Discover more about GPC
below.
GPC...Is the
successor to SPC.
It totally solves the problem of extending the single-variable benefits of SPC
to multiple variables and includes a new multi-variable control chart. It does
all this using geometry instead of statistics so is a truly discontinuous
innovation.
GPC...Advances
the Six-Sigma improvement process.
By
replacing the limited multi-variate methods at the heart of Demings
DMAIC
cycle with new methods that can cope with hundreds of variables
insteadof the
ten or so that are possible today. This means fewer
projects but of wider scope
so reducing overheads and avoiding
sub-optimisation. And no one needs
mathematical spacialisation to
understand it
so everyone on the site
can participate and understand the results.

GPC...greatly
improves process understanding by substantially improving your ability to
extract information from existing process history data.
You always knew that there was a lot of information and process understanding
available in your historical process data. That was how your company originally
justified the purchase of a process historian, wasn't it? But the reality until
now was that you lacked usable methods for extracting all but a fraction of the
information so have probably used your historian mainly for post-event
sequence-of-events confirmation.
Imagine, for instance, how a multi-variable contour chart of many process variables by product grade would simplify process understanding and allow better communication amongst all your staff from the Process Operator to the Site Director.
And all this from existing process history data without requiring disruptive and expensive Plant Experiments and Test Runs. It is now possible with C:Suite Visual Explorer.
GPC...greatly
simplifies Setting Up a process by very quickly finding consistent and therefore
better Operating Procedure Limits for both continuous and batch processes

Today you find Control Limits or Operating Procedure Limits one variable at a
time and continuously revise them to try and get better plant performance. We will show you why the
one-at-a-time methods can't work and give you Inconsistent Limits. Then we'll
show you how much easier and faster it is to find many Consistent Limits at the
same time and calculate the process improvement that you will obtain.
GPC will also guide further investments in process improvement by clearly
identifying which variables need improving next.
Used at Commissioning time for a new plant these methods can reduce the time to reach acceptable performance levels.
GPC...Provides the
first ever method to calculate values for Alarm Limits.
There has never been a method of calculating values at which to place alarm
limits yet a little thought quickly reveals that wrongly set Alarm Limits are
where all the problems
of Alarm Systems begin. The mathematical basis means that the Alarms generated
by GPC are of much higher quality than the Alarms you have been used to so there
are very few false alarms. This means GPC Alarms are in 'closer' and thus
annunciate earlier before process disturbances have had time to become
established so much smaller corrective action is needed and the operator has
much more time to respond. We were proud to receive the 2003 EPSC Award for the
biggest single contribution to improving process safety for this first-ever
fundamental definition of process alarms.
GPC...does
Condition Monitoring without a first-principles process model.
The Operating Envelope of a process can be made as wide as the normal
operating span of a process so that
operator intervention or automatic control action is only needed in
extreme, hence rare, non-normal situations. The non-linear nature of GPC
models makes them very effective at spotting small single-variable
anomalies even though the magnitude of the anomaly means that it is
still well inside the fixed normal operating range for the variable.
The detected anomaly is effectively a slight change in the relationship
to other variables and is
thus an early and sensitive detector giving increased time for reaction.
|
GPC...for Fewer Designed Experiments The non-linear
GPC models are able to create a model of the Response Surface
from the results
of the first round of Designed Experiments (DOE). The Surface
can be visually explored by the experimenter himself to
determine which additional experiments to perform. Todays
methods require multiple linear regression to fit the surface
with skilled mathematical interpretation of the results. This
cannot usually be done by the experimenter himself leading to
delay and additional cost which are eliminated by using GPC.
Additionally, the experimenter has better domain knowledge and
can sometimes use this advantageously to eliminate an
experiment. Note that half-factorial experimental designs should
not be used.
GPC...for
faster Formulations Allows Formulators
to visually explore and interpolate between formulations and
predicts the ranges of properties GPC...for High Throughput Experimentation The interesting experiments are the ones with different results but it can be time-consuming to recognise one or two amongst several hundred while also being certain that none have been missed. The visual methods of GPC make it a breeze. Use the time saved to do more experiments and increase your chances of making a big discovery sooner. If you have 800 frequencies in your spectra how do you begin to compare one sample with another? Try doing it visually with GPC and you may be surprised by how much additional information you obtain
|