Putting Tensile specimens to the Test.

Studies in aluminum iron tensile testing reveal discrepancies the test specimen and casting and offer solutions to counter these problems.

Tensile testing, although commonly used, has its limitations. Foundries and their customers use it primarily to determine if their castings meet strength specifications. In reality, however, this test only should be one tool to certify casting strength.

The limitations of tensile testing begin with the fact that the test is performed at room temperature at a slow rate of strain. Most "engineered" castings (especially ductile iron castings), however, are subjected to a wide variety of temperatures during performance, according to P.H. Mani, a consultant with MRA, Springfield, Illinois, who evaluated ductile iron tensile testing with Lyle R. Jenkins, the former technical director of the Ductile Iron Society, North Olmsted, Ohio. As a result, the mechanical properties of a casting change with time/temperature and in different environments, such as a highly corrosive one. Conventional testing does not take this into account.

In his studies of microporosity, Franco Chiesa of the Dept. of Metallurgy at College de Toris-Riveres, Quebec, found mechanical property discrepancies between aluminum castings and their tensile test counterparts. He concluded the test was not inaccurate, but it didn't account for the metallurgical structure differences between the casting and the separately cast specimens. In one of his studies, Chiesa found that the castings usually solidified at a slower rate than the separately cast or attached specimens, resulting in a coarser microstructure.

Despite its limitations, Mani and Chiesa consider tensile testing vital to satisfy customers, because most casting specifications list mechanical properties based on tensile tests. For structural castings, the tensile test is "the" practical test to assess overall mechanical strength, according to Chiesa. The key is for foundries to communicate the limitations to their customers to ensure the castings being produced meet the proper mechanical property levels.

This article examines two limitations of tensile testing--controlled testing conditions and variations in casting test specimen properties--and solutions to overcome those limitations.

Misleading Information

The technique (Fig. 1) for testing tensile properties is identical for all metals. The data relies an the test sample or specimen. Data from the test identifies a specimen's elongation (the stretch during the fracture), yield strength (the stress at which a specified limiting permanent strain begins to manifest) and ultimate tensile stress (the maximum tensile stress).

Tensile specimen bars (Fig. 2) may be machined from a casting' itself or cast separately. In the case of ductile iron, test bars maybe machined from a separately cast keel or Y block.

Through his work, Mani found that the tensile test can be misleading in the assessment of how a casting will perform in service. The subtle changes in the process controls for making ductile iron castings can show substantial differences in the service performance of a casting when it is tested dynamically. Conventional tensile testing under a static uniaxial loading at a slow strain rate may mask this fact by recording the strength as "acceptable," whereas under a dynamic test, such as impact or fatigue, differences may be detected.

Mani points out that most "engineered, critical, life/limb" castings produced are subject to dynamic loading with cantilever beam stress, three-point beam stress, torsion stress or a combination of stresses, while in service. The stressed areas in the casting may contain sharp fillets or notches from threads. The conventional tensile test doesn't take any of these into account.

Furthermore, castings may be subjected to stresses from temperatures ranging from lower than -40F (-40C) to higher than 100F (38C), which would change the castings' mechanical properties. In other words, the conditions of the tensile test rarely match the service performance conditions of the casting's use in the real world. Unlike the tensile test, few castings are subjected to a uniaxial tensile loading at a slow strain rate at room temperature.

Alternatives

Mani suggests, as one option, using a compact tensile specimen to determine the number of cycles and the amount of strain needed to start a fatigue crack and its rate of growth. This may provide some valuable information on the strain rate sensitivity of the material.

Many engineers recognize the strain rate sensitivity of ductile iron, says Mani. Strain rate sensitivity is the property that explains how the material behaves under different strain rates and different rates of loading. Foundries should be aware of the factors that affect the strain rate sensitivity of ductile iron and the critical strain rate, which produces the brittle fracture. Some of the factors are chemistry--especially the content of silicon, phosphorus, nitrogen and some trace elements--nodule size and distribution, nodule alignment, matrix structure (as-cast vs. heat treat) and matrix cleanliness.

