Continuous development of the national standard for vibration

Transcription

1 Continuous development of the national standard for vibration by Ian Veldman, National Metrology Laboratory, CSIR In this fast technologically-advancing world, higher measurement accuracy is part of the development criteria. This is also true where vibration measurements are concerned. Huge technological advances are being made using various laser techniques to measure vibrations with sub-nanometre displacements at increasingly higher frequencies. Such laser systems and techniques enable us to measure vibrations in a spot, in grids and planes, all on static and rotating objects. This sophisticated laser equipment allows the user to measure vibrations in one, two or three axis simultaneously, using only a single spot laser beam [1]. As technology advances, these laser systems become smaller and more compact, to the extent where it is possible to manufacture a laser interferometer head consisting of only a single laser diode and a focussing lens [2]. Such small, high end technology devices have numerous new applications, for instance the accurate description of the 3D vibration of the middle ear [3]. that best suited the requirements as the national vibration standard was laser interferometry. The first interferometer system implemented used the well established ratio counting technique as described in ISO method 1 [4]. This system provides the NML with primary accelerometer calibration capability (magnitude only) over the frequency range 10 Hz to 800 Hz with a measurement uncertainty of 1%. This was the first step towards satisfying the country s vibration measurement requirements and ensuring that the standard is comparable internationally. Following this initial steppingstone, further system developments were These advances in laser measurement technology has brought about improvements in accelerometer calibration capabilities, which are directly traceable to the international system of units (SI). The available technology at the time 14 Existing technologies - Homodyne laser interferometry - Heterodyne laser interferometry - Heterodyne interferometer with timeinterval analyzer Industry s current accelerometer calibration needs - 1 Hz to 10 khz Industry s future accelerometer calibration needs - 0,1 Hz to 20 khz, complex calibration (magnitude and phase shift) - Shock calibration of accelerometers - Rotational vibration calibration Commercial laser interferometer calibration International trends. Laser interferometer implementation In the early 1990s, the CSIR National Metrology Laboratory (NML) embarked on a journey to realise the national measurement standards on a primary level, where feasible. At the time, the national measurement standard for vibration was realised at a single frequency (160 Hz) using the reciprocity method. Though the reciprocity method is defined as a primary method, this realisation method was inadequate in meeting the vibration requirements of the country and this method is not comparable with international vibration standards as the reciprocity method is not directly traceable to the SI units. initiated. A homodyne laser interferometer system based on method 3 as described in ISO was established. This system and method was decided upon after the careful consideration of: Fig.1: Back to back accelerometer results at 160 Hz. Method 3 of this ISO standard employs the sine approximation method (SAM). The implementation of this system extended the NML s accelerometer calibration capability from 10 Hz to 10 khz with an improved uncertainty of measurement. To be able to realise the vibration standard with such small uncertainties, a new air bearing vibration exciter had to be acquired to complement the interferometer system. Realising the national vibration standard Fig. 2: Single ended accelerometer results at 160 Hz. With this newly-developed system, adequate validation of the system s performance was required before it could be implemented as the national standard for vibration. Such extensive validation is required to ensure confidence in the measurements delivered by the

2 Participating laboratory Acronym Country Metrology region Calibration period CSIR National Metrology Laboratory Physikalisch-Technische Bundesanstalt system, and to ensure that the stated estimated uncertainty of measurement could be achieved. As part of the validation process, the system was compared against similar systems from other national metrology institutes (NMIs). The two main comparisons that was participated in were CCAUV.V-K1 [5] and SADCMET.AUV. V-S1 [6]. CCAUV.V-K1 CSIR-NML South Africa RSA During August and September 2000, the NML participated in the international comparison, CCAUV.V-K1, arranged by the Consultative Committee for Acoustics, Ultrasound and Vibration (CCAUV) of the International Bureau of Weights and Measures (BIPM). Twelve NMIs from the five regional metrology organisations (RMOs), (APMP, COOMET, EUROMET, SADCMET and SIM) participated in the key comparison. Fig. 4: Difference in sensitivity for the phase shift response for both SADCMET September 2003 PTB Germany DE EUROMET August, October 