| NDT.net - September 2002, Vol. 7 No.09 |
During the last 20 years PETROBRAS has been attempting to use Acoustic Emission (AE) as an inspection tool. In this period the AE concept has changed from a revolutionary method to a way of finding areas to make a complete inspection.
PETROBRAS has a lot of pressure vessels inspected by AE and with other NDTs techniques to establish their relationship.
The R&D Center has conducted destructive hydrostatic tests and helped PETROBRAS Refineries to develop a new procedure to analyze AE data.
This article shows the R&D projects undertaken during these last 20 years and the new approach proposed by PETROBRAS - R&D Center to accept the use of Acoustic Emission.
"Acoustic emission is a naturally occurring phenomenon which man has observed from early times." This sentence written by Thomas F. Drouillard was the reason to write this article. Some engineers and researchers in Brazil have been studying acoustic emission (AE) since 1979 and PETROBRAS/R&D Center has been a great partner in this process. This article tries to summarize the projects undertaken and the new research in progress.
The first report made by PETROBRAS researches dates back to 1979, by Allevato, in a carefully written paper, described a short history of AE since the "tin cry", passing by Joseph Kaiser (Germany - 1950), Schofield and Tatro (USA - middle 50's) and other scientists in the 60's. Allevato report applications in material research, structural integrity of pressure vessels and reactors. In this period the system described and known to PETROBRAS was the ACOUST from Exxon Nuclear, transducers used were from Dunegan Corporation with pre-amplifier Trodyne. Signal analyses were made in oscilloscope.
Allevato reviewed the basic principles, instrumentation and general application of the technique of non-destructive tests by AE ("spontaneous elastic waves produced by materials under mechanical stress") in 1980. The emphasis was on the evaluation of the structural integrity of pressure vessels and reactors in general. Examples were given of the use of this inspection technique in the petroleum industry, with graphs, diagrams and photographs. The application field of the technique was expanding with the development of improved transducers, allowing the study of the mechanical performance of materials at high temperatures. Theoretical studies in stress oriented hydrogen induced cracking, applied to structures were in initial steps and could be significantly extend to in-service complex structures, in a great variety of environmental conditions.
In 1980, Allevato and Ramos wrote an article to discuss applications of acoustic emissions in corrosion tests. In corrosion studies no stress was needed, because the source was electrochemical reactions that produce gases. Graphics were shown with the relationship between corrosion and AE rate, cumulative counts versus collected hydrogen, galvanic effect in AE and the inhibiting effect of Na2CrO4 toward decreasing of AE activity. In 1980, AE was not fully developed to study corrosion phenomena, but it was possible to foresee that it could be used for studies in corrosion associated with mechanical effects, such as stress-oriented hydrogen induced cracking.
In 1982, Allevato and Carneval showed the basic principles, instruments and general applications of a non-destructive method by AE, emphasizing the structural integrity of pressure vessels and reactors. At that time, Allevato wrote that the AE field was increasing with sensor developments that allowed a study of the performance of materials under high temperatures and monitoring and detection of fatigue cracks in offshore structures.
In 1983, Allevato witnessed what could have been the first test of AE made in Brazil. This test was performed in Oxiteno do Nordeste in Bahia. It was conducted by DNV-ITL, and the system used was a D/E 1032 with 32 channels.
Mourão, Siqueira, and Allevato reported in 1984 the first research project in PETROBRAS for off-shore rig structural monitoring using a lot of techniques including AE.
During the year 1984, inspections were performed using AE in two PETROBRAS refineries. This report was not found in R&D Center Library, but in PETROSIN (software that control documents). In it, there appears a message informing that this test was carried out in sphere TQ 16J in REFAP and sphere EF-4418 in REPAR. Said report was written by Claudio Allevato, Gilberto Borges and Marcos Mattos.
In 1986, Feres and Carvalho published an article about AE inspection performed on the K-3401 A and B reactor located in the Cubatão facility (RPBC). This paper was divided into two sections. The first section was the result analysis for crack detection and location. The second section was destined to find out which phase of the operation cycle brought about the possibility of creating severe conditions to start and propagate existing defects.
In 1986, Camerini wrote that the concern of the industries with operational costs and values associated with equipment inspection had played an important role in this context. Acoustic emission has been developed to reduce inspection costs and assess the integrity of equipment or part of it. These things would be possible by dynamic evaluation of stress wave propagation into the material, fast exam and in-service inspection. CTOD tests were made and pressure vessels were submitted to hydrostatic test. The pressure vessel contained artificial defects like milling wall loss and cracks made by cryogenic temperature. It was concluded that the AE showed good results in fatigue crack linear location but in pressure vessels the results were not good due to great dispersion. Acoustic emission sensors did not detect milling defects and cracks, but cracks led the vessels to fail in hydrostatic tests.
