Abstract. 1. Introduction

Transcription

1 IBP1299_11 SMARTBALL: A NEW PIPELINE LEAK DETECTION SYSTEM, AND ITS SURVEY OF TWO PETROBRAS / TRANSPETRO PIPELINES FIELD TESTS. Felipe Oliveira 1, Timothy Ross 2, Adriano Trovato 3, Muthu Chandrasekaran 4, Fábio Leal 5 Abstract Significant financial and environmental consequences often result from line leakage of oil product pipelines. Product can escape into the surrounding soil as even the smallest leak can lead to rupture of the pipeline. From a health perspective, water supplies may be tainted by oil migrating into aquifers. A joint academic-industry research initiative funded by the U.S. Department of Transportation s Pipeline and Hazardous Materials Safety Administration (PHMSA) has led to the development and refinement of a free-swimming tool called SmartBall, which is capable of detecting leaks as small as liters per minute in oil product pipelines. The tool swims through the pipeline being assessed and produces results at significantly reduced cost to the end user compared to current leak detection methods. GPS synchronized GIS-based above ground loggers capture low frequency acoustic signatures and digitally log the passage of the tool through a pipeline. A tri-axial accelerometer system gives the odometric position of the ball, and has the accuracy of standard instrumented pigs; and several other types of sensors temperature, pressure, etc. are present in the ball. Petrobras and Transpetro decided, after the SmartBall was presented to them by Nortech/Pure Technologies, to test the capabilities of the tool on two pilot inspections, with simulated leaks and real operational conditions, without changing anything on the production profile (low pressure, low flow rate, multiphasic product) of the pip eline. This paper presents a briefing of the development of the SmartBall, the full capabilities of the tool and the results of the pilot inspections, which validated the capabilities of the tool in detecting small leaks, without the need for any new infrastructure, besides the common pigging facilities. 1. Introduction Early detection of leaks in hazardous materials pipelines is essential to reduce product loss and damage to the environment (Carlson 1993; Zhang 1996). Small leaks, if undetected, can result in elevated clean-up costs and have the potential to grow to more serious failures (Figure 1). As an example, on July 30, 2003, an 8 steel pipeline carry fuel between Tucson and Phoenix, AZ ruptured resulting in a spill of over 50,000 gallons of gasoline. This resulted in contamination to the soil and groundwater. Fortunately, an adjacent 68 lot housing subdivision was still under construction at the time and thus did not have any residents present. This incident not only resulted in physical/environmental damage, but also economic consequences from the extremely high fines imposed by the environmental regulations. SmartBall is a new approach that combines the sensitivity of acoustic leak detection with the 100% coverage capability of in-line inspection. The free-swimming device is spherical and smaller than the pipe diameter, thus allowing it to roll silently through the line and achieving the highest detection of small leaks. It is typically launched and retrieved using conventional pig traps; however, its size and shape allow it to be launched by other means and to negotiate obstacles that could otherwise render a pipeline unpiggable. 1 Physicist, Engineering Manager NORTECH 2 Project Engineer, Project Manager PURE TECHNOLOGIES 3 Mechanical Engineer, Director NORTECH 4 Project Engineer, Marketing Manager PURE TECHNOLOGIES 5 Engineer, Junior Engineer TRANSPETRO

