top of page
  • Writer's pictureNed Patton

How to Keep it From Getting Broke

I want to talk a little here about how to design a composite part so that it can tolerate a little damage and still function – in other words a little damage doesn’t completely break the part. This is what is called “damage tolerance” in composites design, and is one of the fundamentals of good design with composites. The history of fault tolerant design of course came from the aviation industry, which is why airplanes rarely if ever fall out of the sky.

What I’m talking about is being able to avoid something like this to one of your designs.

There are several different ways to approach making your composite design not break in normal service, or even in some extreme service cases – like the aircraft that you see to the left here. It all starts with your ability to foresee the future a bit and create scenarios that you can analyze that could potentially cause your part, or your airplane, or your boat, to break. This does take a little bit of dreaming up ways that whatever you are making could be caused to break by a user that isn’t paying attention. Or by some event that should not happen but could happen. It is best to not try to think too far and wide about things unless you are designing something like armor where you need to protect your user from advertent attempts to break your composite design.

Once you have dreamed up all of the potential means whereby your composite design could break, it is accepted practice to do some sort of analysis to determine exactly how your composite design might break. This is the first step in starting a fault tolerant design. The analyses that you do should tell you where you need to add material or blend out a curvature to make it more gentle or any number of things that you can do to make your composite design not break – on the computer anyway. As I talk about simulation or analysis here, I am talking about creating a 3D model of the part on the computer, applying the material properties that you need to use for your composite design, and then running what is called a Finite Element Analysis. What this means in terms most can understand is to break up the part into smaller pieces on the computer, assign material properties to each little part, tie them all together so that they look like your part, fix some of the little pieces in place where it makes realistic sense to do that, and apply loads to the entire part. This used to be a fairly arduous task, but now it has been made much easier thanks to the advances in both the computer speeds and memory as well as the development of point and click interfaces for all of these computer based tools.

The next logical step, once you have done all of your analyses and made all of the design alterations that you have made because of your analyses, is to make a prototype and test it to see if you did it right.

This method is the most cost effective of all of the methods that I am going to present in this blog. Since making prototypes out of composites is relatively costly compared to just doing calculations on a computer, this method generally wins the day in a composite design development project. And it also saves you time because you have already done your homework with the design and you have high confidence that your composite design won’t break under not only normal service conditions, but also under some of the more extreme, accident-like conditions that your composite design will see.

This is in fact the way most composite designs are developed today. Our analytical tools have become sophisticated enough that simulations of composite designs is quite practical. This of course was not the case when I started out in this business about 40 years ago. At that time those of us that needed to analyze composites had to develop our own tools. This changed in the 1990’s, and by the early 2010’s the tools had matured to the point that they had mostly point and click, intuitive interfaces that allowed the composite designer to come to a robust design fairly quickly.

Another method that has pretty much fallen out of favor except for certain cases where the loads are actually either mostly unknown or difficult to simulate. I like to call this the “Edisonian” approach because it is largely trial and error. It is, however, useful for when the design is critical for safety and yet needs to be optimized for weight. In these sorts of designs, you are working right at the hairy edge of failure in a very highly loaded environment, and you have to get it right or someone dies. NASA has been dealing with this problem ever since the early days of manned space flight. There were lots of test firings of rockets designed to carry people into space that blew up on the launch pad, flew off course and blew up in mid-air, or just completely failed to take off in the first place. In every single case the engineers at NASA knew that they needed to learn from every failure if they were ever going to get anywhere.

Still another way of closing on a design to make sure it doesn’t break is if you are designing something destined to replace a metal part. To make the best use of your resources, it pays to first do a complete simulation of the metal part to see how it breaks. Finding already broken metal parts to inspect as long as you know how they broke is the first step in this process. Then once you have figured out what makes the metal part break, simulate it in that scenario and see if your simulation shows that it breaks. This benchmarks your simulation and you can now go on to developing a design for a replacement part made out of composites. And you can go back to the first scenario I talked about as long as you have confidence in your modeling and simulation skills. You will know the loads that made the metal part break, so you can design around those with your new composite design with full confidence that your first prototype will not break even under severe conditions.

Well, that’s my take on how to develop a composite design to keep it from getting broke in service. Hope you enjoyed this one. If you think that this worth reading, please sign up so that you get a new blog each week. And, if you want to see this on LinkedIn, it will be my latest post there -

See all of you next week.


Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page