Camerini, Soares and Sant'Anna reported in 1987, that lab tests confirmed that AE had a good potential to be used in equipment inspection, but the field tasks did not reach the same level, and so developments were required to be made. A PAC-3004 equipment with R-15 transducers and 1220A pre-amplifier were used in those works and the coupling was provided by silicon grease. Three lab tests with fatigue cracks and 5 pressure vessels with artificial defects were conducted. For vessels use was made of a PAC system with 12 channels and R15I sensors. The last test was a simultaneous test with two different service companies conducting the inspection at the same time. Unfortunately, the results were different too.
The Research Group from the Brazilian Society for Nondestructive Testing (ABENDE) prepared a program in1988 with the aim of developing AE in Brazil. This group consisted of PETROBRAS, other users and national service companies. The first results indicated a great potentiality due to lab tests, but field tasks did not reach the same level. Based on substantial advances in instrumentation in the period 87-88 and on newly written procedures, another evaluation was made. Then, 5 tubular specimens were built to re-test this new approach. It was found that leak detection had a high sensibility and was proportional to pressure. Sources classified as grade A and B were identified in advance when related to final collapse, this meaning that the proposal criteria in 1988 had satisfactory results. Despite of the fact that fractures did not happen in the marked places, specimen showed failure in locations with structural damages, which were previously detected and localized. Linear and planar locations were not evaluated as a function of the specimen geometry, but zonal location showed good results.
Castelo Branco presented the results obtained from a program for evaluating the reliability of the use of AE technique in industrial scale, in 1988. This program consisted of lab tests in CTOD specimen (15), hydrostatic tests until failure in small pressure vessels (30) and field inspections (60). 16 spheres, 13 reactors, 10 vessels, 9 cylinders, 5 towers, 2 heat transfers and 5 piping were inspected. Five conclusions were established, as follows:
The partnership Brasitest (service company) and PAC-Spartan (equipment) didn't attain good results in 1989 when tests were made in tubular joints under fatigue cycle and pressure vessels submitted to hydrostatic test up to rupture. Camerini and Soares described other conclusions. For example: by means of a pulser it was found that sensors did not have the same behavior. This statement was very important because the analysis criteria were based on parameters obtained from acquired signals. Finally, in 1989, it was suggested that AE should not be used with the instrumentation that was available in the market until deep changes were introduced in hardware and software, and new evaluations could be made.
Castelo Branco and Camerini described in an article published in 1989, the main polemic points in AE techniques. The discussions were focused on the need or not, of an exact source location, sensor performances, global versus specific procedures and over pressure or continuous monitoring during the industrial process. At this point, AE was not a reliable technique, and then new efforts were proposed to develop a technique for use by PETROBRAS. The solutions proposed were the following: to develop exact and on-line location; to increase sensor reliability and fidelity. It was stated too that specific procedures that perform monitoring during real operational conditions have shown better results as compared with over-pressure procedures.
Camerini reported in 1991 some aspects of studies made in AE in PETROBRAS. These aspects were divided into four items. First, in 1998, during a 3-month trip to USA it was concluded that there was an increasing use of AE to verify structural integrity and big companies like Monsanto and Exxon were using it almost as a routine. Despite of the progress, AE was not acceptable and reliable like other NDT techniques. The data acquisition and exact location capability should be improved. Acoustic emission should be used as a complementary exam, not as substitute way. The standards in 1991 were poor and in a low number. The MONPAC program was the most powerful package to perform inspection. Brazilian and PETROBRAS stage was synonymous with the USA market as far as use and doubts are concerned, and AE inspection could be allowed in specific cases. The second item was the conduction of two tests, one in a tubular joint under fatigue cycle and the other in a pressure vessel. Those tests showed that conventional AE was not able to detect signal generated by cracks. The third part was a Master of Science Thesis to study the relationship between AE and a structural steel (A516 Gr-60) during a CTOD test. The main conclusion of this thesis was that signal energy (MARSE) had low intensity, around 25 dB, and the threshold used in field procedures was set to 40 dB; then, it was impossible to detect events from cracks and discontinuities in propagation. The fourth part was a test performed with a Hartford Steam Boiler in the REDUC facilities. In this test it was concluded that 10 areas indicated as sources did not have metallurgical or structural evidences that justified the result. The test report was written after a maximum pressure in hydrostatic test had reached 25 kgf/cm2 and the vessel was led to failure in 28 kgf/cm2. Cracks having a size capable of becoming dangerous to the vessel were not detected.
