Good defect detectability depends on an outside source of energy producing a sufficient level of strain across a defect to force some physical property change or movement in its vicinity. The outside source of energy can take several forms, such as; pressure changes, mechanical tension/compression changes, or as in this case, thermal excursions.

These thermal excursions, either up or down, produce temperature gradients through the thickness of the material, which, in combination with the strain due to the existing internal pressure, produce circumferential and radial strain in the piping system. This strain is the driving force used to produce a physical change around a defect.

The physical change that takes place around a defect can take several forms, such as crack extension or growth, rubbing of crack face surfaces, breaking of corrosion product in a crack, or yielding of the material around a crack tip. Each of these mechanisms will produce transient elastic stress waves (AE) that move away from the defect at the speed of sound.

Movements of these elastic waves are similar to those produced by dropping a one pound steel ball into a still pool of water, that is, as the ball enters the water it will cause a vertical surface displacement that translates into a surface wave front moving away from the epicenter in all directions.

Stress waves produced in steel can be detected several feet from a defect by sensors that are used to convert the surface motion of the elastic wave into an electronic signal. The signals detected by the sensors are recorded by the data acquisition system in a digital format data file that contains general information about the elastic wave, such as its amplitude, duration, time-of-arrival at each sensor, and relative energy or Measured Area of the Rectified Signal Envelope (MARSE). MARSE can be viewed as relative signal magnitude.

When multiple sensors are mounted on the item, such as this steam line, the approximate location of a defect can be calculated by using the difference in times-of-arrival (Delta-Time) compared to the physical distance between sensors. This technique is similar to that used to locate the approximate epicenter of an earthquake.

Unfortunately, most of the acoustic emission activity recorded during a typical Hot Reheat steam line test is produced by noise sources other than defect growth. This extraneous noise comes from steam flow and mechanical movement of the piping.

Recognition and identification of these extraneous noise sources is the key to effective analysis.

Using the above mentioned "pool of water" analogy, steam flow would be similar to droplets of steady rain falling into the pool, producing many thousands of small wave fronts (Background Noise) traveling in all directions. Hanger noise, on the other hand, would be like a ten pound steel ball splashing into the pool, producing a very large wave front that will travel for long distances in all directions.

Steam flow noise, like the rain drops, shows up as thousands of low amplitude AE signals striking the sensors randomly throughout the duration of the test. The steam flow problem can be combated by using higher frequency band pass filtering with the sensors and by adjusting the detection threshold of the data acquisition system slightly above the level of the steam flow noise.

Hanger noise, like the ten pound ball, shows up as high amplitude, large magnitude AE signals that can travel the entire length of the pipe line. Since there is no effective way to keep from recording this emission, the best that can be done is to calculate the approximate location of the high amplitude emission and compare this to existing hanger locations and external penetrations in the pipe line.

Currently, standard AE data analysis methods used on pressure vessels depends on analysis of independent channel information to determine the presence of a defect. This method, which considers the data acquired by each channel as if that channel existed independently in space, works very well when background noise is not a problem and the loading method can be highly controlled. The results obtained from previous AE tests performed on Hot Reheat steam lines have not been good when the independent channel analysis method has been used.

The independent channel analysis method depends on a defect showing increasing emission magnitude in response to the external loading imposed on the item under test. The problem encountered when using this technique on a steam line is that the steam flow noise can increase, decrease, or remain steady while the pipe is changing temperature. If the steam flow emission is included in the analysis, the results can become unreliable.

The analysis procedures that we have developed improve the reliability of the test, this method uses a combination of spatial event filtering and event characteristic correlation.

Spatial filtering, also known as source location, looks at emission that is large enough in magnitude to strike at least two adjacent sensors on the pipe, in a time frame dictated by the speed of sound in the material and the distance between the sensors. This method eliminates a large part of the random emission generated by the steam flow because this steam flow emission usually does not strike two adjacent sensors in the required time frame. Even with the use of spatial filtering, if enough random emission strikes the sensors, eventually, the system will find some events that show initial locations. Note: Sound travels about 12000 feet per second in this pipe.

Event characteristic correlation is a method where each locatable event is plotted on a scatter plot showing the duration of the event compared to the amplitude of the event. This method is used to discriminate between emission from steam flow, mechanical noise and actual defect growth.

Generally speaking, emission from the typical sources detected during a steamline test show the following characteristics:

1. Steam flow emission will be very low in amplitude and short in duration, showing just above the detection threshold settings.
2. Defect emission will be somewhat higher in amplitude and longer in duration.
3. Mechanical emission will be very high in amplitude and very long in duration.

An acoustic emission "Event" would be defined as a physical change in the item under test that would give rise to a transient stress wave. This stress wave would travel away from the "point of origin", along the pipe at the speed of sound for the specific material. A "Hit" would be the signal detected by each sensor as the stress wave, traveling along the pipe, arrived at each sensor position. In the example below, excluding possible reflections, the acoustic emission data acquisition system would record six (6) "Hits", one (1) on each channel, which were generated by one (1) physical "Event" located at channel one (1).

Event source location uses the timing information gathered from the several sensors scattered down the length of the piping systems to point to the origin of the physical event which generated the acoustic emission activity detected during the test.

As in the figure below, each sensor would be connected to an individual channel in the acoustic emission data acquisition system. Each channel will record the "Time of Arrival" of each acoustic emission hit detected by the sensors. The approximate distance from the "First Hit" sensor (in this case channel #2) to the source of emission is given by the formula below.

Some confusion can occur when the data acquisition system records emission from sources that are located outside the coverage of the sensor array. Ideally, this emission would simply "Pile up" at the edge of the array, but this is usually not the case. The location events generated by external sources will show some positions slightly within the edges of the array, as shown below.

The AE data acquisition system records a HIT every time the signal level from a sensor rises above the individual channels detection threshold level. When the hit is recorded, several measurements are taken relating to the size and shape of the incoming signal. Some of the measurements are:

1. Threshold trigger time in 250 nano-second increments.
2. Peak Signal Amplitude in decibels.
3. Signal Duration in micro-seconds.
4. Risetime in micro-seconds.
5. Counts or Number of threshold crossings.
6. Energy Counts or MARSE.

The figures below are schematic representations of some of these AE hit parameters.

The American Society of Mechanical Engineers has introduced the term MARSE in Section V, Article 12 of the Boiler and Pressure Vessel Code (8). MARSE is defined as the Measured Area of the Rectified Signal Envelope. The term MARSE should not be confused with the term MONPAC, which has nothing to do with this type of test. This AE signal parameter is also referred to as relative energy and energy counts. Note that MARSE is the area under the voltage signal and not the decibel signal.