APM.FILTER - APMonitor Option

 Type: Floating Point, Input
 Default Value: 1.0
 Description: Measurement first-order filter: (0-1)

FILTER determines how much of the raw measurement is used to update the value of MEAS. A filter of 0 indicates that the measurement should not be used in updating the MEAS value. The FILTER parameter applies to all inputs into the model from the CSV (data) file and also from the DBS (database) file. FILTER at 1.0 uses all of the measurement and ignores any prior measurement value and 0.5 uses half of each.

 CSV (data) file
 Model = Model * (1-FILTER) + Measured * FILTER

 DBS (database) file
 MEAS =  MEAS * FILTER + LSTVAL * (1-FILTER)

The FSTATUS parameter is used to adjust the fractional update from new measurements when using apm_meas(s,a,'name',value) but this only applies to a single entity at a time. FILTER applies globally to all inputs. Both FSTATUS and FILTER are useful to control the flow of information into the model. It is sometimes desirable to set FILTER or FSTATUS to a low value (close to 0) when the solver is not able to find a solution because of big input changes. A drawback of a filter on data is that raw inputs are not used in the application and it takes several cycles to reach true input values.

See also FSTATUS, MEAS, Estimation Objective Tuning

💬