The preload that cannot be taken back

The backprop jacked tight lands on the floor below followed a frame of 250 mm flat slabs cast a floor a week and found that a preload in its backprops does not reduce the load in the frame — nothing can — but decides which slab carries it. The young slab is relieved in the week after it is struck, and the slab below pays, because a jack pushes down exactly as hard as it pushes up. The best preload balanced the two, and it was small: a fourteenth of a slab’s weight, a firm turn of the collar.

That essay ended on the sequence a temporary-works engineer would reach for next. Jack the backprops hard when the young slab is struck, to give it relief when it needs it most; then slacken them back to snug before the next slab is struck above, so that the preload is gone before it can land on the slab below. The idea treats a preload as a loan to be repaid on a date. Whether it can be — whether a collar turned back hands the young slab only what the jack lent it, and whether it does so before the slab below has felt it — is a question about what a prop’s force is made of, and when each slab is at its worst.

The same frame, and a collar turned back #

The frame is the earlier essays’: slabs 250 mm thick on a 7.5 m column grid, a floor every seven days at 20 °C, one level of shores under the newest slab and one level of backprops under the slab beneath that. When a slab reaches seven days its shores are struck and backprops put in, and in this frame the backprops are jacked to 0.15 of a slab’s weight before they are locked. The slabs creep, by Eurocode 2’s creep function at their equivalent age, and the props are springs: the slabs and props are the same creeping stack that found a snug backprop takes load nobody gave it.

Slackening is modelled in the two ways a site can do it. Slackened to snug: the collar is turned back until the prop is loose and then turned forward until it just touches, which is what a gang with spanners does. Let down by the preload: the force in the prop is reduced by exactly the amount it was jacked to, which needs a prop with a load cell or a hydraulic jack with a gauge and is not what a gang with spanners can do.

What a prop’s force is made of #

The preload is not all a jacked backprop carries. The moment it is locked, the next slab is cast on shores above the young slab, and the backprop takes its share of that wet slab’s weight — about half of it — as a snug one would. Then creep adds more: the young slab creeps down onto its backprops under the weight it is still carrying. A day after jacking to 0.15 w, the backprop is carrying 0.81 w. Of that, 0.15 is the preload, nearly half a slab’s weight is the wet slab above, and the rest is a day of creep.

A force in a prop has no labels. Turned back until it is loose, the backprop lets go of all 0.81 w, and every bit of it goes back to the young slab above, which is now carrying the wet slab on its shores and has nothing under it. Its load jumps from 1.19 to 2.00 times its own weight in an instant. Snugged up again, the backprop starts collecting creep from nothing, and by the end of the week it is back to 0.35 w — half what a backprop that was never touched would have.

A snug backprop is not a jacked one with its preload removed. It is a prop that carries its share of the slab above and whatever creep has put into it since it went in. Slacken it to snug and all of that is lost, not only the jack’s contribution.

How much worse the frame gets #

The measure that decides whether a construction sequence is acceptable is a slab’s use of strength: its load divided by its strength at that age, as a fraction of its 28-day strength, because what a slab can carry on the day depends on how old it is. With snug backprops the frame’s worst use is 1.95, at the moment the young slab receives the next wet slab. Jacked to 0.15 w and left, 1.93.

Slackened to snug on any day of the week, the frame is much worse: 2.55 if the backprops are slackened six hours after jacking, 2.49 a day after, 2.23 if left until the day before the next strike. The later the slackening the less bad, because the young slab is a little stronger each day it waits. But even at six and a half days the young slab, given everything back at once, reaches a use of strength fifteen per cent higher than it would have had if nobody had touched a collar.

The sequence that was meant to protect both slabs overloads the one it was meant to protect. It does so because the thing the gang can do — slacken to snug — is not the thing the sequence needs.

The preload, and only the preload #

The sequence can be rescued in principle. A prop with a load cell can be let down by exactly its preload, leaving it carrying everything else, and a hydraulic jack with a gauge can do the same. That is the version the earlier essay’s question had in mind, and it is the second line in the figure: the worst use stays between 1.92 and 1.94, against 1.93 for leaving the preload in. It does essentially nothing.

At every preload the two lines lie together, and letting the preload down is very slightly worse. The young slab gets its relief at the moment it needs it, the cast above, and the letting down returns the preload to it a day later when it is a little stronger, which costs it little. But the frame’s worst use is not the young slab’s once the preload is more than about a tenth of a slab’s weight. It is the slab below’s, and letting the preload down does not reach it in time.

When the slab below is at its worst #

The slab below receives the preload at the instant the jack is turned, because the jack pushes down on it exactly as hard as it pushes up on the young slab. At the same instant it receives its share of the next wet slab, cast on the shores above. And at the same instant its own backprops — the ones that were under it in its first week — are taken out, because the frame carries only one level of them. Its load steps up and its use of strength steps from 1.44 to over 1.8, and then creep in the young slab above, pressing down through the backprops, adds a few per cent over the following hours.

That is when the slab below is at its worst: within hours of the jacking. A letting down at six hours cuts the rise short and the slab’s use falls back toward the snug line, but its peak, 1.92, has already been reached. A letting down at a day comes after the peak, and the peak is 1.94. There is no later moment at which a release can help the slab below, because its worst moment is the jack’s own moment. It is the same kind of instant as a prop jacked against a finished structure, a support that carried a moment nobody computed until it was installed: the load it brings is decided when it goes in.

This is the structure the earlier essay’s question did not see. The sequence assumed that the young slab’s worst moment and the slab below’s worst moment come at different times, so that a preload can be present for one and absent for the other. They do not. Both arrive at the cast above the young slab, which is the same moment the young slab is struck and its backprops jacked. A preload is present at both worst moments or at neither.