Looking for a standard AE behavior from a crack flowing in a structural steel, CTOD tests monitored by AE were performed by Camerini, Soares and Rebello. The results were published in 1992, and showed the AE capability to detect a crack before its propagation, that is, before the value of CTOD initiation. Structural steels for application in pressure vessels have been widely studied as regards their fracture behavior, in view of the high level of reliability expected from these constructions. In this case, the non-destructive tests are normally conducted by ultrasonic and radiography tests for internal flaws. When an AE test is adopted, the vessel is loaded by internal pressure, in order to open the tips of the flaws, and to generate AE signals. One can pose two obvious questions: what is the required level of the load? And will such loading give rise to flaw propagation and endanger the vessel integrity? These questions were answered for the specific structural C-Mn steel when loaded to fracture in a CTOD test, throughout a test being monitored by an AE signal detection. The following conclusions could be drawn: during the CTOD test two characteristic peaks corresponding to an intense acoustic activity were observed; the second peak to occur was related to the fracture of the test piece, and was considered meaningless for the AE test; the first peak was shown to be related to the moment when the crack starts the stable propagation. A relevant fact was that an intense acoustic activity was detected before the first stable crack extension had occurred. On a quantitative basis, it was found that the CTOD value corresponding to a far detectable AE from the loaded crack tip is lower than the CTOD for the initiation of the stable crack propagation; and the AE test may be reliably performed at sufficiently low CTOD values, where the fracture risk may be discarded.
In 1994, the Duque de Caxias Refinery (REDUC) called upon PETROBRAS R&D Center (CENPES) to search for a device and technique that would allow to detect leaking in unit process valves in operation. This study was necessary due to the fact that REDUC replaced around 1000 valves during overall maintenance every 4 year. Two devices were tested using two different techniques. The first device was a PAC equipment to inspect vessels using AE, and the second one was the Microsonic equipment using an ultra-sonic method. A small device provided with 3 valves was built, and a pressure range used between 5-kgf/ cm2 and 40-kgf/ cm2. It was concluded that the second device (Microsonic) should be used in refineries and offshore rigs to detect leaking due to its low price, easy handling and good performance.
The research project "Integrity Evaluation of Equipment in service with H2S" conducted by PETROBRAS/CENPES generated a report written by Soares in 1985. In this project CENPES joined a multi-client project "Reducing Duration of SSC Qualification Testing through AE Monitoring" that was being conducted by CLI International. This report shows the main results obtained since October/83 until July/84. This project had for its goal to evaluate different techniques to perform analysis of stress corrosion cracks (SCC) to help to reduce the time period required to qualify high strength steel for H2S environment applications. CLI International stated that it succeeded in identifying increasing AE energy rate, although it should be considered that only 14 out of 70 samples were accepted for analysis. Another important conclusion was that all samples that failed by the SCC had an increasing energy rate higher than 610 units. The main reason for the data loss was easy contamination by spurious sources and failures in small duration tests; these problems impose a limit to AE application.
The research project "Integrity Evaluation of Equipment in service with H2S" conducted by PETROBRAS/CENPES generated another report written by Carneval in 1996. In this report it was described a non-destructive evaluation conducted in a pressure vessel called V-2058 out of service from REPLAN refinery due to hydrogen damage (confirmed, by historic inspections during life time) that was dangerous for safety equipment operation. This report showed conventional ultrasonic inspection based on ASTM A578 in all vessel plates to detect stepwise crack, semi-automatic ultrasonic inspection using procedure for corrosion mapping, hydrostatic test monitored by AE to identify critical areas. The main conclusions were the following: conventional ultrasonic procedures did not cover all vessel areas and spent much time; semi-automatic systems showed a good performance, but at a high price, in addition to lack of specific procedures for hydrogen damage it could be a negative point for this technology; the ultrasonic inspection results confirmed the field inspections performed in the REPLAN refinery, but the results were not confirmed in destructive tests. After completion of hydrostatic test, regions indicated by AE and ultrasonic inspection were re-inspected with the idea of finding whether discontinuities had grown or not. The extreme conditions imposed during hydrostatic test (300 % higher than the hydrostatic test), although the material was not in embrittlement condition, showed us that the intrinsic safety vessel design condition, due to the fact that the vessel did not fail during the hydrostatic test, although a plastic deformation occurred. Acoustic emission indicated different areas when compared with ultrasonic inspection. There are two ways to evaluate this: first, discontinuities detected by ultrasonic inspection were not critical for the structure; and, second, the indicated regions did not show any detectable discontinuity by conventional, non-destructive methods. In view of this, the AE results were classified as non-conclusive.
