·Table of Contents
·Materials Characterization and testing
Non-destructive characterisation of Nodular Cast Irons by Ultrasonic methodC. Hakan GÜR, Birand AYDINMAKINA
Middle East Technical University,
Metallurgical and Materials Engineering Department, 06531 Ankara - Turkey
Tel: 00-90-312-2105916, Fax: 00-90-312-2101267,
Email : firstname.lastname@example.org
It was concluded that ultrasonic techniques are sensitive to the changes in the graphite morphology, thus they can also be used to predict the mechanical properties. The ultrasonic techniques eliminate the need for preparation and destructive testing of specimens, furthermore, they can be carried out in a few minutes, and are cost effective.
Keywords: Ultrasonics, Ductile Iron, Graphite Nodularity, Sound Velocity, Attenuation
Ductile cast iron (nodular cast iron) is cast iron in which the graphite is present as tiny spheres. Because of additives that are introduced in the molten iron just before casting, the graphite grows as spheres, rather than as flakes of any of the forms that are characteristic of gray iron. Cast iron containing nodular graphite has markedly greater strength and greater ductility than gray iron of similar composition. The relatively high strength and toughness of ductile irons give it an advantage over gray iron in many structural applications. Some typical applications of nodular iron are automotive and diesel-crankshafts, pistons, and cylinder heads, electrical fittings, switch boxes, motor frames and circuit breaker parts, mining-hoist drums, drive pulleys, flywheels, and elevator buckets, steel mill-work rolls, furnace doors, table rolls, and bearings, tool and die-wrenches, levers, handles, clamp frames, and miscellaneous dies for shaping steel, aluminium, brass, bronze, and titanium.
Most of the specifications for standard grades of ductile iron are based on properties-that is, strength and/of hardness is specified for each grade of ductile iron, and composition is either loosely specified or made subordinate to the mechanical properties. Many ductile iron castings are used as cast, but in some foundries, 50% or more are heat treated. Heat-treated ductile iron usually has more uniform mechanical properties than as-cast ductile iron, particularly in castings with wide variations in section thickness . Many castings are given a ferritizing anneal or are normalized to produce a uniform pearlitic microstructure. Other castings are heat treated to produce bainitic or martensitic matrixes. As the matrix structure is varied progressively from ferrite to ferrite plus pearlite to pearlite to bainite and finally to martensite, hardness, strength and wear resistance increase, but impact resistance, ductile and machinability decrease.
Quality control in the production of nodular cast irons is very critical, since it is very sensitivite to the process variables influencing the microstructure and mechanical properties. Frequent chemical, mechanical and metallurgical testing is needed to ensure that the required quality is maintained and that specifications are met. However, the results of metallographic examination and tensile tests may not be truly representative of the real castings, because most of these procedures include testing samples from large heats of iron, rather than the actual castings. As a result, the information received is not enough to guarantee that all the castings are of desired quality meeting the minimum requirements. These methods are destructive, labour intensive and time-consuming. It is necessary to develop an NDT method that could identify the nodularity, and the matrix structures for quality control and product reliability purposes .
While there are no discrepancies about the predominant role of graphite morphology on ultrasonic velocity, the relationships between indications of NDT measurements and metallic matrix structure, especially after various types of heat treatment, are not clear enough. Most of the researchers believe that the matrix structure does not affect these measurements. A similar conclusion was made from the ultrasonic velocity measurements, that prediction of nodularity was independent from metallic matrix structure. For metallic matrix evaluation, it is recommended to combine the ultrasonic velocity with a second measurement that is sensitive to changes in the metallic matrix, such as hardness.
2.1 Production of Specimens
To produce cast irons with different graphite nodularities, three groups of samples were cast into the Si-based sand molds by applying different procedures. Because of the reaction between metallic magnesium and molten iron is violent at this temperature, "sandwich process" was used. The casting temperature of Group-I and II was 1430oC, however, their pouring times were different. For Group-I, it was aimed to have maximum percentage of nodularity, therefore pouring time after Mg inoculation was chosen as 2 minutes. For Group II, since the nodularity percentage was aimed to be lower than that of Group-I, the molten metal was poured into the molds 8.5 minutes after Mg inoculation. Since the aim was to produce grey cast iron in Group-III, the casting temperature was 1360oC and no post inoculation was exposed. After casting, a representative sample was analysed with a spectrometer.
