GUIDELINES FOR CALIBRATION OF FALLING WEIGHT DEFLECTOMETERS

Transcription

1 1 GUIDELINES FOR CALIBRATION OF FALLING WEIGHT DEFLECTOMETERS RENE CLEMEN, PRODUCTION MANAGER Fuglesangsallé 16, 6600 Vejen Denmark Tel.: Fax.: ABSTRACT The interest and need for systematic pavement maintenance has been growing tremendously over the last decade. This has increased the demand for specialised equipment for planning and maintenance in practice and for pavement condition survey equipment. The FWD is today the most used non-destructive testing (NDT) device for testing of load bearing capacities of road and airfield pavements. The FWD equipment has undergone an impressive development and is a true state-of-the-art device using the latest electronics and computer technology. In Europe and the USA research work has been done in order to harmonise test methods and calibration procedures for the FWD. In Europe this work is done by the Centre for Research and Contract Standardisation in Civil and Traffic Engineering (C.R.O.W.), the Netherlands. A special group - the Study Committee P7 Falling Weight Deflection Testing - has developed preliminary guidelines for Falling Weight Deflectometer calibration. The calibration protocols are partially developed by the Study Committee P7 and partially with amendments from the protocols developed by Strategic Highway Research Program (SHRP), National Research Council at Washington, DC. Different levels of the calibration programme have been developed, ranging from simple to difficult, and from low-priced and convenient to more expensive and comprehensive. This paper will discuss new equipment, which has been developed to comply with the new calibration tests performed according to the new protocols at verification centres.

2 2 INTRODUCTION In Europe there are 3 manufacturers of Falling Weight Deflectometers FWD, namely PHØNIX Pavement Consultants, KUAB and Dynatest. All have their own FWD calibration programmes. Normally deflection transducers, load cell and temperature sensors are calibrated. CBPC has as the only manufacturer chosen to base its calibration programme on CROW (Preliminary Guidelines for Falling Weight Deflectometer Calibration). First of all this programme was chosen because it has been developed for FWDs operating with load pulse widths within the range of msec. and our opinion is that this procedure gives the best calibration result - and it is an European standard. CROW operates with the following calibration procedures (protocols A-G): A. Relative Calibration Verification of FWD deflection sensors. B. FWD short-term repeatability verification C. FWD long-term repeatability verification D. Reference LVDT calibration procedure (Linear Variable Differential Transducer) E. FWD deflection sensor calibration verification F. FWD group field calibration procedure G. FWD field calibration When CBPC calibrates FWDs this is done on the basis of Protocols A, B, D and E, which normally give a good and secure calibration result. (Protocols C, F, G are for more comprehensive analysis programmes in which CBPC takes part every time the Study Committee P8 and CROW arrange comparison tests among various FWD brands). In stead of just using Protocol A, B, D and E, CBPC has taken a step further in the endeavour to obtain the optimum calibration and thus the most precise and correct results. CBPC Calibration Equipment CBPC has developed calibration equipment, which works with the newest technology and principles within this sphere. Furthermore the equipment has a state-of-the-art design, which makes it possible to bring along the equipment to the customer to be used at site.

3 3 BASIC THEORY When a geophone (deflection sensor) is calibrated the problem is that the curve is not linear in the low frequency range from 7-0 Hz. Geophone frequency response Hz db Geophone frequency response This problem may be compensated for by adding a fast filter at a natural frequency of 4.5 Hz (- 3 db), which is similar but in opposite direction of the geophones own frequency characteristics (see below figure). Filter compensation Hz db Geophone frequency response Filter compensation

4 4 In this way a resultant characteristic is obtained which is linear from approx. 0 Hz to 300 Hz. To compensate for geophone unlinearity within the low frequency area in this way is not a new method but a method, which has been used for many years. Geophone frequency response with filter compensation 50 Hz db Geophone frequency response Filter compensation Frequency response w ith filter compensation CBPC differs from other FWD manufactures in that CBPC provides for the fact that geophones are not uniform in their characteristics. The natural frequency has according to the geophone manufacturer a tolerance of +/- 0.5 Hz. This means that it may vary from 4 to 5 Hz. It is very important to consider this tolerance as it may give an inaccuracy on the peak value of up to 5%, which will again result in a large inaccuracy when back calculating. If the geophone has a natural frequency of 4 Hz and a filter compensation of 4.5 Hz is used, this means that the frequency characteristic and thus the geophone output will be considerably wrong. That is up to 5% too high. Geophone frequency response with wrong filter compensation HZ db Geophone frequency response Filter compensation Frequency response w ith filter compensation

5 5 If a geophone has a natural frequency of 5 Hz and the filter compensation is 4.5 Hz, this will mean that the frequency characteristic and thus the output of the geophone will be noticeably wrong. That is up to 5% too low compared to the reference. Geophone frequency response with wrong filter compensation Hz db Geophone frequency response Filter compensation Frequency response w ith filter compensation EQUIPMENT FOR ABSOLUTE CALIBRATION When CBPC makes an absolute geophone calibration, specially developed calibration equipment is used that is the newest technique within this area. The system consists of a flight case, which is easy and handy to transport worldwide and a Notebook Personal Computer with Windows 95 software.

