Tieback Testing w/ Load Cell 1

[born2drill]

I'm dealing with a situation on a DOT project where the State does not want to accept several anchors due to differences between the jack calibration and load cell readings. There are significant descrepancies - the load cell showing as little as only 65% of the test load in the worst case.

These are large (260kip - 375kip) strand anchors supporting a sheetpile wall, performance tested to 150% DL with a 10-minute creep hold. All were tested with the same jack and load cell on the same day. The tests met the criteria for elongation (above PTI min elongation) and creep (under 0.040" in 10 minutes).

Of course, the DOT is taking the load cell readings as gospel and recommends retesting the anchors unless I can provide an adequate explanation for the differences.

Does anyone have any experience with this situation? I'd appreciate any suggestions.

 

[InDepth]

Try this guy, he may be able to direct you to someone in RST

RST Instruments Ltd., per
Chris Bray, EIT
Sales Engineer
Tel: 604.540.1100

 

[PEinc]

I do not know what your project specifications say about the load cell readings vs. the pressure gauge readings. However, FHWA's Geotechnical Engineering Circular No. 4, Ground Anchors and Anchored Systems, Section 7.3.3.3 Pressure Guages and Load Cells says, "The readings on the jack pressure gauge are used to determine the absolute value of the applied load. For extended load hold periods, load cells are used as a means to monitor a constant applied load while the pump is incrementally adjusted. Over extended periods of time, any load losses in the jack will not be reflected with sufficient accuracy using a pressure gauge. For proof tests and for lift off tests, a pressure gauge alone is usually used for measuring load. For all tests involving extended load hold periods (i.e., all creep tests), a load cell should be used in concert with a pressure gauge."

Load cells are more affected by weather, moisture, temperature, etc. than are pressure gauges. In addition, the contractor should have two calibrated gauges, one being a spare. If you have a problem with a pressure gauge, it can be switched easily. You can't switch a load cell easily. The test would have to be stopped and then be rerun. The load cell is usually used only to check if the jack is losing load during an extended hold. If the load cell's digital counter is decreasing, the jack can be pumped to take the load back up to the desired pressure gauge reading.

 

[DRC1]

What type of soil are the anchors in?
Consider recalibrate the jacks and the load cell. We use Dudgeon in Bridgeport CT. They ship all over, very good calibrations. One sticking point is always the seating load, but I don;t think that would account for that much dicrepancy.

 

[born2drill]

The anchors are bonded in gravelly sand, 7" diameter holes.

I don't think it's a calibration problem. The test setup was calibrated just days before being used. (I use WB in Wood-Ridge, NJ - they're every bit as good as Dudgeon)

I'm thinking more along the lines of PEInc that load cells are finicky and affected by things like weather and especially can give funky results if loaded slightly eccentrically as can be common for multi-strand tiebacks. I was looking for some backup that the pressure gage readings, and not the load cell, should be relied upon to indicate applied load.

DRC1, what do you mean by seating load?

 

[PEinc]

I just performed a 200 ton load test on a 100 ton DL micropile. We used a pressure gage and a load cell, both calibrated. The load cell read pretty close to the calibration values. However, we loaded the jack according to the pressure gauge readings. We checked the load cell at many of the load increments. We used the load cell to monitor load loss during the hold periods. The digital load cell readings were fairly close, but not exact, to where they were supposed to be. When the digital readings dropped off, we pumped the jack back to the desired pressure gauge reading. If your load cell was reading 65% of the pressure gauge value, something was wrong. I assume there wasn't a calibration factor mixing of tons and kips? That could give you a 50% difference and the rest could be accuracy of the pressure gauge readings.

 

[PEinc]

P.S. We used WB also.

 

[PEinc]

born2drill,

You said, "I was looking for some backup that the pressure gage readings, and not the load cell, should be relied upon to indicate applied load." The FHWA reference I quoted above should be the backup you want.

 

[aeoliantexan]

I had a similar project where the load cell and jack didn't agree. The deflections were too small, making us suspect the jack was off. We discovered that the quick-connects on the jack and the return hose to the pump weren't a matched pair; they would connect, but not open both check valves. Pressure was being trapped in the return side of the jack, reducing the actual force delivered.

This problem might not be revealed by a calibration in a testing machine that allows very little extension of the jack during the test.

Call me conservative, but faced with two conflicting pieces of data, I am slow to adopt the optimistic one.

 

[born2drill]

If you have the return line on a double acting jack blinded off, you'll get practically no movement and you also won't be able to stroke the cylinder back in when you go to retract it. This is a simple problem that should be instantly recognized in the field, and something a jack rental company would certainly not do. A proper calibration also should average pressures at different jack ram extensions.

Also - I don't have 2 pieces of data, I have 3. Jack pressure, load cell readings, and anchor elongation. 2 of them say I had the loads right.

 

[STVU]

I had a project about 4 months ago where the backup readings were off after the 2nd cycle of readings. We concluded that the instrument was knocked off by wind or someone & only used the readings from one instrument. Could you test the load cell in a controlled near future field test and show that mechanically it is sound-then propose the accidental bump as a viable solution?

 

[PEinc]

STVU,

If the first gauge's readings look good, forget the backup readings and use the first set. That's why you have a "backup" gauge - in case something happens to the other gauge. If you really needed two sets of readings, they would need to specify three gauges, with the third being the backup gauge.