2003 Table 1: List of participating institutes Fig. 3: Difference in sensitivity for the magnitude response for both In the field of vibration and shock, this first key comparison was organised in order to compare measurements of sinusoidal linear accelerations in the frequency range from 40 Hz to 5 khz. Moreover, the CMCs of the NMIs for accelerometer calibration were examined and compared. It was the task of the comparison to measure the charge sensitivity of two accelerometer standards (one of single-ended design (SE) and one of back-to-back design (BB)) at different frequencies and acceleration amplitudes. The charge sensitivity was calculated as the ratio of the amplitude of the accelerometer output charge to the amplitude of the acceleration at its reference surface. The charge sensitivity was given in pico coulomb per metres per second squared (pc/(ms - ²). A calibrated charge amplifier was used to measure the output charge of the accelerometer standards. Comparison results The 12 NMIs measured the charge sensitivity o f t w o d i f f e r e n t transfer standards at 22 frequencies from 40 Hz to 5 khz (third-octave frequency series). The results of the CCAUV. V-K1 are a set of key comparison reference values (KCRVs), their u n c e r t a i n t i e s a n d degrees of equivalence regarding the KCRV a n d r e g a r d i n g a l l the laboratories with respect to one another. Forty four matrices of equivalence were c o m p u t e d. F r o m this complete set of results, four matrices o f e q u i v a l e n c e were selected and d e m o n s t r a t e d b y g r a p h s w h i c h a r e included in the BIPM k e y c o m p a r i s o n database. The results are accessible through the following URL: AppendixB/ KCDB_ ApB_search.asp. The BB accelerometer was given preference because only the back-to-back accelerometer converts the acceleration sensed by the laser interferometer at the reference surface (optically reflecting top surface) into an electrical charge, both quantities being used for the calculation of the charge sensitivity as a measur and in the key comparison. In each case, the calibration results obtained for the BB and the SE accelerometer represented the current calibration capabilities of the participating laboratories for the charge sensitivity of back-to-back accelerometers and single-ended accelerometers, respectively. However, the calibration results for the BB accelerometer represent the measurement capabilities for the physical quantity of acceleration. The systematic deviations of SE accelerometer calibrations at high frequencies as identified by the key comparison have assisted the laboratories concerned in investigating and improving their calibration facilities. At the reference frequency of 160 Hz (specified in ISO :1999), all participating laboratories calibrated both transfer standards with a relative expanded uncertainty (k = 2) smaller than 5x10-3, i.e. the limit specified by the ISO standard as depicted in Figs. 1 and 2. SADCMET.AUV.V-S1 Following participation in CCAUV.V-K1, the CSIR NML s vibration laboratory extended the SAM accelerometer calibration system capability. With this system extension, a complex accelerometer calibration capability was established over the frequency range 10 Hz to 10 khz, as compared to the previous capability of 40 Hz to 5 khz. In a collaboration project with the Physikalisch- Technische Bundesanstalt (PTB), this newly established capability was validated and culminated in the supplementary comparison, SADMET.AUV.V-S1, registered with the (BIPM). This was the first international publication of phase shift accelerometer calibration comparison results between NMIs. Participants Two NMIs from the two RMOs, SADCMET and EUROMET, participated in the supplementary comparison, SADCMET.AUV.V-S1. Task and purpose of the comparison In the field of vibration and shock, this supplementary comparison (SADC.AUV. V-S1) was organised to compare complex measurements of sinusoidal linear accelerations in the frequency range from 10 Hz to 10 khz. Moreover, the complex (magnitude and phase) CMCs of the NMIs for accelerometer calibration was examined and compared. It was the task of the comparison to measure the magnitude and phase shift of the complex charge sensitivity of two accelerometer standards (both 16

3 single ended in design) at different frequencies and acceleration amplitudes [7]. The charge sensitivity was calculated as the ratio of the amplitude of the accelerometer output charge to the amplitude of the acceleration at its reference surface. The reference surface was defined as the mounting surface of the accelerometer. The magnitude of the complex charge sensitivity was given in pc/(m/s - ²), while the phase shift was given in degrees ( ). A calibrated charge amplifier was used to measure the output charge and phase shift of the accelerometer standards. For the calibration of the two accelerometers, both NMIs applied laser interferometry in compliance with method 3 of the international standard ISO :1999, in order to cover the entire frequency range chosen, within a specified range of the acceleration amplitude with specified uncertainties. Transfer