A prototype inspection system was designed and manufactured by SAIC and is currently undergoing acceptance testing. Laboratory and on-aircraft testing to date have proven the prototype system has a detection threshold potential of (0.050 inch crack lengths in the C-141 splice-joint second layer. A Probability-of-Detection (POD) Study has been designed to formally quantify the inspection reliability of the prototype process. POD data will allow the C-141 SPD to confidently determine future inspection intervals and requirements. The new automated ultrasonic scanning process has the potential to significantly reduce splice-joint inspection labor costs as compared to the costs of BHEC inspection. This paper describes the inspection development process, including feasibility studies, design and fabrication of prototype, functional testing, POD testing, and Air Force implementation. Typical ultrasonic C-scan images will be presented, and the POD test plan will be discussed.
The spanwise splice cracking problem dictated the need for inspection of the second-layer or inner tab of the splice-joint. Bolt hole eddy current (BHEC) inspection is the only immediately fieldable method available, capable of providing a reliable inspection of the splice-joint inner tab. BHEC has the proven capability of detecting 0.070" cracks in the inner tab fastener holes, but its use would require the removal of every splice-joint fastener. Cost estimates for BHEC testing of the C-141 spanwise splice-joints have ranged up to 9,000 man-hours per aircraft. Potential risks associated with BHEC include inducing damage to fastener holes during large scale removal and reinstallation operations, and the degradation of fuel tank sealant integrity. Based on the associated costs and risks, BHEC is an unacceptable alternative for large-scale inspection of C-141 spanwise splice-joints. As an interim safety measure, the C-141 SPD issued a Time Compliance Technical Order (TCTO) in July '96 which directs the performance of an eddy current surface scan on the exterior surface of the lower inner wing spanwise joints. This inspection will remain in effect until a new method; capable of detecting smaller inner tab cracks can be developed and validated.
The C-141 SPD requested that WR-ALC/TIEDM develop an alternative inspection method with the following capabilities: a) proven detection of 0.125 inch second-layer cracks, b) inspection from aircraft exterior without fastener removal, c) archived inspection records, d) automation to remove operator subjectivity, e) portability, f) significant return-on-investment. Based on inner wing damage tolerance analysis data, 0.125 inch crack detection capability would allow a five-year inspection interval to be established for the spanwise joint.
TIEDM immediately explored and evaluated low frequency eddy current (LFEC) and manual shear wave ultrasonic inspection techniques to satisfy the C-141 spanwise joint inspection need. A LFEC technique developed by TIEDM and a manual shear wave ultrasonic technique developed by Lockheed were both proven to be acceptable for detection of a 0.375 inch crack that extends forward from an inner tab fastener hole to the panel edge. Due to a combination of joint geometry, panel thickness changes, and/or fastener variables, neither technique has yet proven to be an effective means for detection of shorter length cracks in the forward direction or for any crack length in the aft direction. Subsequently, neither technique will satisfy the specific C-141 SPD need for 0.125 inch crack detection in the spanwise splice-joint inner tab.
At the time of the C-141 SPD request, the FAA, under its Aging Aircraft Program, was funding the development of a multi-layer corrosion detection technique for the DC-9 wing box (Ref. 3). Successful on-aircraft demonstration of the DC-9 wing box inspection was performed in the first quarter '95. Based on similarities between the DC-9 wing box and the C-141 splice joint configuration, i.e., multi-layers bonded with aircraft sealant, TIEDM decided to pursue a proof-of-concept study to explore the feasibility of using a scanning, ultrasonic imaging technique to solve the C-141 spanwise splice-joint problem.
FIGURE 1. C-141 Lower Inner Wing Detail.
FIGURE 2. Splice-Joint Configuration.
