Dealing with Stress and Interference

The design engineer and simulation

—by Alan Wegienka, Product Manager Dynamic Designer, MSC.Software Corporation
and Baskar Rajagopalan, Analysis Leader, Dassault Systèmes

Under intense pressure to turn out better designs in less time and at a lower cost, design engineers are looking towards a new generation of simulation tools tailored for their needs earlier in the design process. Once considered the domain of analysts with Ph.D.’s and sophisticated work stations, today’s PC-based simulation software empowers the design engineer with proof-of-concept, optimization and testing long before expensive, time-consuming physical prototypes are built. As a result, design engineers gain a better understanding of product/assembly behavior, resulting in fewer errors, better products, lower design, manufacturing and warranty costs, and faster speed to market.

Typically, design engineers use simulation to check stress levels and loads on bearings and actuators, identify hidden stresses caused by motion, and ensure life expectancy meets specified objectives. Simulation follows three phases of the product development process — Proof of Concept, Design Optimization and Simulation of Physical Test. In Proof of Concept, determination of whether the design will work (and not break) permits maximum freedom to find the best design solution, because related constraints are at a minimum. As the design progresses, added constraints make it more difficult and/or expensive to introduce change. By Design Optimization, the product is merely fine-tuned to meet established performance goals and minimize material and manufacturing costs. Simulating physical test is the last check of design performance before building a physical prototype.

As already mentioned, simulation can replace the substantial expense and time required for creating and testing physical prototypes. Moreover, a physical prototype is, basically, a pass/fail test. Strain gages are attached to perceived stress points, so the margin by which a physical prototype passes or fails a test is only known if and where measurements are taken, which in turn assumes they were taken at the critical areas. Simulation, on the other hand, enables the design engineer to run a series of what-if tests that identify which features affect functionality and how they do so. This provides information and understanding to create better performing products with a minimum of expense and time.

The difference in the objectives between design engineers and analysts is driving ease of use and other differences in simulation products. As a result, simulation products for the design engineer have been developed where some or all of the steps are performed in the background, transparent to the user. Design simulation products can be loosely grouped as those embedded within a CAD system, standalone products and automated products.

The simulation process

The basic finite element (FE) simulation process consists of five steps, some of which can be made transparent to the user to facilitate use:

1. A geometric model is created in a CAD system that can produce a precise solid and surface geometry.
2. An FE model is created by dividing the geometric model into discrete, predictable elements (also called meshing). Element and material properties are assigned.
3. The operating environment is defined by applying loads and boundary conditions.
4. The structural response (deflection, stress, motion, temperature, etc.) is processed.
5. The results are compared to the design criteria and the process repeated, if necessary.

Some engineers may perform loads and restraints first and then create the mesh, but in most FE simulation programs, it doesn’t matter. In the motion simulation process, the steps are simpler because a mesh does not need to be created; e.g., create geometry, define constraints and motion drivers, run simulation and review results.

To make simulation tools more accessible to design engineers with novice-to-expert levels of experience, features have been built into the software that typically involve automating some or all of the simulation steps listed above. Process automation, such as this, reduces training time, enabling faster simulation throughput. The main issue is matching the type and amount of automation to a particular application.

Most of the simulation necessary for the early design stages is motion simulation and/or stress analysis. Functionality in stress and motion simulation can be highly automated, including opening 3D solid model geometry, meshing, specifying material properties and identifying constraints. It can be very time-intensive to rebuild a model for analysis, determine the proper size mesh or figure out what constraints to apply. Simulation software today eliminates the guesswork by automating these assumptions, gathering information from how the model was built in the CAD system and speeding up what used to require a lot of manual calculations. Where the simulation process is the same for different parts, automation tool kits and customization can simplify the process based on existing company knowledge.

Embedded software

The goal of embedded simulation software is to make the technology transparent to the designer; that is, working right inside the CAD program with the same familiar interface and operating directly on the CAD-built geometry. The constraints on the assembly model are used directly in the motion simulation model, and the same geometry is used to directly build the FE model. Ultimately, the design engineer sees the motion simulation or FE model as results viewed inside the CAD system — animation, interference checking or XY plot of velocity of stress contour tasks can all be displayed without requiring the engineer to learn a different interface. This minimizes training time, allowing the design engineer to become productive much quicker and speeding design revisions.

Standalone software

Standalone simulation products fall into two categories — those that open native CAD geometry files, automatically meshing and assigning constraints to the model; and those that import geometry and require some or all of the meshing to be done manually. In the past, all meshing was performed manually, incurring a substantial amount of time.

Some highly automated standalone products enable both motion and stress analysis, so engineers can use physics-based motion to test the dynamics of an assembly, as well as identify stress failure areas. For the expert design engineer, there are standalone products with automeshing that analyze the stress, vibration, dynamic, nonlinear, heat transfer and fluid flow characteristics of a wide range of structures and mechanical components.

Automated solutions

Automated solutions are divided into standard and custom products, each capturing expert knowledge and making it available to the design engineer. Typically, standard solutions are useful for situations where the simulation process is the same from product to product, or where there are a large number of repetitive tasks to be performed.

In addition to saving time, a major benefit of an automated solution is knowledge-capture process automation, obtaining best of class practices and standardizing processes. Specific knowledge acquired is used to build a template that prompts the design engineer for information. These solutions require little if any training to perform simulation, because geometry is constructed and meshed, and materials properties and boundary conditions are assigned, automatically.

Providing design engineers with simulations early in the product development process establishes proof of concept and design validation prior to physical test, resulting in greater innovation, better quality prod-ucts and lower costs. Simulation applications give more reliable, easily testable trends than physical test data, and therefore can readily guide and improve a given design. Simulation can also support knowledge-capture process automation, which facilitates best-in-class practices and standardized processes. All of this puts the design engineer in the driver’s seat by cutting costs, improving product performance and quality, reducing time to market, and ultimately, increasing profits. In the end, design engineers enhance their over-all skill sets, and become more valuable to the organization.

“We are just beginning to be able to do complete simulations,” adds Hofferman. “Once we have a solid model in the computer, we can do stress, thermal, and flow analysis on the part. Then, we might want to make a few small changes to fine-tune the design.” The software also enables the engineers to do some long-term simulations such as structural analysis. “We do a lot of engine packaging and moving compo-nents around for greater efficiency, such as radiator design changes and modifying ducting,” he concludes. “We are always trying to use fewer parts, lighter parts, and more durable parts.”