standards During the preparatory stage, the PTB thoroughly investigated the characteristics (long-term stability, linearity, etc.) of various reference standard accelerometers considered to be candidates for the transfer standards to be used in the supplementary comparison. The following two accelerometers were selected: Accelerometer A A transfer standard accelerometer; Brüel & Kjær model 8305 WH 2335 Serial number: nominal charge sensitivity (magnitude): 0,13 pc/m/s² Accelerometer B A reference standard accelerometer; Endevco model 2207M8 Serial number: AC08 Nominal charge sensitivity (magnitude): 0,22 pc/m/s - ² A complete measurement series was carried out on different days under nominally the same conditions, except that the accelerometer was re-mounted and the cable re-fixed. The (mean) result of the all the measurement series was given as the final measurement result by PTB. Measurement results In this bi-lateral comparison between the CSIR NML and the PTB, calibrations of the magnitude and phase shift of the complex sensitivity of two reference accelerometers, the SAM method specified in ISO (method 3) was applied in three versions: Version 1: Homodyne interferometer with two output signals in quadrature (i.e. phase-shifted by 90 degrees) Version 2: Heterodyne interferometer with quadrature signals generated by digital data processing Version 3: Heterodyne interferometer with timeinterval analyser Both the CSIR NML and the PTB used Version 1 as specified in ISO , with a modified Michelson interferometer as depictured in Fig. 4 of that international standard, as a subsystem of the calibration equipment. The special techniques and procedures developed at the CSIR NML (standard measuring equipment with vibration exciter, interferometer, data acquisition and signal processing system etc.) are described in detail in [8]. For the calibrations performed at the PTB, versions 2 and 3 were applied in addition. The standard measuring equipment developed at the PTB is equipped with a heterodyne Mach- Zehnder interferometer head in conjunction with sub-systems for frequency-conversion, data acquisition and digital signal processing. The standard measurement equipment is described in [9], [10], [11]. Supplementary comparison reference value (SCRV) The weighted mean was agreed upon by both laboratories to calculate the SCRVs for the SADCMET.AUV.V-S1 data. SCRVs are calculated separately at each frequency point measured (37 points per accelerometer in total). Calculation of SCRVs using the weighted mean method. For each laboratory i the data are x i,f : best estimate of sensitivity at frequency f; and u(x i,f ): associated standard uncertainty of sensitivity reported at frequency f. For each of the two transfer standards and at each frequency f, a SCRV value x R,f was determined as the weighted mean of the results of n laboratories (for this comparison, n = 2) according to (1) (2) The degree of equivalence, DNMI-WM, and UNMI-WM, was determined for the magnitude as well as the phase shift measurements for both accelerometers using (3) where x NMI represents the measurement results 17

4 obtained by the laboratory at each frequency point for the magnitude and the phase shift and x WM represents the reference value (SCRV) calculated as the weighted mean using Eq. (1). U NMI-WM is the uncertainty of measurement associated with the calculated D NMI-WM for k = 2. SADCMET.AUV.S1 conclusions Two NMIs measured the complex charge sensitivity of two different transfer standards (single-ended accelerometer at 37 frequencies from 10 Hz to 10 khz). The results of the SADCMET.AUV.V-S1 are a set of SCRVs, their uncertainties and degrees of equivalence regarding the SCRV and regarding the laboratories with respect to one another. In the calibration of the single-ended accelerometer, the reference surface (mounting surface) is not accessible to the laser light beam. Relative motion between the acceleration acting on and converted by the single-ended accelerometer and the acceleration sensed by the laser interferometer (close to the accelerometer) may have been present. In each case, the calibration results obtained for the SE accelerometer represent the current calibration capabilities of the participating laboratories for the complex charge sensitivity of single-ended At the reference frequency of 160 Hz (specified in ISO :1999), the participating laboratories calibrated both transfer standards with a relative expanded uncertainty (k = 2) smaller than 0,5%, i.e. the limit specified by the ISO standard [4]. In the most important frequency range of 40 Hz to 5 khz, covered by the key comparison CCAUV.V-K1, the deviations between the corresponding PTB and CSIR NML results were smaller than 0,4 for the phase shift measured (22 measurement points) for the Endevco accelerometer. For the Brüel & Kjær accelerometer, the deviations between the Table 2: Previous vibration calibration capability. Table 3: Improved vibration calibration capabilities over the extended frequency range. corresponding PTB and CSIR NML results were smaller than 1,2 for the phase shift measurements (22 measurement points). For the frequency