Strain rate sensitivity must be considered in casting design and application and with the process used in the production of the final part. For example, if the ductile parts are subjected to coining or pressing, it should be done in a slow hydraulic press as opposed to a high load rate stoke press, advises Mani. Similarly, if one were to roll thread the parts, it should be done at a slow rate to avoid cracks. The same applies to fillet rolling as well.

A simpler and practical test for acceptance criteria for castings could be the assessment of the degree of brittle (cleavage) fracture from a dynamic tear test bar. This test provides qualitative information about the dynamic properties of the material. Such a test could provide the end user of the casting valuable information.

Mani also recommends that foundries work with the end user to specify the amount of cleavage fracture acceptable at a specified low temperature for the grade of iron based on the performance criteria of the casting. The load/stress, strain and strain rate that the part is subjected to under the worst possible conditions needs to be taken into account. The specimen could comply with the dimensional standards set by the American Society of Testing and Materials (ASTM). The method of measurement of the cleavage fracture also needs to be spelled out. For example, a customer could specify that at -20F (29C), the amount of cleavage fracture should not exceed 50% on a fractured sample for as-cast ferritic grade ductile iron.

Tracking the Incongruities

In a study, Chiesa tested aluminum tensile specimens. Using flat specimens cut from a plate, round specimens attached to the gating system and round specimens separately cast, he showed tensile properties poured from the same melt change with the process and vary within the same casting.

The study also revealed that:

* tensile properties of the round specimens are substantially higher that that of the flat specimens;

* tensile properties of separately cast round specimens are slightly higher than that of the round specimens attached to the gating.

These improved properties of the separately cast round specimens may be attributed to shorter solidification times, resulting from the higher radiant heat losses at the outer surface of the shell of the separately cast specimens.

If such discrepancies exist between test specimens from the same pour, then Chiesa concluded that cast specimens, whether separate or attached to the gating system, could exhibit mechanical properties higher than the casting. Therefore, defining the geometry and the location of the specimens on which a specification will be based takes on the utmost importance.

If a given minimal mechanical property is required at a designated area of the casting (or for the entire casting), then it must be specified, as allowed by standard ASTM B686: "Standard Specification for Aluminum Alloy Castings, High-Strength." The foundry must make sure the user and designer are aware of unavoidable fluctuations in mechanical properties throughout a casting. As a result, the foundry may have to use special devices such as chills to meet these properties in specific problematic areas of the casting.

What Foundries Must Tell Users

To avoid any confusion or misunderstanding of the performance of a casting and tensile testing, foundries need to emphasize some points to their customers.

Mani suggests that foundrymen and the end user should understand:

* the tensile test does not guarantee the actual functional integrity of the castings they produce;

* tensile tests on test bars from actual castings give more accurate information about the product than test bars taken from keel blocks or other methods;

* the results can vary very widely depending on how and where the castings are taken during the process;

* foundries should not completely rely on tensile testing, especially if it is making "critical/safety, life/limb" parts.

Chiesa adds:

* mechanical properties within a part must be expected to vary;

* the "universal" mechanical properties for a casting alloy, as determined on a standard shaped specimen, will not be exactly reproduced in the casting because the solidification conditions are not identical;

* mechanical properties inside a cast product can be insured via the proper use of standard ASTM B686.

Modeling Reveals Variations in Tensile Specimens

Computer solidification modeling is helpful in predicting/adjusting the thermal factors that determine the fineness of dendritic structure for castings and tensile specimens, according to Franco Chiesa of the Dept. of Metallurgy at the College de Trois-Riveres, Quebec. Since the fineness of the metallurgical structure, as indicated by the dendrite arm spacing (DAS), is directly related to the solidification time, the DAS distribution can be predicted inside the casting via solidification modeling.

Chiesa modeled the solidification of flat tensile specimens cut from a plate, round tensile specimens attached to the gating system and round tensile specimens separately cast.

In the study, the solidification model (Fig. 3) predicted a coarser structure at the location of both flat tensile specimens than at the attached as-cast round specimen. The round specimen also is favored by a higher temperature gradient (Fig. 4). This will provide better feeding and result in a lower level of microporosity.

What this means is casting properties are predictable, so foundries can explain the variation in tensile properties by the differences in solidification conditions, according to Chiesa.