Engine Titanium Consortium - Center for Nondestructive Evaluation ...

More documents

Recommendations

Info

Preliminary measurements of backscattered noise levels through the (lower) flat surfaces of the blocks were carried out early in the quarter. Selected results are shown in Figure 9. Although the average noise levels were generally similar from block-to-block, the differences were deemed to be large enough to require further measurement and documentation. In addition, the preliminary measurements revealed that the flat block, which was cut from the forging at a later date than the other five coupons, had been cut out “upside down” (i.e., with “TOP” and “BOT” faces reversed in Figure 8). This error, and the fact that the noise level in the ID portion of the forging increases from “TOP” to “BOT” is responsible for the large difference seen in the right-hand portion of Figure 9 between the noise level in the flat block and those of the other five coupons. Although unfortunate, this fabrication error was not deemed to be a major problem for testing the noise models. The principal noise measurements could be confined to the thicker zones of the blocks, and the flat block could simply be “flipped” when measurements were made. During the quarter, additional noise measurements were made through the flat surfaces of the blocks to: (1) carefully document the effect of microstructural differences between blocks; and (2) to test noise model predictions for inspections through flat surfaces. The measurements themselves, and subsequent analysis to calculate relevant noise statistics, were performed at Pratt&Whitney. Additional analyses and model predictions were then carried out at ISU. This second round of measurements was made using a “characterized” 10-MHz transducer whose effective focal properties had been carefully determined (effective diameter = 0.344”; geometric focal length = 3.95”). Measurements through the flat surfaces of the blocks were made for two water paths: 2.4”, which put the actual focal spot just under the entry surface; and 1.0”, which put the focal spot approximately 0.5” deep in the metal. In each case the probe was scanned over a 3” x 0.5” (radial x hoop) region using a step size of 0.015” in each direction. The noise A-scans acquired at each site were stored for later processing. Each day that noise data was acquired, a reference echo from a #1 FBH in a calibration block was also measured. The reference echo was used to correct for minor differences in measurement system efficiency from one measurement trial to the next. In addition, the electronic noise level was measured so that it could be properly accounted for when noise model predictions were compared with experiment. The manner in which the measured noise level depends upon depth is shown in Figures 10 and 11 for the 2.4” and 1.0” water paths, respectively. In the left-hand panels, the rms noise levels, averaged over the lateral scan positions, are shown for each block. Block-to-block differences in noise levels can be seen which presumably result from variations in the forging microstructure with circumferential position. Also shown in Figures 10 and 11 are rms noise levels predicted using an ISU model. The model assumes a uniform microstructure, and requires ultrasonic velocity, attenuation, and grain noise Figure-of-Merit (FOM) values as inputs. Initial guesses for the model inputs were obtained by averaging values measured in a suite of smaller coupons cut from one section of the same forging (see Figure 8b). The estimated FOM value was then increased by 10% to bring the predicted rms noise level more into line with the average of the measurements on the six noise curvature blocks. In Figures 10-11, there is good agreement between the model prediction and the average experimental result in the central depth region, i.e., away from the influences of the front-wall and back-wall echoes. Also note that the location and width of the focal maximum in Figure 11 is well predicted by the theory. Quarterly Report – January 1, 2002 –March 31, 2002 print date/time: 6/6/2002 - 8:39 AM – Page 43