The C-141 wing is constructed of five major assemblies, the center wing box and the left- and right-hand inner and outer wing sections (see Figure 1). The inner wings are joined to the center wing at Wing Station (WS) 77.7 and to the outer wing at WS 405. The inner wing is designed to carry fuel and is divided into four separate fuel tanks. The inner wing lower surface consists of 11 wing panels numbered 1 through 11, AFT to FWD. Panels 1 through 6 span the entire inner wing length (approximately 27 feet). The length of Panels 7 through 11, taper with the leading edge of the wing
The lower inner wing panels are attached together at spanwise splices with 0.250 inch to 0.375 inch taper-lok fasteners. The fasteners are installed wet with sealant, and a faying surface sealant is applied in the panel lap joint. The sealant applied at manufacture was MIL-S-81733. The splice-joint configuration is shown in Figure 2. Total joint thickness (two layers) ranges from approximately 0.275 inch to a maximum of approximately 0.825 inch. The thickest locations are in the transition zones near the outer wing and center wing attach points. Inner tab cracking initiates at the forward and/or aft side of the splice-joint fastener holes, typically on the interface side mated with the outer tab. Cracks initiating on the forward side of a splice joint fastener hole propagate until the edge distance is traversed and the ligament is severed. Cracks initiating on the aft side of a splice-joint fastener hole propagate until they are surface breaking, as shown in Figure 2.
The current TCTO inspection performed on the C-141 spanwise splices is an eddy current surface scan which detects these surface breaking cracks. The procedure consists of the following steps: 1) Removal of the lap-joint sealant bead as necessary in the area of inspection, 2) Calibration on lap-joint standards, and 3) Surface inspection. The surface inspection entails scanning a surface probe along the aft lap-joint panel the length of each spanwise joint, #'s 2 through 9, as noted in Figure 2. The procedure is designed to detect a surface crack < 0.100 inch in length measured from the panel edge. Total crack length at this point of detection (as measured from the inner tab fastener hole) is approximately 1.0 inch. This inspection is repeated every 120 days on every C-141 aircraft and has become a major manpower burden on field-level NDI personnel. Some high time aircraft are inspected every 30 days. To date, one splice-joint crack has been found with this TCTO.
FIGURE 3. Feasibility Study Test-Article Configuration.
The feasibility study commenced with an ultrasonic immersion evaluation of the splice-joint test article. Immersion scanning was used to take advantage of the resolving capabilities of a focused probe and to eliminate mechanical contact scanning affects. Initial evaluations focused on the sealant bondline. The influence of the sealant bondline on the propagation of ultrasonic shear waves ultimately controls the inspection sensitivity capable by ultrasound on the splice joint configuration. Immersion testing yielded vital bondline information. It was determined that the sealant bondline provided ample sound coupling through to the inner tab of the splice-joint. Further immersion evaluations performed on the test article determined acceptable probe angle, incident angle, transducer frequency and depth gating requirements. A 20( incident angle provided the best signals from the fastener holes and saw-cuts. A typical immersion C-scan image of the test article is shown in Figure 4.
Typical Immersion Scan Results.
The scan was performed using a 5 MHz, 0.50 inch diameter, spherically focused (1.5 inch focus) immersion transducer, and a 0.010 inch scan resolution. Separate scans were performed with the transducer aimed inboard, then outboard. The separate inboard and outboard files are pasted together in Figure 4. Time-of-flight and amplitude data are collected with each scan. Amplitude data only is shown in Figure 4. All artificial defects in the test article were successfully detected, as noted in Figure 4, and in the zoomed image shown in Figure 5.
To illustrate the quality of the ultrasonic signals received from the inner tab of the test article, captured A-scan results from Hole #9 are shown in Figure 6.
