Simultaneous Increase of Productivity, Capacity and Quality in NDT of Engine Components Alona Miltreyger Israel, elena@scanmaster-irt.com Michael Bron Israel, misha@scanmaster-irt.com Steve Saitowitz Israel, shmuel@scanmaster-irt.com ScanMaster Systems Ltd. Israel 1
2 Contents Contents... 2 List of Tables... 2 List of Figures... 2 1 Abstract... 3 2 Introduction... 3 3 Case Study... 4 4 Methodology for achieving improvements... 5 5 Results... 8 6 What s next... 8 7 References... 9 List of Tables Table 1 Time consuming disc inspection procedures... 5 Table 2 Comparing time consuming for regular and spiral inspection modes... 6 Table 3 Comparing time consuming for regular and constant linear inspection modes... 6 List of Figures Figure 1 Typical Jet-engine disk and its Ultrasonic shape [1]... 4 Figure 2. Spiral inspection trajectory.... 5 Figure 3. Pixel arrangement for constant linear speed inspection mode... 6 Figure 4 Calibration blocks arrangement for automatic sensitivity calibration... 8
3 1 Abstract The aerospace industry has experienced a rapid ramp up in production over the last few years and this ramp up will increase in the coming years. The demand for component supply especially engine disk and related hardware, is forcing manufactures to find ways to address the bottle necks created at the Ultrasonic NDT stage of production and also maintaining high standards of quality. Increasing capacity does not necessarily increase production or quality. Finding ways to get more out of available equipment and personnel does increase production and with the right technology improves quality. It is not enough to only automate. Automation does undoubtedly speed things up however the actual increase in production is still not at a level most manufacturers would like. Therefore technology that automates, allows for employees to be more productive, improves quality and increases capacity is the key to dealing with the expected ramp up in an expedient and cost saving manner. ScanMaster s developments over the last 8 years has been focused in this direction achieving positive results and enormous production increases, higher quality inspections and tremendous cost savings. These developments include improving the following: Transportability of part programs between multiple systems after importing geometry from existing CAD drawings Automation of every day procedures such as transducer normalization, sensitivity calibrations and material attenuation Automation of adaptation of sound beam entry on surfaces High speed data collection and automatic file compression. Easy transducer change mechanisms for multi-zone type inspections. Solutions for automatic loading/unloading of the inspected parts Sophisticated algorithms to auto evaluate captured data while working according to OEM procedures and specifications. Rapid Data exchange solution Unattended scanning: one operator running multi systems. These developments result in increased capacity, quality, and production, as well as reduced overhead costs. 2 Introduction NDT of Engine Components has passed several development stages during last decades. While the ability to recognize the defects in critical aircraft components was always a priority, the methods used for this purpose have differed. Starting with manual ultrasonic inspection of discs and blades where most of the hardware used for this purpose was analogue instrumentation with a single probe handled by the operator, NDT has advance to fully automated systems controlled almost entirely with software. Auditing of the inspection procedures, reliability, and operator level of experience was key factors to ensuring not only for the quality of the inspection but also for inspection productivity. The progress in different fields such as electronics, CNC, software allowed bringing to the field new type of inspection based on different Cartesian robotic systems. These systems allowed assuring inspection of almost 100% of the volume of the inspected objects with relatively high reliability and inspection speed. Not having automatic registration and evaluation means, these systems
4 relied on the operator to recognize possible defects. This made an operator role even more critical and challenging. Despite very big increase in inspection productivity, the quality and inspection time remained a function of the operator s personal abilities. Most Jet-engine OEMs invest a lot into setting up correct procedures and specifications, however carrying out these procedures were still 100% human-dependent. Figure 1 Typical Jet-engine disk and its Ultrasonic shape [1] Introduction of computer-based Ultrasonic inspection systems with ability to present and store results of inspections allowed moving the focus from the operator to the system. Using different methods of data storage at high speeds for example 100-500MHz while using data compression, and view conversion from analogue to digital A-scan, B-scan and C-scan provides for new possibilities not only for inspection results traceability, but also for the automation of the different stage of the inspection process. Automated inspection processes relieve the human operator from routine tasks, without relieving them from the overall responsibility of decision making. 3 Case Study A case study was conducted in order to determine the most time consuming and critical routine procedures performed by the operator as a part of the ultrasonic inspection process of jet-engine components namely discs. Procedure Duration Frequency Notes Inspection plan 1-2 days Once per new part Using teach-in preparation and any part version procedure Ultrasonic transducer normalization, sensitivity calibrations and material attenuation 15 45 minutes depending on the relevant specification change Once per shift Highly depends on operator experience Verification of proper transducer positioning on each inspected surface 1-5 minutes per surface There are 20-30 surfaces in typical disc shape, so it makes the procedure frequency between 20-30 times per one part Highly depends on operator experience, parts and systems tolerances
5 Transducer change 5 minutes At least once per each inspection zone typical 1-2 times per part. Some multi zone procedures can require from 1 to 12 times per part Data evaluation 5-10 minutes At least once per each surface. Typical part with 20 surfaces Parts loading/unloading Inspection reporting Result 5-10 minutes Two times per part for inspection, once for each side. Requires running normalisation procedure Human evaluation is subject to human errors Includes parts loading and centring 30 to 45 minutes Per part Tedious and prone to human error Table 1 Time consuming disc inspection procedures The inspection process itself was studied separately in order to determine the areas where productivity could be improved without compromising major inspection constrains. The following areas were deemed to have potential for improvement: - Excluding index time, i.e. the time necessary for movement of the transducer from one inspection line to another - Keeping constant linear inspection speed - Evaluation of data during ongoing inspection 4 Methodology for achieving improvements A set of software, mechanical and systems tools were developed and tested. Major attention was given to reducing inspection time by introduction of new modes of motion control during inspection. It was determined that helical scanning using synchronized axes motion eliminates led index time, without compromising inspection coverage. This mode of inspection motion control is mostly suitable for vertical surfaces with the same diameter or with horizontal surface that have vertical sloping. Figure 2. Helical inspection motion.
