MECHANICAL EVENT SIMULATION AT WORK

BOSCH ENGINEERING USES MECHANICAL EVENT SIMULATION TO PRODUCE FAST, ACCURATE ANALYSES OF A HIGH VOLTAGE CIRCUIT BREAKER COMPONENT.

By Gerhard Bosch

This screen capture shows von Mises stress results for the first timestep of the event simulation of a switch component. The goal of the simulation was to determine the closing time and acceleration of the switch and prove that the design could be optimized by altering material properties. Note that the unit of measure used is N/m2. To view .avi files of the switch event simulation, please visit Algor’s Software in Action section.	The switch is shown at the moment it contacts the impact wall, which simulates the switch housing. Unique to Accupak/VE software from Algor, Inc., impact walls automatically compute the impact, bouncing, and sliding behavior of the contacting object.
The switch undesirably contacts the impact wall beneath its resting position because too much energy was retained by its spring. This behavior could not have been established using traditional FEA and would still have been a surprise in the prototype-testing phase despite previous computer-aided analyses.	Stresses result as the switch contacts an impact wall as it reaches its resting position. The green outline indicates the initial position of the switch.

June 16, 1998, Pittsburgh, Pennsylvania -- Today’s engineering requires that we create better, safer designs in a shorter time frame. Engineers of all disciplines use computer-aided engineering, including traditional FEA, to meet such requirements by attempting to simulate physical behavior.

Often, my colleagues and I at Bosch Engineering, a consulting firm in Oldenzaal, The Netherlands, have found that traditional FEA enables us to make assumptions about an object’s behavior, but does not adequately simulate its physical behavior. Accupak/VE Mechanical Event Simulation (MES) software (Algor, Inc., Pittsburgh, PA), however, combines traditional FEA with physics, taking into account time, motion, and impact. These factors were critical in the following situation where we used Accupak/VE to test a switch component design for a high voltage circuit break.

Accupak/VE offers the unique ability to conduct "virtual prototyping" with a computer. This capability provides comprehensive analysis results about an event, shortens the product design process, and reduces the need for iterative laboratory testing. As you will see, this was true in the case of the switch component.

The Switch Component

The switch is comprised of a spring with a stress relief hole to ensure the correct amount of stress in its initial position and to control the switch’s potential energy when it returns to its resting position upon release. The contacting element is linked with a single joint and is free to rotate within the realm of the circuit housing.

Accupak/VE expands the scope of traditional stress analysis to include parts of the event that produce force, and carries the event through as much motion as is of interest. To simulate events, Accupak/VE accounts for interaction between motion and stresses by incorporating both Newton’s second law of motion and Hooke’s law.

The combination of motion and stress analysis enables us to see motion and its resulting aftermath such as impact, buckling, permanent deformation, fracturing, and additional motion. Without Accupak/VE in this situation, we would have spent days, instead of hours, gathering input data and conducting separate linear, nonlinear, and kinematic analyses. Besides the disadvantages of doubling evaluation efforts and transferring the results from one analysis to the input of another, it would have been difficult to establish the correct spring stiffness of the switch, which is not constant in every position.

The goal of our event simulation was to simulate the closing time and acceleration of the switch and to prove that the design could be optimized by altering material parameters, such as damping and stiffness, or the geometry. Furthermore, we had to prove that small changes in these parameters due to manufacturing differences or wear and tear during the lifetime of the switch assembly did not influence its functionality significantly. It was important to simulate the effects of all dynamic responses of the components themselves, as well as the change in the stiffness of the spring with a tolerance-free link.

Setup and Design Considerations

Designing an Accupak/VE simulation is similar to designing a traditional FEA problem, with one key difference: traditional FEA requires that you make assumptions about forces and boundary conditions in the problem setup. Accupak/VE does not demand these assumptions though it does require inputs for motion, time, and the position of the object in space. While this may take more setup time, we find that our analysis results are more comprehensive, giving us information about resulting forces and the way the design reacts to them.

All switch components were modeled using Superdraw, Algor’s CAD for engineering, with all elastic and free joints defined with no rotational degrees of freedom. In this situation, the links could not be modeled as rigid elements as they would in a kinematic analysis because their stiffness would influence the behavior of the total mass/spring/damper system. Therefore, not only mass and motion effects were taken into account, but also the stiffness of parts.

We defined impact walls, a feature unique to Accupak/VE MES software, to simulate the contact of the switch with the circuit housing. Impact walls compute the impact, bouncing, and sliding behavior of a contacting object. Without this feature, we would have been required to make more assumptions about contacting object behavior. Accupak/VE also requires that we define the period of time over which the analysis will take place. This additional input often increases the analysis processing time of an event, but it can provide invaluable information, as you will see, during the analysis process.

Analyzing the Event

Again, we note that processing an Accupak/VE simulation is similar to traditional FEA; however, we have found many advantages to using Accupak/VE. It offers a monitoring option that, when activated at the start of the analysis, enables us to view analysis results during processing. We can verify that the analysis is running properly, and thus save time by identifying potential problems early on. Also, the option exists to suspend the analysis and later restart it when the computer is not needed, overnight for instance, since the number of calculations required could take several hours to complete.

We found the ability to analyze stresses during the analysis of motion critical in the switch design because fatigue life needed to be estimated accurately. Finding the most critical timestep or position with respect to stresses would always be a guess without this capability. Because we received stress and acceleration values for every timestep, we eliminated the possibility of unexpected outcomes when we reviewed the results.

Viewing Analysis Results

Upon reviewing the initial analysis results, we found that the switch was forced beyond its projected rest position, undesirably hitting an impact wall beneath that position. This behavior could not have been established using separate linear, nonlinear, and kinematic analyses and would still have been a surprise in prototype testing despite previous computer-aided analyses.

Accupak/VE analysis results have proven to be very accurate. In our experience, they are more accurate than multiple FEA problems performed separately and then combined. With Accupak/VE, more inputs are entered in the problem setup, therefore less assumptions are needed. For example, we would not have to make assumptions about how an acceleration found in a separate kinematic analysis might influence the stresses and deflections found in a linear stress analysis of a part. Also, less of the responsibility of determining accuracy is left to the perception and experience of the individual engineer.

Once our clients were pleased with the analysis results, a simple, hand-machined prototype of the design and its parts were tested. Our analysis results were confirmed with a greater than 99 percent rate of accuracy.

As consultants and product developers, we at Bosch Engineering have always found Accupak/VE MES software easy to use and reliable with the most powerful meshing tools and links to virtually any CAD system. Algor’s fully graphical environment allows us to visualize an analysis of any design and easily identify the strengths and weaknesses of it. In the end, designs need to be backed up by less physical prototype testing, so costs of the design process are much less. Final prototype testing is almost always performed, but 99 times out of 100, testing proves that Accupak/VE is accurate.

The bar graph shows comparative hypothetical new design processes: one using traditional FEA and one using Accupak/VE Mechanical Event Simulation software. Because Accupak/VE combines traditional stress analyses with real-world physical data, we receive actual data about the forces acting upon the product. Thus, we are able to get more precise results with less prototype testing. This shortens the product design process and decreases time-to-market.

My colleagues and I at Bosch Engineering and our clients benefit from these advantages with an enormous decrease in design processing time and thus costs. We use Algor FEA and MES in virtually every corner of the developers’ market, from shipbuilders and plastics developers to tank and vessel manufacturers.

Gerhard Bosch is an independent consultant and president of Bosch Engineering, Oldenzaal, The Netherlands. Contact him via e-mail by clicking here.