CogniSim
Background:
CogniSim is the trademarked name for Zetacon's unique proprietary
control architecture. This is filed for patent under the heading: "The
Predictive Control Method" which is also known as Generic Cognitive
Predictive Control Architecture (GCPCA).
The basic CogniSim architecture is shown here:
The basic motor model for an induction machine
is shown here.
The basic motor model for a PM synchronous
machine is shown here.
The most popular three phase voltage source
inverter and associated relationship is shown here.
CogniSim configured for use with an induction
motor and a three-phase voltage source inverter is shown here.
CogniSim configured for use with a PM synchronous
motor and a three-phase voltage source inverter is shown here.
As you can see, the differences are minor. This is the object-oriented nature
of CogniSim and is the key to maximized reuse, flexibility and versatility
between motor classes.
All power circuits consist of switch states that can be defined. CogniSim
utilizes state transition constraints to establish the allowable choices of
switch vectors available during every h-step. (An h-step is a CogniSim time
slice; see patent for detailed definition).
This diagram shows state transition constraints for the three
phase voltage source inverter.
These switch states can be translated to the
voltage vector representation.
Each and every h-step, CogniSim applies key measured data from the A/D
currents and DC bus voltage to the embedded model. Each and every h-step
CogniSim simulates and synchronizes the embedded model to the real world.
If (and only if) a variable to be controlled (example torque and flux) is
presently contemplated and it is determined to potentially warrant the
change of the switch state, then (and only then) does CogniSim proceed to
execute a next state contemplation.
A next state contemplation consists of the building and seeding of several (4
in this case) virtual simulations. A single synthetic time step is simulated for
each case, and a patent pending set of techniques are utilized to determine the
"best" next state.
Because CogniSim only switches when necessary, one can view CogniSim as
searching for opportunities not to switch. This aspect of CogniSim
results in a lower average switching frequency compared to traditional PWM
(Pulse Width Modulated)
control systems.
Typically the switching frequency is at least 30% less by comparison in
periodic applications, resulting in meaningful reduction in heat sink and/or EMI
suppression components.
In non-periodic applications, where the signal varies randomly, such as audio,
active vehicle suspension, and motor drives with impulsive or complex load
cycles, the savings can be significantly higher.
The conceptual simplicity of CogniSim in principle is clear.
Some of the other technical advantages are summarized below (click image
to enlarge view)
CogniSim Exclusive
Opportunities
CogniSim offers, in some cases the capability of continuing to operate if
and when a power switch fails. For example, if a single switch fails to operate, most
traditional control systems will become instantly unable to function. This is
because they have been designed explicitly to operate with one modulator,
and the entire control system is tuned around the correctly operating
expected modulator. It is no wonder a single point failure can bring the
entire controller down.
In advanced CogniSim applications, the knowledge that a switch has
failed can be determined, and when it has determined this
information, the CogniSim State transition constraints can be reloaded with the
correct remaining achievable states. Instantaneously, CogniSim can
continue operation (albeit at a reduced power quality level) but with the
Cognitive engine only simulating the available states that are achievable.
The best switching sequence possible that delivers closest to the
performance objectives will be chosen.
|