6 6 NoteBook Computer MS Windows 95 sw. Geophone 1-9 GEO Shaker The flight case is constructed around a mini-shaker, which is capable of generating pulses within the entire geophone frequency and displacement ranges. A highly accurate LVDT is mounted on the mini-shaker, which is used for reference and for control of the shaker s fluctuation. The control of the computer and the calibration process is taken care of by a CPU built into the flight case. LVDT CPU Card A/D Converter D/A Converter AMP Amplifier FWD Analog Card CBPC CALIBRATION PROCEDURE When CBPC calibrates geophones (displacement transducers), three different calibration parameters are used: Off set Cut off frequency Slope The off set and the natural noise, which influences the analog card and the channel, and for which the system compensates automatically. This means that in practice the value is 0. Cut off frequency is exactly the frequency where the system shall compensate for the non-liniarity in the low frequency range. Slope is the amplification required to make geophone and calibration reference show the same peak value. These three software values have the great advantage that they do not change over time and the current development of the calibration may be followed. Calibration is done at a given rise time (frequency) and peak value which may be changed by the operator as required. When the system has automatically found offset, cut off frequency and slope, a test drop is made with the found calibration data. The operation may now decide whether the found values shall be accepted.

7 7 Normally the values will be within the below range: Min. Max. OFFSET ADC CUT OFF FREQ Hz SLOPE gg The found calibration values may now be tested in various ways to find out if they are acceptable. A number of test drops may be made to check that max. deviation ratio at peak does not exceed a fixed value. Normally the deviation from the reference is below 0.1%. This test does however not indicate whether the cut off frequency is correct but only whether the slope is correct. In order to check the cut off frequency a frequency response test must be made. Frequency response After a correct calibration a frequency response should look like this response. Under frequency response it is possible to see which frequency characteristic the geophone has within the range of Hz. Here is the same geophone shown with the same calibration parameter (offset and slope) but without compensation filter. As it clearly appears, the geophone frequency response is no longer linear. Here is the same geophone again with the same calibration parameter. However the cut off frequency has been changed with -0.5 Hz. As it appears, the geophone becomes unlinear from approx. 7 Hz.

8 8 Here is the same geophone again with same calibration parameter. However cut off frequency has been changed with +0.5 Hz. As it appears, the geophone is unlinear from approx. 7 Hz but this time in the opposite side. Peak value In the below table it appears how large an influence a correct cut off frequency has on the peak value. Peak value test at diff. cut off frequencies Cut off frequency Drop. REFERENCE GEOPHONE DEVIATION DEVIATION Hz. No. µ m. µ m. µ m. % Correct cut off frequency at 4.75 Hz Incorrect cut off frequency at 4.0 Hz Incorrect cut off frequency at 5.2 Hz

9 9 Linearity test After the frequency test the system automatically performs a linearity test in accordance with the conditions set forth in C.R.O.W. protocol E. The linearity test gives a good picture of the quality of the calibration as more combinations of rise time and amplitude for the pulse are tested. Data analysis software CBPC has also developed an advanced user-friendly software which can be used for data analysis and manipulation. For many years it has not been quite clear what time history data could be used for. So far only peak/max. values have been used for back calculation. However a tool is now available, which makes it easy and quick to make FFT analyses (Fast Fourier Transformation - frequency analyses) and in this way see what the frequency content of a drop and the peak value. Another facility of the tool is low and high pass filter. These filters may among other things be used for noise filtering of time history signals or for comparison with data from other FWD manufacturers using filter manipulation in data collection software (see below).

10 10 CBPC finds that original signals should not be filtered before presenting it. This is a possibility when analysing data. In this way it is always possible to return to the raw original signals. The CBPC analysis software has been fully integrated with the CBPC data collection software and may automatically be integrated with Microsoft EXCEL spreadsheets. RECOMMENDATIONS With more than 20 years experience as a manufacturer of falling weight deflectometers and as a user of back calculation CBPC has learned how important it is to have data collection equipment (FWD) that measures as accurately as possible. The data collected by the FWD should not only be approximations, which are e.g. added a lot of filters (60 Hz) in order to have equipment with a reasonable repeatability performance. CBPC is of the opinion that the data collected by a FWD shall be as realistic as possible. That means without deliberate data manipulation. The only filters and compensation that are accepted are those required by the physics and mathematics in order to present correct data. As a consequence it is vital to know not only the off set and amplification as calibration parameters but also the cut off frequency varies from geophone to geophone due to the way geophones are constructed. These values may vary up to 1 Hz, which is of utmost importance for the output. If the FWD (only the types using velocity transducers) does not consider this parameter, which is the case for much equipment, CBPC strongly recommends that precautions are taken (maybe via the FWD manufacturer) to provide for this inaccuracy of the geophone in future. REFERENCES E. BEUVING, CROW Preliminary Guidelines for calibration of falling weight deflectometers, C.R.O.W. Centre for Research and Contract Standardization in Civil and Traffic Engineering. SENSOR NEDERLAND BV, SM 6 Miniature digital grade long travel geophone, Input/output inc. specifications for SM-6 geophones.