ISO 13374 — Overview

What is ISO 13374

ISO 13374 is the international standard for condition monitoring and diagnostics of machines. It defines a six-level data processing architecture that transforms raw sensor data into actionable maintenance recommendations.

Rela AI implements all six levels of the standard, allowing organizations to progressively evolve from simple data collection to automatic generation of predictive recommendations.

The 6 levels

L1 — Data Acquisition

Collection of raw data from event sources: sensors, PLCs, SCADA systems, IoT devices. Data is normalized and stored with its original timestamp.

Supported protocols (6):

Implementation in Rela AI:

• Event sources configurable per protocol.
• Real-time ingestion with schema validation.
• Storage in MongoDB with temporal indexes.

L2 — Data Manipulation

Statistical processing of raw data to extract trends, averages, and detect anomalies. Converts point data into continuous information.

Implementation in Rela AI:

• Rolling statistics (min, max, avg, stddev).
• Linear regression and moving averages (SMA/EMA).
• Anomaly detection via z-score and bands.

L3 — State Detection

Compares current data against learned baselines to determine equipment condition state. Identifies operating modes and deviations.

Implementation in Rela AI:

• Automatic baseline learning.
• Operating modes: normal, startup, shutdown, standby.
• Condition evaluation: good, acceptable, unsatisfactory, unacceptable.

L4 — Health Assessment

Combines multiple indicators into a single health index (AHI) that summarizes the overall asset condition with a grade from A to F.

Implementation in Rela AI:

• Asset Health Index (AHI) from 0 to 100.
• Four sub-indices with configurable weights.
• Deterministic recommendations per grade.

L5 — Prognostics

Estimates the remaining useful life of equipment and predicts when maintenance will be required based on degradation trends.

Implementation in Rela AI:

• Remaining Useful Life (RUL) estimation.
• Degradation rate (AHI points/day).
• Condition-based maintenance (CBM) triggers.

L6 — Advisory

Generates specific action recommendations using artificial intelligence, considering the full asset context.

Implementation in Rela AI:

• Recommendations generated by Gemini.
• Managed lifecycle (open, acknowledged, resolved).
• Automatic creation of urgent tasks.

Summary table

Progressive activation

You do not need to implement all six levels at once. The recommended path is:

1. Phase 1 — Configure event sources (L1) and trend analysis (L2).
2. Phase 2 — Learn baselines (L3) and activate health assessment (L4).
3. Phase 3 — Enable prognostics (L5) and AI recommendations (L6).

Predictive Maturity Levels

Each asset automatically progresses through 4 levels as it accumulates data:

Progression is automatic — the system evaluates each level's requirements and promotes the asset when they're met. No manual configuration needed.