Design engineers with considerable experience in metals may overlook the effect creep can have on the long-term performance of a plastic part, particularly when applications involve high operating temperatures.
When a stress is applied over a long time, the observed strain is greater than that predicted by the short-term modulus. This additional strain is called creep. Creep occurs in metals, but is a much more significant design factor in plastics. It can generally be stated that the closer the applied stress is to the ultimate stress (or failure point), the more significant creep becomes for plastics.
“Creep is a phenomenon that must be taken into account when designing with plastics, especially if the application is subjected to thermal cycling,” comments K.C. Desai, CAE manager for Solvay Advanced Polymers, Alpharetta, GA. The polymer structure, as well as reinforcement and fillers, affect the material’s tendency to creep. Unfortunately, creep cannot be determined directly from data sheet information. A consequence of creep is stress relaxation.
Consider a plastic part attached with a bolt torqued to achieve sufficient compressive stress. Over time the plastic will creep, which in turn lowers the compressive stress. As the stress drops, the tendency to creep is diminished until an equilibrium is achieved at a lower stress level. Because of this, it is usually necessary to increase the initial torque to avoid a loss of bolt tightness.
Issues that can result from a poor understanding of creep range from premature part failure to excessively high safety factors due to overcompensation, such as sections that are too thick. A graph of apparent modulus versus time should be used as a reference before selecting a material for a given application. Apparent modulus is the ratio of applied stress to creep strain.
Creep properties are evaluated through the measurement of strain as a function of time while a test specimen is subjected to a constant tensile, compressive or flexural stress at specified conditions. The primary caveat of these tests, however, is that the tests occur under ideal conditions. Understanding the effects of temperature and chemical atmosphere is extremely important.
A useful tool for analyzing creep is the isochronous (equal time) stress versus strain curves, shown here.
Plot the stresses and the resultant strains for a specific period of time, such as 100 hrs, then draw a smooth curve through the points. Repeat the process for other time intervals and other temperatures. The apparent modulus at any point can be calculated by dividing the stress by the indicated strain.
Do not infer that these stress rates will occur over the entire part. Engineers who do make that mistake, design parts that are too thick, thus wasting material, slowing cycle times during molding, and limiting design opportunity. Use visco-plastic models in CAE programs to predict the specific points of stress where ribs or other techniques can be used for added strength.
For clamped parts, the compressive creep can affect bolt torque over the life of a part, especially when cyclical elevated service temperatures are encountered. Plastic parts bolted in vehicle engine areas are perfect examples. Typically, plastics have coefficients of thermal expansion that are greater than those of metals. Higher operating temperatures will promote higher compressive stresses in the plastic which results in higher creep.
If a bolt/washer combination is used to secure a plastic flange to a steel frame, you must consider compressive stresses in the plastic located under the washer. When the assembly cools, the plastic will shrink more than the bolt, relaxing bolt torque. As the part vibrates, the bolt could even loosen. A formula that will predict the stress can be found on page 86 of the Amodel® PPA Design Guide.
One solution is a design in which a metal sleeve is placed in the hole of the plastic flange in the accompanying illustration. The sleeve can be insert molded or put in place in the molded hole by heat staking, ultrasonic insertion, or press-fit. Cyclical stress is another factor designers must consider for applications such as gears. As a driving gear causes a driven gear to rotate, each tooth is subject to stress. As the wheel turns there is no stress on that tooth until the gears engage again.
When selecting a material for the gear teeth, look at the maximum stress in the gear and the allowable stress level for the material for number of cycles anticipated. Endurance limit, that is the stress level that can be tolerated for an infinite number of cycles, is not valid for plastics or many non-ferrous metals. S-N curves (stress vs. number of cycles to failure) are necessary for proper material selection.
More detailed data on S-N curves, stress/strain curves and apparent modulus can be found in our design guides. Careful engineering analysis takes the stress out of designing load-bearing parts in plastics.
The Inside Skinny
Failure to account for creep can lead to failure in a plastic design, particularly when heat is encountered. Other leading pitfalls include the following:
- Mold shrinkage is listed as xxx in/in in flow direction and yyy in/in in transverse direction. Engineers try to predict exactly what the shrinkage will be on a part that combines both flow and transverse patterns, then cut their tool exactly to dimension. What happens is that they are inevitably wrong and have cut too much steel and have to start over.
- A mold fill analysis will give a gate recommendation as “minimum gate size to fill” the part. So the gate is cut to that size and they have fill problems. The “minimum size to fill” is an ideal world; in the real world, you always need more than the minimum.
- Steel prices have been near record highs and people try to save a few bucks by using less expensive steel. Bad decision. Even though a steel grade may be rated for the mold operating temp, say 135°C (275°F), what the steel actually “sees” every shot is the melt temperature of the plastic, which is significantly higher.
- The conventional wisdom is that more cavities in a single tool equal lower cost per part (a single eight cavity tool vs. two four-cavity tools, for example.) That’s not always true in the big picture. If your only eight-cavity tool has to go out for repair, you are out of business. With two smaller tools, at least you can get some production out when one goes down.
- A common perception is that a hot runner system will completely eliminate regrind with no problems. What happens is that there is often an issue with a cold slug or drool, especially with sharp melting point semi-crystalline materials. To remedy this, people try to install valve gates, which are high maintenance and expensive. The better alternative may be use of a “hot-to-cold” runner system and cluster the parts. This cuts way down on the cost, eliminates the cold slug/drool issues and only generates a tiny amount of regrind.