2 Figure 1. Petroleum Pipeline Leak 2. Current State of leak Detection Techniques There are a variety of methods that can detect leaks in natural gas and petroleum product pipelines, ranging from manual inspection to advanced satellite based hyper-spectral imaging (Carlson, 1993). The various methods can be classified into internal and external systems. Internal systems include: 1) computational pipeline modeling; 2) pressure testing; and 3) inline inspection. External systems include: 1) optical remote sensor methods; 2) acoustic emissions; 3) fiber optic sensing; 4) liquid sensing; and 5) vapor sensing. No single system is universally accepted as the preferred method as all have strengths and weaknesses, though some are far more commonly used than others. Typically, most operators opt for a combination of Computational Pipeline Modeling (CPM) and direct observation methodologies including aerial and/or ground patrols. Permanent monitoring sensors based on acoustic or other technologies are also available; however, these methods can be costly and none can reliably detect small leaks regardless of their location in the pipeline. Most pipeline operators also employ in-line inspection programs using smart pigs that detect cracking, wall thinning, and other anomalies (Turner 1991). However, a majority of pipelines are typically only inspected at intervals of several years, and all in-line inspection technologies have limits on the minimum cross-sectional area of the defects that they can detect. Furthermore, they are unable to differentiate between a deep defect and a leak. Through-wall pinhole leaks resulting from aggressive corrosion mechanisms such as microbiologically influenced corrosion are not typically detected. Also many lines, particularly in the gathering and distribution sectors, are never pigged. The desirable attributes of a leak detection system as defined by API are as follows: Possesses high sensitivity to product release; Allows for timely detection of product release; Offers efficient field and control center support; Requires minimum software configuration and tuning; Requires minimum impact from communication outages; Accommodates complex operating conditions; Is available during transients; Is configurable to a complex pipeline network; Performs accurate imbalance calculations on flow meters; Is redundant; Possesses dynamic alarm thresholds; Possesses dynamic line pack constant; Accommodates product blending; Accounts for heat transfer; Provides the pipeline system s real time pressure profile; Accommodates slack-line and multiphase flow conditions; Accommodates all types of liquids; Identifies leak location; Identifies leak rate; Accounts for effects of drag reducing agent. 2

3 The U.S. Department of Transportation (USDOT) regulates the Transportation of Hazardous Liquids by Pipeline under federal regulations in 49 CFR Part 195. Part 195 recognizes computational pipeline monitoring (CPM) as the acceptable standard for leak detection systems on hazardous liquid pipelines, and that each CPM system must comply with American Petroleum Institute (API) Standard CP M systems employ software modeling that dynamically evaluates flow monitoring devices measuring the rate of change of pressure or the mass flow at different sections of the pipeline. If the rate of change of pressure or the mass flow at two locations in the pipe differs significantly, it could indicate a potential leak. The major advantages of the system include its ability to monitor continuously, as well as non-interference with the operation of the pipeline. Two disadvantages of the system include the inability to pinpoint the leak location, and the high rate of false alarms. These systems are also very expensive for monitoring a large network of pipes. API recognizes that detectable limits using CPM are difficult to quantify because of the unique characteristics presented by each pipeline (API 1130 Document Information, Nov. 1, 2002). The Alaska Department of Environmental Conservation has expanded on 49 CFR Part 195 and API 1130 by establishing a 1% of daily throughput leak detection standard. For example, this means that a 24 diameter pipeline transporting refined or crude oil at 2.5 m/s, would require a leak rate of more than 380 liter/min to occur before the CPM system would detect the leak at the 1% detection limit. Given the high rate of false alarms associated with CPM systems, many pipeline operators do not initiate action until an even higher detection limit is discovered. 3. Technology Background When pressurized product leaks from a pipe, it creates a distinctive acoustic signal that is transmitted through the product flowing in the pipeline. Acoustic leak detection equipment identifies the resulting sound or vibration (Klein 1993). Such systems have proven to be one of the most effective and reliable means for identifying leaks in pipelines, but fixed sensors have limited range due to the rapid attenuation of sound. A free-swimming, un-tethered acoustic leak detection device avoids this limitation by passing directly by any leak, regardless of its location along the pipeline. By operating inside the pipe, the device is not affected by ground cover or environmental conditions, and there is never any concern that sections of the line could be missed. Recognizing the value offered by acoustic technology, but also the range limitations for a fixed sensor system, a research and development program was undertaken to develop a free-swimming acoustic leak detection device. The goal of the program was to develop a technology that would provide operators with the ability to detect small leaks at an early stage and determine the location with sufficient precision to enable remedial work to be performed without additional surveying. It also had to be available at an affordable cost without the need for major changes in pipeline construction or operation. The resulting system, named SmartBall, was initially developed in 2005 for water transmission pipelines because of the magnitude of the leak problems in the water industry and the lack of alternative solutions. Because the SmartBall principle was equally applicable to oil and gas pipelines, the research and development program was extended in 2007 to modify the technology for use in the hydrocarbons industry. This led to a two year research project funded by the U.S. DOT s Pipeline and Hazardous Materials Safety Administration (PHMSA) in collaboration with Pure Technologies Limited and Arizona State University. The SmartBall device diverges from the traditional cylindrical shape of in-line equipment or pigs. The spherical shape greatly reduces the noise produced by the device as it passes along the pipe, thus permitting the sensitive acoustic sensor to operate free of external interference. The net result is a tool that is sensitive to very small acoustic events and therefore, very small leaks. An additional benefit of the spherical design is that it allows much greater flexibility in the methods of deployment and retrieval compared to a cylindrical tool, and it is able to negotiate a much wider range of bore changes, small radius bends and other obstacles that may exist within the pipe. SmartBall can run in lines without pig traps, and in a wide range of traditionally unpiggable locations. Both liquid and gas pipelines can be inspected. With the current series of tools, any pipeline size of 4 diameter or greater is accessible. The device is inserted into an operational pipeline in the same manner as a smart pig and can monitor many miles of pipeline during a single deployment. The SmartBall s position is logged internally by on-board accelerometers and the ball can also be tracked by GPS synchronized surface sensors in the same way a conventional pig. This combination of tracking systems allows the location of any leaks to be determined within ±1 meter. The SmartBall is fully sealed with no electronic components exposed to the pipeline environment. This allows us to achieve high levels of reliability in a wide range of hostile conditions, and ensures that the device is intrinsically safe for use in flammable products such as oil and gas. Figure 2 illustrates the internal components and outer shell. 3

