Failure Simulation

[thepilot 578]

Hi all, I have a question regarding failure simulation that I couldn't quite resolve reading the manual.

 

First off, service-based failures. I have those activated and I understand that some items will fail if the plane is either subjected to adverse operational conditions (i.e. hard breaking will lead to break overheat, then reduced brake effort or brake failure) or due to normal "wear & tear" over time, also requiring the airplane to be serviced regularly (though I am at a loss to say what exactly would fail).

 

Then, random failures. Those would come "out of the blue", correct? Which would mean that even a properly serviced aircraft that is operated within the operational limits could sustain a failure, as happens sometimes in the real world, too. In the literature I could find that safety-critical aircraft systems have to be so reliable that a failure only occurs with a likelihood of 10^-9 per flight hour. So, very seldom. In the PMDG CDU menu I have activated random failures. I see you can also enter a value for events per ten hours. I've tried to enter 0.1 or 0.01 (= one event per 100/1000 hours) but it will only take (positive) integers. One event per 10 hours is too likely for my needs, though.

 

I'm a bit puzzled right now what to expect in regards to failures. I duly complete the entire (Preliminary) Preflight Procedure according to the respective FCOM before every flight even though the simulation might never provide a failure scenario in this constellation.

Basically what I'm trying to achieve is a realistic chance of a system failure (whichever it may be) – not a failure simulation. So a malfunction should not be impossible, but less than "likely", so that you can't plan for it.

 

For example, I could imagine that during the preflight checks, the APU fire warning is rendered inoperative, so that you either need maintenance to fix it (i.e. through the CDU), or it's a MEL item and you can't use the APU, so a start using ground air/cross bleed becomes necessary, etc. I wonder if this would be possible to achieve using the NGX alone or some third party addon; it would make flight simulation all the more exciting.

 

Thanks a lot for any input!

[Rhinozherous 90]

* BUMP *

- This is a very interesting question, I also wanna know.

[kevinh 409]

I would try a few random failures at the maximum 1 per 10 hour rate. That could represent several flights before a failure so it would still take you be surprise when it activated. Concentrate on failures in the area you wish to study. Not ideal but better than nothing.

 

I don't believe the NGX failure system is designed in the way you want it to work. However someone on this forum did write an external script to control engine failure so it should be possible for some enterprising programmer to create an addon to control all failures in more complex and varied ways.

[rondon9898 84]

Do please correct me if I'm wrong chaps, but isn't what you're trying to achieve, Marius, covered by the 'Service Based Failures' option? I mean, the nature of the failures generated by this would be providing the randomness you're trying to achieve, even though the aircraft may be close to being due for a 'service' (which is a purely theoretical, imaginary concept in the sim anyway). Or am I misunderstanding?

[thepilot 578]

Concentrate on failures in the area you wish to study.

 

I have done some failure training in the simulator, activated them right through the CDU, which worked very well.

The flights that I do in the simulator are about 1 1/2 hours on average. That means if I activate 1 failure/10 hours, then something will happen over the course of the next 5-7 flights I do.

It would basically be like getting in your car, being certain it will brake down very soon. Or a pilot taking command knowing that something in his airplane will definitely fail.

Not so realistic when your aim is to simulate a regular airline environment where (serious) faults have become a rare occurence, measured by the amount of flights operated every day. In that respect, I wouldn't mind not experiencing a failure in the next 1000 hours or even more.

 

 

 

but isn't what you're trying to achieve, Marius, covered by the 'Service Based Failures' option?

 

It probably is, that's also what I was trying to find out.

[rondon9898 84]

In that case Marius, stick with the Service Based Failures and you will get what you are looking for. Leave the random failures off.

[thepilot 578]

I will. Activated and tracking closely when the airplane needs to be serviced. :smile:

 

One thing that makes me wonder though is that you can activate random failures, but you can leave the events/10 hours at 0 at the same time. Somehow that doesn't make complete sense to me.

[kevinh 409]

I have done some failure training in the simulator, activated them right through the CDU, which worked very well.

The flights that I do in the simulator are about 1 1/2 hours on average. That means if I activate 1 failure/10 hours, then something will happen over the course of the next 5-7 flights I do.

It would basically be like getting in your car, being certain it will brake down very soon. Or a pilot taking command knowing that something in his airplane will definitely fail.

Not so realistic when your aim is to simulate a regular airline environment where (serious) faults have become a rare occurence, measured by the amount of flights operated every day. In that respect, I wouldn't mind not experiencing a failure in the next 1000 hours or even more.

No different to real airline crews training in the simulator. They know they will get failures they just don't know what and when exactly.

 

It probably is, that's also what I was trying to find out.

Service based failures will take a long time to develop. They are based on reliability data so the more likely failures might not be the most interesting..

[thepilot 578]

 

 

No different to real airline crews training in the simulator. They know they will get failures they just don't know what and when exactly.

 

In the simulator, they know that they will get something within their simulator sessions, which is how long? Maybe one or two hours? And then it's probably the nasty stuff (engine failure at V1, cargo fire,...). In the real world, they probably won't set their mind to experiencing a fault like this. I'm more interested in simulating these actual line operations. :smile:

 

 

 

Service based failures will take a long time to develop. They are based on reliability data so the more likely failures might not be the most interesting..

 

Can't ask for more. I'm not at all intent on experiencing a malfunction, but the sole possibility of it happening drives me to be more attentive and accomplish all flows, checklists etc. as diligent as possible.

[Rhinozherous 90]

Can't ask for more. I'm not at all intent on experiencing a malfunction, but the sole possibility of it happening drives me to be more attentive and accomplish all flows, checklists etc. as diligent as possible.

 

That is the point, for me too! In other aircraft I wouldnt do a lights test and really LOOK at the lights :-)