What does help #

The earlier essay said that where the young slab genuinely needs more relief, the answer is a second level of backprops. The creeping stack confirms it with numbers. One level of snug backprops gives a worst use of 1.95; jacked to its best preload, 1.87. Two levels of snug backprops give 1.79, better than any preload on one level. Two levels with the same light preload give 1.70.

The difference is the one between moving a load and spreading it. A preload moves load from one slab to the next, and the best it can do is to balance the two slabs it sits between. A second level of backprops ties a third slab into the share, so that each wet slab is carried by three slabs rather than two, and every slab in the frame carries less. It costs a second level of props and the floor space they occupy for an extra week. It costs no measurement, no load cell, and no collar turned at a precise moment by a gang that has other things to do.

Snug again, and slack again #

There is a second cost of slackening to snug that the figures do not include, and it is the one the essay on snug backprops found to be the largest effect of all. A backprop put in by hand is rarely truly snug: a collar turned until it feels tight leaves a fraction of a millimetre of slack, and a millimetre of slack under a 250 mm slab is more than half that slab’s deflection under its own weight. A backprop with slack carries nothing until the slab above has sagged onto it.

The jack’s real virtue, that essay and its successor found, is that it removes slack: a prop jacked to a force has none, whatever gap it was installed with. A sequence that jacks a backprop and then slackens it to snug gives that virtue back. The slackened prop is exactly as snug as the gang’s hands make it, and the young slab, which has just been handed the whole of the backprop’s force, now has to sag onto its props again before they take any of it back. Every number in the slackening figures assumes the re-snugged prop has no slack at all. Any slack it does have makes them worse.

Three instructions for the site #

The findings come down to three instructions, each shorter than the reasoning behind it.

Do not slacken a jacked backprop during the floor cycle. Turned back to snug, it releases its share of the slab above as well as its preload, and the young slab takes all of it at once. There is no day in the week on which that is better than leaving the prop alone.

Do not specify a release by the preload. Done exactly, with a load cell, it changes nothing that matters: the slab below has carried the preload at its worst moment before anyone can turn a collar. Done by feel, it is a slackening to snug by another name.

If the young slab needs more relief than a light preload gives, add a level of backprops. It is the only change here that lowers every slab’s use of strength rather than moving it from one slab to another, and it needs no measurement. The check it relieves most is the punching check at the columns of the slab below, which carries the most at a fortnight and is the check a construction-stage calculation most often makes only on the youngest slab.

A preload is a transfer, not a loan #

The general lesson is about what kind of thing a preload is. A preload looks like a loan: a force lent by a jack that can be taken back when it is no longer needed. In a frame of slabs and props it is a transfer, made at an instant, between two members of nearly equal stiffness, and what happens next depends on everything else that arrives at that instant. In this frame three things arrive together — the preload, the next wet slab, and the removal of the lower backprops — and once they have arrived the frame’s worst moment is already past for one slab and arriving for the other.

A bolt that was already stretched behaves differently, because it clamps plates far stiffer than itself, so a load added afterwards mostly unloads the plates and the bolt hardly notices. A backprop clamps nothing stiffer than itself; it links two slabs that creep, and a force put into it is shared by both of them at once. Whether that force can usefully be removed later depends on whether anything has happened in between that made it worth having, and in a floor cycle the thing that made it worth having is the cast, which happens at the same moment.

The jump and the step, by hand #

The jump when a backprop is slackened to snug is the backprop’s force at that moment, 0.81 of a slab’s weight, added to the young slab’s 1.19. At eight days the slab has 80 per cent of its 28-day strength, so its use is 2.00/0.80 = 2.49, the figure’s value.

The step in the slab below at the moment of jacking is the sum of three things: the preload, 0.15; its share of the next wet slab, about half, less what the young slab carries; and the load its own backprops were carrying, about 0.7, which is now its own. Its use of strength steps from 1.44 to 1.83 or so, at 90 per cent of its strength, and the preload adds 0.15/0.90 = 0.17 more, which is the difference between the snug and jacked lines in the slab-below figure, near enough.

Where the creeping stack stops #

The props are springs and every prop at a level is the same. In practice the force a collar turn achieves varies from prop to prop, and a slackening by a gang varies too; the model’s single number for each level is an average over the panel.

The release is instantaneous. A level of backprops is slackened one prop at a time over an hour or two, and the slabs share the released force as the props go. The sum is the same; the young slab’s peak may be a little less sharp than the model’s.

The creep at a few hours old is extrapolated, as the earlier essays noted, and the slab below’s rise over the hours after jacking is mostly the young slab’s first-day creep.

The load cell, the gang and the trades below #

They cannot show a prop with a load cell, which costs more than a backprop and is not used on ordinary frames, nor the time a gang spends on a floor of several hundred props, which is what turns a sequence on paper into a sequence on site. And they cannot show the second level of backprops’ cost in the trades below, who lose a week’s use of another floor while the props are in it — which is what a preload was meant to save.

Still open: the floor cycle that is not a week #

Every frame here casts a floor every seven days. A fast cycle — four or five days, with heated formwork to bring the young slab to its striking strength sooner — puts the young slab under the next wet slab younger and the slab below under its preload younger as well, and changes which of them governs. Whether a faster cycle makes a preload more useful, because the young slab is weaker relative to the slab below, or less, because the slab below is also younger — and whether the second level of backprops is then the only answer, or the shortest cycle a frame can sustain on one level — is the question the week leaves, and the strength a slab has when its load arrives is the arithmetic it turns on.