The C-scan is comprised of peak amplitude information collected from a precisely gated signal in the second layer of the test article splice-joint. Hole #9 has two saw-cuts in the inner tab fastener hole; an 0.045 inch, 45( corner notch at the interface surface, and a 0.040 inch through-the-thickness cut. The image was generated with the transducer aimed in the outboard direction. In the C-scan, higher amplitudes are shown as lighter shades. The cursor is located over the corner notch image with the corresponding A-scan shown above. The test yields excellent signal-to-noise ratio for this small indication. The through-the-thickness 0.045 inch cut image is located in the lower portion of the C-scan, just to the left of the vertical cursor. The large oval image centered in the C-scan corresponds to a signal from the inner tab fastener hole. The small oval image is skip information, inherent to angle-beam inspections, from the first-layer (outer tab) fastener hole. Note that the saw-cut images are separated and appear at an angle to the fastener image. This distinct separation and location will allow crack signals to be readily identified. This phase of the study proved that ultrasonic inspection coupled with imaging technology provided adequate inspection sensitivity for detection of small fatigue cracks in the splice joint inner tab, provided the sealant bondline is intact.
The next phase of the project involved determining the feasibility of converting the successful laboratory immersion techniques to a field-level process. This conversion involved the use of automated scanning equipment and contact transducers. Contact inspection introduces a variety of disadvantages versus immersion inspection, including mechanical scanning affects and loss of the focused probe resolution. These disadvantages must be overcome with appropriate selection and design of high precision scanning equipment and contact probes.
The reference sample was scanned using an SAIC manufactured, 45( shearwave, 0.250 inch x 0.250 inch, 5.0 MHz contact transducer, mounted in a standard Ultra Image semi-automatic scanner assembly. Motion encoders automatically provide the computer with precise transducer location while scanner movement is performed manually. A transducer holder was assembled that would hold transducer skew to ( 10(. The holder was designed such that the transducer is slightly recessed from the inspection surface, allowing the transducer to slide on a thin film of couplant applied by an automated couplant delivery system. A portable ULTRA IMAGE IV system was programmed for scan parameters, and ultrasonic settings were optimized for contact transducer scanning. The scan resolution was set for 0.020 inch. Again, two scans were performed, one with the sound beam aimed in the inboard direction, the other with the sound beam aimed outboard. Redundant inspection from both sides of the fastener reduces the probability of missing data due to a variety of variables, including poor surface conditions, bondline voids, or crack skew. Figure 7 shows typical C-scan data collected during a semi-automatic contact scan.
The image was generated with the transducer aimed in the inboard direction. Due to the physical limitations of contact scanning at the end of the test article, only 13 fasteners could be inspected per scan. The image displays time-of-flight (TOF) data at the top and peak amplitude data at the bottom. In the TOF scan, fastener holes and saw-cuts are clearly shown with the gradient scale corresponding to decreasing TOF, light to dark. In the amplitude scan, the saw-cut images are not as distinct, but can be distinguished by their lateral position at the top and bottom side of the fastener hole image. Note that saw-cut image positioning is the same as detected during immersion scanning (Ref. Figure 4), although signal separation between the fastener hole and saw-cut images is not as distinct. All through-the-thickness saw-cuts ( 0.040 inch were detected with the semi-automatic scanning set-up. Corner notch detection proved more difficult, as only one corner notch (0.075") at the interface was consistently detected.
The feasibility study was completed by SAIC in December '95, and their results were reported to TIEDM. The study conclusively demonstrated the feasibility of fielding an automated ultrasonic inspection process capable of second-layer fatigue crack detection in the C-141 lower inner wing spanwise splice-joints. It was further demonstrated that, with further process development, fatigue cracks as small as 0.050 inch may be targeted for on-aircraft inspection. These results far exceeded the C-141 requirement for 0.125 inch crack detection. TIEDM reported the results to the C-141 SPD and concluded that automated scanning, coupled with ultrasonic imaging technology, offered the best near-term solution for C-141 spanwise splice-joint inspection needs. TIEDM recommended pursuing further development and optimization of this inspection technology for the C-141 application.