4 Figure 2. Internal Components and Outer Shell Following a run, the data are downloaded to a computer for analysis. This process is relatively fast; subsequently the initial decision on whether a leak exists in the line can often be made before the field crew leaves the inspection site. The SmartBall log comprises of a continuous, high-fidelity acoustic record of the inspected section over a wide range of frequencies. Leaks are identified by a combination of increased acoustic power and a characteristic frequency spectrum. Subsequent logs may be overlaid to identify changes between inspections, further enhancing sensitivity. 4. Laboratory Testing and Development As part of the R&D program, SmartBall underwent a rigorous laboratory testing program to ensure operability in a hazardous and damaging environment inherent in live oil and gas pipelines. Subsequently, a series of tests were performed to ensure that the device would perform as intended, in a safe manner and be sufficiently robust to resist damage during operation. The tests performed included the following: External Pressure Test; Rolling Noise Test; Inclined Line Propulsion Test; Impact (Drop) Test; Thermal Test; Sensor Attachment Integrity Test. All six tests were successfully completed prior to field application. Details of the Inclined Line Propulsion and External Pressure Tests are presented in the following section Inclined Line Propulsion Test The SmartBall will need to be able to traverse varying inclinations involved with pipelines dependent on the topography associated within varying regions. Ensuring the tool keeps moving along the line is less certain than for conventional full-bore pipeline pigs since the SmartBall does not seal across the pipeline, rather it is pushed by the product flow with a certain degree of product bypass. Critically, it is necessary to ensure that the ball can be propelled up the steepest incline in the pipeline under the available product pressure and flow speed. The purpose of the incline test was to determine minimum flow rates required to negotiate different slopes. The data accumulated from the test were then used to validate the developed computational fluid dynamics (CFD) flow modeling software used to determine the minimum flow velocities required to move the tool up varying slopes. Prior to running the SmartBall in an operational line, the validated CFD model can be used to ensure that the line conditions are suitable for propulsion of the ball. A short section of inclined pipe was used to represent an uphill section in an operational pipeline as illustrated in Figure 3. The test pipe was made from a clear polymer to allow the movement of the ball to be observed. The test rig allowed the angle of incline to be varied. A portable compressor was used to create a variable rate of flow in the line. The flow speed was measured by means of a velometer. Test results are presented in Table 1. 4