6 In order to determine the gains in productivity, the test was conducted on a 40 OD surface in two modes of inspection motion regular (scan and index) and helical modes. Radius Start Finish Regular travel Size (in) Time per indexes PRF (Hz) Rotation RPM Latency between Surface Time (Min) 40 40 4 0.025 160 600 55.314 1.085 1 150.2 Radius Start Finish Helical Synchronized Axes Scanning travel Size (in) Time per indexes PRF (Hz) Rotation RPM Latency between Surface Time (Min) 40 40 4 0.025 160 600 55.314 1.085 0 147.5 Table 2 Comparing time consuming for regular and spiral inspection modes Constant linear speed was recognized as a most suitable for inspection of web areas of the disc, since it allows keeping the most optimal inspection speed. The rotation speeds up as the linear axis moves inward from the OD towards the ID The test was conducted on a surface starting from 40 to 36 where the regular mode was compared with constant linear speed: Radius Constant Linear speed Time per Time per Latency Start Finish Rotation RPM at Rotation RPM at between travel Size Surface indexes At Start Start At Finish Finish (in) Time (Min) 40 36 4 0.025 160 55.314 1.085 49.78 1.20 1 142.8 Radius Start Finish Figure 3. Pixel arrangement for constant linear speed inspection mode travel Size (in) indexes Regular scanning Time per Rotation At Start RPM at Start Time per Rotation RPM at At Finish Finish Latency between Surface Time (Min) 40 36 4 0.025 160 55.314 1.085 55.314 1.085 1 147.50 Table 3 Comparing time for regular and constant linear inspection modes
7 The conclusion to be drawn from the above testing is that the correct application of inspection mode can gain about 2-3.5% of time saving per surface/part. Further gains in productivity can be made on horizontal surfaces including sloped or contoured surfaces by combining the synchronized axes helical mode with the linear speed mode. In the table below it is noted that on a 6 long surface a further 2 minutes and 17 seconds were saved.
8 A method of automatic sensitivity calibration (TCG setting) was developed. The method is based on scanning of set of calibration blocks installed in one cassette with the same height. The system doesn t have to make an adjustment of water path for each block. Calibration Blocks with FBH Cassette Figure 4 Calibration blocks arrangement for automatic sensitivity calibration Once the raw data is collected it passes through a special algorithm which generates for each block the relevant sensitivity correction factor per specific depth. This information is either used for verification or adjustment of the TCG sensitivity. The time saving estimation for using of this tool is not straightforward, since the manual calibration procedure is very much defendant on, operator experience, probe quality and specification methodology requirements. Notwithstanding the pre-mentioned parameters the test carried out on set of 12 FBH#1 blocks demonstrate ability to perform sensitivity adjustment/verification within less than 5 minutes compared to typical manual 15 to 45 minutes. 5 Results The above-mentioned developments and others such as mechanical fast exchange solution, automatic data evaluation according to different OEM standards, ability to perform evaluation in parallel to inspection, automated reporting, motorized part clamping and automated parts loading/unloading reduce dramatically the inspection time. Many benefits are realized as a result of this approach: Risk is reduced. People tend to make mistakes especially when they are tired. Thanks to automatic sensitivity setting, automatic evaluation the risk for mistakes at those stages of the inspection process is almost entirely eliminated Improved labor utilization is realized due to the high level of automation and quality controls. This results in one operator running multiple systems. This approach has already been adopted by one of the largest aerospace OEMs and has been implemented in their supply chain. 6 What s next In addition to new developments related to using Phased Array hardware in the NDT ultrasonic process, progress in the field of artificial intelligence, machine learning, etc. opens new
9 opportunities for further improving of engine component inspection systems. Formalization of relevant NDT specifications into the form suitable for artificial intelligence probably will lead to excluding of the operator totally from inspection and evaluation process, resulting in human supervision of the intelligence only. 7 References 1. N. Brierley, T. Tippetts, P. Cawley Data fusion for automated non-destructive inspection, Proceeding of The Royal Society, Published 14 May 2014.DOI: 10.1098/rspa.2014.0167 2. CHARLES J. HELLIER In the Beginning... A brief history of the technology of nondestructive testing. American Welding Society, (https://app.aws.org/itrends/2005/07/021/) 3. See more info at: http://scanmaster-irt.com/industries/aerospace-industry/