**2. Experimental Approach**

The goal of the experiments performed was to monitor fatigue crack propagation in standardized specimens. The test samples for both Al7075-T6 and Ti-6-4 were supplied in the form of compact tension (CT) specimens, based on ASTM standard E647 [18]. The test specimens were manufactured from 0.125 in (3.175 mm) thick plates. The geometry and dimensions of the specimens are shown in Figure 2.

A servo-hydraulic Material Testing System (MTS) fatigue machine retrofitted with an Instron 8800 controller was used to define and control the loading conditions. An advanced DiSP-4 AE system supplied by the Physical Acoustics Corporation (now MISTRAS Group), was used to record the AE signals. The setup included one 100–900 kHz wideband differential AE sensor (WD) to collect the signal, a 40 dB preamplifier to amplify them, a data acquisition module to perform primary filtration and record the signals, and a software module to visualize the data and to perform feature extraction.

**Figure 2.** Technical drawing of the CT specimen (dimensions in inches).

A ruler with smallest divisions of 0.01 in. (0.254 mm) was placed on the specimen so the actual crack length could be measured during post-processing of the images. The AE sensor was attached to the upper left-hand side of the specimen with silicone grease and held in place with a C-clamp (Figure 3). The instrumented CT specimens were mounted on the 5 kip MTS machine which applied sinusoidal loading cycles until a fatigue pre-crack of adequate length and straightness (according to ASTM E647) could be detected. Then, the standard crack growth tests were conducted using the same MTS machine. All the experiments were continued until fracture.

**Figure 3.** Standard CT specimen with a mounted AE sensor.

A 40 dB preamplifier amplified the AE signals received from the sensor. A band pass filter was used in the amplifier, and the amplified signals were analyzed using the DiSP-4 system. The AE features that were measured and recorded included AE counts, energy, and the time of the event. A digital close-up camera took time-lapsed high-resolution pictures of the crack growth, and the crack length could be measured based on the ruler attached onto the specimen within the accuracy of 0.01 inch (0.254 mm). As a result, considerable amounts of data were captured and stored for further processing.

Experiments performed on several CT specimens were subjected to cyclic loading with different loading ratios. For each case, the loading ratio was changed, with all other test parameters remaining constant. Table 1 lists the experiments performed for the corresponding load ratios:


**Table 1.** Details of experiments and load parameters.

For the three Ti-6-4 tests, the loading was increased because Ti-6-4 has a higher yield strength of about 120 ksi (830 MPa) compared to the yield strength of Al7075-T6 of about 67 ksi (460 MPa) [19]. Additionally, the frequency was decreased to 5 Hz in order to better distinguish and process AE events during post-processing [20].