5 Figure 3. Inclined Line test Apparatus Degree of Inclination Velocity Required m/s 0 0, , , , , , , , , ,427 Table 1. Velocity Required by Degree if Inclination The incline test was not performed using liquid because gas is the primary concern with regard to propulsion. Validation of the CFD model in gas provides confidence of the model pertaining to use in liquid lines. Additionally, there are numerous live runs that have been conducted in liquid (water) to use for validation External Pressure Test The SmartBall must be able to withstand pressures realized in a normally operating pipeline. Any breach of the ball in service would result in damage to the internal electronics and a potential safety risk in the case of an inflammable product. The pressure test was designed to confirm the SmartBall s capability to withstand pressures of 2000psi (138 bars). Though the tool is designed for pressures higher than 2000psi, due to factor of safety requirements, the quoted maximum operating pressure for SmartBall is 2000psi. The tool is not anticipated to be used in line pressures greater than 2000psi; with this consideration, a test vessel was designed and built for the purpose of pressure testing all SmartBall s that are produced. It comprised of a sealed container sized to fit the largest SmartBall under production. Pressurization was achieved by means of a standard air compressor. The internal pressure was monitored continuously throughout the test. The vessel is illustrated in Figure 4. SmartBall sizes of 4, 6, 8, and 10 diameters were pressure tested at 2,000psi for a duration of 12 hours. All sizes were able to successfully withstand the 2,000psi pressures. Figure 4. Pressure Vessel Testing 5

6 5. SmartBall Logistics and Testing on Petrobras/Transpetro Facilities On 2008, Nortech became the representative of Pure Technologies for South America. After presenting the SmartBall tool for Petrobras engineers at CENPES (Centro de Pesquisas Leopoldo Américo Miguez de Mello, Petrobras greatest R/D facility), they decided to test the capabilities of the tool on the loop test facility of CTDUT (a Pipeline Technology Center). There the capabilities of the tool would be tested on a 14, 100m long 14 pipeline, propelled by water; on these tests, several simulated leaks were created, ranging from 0.1 lpm to 0.45 lpm. Although the SmartBall was able to detect all the simulated leaks, the test signal results were not as good as expected, due to: 1) Low flow velocity around 0.15 m/s; the minimum recommended is 0.3 m/s - ; 2) The great noise captured by the SmartBall, that came from the water truck that was supplying the loop and a construction near on a real simulation, there are almost no external noise, as the pipelines are usually berried or underwater -. One of the simulation leaks can be seen on Figure 5. Figure 5. Simulated 0.2 lpm Leak on CTDUT Facility Regardless the difficulties found on the tests, the end report was satisfactory, and the capabilities of the tool on detecting the simulated leaks were proven. Now, a simulation on a real pipeline, with the day-to-day basis operation condition, was needed to assure Petrobras of the applicability of the SmartBall on their leak detection inspections. In 2010, two pipelines were selected for these tests; one was an 18, 100 km long pipeline in Transpetro an 18 ; the second one, a 12, 20.9 km long in Petrobras. The tests were conducted by the Nortech team, supervisioned by a Pure Technologies engineer, and supported by the local Transpetro/Petrobras Teams, who provided all the SMS procedures, all operational supporting regarding the pipelines and the handling of the simulations mechanisms. 6. The Transpetro Test The Transpetro group of Petrobras retained the services of Nortech/Pure Technologies to perform a demonstration SmartBall survey of the selected 18 pipeline. As the purpose of the SmartBall survey was to demonstrate the leak detection capability of the SmartBall tool by means of the creation of multiple simulated leaks along the length of the pipeline. During the survey the SmartBall device was inserted into the pipeline through the pig launcher and released into the flow of the pipeline. It traversed the pipeline with the flow, and in so doing acquired acoustic, positional, temperature, pressure and magnetic data. This data was evaluated to identify the acoustic activity associated with leakage. During the evaluation of the data, Pure detected 3 acoustic anomalies that resembled leaks in the main at the time of inspection. Two of which were known simulated leak points, with the remaining acoustic anomaly to be an leak created by the Transpetro team, without the Nortech/Pure crew knowing if a leak was or not created, and, if so, what was the leak rate the tool was able to identify and quantify this leak, and this information was confirmed with the Transpetro team on the presentation of the final report -. This was made with the purpose of evaluation the measurement capabilities of the tool. The survey and the results are summarized in Table 2. 6

