Hi,
We are trying to back check our Ram Concept model, but we're running into quite a few "black box" calculations. We've created a simple baseline model of a continuous 3-span strip of concrete with uniform loading to try to determine what some of the deflection calculations are doing. We've simplified this slab run as much as possible to try and verify with hand calcs/ spreadsheets. We've zero'd out the shrinkage restraint and shrinkage strain to try and eliminate the shrinkage parameter from the calculation. We've also only analyzed self-weight to stay below the cracking moment in the slab.
1. My question today is regarding the "Duration (days)" in the load steps. We've modified or created Load steps and combinations to only include slab self weight, no ADL or LL was used in these Load History combinations. When we change the duration from 1 or 2 days (doesn't affect results) to 3 days for the "Instantaneous DL" load history step, the deflection increases by 25% seemingly overnight. No other single day change creates that effect so we're wondering if there is a factor that kicks in at that duration or total age?
2. Also, the sustained deflection (under only self weight) appears to peak somewhere between 3-5 yrs and then starts to improve? Any guesses?
Our ultimate goal is to be able to compare Concepts results with our calcs that are based off ACI, and be in the ballpark.
Try pushing out your initial load application a few more days. Do you still see the big change in the early days after load application? If that's not the issue, send the file.
We uploaded the file with the original post. We played around with different combinations of duration and initial load application and we seem to get similar answers.
I don't see any files in there. Try zipping it first.
Ram Concept Deflection Tests.cpt
That is a little curious. I added extra steps tp have results at day 1, 2 and 3 and then things are essentially constant. I'll see if I can tell anything more from the original logs.
I checked with the development lead and he explained a few things that might be making the difference,
"There is a little bit of magic around that 3 day value that is causing the behavior they are seeing. We somewhat arbitrarily chose that anything less than 3 days would be considered “instantaneous” for the purposes of our calculations. This was primarily for the savings of the computational expense. For loadings of duration 3 days and longer we do both an instantaneous portion iteration and then a sustained portion iteration – and the sustained portions generally take more iterations to converge. For less than 3 days we only do the instantaneous portion and skip the sustained portion altogether. While this can obviously cause some difference in the calculated deflection, the expectation was that the change would be small.
There is no way to predict deflections with any reasonable degree of accuracy with the creep/shrinkage that might occur over a couple of days, which is why this granularity of time was thought to be insignificant."