Robert;
a) when the decision is taken to avoid potentially hazardous weather systems, is it necessary to switch the autopilot off?
No. Weather avoidance is a routine flight operation requiring no special handling. The autoflight system on the A330, like all transport category aircraft, is designed to handle moderate turbulence should it be encountered. The autopilot almost always gives a better ride and less 'g' loads to the airframe.
In heavy turbulence which may be beyond the autopilot's design capabilities, the autopilot is disconnected to permit the aircraft to ride up or down encountered severe vertical drafts without the autopilot trying to maintain altitude. The autothrust is similarly disconnected. In such encounters the goal is to maintain the last settings before disconnection and ride the turbulence out without changing pitch or power, permitting small excursions of same to ease the load on the airframe.
Such conditions and responses are rare because avoidance techniques are so successful.
b) same question regarding autothrust.
Same answer, same reasons, essentially. If the autothrust chases the speed however, it is disconnected while the autopilot may or may not be left on. A situation where the autopilot is on and the autothrust is off means that the airspeed must be monitored very carefully. Too fast is far better than too slow, even overspeeding the airframe by as much as 20kts past VMO/MMO (Velocity Maximum Operating/Mach Maximum Operating)
c) assuming that the answer to both these questions is YES, and given the final known position of the ill-fated flight (I understand that it was relatively close to an active weather system), is it surprising that both autopilot and autothrust were ON at 0210? That is, if they had been turned off temporarily to evade weather, is it likely that they would have been turned on again in relative proximity to the weather?
The question is almost certainly immaterial as are the assumptions. There is the possibility that one or both were disconnected and the aircraft manually flown but there are no ACARS messages indicating this. A manual disconnection would generate such a message through the FIDS, (Fault Isolation and Detection System) and the CMC, Central Monitoring Computer and ACMS, Aircraft Condition and Monitoring System, which sends maintenance messages to the ACARS. The actual process of warning/message generation is far more complex than this as I'm sure you appreciate but that is the essence of it.
c) could the aircraft continue to fly more or less in a stable way for some time if there was no pilot control after the loss of autopilot and autothrust?
This is a highly speculative, hypothetical question and, necessarily, any response is the same. Keep in mind, the flight path is a result of all forces acting on the aircraft - smooth air, less forces, turbulent air, more forces, etc.
Hypothetically speaking, in smooth air where the dynamics are relatively undisturbed, because of the autoflight system fbw - fly-by-wire design, would remain fairly stable for some time, (no inputs mean no output - autoflight retains last set attitudes). Eventually however, without altitude hold (lost upon a/p disconnect), the flight path would tend to change with changes in temperature, (very slightly reduced/increased lift) and would also begin to lose speed stability. With other transport aircraft which are not fbw, loss of the flight path would occur much more rapidly without autothrust/autopilot connection.
In the kind of weather this aircraft was "in", loss of a stable flight path would occur more quickly only because of the vertical/lateral gusts associated with the surrounding (or entered) CB's.
In either case, with autopilot off and autothrust off (leaving the engines at the cruise power setting, the ultimate outcome of a "pilotless" transport is a very high-speed dive, possibly spiral dive. In other words, a flight path a very long way from the description offered in the BEA report. The wreckage pattern alone obviates any of this scenario.
d) if a Mayday was sent, is it likely that it would have been received by other aircraft in the vicinity?
Theoretically, yes. All aircraft monitor 121.5 and the common frequency. Even an open mike would have been "heard" by other aircraft and would be on their CVRs, (which would be a matter of routine investigation to check, I should expect). Almost certainly, (and this has been pointed out dozens of times by other pilots on this thread), the crew was likely very busy handling their aircraft in what seems to have been a rapidly degrading set of circumstances. They would not have had a moment to send out a distress call nor possibly to even key the mike to broadcasts "in the blind". The RMP, radio management panel, would be set to VHF 1 on #1 and VHF 2 on #2 but they were also in the process of contacting DAKAR (if I recall) and may have been on HF1 when trouble hit.
It may have been impossible to hang onto the sidestick to key the mike - the other switch is on the audio control panel and is tiny so placing it in the "transmit" position would be difficult - the trigger switch on the sidestick would be the more obvious choice. Neither occurred.
e) how much redundancy is there in systems to detect loss of cabin pressure on this model of aircraft?
You have to bear in mind that detection of this does rest with redundant systems but such systems rely upon their correct operation. As I have posted twice now, the schematic of systems which receive input from the ADRs, if an ADR is malfunctioning that may send false information to the two CPCs, cabin pressure controllers, which in turn may degrade and possibly send BITE - built-in-test-equipment, an internal self-test mode, messages to the CMC. In other words, as the BEA report states and as I stated around June 4th or so, on the first thread, the ACARS messages may not have occurred in the order that they appear in the now-infamous ACARS list of events. Also, in these circumstances, (rapidly degrading airplane due pitot/ADR/IRS issues), the actual cabin rate may or may not have exceeded the 1800fpm needed to trigger the message. For these reasons it is exceedingly important to avoid coming to conclusions, without extensive knowledge of the aircraft, using just the ACARS messages.
My questions are partly motivated by discussion a while ago about the possibility of flight crew incapacitation.
This notion has been posited a number of times throughout the thread. While it is acknowledged that "anything is possible", (bomb, meteor strike, EMF, terrorist intervention), other indications (the above discussion re loss of control with ap/at disconnected, condition of the recovered wreckage) do not support incapacitation theories. However, as singpilot as wisely stated, the CVR will help a great deal in winnowing all theories.
Hope this helps.
PJ2