7 Total Length of Pipe Surveyed: m Pipe Material: Steel Diameter of the Pipe: 18 Nº of Simulated Leaks: 2 Nº of Unknown Leaks Detected: 1 Duration of Survey: 2d, 10 hrs. and 42 min Average SmartBall Velocity: 0.5m/s Table 2. Transpetro Pipeline Details / Survey Results 6.1 The Preparing the test site. As the pipeline had a bigger internal diameter then the major SmartBall, a technique baptized by overshell was adopted. A 16 sphere special foam was prepared for tis inspection; that is a system composed of a great mesh system, made of a low liquid absorption rate material, making it perfect for the SmartBall, as it doesn t muffles the sound of the pipeline. With this foam, the area occupied by the SmartBall inside the pipeline greatly increased, allowing it to travel at ~95% of the flow speed and trespassing the great slopes of this pipeline, some of those with inclinations of 22 degrees. A photo of the SmartBall and its foam, at the extraction point, can be found at Figure 6. Figure 6. SmartBall Extraction The leaks were simulated on vents near the existing valves of the pipeline. A system was created for the leak rate could be started/stopped and measured. Two sequential needle valves were installed after the vent. On the first one, we selected the volume of leak desired. Once it was selected, we left the first valve on that state and the second one was used as a start/stop device. For environmental reasons, and for losing the minimum amount of product as possible, the leaks were activated just minutes before the passage of the SmartBall. The Nortech / Pure crew placed seven (7) SmartBall Receivers along the line to track the progress of the ball as it traversed the pipeline, and check if the time of the actual tool passing corroborated with the scheduled time, calculated previously, based on the flow velocity, that was near constant. These receivers are capable of sensing the proximity of the SmartBall for as long as 500m. The SmartBall Receivers were placed on the pipeline at the locations indicated in Table 3. The inspection was complete in 2 days, 10 hours and 42 minutes. The time at which the SmartBall device passed the SmartBall Receiver locations is also summarized in Table 3. Distance from Start Elapsed Time Since Launch (hh:mm:ss) Location Description SBR# 0m 00:00:27 Insertion SBR1 19,923m 41,726m 11:23:43 23:31:47 Valve 29 Valve 28 SBR2 SBR3 56,183m 70,201m 32:28:39 41:14:35 Valve 27 Valve 26 SBR4 SBR5 88,770m 100,566m 51:25:52 58:42:50 Valve 25 Extraction SBR6 SBR7 Table 3. SmartBall Receiver Locations 6.2 Summary of Leaks 7

8 The acoustic data recorded by the SmartBall has been analyzed and cross-referenced with the position data from the SBR to determine a location for each anomaly. A summary of the leaks and pockets of trapped gas identified during the SmartBall survey is provided below on Table 4. Distance from Start Time Since Launch Description 0m 00:00:27 Insertion 19,923m 11:23:43 Valve 29 19,925m 11:23:46 Leak (0.4 LPM) 41,726m 23:31:47 Valve 28 41,729m 23:31:54 Leak (0.6 LPM) 56,183m 32:28:39 Valve 27 70,201m 41:14:35 Valve 26 70,204m 41:14:43 Leak (0.14 LPM) 88,770m 51:25:52 Valve ,566m 58:42:50 Extraction Table 4. Transpetro Inspection General Points of Interest 6.3 Leak Analysis Details As described above, 3 leaks were detected during the survey. The acoustic sensor inside the SmartBall acquires the leak indication power, and, with that information in hands, an comparison tool inside the analysis software compare the leak indication power of a detected leak with that of a known leak rate to assist in identifying and quantifying the leak rate of any leak. Known leak rates and corresponding leak indication power are developed by holding the tool in the launch trap at the start of the survey run. The leak indication power is the single most important indicator of a leak s presence and size. To validate the leak detection capability of the SmartBall, several projects were undertaken where leaks were simulated at predetermined locations along the pipeline. Leak information is illustrated in Figure 7 that reports the simulated leak 2, of 0.6 liters per minute. Table 5 provides the full information of the leak. A whole myriad of other information, such as full data from a magnetometer, pressure sensor, temperature sensor and velocity profile, are placed on the final reports. Figure 7.1. Leak Indication Power of Leak Figure 7.2. Frequency Spectrum of Leak Distance from Insertion Point: 41,729.2 m Distance to Nearest Sensor: 3.2 m after SBR 2 Time Since Start of Rolling: 23:31:54 Time Since Tool Activated: 27:42:06 Time of Tool Pass (GMT-3:00): 06:02:35 PM November 2 nd, 2010 Approximate Latitude: Approximate Longitude: Leak Indication Power: db 8

