"Let's simplify the art of Structural Engineering"

Why Bolt and not Weld? Why Weld and not Bolt?

July 31, 2020 Posted by Parishith Jayan , , , 10 comments

As a practicing design engineer, during the initial stage of the career, each and every one of us would be subjected to this particular dilemma.  

"Choosing between bolted and welded connections."

 Where should we adopt bolted connection and at what circumstances we should incorporate welded connection? 

Over experience, we would have figured it out. Let me share my own experience with it so that it would be beneficial for upcoming engineers.

Choosing between a bolted connection and a welded connection is not so difficult. 

Design engineers choose their connection type (bolted or welded) based on one important factor. “FEASIBILITY OF FABRICATION AND ERECTION”.

One important rule to follow.

Designing a structure might involve several design constraints and complexities based on the loads that are about to act on the structure. The solution to these difficulties should be “SIMPLE” & “FEASIBLE”.

Having the above-mentioned rule in concern, the choice of bolted or welded connection should be decided based on the feasibility of production, transportation, and erection process.

To have a clear understanding. Let us say we are designing an industrial building whose cross-section looks similar to this. (clear span gable frame)

The connection between the column and the rafter is usually the bolted connection because providing a welded connection, in this case, is not practically possible. Say the height of the column is around 8 m. In site, one cannot lift all the welding equipment to such height and provide proper weld. If you suggest going for welding at the fabrication shop itself, then, in that case, it is very hard to transport the entire arrangement. So, a practically possible and simple solution to this problem is adopting “bolted connection”.

Now, let us go for a building with a roof truss.

As you see the above picture, the rafter is a truss arrangement that constitutes a top chord member, bottom chord member, diagonal, and vertical truss members. The knee connection (column and truss connection) will be bolted connection similar to the previous case. Whereas, the member connections within the truss will be welded connection.

Because individually bolting each of the truss members is a tedious process and it employs erection difficulties. The simplest solution to this case will be welding the truss members (top chord, bottom chord, vertical and diagonal members) in the fabrication shop itself up to transportable length (say 12 m). Then the truss members can be joined together with bolted connections in the site to form the complete truss.

So, the choice of bolted and welded connection is not only a design consideration. While doing so, the practical feasibility of the connection has to be addressed to come up with a simple solution.

Shifting the position of a single rod matter a lot in design...

July 31, 2020 Posted by Parishith Jayan , , , 3 comments

If you are a professional design engineer, then most probably you would have heard about this particular collapse dated a few decades back.

July 17, 1981, is an unforgettable date similar to 9/11. One of the deadliest structural failure recorded in American history which took place in the city of Kansas, Missouri.

Yes, I am talking about the Hyatt Regency hotel walkway collapse.

But, WHY?

Why this particular collapse?

Why it is so important?

If someone wants to understand that simple misunderstanding can cause severe collapse, then they should definitely go through the failure study of the Hyatt Regency walkway collapse.

Let me depict the story first, and we see the structural aspect of it after.

So, it goes like this. The hotel was opened to the public by 01-July-1980. Exactly one year and 17 days later from its inaugural, there was a huge gathering in the evening of 17-July-1981, for a tea dance in the atrium, and that is when the tragic event occurred.

There were three walkways that connect the north and south wing of the hotel. The fourth-floor walkway lies just above the second-floor walkway and the third-floor walkway runs at some offset from these two. (shown in the figure below)

At some point, the fourth-floor walkway collapsed over the second floor and both together fell down taking the lives of 114 people and 216 were severely injured. It is recorded as the deadliest structural collapse in the US until the collapse of World Trade Center which happened after 2 decades.

Why did it happen?

The main reason for the collapse is stated as the mismatch between the original design and the one which has been executed.

Moving to the structural aspects:

The original design of the walkway incorporates a steel hanger that supports the two floors of walkways, fourth floor lying above the second floor.

The floor is supported over the cross beams and these cross beams are inturn were hanged from the ceiling with the help of hanger rods and nuts. One important detail to note is that this hanger rod supports two walkways. As per the original design, in order to support the fourth-floor walkway, the nut has to be screwed from the bottom (i.e. from the second floor) up to the fourth floor. The entire hanger rod between the fourth and second floors needs to be threaded.

This triggered the steel manufacturers to go for revision. Instead of threading the whole rod, they made it as two separate rods. First rod stopping at the fourth-floor walkway and the next rod hanging from the fourth to the second floor. It is said that the designer approved the revision over a telephonic conversation.

This simple change from the steel manufacturer affected the whole design leading to the collapse of walkway.

This change had doubled the load that comes to the fourth-floor cross beam.

Let us focus on the fourth floor since it is where the collapse initiated.

As we see detail 1, it is very clear that the cross beam is supported by the nut. Say, if the floor load is “P”, then the hanger rod on each side would experience a force of “P/2”.

