Hello PJ2,
yes i'm very cautious to the sim experiments, and especially due to the facts that such experiments or their results are not known from BEA or AB or from the main players in Simulator software or applications. There should be reason that they are not confident in the validity of any results after the stall was beyond its initial state.
PJ2
To focus on the item under discussion, the THS does not change its behaviour in a stall; it is responding to orders as before. I think it is reasonable to assume, (until proven otherwise by AB, an AMM or a true authority on the A330), that the simulator replicates the THS behaviour accurately.
I have no doubt that the simulator will react on SS inputs like the real jet, but only in those regions where the input datas are known and programmed into the data base. If those important values are not known and therefore are not in the database, then some default values would be used that the system is not going to tilt. Let me make an example concerning AOA (There would be more to it for sure)
. We know, that AF447 encountered AOA up to 50°, and that there are no real air datas available for such extreme AOAs and the behaviour of the airframe under such conditions. As nobody expected such kind of extreme AOAs would be possible for more than few seconds in an air transport catagory aircraft and no legal regulation requests for sampling data for such situations, the default value for the sim software in order not to crash the sim would be some default value. I have my optimistic day today, so let us assume the default would be something like highest tested and confirmed stall value + 10°, let's again assume that value would then be an AOA of 20°.
If that would be the case, and i do not know if it is, then you would start your sim experiment stall recovery always with the same default data of 20°AOA and not from actual AOA. That is the typical "garbage in, garbage out" case.
OK465
I'm completely out of the business now, but with respect to all this discussion of simulator post stall behavior and subsequent recovery in a simulator, the last communication I read from a noted sim manufacturer stated that beyond a relatively early point in the sequence, as opposed to 'flying', one is simply operating a high dollar video game.
For me that sums it up pretty well.
PJ2
On the sink-rate 'g', I think it was very close to 1g all the way down - isn't that the definition of "falling"? Other than slight changes in pitch slightly increasing or decreasing 'g', why would it be "most times below 1g"?
Good point, i should have been more specific.
From 2:10:27 until 02:10:50 (zoom phase) and from 02:11:03 until 2:12:00 the loadfactor was below 1 g. The timeframe with the positive g (02:10:50 until 02:11:03) kicked AF447 into the stall by some noseup stick and TOGA thrust. At 02:12:00 the jet was falling through FL 310 with an AOA of 45° After that point the loadfactor varied around 1 g.
The sinkrate was increasing until passing FL 310. In normal flight that would be longitudonal acceleration, with high stall AOA it is sensed as vertical acceleration. If my math is right, then with AOA 30° sinus is 0,5 and with AOA 60° sinus is 0,866. Therefore increase of descent rate and increase of AOA both contributed to the loadfactor below 1g.
I hope that clears up my wrong statement "all the way down".