Model Alarming block
Description
The Model Alarming block monitors an input field and it generates an alarm if the field passes outside of the user-defined limits for longer than a user-defined time period. It is able to extract useful information regarding the relationships between the variables in the process being modelled by performing analyses on the model block. It is able to determine the causes for deviations (i.e. which of the variables is causing the deviation to occur) in the monitored variable. The analysis can recommend which adjustable variables should be altered (increased or decreased) in order to correct the deviation It is also able to function as an alarm without performing analysis on a model block.
Diagram of the Model alarming block and its various ports
Block Type
Rules and Models block
Input/Output ports
Inputs
-
The top input port of the Model alarming block is intended for the monitored variable
-
The second input port from the top is intended for the model inputs (if a model is to be used), the block will automatically connect to this port when the model is specified in the Model alarming block
-
The third input port from the top is intended for the model outputs (again, only if a model is to be used), the block will automatically connect to this port when the model is specified in the Model alarming block
Parameter ports
-
The first parameter port on the left is intended for external alarm limits, if available. This port may contain variable set-points and variable lower and upper limits. If the set-points and/or limits are fixed, they may be configured in the Model alarming block without having to make use of the Alarm Limits port.
-
The parameter port on the right is intended for the Alarm Codes, if available.
Outputs
-
The top output port of the Model alarming block gives the alarm messages. There are 4 fields available; IndicatorsValid (this will indicate whether the cause and leverage indicator ports contain possible valid values or not), Message (the message of the alarm), MessageCode (this is an integer value that represents each message, 0 indicates no message) and MessageSeverity (the severity of the alarm).
-
The second last input port holds the Causes field
-
The bottom port holds Leverages fields.
-
The Top N Causes and Leverages are found in the second and third output ports. N is defined by the user and is <= number of model inputs. It represents how many of the causes (and leverages) with the highest contribution (influence) to (on) the deviation in the monitored variable should be reported.
Functions performed on tags
-
On the values - The input values are used to determine the values of the various output fields. The output fields include;
-
On the timestamp - The time stamp of the output fields will be set to the newest time stamp of the input fields.
-
On the quality - If any one of the input values is of bad quality, the output value(s) will also be of bad quality. If the Enable Input Range Validation box is checked, then the output will be set to bad quality if any of the input values are not within the range of the input minimum and maximum of the data it was trained on.
Related topics: