More info about this article: http://www.ndt.net/?id=21218 Elimination of Pneumatic Noise during Real Time Acoustic Emission Evaluation of Pressure Vessels Binu B*, Yogesh, Praveen.P.S, S Ingale, KK Purushothaman, Isac Daniel & J Philip Structural Testing Group Structural Engineering Entity, Vikram Sarabhai Space Centre, Indian Space Research Organization, Thiruvananthapuram, Kerala, India 695 022 (* corresponding author :b_binu@vssc.gov.in) Abstract Acoustic Emission (AE) technique, being a real time global NDT method, is a potential tool for the detection of active flaws and can provide the full coverage of the pressure vessels. However, AE monitoring currently has difficulties during pneumatic pressure testing of components. Continuous noise signals with maximum duration are generally recorded by the system during pneumatic charging of the tank. Since such noise corrupts AE genuine signals, the structural integrity evaluation of the hardware in real time and offline becomes difficult. It calls for a suitable method to eliminate this pneumatic noise during testing. Different methods, as described in this paper, have been attempted which provide encouraging results. This paper reports the cause for the generation of pneumatic noise, its implications during AE signal interpretation and details of the proposals for its elimination. Keywords: Acoustic Emission (AE), Pneumatic noise, Helmholtz resonance Introduction Acoustic emission (AE) is widely used in the aerospace industry for the real time structural integrity evaluation of metallic pressure vessels used especially in launch vehicles. These pressure vessels do have discontinuity due to fabrication deviations which can propagate under loading and can cause catastrophic failure during operation. All the pressure vessels used in launch vehicles are normally acceptance tested before being put to service. Since Acoustic Emission (AE) is one of the most sensitive global NDT techniques used for remote monitoring, it is widely employed during the proof testing of 457 Non-Destructive Evaluation 2016
pressure vessels [1]. During proof pressure tests of such hardware, continuous monitoring provides an increased reliability of behavior of discontinuity with the help of AE technique. Moreover, incipient failures can be detected and tests can be aborted in real-time to salvage the hardware and the facility due to bursting. AE can also locate the degraded locations of the hardware during post-test analysis. These locations can be cross verified with the help of other conventional NDT techniques and even repair schemes can be attempted resulting in huge cost saving. Despite the advantages, successful use of AE technique for structural integrity assessment applications has several challenges. One of the major challenges is the presence of noise signals which can directly affect the diagnosis results [2]. Noise is any unwanted disturbance within a useful frequency band, such as undesired electric waves in a transmission channel or device. Noise may be erratic, intermittent, or statistically random. Noise is further defined as electric/ electromagnetic noise and acoustic noise as shown in Figure 1. 1a. 1b. 1c. Figure (1a) electromagnetic noise (1b) electric noise and (1c) narrow-band harmonic/ acoustic noise Acoustic noises are unwanted signals generated by (i) extraneous noise, (ii) loading fixture noise and (iii) internal pressurizing fluid noise. The turbulence caused by the flow of a pressurized fluid through an orifice produces energy waves of both sonic and ultrasonic frequencies due to cavitation as shown in Figure 2. Fluid noise may be generated when orifice size and fluid flow velocity form an effective Helmholtz resonator [3]. Helmholtz resonance is the phenomenon of air resonance in a cavity or chamber that contains a gas. When fluid is forced through a cavity to a chamber, the pressure inside increases. Due to the formation of vortex and turbulence the cavity will be left at a pressure slightly lower, causing the fluid to be drawn back in. This process repeats and causes the oscillations of fluid at the neck [4] which in turn generates the acoustic Figure 2 Fluid flow through orifice emission signals. The Aluminium pressure vessels are tested with pneumatic media to avoid contamination and corrosion [1]. The high pressure, high velocity gas is let in to the pressure vessels through an orifice and 458 Non-Destructive Evaluation 2016
the turbulent flow generates Acoustic Emission signals. These are low frequency, continuous type signals with maximum AE duration and are always well above the set threshold level as shown in Figure 3 & 4. Due to the low damping nature of the Aluminium material, all the AE sensors fixed on the pressure vessel will capture the noise AE signals generated with a small decay in the magnitude of AE parameters. Since the amplitude levels are quite high, these signals are observed throughout the testing of the pressure vessels. Figure 3a. Genuine AE signal waveform Figure 3b. Pneumatic Noise waveform Figure 4a. Power Spectrum Genuine AE Figure 4b. Power Spectrum Pneumatic Noise AE data acquisition system records these noise signals with maximum duration, stopping for a hit lockout time (HLT) - a system requisite for signal processing, and hence such signals are continuously recorded by each individual AE sensors. The set value for maximum duration in PAC DISP/ SAMOS Express cards are 1000 milli sec and 350 milli sec for PAC Samos cards. Hence the system will register continuous pneumatic noise signals with duration of 1000millisec for DISP cards and 350millisec for SAMOS cards which indicates only one hit in each 1 second for DISP cards and 0.35 sec for Samos cards as shown in Figure 5. 459 Non-Destructive Evaluation 2016
