More Info at Open Access Database www.ndt.net/?id=15214 A Method for High Sensitive, Low Cost, Non Contact Vibration Profiling using Ultrasound Haneesh Sankar T P 1, a, Subodh P S 2, b, Mathew J Manavalan 3, c 1 Centre for Development of Advanced Computing, Thiruvananthapuram, Kerala, India 2 Centre for Development of Advanced Computing, Thiruvananthapuram, Kerala, India 3 Centre for Development of Advanced Computing, Thiruvananthapuram, Kerala, India a haneesh@cdac.in, b subodh@cdac.in, c mathewjm@cdac.in Keywords: Non contact vibration profiling, Ultrasound, NDT Abstract Vibration profiling is an important tool for Non destructive testing of materials and equipment, since it helps to give an early indication of faults in machinery, continuous health monitoring of bridges, engine vibration profiling etc which ultimately helps in preventive maintenance leading to cost and time savings thereof. A novel method is employed for non contact vibration profiling using ultrasound, where the technology used paves the way for drastic cost reduction without a compromise on the sensitivity of measurement. Existing systems employed for high sensitive measurements, depend on Laser vibrometry, which is not feasible for simultaneous survey of large area, primarily due to the huge costs involved. In this context, a low cost non contact vibration measurement system can considerably widen the range of applications for vibration profiling systems The theory, implementation of the method, tests conducted, results achieved, conclusions and inferences along with the possible application areas are discussed in detail in this paper. The experiments whose results are discussed here were conducted to analyse surface vibration using a target with varying vibration frequency and amplitude, to find out the measurable vibration limits. Experiments were also conducted to analyse the impact of distance between target specimen and the sensor on vibration profiling. Subject of study also encompasses a comparison of the non contact method vis-a-vis the contact method in the above scenarios. The cost effectiveness of the said method, compared to other systems having comparable sensitivities was also found out. I. INTRODUCTION A non contact vibration sensor operating in ultrasound frequencies has been realized and subject to various measurements using a vibrating target for its comparison against a contact sensor. The experiments conducted and the results shared here provide an insight into the applications where ultrasound based non contact sensor can be faithfully deployed for vibration measurements. The most popular solution currently available for non-contact vibration measurements are the devices based on laser technology. These systems rely on a laser, which is beamed onto the surface of interest and the frequency of vibration of the surface is extracted from the Doppler shift of the reflected laser beam. The huge cost of such sensors makes development of array systems based on lasers unfeasible. Another disadvantage of laser method is the distance dependency in measurement sensitivity. The method discussed here relies on ultrasonic technology combined with innovative algorithms employing frequency down conversion to achieve high sensitivity. The frequency down conversion method is suitably adapted for surface vibration measurement. This set up consists of a sensor subsystem, for transmitting ultrasonic signals and reception of the modulated ultrasonic signals, a front end signal conditioning system, processing modules for identification of the modulating signals which corresponds to the vibration of the surface being monitored, a software running on a computer, for the five dimensional surface profiling, the said dimensions being amplitude of vibration, frequency of vibration, x coordinate of points on the vibrating surface, y coordinate of the same points and time. Ultrasound based non contact vibration measurement technique, yields consistent results with high sensitivity.
II. THEORY The signal transmitted to the vibrating surface can be represented by Tx=, (1) A: Transmitted signal amplitude, : angular frequency of carrier The received phase modulated signal, Rx = (2) B: Received signal amplitude, : Phase variation due to the target vibration Rx = (3) C: Vibration amplitude, : angular frequency of vibration signal, : wavelength of the carrier signal Rx = (4) fc : frequency of the carrier signal, fv : frequency of the vibration signal A frequency down conversion process ensues, which results in the down converted signal, Rxd = (5) The vibrating signal (modulated signal) is also fed to another down conversion process using the quadrature component of the carrier signal, creating the quadrature of signal given in the equation 5. Low pass filtering on the down converted signals results in the following signals. RxI = (6) RxQ = (7) Equation 6 and equation 7 is solved to find out the target vibration frequency and amplitude, which is done on a continuous basis, the output of which is used to create a continuous time varying graphical representation of the vibrating parameters spread spatially, corresponding to the sensor position in the array. III. SYSTEM DESCRIPTION Ultrasound Non contact Vibration Sensor Array is an array of sensors, each one of which is capable of transmitting high frequency ultrasound signals to the target vibrating surface, and receive back the modulated signals. The modulated signal is fed to a frequency down converter subsystem, whose twin channel output contains the sum and difference frequencies of its inputs, which is Low pass filtered and fed to the multichannel digitization module for vibration detection. The functional block diagram of the method used is shown in fig 1.
