Read STYLWAN text version
June 14-16, 2000 South Shore Harbour Resort League City, Texas USA
THE INSPECTION OF MARINE RISER SYSTEMS
Rafik Sabet TRANSOCEAN SEDCO FOREX Steve Papadimitriou - STYLWAN
The marine drilling riser (Riser) provides a conduit for the drilling fluids from the subsea BOP up to the drilling rig or the drillship. A Riser joint is selected for use in the field based on minimum material strength requirements. The selection is complemented by safety factors in order to further reduce the risk of a failure. The safety factors reflect the experience of the particular operator and its tolerance level to adversities. A failure occurs when the application requirements exceed the actual material strength. As with everything else manmade, it is not a matter of "IF a failure occurs" but a matter of "WHEN a failure occurs". Inspection can further minimize the probability of a failure and delay, with a bit of luck indefinitely, the inevitable "WHEN".
IMPERFECTIONS IN MARINE RISERS
Deterioration of Marine Risers occurs for different reasons, some of which are summarized in Table 1.0 below (typical operators survey).
Operation induced fatigue, wave action, Cross-Flow Vortex Induced Vibration (VIV) can generate very high fatigue build-up Impact loading during handling, deployment and retrieval Mechanical damage, Fatigue, High tension accelerates fatigue crack growth Deterioration of the metal due to chemical or electrochemical reaction. manifests as wall thinning (uniform or localized)
3D MFL 3D MFL 3D MFL 3D MFL Wall, 3D MFL Wall, Mechanical Wall, Mechanical Wall, Mechanical, 3D MFL Wall, Mechanical, 3D MFL Mechanical
Fatigue damage Vibration (lower fatigue build-up than VIV)
Wear Wall loss due to mechanical wear and fluid movement Erosion Corrosion
Key Seating wear from the drill string Material movement increases both the wear and the rate of corrosion When tube bends, the cross-section flattens and becomes oval in order to reduce the stored energy reduces collapse pressure
Random mechanical damage Ovality
Where: (3D MFL) Three-Dimensional Magnetic Flux Leakage. (Wall) Wall Thickness measurement. (Transverse / Longitudinal) Perpendicular / Parallel to the material length. (Mechanical) Dimensional measurements.
INSPECTION PROCESS and EQUIPMENT
Inspection is carried out in order to verify that the material exceeds the minimum strength requirements for the application. The strength is influenced by many variables like fatigue, stress concentration, variable loading, torsion, shock, internal, external and mid-wall imperfections etc. The designer must consider all of those factors, so should the inspector. Inspection will eliminate MOST of the material that does not have the minimum design strength, but not ALL, and it will not prevent failure from a bad design, from mishandling or from misuse. There are three parts in any inspection process. · · · Detection is the process that locates an imperfection. Evaluation is the process that isolates the imperfection and measures pertinent parameters. Classification is the process that assigns Type & Value to the imperfection. Imperfections that exceed a Type & Value threshold become flaws (Defects).
The designer of the inspection equipment must now choose between Measurement or Testing. · · Testing is not as rigorous as measurement and it is only appropriate when a simple "pass / fail" indication may suffice. Testing may utilize assumptions and simplifications. The results of Testing are often slanted because they depend heavily on the operator's interpretation. Measurement gives quantitative results based on accurate and traceable calibration.
Countless volumes of "Strength of Materials" literature are available at the closest library. For centuries, practicing engineers have depended on that knowledge to make important decisions. After extensive research, STYLWAN based its 3D-FEI on that Knowledge which demands Measurement. Centuries of knowledge warn that Testing will not suffice, especially when safety is also a concern.
The very first step of any inspection unit design is to define its Detection capabilities. This critical design step must exploit in-depth knowledge of the different types of imperfections and the available state-of-the-art means for their reliable Detection. For Evaluation and Classification to be accomplished, the imperfection must first be Detected. One would think that the most sensitive technique should be employed for Detection. However, as many have found out the hard way, the most sensitive technique may not work reliably in the hostile oil field environment. In fact, many techniques do not work at all outside the laboratory. STYLWAN selected Magnetic Flux Leakage (MFL) because it can be used reliably, both in the laboratory and out in the hostile oil field environment. STYLWAN then applied measurement techniques and modern computer technology to overcome many of the MFL shortcomings and improve its strengths. Imperfections in the path of the magnetic flux act as obstacles. When the magnetic flux encounters an imperfection, the flux is forced to go around the imperfection (the obstacle). If the applied magnetic field is properly set, some of the flux will exit the Material-Under-Inspection (MUI) around the imperfection and will reenter the MUI after the imperfection (the flux is diverted). A properly located sensor (or sensors array) will detect this flux change around the MUI. There are three main steps in the MFL inspection: a) Magnetize the MUI; b) Scan the MUI with appropriate sensors and c) Process and present the signal from the sensors to the operator in a meaningful manner.