|Scanner design efforts focused on reducing weight while maintaining speed, resolution, and reliability. The scan envelope (36" X 6") was designed to accommodate the long, narrow area of interest. The width of the splice-joint area in need of inspection is only 2 inches, so the Y-axis extension was intentionally kept to a minimum to reduce the weight of the scanner. Because of the procedural numbering sequence chosen, all scans will be less than 36 inches in length. The scanner is vacuum attached to the aircraft skin with three, six-inch cups spaced along the scanner length. The scanner axes are powered by two AC-induction motors. The motors and accompanying controller provide high resolution and positional accuracy to ( 0.001" at speeds up to twelve inches per second. Integral scanner safety measures include visual gauge indicators and audible alarms. Supplemental safety measures include safety straps and a portable catch basin. Total scanner weight is approximately 40 pounds.|
|Support functions required by the scanner are provided by the scanner utility system. The utility system supplies regulated air pressure, automatic coupling, and vacuum control. These systems are integrated inside the portable cart as a slide-out module. All utilities are monitored and controlled by the operator from a control panel on the cart exterior. Utility specifications are provided as follows: Air (0 to 45 psi), Water (0 to 35 psi), Vacuum (18 to 22 inches Hg).|
|The cart is totally self-contained and requires only access to 120V external power. The cart houses the system computer, scanner motion controller, couplant reservoir and utility box. The cart also provides for monitor, cables, and miscellaneous parts storage during transport. The cart also contains an integral utility control panel and pullout keyboard. Auxiliary power outlets and a printer port are also provided. While in operation, cooling is provided by two fans which circulate air through the cart. Thirty-five foot scanner attachment cables provide numerous options for cart placement during inspection operations. The cart has large, steerable pneumatic wheels and a tow bar to aid in portability.|
|Computer Control System|
|The system controlling software is ULTRA IMAGE IV/Scanmaster Version 9. The software provides the operator with a Windows-like operating environment for all motion control, ultrasonic, data analysis, and reporting functions. The computer is 120 MHz Pentium-based with a 540 MB hard disk. The ultrasonic system supports three channels with fully programmable pulser, receiver, and gate parameters. The system is capable of real-time display of A-scan, B-scan, and C-scan information of all three channels. Each channel can be programmed for A-scan waveform capture. All inspection results are completely archived on hard disk or on a 1 GB JAZ drive provided for external data storage.|
|Tandem Transducer Head|
A tandem transducer head (See Figure =>) is used to hold two opposing shearwave, pulse-echo transducers. Each transducer is separately gimbaled to allow travel over minor surface irregularities without loss of coupling. Transducer spacing is adjustable, which allows ultrasonic sound paths to be optimized for shear wave inspections. The transducer housing is ported to provide couplant flow to the inspection surface. The housing also allows the transducer to be recessed so that it slides on a thin film of couplant. The holder also provides two inches of vertical displacement by means of an integral pneumatic actuator which provides slight contour- following capabilities. Air pressure provided by the actuator maintains constant transducer pressure against the inspection surface. The transducer head is configurable to a single or triple transducer arrangement. |
FIGURE 8. Transducers in scanning position on-aircraft.
FIGURE 9. System On-Aircraft.
The equipment was delivered to the hangar and forklifted onto the wing stand. Figure 9 shows the scanner and cart in place during scanning. Three spanwise splice-joints, #'s 3, 5 and 7, were arbitrarily selected for inspection. Splice-joints #3 and #5 span the length of the inner wing and contain approximately 300 fasteners each. Splice-joint #7 is slightly shorter and contains approximately 260 fasteners. Each splice-joint was laid out and marked, in accordance with the preliminary inspection procedure, to define the scan areas and to facilitate scanner placement. The equipment was energized and scanning started on splice-joint #3. A previously stored calibration scan file was used for ultrasonic parameters. Joints #5 and #7 were subsequently scanned in the same manner. On-aircraft scanning was completed over a period of three days. A total of 51 scans was performed, with a total of 850 fastener sites inspected. An average of 160 holes were inspected per hour. Twenty to twenty-four fasteners were inspected per set-up. Scanner positioning, scanning and data analysis took approximately 7 minutes per scan. The original goal of the program was 100 fasteners per hour.