Ultrasound Non contact Vibration Sensor Digitization Processing Visualization & Display Fig.1 Architecture The digitization subsystem is capable of digitizing analog input data, the said system comprising of analog to digital converters and processing elements which contains algorithms to solve the equations corresponding to the low pass filtered in-phase and quadrature components to compute the frequencies and their corresponding amplitude of the vibrating surface. The digitization subsystem comes with a network interface which provides the option of transmitting the captured data over a network to a remote location. Visualization algorithms are implemented on a computer system, the said system taking inputs from the network interface, which corresponds to the real time vibration parameters corresponding to each sensor array, whose spatial position is known, facilitating the creation of a graphical image of the vibration parameters depicted by the frequency which is represented by the color, its amplitude, drawn in the z dimension, the sensor position, represented by the x and y coordinates, the said parameters being varied continuously according to the data received IV. EXPERIMENTS TO EVALUATE FEASIBILITY A series of experiments were conducted to find out the feasibility of applying the ultrasound based non contact vibration measurement in real world. The experiment set up is as shown in the figure Fig 2 showing vibration table, accelerometer, sensor, display Display UNC Sensor System Accelerometer Vibration Table Accelerometer Reading Fig. 2 Test setup A calibrated vibration table was used as the target specimen. The calibrated contact sensor was used a reference for this study. The experiments can be classified into 3, based on the objectives i) to find out the measurable vibration limits, by using a target with different vibration frequency and amplitude ii) to analyse the impact of distance between target specimen and the sensor on vibration profiling iii) Comparison of the non contact method vis-a-vis the contact method in the above scenarios
i) Vibration measurement limits a) Vibration amplitude measurement limits. The target vibration amplitude is varied keeping the vibration frequency constant to find out the vibration measurement range of ultrasonic non contact sensor (UNC) and the error in each measurement. The results are tabulated below. Target vibration frequency = 100Hz, UNC to vibration table distance = 10cm Target vibration amplitude [mg] Table 1 Vibration amplitude measurement range UNCS measured UNCS Measured Amplitude [mg] Error in % < 5 Noisy Noisy - 5 100 6 20 10 100 12 20 50 100 58 16 100 100 110 10 200 100 210 5 300 100 310 3 400 100 408 2 500 100 490-2 800 100 780-3 1000 100 900-10 1200 100 1000-17 > 1200 300 ( not correct) - - Max error in reading = +/-20%, Measurable Target vibration amplitude = 5mg to 1200mg Fig. 3 Variation of vibration amplitude measured using UNCS b) Vibration frequency measurement limits. The target vibration frequency is varied keeping the target vibration amplitude constant to find out the ultrasonic non contact sensors (UNC) vibration measurement capability over a frequency range. Target vibration amplitude = 100mg, UNCS to vibration table distance = 10cm
Target vibration Table 2: vibration frequency measurement range UNCS measured UNCS measured amplitude [mg] Error in % < 100 - - - 100 100 118 18 200 200 114 14 300 300 110 10 400 400 108 8 500 500 104 4 800 800 97-3 1000 1000 95-5 2000 2000 90-10 5000 5000 82-18 10000 10000 81-19 >10000 - - - Maximum error in reading = +/-20%, Measurable vibration frequencies = 100Hz to 10 KHz Fig. 4 Variation of vibration frequency measured using UNCS ii) Target distance Vibration table amplitude = 100mg, UNC to vibration table frequency = 400Hz, UNCS to vibration target distance [cm] Table 3: Vibration amplitude at various target distances UNCS Measured UNC Measured amplitude [mg] Error in % 5 100 105 5 10 100 108 8 15 100 109 9 20 100 112 12 30 100 118 18 Maximum error in reading = +/-20%, Maximum / minimum distance between sensor and target for valid vibration measurements = 5 cm / 30cm