The average noise characteristics, as reflected by the rms-noise-versus-depth curves, are somewhat different for the different blocks. For each block the following procedure was used to determine a depth-dependent “microstructure difference factor”. (1) The measured rms-noiseversus depth curve was adjusted to remove the (minor) effect of electronic noise. (2) The six rms noise curves (one for each block) were averaged to determine the mean curve for the suite of coupons. This was done separately for each inspection water path. (3) The rms noise curve for a given coupon was divided by the mean curve to determine the fractional deviation from the mean at each metal depth. As shown in Figure 12, these difference factors were found to be nearly identical for the two inspections at different water paths, indicating that they do indeed arise from microstructural variations. (4) The difference factors measured for a given block using the 2.4” and 1.0” water paths were then averaged. In future analyses of grain noise measured through the curved surfaces of the blocks, these difference factors will be used (with the distance scale inverted) to “correct” noise quantities for microstructural differences. Remaining differences in noise properties will then presumable be due to surface curvature effects. From the stored grain noise A-scans, other noise statistics can be readily computed. For example, for one typical block, Figure 13 shows gated-peak noise C-scans reconstructed from the flatsurface A-scan data for four choices of the time gate. As the gate is widened, keeping its center at the same depth, the average noise is seen to increase, as expected. Similar C-scan images were generated for all six coupons, and the average and maximal noise amplitude in each image was tabulated. Results are shown in graphical form in Figure 14-15, where they are compared to the predictions of the ISU noise model (rightmost portion of each plot). The model the does a good job of predicting the average gated peak noise amplitude for each of the four gates. However, the theory tends to underestimate the maximum noise amplitude seen in a given C-scan. The current noise model assumes a uniform “average” microstructure. The actual microstructure of a typical block varies only modestly with axial depth, but varies quite significantly with radial position, as can be seen in Figure 13. In the future the model for gated-peak noise statistics will be modified to account for such microstructural variations. This should improve predictions of maximal gated-peak noise amplitudes. Quarterly Report – January 1, 2002 –March 31, 2002 print date/time: 6/6/2002 - 8:39 AM – Page 44
Page 1 and 2: Engine Titanium Consortium Quarterl
Page 3 and 4: Project 1: Task 1.1: Subtask 1.1.1:
Page 5 and 6: findings for the first billet studi
Page 7 and 8: measured P(L) curve, and tabulate t
Page 9 and 10: 1.2 1.0 0.8 V-DIE-A coupon "D" (axi
Page 11 and 12: Attenuation (dB/in) . Waspalloy Att
Page 13 and 14: Deliverables: Fundamental propertie
Page 15 and 16: Small Diameter Billet Assessment: A
Page 17 and 18: Under normal circumstances one woul
Page 19 and 20: Plans (April 1, 2002 - June 30, 200
Page 23 and 24: system will be at GE, and evaluatio
Page 25 and 26: fixed-focus transducers that were n
Page 27 and 28: Back-Wall Scan FBH scan 20% 57% 49%
Page 29 and 30: 63% Back-Wall Scan 74% FBH scan 58%
Page 31 and 32: Figure 6. Noise and target amplitud
Page 33 and 34: Original Date Revised Date Descript
Page 37 and 38: Objective/Approach Amendments: Obje
Page 39 and 40: (a) Transducer is: 15 MHz 0.5”-Di
Page 41 and 42: FOM (std. units) 0.20 0.18 0.16 0.1
Page 43: Peak> Range of Ave. Range of Peak G
Page 47 and 48: Noise C-scans through the “Flat S
Page 49 and 50: 4.0” Convex Block; Flat Side Insp
Page 51 and 52: % FSH Noise Measurements from flat
Page 53 and 54: Original Revised Description Status
Page 57 and 58: possible candidates. Development wo
Page 59 and 60: Forging Calibration Standard #1 1.7
Page 61 and 62: Surface Finish Studies The plans fo
Page 63 and 64: Figure 4. Beam diameter determinati
Page 65 and 66: size was adequate to support a 0.04
Page 67 and 68: phased array instruments. The desig
Page 71 and 72: Task 2: Task 2.1: Subtask 2.1.1: In
Page 73 and 74: Objective/Approach Amendments: Obje
Page 75 and 76: Figure 2. A preliminary design of t
Page 77 and 78: Publications and Presentations: Dat
Page 79 and 80: Survey of common practices (industr
Page 81 and 82: include air drying, oven drying, an
Page 83 and 84: aluminum oxide blasting is conceali
Page 89 and 90: the methodology. These may be thoug
Page 91 and 92: model parts across 17 micrograph pl
Page 93 and 94: Z Y X Figure 4. Angle View of Void
Page 95 and 96:
Original Date Revised Date Descript
Page 97 and 98:
A report documenting the relative e
Page 99 and 100:
Project 3: Task 3.1: Subtask 3.1.2:
Page 101 and 102:
naturally occurring flaws in forgin
Page 103 and 104:
Milestones: Revised dates of TBD ar
Page 105 and 106:
Original Date Revised Date Descript
Page 107 and 108:
Project 3: Task 3.1: Subtask 3.1.3:
Page 109 and 110:
7a-7l 6 8 9 10 Key additional ingre
Page 111 and 112:
incorporating the results of the pr
show all

Engine Titanium Consortium - Center for Nondestructive Evaluation ...

Create successful ePaper yourself

Delete template?

Save as template?