DISP Card Samos card Figure 5. Pneumatic noise with saturated duration observed with different systems The genuine AE bursts emitted during this period will super impose with this noise and all the characteristics of the genuine AE signals will change due to this and will not register in time. These unwanted AE signals corrupts AE data. Further it makes the hardware difficult to assess its performance through AE data both online and offline. The data acquired from an Aluminium propellant tank is shown in Figure 6 & 7. Figure 6. Aluminium Alloy Propellant tank Figure 7. Typical AE data of Al.alloy prop.tank with pneumatic noise In practical applications it is difficult to discriminate the genuine AE signals from noise if the signal is reduced to a few parameters. The parameters of AE signals are strongly related to the material and the geometry of the structure. The reflected waves also will superimpose the genuine signals and enhancing its complexity. Due to the narrow differences in parameters for signals from various sources the complexity is more in data recorded using resonant sensors [5]. This yields to make use of suitable solution for eliminating this noise during the time of testing itself. Different methods, as described in this paper, have been attempted which provide encouraging results. 460 Non-Destructive Evaluation 2016
Methods attempted The methods attempted so far for the elimination of pneumatic noise and the proposals are described below. a. Pressurization Rate The rate of pressurisation has a clear role in generating the pneumatic noise. Higher rate of pressurisation will create more velocity which in turn increases the turbulence and thereby the noise level. This was demonstrated by the testing of similar type of Aluminium alloy propellant tanks at two different pressurisation rates as shown in Figure 8a & 8b. Figure 8a shows the test with higher rate and Figure 8b with slower rate. The test with higher rate is showing continuous pneumatic noise signals with saturated duration (350milli sec) throughout the testing, while the test with slower rate, the saturation of duration of AE signals are changed at lower pressure itself and the whole pneumatic noise has vanished after certain pressure. This indicates, the use of an optimum pressurisation rate for each type of hardware will help to reduce the pneumatic noise. Figure 8a. Typical AE data with higher Pressurization rest Figure 8b. Typical AE data with slower Pressurization rate b. Orifice Size The change in diameter of the orifice will reduce the velocity of the jet and there by the pressure, which in turn reduce the backflow and thereby reduce the noise. The phenomenon was 461 Non-Destructive Evaluation 2016
demonstrated by the testing of similar type of Aluminium alloy propellant tank by using two different sizes of orifices. The data shown in Figure 9a is of a propellant tank in which a smaller orifice of 14 mm diameter was used and Figure 9b shows the data with higher diameter orifice of 19 mm used during testing. In the first case, the pneumatic noises with saturated duration signals continued throughout the test. While during the second test the pneumatic noise was reduced and vanished at lower pressure. This shows, an optimum size of the orifice for each type of hardware is to be identified for reducing the noise. Figure 9a. Typical AE data - Pressurization through smaller orifice (14 mm) Figure 9b. Typical AE data - Pressurization through bigger orifice (19 mm) c. Software Filters The frequency characteristic between a pneumatic noise and genuine AE signals are distinct. Figure 10 shows the test data of an Aluminium Tank with two different frequency filters. Pairs of AE sensors with two different frequency filters were mounted close to each other at four locations of the hardware. Figure 10a shows the data of AE sensors with an analogue frequency filter setting of 20-400 khz range while the Figure 10b is the test data of AE sensors with analogue frequency filter range of 100-400 khz. Significant reduction in saturated duration AE signal was observed in Figure 10b corresponding to sensors with analogue filter 100-400 khz setting. The result shows the effectiveness of using the filters to reduce the pneumatic noise. 462 Non-Destructive Evaluation 2016
Figure 10a. Typical AE data acquired with Figure 10b. Typical AE data acquired with 20-400 khz filter setting 100-400 khz filter setting d. Diffuser tubes Figure 11 shows the data acquired from a dual compartment Aluminum tank proof pressure test. In actual test condition the compartments experience pressures with a small difference. During the pressurization, the higher pressure compartment gets pressurized through a diffuser tube as shown in Figure 11 and the second compartment is charged directly. Figure 12a shows the charging through diffuser tube and Figure 12b shows the direct charging without a diffuser. The pneumatic noise signals are very minimum during the pressurization through diffuser tube and more for direct pressurization. This is because the diffuser will reduce the turbulence near the neck and thereby the noise generation. Figure 11. Diffuser 463 Non-Destructive Evaluation 2016
Figure 12a. Typical AE data- Pressurization with diffuser Figure 12b. Typical AE data-pressurization without diffuser Conclusion The causes and implications of noise generated during the AE testing of pressure vessels are studied in details. The methods attempted so far improves the capability for the real time structural integrity evaluation pressure vessels during pneumatic testing. Future studies are also planned to improve better evaluation and interpretation of AE data in real time and offline. Reference: [1] Structural Integrity Assessment of Aluminium Liquid Propellant Tanks during Proof Pressure Testing Using Acoustic Emission Technique - B Binu,K K Purushothaman, Annamala Pillai and Jeby Philip, APCNDE 2013. [2] Christian U. Grosse, Masayasu Ohtsu (Eds.) Acoustic Emission Testing. [3] ASTM E750 10, Standard Practice for characterizing Acoustic Emission Instrumentation. [4] Sukanta Biswas and Amit Agrawal, Department of Mech.Engg, IIT, Bombay Noise reduction in a large enclosure using single, dual and ensconced Helmholtz resonators [5] M o gh o Ha, Lehigh U i e sit ou d edu tio a Hel holtz eso ato 464 Non-Destructive Evaluation 2016