Statistical validation of manrating?

[mikejz]

I have been wondering, given that the most reliable rockets out there are NOT man rated (Soyuz, Delta II, Atlas II), have any studies been performed that validate the concept of ‘man rating’ a booster in terms of actual flight reliability?

 

[vogon13]

Statistics aren't all they are cracked up to be.<br /><br />Shuttle was 'man rated' to a degree prior to flight, actual experience with the vehicle demonstrated otherwise.<br /><br /> <div class="Discussion_UserSignature"> <p><font color="#ff0000"><strong>TPTB went to Dallas and all I got was Plucked !!</strong></font></p><p><font color="#339966"><strong>So many people, so few recipes !!</strong></font></p><p><font color="#0000ff"><strong>Let's clean up this stinkhole !!</strong></font> </p> </div>

 

[drwayne]

The Saturn 5 was manrated after a couple of flights, the last of which was not *exactly* a complete success.<br /><br />Wayne <div class="Discussion_UserSignature"> <p>"1) Give no quarter; 2) Take no prisoners; 3) Sink everything."  Admiral Jackie Fisher</p> </div>

 

[tap_sa]

<font color="yellow">"actual experience with the vehicle demonstrated otherwise. "</font><br /><br />This shows that 'man rating' as such means only anal-retentive papertrail which itself does not guarantee that the vehicle really is safe.<br /><br />The term is outdated but a neat way to rip off customers even more: a valve $100, a spacerated valve $1000, a manrated valve $10000. The hardware in all cases is the same <img src="/images/icons/wink.gif" />

 

[vogon13]

I know all about the anal-retentive paper trail. Having multiple sets of eyes watching all the assembly procedures helps, but design problems, or 'concept' problems still creep in.<br /><br />If a particular widget is percieved as being failure prone in 3 ways, 2 of which are benign to the vehicle, but the third results in loss of vehicle, that third failure mode is going to recieve a great deal of attention. If the analysis of failure 2 is wrong, or incomplete, you may not find out till the vehicle is in space.<br /><br />Hence the paper trail, and having more sets of eyes looking at everything, sieving out (all) the problems.<br /><br /> <div class="Discussion_UserSignature"> <p><font color="#ff0000"><strong>TPTB went to Dallas and all I got was Plucked !!</strong></font></p><p><font color="#339966"><strong>So many people, so few recipes !!</strong></font></p><p><font color="#0000ff"><strong>Let's clean up this stinkhole !!</strong></font> </p> </div>

 

[drwayne]

This is similar to another discussion here a week or two ago, in which I basically stated that when paper is a by-product of a good process, then the situation is healthy, but when paper is thre process - then the process is at least as bad as worthless, maybe worse - as shutting down of thought processes can occur.<br /><br />Wayne<br /><br />p.s. Saturn was one of the first big systems to undergo "all up" certifications, in which the components were tested as a unit, not as much as individual systems.<br /> <div class="Discussion_UserSignature"> <p>"1) Give no quarter; 2) Take no prisoners; 3) Sink everything."  Admiral Jackie Fisher</p> </div>

 

[CalliArcale]

<blockquote><font class="small">In reply to:</font><hr /><p>This is similar to another discussion here a week or two ago, in which I basically stated that when paper is a by-product of a good process, then the situation is healthy, but when paper is thre process - then the process is at least as bad as worthless, maybe worse - as shutting down of thought processes can occur. <p><hr /></p></p></blockquote><br /><br />It's very very true. A lot of people complain about the futility of ISO or CMMI certification, because it means so much paperwork that it impedes their ability to work. But that's really a symptom of a bad implementation of process. Getting a good process does require paperwork, but every single bit of paperwork has got to be justified. For instance, CMMI certification at certain levels requires that you collect measures of your work and look at them regularly with. If you pick the measures at random and then don't put enough thought into how much work it'll be to maintain those measures and then to get any use out of them, then it's worse than useless because it'll become nothing more than waste of time and resources. Instead, you need to pick measures that are meaningful, useful, and not too cumbersome to collect.<br /><br />There's no point implementing process unless you do it right. But if you do it right, there's a lot to be gained by it. If the widget you build goes wonky later on, you'll have a paper trail of the right artifacts to zero in on the reason why it went wonky. If your factory burns down, you'll be able to pull records from your offsite storage facility to rebuild your entire fabrication process. Stuff like that. And if you build in a very good review and change control process, you will greatly reduce the number of defects and enhance overall understanding of the product. If you do it right, the process will help you. If it's not helping you, that's a good sign that something needs fixing in the process. <div class="Discussion_UserSignature"> <p> </p><p><font color="#666699"><em>"People assume that time is a strict progression of cause to effect, but actually from a non-linear, non-subjective viewpoint it's more like a big ball of wibbly wobbly . . . timey wimey . . . stuff."</em>  -- The Tenth Doctor, "Blink"</font></p> </div>

 

[radarredux]

> <i><font color="yellow">Also man rating also adds "9's" to the reliability number.</font>/i><br /><br />The problem is that we do not have enough data points to make any statistical claims on such complex systems.</i>

 

[trailrider]

That IS a major problem with limited production items. In addition, one needs to test the entire assembly under as close to operating conditions as possible, with as many samples (flights) as possible.<br /><br />What is needed for maximum reliability of any flight vehicles, but especially with limited production manned space vehicles, is a conservative design, with as much redundency as practicable.<br /><br />The design must be preceeded by a thorough Failure Modes And Effects Analysis (FMEA), by as paranoid and immaginative systems engineers as you can find outside of an insane asylum. And then...you have to listen to them!!!<br /><br />And after that, you have to figure that a guy named Murphy is really running the program, and it is the engineers' job to keep the guy out of the picture!<br /><br />So far as I am concerned, you need to forget statistical reliability in the mass production sense, and concentrate on building the beasts for REAL reliability by not getting real clever with the design! KISS! And don't push the envelope!<br /><br />Ad Luna! Ad Aries! Ad Astra!<br /><br />