7 Ways to Predict Fatigue Failure Before You Manufacture a Part

There is a wonderfully annoying thing about fatigue failure: a part can look completely fine right up until it is not. 

It can survive the first load, the hundredth load, and the ten-thousandth load without giving you much reason to worry. Then a tiny crack starts near a fastener, a bend, a notch, or some other highly stressed area. That crack grows a little every time the part is used until something that seemed perfectly strong suddenly fails. 

That is what makes fatigue such an interesting engineering problem. You are not simply asking, “Will this part hold?” You are asking, “How many times can it do this before something changes?” 

Traditionally, one of the best ways to answer that question has been to build the part and test it. Put it on a fixture, cycle it, inspect it, and keep going until something breaks. That still matters, but discovering a durability problem after the part exists physically is a much more expensive moment to learn that lesson. 

With SOLIDWORKS Simulation and the right fatigue data, engineers can start asking those questions while the design is still digital. 

Here are seven ways to do it. 

1. Simulate What the Part Will Actually Experience 

Before you worry about fatigue curves or cycle counts, start with the thing itself. 

What does this part actually do? 

During a recent GSC Power Hour, Chris used a roughly 50-year-old kitchen turner as an example. It has spent decades sliding under eggs, fish, burgers, and just about anything else that needed to come out of a frying pan. Somewhere along the way, a crack started forming near the riveted connection between the handle and the blade. 

At first glance, you might assume the weight of the food is the important load. Maybe it is. But it could also be the small bending motion that happens every time the edge of the turner presses against the pan and flexes as it slides underneath something. 

That is the interesting part. 

The load that eventually causes fatigue is not always the most dramatic load. Sometimes it is the boring little event that happens thousands of times. 

A good fatigue study starts by identifying those real use cases. For a machine component, that may be vibration. For a clip, it might be opening and closing. For a bracket, repeated bending may be the problem. In a more complicated system, several different loads may matter at different points in the operating cycle. 

The closer the model gets to what the product actually experiences, the more useful the prediction becomes. 

2. Find the Places Where Stress Likes to Collect 

Fatigue cracks tend to show up in familiar places. 

Holes, notches, sharp corners, rivets, welds, fasteners, and sudden changes in thickness all deserve attention because they can concentrate stress in a relatively small area. 

In our kitchen-turner example, the crack appears around the riveted connection. That makes sense. Two parts come together there, the geometry changes, and the surrounding material flexes every time the tool is used. 

This is where a structural simulation can be incredibly useful. Instead of waiting for a crack to tell you where the weak point is, you can look at the stress distribution in the CAD model and start asking questions while the design is still easy to change. 

Could a larger radius help? Could the connection move? Would a little more material spread the load better? Is the problem really the geometry, or is the way the part is restrained creating a misleading result? 

That last question matters because simulation can also create stress hotspots that look terrifying but are partly caused by the model itself. Mesh density, rigid connectors, restraints, and idealized geometry can all influence the numbers. 

So when you see one bright red area on a stress plot, do not immediately rebuild the entire product. Look at the surrounding stress field and ask whether the result makes physical sense. 

The software gives you the numbers. The engineer decides whether the numbers are telling a believable story. 

3. Ask What Happens When You Do It a Million Times 

A standard structural analysis can tell you whether a part survives a load. 

Fatigue analysis asks the more interesting question: What happens if I keep doing it? 

Once you understand the stresses created by a load, SOLIDWORKS Simulation can evaluate what happens when that same event is repeated over thousands or millions of cycles. 

That distinction sounds small, but it changes the entire problem. 

A bicycle crank may easily survive one pedal stroke. A machine bracket may survive one vibration cycle without giving you anything to worry about. A spring clip might work perfectly the first hundred times someone uses it. 

The trouble is not always the load itself. The trouble is repetition. 

Go back to the kitchen turner. One omelet is nothing. Fifty years of omelets apparently becomes an engineering case study. 

Fatigue analysis lets you fast-forward that story and begin estimating where repeated loading starts consuming the life of the product. 

4. Compare the Design Against What the Material Can Actually Handle 

This is the part where fatigue analysis can start sounding much more complicated than it really is. 

Engineers have been breaking material samples on purpose for a very long time. 

Take a sample of steel. Load it repeatedly at a certain stress and count how many cycles it survives. Then test another sample at a different stress. Keep doing that and eventually you can build a relationship between how hard the material is being worked and how long it tends to last. 

That relationship is called an S-N curve, or stress-life curve. 

The basic idea is exactly what you would expect. Work the material harder and it generally survives fewer cycles. Reduce the stress and it generally survives longer. 

SOLIDWORKS Simulation can use that fatigue data to compare the stresses in your design with known material behavior. Now you are not just staring at a stress plot and trying to decide whether red looks bad enough to redesign the part. You are connecting the stresses in the design to data about how that material behaves after repeated use. 

The quality of that data matters, of course. A solver can produce a wonderfully precise answer from questionable inputs. Matching the fatigue data to the actual material and application is still part of the engineering work. 