From each slab, four cubic specimens were prepared. After proving that the ultrasonic values were not dependent on the direction of the measurement, the cubes were cut into two parts. The specimens were machined and ground to obtain the parallel surfaces (at least ± 3o). The dimensions of the specimens are given in Table 1.
|Element||Gr. I||Gr. II||Gr. III|
|Table 1: Compositions of Group-I, II and III, and specimen preparation|
2.2 Heat Treatment
To investigate the effects of matrix phases on the mechanical and acoustic properties, various heat treatment procedures were carried out on the specimens selected from each group. Specimens were heated to the austenitization temperature in the Wild Barfield furnace. After holding for 1 hour at 860oC, the first group of the specimens selected from Group-I, II and III was normalized by cooling in air (N), the remained specimens were annealed by furnace cooling down to room temperature (A).
2.3 Metallographic Investigation
After surface preparation, the specimens were examined under the optical microscope with the magnifications of x100 and x200, first without etching to see the nodularity of graphite, then, after etching in 2% HNO3 solution to determine the matrix phases.
2.4 Ultrasonic Measurements
Pulse-echo technique with A-scan presentation was used in the measurements. The longitudinal ultrasonic waves were generated and applied by KrautKramer straight-beam probes with the diameters of 10, 24 and 24 mm, respectively. The wave was transmitted into the specimen through oil-base couplant. The signals were amplified and displayed on the analogue KrautKramer Branson equipment. A constant pressing force on the probe was applied to maintain a constant thickness of couplant films at the probe-specimen interface. Time-base calibrations were performed on miniature IIW angle-beam block (K2).
The propagation velocity of the sound was determined using the formula given below,
where d is the thickness of the specimen, Vr is the velocity of sound in the calibration block, k is a measure in mm representing one scale on the time-base range of the screen, and T is the scale value of the peak from backwall. Measurements were repeated three times on each surface, and then, the mean value of the velocity was calculated.
Attenuation measurements were performed in accordance with ASTM E 214 and ASTM E 664. It was aimed to obtain 10 backwall echoes on the screen. The first backwall echo was considered as the reference peak and its amplitude was adjusted to a constant value to allow other peaks to be compared with. Three independent measurements were done on each surface and the average value was calculated. The apparent attenuation is calculated according to the following formula,
where Am and An are the amplitudes of the mth and nth back reflections in decibel (dB) where n>m and t is the thickness of the specimen. In this study, m is chosen as first backwall echo height and n is chosen as second backwall echo height.
2.5 Hardness Tests and Determination of Density
The hardness tests were performed by Heckert hardness device. Vickers hardness was taken with 30 kp preload for Group I and II specimens and 15 kp preload for Group III specimens. For the sake of systematic study, the indentations for hardness have applied over some specific regions. Five indentations for hardness have been applied on each specimen in a systematic way, and the averages of these values were calculated.
Density values were determined by mass/volume formulation using 8 digit digital weight measuring device and 0,01 mm accurate mechanical micrometer. After calculating the densities of the specimens, the mean value of each group was calculated.
3.1 Nodularity and Matrix Phase
In the metallographic investigations, the approximate percentages of the phases present in the matrix, the percent nodularity and average sizes of graphite nodules was determined by the image analyzer (Buehler Omnimet Advantage Model).
Three groups of cast iron was used with nodularity percentage of 90%, 50%, 0% and these were named as Group-I, II, and III respectively. Graphite size was determined three times at the randomly selected points on each specimen, and the averages were calculated. In the case of as-cast specimens, the matrix consists of 65% Ferrite and 35% Pearlite for Group-I, 80% Ferrite and 20% Pearlite for Group-II, and 5% Ferrite and 95% Pearlite for Group-III. In the case of heat treated specimens, the matrix consists of approximately 100 % ferrite in the annealed specimens; and 100 % pearlite in the normalized specimens.