One-Dimensional (1D) MFL TECHNOLOGY BRIEF REVIEW
TYPICALLY 8 CHANNELS
AMPLIFIERS & FILTERS
SELECT HIGHEST SIGNAL
HIGHEST UNCORRECTED MAGNETIC ANOMALY
MFL has been the backbone of the OCTG 1D inspection industry since its inception in the 1940s and very little has changed since then. Modern MFL inspection units, including 4-Function units, Drill Pipe buggies and Coiled Tubing units, use a number of sensors to surround the MUI in the 1940s formfitting fashion. Even today, a set of form-fitting sensors is required for each material size and shape. Developed in the 1940s, the 1D signal processing is "Based on the ASSUMPTION that the largest signals originate from the most serious (deepest) flaws". Experience and FEA easily demonstrates that this Highest Magnetic Anomaly assumption is inaccurate, especially for the inspection of materials that are used under dynamic loading. Nevertheless, it is still the state of the art for the 1D Non-DestructiveTesting and is promoted as such. Remember, Testing allows for assumptions, Measurement does not. Today, just like in the 1940s, the signal from each sensor is amplified individually and then the highest signal is selected for presentation to the operator (see picture above). This type of signal processing does not detect the imperfection's Geometry or Neighborhood and results in a one-dimensional inspection system. The 1940s chart recorder has been replaced with a modern day computer and the signals are presented in colorful displays and printouts. However, there is no signal processing backbone behind these skin-deep beauties. Take away the colorful displays and the Bells and Whistles and all you are left with is a signal that does not reflect the failure potential or the depth of any imperfection. 1D READOUT: Since 1D cannot detect geometry information to segregate imperfections, 1D presents all the Highest Magnetic Anomaly signals on a single trace in an "Imperfection and Flaw soup". Thresholds, therefore, apply equally to every indication in this "Imperfection and Flaw soup" regardless of its origin. Pits, cracks, hardness changes etc are all treated incorrectly as equals. The Classification process for1D is then reduced to "if every signal originates from a fixed geometry pit that is alone in the neighborhood, then the depth will be..."; which is not quite relevant to the question on hand: "What is the failure potential of the imperfection under loading?". The unit operator is the "Imperfection and Flaw soup" interpreter and the inspection conclusions are always influenced by his/her education, ability, mood swings, lack of sleep etc. Most of his/her training and focus is aimed at overcoming the 1D inspection unit shortcomings, not the customer's needs.
STYLWAN Three-Dimensional (3D) TECHNOLOGY BRIEF REVIEW
SN1 S11 . . . . . . . . . . S1X KN1 ... K K11 ... K1X . . . . . . . . . . KX1 ... KXX
STYLWAN FLAW SPECTRUM
At STYLWAN, we never believed that repackaging, somehow magically, instills new abilities or improves the capabilities of the 1940s 1D technology. Neither do computers that mimic the old chart recorder. Mimicking the analog world is not always the strong point of computers. Carrying out mathematical calculations, like Finite Element Analysis, is what computers do best. In order to overcome most of the 1940s MFL shortcomings, STYLWAN pioneered the development of 3D inspection technology. 3D uses a number of 2D sensor arrays distributed along the length of the MUI (see picture above). Each 2D sensor array uses a number of sensors to surround the MUI. The sensor signals are processed through the Extraction Matrix. 3D solves systems of equations by using measured quantities and applies stress concentration and Neighborhood correction factors in real time. 3D enjoys significant advantages over the 1D 1940s technology in resolution, sensitivity, Signal-toNoise ratio (S/N), tolerance to adversities, reliability of Detection and application flexibility. 3D has no analog counterpart and it did not exist until STYLWAN defined it in the digital realm. 3D READOUT: As opposed to 1D Testing, STYLWAN 3D-FEI measures the imperfection's geometry and segregates imperfections in the FLAW SPECTRUM which spans continuously from FATIGUE (microcracking) to WALL THICKNESS changes. The FLAW SPECTRUM automatically performs the classical EVALUATION process based on measured parameters and it is free from the subjective interpretation of the operator. Although STYLWAN Measurement is not infallible, it is by far more dependable than the most promising assumption-based operator-interpreted Testing. 3D-FEI Reference is automatic on the MUI and always takes 5 seconds.
