1、01FTM3Automated Spiral BevelGear Pattern Inspectionby: S.T. Nguyen, A. Manesh, INFAC/IIT Research Institute,K. Duckworth and S. Wiener,Honeywell Engines and SystemsTECHNICAL PAPERAmerican Gear ManufacturersAssociationAutomated Spiral Bevel Gear Pattern InspectionS.T. Nguyen, A. Manesh, INFAC/IIT Res
2、earch Institute, K. Duckworth andS. Wiener,Honeywell Engines and SystemsThestatementsandopinionscontainedhereinarethoseoftheauthorandshouldnotbeconstruedasanofficialactionoropinion of the American Gear Manufacturers Association.AbstractSpiralbevelgearsaretypicalcomponentsfoundinmostgasturbineengines
3、thatareusedinawidevarietyofmilitaryandcommercialapplications. Thesegearsareamongthemostdifficultandcostlycomponentstodevelopandmanufacture.Manufacturingprocessesrequiredtoproducespiralbevelgearsarehighlyoperatorintensive,makingthemparticularlycostly in todays small lot production environment. Compou
4、nding these problems are requirements to producereplacementpartsforoperationalsystemsthathavebeenoutofproductionformanyyears.Thisisparticularlytrueiftheoriginal equipment manufacturer (OEM) no longer supports the system.To overcome these multiple issues, the gear production center at Honeywell Engin
5、es 2) multiple loops of trial and error machining; and 3) inspection based on physical master gears, inconsistent test equipment and operator interpretation-based evaluations yielding Pass-Fail designations. In the existing gear design process, the design engineer selects initial gear set parameters
6、, (i.e. gear ratio, pitch, pressure angle, etc.) based on the specified performance requirements for the gear train. Based on the information, a load bearing analysis is performed using a computer model. This computer model also generates an “undeveloped summary”, which is a set of kinematic instruc
7、tions for specific generators and grinding machines. This undeveloped summary is used as a starting point by manufacturing engineers to manually produce a bevel gear. However, the undeveloped summary typically does not produce satisfactory bearing patterns and requires extensive iterative developmen
8、t to obtain a good gear set. All the gear teeth have to be cut in order to evaluate parts on the existing inspection equipment. This is a costly and a time consuming process. In aerospace gearing there are currently no fully automated methods to provide feed back of error conditions and generate cor
9、rections when errors are found in the first trial cuts of spiral bevel gear teeth. Identification of proper machining changes is dependent upon the skill of the engineer or operator and their interpretation of the design intent. The process for pattern, run-out, and backlash inspection using industr
10、y-standard Gleason single flank roll-testers still involves the interpretation of colored marking compounds on mating gear teeth. The roll test requires that a working master gear be meshed with a production gear on a gear tester as shown in Figure 1. Prior to running, the gear-marking compound is a
11、pplied lightly to the gear and pinion teeth. The machine can be hand cranked or motor-driven with a light load applied to the gear. The resulting contact pattern, Figure 2, on the gear and pinion flanks due to the surface contact between the mating teeth and the gear compound are examined. The horiz
12、ontal and vertical (H and V) offsets to the pinion axis with respect to gear axis are also adjusted to span the maximum and minimum allowed per the specification. Contact patterns are then compared against the Bevel Gear specifications to ensure that they are within the acceptable range. This type o
13、f subjective evaluation of pattern comparison makes it difficult to identify a consistent method for adjustments. Since the industry-wide production acceptance methodology requires a sample of each bevel gear production lot to be mounted and tested in a similar fashion described previously, variatio
14、n in gear quality are possible. Figure 1. Universal Gear Tester Figure 2. Zerol and Spiral Bevel Gear Mating Contact Patterns Are Difficult to Interpret Properly. A master spiral bevel gear is an inspection tool that has most of the features of a production gear. It is produced on the same type of e
15、quipment used to manufacture the production gear design. It is a control tool that defines the shape of the gear teeth. The specific elements 3being controlled are the tooth flanks and tooth thickness. All subsequent production members are to duplicate the master control gear in the area of contact
16、pattern and size. The master gear and master pinion of a gear set are made to be identical to gears which had been developed, benched tested, and proven to provide the best possible meshing conditions in an actual gear box mounting and tested through an appropriate range of loads and temperatures. T
17、he terminology for master gears varies among companies, but commonly there are three tiers of physical Master Gears. They are Grand Master, Surveillance Master and Working Master. Working Masters are generally used for production roll testing while Surveillance Masters are used for periodic calibrat