3.2 Initial Measurements to Control the Isotropic Behaviour
To investigate the directional dependency, measurements were performed at the three perpendicular surfaces of the cubes. After cutting the slabs into cubes (40 mm side length), the surfaces were ground and polished. Velocity measurements were carried out by using a straight-beam probe of 4 MHz with 10 mm diameter. Standard deviations were very small (about 10 m/sec) compared to the mean ultrasonic velocity. Therefore, there are not significant differences in the velocities depending on measurement direction. Apparent attenuation was measured at 2 MHz by considering first and second backwall echoes. While the average values are within the range between 0,038 and 0,068, the standard deviations fluctuate between 0,001 and 0,016. This difference can be related with the variations in the grain size throughout the cast piece. The average of apparent attenuation values is 0,045 dB/mm, and the average of standard deviations is 0,006. The average of apparent attenuation for surface A is 0,041 dB/mm, that for surface B is 0,044 dB/mm and that for surface C is 0,049 dB/mm. There is no directional dependence of attenuation in the specimens concerned.
3.3 Effect of Nodularity on Ultrasonic Velocity in As-Cast Specimen
The thickness of the specimens was measured by a micrometer having 0,01 mm accuracy. Time-base range (SB) of the test equipment was calibrated to 20 mm on the K2 reference block since the thickness value is about 15.5 mm. To determine the effect of the testing frequency on the results, the velocity measurements were repeated for the frequencies of 4, 2 and 1 MHz.
In each group, the velocities were measured at least on 15 specimens and twice on each specimen. Then, mean value and standard deviation for the respective group were calculated (Table-2). The standard deviation values indicate that fluctuations in the velocity values within the groups are very low, i.e., the specimens within the groups have consistent properties.
|Testing Frequency||Velocity (m/sec)||Standard Deviation||Velocity (m/sec)||Standard Deviation||Velocity (m/sec)||Standard Deviation|
|Table 2: The average of the ultrasonic velocity values for each group and corresponding standard deviation obtained at different testing frequencies.|
There is a direct relation between the acoustic velocity and the nodularity percentage, and testing frequency has no influence on this relationship (Figure-1). The results indicate that there is good correlation between nodularity and ultrasonic velocity, therefore ultrasonic velocity in cast irons, to a great extent, reflects the graphite morphology. In the same figure, there is a sudden drop of ultrasonic velocity below 25% of nodularity. This was also figured out in the literature  and this represents the velocity of longitudinal wave propagation in gray cast iron in which the graphite appeared in flake form. This sudden drop indicates that the material is unable to be classified as ductile iron. For the nodularity values less than 25%, the sharp edges of graphite cause critical stress concentrations; which results in a decrease in Young's Modulus.
|Fig 1: Variation of longitudinal ultrasonic velocity as a function of % nodularity at various testing frequencies (velocity value for 25% nodularity was taken from literature).|
3.4 Effect of Nodularity on Acoustic Attenuation
With decreasing nodularity percentage, possibility of getting further backwall echoes decreases. This is due to the scattering of ultrasonic waves by flake graphite much more readily than spheroidal graphite. Thus, ductile iron with lower nodularity percentage resulted in higher attenuation of sound energy in terms of echo amplitude.
Apparent attenuation calculated from 1st and 2nd backwall echoes do increase with decreasing nodularity percentage. The mean apparent attenuation values are 0,173 dB/mm, 0,248 dB/mm and 0,588 dB/mm for groups I, II, III respectively. This is why, graphite flakes are much more appropriate to reflect and scatter the wave.
Figure 2 gives the variation of relative amplitude with respect to number of backwall echoes for 4 MHz. Since the thicknesses of the specimens are almost the same (maximum variation is about 0,2 mm), the decrease in the relative amplitude can be correlated with backwall echoes.
|Fig 2: The attenuation of ultrasonic wave for various nodularities at 4 MHz|
With decreasing nodularity percentage, the relative amplitude decreases in other words attenuation increases. The mean apparent attenuation values for at 4 and 2 MHz given in Table-2 can be compared to evaluate the effect of frequency on attenuation. Lowering the frequency will result in a decrease in attenuation.
|Testing Frequency||Apparent Attenuation (dB/mm)|
|Group I |
|Group II |
|Group III |
|Table 3: Average apparent attenuation values for two different testing frequencies|
3.5 Correlation between Microstructure, Hardness and Acoustic Properties
The variation of ultrasonic velocity as a function of nodularity percentage can be explained by density and modulus of elasticity. Densities were measured on the representative specimens. The values are 7,055 and 7,061 gr/cm3 for Group-I specimens; 7,02 and 7,024 gr/cm3 for Group-II specimens, respectively. The differences are negligibly small, thus the reason for the variation in the ultrasonic velocities is the mechanical property.
