Appendix D.Load Cell Calibrations: Effects Of Bearing Plate Warping
Load cells used to measure loads during testing of tiebacks, driven piles, and drilled shafts give calculated loads that are frequently in disagreement with loads calculated based on hydraulic jack pressure and piston area. Because of this, there is a general lack of confidence in load cell data and the fault is often ascribed to manufacturing defects, or to improper, inaccurate calibration procedures.
Nevertheless, it is well-known throughout the industry that load cell readings can be affected profoundly by eccentric loading, uneven and/or warped bearing plates, or frictional effects on the bearing surfaces. The purpose of this technical note is to provide some insight into these effects.
D.2Load Cell Calibration Procedures
The usual calibration procedure is to use a testing machine to apply a load to a load cell. The measured load cell output is then correlated against the known applied load as measured by the testing machine. Usually, the testing machine has a hydraulic pressure applied to a piston of known cross section area. The testing machine is checked out periodically by running tests on a load cell traceable to NIST. Generally there is little doubt about the accuracy of the testing machine. Accuracies of ¼% FS, ½% FS, or 1% FS are normal.
The calibration tests are usually performed between large, flat, parallel platens in the testing machine (sometimes with the addition of a spherical seat). The calibration is performed in this manner so that there is little or no eccentric loading, and no bending of the platens, only the elastic compression in the zone immediately bearing against the load cell.
Such a state of affairs may not exist on the job site, since the bearing surfaces next to the load cell are usually much less rigid and more at risk to bending.
This bending is particularly apparent if there is a mismatch in size between the load cell and the hydraulic jack. If the hydraulic jack is larger than the load cell there is a tendency for it to try to wrap the intervening bearing plate around the load cell. If the hydraulic jack is smaller than the load cell it will try to push the intervening bearing plate through the hole in the load cell.
Thicker bearing plates will bend less, but the effect will never be entirely eliminated. The consequence of this bending can be quite large since the effect on the load cell is to cause it to either barrel out at its midsection if the jack is too small, or pinch in at its midsection if the jack is too big. For electrical resistance strain gauge load cells, the gauges are usually located on the outer surface of the load bearing cylinder at its midsection.
D.4Effects of Jack Size on Load Cell Reading
A series of tests were conducted in a testing machine to investigate the magnitude of the effect of jack size on load cell readings.
Load cell bearing surface: 100 mm (4") ID, 146 mm (5¾)" OD
Simulated jack A bearing surface: 50 mm (2") ID, 100 mm (4") OD
Simulated jack B bearing surface: 100 mm (4") ID, 146 mm (5¾") OD
Simulated jack C bearing surface:150 mm (6") ID, 200 mm (8") OD
Maximum applied load: 150 tons
table 6: Effects of Jack Size on Readings
It can be seen from the results that if the jack is smaller than the load cell, the load cell will over-register, while a jack bigger than the load cell will cause the load cell to under-register. The effect is bigger if the bearing plate between jack and load cell is thinner.
The correct bearing plate thickness will depend on the extent of the mismatch between jack and load cell. However, as a rough rule of thumb the following thicknesses should be required:
Up to 100-200 kip load: 25 mm (1") thick
Up to 400 kip load: 37 mm (1.5") thick
Up to 1000 kip load: 50 mm (2") thick
Up to 2000 kip load: 75 mm (3") thick
The consequences of the above would seem to indicate that for best results the load cell calibration should be performed in the field with the actual hydraulic jack that will be used. Alternatively, if the calibration is done in the laboratory, the load cell and the jack should be placed in the testing machine at the same time. Failing that, the load cell should be loaded through a ring, having the same dimensions as the bearing surface of the hydraulic jack, positioned on the other side of a bearing plate of the correct thickness. In this way, some of the variables affecting the agreement between load cell readings and hydraulic jack readings can be removed and the agreement should be that much closer.
This technical note has addressed only the subject of the size mismatch between load cells and hydraulic jacks. Other factors affecting the agreement between load cell readings and hydraulic jack load are important, thus frictional losses within the hydraulic jack can cause under-registering of jack load indications by as much as 15%. (Dunnicliff 1988' Section 13.2.6)
Also annular style load cells are susceptible to end effects and eccentrically applied loads. The height of the load cell should exceed four times the wall thickness of the annulus and at least four strain gauges should be used (Dunnicliff 1988' Section 13.2.6) increasing to eight or 12 in number as the size of the load cell increases.
References:
J. Dunnicliff. 1988. Geotechnical Instrumentation for Monitoring Field Performance, John Wiley & Sons, New York, NY: 577pp.