Emerson Rosemount 3144P, Справочное руководство

Страница: 171 / 236

161
Reference Manual
00809-0100-4021, Rev JA
Configuration
April 2017
Configuration
4.10.4 Statistical process monitoring (SPM)
SPM algorithm provides basic information regarding the behavior of process measurements such as PID
control block and actual valve position. The algorithm can monitor up to four user selected variables. All
variables must reside in a scheduled function block contained in the device. This algorithm can perform
higher levels of diagnostics by distribution of computational power to field devices. The two statistical
parameters monitored by the SPM are mean and standard deviation. By using the mean and standard
deviation, the process or control levels and dynamics can be monitored for change over time. The
algorithm also provides:

Configurable limits/alarms for high variation, low dynamics, and mean changes with respect to the
learned levels

Necessary statistical information for regulatory control loop diagnostics, root cause diagnostics, and
operations diagnostics
Note
F
OUNDATION
Fieldbus devices offer a wealth of information to the user. Both process measurement and
control is feasible at the device level. The devices contain both the process measurements and control
signals that are necessary to not only control the process, but to determine if the process and control is
healthy. By looking at the process measurement data and control output over time, additional insight
into the process can be gained. Under some load conditions and process demands, changes could be
interpreted as degradation of instruments, valves, or major components such as pumps, compressors,
heat exchangers, etc. This degradation may indicate that the loop control scheme should be re-tuned or
re-evaluated. By learning a healthy process and continually comparing current information to the known
healthy information, problems from degradation and eventual failure can be remedied ahead of time.
These diagnostics aid in the engineering and maintenance of the devices. False alarms and missed
detections may occur. If a reoccurring problem in the process exists, contact Emerson for assistance.
Configuration phase
The configuration phase is an inactive state when the SPM algorithm can be configured. In this phase,
the block tags, block type, parameter, limits for high variation, low dynamics, and mean change
detection can be set by the user. The “Statistical Process Monitoring Activation” parameter must be set
to “disabled” to configure any SPM parameter. SPM can monitor any linkable input or output parameter
of a scheduled function block that resides in the device.
Learning phase
In the learning phase of SPM, the algorithm establishes a baseline of the mean and dynamics of a SPM
variable. The baseline data is compared to current data for calculating any changes in mean or dynamics
of the SPM variables.
Monitoring phase
The monitoring phase starts once the learning process is complete. The algorithm compares the current
values to the baseline values of the mean and standard deviation. During this phase the algorithm
computes the percent change in mean and standard deviation to determine if the defined limits are
violated.