Figures 1-4
Figures 5-7
Figures 8-10

Evaluation of pipe wall integrity while the pipe remains in service is an important preventive maintenance issue for many industries, but it is especially critical for applications in power generation and petrochemical facilities. Pipe failure at these plants can have both a significant cost and environmental impact and safety implications. Large gas and oil transmission pipes are traditionally inspected from the inside by pigs which travel with the flow of the product, but for most piping this method is not practical. The majority of pipe must be inspected externally with methods that monitor the internal and external condition of the wall. The main methods in use today for external evaluation of pipe wall are film radiography and ultrasound. These methods are good for sampling inspection, but industry needs cost effective alternatives which can provide more complete information about the piping system.

In the United States during the past few years alone, several refineries and chemical plants have had explosions, fires and loss of lives due to spillage of flammable and other hazardous fluids from the pipelines. Even power generation plants have had their share of problems. In February 1995, in the Pleasant Prairie power plant of Wisconsin Electric and Power Company (WEPCO), a super heated high pressure water feed line to a boiler had burst and killed two employees on the spot. This not only caused loss of human life, but also caused a plant shutdown for several months. In 1996, a similar accident happened in a small co-generation plant in Green Bay, Wisconsin with loss of several lives.

---

Today, this preventable cost may be more than $100 billion plus saved lives and pollution prevention.

---

 

It is conservatively estimated (Dean, 1989) that the total cost of corrosion in the United States was $70 billion in 1975 and $143 billion in 1982, out of which the author estimated that $10 billion in 1975 and $21 billion in 1982 was preventable with better detection and prevention technology. Today, this preventable cost may be more than $100 billion plus saved lives and pollution prevention.

A recent development in gamma ray detector technology has made an external, online, real time radiography system practical. This system (ThruVU) uses a linear array of solid state gamma ray detectors to measure wall thickness of bare or insulated piping systems. The system uses a low intensity Ir-192 gamma ray source and a solid state detector array mounted on a computer controlled robotic crawler. The Ir-192 gamma ray source is located on one side of the piping components and the solid state detector array is on the other side. The individual sensors of the detector array measure the intensity of the gamma rays after passing through the inside product, the walls and the insulation of the piping component under measurement. The output signal of the detector array is connected to a high speed laptop personal computer system through a long cable. The computer system collects and analyzes the data from the detector array in real time as the robotic crawler travels over the piping system. The system measures, computes, and provides the actual double wall thickness values instantaneously. This new technique continuously measures the wall thickness of the pipeline as the crawler travels over the pipe at a rate of 609-1,200 mm (2-4 ft) per minute.

The first demonstration of this technology was made at WEPCO Pleasant Prairie plant in May 1995 under a research contract from Electric Power Research Institute (EPRI). Since then, this system has been used at various refineries and oil and gas production facilities of Amoco, Exxon, Mobil, Chevron, Valero, Du Pont, Shell, and others in the US and Canada. Recently, a successful demonstration of this system was also made at the Japan Energy Refinery in Chita City, Japan. This system has proven that it can detect small corrosion points even in very large in-service pipes. The system has been successfully used on several large 600 mm (24 in.) diameter in-service crude oil pipelines. At present a system is in use at the oil production facility in Prudhoe Bay, Alaska.

Within the past 18 months, under a contract from the EPRI/NDE Center, a single wall tangential scanning method has been developed (tangential or shadow radiography) using the similar detector array system. Initial results with this tangential radiography system are very encouraging. The first measurements were made on a section of 100 mm (4 in.) insulated pipe removed from service at WEPCO. This pipe was down stream from an elbow in a high pressure water line and had lost a significant amount of its wall on one side. Figure 1 shows a picture of the laboratory set up used to make the tangential measurements on this pipe. The tangential scan in Figure 2 shows both the double wall and single wall features of this pipe. A digital profile in Figure 3 provides double wall measurements in the center section and single wall thickness measurements on both edges of the pipe. With some additional robotic design effort, such a tangential measurement system can also be incorporated in the present double wall inspection crawler. The resultant robot will crawl over the pipeline providing routine double wall measurements. At selected locations at the option of the operator, single wall tangent measurements can be made by stopping the crawler and activating a second motion of the detector array.