18、ion of the Working Masters. A Grand Master is used exclusively to calibrate Surveillance Masters. Objective: The objective of this project was to significantly advance the state of the art in US Aerospace spiral bevel gear production development, manufacturing and inspection. Specifically, the proje
19、ct developed a closed-loop manufacturing process that would reduce development times for new designs, implemented quantitative inspection system, reduced variation and reduced development and production cost. Additionally, the use of digital electronic master data for acceptance of production bevel
20、gears in lieu of physical gear master was developed and topography-based tolerancing limits for new zerol and spiral design was established. Approach: To accomplish the objective, tight integration of hardware and software capable of closing the loop from designing, manufacturing, inspecting to corr
21、ecting spiral bevel gear was deemed necessary. Once completed, a process that uses a digital gear master for production part acceptance instead of the physical gear master could be developed. In contrast to the existing gear manufacturing process described in the Background Section, it was desirable
22、, through the use of gear designing software, to be able to generate spiral bevel gear data model and automatically convert it into machine instructions after gear loads and contact pattern are successfully simulated. The machine instructions would then be downloaded directly to a gear generator or
23、a grinder for tooth generation. All inspections would be completed on an automated gear inspection system (AGIS) against an electronic master database thereby providing quantitative measurements. Through a separate software module, corrections for discrepancies due to machining inaccuracies would be
24、 linked back to the first part cutting instructions to permit closed-loop corrections. The closed-loop correction will ensure that all parts produced after the first article acceptance will meet the design intent. The desired end state system and the automatically information flow are depicted in Fi
25、gure 3. In order to develop a process that uses digital gear master for production part acceptance instead of the physical gear master, the following issues need to be addressed: 1) How does the bearing pattern simulated by the software compare with the bearing pattern generated from the actual roll
26、ing test for a given design? 2) What inspection outputs from the AGIS inspection system need to be controlled to ensure the desired bearing pattern? 3) Based on an existing Bevel Gear Specifications, how do the tolerances allowed on the bearing pattern tapings translate back as tolerance limits on t
27、he identified output controlled parameters? Figure 3. Information Flow Through Closed-Loop System. A multi-phase approach, which involves system evaluation and selection, system set-up, 4production readiness evaluation, and electronic inspection qualification, was used to select the appropriate equi
28、pment and to develop the closed-loop process for manufacturing of spiral bevel gears described above. For part acceptance using digital gear master process development, computer simulation was used. The approach of actually machining various bearing pattern deviations was deemed too time-consuming a
29、nd costly to be performed. The simulation was structured in a Design of Experiments (DOE) format. Since the computer model describing the gear tooth surface was based on actual machine settings, a wide array of those machine settings was selected as the initial input factors for the DOE. Computer-ge
30、nerated bearing patterns were selected as output factors. Highly skilled inspectors were used to evaluate these bearing patterns to rate and rank them in several groups of acceptable or unacceptable conditions. A second set of simulations was also performed, in which the modified computer model for
31、the DOE was compared to the nominal baseline-model to simulate an AGIS inspection. Using the same data for creating simulated bearing patterns and simulated topography inspections it was possible to establish a correlation between current visual inspections and proposed AGIS inspections. The DOE was
32、 executed in several steps, summarized as follows: Ga7G20 Step 1: Reverse Engineer Grand Masters To Produce Perfect Electronic Master Ga7G20 Step 2: Create Modified Summaries for DOE execution Ga7G20 Step 3: Perform Tooth Contact Analysis (TCA) Based On Modified Summaries Ga7G20 Step 4: Rate Simulat
33、ed Bearing Patterns using ordinal ranking system Ga7G20 Step 5: Simulation of AGIS Inspection of the Modified Summaries Ga7G20 Step 6: DOE Analysis The results of the visual evaluation and the topography simulation were arranged in a matrix format for analysis, along with other crucial DOE informati
