Meeting with Gary

Kurt,
As requested here is a brief overview of the discussion between Mark Fisher
, myself and Gary Verklereen that took place on 6/13/02 & 6/14/02 regarding
availability on the Millrun & Somerset sites. Mark did a great job of
demonstrating the usefulness of the Mon.db file to Gary for the purpose of
calculating availability, and also the method used by GE wind to determine
availability. By the end of the second day Gary expressed he had a new found
confidence in the accuracy of the Mon.db file data and also in general agreed
with the method GE Wind used to calculate availability. Gary did express a
few concerns which are listed below.

Gary was concerned that we may continue to overlook the fact the turbine may
have been put into virtual mode. Mark reassured Gary that new measures were
being taken to prevent this situation from being overlooked in the future and
during the recalculation of the availability.

Gary also expressed that he could see no problem with the 4 hours of
maintenance time that was being added each month as long as, in the monthly
report the actual availability is list in a sperate column as the adjusted
availability.

Gary felt that in the recalculation of the availability it may be warranted
to count the many breaker operations that occurred in the first few months of
the project as downtime rather then Line Out Hours as they may have been
counted. This was due to the fact that we made a adjustment to the breaker
that later prevented it from tripping as often. This suggested to Gary that
the breaker trips were in fact due to improper settings.

Another suggestion by Gary was that after a fault, the turbine could take as
long as 1 to 2 minutes to began producing as the turbine performed its run up
sequence. Gary suggested that during the winter months when the turbines
experienced hundreds of Low Gearbox Oil Pressure faults, that this 1 to 2
minute run up sequence would equate to several hundred minutes were the
turbine reported available, but was not truly available to produce. He
concluded that in the recalculation we may want to count this time as
downtime.

When asked what method could be agreed upon during the recalculation, Gary
stated that after reviewing the method we used he could see no problem with
it as long as we did not overlook things like the virtual mode. Gary also
stated that as long as the final availability was within 1% of his
calculations he would have no complaints. If the final availability numbers
were greater then 1% of his calculations then he would need a break down of
what was counted and what was not counted for review.