 

Basic Concept
The detection system uses a solid state gamma ray detector array and an Ir-192 radioisotope gamma ray source mounted on a remotely controlled robotic crawler to inspect large lengths of pipeline. At the crawler itself, the detector array enclosure incorporates electronics to amplify, integrate, and multiplex the detector signals so that the output of all detector channels is sent on a single pair of conductors. The multiplexed signal from the detector array is transmitted to a laptop PC which also includes a 12 (or 16) bit analog-to-digital converter (ADC) board. The detector array is connected to the laptop computer via several hundred feet of flexible multiconductor cable. This cable carries the detector signals to the ADC board at the computer and provides power and logic signals to the detector array.

Figure 4 shows a concept of the double wall system with the source on one side of the pipe and the detector array on the other side. Figure 5 shows a block diagram of the systems with various subsystems.

The length (number of detectors in the detector array) of the detector array is chosen to cover the cross section (diameter) of the pipeline to be inspected. A 64-channel detector array examines about a 200 mm (8 in.) wide cross section simultaneously. While a 128-channel detector array lets us examine a 400 mm (16 in.) cross section at a time. The detector array collects data from a narrow strip (about 0.5 mm [0.020 in.] wide) of the pipeline at one time. With a motion along the length of the pipeline, the detector array continuously measures the wall thickness of adjacent narrow circumferential strips of the pipeline. Thus, data collected in strips is laid side by side to form a raster scan image of the pipe. As the crawler crawls over the pipeline, a two dimensional data set of the wall thickness is generated that covers the entire pipeline at a rate of about 609-1,200 mm (2-4 ft) per minute.

Figure 6 shows a picture of an older crawler system during actual measurements (in 1995) on a 600 mm (24 in.) pipe-line in the tank field at Amoco's Whiting refinery. Figure 7 shows a thickness image of a section of this crude oil pipeline indicating significant corrosion areas caused by sulfite reduced bacteria. Figure 8 shows an image of a 600 mm (24 in.) in-service insulated pipe showing significant amount of solid deposits inside the pipe. Figure 9 shows a comparison of double wall thickness measured by the real time radiography system and the standard UT measurements on an in-service 100 mm (4 in.) diameter insulated pipe at WEPCO during 1995.

The operation of the system is quite simple. After the crawler has been placed over the pipe under inspection, the Ir-192 source is remotely inserted inside the tungsten collimator at the crawler using a standard iridium radioisotope camera exposure system. The gamma rays from the Ir-192 source penetrate through the pipe-line to reach the detector array which is on the other side of the pipe. Each sensor of the detector array generates an electrical signal proportional to the intensity of its gamma ray excitation. The detector signals are digitized by the analog-to-digital converter at the laptop PC. The laptop computer collects and analyzes the data received from the detector array.

In real time, the computer system processes the signal from each detector channel based on its own calibration. The data for the calibration of the detector array is collected prior to actual measurements of the pipelines. During pipeline scanning in real time, the computer generates a two dimensional digital image of the actual wall thickness of the pipeline and highlights questionable areas. Thus collected data and images provide a 2D quantitative view of the pipeline showing actual wall thickness at every location. This data set shows all features of pipeline including corrosion/erosion on the outside and inside walls. The intensity of the image provides the quantitative wall thickness and the coordinates of the image feature reveals its location. The inspection data (digital image) is stored in the computer for further analysis, comparison or later retrieval. Since the data is already stored in the computer, other advanced analysis and display tools (edge enhancement, mask subtraction, etc.) can be utilized to enhance and improve the data analysis and display.

 

Single Wall Measurements
Generally, this system is used to measure double wall thickness of the pipeline in service. Single wall (or tangential) measurements can be made by aligning the detector array parallel to the pipe. For tangential scan, the detector array is traversed from one side of the pipe to the other side. This new data collection method provides the usual double wall measurements in the center section but extends the measurement across the full pipe width to the outside wall tangents. This provides the single wall thickness profile of each side wall. The width of each side wall profile provides a measurement of its side wall thickness. Figure 1 shows an experimental setup used to collect single wall data. Figure 2 shows a display of the data from the side wall measurement technique. Figure 3 shows the line profile of this data where the wall thickness of each side of the pipe can be easily measured by counting the number of pixels from the outside edge to the highest point of the thickness profile. The highest point represents the longest path length in the pipe and thus the inside wall of the pipe. The system automatically calculates the thickness based on the pixel size calibration.