The pictures above were obtained during a direct comparison test between an NDT-2000 drill pipe buggy and a STYLWAN 3D inspection system (as shown on the right). The NDT-2000 is an original drill pipe inspection unit that has been copied widely by the parasitic NDT equipment manufacturing industry, so, the comparison results apply to many other drill pipe buggies as well. Notice that the STYLWAN inspection heads are designed to readily accept 1D drill pipe sensors also (a set of sensors and centering mechanisms per pipe size). During such tests, STYLWAN has documented 28% erroneous Classification by 1D, notably the inability to detect internal imperfections and fatigue.
THE STYLWAN FLAW SPECTRUM
STYLWAN 3D-FEI does not assume anything about the imperfection Magnitude, Geometry or Neighborhood. It measures it and then it derives the Flaw Spectrum. The STYLWAN 3D-FEI exists in a continuum and does not recognize different inspections (wall, rodwear, pitting etc) which, after all, are only 1D Testing technology artifacts. These artifacts were created to cope with the inability of 1D to delve in-depth and extract all the needed inspection information from a single technique or a unified sensor arrangement. This forced 1D to scratch the surface of multiple techniques, some of them outright dangerous (Radiation), to obtain inspection data. STYLWAN divides the flaw spectrum as follows (STYLWAN trademark sequence and color):
2-D (purple) 2-d (red) c (cyan) 3-d (green) 3-D (blue) Wall (Black or Yellow)
Each category is further subdivided to (alpha), (beta) and (gamma) subcategories excluding the Wall. designation is toward 2-D and toward 3-D. The spectrum and its meaning is:
2-D Fatigue 0 to ~50% of life cycle Fatigue ~50% to ~75% of life cycle, Identify future failure locations Fatigue ~75%+, cracks visible with magnifying glass Crack visible with naked eye, cracks in hardbanding Pinholes (coiled tubing definition), crack initiation seeds CO2 type pitting, crack initiation seeds ~1/8" pit, Default for failure to calculate (saturation or failure to converge) Pinholes (coupled tubing definition), small gouges H2S type pitting, ~1/4" conical pitting Mechanical damage, Hardness changes (~up to Rockwell A-Scale 65) Mechanical damage, Hardness (~Rockwell A-Scale greater than 65) Taper, corrosion bands, erosion, rodwear etc STYLWAN STANDARD ACCEPTANCE TEST DNV Certificate No: HIO-00-0077 One of the tests uses three test joints. A) a 1.75" x 0.095" flawless tube (outer material in concentric configuration); B) a 1.25" x 0.095" tube with a 1/8" drilled through hole and a 6.00" wall loss and C) a 3/8" OD coated steel cable with removed (broken) strands. The test joints are run in a concentric configuration as shown on the right. Notice the clarity and the exactness of the inspection traces on the left. The traces show the imperfections in the order they were described above. The 1/8" hole is Classified as 2-d and the cable broken strand (crack) as 2-D. The wall loss is on the left and removing the steel cable is shown on the right.
For STYLWAN 3D-FEI, the FLAW SPECTRUM automatically performs the classical EVALUATION process based on measured parameters. On the other hand, 1D relies on offline manual Evaluation using other means (ultrasonic, eddy-current, visual etc) to assign TYPE to the imperfection. Notice that the offline Evaluation takes place only for imperfections that were flagged by the "Highest Magnetic Anomaly". When offline Evaluation is not possible, 1D is forced to assume (again) an imperfection morphology. This TYPE assignment is then based on a second order assumption that is unforgivable and defenseless anywhere. It is easier to defend a decision "not to inspect" than to defend a second order assumption, especially when there is ample evidence that the first assumption, the Highest Magnetic Anomaly, has a very high probability of error. The spirit of inspection is to be pro-active. For the inspection to fulfill its pro-active role, the inspection must Detect failure seeds and Evaluate them as such. While 1D claims the ability to detect fatigue cracks, the STYLWAN 3D-FEI will detect the seeds (including Fatigue) before the cracks develop. By the time Fatigue progresses to cracking, it is normally too late. When possible, the removal of the failure seed or an operation modification will extend the useful life of the material and eliminate a failure. The STYLWAN 3D-FEI adds other dimensions to the Evaluation process that are unthinkable with 1D Testing techniques. 3D-FEI exports the inspection data for use by Spreadsheets and Finite Element Analysis engines and facilitates the creation of meaningful Databases.
