12.5Ultrasonic Methods

12.5.2.1 Overview

Extensometers employ a technique based on the transit time of an ultrasonic pulse traveling along the fastener’s axis, but with allowance for the reduction of sonic velocity with increasing stress.

The principle of operation is as follows:

• a. A pulser within the instrument shocks a transducer that then delivers a brief, highly damped pulse of ultrasound to one end of the fastener.
• b. This pulse travels down the fastener, echoes off the far end, and returns to the original transducer.
• c. The instrument measures and records the transit time of the pulse.
• d. The fastener is then tightened, increasing the length of the path to be covered by the signal and decreasing the sonic velocity.
• e. The new increased transit time is recorded.
• f. Using this data, the instrument (typically incorporating a microprocessor) computes the total change in transit time, computes and discards that portion of the change which resulted from the change in velocity, and then displays the actual change in fastener length.

A further important factor, which has to be considered, is the effect of temperature. A temperature change can affect the length of the fastener and the sonic velocity. If temperature of the fastener changes between the unstressed and stressed readings, it is good practice to factor out its influence on length and velocity by the measuring equipment.

Therefore, before commencing preload measurement, three factors can be set into the instrument. These are 'velocity' and 'stress' factors for the fastener material and a 'temperature' factor. It is important to determine these factors experimentally determined, since theoretical calculation of the relationship between preload and change in transit time is not possible. Extensometer manufacturers have determined factors for many materials and these are available to the user. In cases where the data is not available, the instrument can be user-calibrated on a sample of the fastener. Subsequent recalibration is usually done by means of a calibration block provided with the instrument.

12.5.2.2 Using an Ultrasonic Extensometer

After calibration a typical procedure (Reference 12.1) for tightening a joint is as follows:

• a. The instrument is set in the length mode.
• b. The temperature of the fastener to be tightened is measured and the appropriate temperature factor input into the instrument.
• c. An acoustic couplant such as glycerine or oil is applied to the fastener and the transducer set in place. This can be attached to either end of the fastener.
• d. The initial length of the fastener is ultrasonically measured and recorded for future reference.
• e. Steps 3 and 4 are repeated for other fasteners in the joint.
• f. The instrument is switched to the change of length mode and the extension necessary to give the desired preload is calculated.
• g. The joint is pulled together by tightening each fastener to approximately 30% of the required preload. Torque control is commonly used for this pass. The length of the fasteners can be monitored at the end of the pass to confirm that results are reasonable.
• h. Additional passes are made, tightening the fastener to a higher percentage of the required preload. A small sample of fasteners is measured with the extensometer at the end of each pass, but torque is still used for control.
• i. The extensometer is used to control the tightening process for values of preload above, typically, 70% of the required preload.
• j. It is recommended to remeasure every fastener in the joint after the final pass to monitor the relaxation effects such as embedding.

12.5.2.3 Practical Considerations

In laboratory conditions instruments are capable of resolving to between 2,5 mm and 0,25 mm. However, in the field the performance can be significantly worse, by an order of magnitude or more.

The accuracy to which preload can be controlled is a function of fastener length. It is important to take into the account the fastener diameter if it is less than ten times the sonic wavelength.

As for the angle controlled tightening method, variations in fastener and joint stiffness are reflected in preload scatter and are to be considered.

Transducers vary in frequency, diameter, signal strength, acoustic damping, etc, and at present, selection is still largely a matter of trial and error.

Good and consistent acoustic coupling of the transducer is necessary. However, it is not possible to couple all fastener configurations since a sufficiently large mounting area is available for the transducer and the opposite end provides a substantial signal reflection. Hexagon headed fasteners usually work well but socket head cap screws, do not permit this method unless they are large enough to allow the transducer to be mounted in the wrench socket, or beside it. If the transducer is too large, it can receive unwanted reflections from the underhead interstice, etc.

Fastener identification markings can prevent a satisfactory acoustic coupling and machining of the head can be necessary to eliminate air, which creates acoustic impedance. Even with perfectly flat surfaces, consistent coupling can be difficult to achieve, since varying the contact pressure between the transducer and the fastener can significantly alter results. In critical applications, it can be necessary to accurately control the pressure between the surfaces, as described in Section 12.4.2.