range 10 Hz to 10 khz, the deviations between the PTB and NML results were smaller than 0,6 for the phase shift measurements (37 measurement points) for the Endevco accelerometer. With the Brüel & Kjær accelerometer for the frequency range 10 Hz to 10 khz, the deviations between the PTB and CSIR NML results were smaller than 2 for the phase shift measurements (37 measurement points). The uncertainties calculated for the phase shift difference values of PTB ranged from 0,4 to 0,7. The corresponding uncertainties calculated for the phase shift difference values of the CSIR NML ranged from 0,4 to 1,2. In conclusion, the degrees of equivalence calculated from the data submitted by the two laboratories supports the uncertainty of measurement reported by the two laboratories for the calibration of the complex sensitivities of accelerometers, over the frequency range 10 Hz to 10 khz. Accreditation To underpin the competency requirement for realising an internationally acceptable national standard, an internationally recognised quality system is maintained. The CSIR NML opted for the method of third party accreditation as to the alternative of self declaration. For its laboratory accreditation, the NML contracts the internationally recognised accreditation body, SANAS. During 2001, the vibration laboratory was the first laboratory of the CSIR NML to be accredited by SANAS to ISO [12]. On completion of the validation of the new vibration capabilities, application for an extension of the accreditation scope to include the new calibration capabilities, with improved BMCs was submitted to SANAS. During February 2005 SANAS, with the assistance of a local and an international assessor, performed the vibration laboratory s first full reassessment. During the assessment, numerous vertical assessments as well as witnessing of various calibrations were performed to ascertain the competency of the staff to disseminate the newly established vibration standard at the improved accuracy levels. After the three day assessment, the assessment team concluded continued accreditation as per the newly submitted accreditation schedule. This conclusion was only reached after the careful scrutinizing of supporting documentation of the extensive validation process that was followed by the CSIR NML as well as supporting data of comparison measurement results. Tables 2 and 3 list the previous and current vibration calibration capabilities of the laboratory, respectively. This improved vibration standard enables the CSIR NML to calibrate accelerometers over the additional frequency range of 10 Hz to 40 Hz with a best measurement capability of 0,5%. This improved uncertainty of measurement is also achievable from 40 Hz up to 1 khz. Before the system improvements, no capability existed below 40 Hz and the uncertainty of measurement from 40 Hz to 1 khz was estimated to be 1%. Similar improvements were made on the upper frequency calibration range. The frequency range over which the CSIR NML was capable of providing accelerometer calibrations was extended from 5 khz to 10 khz, with improvements to the BMC of the previous frequency range of 2,5 khz to 5 khz. In the frequency range 2,5 khz to 4 khz the BMC was improved from 3% to 1,0% while the BMC at 5 khz was improved from 3% to 1,8%. Not withstanding the frequency range extensions and BMC improvements, the new vibration standard added a new dimension to accelerometer calibration in South Africa, with the added capability of the phase shift calibration of With this addition to the national vibration standard, the CSIR NML became only the second NMI in the world to add complex sensitivity calibration of accelerometers to its CMC list on the BIPM database. A list of all CMCs is available at asp?reset=1&met=auv. Future developments The current national standard for vibration supports industry s current requirements. Some important traceability areas are not yet addressed: Sinusoidal vibration calibration below 10 Hz Calibration of laser vibrometers Shock calibration of accelerometers Rotational vibration calibration of 18

5 The natural progression for extension of the national standard for vibration would be the development of a sinusoidal calibration facility for accelerometer calibration below 10 Hz. This step in the development of the standard is, though in its early stages, already in motion. Through the progression of the system to the current level, the required knowledge to successfully develop this next stage of the standard was established. What is lacking is specialised equipment and the time to develop and validate the facility. Due to the continued support by the Department of Trade and Industry for the maintenance and development of the national measurement standards, a special low frequency vibration exciter was acquired to complete the low frequency vibration system. For the generation of sinusoidal acceleration levels at low frequencies, very large displacements are required. A displacement of 100 mm (peak to peak) at a frequency of 0,16 Hz, produces a peak acceleration level of only 0,05 m/s2. This produces a low accelerometer output voltage that needs to be measured