Ultrasonic Testing of Pipeline Girth Welds

Until recently, pipeline weld inspection has been traditionally solely the domain of radiography. With the advent of mechanized gas metal arc welding (GMAW) ultrasonics has proven to be an effective option to detect nonfusion defects oriented unfavourably for radiography.

Several companies now provide mechanized UT inspection of pipeline girth welds. Techniques for this application utilize many of the advances made in equipment and computerization made possible only recently. These systems incorporate many aspects of technology discussed throughout this book and their description here provides a fitting summary of technology as applied in the field.

Systems Basics

Figure 23

For several years now, ultrasonics has been used as a nondestructive testing method to evaluate the integrity of automatic welded girth welds. For the purpose of defect identification and Engineering Critical Assessment (ECA) the weld is divided into several zones. Each zone covers about 2 mm of thickness of the weld. Ultrasonic transducers are designed and positioned to investigate each zone from both sides of the weld centreline. The array of probes is moved around the girth weld by a motorized carrier which travels along the same track the welding apparatus uses. Signals received by the ultrasonic instruments are monitored by electronic gates and both amplitude of signal and its time of arrival (the point where the signal interrupts the gate) can be collected. The weld region is monitored from just before the design fusion line to just after the weld centreline. The gated output is usually digitized and then displayed for evaluation by the operator. In some cases a computer monitor is the display on which the evaluations are made while in others it is a multichannel strip chart which is evaluated.

In evaluating the scan results the operator makes a decision as to weld acceptability based on the length of a signal exceeding a threshold as set out in the Company Specification.

Although nonfusion is the most likely defect to occur in GMAW it is certainly not the only defect. Therefore the operator cannot call all signals over the threshold lack of fusion. To more reliably evaluate the nature of the signal an operator must use all possible information available on the displays. Figure 10-23 is a block diagram of a UT inspection system for pipeline girth welds

Transducers

The accuracy of zone definition has been made possible by the development of contact focused beam ultrasonic transducers (also called internally focused transducers). (see chapter 3).

Gated Information

Displays (computer monitors or paper charts) provide information about the origin of a signal. Distance from the probe and relative reflecting area of the source of the signal is the only information that is collected by the Ultrasonic (UT) instrument. This must be coordinated with the point the beam enters the metal and the angle it travels in the metal to allow the operator to collect reliable information.

Amplitude of a signal is the displacement from the baseline that a signal makes on the UT scope. This displacement is output as a voltage, from zero for no signal in the gate to a maximum of 10, 12 or 15 volts (depending on the manufacturer) for a full screen displacement. Voltage output is then translated to a point on the strip chart with chart position proportional to the signal height.

Similarly, time can be displayed by using the gate length as the voltage variable. The earliest point on the gate is zero volts and the longest (latest in time) is 10 volts. Again "time voltage" is converted to a strip chart signal whose position is proportional to the time a signal occurs in the gate. In addition to strip chart display, amplitudes and time (or equivalent distance traveled) can be presented as colours. Colour display is used by the operator as an aid in quicker evaluations. Chart displays of gated information is explained by Figure 10-24. Two signals are shown as they might appear on a scope. The amplitudes and times in the gated region are shown as increased voltages on the chart. Signal S1 is about 80% of the screen height and occurs early in the gate. The early arrival is indicated by the low voltage in the time trace, about 2 volts. The second signal, S2, has a lower amplitude and occurs later in the gate. This is indicated by the lower (4 volt) amplitude displacement on the chart and the higher time displacement (about 8 volts).

Figure 10-24

Surface Preparation Requirements

In order for the probes to be positioned correctly, for the beam to intersect the zonal targets, they are located at varying distances from the weld centreline. This requires the surface over which the probe moves to be free of any material that would impede the sound or hinder its movement along the pipe surface. This requires the pipe surface be cleared of all weld spatter and dirt as well as ensuring the pipe coating is removed far enough back to allow the most distant probe to move over the pipe without any part of the probe being on coating. The maximum distance will be determined by a combination of thickness of pipe and the angles used. However, typically the coating cut-back will need to be about 100mm.

Calibration

Calibration of the UT inspection apparatus involves setting up the system to receive signals from targets of a known size and at known position. Primary targets for nonfusion type defects are 2 mm diameter flat bottom holes. These are arranged in a sample section of pipe at points corresponding to the weld preparation geometry. In addition to the primary targets there are targets for undercut (1 mm deep notches) and another target for confirmation of range to ensure the weld is inspected to just past the centreline (a centreline slot or through hole).

These targets are illustrated in Figure 10-25 along with the path taken by the centre of the sound beam. It will be noted that the path taken for inspecting the 1st Fill is different. This involves the use of a tandem pair of probes. One probe transmits the beam which bounces off the ID and OD surfaces as well as the target. Since its path does not return to the transmitting probe another probe is used as a receiver to detect any signals associated with defects in the path of the transmitted beam. Examination of the Fill 2 path indicates the position of the probe is well back from the weld centreline, this results from the pulse echo mode the probe is used in.

