (and we have designed the beam that way) we want to make an end-plate connection with sufficient rigidity. Let’s go through possible options:

  • Solution A: Typical end plate solution – we can wonder whether screws should be on the inside or on the outside of the cross section. Deformations are shown for concentrated force 10kN at the end. Cantilever length: 1m.
  • Solution B: Extended end plate solutions – something that will most likely fit better here, we will use both for comparison. Load and cantilever length is, of course, the same in both cases.
  • With the above solutions, tensile force on the support is carried through the rigid circles representing screws, while the compressive contact is allowed on the entire surface of the end plate. With both geometries, different plate thickness was used (10, 15 and 20mm). The above outcomes are for 20mm thickness.

Presented models are very simple (only one plate, cross-section outline with rigid elements, and rigid cantilever). There is always a temptation to add several elements for them to look “nicer”. Here, for instance, we could add a few centimeters of the cantilever length as shells, simply for a better looks at the outcomes. I would advise against it however as it significantly impacts the results when deformations are so small (0.054mm as maximal deformation!).

As you can see above a certain length of the beam was modeled with surface elements (100mm length to be precise). Note that maximal deformation increased greatly – this is due to deformation of the added cross-section (that replaced rigid element). Since in our final model that cross-section will be there, we shouldn’t take its deformation into account as we discussed previously.

The rigidity of the connections was calculated in the same way I have presented in the previous post – this time, however, there will be no slippage. Outcomes for different geometries and plate thickness are given below:

As I wrote above we will expect the outcomes to be somewhere between the hinge connections and the rigid ones, so let’s try it out. For the first solution A all rigidities were added into the following model: 5 different 10m long HEB 300 beams were implemented with the different stiffness of the supports. Starting from the hinged connection (on the bottom left) and going through all the rigidities up to a completely rigid connection (on the top right).

As can be seen, a model with the plate t=10mm has the bending moment in the middle of the span 50% higher than that for the completely rigid solution. Even plate t=20mm has a difference of 13%. Sometimes such difference is huge, sometimes it is negligible – I will address such considerations in the post summarizing the connection rigidity segment. For now, let’s keep in mind that the differences can be quite high.

Of course, a similar comparison can be done for solution B:

Here differences are smaller, but still quite high in certain cases (up to 142% for plate t=10mm).

Of course, at this point, we could discuss that it is easy to stiffen solution B with additional plates on web extension, etc. This is all true, however, I do not wish to optimize a simple connection, but to show you how big influence the rigidity of the connection may have on the outcomes from static design. I completely understand that such influence is not always important and that in several cases it won’t produce as dire effects – I will discuss such topics in the summary post of the connection rigidity segment.

However, since we are all here, let’s take a look at one more example. The same beam but from a different perspective. Let us assume that we want to have a hinged connection to that concrete wall. Of course, I would pick an end-plate solution as below – for future reference let’s call it Solution C:

With such geometry actual rigidity of the connection is:

Of course, rigidity is significantly smaller than in the case of solutions A and B (that was the point after all), but let’s see how this influence the static design outcomes in a similar way than before:

You can easily see that regardless of the plate thickness bending moment in the middle of the span drastically decreases (even by 50%). This alone is not a problem of course, but:

in static design it is always like this: if conditions for some element improved, surly some other element have worse conditions that before! Nothing is free in static design!

In this example, the “better conditions” element is clearly the beam (less moment in the span), but the connection has it “worse”! There is a bending moment on the support that we haven’t foreseen with assumed hinged connection… and experience shows that if an unexpected moment “appears” in the connection we are in for some trouble!

Summary

As I have shown here rigidity of connections is not only a problem when you try to transfer bending moment… it is also a problem when you try NOT to transfer it. This fact is greatly overlooked in my opinion (well the entire topic of connection rigidity is, I think). We have to be aware that “nothing is free” in static design – if something decreased somewhere, there is a cost: something else must have increased somewhere else!

There is no way to convince a steel connection to act as a hinge or a rigid one, and we so often assume one or the other out of habit. In many cases, this is completely understood and accepted, but in others, this may lead to serious problems.

Thank you for reading, I’m very happy that you decided to take some time to reach this point. In fact, I have a surprise for you : ) If you are interested you can subscribe below to get my free FEA essentials course.

Author: Łukasz Skotny Ph.D.

I have over 10 years of practical FEA experience (I'm running my own Engineering Consultancy), and I've been an academic teacher for a decade. Here, I gladly share my engineering knowledge through courses, and on the blog!

Read more

10 Lessons I’ve Learned in 10 Years!

Get Essential FEA Course for Free!

Comments (8)

gabriel - 2021-08-08 21:27:28

Hi Lukasz, thanks for writing this post. very informative.
Correct me if I'm wrong. The configuration of your connection is what we would usually call a pinned connection. The point you're making is that as we increase the thickness of the plate, the connection attracts more bending moment and it's no longer a pinned connection.
Regarding the connection with only 2 bolts, how would that take bending moment if you cannot create a push/pull force on the bolts (as they are horizontally aligned)?

Reply
Łukasz Skotny Ph.D. - 2021-08-10 13:44:19

Thank you for your kind words :)

As to your question, calling a connection "pinned" doesn't mean that it will stop transferring moment. Pinned actually means that elements can freely rotate against each other - and that is not always obvious. Of course, often you only need minimal rotation angle (since the structure doesn't want to rotate like crazy, but just a little bit) but still, this is a consideration.

As for the 2 bolts connection, the compression can be transferred "for free" where the flanges at the bottom "touch" (all you need is touching to transfer compression, it doesn't require anything else (apart from contact capacity to transfer applied compressive stress of course, but this is rarely an issue in such connections). And the tension can be transferred by the bolts. So you have a pair of forces (compression = flange contact, and tension = bolts). This is how it can transfer bending :)

Hope this helps!
Ł

Reply
paolo - 2020-01-04 13:13:18

great post. tnks Lucas
You've opened my eyes, and unfortunately scared a little.
How many hinge connections have I made wrong in my (fortunally short) career

Reply
Łukasz Skotny Ph.D. - 2020-01-05 18:13:11

Hey Paolo!

I'm glad that you found this useful. I guess we all learn stuff this way, and when you learn more you are afraid about the old projects. I guess this never stops, so one of my strong rules is "never look at your old projects"!

All the best
Ł

Reply
Boguslaw - 2017-02-15 13:14:22

Quite interesting. But there is sthng wrong with this: "model with the plate t=10mm have the bending moment in the middle of the span 150% higher than that for the completely rigid solution" - it should be 50% (30 vs 20).
Rgds
BP

Reply
Łukasz Skotny Ph.D. - 2017-02-15 16:05:31

Hi,

Ah, of course - I was to hasty :) I tried to write that it was 150% of the moment in the rigid beam and somehow "higher" got involved :)

Will fix it right away :)

Reply
Mathias - 2016-12-05 23:02:33

Very interesting set of topics. Good to have some knowledge about it, as it is very often neglected!

Reply
Łukasz Skotny Ph.D. - 2016-12-06 09:09:44

Hi Mathias,

I'm very glad you like it :)

Cheers
Łukasz

Reply

Sign up to newsletter

and get Free FEA Course!