5. Look at Where the Design Is Spending Its Life 

One of the most useful things fatigue analysis can show you is not simply where stress is high, but where the design is actually consuming fatigue life. 

SOLIDWORKS Simulation can produce damage results that show where repeated loading is taking the biggest toll. It can also estimate life, giving you an idea of how many loading cycles the component may survive under the conditions you modeled. 

It is very tempting to focus on that cycle count. 

Suppose the analysis predicts 875,432 cycles. That number looks wonderfully exact. It feels like you should be able to put it on a calendar and make sure nobody uses the part for cycle 875,433. 

Real products are not that cooperative. 

Materials vary. Manufacturing varies. Surface finish matters. Temperature matters. Corrosion matters. And customers have a remarkable ability to find loading conditions the engineering team never imagined. 

The more useful question is often, “Where is this design losing life, and what changes make it better?” 

Maybe increasing a radius makes a dramatic difference. Maybe moving a hole helps. Maybe changing the material buys you several times more life. You may even discover that the load you were worried about barely matters, while another everyday load is doing most of the damage. 

Those are useful things to learn before you have a physical part sitting on the bench. 

6. Try to Break Your Own Assumptions 

Every simulation contains assumptions. That is not a problem. That is modeling. 

The interesting question is how much those assumptions matter. 

Our kitchen turner is a good example. Those rivets were installed decades ago by physically deforming material. Did that process leave residual stress around the joint? Almost certainly. Do we know exactly how much? Probably not. 

Then there is heat from the pan, surface wear, material variation, and decades of use. A computer model is not going to perfectly recreate every minute of that history. 

What you can do is test the uncertainty. 

Run the expected load case, then make it a little worse. Change a fatigue reduction factor. Increase the load. Try another material. Compare different operating conditions. 

If the design still looks healthy when you make the assumptions more conservative, that gives you more confidence. If a tiny change causes the predicted fatigue life to collapse, you have learned something equally useful: the design may not have much margin. 

This is one of the parts of simulation I find most useful. You are not just asking the computer for an answer. You are poking at that answer to see how easily it falls apart. 

7. Find the Problem While It Is Still Cheap 

This is really the reason to do any of this before manufacturing. 

Imagine a fatigue study shows that one small transition in the geometry is consuming most of the predicted life. If the part is still in SOLIDWORKS, the fix may be a larger radius, a little more material, a different feature location, or another material choice. 

That might take an afternoon. 

Now imagine finding the same problem after tooling is complete, prototypes are built, durability testing has been running for several weeks, and the project is already moving toward production. 

Same crack. Completely different problem. 

Catching fatigue risks early gives engineers room to experiment. You can compare several versions of the design, try different materials, evaluate different load cases, and understand what actually improves durability before anybody commits to production hardware. 

That does not eliminate physical testing. It makes physical testing much more valuable because you are testing a design that has already survived a lot more questioning. 

What If the Part Is Actually Bending Permanently? 

Most of what we have been talking about so far falls into high-cycle fatigue, where a part is loaded repeatedly but mostly returns to its original shape. 

If the material starts permanently deforming every time it is loaded, the problem changes. 

Think about a paper clip. Bend it back and forth and you can actually feel the material changing. It takes on a new shape, gets worked harder, and eventually snaps. 

That is closer to low-cycle fatigue, where strain and permanent deformation become much more important. 

For those types of problems, engineering teams may need a different approach. SIMULIA tools available through the 3DEXPERIENCE platform can extend the analysis into strain-based durability when the material behavior becomes more complicated. 

You do not need to memorize every fatigue method or solver. The more important skill is recognizing when the behavior of the part has changed enough that you need to ask a different engineering question. 

The Point Is Not to Stop Testing 

Physical durability testing still matters, and there are things a real manufactured part will teach you that a simulation never can. 

Real parts have manufacturing variation. They have surface imperfections. They have assembly differences, environmental effects, unexpected contact, and all of the strange little realities that are hard to model perfectly. 

The advantage of simulation is that you get to learn a lot before you reach that point. 

You can find a potential weak area while changing the geometry still takes minutes instead of weeks. You can compare materials before ordering either one. You can identify which loading condition is actually driving fatigue and send a much stronger design into the test lab. 

There is a big difference between using physical testing to discover what is wrong with your design and using physical testing to confirm what you already understand about your design. 

That is really the goal. 

You are not replacing engineering judgment. You are giving engineers better information earlier, when they still have plenty of options for what to do with it. 

And if that same thinking happens to explain why a 50-year-old kitchen turner is finally starting to give up after a lifetime of omelets, that is a pretty good bonus. 

Trying to Understand How Long Your Design Will Last? 

Fatigue problems can be relatively simple, like one repeated bending load, or they can involve complicated combinations of vibration, temperature, materials, manufacturing effects, and changing operating conditions. 

The right analysis depends on the product and the question you are trying to answer. 

If your team is trying to improve durability, investigate a recurring failure, reduce unnecessary prototype iterations, or determine whether fatigue analysis fits into your development process, talk with the GSC simulation team. We can help you look at the application, the loads, and the available tools to determine the right approach before you commit to manufacturing.