|
ABNORM_ACTIVE
|
None
|
The indication that a block error condition not selected in
BAD_MASK (on the function block level) is True (Active).
|
|
BAD_ACTIVE
|
None
|
The indication that a block error condition selected in
BAD_MASK (at the function block level) is True (Active).
|
|
BAD_MASK
|
None
|
The set of active error conditions that triggers a
user-defined Bad condition. The user selects a subset of block error
(BLOCK_ERR) conditions in the BAD_MASK parameter. When any of these conditions
is True, the BAD_ACTIVE parameter becomes True. When any of the BLOCK_ERR
conditions not included in BAD_MASK is True, ABNORM_ACTIVE becomes True.
|
|
BLOCK_ERR
|
None
|
The actual error status of the function block, set in the
following manner:
-
Out of Service – Block mode is out of service. This
status is also set when the neural network module scan is greater than one
second.
-
Readback Failed – One or more of the inputs has an
invalid reference path.
-
Input Failure/Bad PV – One or more of the inputs
used by the neural network have a Bad status or were configured with an invalid
path.
-
Configuration Error – The neural network module
scan is greater than one second or none of the inputs have a parameter path
defined.
-
Other Error – An input or output value exceeded the
values in the data range used to train the neural network and is being limited
to this range before being used.
|
|
CORR_BIAS
|
EU of OUT_SCALE
|
The model correction value added to the neural network output
to determine OUT and FUTURE.
|
|
CORR_ENABLE
|
None
|
Sets the neural network model correction value used. If
correction is enabled, the model correction value (CORR_BIAS) is calculated as
the difference between the sampled value and the neural network output, after
limiting and filtering. If correction is disabled, a value of zero (0) is used.
|
|
CORR_FILTER
|
Seconds
|
Time constant of the first order filter used to calculate the
model correction factor. Filter saturation is prevented by calculating the
filtered value only when CORR_BIAS is being computed and using the ratio
between CORR_FILTER and block execution period.
|
|
CORR_LIM
|
EU of OUT_SCALE
|
Absolute value of the maximum difference between the sampled
value and the neural network output. Limited to 10% of scale.
|
|
DELAY
|
Seconds
|
The time elapsed between sampling and lab analysis or sampled
analyzer output being available to the NN block.
|
|
FOLLOW
|
None
|
In AUTO mode, the OUT parameter tracks the SAMPLE input value
and status when FOLLOW is active (True). If FOLLOW is inactive (False), OUT is
calculated based on the block algorithm.
|
|
FUTURE
|
EU of OUT_SCALE
|
The predicted value that the process output would become if
the input values were to remain constant at their current value and status.
|
|
IN_DESCn
where
n is 1 through 20
|
None
|
User-specified unique description of INn. These
parameters appear only after the network has been trained.
|
|
INn
where
n is 1 through 20
|
Determined by source or
EU of PV_SCALE or
EU of IN_SCALE
|
Input value and status. These parameters appear only after
the network has been trained. You can apply a maximum of 16 to the model.
|
|
INSPECT_ACT
|
None
|
Indicates if Inspect is enabled and one or more of the limits
for the block have been exceeded. The normal value is 0. This parameter is set
to 1 if both of the following conditions are true:
-
The
Write To Inspect Alarm context menu
item was selected from Inspect for this block.
-
With the Current Hour filter selected, Inspect indicates
that an abnormal condition exists for Mode, Control, Input, or Variability.
(Note that an abnormal condition will only exist for Variability if both the
Variability Index
and the Standard Deviation have exceeded their
defined limits.)
|
|
MODE
|
None
|
Parameter used to show and set the block operating state.
MODE contains the actual, target, permitted, and normal modes.
|
|
OUT
|
EU of OUT_SCALE
|
The process output value calculated by the neural network
based on process inputs.
|
|
OUT_SCALE
|
None
|
The high and low scale values, engineering units code, and
number of digits to the right of the decimal point associated with OUT.
|
|
SAMPLE
|
EU of PV_SCALE
|
Sampled process output determined by lab analysis or by a
sampled analyzer.
|
|
SAMPLE_DESC
|
None
|
Description of the sampled process output. This description
appears in the parameter list in the Neural application.
|
|
STDEV
|
Percent of scale
|
The standard deviation of PV. For analog control blocks in
AUTO, mean is assumed to be the SP.
|
|
STDEV_CAP
|
Percent of scale
|
The estimated capability standard deviation (measurement of
short term variation). An estimate of the least standard deviation the process
could achieve ideally.
|
|
STDEV_TIME
|
Seconds
|
The time frame over which STDEV and STDEV_CAP are performed.
The default value of zero is good for most processes where the scan rate is no
more than approximately 10 times faster than the time to steady state.
If the process is relatively much slower, it is recommended
that you enter the approximate time it takes for the process to return to
steady state after a change. This ensures that the STDEV and STDEV_CAP
calculations accurately consider the actual time constant of the process.
|