Ford's answers to the NHTSA 6.7 Investigation

NinerBikes wrote:

- cognate field (fluid mechanics, thermodynamics, simulation modeling, etc.)

Huh? I'm lost here... this is the engineering side, computer modelling

- sources of knowledge (field reports, SAE docs, OEM docs, hiring specs, ...)

need industry inside knowledge for this.

Technology / Science is divided into disciplines, like Mechanical Engineering, Civil Engineering, etc.

Then it is subdivided into specialities like thermodynamics, hydrodynamics, tribology, material science, etc.

Each field is subdivided into much finer slices...


Each application area in turn, creates their own simulations such as combustion simulators for gasoline (small displacement), fluid flow simulations, thermal, etc.

Simulators, in turn, are not perfect, but engineering constructs --- so a good investigator almost invariably start with looking at flaws and defects in the simulation.

Or in misapplication of the simulator, or something.... like Sleipner A.


The job of an investigator (or builder / system architect) is to identify the fields of knowledge that is relevant and useful to solving the problem at hand, and then tapping the knowhow in that discipline and integrating it into a useful final product.

For example, are we dealing with supercritical fluid behavior at 2,000 bars?

Don't know!

If we are, supercritical fluids behave very differently from single or dual phased (gas and liquid) in terms of physical, and likely chemical properties.

Are we dealing with emulsions? Another stuff that behaves very differently and is not well understood with public data.



OK, how do we access sources of knowledge once we identify a potential cognate field?

http://www.thefreedictionary.com/cognate


Well, I go to an introductory text book for the field, and skim it for relevant info.

Very quickly, you can pick up what are the basics of the business, metrics, etc.

Then go back and ask.. does it apply?

Repeat for each field, making notes along the way.


As you may have noticed... one of the earliest hits I got is the issue of DLC and reactivity with common organic amines in lubricating oil:

http://www.rv.net/forum/index.cfm/fuseaction/thread/tid/25794401.cfm

That research was not there --- new knowledge --- at the time the CP4 was finalized and ramped up.

So we now know not to blindly go ahead and raise the lubricity with additives --- without thinking through the risks.

Many of the lubricity "improvers" tried out in the field, like adding 2 stroke oil, ATF, lubricating oil, etc. would actually damage the coating...

Based on that science.


Ahhh... research


Do the same exercise for SAE papers, technical publications, and very quickly, you can figure out "what we know", "what we don't know" and what we ought to have known...

Then the opps show up.





Side note for software guys...

A major debug now is to examine the code, sensors... seeing if it is doing something strange.

Example of something that need to be ruled out:

Is the software commanding something wrong.. like a valve closing when it should be opening.. .or not reading the sensor right?

Could there be something like a conversion error?

Fixed to float point.... that brought down Ariane 5?

Or a drift error... that screwed up the Patriot PAC 3s after 8 or so hours?