During a construction stage analysis, I used the option "Constriction Stage Analysis Control Data - Load Cases to be distinguished from Dead Load for CS Output" to create a separate load case to examine the effect of the wet concrete and the construction loading to CS2. However, running the analysis with the load cases separate from the Dead Load (CS) produced a considerably different bending diagram than anticipated. Can you explain why there is a discrepancy, given that I assumed they would produce the same bending moment diagram? The details of the construction stages and the results are attached below.

I am running 4 Construction Stages:

CS1: Erection of abutments, Piers and crosshead beam (28 days)

CS2: Erect the W beams that are simply supported to the piers and the abutments and apply wet concrete and construction loads to the deck. (7 days)

CS3: The composite action began (wet concrete and construction loading are removed. The beams from simply supported become continuous to the deck. (7days)

CS4: same as CS3 but just to see the shrinkage and creep long term effects.

Now when I run my analysis I get to CS3 the following moments (For Dead Load CS) which make sense:

As you can see have on major axis a small relatively hogging and a large sagging. 

Now I go to Analysis - "Constriction Stage Analysis Control Data - Load Cases to be distinguished from Dead Load for CS Output" and create a separate load case to examine the effect of the wet concrete and the construction loading to CS2. Bear in mind that I am not altering any other parameter to my model i.e. the wet concrete and construction loading are added to CS2 and are taken out at CS3.

No when I run the analysis with the load cases separated from the Dead Load (CS) I get a very different bending diagram see below:

Can you please explain why the difference I would expect them to have the same bending moment diagram. 

P.S. Both models are attached. 

The reason for this is that the wet concrete which is a Dead Load type concrete is now separated from the Dead Load (CS) output and so the self-weight of the slab is now applied onto the continuous beams, instead of being applied on the simply supported beams.

To clarify this, let us first talk about how we can consider the wet concrete load. Due to the wet concrete having to be applied onto the girder before the composite action, we need to apply this load manually as a UDL on the beams before the composite action of the section. There are generally two ways of doing this:

1) Apply a UDL to represent the wet concrete on the structure and use a weightless material for the concrete slab;

2) Apply a UDL to represent the wet concrete which is then removed when the slab is activated with its weight.

In your model you have used the second method, and so in the long term you are using the Self-weight function to represent the weight of the slab. This means that the slab weight will be applied onto the system after the composite action is activated. This is not a problem if the wet concrete load is also applied as dead load and is removed at the stage when the slab is activated because this means that the load from the slab is actually applied in early, on the beams alone, and then in the slab stage the wet concrete UDL is deactivated and the self-weight of the slab is activate. Both these loads act on the composite structure, and they are equal and opposite, thus, the combined effect is 0.

However, if the wet concrete load is distinguished from the Dead Load (CS), then:

 - In the Dead Load (CS) first we have the weight of the girder applied on the girder alone, and then we have the weight of the slab applied when the slab is activated on the composite (continuous) beam.

 - In the Erection Load first we have the wet concrete applied on beam only, and then we have the wet concrete removed on the continuous beam. You will see a residual diagram in this stage for the Erection Load even though there is practically no load acting under the wet concrete load case on the structure. This locked in effect should be considered together with the Dead Load (CS) in order to get the actual state of the structure:

Because the wet concrete is not in the same load case as the dead load from the slab, the two cannot counter each other to leave only the original diagram. You need to combine the two loads to obtain the total effect due to dead loads on the structure.