When is Digital Validation the Right Tool?

Taking design simulation beyond a buzzword

You’ve probably heard the term “simulation” tossed around a lot lately. It’s become a huge buzzword in the development field. In fact, it’s touted as the “cure all” to any design problems and promises as a way to slash development times and avoid costly errors. While there is a good amount of truth to this, there are some situations where it’s just not the right tool for the job.

In this blog post we’ll outline simulation’s strength, where and when they are best used, the different levels of difficulty and complexity for common simulation types, and the pitfalls they can have. Let’s get started:

What Is Simulation? What Can It Do For You?

First, let’s examine some background on simulation. The most common types of simulation tools used today are:

• Mechanical stress
• Thermal
• Computational fluid dynamics (CFD)
• Injection molding simulation
• Circuit simulation

Stress-based simulations typically show what happens to an object when forces or loads are applied. Some common examples of this include drop tests, snap fits, user applied loads, or loads from use. This is the broadest and most frequently discussed area of simulation.

Thermal simulations are used both in mechanical and electrical design. Common examples of this are heatsink design for cooling of electronics, visualizing how heat will propagate through a system, and heat generated during use for mechanical systems with friction, or heat generating during the use of an electric motor or combustion engine.

Computational Fluid Dynamics (CFD) are the some of the most complex simulations to perform. They are used to size fans for cooling applications, calculate loads due to aerodynamic drag or lift, optimize enclosure sizing for electronics, and visualize fluid flow within a system such as pluming or filling tanks.

Injection molding simulation is highly specialized and as you can guess can help with the design optimization of injection-molded parts or of their tooling.

And finally, circuit simulation ranges from simple individual circuit simulation to calculate capacitor and resistor values for filters to complex board level simulation to optimize layout for power dissipation and noise suppression.

Let’s talk about how to leverage simulation as a tool
To explain where simulation is useful vs. where is not, we’ll use the example of designing a hoverboard wheel. To show the strengths and weaknesses of mechanical stress simulations, let’s look at the design of the shaft connecting the wheel to the board. In this example, it is important that the shaft be strong enough to support the weight of the rider and withstand the rotational forces applied from the motor to move the rider forward. We will use a static simulation, one where only one instance in time is simulated, which is desirable because its vastly simplifies the simulation from the standpoint of difficulty to set up and especially solve time.

It’s important to keep in mind that the more complex a simulation, the more computing power and more time is required to solve it. An engineer would apply the weight of the rider and the peak torque of the motor to the model, the material of the shaft, and the areas where the shaft is fixed and where the forces are applied. From this, the engineer is able to see what the stresses are throughout the part. For instance, if the stress is above the failure limit for the material, it would not survive and the engineer could either change the shape of the shaft or the material to ensure that the stresses do not cause the part to fail.

In addition, how the shaft bends or moves under load could also be investigated. This data generated from the simulation can be invaluable to a design engineer, especially early in the design process. With this information available to the design engineer, the shape of the part and the material chosen can be optimized for cost reasons. Testing many designs and choosing geometry that is easier to manufacture, yet still have the required strength is one of the hallmarks of simulation technology. Many material choices can also be easily tested so a less expensive material can be chosen.

Another great example of mechanical simulation is the ability to test for fatigue failure over time. Just because a part holds up during initial prototype testing doesn’t mean it doesn’t have a ticking time bomb of failure, and this failure sometimes isn’t seen until the product has already been released. When developing software, these issues are easy to fix later with a software update, in the hardware world this is called a “recall”, which is something that should be avoided at all costs.

Simulation allows the engineer to see the estimated cycles to failure of a part. With this information, the part can be sized so that it will at least survive the expected life cycle of the product, thus avoiding a costly recall and the bad publicity that comes from one. All of this information can be obtained early in the design process. This has an exponential effect on reducing the development time of the overall product.

Knowing, with confidence, the required size and dimensions of individual parts early on reduces the time needed for rework on other parts. If this shaft size in the hoverboard wheel had to change after initial physical prototype testing, it could potentially have a domino effect on other design changes throughout the product from bearings to space changes that could affect the space available for other needed components in other departments, such as circuit boards or wiring harnesses. If each of these other parts also would have to be redesigned due to a change in the shaft, then, of course, the development time just expanded far more than just the time to change the design of the shaft. Avoiding these downstream design changes is the single most cost saving aspect of using simulation effectively.

This simulation is a great example of where simulation shines. Before this technology was used, the engineer would have to make many assumptions to simplify the problem enough to solve by hand, or would have to make an educated guess and over-engineer the product so that it would be way stronger than needed. Using simulation ensures the final product is optimized so that it can be as economical to produce and there is a level of confidence that it will not fail in the field.

Where does simulation fall short?

Now that we’ve explained why simulation has earned its title of the golden child of modern development, it’s time to examine some pitfalls with its use.

Over-Confidence

An issue some engineers run into when using simulation is overconfidence. Many times an engineer designs the part so it will just barely not fail within the simulation. This can lead to issues if the part experiences forces that were not originally accounted for in the simulation or if the simulation wasn’t solved accurately enough.

A good rule of thumb for simulation is to have a safety factor of between 2–5 depending on the product and how well the use case of the product is understood when the initial simulation is performed.

Answering Questions

An even larger issue for developers is to decide when simulation can be used to answer questions in an effective manner. We’ll use the hoverboard wheel again to demonstrate this. Now that the shaft size is determined, we’ll try and size the motor’s needed output torque using simulation.

This type of simulation requires a dynamic simulation, one where multiple instances in time are solved. It also requires the majority of the hoverboard’s model to be analyzed at once. These factors lead to the simulation being very difficult to set up, not to mention costly (in terms of time) to run. An engineer might have to spend weeks of his time setting this up.

Re-Solving

It could also require solving it many times, which often adds up to days per solve to see the results. This is an enormous use of resources just to gain the information to optimize the torque of the motor. This information could be obtained much faster with traditional hand calculations. However, the results from the hand calculations might only get you within 50% of the right answer, they take significantly less resources to obtain.

With a healthy safety factor, the hand calculations would be plenty close enough for use in design, and the added accuracy from the simulation would have marginal gains in cost reduction for the motor.

Knowing when and where simulation can lead to efficiency gains in the development process is important to optimizing its use. As we’ve outlined, it can be easy to spend many resources on performing a simulation that would yield very little gains to the product.

Yes, simulation can be an incredibly valuable tool for development. It has the ability to enable designs not possible before through optimization of designs, reduction in the costs of goods during production, and the shrinking of development cycles.

With the availability and democratization of simulation tools for many different fields of engineering, it is disrupting the development industry by allowing more products to be designed better, make it to market faster, with less need for capital. As with any new tool, it must be understood in order to be used effectively. Utilizing a design firm with experience in simulation can unlock its potential. At Glassboard, we utilize simulation early and often in the development cycle to ensure our customers’ end products are the best they can be.

Grant Chapman is the VP of Operations at Glassboard a hardware focused product development company.