Dear Reader ,

It is another Friday and the end of the work week. I am here writing a new newsletter reflecting on all things computational modelling as well as some sundry matters that interest me.

Thanks to those who continue to read regularly and for anyone who is new here, welcome. In this edition, I will be reflecting on the following:

1. Technical Reflections: A trick to reduce memory load on simulations
2. Behind the Scenes at CM Videos: A slight change in my video making style
3. Quote of the week: Insight and location by Austin Kleon

Technical Reflections

A trick to reduce memory load on simulations

Let us consider a Representative Volume Element (RVE) of a particulate composite subjected to tensile loading. It is important at the end that you extract the stress-strain data from the simulation. This is shown in the figure below.

If you have tried to run ABAQUS simulations in which you want to extract the stress and strain data, there are two possible routes to use:

• Element-by-element consideration: You can ask the simulation to record in history variables all the stress variables (S11, S22, S12) and strain variables (E11, E22, E12) for every element in the model. This will then be used to apply volume averaging to generate the global stress-strain measures. This will require the user to ask ABAQUS to store these six variables for every time step for all the elements in the model. For a system with say 3000 elements (finely meshed model), you are then having to record 3000 * 6 = 18,000 variables (for all time steps). This excessive number of variables. As shown in the figure above, you will be focusing on the meshed section above
• Reference nodal set consideration: Alternatively, you can decide to use a reference node, for example (as shown in figure above), set at XRP or YRP (i.e. x-axis or y-axis Reference Point). If the tensile load is attached to either the XRP or YRP, then you can simply ask that ABAQUS should output reference force, RF1 and displacement, U1 (for XRP), and RF2 and U2 (for YRP), that means you are looking at possibly 4 variables. These variables will be what will be required to be stored for every time step. So, in this case only 4 variables are recorded in memory as against the previous case of 18,000 variables.

I will usually go with the Reference nodal set option above when I am setting up a model. This is very good for memory of the simulation and makes even post-analysis to be easier.

Challenges to the Reference node method:

Although it is good to computational memory, it however throws up a lot of other issues that the user must address namely:

1. The need for kinematic constraint: Applying load on XRP will only cause displacement of the node XRP which will cause simulation instability and no meaningful change will be seen in the body. What you then must do to correct the problem is to use a kinematical constraint equation to link the XRP behaviour with the front edge of the system. This way, you then force the deformation of the front edge to track the deformation of the XRP node.
2. The need for calculating irregular areas: To get stress measure from the simulation, you will need to divide the reaction force, RF1 with the surface area of the edge/surface. This means that the user will have to determine the surface area as part of the post-processing steps in the simulation. This is made difficult if the surfaces/edges are not continuous. Hence, you will have to find a way of determining the actual surface areas.

I showed these in my recent video about Triaxial loading of RVES which you can watch here if you have not seen it before:

​

Behind the Scenes at CM Videos

A slight change in my video making style

If some of you have have been paying attention to my recent videos, you will notice that I have not been showing my face a lot. This is not a major change.

At the start of the channel, I was determined to make sure my contents have the visual element of my face, voice, illustrations and of course the ABAQUS tutorials. I know this is not always the same with channels in the ABAQUS tutorial niche. Most people show contents without their face.

At first, I struggled with it but once I published my first video where I showed my face, and it had positive engagement and reactions from my audience, I stuck with it for the first a year and half of the channel.

Lately, I have been experimenting to see if this perceived advantage is as huge as I had imagined. I have seen some YouTube research around this and it is believed it is not that significant except in creating a personal brand where you want your audience to know you.

The key winner in YouTube is always the content - if you present a good material - the audience will watch whether they see your face or not.

From my reviews, my top performing videos are some of my earlier videos where I do not show my face. So, I am still running the experiment and will continue to combine both of them.

The advantage of not showing ones face is that you do not have to worry about camera quality, settings and lighting. YOu just focus on the content. In that respect, it is quicker to make. This is what I have found and so others report too.

The key disadvantage is that it can be difficult to engage with the audience and it will affect your average view duration (i.e. the length of time people are kept watching ones content). So, it all depends on what you want to do.

What I am deciding to do now are (a) make contents that are long and require careful step by step procedure not to have my face. The focus is always on the content and if it is a topic that resonates with my audience, they will keep watching and there will be no problems at all (b) for shorter contents that are reflective or highlights unique features of some things, then it is better they are made with my face in it. This is particularly true for YouTube Shorts as engaging the audience within the first 5 seconds is a requirement else most viewers will just swipe away.

So, there you have it - the slight changes I am exploring in my video making style on YouTube.

Quote of the Week

Insight and Location by Austin Kleon

Here is my quote for the week:

I found this quote while re-reading Austin Kleon's book Steal Like an Artist. I read it a while ago and I loved it. I have the physical copy in my office library at home and regularly read it for inspiration.

This quote is interesting in the sense that our brain - the centre of our cognitive function and creativity with insights - can become too comfortable. Like the image above shows, it relaxes on a sofa of comfort and so unable to think creatively again. Why would this happen? One factor is that our everyday surroundings have a way of making the brain comfortable.

The implication of this is that when you are interesting in creating new things, you need to be on the verge of discomfort where you allow the adrenaline to work. You need that slight discomfort to allow the creative juices to start working. That means if you want to write better, create more or even programme better, then get into an unusual environment where your brain is forced to be awake and think.

This is why so many have written masterpieces by sitting in a coffee shop. A lot of times people speak of ideas popping up in their minds when they are having a shower - a bit of discomfort creating insight.

I have a personal experience of this. Although I have a home office and work office where I do most of my thinking, when I was working on my book manuscript, I spent a lot of times in the library in my university during the writing phase. I was sat with students - just myself and the computer in front of me. I could not play music or actually take a nap. I was focussed on the task and that led to making dramatic progress. The whole nearly 500 pages of the book was written and formatted in Latex within 18 months. Being a technical book with a lot of original thinking, I have been told that is a quick turnaround for the book. This was made possible because I was not in my everyday surrounding.

For you, if you are looking to write your thesis or do that very important conference paper or journal paper, then leave your room, or PhD office space - go to the library. Sit in a coffee shop, hang out in a place where you are not too comfortable and churn away. You will be surprised how your insignt level is directly propertional to your uncomfortable location. This is what Austin was referring to in my quote of the week.

Thank you very much for your patience in reading this far. I will catch up with you next week and in the mean time, I wish you all a wonderful happy weekend. God bless you richly.