Different materials have different effects on the acoustic signal with regard to absorption and scattering. For example, some stainless steels are grainy and can distort the signal. The American A286 material is believed to contain Titanium platelets within it that cause false echoes and severe attenuation. Irregular hardness, e.g. case hardening, can also warp the signal.

The fact that preload is actually calculated from the strain needs that the length of the portion of the fastener under tension is accurately known. This is usually not the case and therefore, the standard method of calibration can be inadequate in cases where a high degree of accuracy is needed.

12.5.2.4 Increasing the Accuracy of Preload Measurement

Although the procedures described in section 12.2 and 12.3 above enable fastener preload to be controlled far more accurately than is possible by torque control or angle controlled tightening methods, generally the methods lead to preload scatter of greater than 5%. Higher accuracy can be achieved by making two main modifications to the procedure.

• a. Increasing the repeatability of the coupling between the transducer and fastener head. This can be achieved by using transducers that can be screwed into a tapped recess in the head. By tightening the transducer to a predetermined torque, more consistent coupling is achieved. It is important to note, that the fasteners need to be specially machined and the transducers have to be specially made.
• b. Using a direct method for deriving a set of calibration data, specific to the fasteners being used. As mentioned previously, the usual method of calibration involves measuring the fastener extension to calculate preload with the inherent inaccuracies mentioned in section 12.4.2. These inaccuracies can be eliminated by measuring the fastener load and extension relationship as follows:

1. The fastener is passed through two small steel plates sandwiching a suitable load washer, the total thickness being equal to the thickness of the clamped parts of the real joint.
2. The nut is tightened to give various increments of preload.
3. For each increment the preload and extension readings (given by the load washer and ultrasonic extensometer) are recorded.
4. Any desired preload can now be obtained, knowing the required extension reading.

Approximate values of the calibration factors are input into the instrument at the start. These do not need to be exact, since any error in the extension reading is unimportant, as the output has been calibrated against load required. However, it is important to use the same factors for all future measurements for the calibrated joint configuration.

An appreciable amount of data has been obtained using this method (References 12.2 and 12.3) but it tends to be very specific to a certain type and size of fastener. If only a small number of different fasteners are involved, it can be practicable to calibrate each type of fastener for its specific required preload. However, if the number is large this is often not possible. It is possible to use data obtained for a limited variety of specimens to develop empirical formulae that hold true for fasteners of different dimensions, (see Reference 12.4). It has been found that the following relationship can be applied:

where L1 is the plain length of the fastener and L2 is the threaded length of the fastener up to the face of the nut, and, a, b and c are constants, to be evaluated for the particular fastener system.

12.5.2.5 Resonant Oscillator Methods

Examples of these methods include the Reflection Oscillator Ultrasonic Spectrometer (ROUS) and improved Pseudo Continuous Wave (PCW) methods. The first step is to establish a resonant condition in the fastener. As the fastener is stressed the wave velocity through it decreases. The frequency of exciting is held constant by the transducer and thus the wavelength therefore decreases (since the velocity is the product of the wavelength and the frequency). Therefore, to re-establish a resonant condition, it is necessary to decrease the transducer frequency.

The reader is directed to References 12.5 to 12.9 for more detailed information.

12.5.2.6 Scatter

It is not possible to give an accurate "maximum error" figure covering all fastener configurations, since this depends on many factors, including fastener dimensions, fastener material and the type of nut. However, as a guide to the order of magnitude of errors, Table 12-2 shows values of 'B' value scatter, for extensometer readings for fasteners of six different diameters that were tested as described in Section 12.4.2 (see also Reference 12.4). It is important to note that, for each diameter, different lengths of fastener were tested. Since the largest errors are likely to be incurred with the shortest fastener, the minimum length used is quoted.

It can be seen that the worst cases in this set of tests produced scatter values in the region of 6%, while it is possible in some cases to obtain results within 1%.