During September/99, an evaluation was made of the ring fire extinguisher main line of PPE-1A offshore rig using AE. This inspection was performed as part of an internal research project to study acceptance criteria for offshore rig applications of composite materials. The aim was to assure the correct performance of the ring fire extinguisher that was the first application of composite material in these conditions. This pipeline, named PPER01, has the following characteristics: 6-inch diameter, material PSX-L3, 10 kgf/cm2 operation pressure, started operating in 1999, cyclic way, ambient temperature. The inspection was subdivided into two phases: the first phase was performed through visual inspection and the second one using AE. The AE approved a maximum operation pressure for 10 kgf/cm2, and a new inspection was set for one year later.
PETROBRAS refinery in Belo Horizonte (State of Minas Gerais), made a hydrostatic test monitored by AE in three horizontal pressure vessels. The objective of this inspection was to acquire data to evaluate whether there was any relationship between AE and ultrasonic inspection where concepts of fracture mechanics were used. This work was carried out in 1999. The area indicated by AE contained defects, but in view of the great variety of defect depths it was not possible to make a statement about the relationship thereof. Another point was that the vessel was in safety conditions, while the AE provided a result indicating that the vessel should be put out of service.
During cool-down and start-up of two reactors installed in the RLAM refinery significant AE signals were generated. The ultrasonic inspection detected no penetration in welding in both reactors. Severe conditions for reactor operations point to an evaluation of the possibility of fragile fracture and critic discontinuity propagation during start-up, operation and cool-down. A lot of conclusions were established, but in relation to AE it could be said that in future AE tests, a previous study of thermal stress to define a cool-down rate and temperature differential should be conducted. The reasons for this proposal are to provide activity to the discontinuities and not to create new cracks in the structure.
All material submitted to a stress lower than its yield/fatigue limit showed unlimited operational life. The same doesn't happen if together with this stress, a phenomenon that damages the material properties and accumulates defects is present during the material lifetime. In this context, hydrogen associated damage promote loss of mechanical resistant, fracture toughness and failure of material in stress under its yield limit. In 1999, a hydrostatic test made in lab scale, in a pressure vessel out of service due to advanced stage of damage detected in shell plates and also in weld joints, identified the process of cracking and reduction of mechanical characteristics known by step wise cracking. These discontinuities did not allow operational continuity in a safety way. Vessel F-6005 was inspected in service and out of service with semi-automatic ultrasonic inspection system and an AE test was carried out with the SPARTAN 2000 system. The hydrostatic test was performed in PETROBRAS/CENPES. After a pressure range between 10,5 and 15,0 Kgf/cm2 the structure showed a different behavior, indicating a possible point of plastic deformation. It was concluded that this is a real capability to show in advance spots where leaking can be reasonably expected to occur, because the region where the vessel would fail was indicated as 7,9 Kgf/cm2, while the vessel actually only failed with 89,0 Kgf/cm2.
As a specialist dealing with maintenance and inspection activities knows, integrity evaluation cannot be summarized by applying only one technique, but a combination of several inspection methods, allowing detection of existing damage mechanism that, in this way, provide means for analyzing the overall integrity of an equipment. With the AE testing associated with fracture mechanics, both supplemented by metallurgic and material propriety details, new concepts regarding the evaluation criteria are being introduced, enabling the application of methodologies to detect defects associated with degradation mechanisms, instead of a simple detection of discontinuities that do not contribute to equipment deterioration. Examples of this concept are found on papers PD-6493 (BSI) and API-579. PETROBRAS/CENPES are seeking to develop an evaluation methodology. In the year 2000, this methodology was applied to 3 LPG storage spheres, integrating AE, ultrasonic, magnetic particles testing and fracture mechanics, aiming to define the required repairs and point out the discontinuities that must be followed during operation. It was concluded, based on the consistency of the results, that the analysis of AE and fracture mechanics point out to the proper procedure for analyzing the equipment structural integrity.
| © NDT.net - info@ndt.net | |Top| |