Classification for the STYLWAN 3D-FEI is a matter of applying threshold functions to the FLAW SPECTRUM TYPE assignment. The PROFILE MAKER facilitates the creation of complex custom Classification functions (threshold curves). In 3D, even the thresholds are expressed as functions and are implemented mostly as hysteresis thresholds (MULTISLOPE Classification). The STYLWAN 3DFEI, especially the AutoCullTM, uses dynamic-self-adaptive inspection techniques. Instead of relying on fixed threshold levels, the 3D-FEI dynamically adapts the threshold levels to the overall tube condition, the Corrosion Gauge (CG) and the Environment Induced Cracking (EIC) gauge. AUTOCULLTM uses MULTISLOPE Evaluation and Classifies the tube automatically in real time. On the left is a 3D-FEI AutoCull inspection screen. The Wall Thickness loss exceeds the "Rejection Level" and is Red. Pitting is in the gray area (Green). AutoCull scans for "Good Tube" and grades the tube on the first encountered flaw that exceeds a "Rejection Level".
AutoCull is possible because a Measurement can be automated. An Assumption cannot.
Over the years, STYLWAN and its customers have compared STYLWAN 3D inspection equipment with 1D inspection units. Coiled Tubing in particular, with its exceptionally low life cycle, has exposed the 1D Testing severe shortcomings. The life of a Riser is somewhere between 5 and 15 years, so tracking a riser joint is not as enlightening. On the other hand, and under certain applications, the entire life of a coil is use up in one day and it can be tracked. TEST #1: Used Coiled Tubing (CT) was collected from different users for this test. No record existed about the condition of the tube nor a baseline was obtained before the test flaws were machined onto the tube. The inspection traces below show a test joint with an artificial external pit. STYLWAN MGN is shown on the left and a 1D unit reconstruction is shown on the right. At first glance it apMGN 1D peared that STYLWAN had not detected the pit 50 100 as clearly as the 1D unit had. In fact, when 40 80 the test results were 30 60 MGN 1D released, the 1D service 40 20 provider published a 20 10 paper with the state0 0 ment "... detected the flaws". Not true. Appearances are deceiving. STYLWAN MGN detected that this "used" tube came from the junk pile (retired) and the brand new machined pit did not add any significant incremental damage, so it does not stand out. STYLWAN's measurement Classified the tube properly as "scrap" while Testing failed. Does 1D show "good tube"? Would 1D recommend putting this tube back into service? Would 1D pay for the fishing trip? TEST #2: This acceptance test utilized a coil of 1.75" x 0.175" heavily used CT. A 1D unit had just run the same acceptance test and missed significant surface rippling a late fatigue life manifestation (Defined also by API). In fact, this particular flaw was classified by 1D as "less than 5% pitting" another 1D Testing failure. It was classified by STYLWAN as 75%+ of the fatigue life. Quoting verbatim from 3D-FEI 1D the report about the 1D testing: "... Both sec100 100 tions appeared to be 80 80 rippled, but (1D inspec60 60 tion company) did not 3D-FEI 1D notice this. After the in40 40 spection we were given 20 20 nice colourful printouts. 0 0 What I miss is some kind of explanation ...". The explanation is simple. This exemplifies the difference between STYLWAN 3D Measurement and the 1940s 1D Testing and this is why Testing is inappropriate for the inspection of materials that are used under dynamic loading, like Coiled Tubing, Drill Pipe, Marine Risers etc. Colorful printouts and Bells and Whistles do not make up for the 1D dangerous Assumptions or the dubious Simplifications.