 

Basic Detector Technology
The heart of this inspection system is the system's solid state detector array. A period of several years has ensured development of a very sensitive solid state detector system to accurately measure gamma ray intensity. The detector system uses the basic technology developed for medical computed tomography scanner detector systems. The individual detector channels use a thick piece of high Z single crystal scintillator to detect gamma rays. The depth of the scintillator crystal is such that it detects upwards of 50 percent of the gamma rays emitted from the Ir-192 source. An extremely low noise photodiode is used to measure the small optical signal generated by the scintillator crystal. The small current signal generated in the photodiode is amplified by a high quality amplifier circuit. This amplified output signal is proportional to the gamma ray intensity at the surface of the detector.

In the crawler systems pictured in Figure 10, the detector array is mounted at the bottom side of the crawlers. The inspection can actually be set to any of four circumferential positions, 12 o'clock, 3 o'clock, 6 o'clock or 9 o'clock.

The linear detector array is fabricated using 64 individual channels of scintillator - photodiode detector sensors on one single printed circuit board (PC board). On this 64-channel PC board, the output of the 64 individual detector amplifiers is multiplexed by 64:1 by a series of multiplexers. The dimensions of the individual detector sensors are 3 mm (0.13 in.) ´ 6 mm (0.24 in.). Thus the 64 detector array forms a 210 mm (8.32 in.) long ´ 6 mm (0.24 in.) wide active detection region. In addition to the 64 detector sensors, this 64-channel PC board includes all electronic logic to trigger multiplexers, integrators and the analog-to-digital conversion system. Several of such 64-channel PC boards can be butted together to form 128, 192, or even 256-channel detector arrays.

On a single pair of wires, the multiplexed analog signal from the entire detector array is transmitted to a high speed 12 or 16-bit analog-to-digital converter board located inside the laptop computer. The PC system also receives the encoder pulses from the robotic crawler to keep track of the location of the crawler. The laptop computer collects the digital data from the detector array via the ADC board. The PC also analyzes, in real time, the detector data as it is received from the detector array to compute the actual wall thickness of the pipe.

 

Robotic Crawler System
An automated computer controlled robotic crawler system has been designed to crawl over any size horizontal pipelines. The new crawler uses two individual drive tractors for traversing the crawler over the pipeline. One drive tractor is mounted on each sides of the pipe to form the crawler assembly. Each drive tractor consists of two 100 mm (4 in.) diameter soft rubber wheels. Each rubber wheel is driven by an individual low voltage, low speed DC gear motor. Thus four wheels of the crawler are driven by four individual motors, making a four wheel drive system. This four wheel driven crawler overcomes various imperfections in the pipe and insulation surface.

The entire crawler assembly is assembled by using various size aluminum extrusions. The length of various aluminum extrusions are selected so that the rubber wheels are tangent to the surface of the pipeline under measurement. The crawler assembly can be field adjusted to conform to any size pipe from 200 mm (8 in.) to 910 mm (36 in.) diameter by just adjusting or exchanging a few aluminum extrusions. Figure 10 shows the picture of the crawler on a 200 mm (8 in.) outside diameter (with the insulation) pipe.

For precise location of features, the crawler assembly also includes a position encoder and an electronic inclinometer. The position encoder is driven by one of the drive wheels. The encoder is used to provide positional pulses to the computer and the detector system. In fact, the detector system is triggered by the encoder pulses. The encoder generates 200 pulses per one inch of travel over the pipeline. These encoder pulses are sent to the computer system and used for triggering the detector readout. Typically, the encoder pulses are divided by a factor of 4 by the computer system before they are used to generate trigger pulses. Thus, the detector system is typically triggered 50 times for each 25 mm (1 in.) of travel (0.5 mm [0.02 in.] resolution) over the pipe.