The microstructure of Group-I specimens consists of 65% ferrite, 35% pearlite, graphite with 90% nodularity and 0.0266 mm mean size; that of Group-II specimens consists of 80% ferrite, 20% pearlite, graphite with 50% nodularity and 0.024 mm mean size; and finally that of Group-II specimens contains 5% ferrite, 95% pearlite and flake graphite. The average hardness values of the specimens are 219, 161 and 130 HV for Group-I, -II, and -III respectively.
Ductile iron with the highest nodularity, i.e., highest elastic modulus, gave the highest velocity, and as the nodularity percentage decreased the sound velocity of ductile iron also decreased. The consistent difference in ultrasonic velocities and attenuation depending on the nodularity allows determining quickly and accurately the quality of ductile irons.
3.6 Effect of Matrix on Hardness, Ultrasonic Velocity and Attenuation
The influence of matrix structure on ultrasonic velocity was relatively unapparent in the specimens. In literature it was reported that the slope of the stress-strain curve is not affected by the matrix structure, but influenced only by the amount of graphite present. Since the compositions of the specimens were almost similar in this study, the amount of graphite in various matrices was also similar. Thus, the effect of the matrix structure on the modulus of elasticity, and hence ultrasonic velocity would not be significant. It can be stated that cast irons with fully ferritic and fully pearlitic matrices exhibit the same ultrasonic velocity responses against nodularity, although the effect of phases is very clear in terms of hardness values.
If the factors affecting the mechanical properties of the cast irons were not only due to nodularity, then the correlation between ultrasonic velocity and mechanical properties would not be simple. There are slight differences in the attenuation behaviour of different matrices. For a constant nodularity value, the attenuation effect of the matrix phases increases in the order of ferrite, pearlite and martensite. This difference can be seen better in the measurements at 4 MHz, compared to those at 2 MHz. Since the wavelength of the ultrasonic waves excited by the 4 MHz probe is shorter, the measurement becomes more sensitive to the differences in the matrix structure. Using of higher frequencies may lead to the better differentiation of the matrix effects.
At these experimental conditions it is not significant to differentiate the effects of matrix phases on the mechanical properties only by velocity measurements. The attenuation measurements are more sensitive to the changes in the matrix structure. However, further research activities are necessary by repeating the measurements at higher testing frequencies.
|I||Matrix Phase||65%F+35%P||93-100% F||82-90% P|
|Graphite Size (mm)||0,0266||0,0228||0,0260|
|Attenuation Coeff. (dB/mm)||-0,3805||-0,3509||-0,3173|
|II||Matrix Phase||80%F+20%P||95-100% F||90% P|
|Graphite Size (mm)||0,0240||0,0242||0,0220|
|Attenuation Coeff. (dB/mm)||-0,435||-0,466||-0,442|
|III||Matrix Phase||5%F+95%P||100% F||95-100% P|
|Nodularity||Attenuation Coeff. (dB/mm)||-1,1513||-1,2102||-1,1676|
|Table 4: Summarized results of the experimental investigations|
Nodularity of a cast iron can be determined by longitudinal ultrasonic velocity and attenuation measurements. As the % nodularity increases, the ultrasonic velocity also increases, and converges to that of steel. Since the density differences among the specimens are negligible, the main reason for velocity change is the elastic modulus. Changing the probe frequency did not affect the velocity values. It is also possible to detect locally unnodularized zones with ultrasonic technique. If surfaces according to the requirements are obtained, and if there is no side wall effects, The ultrasonic velocity measurements can be used as a quality control tool in a foundry.
Attenuation measurements also predict the graphite morphology of a casting. Any change in the nodularity percentage affects the attenuation of sound within the component. While the % nodularity decreases, the attenuation coefficient increases due to the scattering effects of the non-spherical graphites. Attenuation measurements are sensitive to the testing frequency selected. However, the influence of matrix structures on the variation of ultrasonic velocity was relatively unapparent. The attenuation measurements may give an indication about the matrix phases. But, the formulation of the general behaviour needs further research. For the evaluation of the metallic matrix, it is recommended to combine the ultrasonic measurements with another technique that is more sensitive to the changes in the matrix structure.
|© AIPnD , created by NDT.net|||Home| |Top||