9 Estimated Size of Leak (small/med/large): Small (0.6 LPM) Table 5. Full information of the Simulate Leak # Additional Information The SmartBall also evaluates and presents a significant volume of valuable data to pipeline owners; these data is acquired full time by the SmartBall, and it helps the confection of the report and identification of leaks. Such data can be seen on: Figure 8 ( full velocity profile), Figure 9 ( full pressure profile), and Figure 10 ( full temperature profile). Magnetometer data can also be plotted, but due to the small signal/noise ratio, only on specific points. Figure 8. Transpetro Inspection Velocity Profile Figure 9. Transpetro Inspection Pressure Profile Figure 10. Transpetro Inspection Temperature Profile 7. The Petrobras Test On Petrobras, the same testing protocol was adopted for the simulations. Two leaks were simulated, and the several obstacles were presented to this inspection. The biggest one was that the pipeline did not have a constant flow. Regarding that, the velocity of the SmartBall varied along the pipeline, making tracking a lot more difficult to handle. To counter that, the Petrobras team made constantly made the outgoing product volume. With that information, the Nortech/Pure team updated the schedule table, so we could be at the monitoring points with precision. Another setback that this pipeline was the first multiphasic product inspection of the SmartBall, and the tool proved to have the same accuracy on these conditions, even with the high percentage of water gas (we estimate a volume of approximately 4% water, 20% natural gas and 76% light/medium crude oil and low levels of wax). As the sound speed is different in different fluids / gas, the exactly behavior of the signal was unknown; but the signal was clearly detectable, with just a 15% loss on the amplitude. 9

10 As this was a 12 pipeline, there was no need for an overshell. A 10 SmartBall coated with polyurethane was used on this survey. The survey and the results are summarized in Table 6. Total Length of Pipe Surveyed: m Pipe Material: Steel Diameter of the Pipe: 12 Nº of Simulated Leaks: 2 Duration of Survey: 12 hrs. and 52 min. Average SmartBall Velocity: 0.5m/s Table 6. Petrobras Pipeline Details / Survey Results 7.1 Summary of Leaks The acoustic data recorded by the SmartBall has been analyzed and cross-referenced with the position data from the SBR to determine a location for each anomaly. A summary of the leaks and pockets of trapped gas identified during the SmartBall survey is provided below on Table 7. Distance from Start Time Since Launch (hh:mm:ss) Time of Tool Pass (hh:mm:ss) Description 0m 00:00:00 10:36:26 Insertion 1,500m 00:52:53 11:29:19 SBR 2 1,502m 00:52:56 11:29:20 Leak (Small) 15,897m 09:39:07 20:15:33 Leak (Small) 15,900m 09:39:08 20:15:34 SBR 3 20,900m 12:52:41 23:29:13 Extraction Table 7. Petrobras Inspection General Points of Interest 8. Conclusions With base on this technology, was proven to the Petrobras/Transpetro crew involved on these operations that SmartBall provides a quick, reliable and efficient leak detection tool, with quick response time the information about the simulated leak points was known in within 3 hours - and solid information provided on a fast report. The ability of the tool to go through unpiggable pipelines is in very high demand, and its accuracy has proven to detect leaks of minimum size, preventing environmental / financial damages. Further enhancements include the development of a simplified arrangement for data download that avoids the need to open the ball. This will have the additional advantage of reducing the risk of damage to internal components and help to ensure that the seals remain intact. The customer version of the analysis software is in the process of being developed and modified to increase the level of automation and reduce the need for technical expertise from the user. 9. References Carlson B.N. (1993). Selection and use of pipeli ne leak detection methods for liability management into the 21st century, Pipeline Infrastructure II, Proceedings of the International Conference of the American Society of Civil Engineers, ASCE. Klein W.R. (1993). Acoustic Leak detection, American Society of Mechanical Engineers (ASME), Petroleum Division, Vol. 55, Houston, Texas, pp Turner N.C. (1991). Hardware and software techniques for pipeline integrity and leak detection monitoring, Proceedings of Offshore Europe 91, Aberdeen, Scotland. Zhang, J. (1996). Designing a Cost Effective and Reliable Pipeline Leak Detection System, Pipeline Reliability Conference, November 19-22, Houston, Texas. 10