About STYLWAN (Quoting verbatim from the same report): "The inspector and the designer (STYLWAN) of the tool got all excited. By looking at the signal only, a wall thickness decrease of 0.004 to 0.005" was expected. The transverse signal showed 2D indications, meaning micro cracking. They stopped the reels and started taking measurements and found a maximum wall loss of 0.005". The nominal thickness of the CT was 0.175", while the actual thickness at certain spots appeared to be 0.170". The rippling was attributed to fatigue." Of course STYLWAN got all excited. This is a real life problem that 1D will always miss with the full knowledge of the 1D service providers. "By looking at the signal only" (3D-FEI Flaw Spectrum display), STYLWAN Detected and Classified correctly 0.005" wall loss (2.9% !!) and Fatigue (...micro cracking). No erroneous assumptions, no blind spots, no need for excuses or explanations. TEST #3: In November 1996, STYLWAN equipment was used at the wellsite to inspect 103 lengths of 7/8" sucker rod through 2-7/8"x 0.217" production tubing (concentric inspection shown on the right) and then to inspect the 85 lengths of tubing. Both the sucker rods and the tubing were afterwards hauled off to an inspection facility and they were reinspected with a stationary NDT-5500. The inspection facility results matched exactly the STYLWAN results (Provost, Alberta). TEST #4: STYLWAN compared an AutoCull enhanced 3D-FEI unit to an NDT-5700, which is one of the workhorses of the 1D inspection industry. 105 joints of 2-7/8" x 0.217" were run through both the units. The test focused at the lower end of the flaw spectrum (at or below 20%) because both units, including the AutoCull, have adequately demonstrated competence in detecting flaws over 20% (of the 105 joints, only one was red). The 3D inspection head was placed at the entry racks of the NDT-5700 (as shown on the right). STYLWAN called the pipe since it went through its inspection head first. The call was radioed to the NDT5700 inspector. Discrepancies were noted instantly and inspection traces were compared. There was complete agreement between the NDT-5700 and STYLWAN on all 105 joints for Wall and 102 joints for Transverse. Inspection Note 3 describes the results in greater detail. Only Joint #3 is discussed here. The NDT-5700 inspection trace is shown on the left. A typical 1D "Imperfection and Flaw soup". On the first run the NDT-5700 missed the flaw at 18'. On the second and third run the NDT-5700 matched STYLWAN. Old thick paint was peeling off and folding over within 1.0" from the indication. Paint chips probably accumulated under the NDT-5700 contact sensor lifting the sensor during the first run. This sensor liftoff resulted in a 50% signal loss placing the flaw signal below the "grass-cutter" threshold level, thus the signal was eliminated. Notice that under those circumstances the sensor liftoff (estimated at about 0.010") would attenuate all of the signals by 50%. On the other hand, the STYLWAN 3D systems are reliable and trustworthy because they are highly tolerant to these adversities, a fact also certified by DNV (certificate No: HIO-00-0077).
The STYLWAN 3D-FEI is highly portable and the inspection sensors are built inside the inspection heads. External, specialized sensors, can easily be installed also. With the 3D-FEI, there is no need for centering or a clean pipe surface for competent inspection. An offshore inspection system is shown below left (the small internal sensors or the handheld sensors are not shown). The inspection head on the left has 10.5" ID clearance and the inspection head on the right has 5.5" ID clearance. Capability wise, appearances are deceiving.
Larger inspection heads are also available (16.0" ID clearance is shown above middle). Inspection of tubing through pressure spools is possible (above right). Precise location of collars between the snubbing unit BOPs can be combined with inspection. The STYLWAN 3D-FEI is the only inspection unit capable of concentric material inspection.
DETECTION and CLASSIFICATION of FATIGUE
For centuries, practicing engineers have recognized that subjecting metal to stress cycles will result in fracture although the forces involved were much smaller than the forces required for static failure. In his "Mecanique industrielle", published in 1839, J. V. Poncelet (1788-1867) was the first to discuss the effects of repeated cycles of stress and most likely, he was the first to introduce the term "fatigue". Fatigue is cumulative and it manifests itself as microcracking between the steel grains. With additional cycles, fatigue progresses to cracking as the microcracks grow and bridge. STYLWAN pioneered the fatigue build-up Detection, Evaluation and Classification techniques (Left: Fatigue line tests May 1995 Aberdeen Scotland). During a field Fatigue detection test (right 1.500" Coiled Tubing - 10,000 psi pump pressure 40 cycles life), STYLWAN detected the Fatigue build-up and marked on the tube the exact failure location before any other means could detect anything (March 1995 Yorktown Texas). Most software failure prediction models are aimed at predicting the alpha failure location (L)TM where the rate of fatigue build-up is the highest, hence, it is the location where the first failure is expected to occur. STYLWAN's tests reveal that at least two predominant locations can be identified as early as 50% of the life cycle. However, the most catastrophic form of failure is Early alpha (E)TM failure that is not predicted by any model. Presently, STYLWAN can only identify a few (E) seeds, but it can easily detect the rapid fatigue build-up.