34、on, such as the corresponding input factors and part numbers. A typical output is shown on Figure 4. Figure 4. DOE Outputs Used In Analysis. In order to maintain consistency with current inspection methods, it was decided to use existing Honeywell Engines 60% for gear generator and 80% for bevel gea
35、r grinders Ga7G20 80% reduction in inspection time Ga7G20 Elimination of visual inspection and Working Masters Ga7G20 Elimination of hard tooling cost and maintenance Additionally, the fundamental difference between the previous inspection processes and the new closed loop inspection is the requirem
36、ent that all of the gear teeth are cut prior to inspection. The new process requires only one tooth to be formed for in-process set-up inspection. For final inspection, a finished part will be inspected on three evenly spaced teeth. Scrap costs are reduced for two reasons: Ga7G20 The requirement tha
37、t all teeth be cut prior to any checks leaves minimal room for adjustment. Any significant change in cutter path will cause the gear under test to fail, as there is minimal stock available for recutting or reforming. Ga7G20A first article setup test gear can be used in the new process. Since only on
38、e gear tooth is cut, these gears can be used for numerous setups before being scrapped. Also, if only one tooth is cut, there is ample material to permit larger cutter path changes. Conclusions: This project successfully identified equipment and developed a process that would allow for the closed-lo
39、op manufacturing and in-process inspection of spiral bevel gears for aerospace gear applications. Additionally, the process of using digital master for gear acceptance was established. Based on this study, the following conclusions were reached: Ga7G20 By having the capability to generate bevel gear
40、 data model and automatically convert it into machine instructions after gear loads and contact pattern are successfully simulated, precise gear-machining set-up is now possible. Ga7G20 Because of the precise gear-machining set-up, the closed-loop gear manufacturing process is capable of making prod
41、uction hardware equivalent to Master Gear quality. Ga7G20 Only one tooth is needed to generate machine instructions to match the theoretical gear data as opposed to a full gear required in the existing gear manufacturing process. Ga7G20 Using theoretical data as a baseline, the coordinate measuring
42、machine (CMM) can be used to determine the deviations between the theoretical geometry and the actual gear resulted from the gear machine settings. By using this approach, the in process inspection is based on computer-controlled measurements instead of human interpretation by the red lead pattern i
43、nspection system. Ga7G20 Closed-loop manufacturing of spiral bevel gear is feasible with significant saving in cost while providing a more precise gear. Ga7G20 It was shown that the simulated bevel gear-bearings match actual production parts. Ga7G20 Based on this study, bevel Master Gears should be
44、mapped using a minimum of 15 x 15-grid resolution. The inspection area should cover the entire active flank of the gear tooth. The inspection area was established by reducing the theoretically possible flank in two ways: o Lengthwise reduction of face width by the maximum allowable amount for profil
45、e edge breaks. o Tooth height reduction by the maximum allowable edge break for the tooth top land and maximum producible and/or allowable root fillet radius. Ga7G20 The results for the Grand Master inspection against the respective 13Electronic Master should produce a Sum of Squared below 25 tenths
46、 for a 15 x 15 grid. This ensures only about 10 percent of the total available manufacturing tolerance would be used by the inspection system. Ga7G20 The ratio between limit values for Pressure Angle Deviation (d) and Spiral Angle (d) Deviation was consistent throughout at: d / d = 4 / 1. Ga7G20 By
47、using digital gear master in lieu of physical gear master for parts acceptance, the supporting infrastructure of physical gear masters, calibration tracking system and the universal gear tester could be eliminated. Recommendations: It is recommended that the newly developed close-loop system be impl
48、emented for spiral bevel gear manufacturing to eliminate the manual, iterative, trial and error approach now used. Companies that are not in a position to upgrade their current grinder or generator will still benefit from the qualitative inspection and design improvements by implementing the open lo
49、op process described in the Experimental Results Section. The tolerancing limits established based on this study should be a good starting point for digital gear inspection. However for parties that are interested in establishing their own limits, the DOE and computer simulation approach could be used as a template to conduct the study. Companies interested in the designing and manufacturing of spiral bevel gears should review the information contained in this report to determine what portion of the closed-loop manufacturing technology that could be incorporated at their facility to
