There were serious problems with the old formulae. Boral research people and myself found the problem in 2015 when we were doing some work on some new reduced shrinkage and creep concrete products they were developing. The problem was with the combination of the drying and autogenous shrinkage formulae and the concrete strength term in the drying shrinkage formula.

Using the 2009 formulae, If you do the conversion from 56 day to 30 year for 100MPa concrete, it shows a 30 year basic shrinkage strain as a starting value of about 3300 for the test shrinkage being achieved, instead of the 800 defined in the code. Working the other way, based on the code defined Basic Shrinkage of 800, the formula gave 56 day shrinkage values for a shrinkage test of 290 for 100MPa concrete and Boral testing was giving about 700 for the test samples!

The 2009 formulae had never been checked for concrete strengths above 50Mpa as that was the maximum concrete strength when the code was first written. We found in 2015 that the formulae were wrong for anything above 50MPa concrete and the error increases as concrete strength increases.

We pointed this out to Ian Gilbert and the AS3600 committee in January 2016 and they modified the formulae to give more logic results. The new results match the results being achieved with Australian concretes based on a survey done in 2016 in conjunction with the development of the changes.

Unfortunately the change has not been made in AS5100 yet as I understand it.

I have not compared to other codes.

Thanks for the reply rapt. Quite concerning since I imagine it would affect most tall buildings built in the last 10 years. Then again, restraint from reinforcement may bring shortening down to more agreeable levels.

Quote (rapt)

The new results match the results being achieved with Australian concretes based on a survey done in 2016 in conjunction with the development of the changes.

You wouldn't happen to have this survey on hand would you?

It was done by CCAAA for Standards at CIA and Ian Gilberts request, so it is a standards document. I do not think it was ever published. I think Ian Gilbert presented something on this at one of the CIA conferences, probably Adelaide in 2017. He might have given information on it then. It is also the basis for changing the Basic Shrinkage for all states in Australia to 800 in the 2018 code from the previous variable value depending on location as the latest figures showed little variation compared to 30 or so years ago when the last survey was done.

Creep was not affected and I would think it would be the real worry for tall buildings as it causes differential shortening between different supports (relatively lightly loaded concrete cores compared to highly loaded exterior columns under gravity loading.). I would be more worried about bridges where higher strength concrete is nearly always used and shrinkage is important and AS5100 is still wrong. Remember the error starts at about 50MPa concrete.

PS RAPT has had this close to right since Jan 2016!