INSPECTION OF CHOKE, KILL and BOOSTER LINES
Baseline of new Choke & Kill line is shown on the top left. Baseline can be carried out on new or used lines and is essential for beforeand-after comparison. STYLWAN extends an advantageous price for baseline creation of any material (Main Tube, C&K, Booster lines, Kelly, Drill Pipe, MUX cable, Mooring Lines etc). Concentric configuration is used to verify the 3D-FEI operation capabilities (top left) because it clearly shows that the STYLWAN 3DFEI equipment: a) are highly sensitive; b) can inspect reliably the ID of the material with very high S/N ratio; c) can reliably inspect concentric material of any complex configuration; d) can reliably inspect OD coated material; e) can reliably inspect odd-shaped material (semi-oval sucker rods, Kelly etc); f) can reliably inspect multiple material sizes and shapes with one sensor; g) can reference automatically on any material shape in 5 seconds and h) are trustworthy and reliable because they can tolerate severe adversities. Concentric inspection is also certified by DNV. LEFT: The inspector is using an internal sensor to inspect the C&K and Booster lines onboard the rig. The internal sensor will not damage any coating.
INSPECTION OF MAIN TUBE
Onboard the rig, a STYLWAN 3D-FEI can inspect the risers during transit from location to location, the drill pipe during every trip and all other OCTG as needed, thus averting potential transportation damage after the inspection and delays. Only materials that need repair are then shipped to shore. The quality of inspection and the cost savings are unsurpassed by any other inspection technique. Left: Inspection of Marine Riser Systems onboard the rig. 3D inspection of Marine Riser Systems is normally carried out onboard in less time that it takes to ship the risers to shore. Right top: The inspector is preparing the internal probe for C&K inspection. Right bottom: Weld line inspection. The STYLWAN 3D-MR1 standard includes procedures: a) to establish remaining Fatigue Life; b) body inspection of MT, C&K and Booster lines; c) inspection of couplings, connections, Jewellery and welds; d) Mechanical and Dimensional measurements and e) miscellaneous ancillary techniques and methods.
INSPECTION OF DRILL PIPE, KELLY, COILED TUBING, HOISTING and MOORING CABLES and MUX CABLE
Onboard the rig, the STYLWAN 3D-FEI can be used to inspect many materials that can be inspected with MFL techniques. For example, inspection of Drill Pipe during a trip utilizing the STYLWAN 3DFEI provides a superior inspection compared to the one obtained by shipping the drill pipe to shore as has been established through direct comparison with the 1D workhorses (NDT-5700, 5500, 2000).
Above: Inspection of tubing on the wellhead. Only STYLWAN can inspect tubing sizes greater than 31/2". As can be seen (third from left) nothing touches the pipe. Nothing! During inspection, the pipe is free to travel laterally without any performance penalty (also certified by DNV). No cleaning is required whatsoever. The STYLWAN 3D-FEI is rated for inspection speeds in excess of 1,000ft/min. Below: Coiled Tubing inspection. Only STYLWAN can inspect all of the materials of a concentric system like tube inside tube, armored cable inside tube, steel cable inside tube etc.
Above: Inspection of MUX cable (left & right), flat strip and steel cable (workover rig hoisting cable). STYLWAN, through its measurement based 3D-FEI, offers to the offshore industry a versatile inspection solution that is unmatched by any other inspection technique. What is impossible for 1D, is child's play for the 3D-FEI. All in a small, compact, lightweight package that calibrates automatically in 5 seconds and provides meaningful answers. The inspection data can be exported directly to FEA engines for further analysis to maximize the utilization of the material. Versatility wise, performance wise and cost wise, the STYLWAN 3D-FEI surpasses all others by orders of magnitude.