The inclinometer signal is used to keep the crawler system level. The error signal from the inclinometer is used to control the center of gravity of the crawler system to keep it level with respect to gravity. The position of a movable weight is adjusted by the error signal from the inclinometer to control the center of gravity of the crawler. The weight is mounted on a motion stage controlled by the inclinometer signal. When the crawler assembly shows a tendency to tilt in one direction, the weight is moved in the opposite direction to overcome this imbalance.

 

Other Components
In addition to the detector array, the crawler and the computer system with ADC board, the system uses an interface electronics module with power supplies and two long cables. One of these cables connects the detector array to the interface electronics module. The other cable connects the crawler system (motors, encoder and inclinometer) to the interface electronics module.

 

Data Collection Software
The data collection software routines are in modular form. The entire data collection routine collects the offset and amplification gain readings for each sensor of the entire detector array before actual scanning. The offset and gain readings are used to correct the data collected during the actual scans.

The scan data is collected from the pipe-line as the crawler traverses over the pipe. As described before, the pulses from the encoder at the crawler are used to trigger the detector array. As the detector array receives a trigger pulse from the computer, it sends a data line (of 64, 128, 192 or 256 data points from the entire detector array) to the ADC board. The ADC board digitizes the detector signals and send them to the computer.

Online, the computer subtracts the respective offsets and corrects for detector amplifier gain variation between the sensors. The data is also converted to the thickness of the pipe wall for each reading. This wall thickness measurement data is computed in real time and it is displayed on the computer screen as well as stored in the computer hard drive.

During the actual data collection, the data on the computer display looks similar to the scrolling images on a airport baggage scanner system. The data stored on the hard drive can be used for future viewing and analysis or post image processing,

 

Display and Analysis Software
Storage of the image data in a digital format allows for further computer display and analysis. This display and analysis software provides many options.

Through two sliding bars, the operator can select the starting point and length of region to display and analyze. For example, the operator can select a region say starting at 2,000 mm (80 in.) and 1,000 mm (40 in.) long. The software will immediately load and display the pipe data from 2,000 mm (80 in.) to 3,000 mm (120 in.).

Through two sliding bars, the operator can select the midpoint and the range of the thickness to be displayed within the contrast range (gray level range) of the image. Then the thickness values within the selected range use the entire gray level range of the display. The values outside the thickness range are displayed in pure white or pure black.

The operator has two cursors at his or her disposal. It can move these cursors independently to select a region of interest between these two cursors. The software provides statistical information (like maximum, minimum, mean thickness, and standard deviation) within this region of interest.

When requested by the operator, the software draws the X profile (along the axis of the pipe) and the Y profile (along the diameter of the pipe) of the thickness data along the selected cursor location on a separate graph.

The operator can select a wall loss threshold. If any region within the display meets a criteria of greater wall loss than the selected threshold values, software detects and highlights these regions.

The data files can be exported to spreadsheet or ASCII files for further review and analysis.The software sends the selected image on the computer screen to the system's default printer for printing.

 

Summary
This approach to pipe inspection can provide close to 100 percent inspection of the pipe wall integrity at high speed. Scanning rates of over 107 m (350 ft) per day have been achieved. It determines the wall thickness and highlights areas that are potential problems. The data can be analyzed online during the inspection process as well as later without loss of spatial or contrast resolution. In general the initial evaluation is on a macro scale with automatic machine vision analysis highlighting areas of concern. Later, those areas can be examined in micro detail. In most cases the original data set has sufficient resolution to document the test results. Evaluation of pipe for potential failure areas can normally be achieved with the double wall technique. The areas of concern are marked and appropriate maintenance is implemented. Single wall evaluation is also available for critical areas and applications that are used to forecast pipe life. To accurately forecast the wall strength and determine its remaining life, sometimes single wall evaluations are necessary. For example, EPRI's Checkworks software uses single wall thickness data with other parameters such as pipe size and operating conditions to calculate the remaining life.

 

Reference
Dean, S. W., "Assuring Plant Reliability Through Optimum Materials Selection", Chemical Engineering Progress, June 1989, pp 36-41.

 

* Omega International Technology, Inc., 460 Wegner Road, Lakemoor, IL 60050-8653; (815)-344-5455; fax (815)-344-3336; e-mail omegait@ mc.net.

 

Copyright © 1997 by the American Society for Nondestructive Testing, Inc. All rights reserved.

[ Materials Evaluation ]