Fig. 5 Variation of vibration amplitude and frequency measured using UNCS with vibration target distance iii) Non contact method vis-a-vis the contact method a. Vibration amplitude readings of contact and non contact sensors for different frequencies, keeping the vibration table amplitude constant. Vibration table vibration amplitude = 100mg Table 4: Contact and non contact vibration amplitude readings for different frequencies Vibration target Measured vibration amplitude [mg] - Contact Measured vibration amplitude [mg]-unc 100 106 110 200 105 108 300 95 104 400 94 90 500 98 88 Fig. 6 Comparison of measured vibration frequency using Contact and UNCS method
b. Vibration amplitude readings of contact and non contact sensors for different vibration amplitudes keeping the vibration frequency constant. Vibration table vibration frequency = 100 Hz Table 5: contact and non contact readings for Different vibration amplitudes Vibration table amplitude [mg] Measured vibration amplitude [mg] - Contact Measured vibration amplitude [mg] -UNC 5 Noisy- not measurable 6 10 15 12 50 60 58 100 114 110 200 212 210 Fig. 6 Comparison of measured vibration amplitude using Contact and UNCS method At lower vibration amplitudes contact measurement is more erroneous when compared with non contact sensor based measurement. V. EXPERIMENT SET UP The vibration profile was studied using the setup shown in Fig. 8. Single sensor frequency spectrum plots (Fig. 9) were used to find out the target vibration frequency and amplitude. The 3 Dimensional profile visualization display was used to verify whether the vibration profile accurately followed the target vibrations and to measure the target vibration amplitudes (Fig. 10 and Fig.11). Fig. 8: System set up Fig. 9: Frequency spectrum plot of single sensor
Fig. 10: Vibration profile, maximum amplitude on corner Fig. 11: Vibration profile, maximum amplitude at middle VI. APPLICATION Applications of vibration profiling includes but not limited to industrial, agricultural, infrastructure etc. contact vibration sensors used for fault diagnostics of machinery using vibration information is prone to measurement errors due to mass effect and coupling issues due to irregular target surface. Non contact sensors based on the technology described here is suitable for such applications as it is free from measurement errors and coupling issues, being non contact measurement. The increased measurement sensitivity results in more precise diagnosis of machinery faults. Continuous health monitoring of major infrastructure like bridges, buildings, dams etc using vibration profiling can be achieved by a network of high sensitive vibration sensors. VII. CONCLUSIONS The set up was used to analyse surface vibration using a target which whose vibration frequency ranged from 100 Hz to 10 KHz. A minimum measurable vibration amplitude of 5 mg at the lowest frequency point was achieved with a measurement accuracy of +/- 20%. The distance between target specimen and the sensor array was varied between 5 cm and 30 cm and the minimum measurable vibration amplitude results were consistent. At lower vibration amplitudes, non contact sensor was found to have better performance than contact sensors. The cost of the total set was in the order of 10K INR which is hundred times less than the comparable vibration measurement devices based on laser technology. The application of frequency down conversion method for high sensitive vibration measurement using ultrasound was found to be feasible and the specifications achieved are comparable with that of other high sensitive vibration measurement devices, whose cost is many times more than the described system. VIII. ACKNOWLEDGMENT The authors would like to express their appreciation for the financial support by our organisation Centre for Development of Advanced Computing, an autonomous society under the Department of Electronics and Information Technology, Govt. of India. REFERENCES [1] Rodriguez, Ponciano;INAOE, Puebla;Trivedi S. ; Feng Jin ; Lorenzo, J. ; Chen-Chia Wang ; Zhongyang Chen ; Khurgin, J. ; Libbey, B. ; Habersat, J High sensitivity pulsed laser vibrometer for surface vibration monitoring. [2] Collins, E.G., Jr.Dept. of Mech. Eng., FAMU/FSU Coll. of Eng., Tallahassee, FL, Coyle, E.J, Vibration-based terrain classification using surface profile input frequency responses. [3] Johan Kirkhorn, IFBT, NTNU, Introduction to IQ-demodulation of RF-data. [4] Dionysius M. Siringoringo & Yozo Fujino, Department of Civil Engineering, The University of Tokyo, 7-3-1 Hongo Bunkyo-ku, Tokyo 113-8656, Japan, Noncontact Operational Modal Analysis of Structural Members by Laser Doppler Vibrometer.