Icon

04_​Creating_​a_​ControlChart_​of_​a_​Time_​Series

Anomaly Detection. Control Chart
- Read 313 spectral amplitudes time series produced by the Pre-processing/Time Alignment & Visualization workflow - missing value replaced by last value in time series - loop on all columns to calculate 1st level alarm Training - calculate cumulative average and cumulative standard deviation Deployment - throws first level alarm if signal wanders outdside boundaries (=average +- 2* stddev) - throws second level alarm if alarm 1 averaged across all frequency bands and all signals > 0.25 Take Action - if second level alarm fires => email for mechanical check up - plot of alarm level 2 for the whole time series to see when it was the earliest moment alarm level 2 would fire. Anomaly Detection: Control ChartThis workflow detects anomalies just by checking the wandering off of the signal from a band centered around the time series "normal conditions" averageand large as 4 times the corresponding standard deviation. Node 40Node 41remove datefrom loop on columnsNode 306line plotswith date on x-axisaverage alarmon y-axislearning boundaries andthrow 1 if signal isoutside of boundariesmean(alarm 1) across signals at same timeif mean(alarm 1) > 0.25 => alarm 2Node 340resolve missingswith last value Table Creator String To Number RowID Image Writer (Port)(deprecated) Generate LinePlot 1st level alarm 2nd level alarm Fire if level2 Alarm Missing Value - Read 313 spectral amplitudes time series produced by the Pre-processing/Time Alignment & Visualization workflow - missing value replaced by last value in time series - loop on all columns to calculate 1st level alarm Training - calculate cumulative average and cumulative standard deviation Deployment - throws first level alarm if signal wanders outdside boundaries (=average +- 2* stddev) - throws second level alarm if alarm 1 averaged across all frequency bands and all signals > 0.25 Take Action - if second level alarm fires => email for mechanical check up - plot of alarm level 2 for the whole time series to see when it was the earliest moment alarm level 2 would fire. Anomaly Detection: Control ChartThis workflow detects anomalies just by checking the wandering off of the signal from a band centered around the time series "normal conditions" averageand large as 4 times the corresponding standard deviation. Node 40Node 41remove datefrom loop on columnsNode 306line plotswith date on x-axisaverage alarmon y-axislearning boundaries andthrow 1 if signal isoutside of boundariesmean(alarm 1) across signals at same timeif mean(alarm 1) > 0.25 => alarm 2Node 340resolve missingswith last value Table Creator String To Number RowID Image Writer (Port)(deprecated) Generate LinePlot 1st level alarm 2nd level alarm Fire if level2 Alarm Missing Value

Nodes

Extensions

Links

Download

To use this workflow in KNIME, download it from the below URL and open it in KNIME:

Download Workflow
On NodePit since: 2021-04-27
Last update: 2024-09-06
Created with KNIME version: v4.3.1
Tags: anomaly detectiontime series analysisauto-regressive modelstime alignmentIoTInternet of Things

Deploy, schedule, execute, and monitor your KNIME workflows locally, in the cloud or on-premises – with our brand new NodePit Runner.

Try NodePit Runner!