Sign In Request Pilot Access
Fault Isolation & Root Cause

From 64 stations to one root-cause station.

Intuigence AI correlates fault signatures across your full station network and surfaces the one station most likely responsible — with the signal-level evidence to back it up. The engineer reviews the hypothesis and the supporting tag-value deviations before any work order is created. Synthetic colleague, not autonomous decision-maker.

044 Root-cause station (94% conf.) Healthy / downstream stations STN-044 isolated Hydraulic pressure drop WO drafted · 3 min 42 sec
How The Engine Works

Systematic correlation, not pattern matching.

The Intuigence AI root cause engine doesn't match faults to a pre-built failure mode library. It correlates signal timelines across your specific station network to find the causal station in your current configuration.

01

Fault Event Detected

OPC-UA or MQTT stream flags an alarm. The engine begins capturing the pre-fault and post-fault signal window across all connected stations — typically a 15-minute window.

02

Temporal Correlation

Correlates signal deviations across stations by time offset. Stations whose signals deviated before the primary fault event are ranked higher as candidate root-cause sources.

03

Evidence Package

The top-ranked station is presented with its supporting evidence: specific signals that deviated, the time offset, and a confidence interval. Engineer reviews and confirms before work order creation.

Transparency

Signal evidence, not black-box scoring.

Process engineers need to understand and verify AI outputs, not just accept them. Every root-cause hypothesis includes the signal-level evidence that generated it.

Signal-Level Citations

Each hypothesis lists the specific PLC tags and deviation windows that triggered it. Engineers can navigate directly to those signals in the trace view.

Confidence Intervals

Station rankings include confidence percentages and ranking margins. Engineers can see when the evidence is strong (92% STN-044) vs. ambiguous (54% STN-044, 48% STN-046).

Engineer Override

Engineers can accept, reject, or substitute the AI's root-cause hypothesis before the work order is created. The override is logged and feeds model improvement.

Historical Pattern Lookup

The engine cross-references prior fault events at the same station to surface similar cases and their resolution outcomes — giving engineers historical context alongside the current hypothesis.

Ready to see fault isolation on your line data?

Pilot access includes a supervised integration with your current PLC and station configuration.