|
Control
Arts article on controller performance assessment is the most downloaded
paper in the prestigious Journal of Process Control (more...)
AlarmCapture
V5.0 just released! New interface, new capabilities, and support for more
DCS systems (more...).
Control
Arts receives 4 patents for controller performance and alarm management
(more...)
Now
available - three courses for the advanced control engineer on alarm engineering,
model identification, and controller performance assessment (more...)
|
Controller Model
Identification / PID Tuning |
Most loops
are still controlled with the PID algorithm, and it's still necessary
to ensure that these are tuned properly. The Control Arts Model
ID and PID Tuning package lets you efficiently obtain dynamic models
of the process and quickly calculate the best PID parameters - all
from a few simple plant tests.
Controller
Model ID Features:
- Fit low-order
Laplace or z-transform models
- Fit step-weight
models
- Cut bad
data from anywhere in the data set
- Apply non-linear
transforms to inputs and/or outputs
- Able to
fit up to 20x20 systems
- Extensive
diagnostics to ensure model fidelity over the entire data set
- Interaction and conditioning analysis to check for benefits
of multivariable control
- Low-pass and butterworth filter design
PID Tuning Features:
- Applicable to a wide variety of DCS systems
- Direct calculation of PID parameters from Laplace transfer function
model
- Ability to design and simulate both SISO and MIMO systems
- Compare PID performance to model-based controllers
|
| Drag and drop tags to configure the system, then
specify the transform type (first order, underdamped, overdamped,
inverse response, or integrating). The program will automatically
fit all the models and disturbance model, and show the results
in the continuous or discrete domain, or as a step-weight model. |
 |
|
 |
Data sets sometimes contain bad data - either the point was
on control, the measurment failed, or the process was hit by
a large disturbance. The ControllerModelID package lets you
graphically mark bad data, and the program will automatically
identify around the bad data segments. |
|
| The first step to implementing a multivariable controller
is to check whether it is needed and if it will work. A good
check whether multivarable control is needed can be determined
from RGA test - a standard test that indicates the coupling
in the system. And it's also necessary to check the feasibility
of the multivariable controller - the SVD test will indicate
if the multivariable system will be too sensitive to model mismatch
to control adequately (by any controller, either single or multi
loop). |
|
|
 |
PID controller parameters are easily obtained once a model
is fit. The PID controller design module directly calculates
these parameters for a large variety of DCS types, and simulates
the regulatory or servo response. |
|
Price (includes one-year of support
+ maintenance upgrades):
Single User License: $700US
Site License: $2500US
|
|