Scan Image From Splice 7 During Functional Test.
Sample images of the inspection results are given in Figures 10 and 11, which depict the analysis mode of the inspection process. Figure 10 is an image from the scan of Splice 7, Segment 4, holes 18 through 34. Both channels' data have been pasted together; therefore, two rows of fastener site images appear. All fastener sites are displayed within the scan area, indicating complete coverage. Figure 11 is a zoomed view of Figure 10. Hole #'s 7420 and 7421 are shown; Hole 7420 has a crack indication as noted, which was subsequently confirmed with BHEC. After the image is zoomed, the operator can pan and scroll the entire scan length. Transducer location coordinates are maintained, with the image so that any point along the scan can be correlated back to a specific location on the aircraft.
A total of seven indications were detected. An area of what appeared to be corrosion was also identified. Five of the seven fasteners where indications were noted were removed so that a confirmation bolt hole eddy current inspection could be accomplished. Two of the indications were confirmed as cracks with bolt hole eddy current. Further investigation is pending on the source of the remaining three indications. Several areas of procedural improvement were identified and remedied.
TIEDM briefed the Ad Hoc Team in August '96 on the status of the ongoing spanwise splice joint inspection development. Based on the maturity of the technologies to be used for the splice-joint process and favorable inspection results already achieved, the Ad Hoc Team determined the technical risk of failure of the process was low. The Ultra Image process was approved by the team as the long-term NDI solution for C-141 splice-joint inspections. The team suggested a Probability-of-Detection (POD) study be performed on the inspection process and approved the expenditure of funds on a POD study effort.
A modification was made to the existing SAIC contract, requiring the development and performance of a POD study. An initial POD Technical Coordination meeting was held in October '96, with attendees from WR-ALC/LJ/TI, SAIC, Lockheed/Martin, and Sandia National Laboratories. SAIC presented an initial outline of the POD study plan. All in attendance approved the basic elements of the proposed POD plan.
SAIC will use FAA Report No. DOT/FAA/CT-92/12,I "Reliability Assessment at Airline Inspection Facilities, Volume 1: Generic Protocol for Inspection Reliability Experiments" (Ref. 5) as a guideline in the design and execution of the Spanwise Splice POD Study. The generic protocol will be modified specifically for the C-141 spanwise splice-joint inspection process. SAIC subcontracted with Sandia National Laboratory and its FAA Airworthiness Assurance NDI Validation Center as consultants on the POD study. Specific elements of the POD Plan are detailed below:
FIGURE 12. Typical POD Test Specimen. Second Layer Only Shown.
Twelve to fifteen segments will undergo cyclic fatigue testing to obtain a 4:1 uncracked-to-cracked fastener site ratio. Approximately 60 total crack sites will be propagated to cover a crack length range of 0.030" to 0.250". Some larger cracks will also be included, even one or two cracks to the tab edge. Distribution of the number of cracks and crack lengths in each segment has been statistically determined. Uncracked segments will also be included in the experiment.
Specimen Characterization: Each crack will be completely characterized and documented for length and orientation with a combination of high magnification optics (during and after cracking) and contact and immersion ultrasonic imaging techniques. Following initial characterization, the splice segments will be sealed, assembled, and painted at WR-ALC with established C-141 maintenance procedures and materials. A second characterization will be carried out with immersion ultrasonic inspection following assembly.
Calibration Standards: A calibration standard has been manufactured from a typical spanwise joint segment (as shown in Figure 12) using EDM notches to simulate second-layer fatigue cracks. EDM notching allows closer control of crack lengths, orientation, and surface condition. Notch lengths in the standard range from 0.030 inch to 0.100 inch. This standard will be used to calibrate the equipment prior to conducting the POD inspections and will insure the consistency needed from calibration to calibration. The calibration standard will be assembled and painted in the same manner as the cracked specimens.
POD Testing & Variables: Since the splice-joint inspection process is automated to a great extent, human and environmental factors will not be included in the POD study. Initially, a Laboratory Baseline of the process will be conducted to characterize the detection reliability under extreme favorable conditions with an extremely skilled operator. This exercise will focus on procedural and process variables that will require inspector input or adjustment during the course of the POD study. Variables identified to be included in the laboratory baseline are: transducer, receiver gain, calibration standard, couplant pressure, gate parameters, probe pressure, and scanner misalignment.
Actual POD inspections will take place on crack specimens mounted overhead to simulate the C-141 lower inner wing. An actual full length C-141 wing panel will be mounted on a framework to simulate work on a typical C-141 wing stand platform. The crack specimens will be attached to the wing panel to allow specimen placement to be varied along the wing panel length. The inspector will mount the scanner to the wing panel and inspect the crack specimens in the same manner as would be accomplished on an aircraft.
Inspectors & Training: Approximately 15 inspectors will participate in the POD experiment. The inspectors will represent a cross-section of experience and training to determine if and how these factors affect inspection results. The inspector pool will be a mix of the following: Six experienced equipment operators will receive only procedural specific training (3 days); Six Air Force NDI technicians with no ultrasonic scanning/imaging equipment experience will receive extensive equipment operation training plus procedural training (2 weeks); Three Air Force NDI technicians with limited ultrasonic scanning/imaging equipment experience will receive equipment operation refresher training plus procedural training (1 week). Specific training details and curriculum will be detailed in the final plan.
POD Protocols: The protocol package will detail all aspects of the experiment to ensure consistency throughout all phases. The protocols are comprised of step-by-step procedures and descriptions which control the following: Monitors Duties, Inspectors Duties, In-briefs & Out-briefs, Specimen arrangements & control, Data Recording, Data Analysis.
Data Analysis: Data will be presented using the methodologies as covered in MIL-STD-1823 and will include POD curves and ROC curves. Data will be presented in such a manner that any number of variables affecting the inspection results can be extracted.
Cracks were initially characterized by visual, optical means and ultrasonic immersion testing. Following crack measurement, cracked holes are oversized to remove starter notches. The panels will then be reassembled with sealant and the appropriate size taper-lok fasteners and painted per C-141 standards. A final immersion ultrasound characterization is scheduled for each assembled panel. A complete characterization package will accompany each test specimen.
Typical crack length characterization results are shown in Table 1 for segment #7200. As noted in Table 1, cracks did not initiate in all notched holes. There were also instances in other segments when cracks initiated in unnotched holes. For these reasons, the initial crack distribution plans which specified crack locations and sizes within the segments was adjusted as necessary to achieve the original crack distribution goals. A total of 60 cracks have been generated. Following fastener hole reaming operations to remove the crack starter notches, the final crack length range will be 0.025 inch to 0.375 inch. Cracks are evenly distributed in the forward and aft directions.
Segment 7200 Fatigue Crack Data
|Section & Hole #||Direction||Starter Notch||Hole Dia.||Length Actual||Taper-Lok Dia.||Length Final|
When fleet-wide inspection has been completed, the SPD will subsequently establish a 5-year depot-level inspection interval concurrent with the C-141 Programmed Depot Maintenance (PDM). Once a PDM requirement is implemented, field-level inspection of splice-joints will no longer be required. The new splice-joint inspection process can be accomplished during PDM without any addition of flow-days. Additional splice-joint inspection costs at PDM will be more than offset from savings generated by elimination of the ongoing 120 day interval field-level inspection requirement.
Mr. Roy T. Mullis|
Warner Robins Air Logistics Center
Technology & Industrial Support Directorate
Materials Analysis Team (TIEDM)
Robins Air Force Base, GA 31098
Mr. Timothy MacInnis|
Ultra Image International
A Division of Science Applications International Corporation
New London, CT 06320
For product information e-mail Glenn A. Andrew: email@example.com
SAIC Home Page: http://www.saic.com
For more information see: NDT in Aerospace - UTonline 11/97