very accurately. For instance, an accelerometer with a sensitivity of mv/g will produce an output voltage of only 5 mv (peak) with an applied acceleration of 0,05 m/s 2. A further stringent requirement for the vibration exciter system is a frictionless, extremely linear motion over these large displacements. This can only be achieved with air bearing exciter systems. Two options are available to the CSIR NML for measuring the displacement of the vibration exciter. Both systems employ a Michelson interferometer system. Option 1 is a single beam Michelson interferometer using the ratio-counting technique as described in method 1 of ISO Advantages A less complicated interferometer Single optical output Less complicated measurement system a standard frequency counter is required Direct available measurement result. Disadvantages Only the magnitude can be measured, no phase measurement capability. Option 2 is a quadrature laser interferometer system implementing a SAM system as described in method 3 of ISO Advantage Complex sensitivity measurement result. Disadvantages Require a quadrature interferometer system Dual optical outputs High sampling rate analogue to digital card with deep memory capacity is required Measurement results only available after post-processing. The decision was made to implement option 1 as it is less complicated, easier to validate, and will satisfy industry requirements. The implementation of option 1 can also be done in a shorter time frame. Conclusions The national standard for vibration has been continuously developed and improved upon to ensure that the standard adequately supports local industry s vibration measurement requirements. The standard is further maintained at a level that is internationally comparable and acceptable. Through active participation in the BIPM key comparisons, the vibration laboratory has demonstrated that the new system can be used to calibrate accelerometers at the level stated in the laboratory s CMC submission. This is underpinned and supported by the SANAS accredited quality system in compliance with ISO References [1] Stanbridge A.B. et. al Measurement of total vibration at a point using a conical-scanning LDV. Proceedings 2nd International Conference on Vibration Measurements by Laser Techniques, Ancona, Italy, [2] Giuliani G. et. al Self-mixing laser diode vibrometer with wide dynamic range. Proceedings 5th International Conference on Vibration Measurements by Laser Techniques, Ancona, Italy, [3] Decraemer, W.F.S. et. Al, The integration of detailed 3-dimensional anatomical data for the quantitative description of 3-dimentional vibration of a biological structure. An illustration from the middle ear.. Proceedings 5th International Conference on Vibration Measurements by Laser Techniques, Ancona, Italy, [4] ISO part 11, 1999, Methods for the calibration of vibration and shock transducers. [5] von Martens, H.-J. et al, Final report on key comparison CCAUV.V-K1, 2003, Metrologia, 40, Tech. Suppl [6] Ian Veldman et al, Final report on supplementary comparison SADCMET.AUV.V-S Metrologia, 41, Tech. Suppl [7] Technical Protocol of the Supplementary comparison SADCMET.AUV.V-S1 (Vibration). CSIR-NML, C.S. Veldman, July 2003 [8] Veldman C.S., A novel implementation of an ISO standard for primary vibration calibration by laser interferometer, Metrologia 40 (2003), pp [9] von Martens, H.-J., Current state and trends of ensuring traceability for vibration measurements, Metrologia 36, pp , [10] von Martens, H.-J. et. al, Traceability of vibration and shock measurements by laser interferometry. Measurement 28 (2000), pp [11] von Martens, H.-J. et. al, Recent advances in vibration and shock measurements and calibrations using laser interferometry. Proceedings 6th International Conference on Vibration Measurements by Laser Techniques, Ancona, Italy, [12] ISO/IEC 17025, 2000, General requirements for the competence of testing and calibration laboratories. Acknowledgement This paper was presented at the 2005 T & M Conference 5-7 September 2005, in Gauteng, and is republished with permission. Contact Ian Veldman, CSIR, Tel (012) , csveldma@csir.co.za Two-colour pyrometers For manufacturers who rely on temperature measurement to control product quality, noncontact infrared thermometers are essential tools. Raytek offers two-colour technology with their Marathon series. Unlike standard (one-colour) thermometers, the accuracy of ratio (two-colour) instruments does not depend on the target filling the field of view of the instrument. Ratio thermometers are also more accurate than standard instruments in applications where there is dust, steam, or particles in the atmosphere, where the field-of-view of the instrument is partially obscured, or where the emissivity of the target varies. In contrast to one-colour IR thermometers, ratio instruments pass the captured IR energy through two IR filters that limit the IR radiation to two different wavelength bands. The IR radiation for each filter is then directed to two separate IR detectors where it is converted to an electrical signal. Contact Steve Edwards, R&C Instrumentation, Tel (032) , stevee@randci.co.za 19