Now, look into detail 2, there is a rod that is supporting the second floor gets attached to the cross beam on the fourth floor. So, the load from the second floor gets transferred to the cross beam in the fourth-floor through the hanger rod. Considering the load on second floor as “P” and since there is a hanger on two ends, each hanger would transfer a force of “P/2″ to the fourth floor. There is already a floor load of “P/2” on one side of the hanger that would be due to the fourth-floor load.

Summing up the entire loads acting on hanger at the fourth floor gives (P/2 + P/2) = “P” (For the original design, it was P/2).

This change in load path was the major issue which is not addressed during the revision..

Original Load path: Floor - Cross beam - Steel hanger rods (both floors have a similar load path)

Revised Load Path: Second floor - Cross beam on second floor - steel hangers - Cross beam on Fourth floor - steel hangers.

This is where the loads get doubled. 

What will happen to the structural components because of this increased load?

The nut which holds the beam in place won't fail easily, similarly the force on hanger rod as per the original design and the modified one is same since it has to carry loads of both the floors in two cases, it won't fail.

The only element which is susceptible to be affected by this change is the cross beam on the fourth floor, upon which double the times of actual design load is about to act.

One more amplifying factor is that the cross beam is a built-up box which is made from two C-sections that are welded face to face throughout its length. The hanger rod passes through the plane of a weld (which is the weakest portion in the build-up member) which resonated the speed of collapse.

Final takeaway - If not properly addressed, shifting the orientation of a single rod in design could result in a catastrophic failure.

Bending Strength of the Laterally Supported Beam

July 30, 2020 Posted by Parishith Jayan , , , 1 comment

Laterally Supported Beam – In general, a beam that does not move nor rotate laterally is termed as “Laterally Supported Beams”. This lateral restraint can be possibly obtained by several means. Few of them are,

·         Compression flange of the element embedded inside the slab
·         Compression flange connected to the slab by means of shear connector
·         Lateral braces provided in the beam

Determining the bending strength of the laterally supported beam is quite straight forward, before moving into it, let's get ourselves clear about the failure modes for the laterally supported beams, from which we can understand which factor governs the bending strength of the beam.

A typical laterally supported beam could fail by anyone of the following failure modes:

·         Shear failure of the section

·         Flexural failure of the section (bending failure)

·         Web crippling/web buckling (local failures)

·         Deflection

Of the above-mentioned failure modes, Flexural failure of the section is a bending failure, occurs when the applied load produces an internal bending moment, which is pretty much higher than the bending strength or moment capacity of the beam.

If we look into the above statement, there are two important terms,

1.    The internal bending moment generated because of the applied load.

2.    Bending or Moment capacity of the beam.

The first one can be determined by simple mechanics. For example, if we have a laterally supported pinned beam of length L and which supports a uniform load of W kN/m. Then the bending moment produced will be WL*L/8.

Our main interest is the Moment capacity of the section. It depends upon the cross-section of the beam as well as material grade.

In general, the Moment capacity of the section equals the product of Section Modulus and Yield strength of the material.

M = Z * fy * (factor of safety as per the specified code)

Thus, we have our required bending capacity or strength of the section.

What is so intriguing about Mono-pole Traffic Signboard? - Bi-Axial Bending...

July 29, 2020 Posted by Parishith Jayan , , , 9 comments

When it comes to structural behaviour, the technical terms might sound a bit complicated, but actual behaviour is NOT.

I am moving forward with the assumption that you know the basic concept of bending.

“Biaxial bending of a member is nothing but the bending of a member in both the axis simultaneously because of the applied load.”

In general, it is same as a beam or column bending which takes place because of the applied transverse load. The only difference is that it occurs in both the principal axis, hence the name “Biaxial Bending”.

This can be easily seen in columns. Especially in the monopole structures. (Monopole is a structure which is composed of only one column. For example, traffic signboard, transmission towers, street lights, etc.)

For our easy understanding, let us discuss this particular behaviour with an example.


The above image shows a traffic signboard with a monopole column. Moving little deeper into the behaviour of the column, lets first fix the principal axis.

For the general convention of the axis, let's take the fog line on the road (continuous white line on the side of the road) as global Z-axis, and the one perpendicular to it is the global X-axis.

Now, let us see how bending would arise in both these directions.


Firstly, bending about global Z-axis would occur because of the cantilever beam which is rigidly connected to the column end. This cantilever beam holds the signboard, that means, those dead load of the signboard will act on the cantilever beam which in turn gets transferred to the column top joint as “vertical load” and “equivalent moment” along Z-axis. This moment induces a bending of the column about that axis (Z-axis).


Moving to the X-axis, as you can visually see, the surface area of the signboard which is likely to subject to the wind load. Assume that, the heavy wind is blowing, the widespread signboard is standing against that. What probably would happen?

The wind that hits the signboard in the direction parallel to Z-axis would result as “horizontal load” and “moment” along X-axis. This is how the bending about X-axis arises in the column.

As it is evident that these two load cases will simultaneously occur in the signboard structure.

What would happen next?

It will bend in both the directions (biaxial bending).