Folks,
I thoroughly enjoyed reading Tom's lessons
learned. It brought out some important points that needed emphasis.
The role of design integration is critical and should be a leadership role
within the project. Tom and Mike did an outstanding job in this
role. At PPPL, Tom was the only engineer at PPPL who really had the
competance and experience to perform this work at the project level. We
always felt that someone should be assigned to Tom to be mentored by him for two
purposes - to help share the load and to improve our core competance in this
critical discipline. After I left for ITER, I believe Phil was finally
given the nod to hire someone to do this. It was long
overdue.
PPPL should move in the direction of training mechanical engineers
in both CAD design and analysis. One of the more compelling reasons
to give ORNL the stellarator core at the beginning of the NCSX project was
because their mechanical engineers (Nelson, Cole, Williamson, Freudenberg) were
capable of doing both CAD design and analysis and could more efficiently
conceptualize the stellarator core. History proved the wisdom of that
decision.
PPPL has an excellent engineering staff but PPPL mechanical
engineers tend to have narrower competancies, i.e. they are analysts or
design engineers (not necessarily CAD competant) or operations engineers.
The PPPL engineers who are competant doing both - Ellis, Brown, and Loesser -
are always in high demand. I am not saying that being an expert analyst
for instance is not valuable - it certainly is and we need experts - but
institutionally, we would benefit by having more engineers who could fill
broader roles like Tom. This takes purposeful employee development, not
just assigning folks based on existing capabilities. It must be a
strategic initiative. If PPPL had such a strategic initiative in place, it
would have been much easier to supplement the design integration effort on
NCSX.
Simpler drawing and document control is a laudable goal.
Institutionally, we might take a look back at NCSX and see how this could be
improved. Certainly, if the folks in line for approving models and
drawings were CAD-competant, it might facilitate adopting a more streamlined
process. Having a small, core team at the start (a skunkworks so to speak)
to develop a workable concept is a better way to launch a project than to start
with a full organization. It certainly simplies drawing and document
control in the developmental phase. Also, if we could bring the
design to a greater level of maturity before filling out the organization and
baselining, much of our cost and schedule uncertainties would be resolved. The
involvement of qualified manufacturing engineers early on as Tom suggests should
also be encouraged. This is true concurrent engineering. Doing the
design while doing manufacturing (a trap we fell into) is what we wrongly called
concurrent engineering.
Wayne