Icon

RUL_​LAB_​Part2b_​TwoBaselines

RUL LAB Part 2b - two health indices with different memory, combined with OR (KNIME 5.x, metanode version). Double-click a metanode to open it.
1 Prepare dataKeep cycles 0-6, drop RECOVERING hours.Drop sensor_50 and the redundant block -> 36 sensors.Fill gaps with the previous value.The same clean table feeds index A and index B.
2 Index A - short memoryInside one failure cycle: rolling 48 h median/MAD,shifted +24 h -> z -> HI_A = count |z| > 3 -> HI6_A.Forgets fast: catches sudden changes,drifts along with slow degradation.Alone: 2 of 7 failures (cycles 2, 4).
3 Index B - long memorySplit hours into runs (pump running, no restart).Inside one run: EXPANDING median/MAD since run start(pool from age 12 h), shifted +24 h -> HI6_B;first 48 h of a run suppressed.Remembers the whole run: sees slow drift,needs a long enough run. Alone: 3 of 7 (cycles 3, 4, 5).
4 Combine with OR and evaluateALARM = (HI6_A > 9.8 for 2 h) OR (HI6_B > 9.0 for 2 h)thresholds = 99th percentile over healthy hours (rul > 72)lead time = max rul while ALARM = 1 and rul <= 72false-alarm hours = ALARM = 1 while rul > 72Result: 4 of 7 failures (2, 3, 4, 5), leads 62 / 62 / 68 / 24 h,15 false-alarm hours in 2,160 hours.
Read pump_rul_hourly.csv(data folder of this workflow)
CSV Reader
features_or.csv(timestamp cycle rul HI6_A HI6_B ALARM)
CSV Writer
Both indices over time:short vs long memory
Line Plot
Per cycle: max lead timeand false-alarm hours
GroupBy
eval_or_per_cycle.csv
CSV Writer
1 Prepare data
3 Index B - long memory
2 Index A - short memory
4 Combine A and B (OR)

Nodes

Extensions

Links

Download

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

Download Workflow
Created by: anos
Created at: 2026-09-24
On NodePit since: 2026-09-25
Last update: 2026-09-27
Created with KNIME version: v5.12.0

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

Try NodePit Runner!