Figure 10-25
Hot Pass 1 & 2
Land Area (LCP)
Fills 1 and 2 and OD Undercut
Root Area
	Thanks to RTD Quality Services Ltd. for the above figures

Gates are arranged such that sufficient material is covered to ensure the fusion line is included in the range monitored as well as the weld metal to a point about 1mm beyond the centreline. This will usually require the gate start be arranged to be about 3mm before the fusion line. The extra before the fusion line allows for some misalignment due to mismatch and defects that might occur earlier in time such as burnthrough. Figure 10-26 illustrates the regions of concern covered in the gated time by two of the probes.

Calibration Charts

Differences in equipment and displays exist. However a good example of the principles is shown in Figure 10-27, (courtesy of Shaw Pipeline Services). The chart is divided into four regions. On the right side is identification information. Location, customer, weld and chart resolution are indicated as well as a numbering system for the target identification, e.g. highlighted indication #2 is the Fill Hole target. Along the top the dark lines within the bars identifies locations where the amplitude in the particular gate has exceeded the preset threshold (40%). Similar dark lines along the bottom of the chart identify locations where coupling was not maintained. The main area of the chart is taken up with voltage output (vertical) versus position (horizontal) for each gate in the system. The middle two traces are the gated time outputs for the root and the others are gated amplitude outputs. The traces are symmetric about the middle, the upper seven traces dedicated to the downstream probes and the lower seven to the upstream probes (i.e. the side of the weld inspected is identified as up or downstream).

Evaluation of Signals

Several sources of signals can occur in the ultrasonic testing of GMAW joints. Not all signals occurring in the gated region are the result of defects and not all defects are called for repair. Signals may be a result of geometry being different from the ideal (such as results from mismatch of pipe) or the defects may not achieve a length or amplitude sufficient to require a repair. The ultrasonic operator must assemble all the available information to evaluate the source and severity of the signals.

Figure 28: SCHEMATIC REPRESENTATION OF GMAW DISCONTINUITIES

The Schematic illustrates the main discontinuities found in GMAW welded joints. The descriptions that follow are the foundations for the operator's evaluation of the signals. Open this Schematic as Image Map together with the table Separate Page

Different Manufacturers Pipe (i.e. Different from Calibration Piece):
Studies have shown that acoustic velocity varies from one type of pipe to another (even though all are called X70 steel). The ratio of the velocity of the plastic wedge the ultrasonic probe sits on to the steel being tested determines the refracted angle the beam makes in the steel. If the steel's sound velocity changes the ratio changes so the refracted angle changes. This could mean that if a steel with higher acoustic velocity than the calibration block was to be tested (without changing the probe or instrument settings) the zones investigated would not be those calibrated for. e.g. the LCP probes' beam might now be bent up so it was directed at the Fill 1 area.

Temperature Effects

During winter work projects it had been found that numerous adjustments were being made in calibrations. Investigations indicated it was critical to maintain temperature within a narrow window (+/- 10 C°). The reason for this concern is acoustic velocity changes with temperature. For plastics the effect of temperature change is much faster than for steel. In the working range of inspection (+40°C to -40°C) the velocity changes in steel are negligible compared to the velocity changes in plastic over the same range of temperatures. These changes are more pronounced for greater refracted angles, i.e. refracted angle > 60°.

To ensure control of temperature conditions several precautions are required. Thermocouples are placed in the wedges to ensure operating temperature is within the tolerances permitted by the specification. Temperature regulation is accomplished by either using a heating pad under the calibration block or by circulating warmed couplant through the probes. The heating pad is somewhat neater in that couplant is not running all the time but unless the probes are centred on the calibration block heating may not be even. Also, when the probes are taken off the pad they begin to cool. From the start of the scan to its end the plastic may cool down below tolerance. Re-running a weld twice in a row may require the probes be re-placed onto the heating pad to warm up again prior to the second run.

Figure 29: Temperature and Velocity Effects on Refracted Angles

Figure 10-29 illustrates the effect of changes in both steel velocity and temperature of the plastic wedges on the position of the beam in the test piece. When the velocity of the steel is faster than was calculated for the plastic being used to refract the beam, the angle of refraction increases. This causes the beam to move up to a new zone. In cold weather the plastic gets harder and transmits sound faster. This means a smaller velocity difference exists between the plastic and steel. Under these conditions a beam will bend less on refraction and the zone investigated will actually be lower than the zone intended.

The beam on the Left side shows an intended beam position for the LCP area. The Right side indicates a resultant beam when steel velocity is greater than predicted (e.g. different manufacturers pipe).

Conversely the right side may show a beam intended to hit the Fill 1 zone (as in the tandem technique shown in the calibration section). Unless corrections are made to keep the probe at the temperature it was designed for the cold temperatures of winter can cool the plastic and the acoustic velocity increases. This reduces the refracted angle to something like the beam on the left thereby placing it in the LCP region instead of the intended Fill 1.