Skip to content

EXPLORE.md

Latest commit

 

History

History
201 lines (151 loc) · 6.79 KB

EXPLORE.md

File metadata and controls

201 lines (151 loc) · 6.79 KB

Part A: Data Exploration in Aura

Open console.neo4j.io, sign in, select your instance, and click Query to open the query interface. Use the queries below to explore the aircraft digital twin graph.

Aircraft Fleet Overview

List every aircraft with its model, manufacturer, and system/component counts:

MATCH (a:Aircraft)
OPTIONAL MATCH (a)-[:HAS_SYSTEM]->(s:System)
OPTIONAL MATCH (s)-[:HAS_COMPONENT]->(c:Component)
WITH a, count(DISTINCT s) AS systems, count(DISTINCT c) AS components
RETURN a.tail_number AS tail, a.model AS model,
       a.manufacturer AS mfr, systems, components
ORDER BY a.tail_number

The fleet contains 20 aircraft across four models: A320-200 (Airbus), A321neo (Airbus), B737-800 (Boeing), and E190 (Embraer). Each aircraft has 4 systems and 16 components.

System-Component Hierarchy

Expand a single aircraft to see its full system and component tree:

MATCH (a:Aircraft)-[:HAS_SYSTEM]->(s:System)-[:HAS_COMPONENT]->(c:Component)
WITH a, s, c ORDER BY s.name, c.name
WITH a, s, collect(c.name) AS comps ORDER BY s.name
WITH a, collect({system: s.name, components: comps}) AS systems
RETURN a.tail_number AS tail, a.model AS model, systems
LIMIT 1

For a B737-800 (e.g. N95040A) this shows:

  • CFM56-7B #1 / CFM56-7B #2 — Compressor Stage, Fan Module, High-Pressure Turbine, Main Fuel Pump, Thrust Bearing
  • Generic Avionics Suite — Air Data Computer, Flight Management System, Navigation Receiver
  • Main Hydraulic System — Flap Actuator, Hydraulic Reservoir, Main Pump

A320-200 aircraft have V2500 engines, A321neo aircraft have LEAP-1A engines, and E190 aircraft have GE CF34-10E engines.

Sensors

View the sensors attached to each aircraft's systems. Every engine system has four sensor types:

MATCH (a:Aircraft)-[:HAS_SYSTEM]->(sys:System)-[:HAS_SENSOR]->(s:Sensor)
RETURN a.tail_number AS aircraft, sys.name AS system,
       s.sensor_id AS sensor, s.type AS type, s.unit AS unit
ORDER BY a.tail_number, sys.name
LIMIT 10

Sensor types: EGT (°C), Vibration (ips), N1Speed (rpm), FuelFlow (kg/s).

Flight Operations — Top Routes

Find the most frequent routes across the fleet:

MATCH (f:Flight)-[:DEPARTS_FROM]->(dep:Airport),
      (f)-[:ARRIVES_AT]->(arr:Airport)
WITH dep.iata AS origin, arr.iata AS dest, count(f) AS flights
RETURN origin, dest, flights
ORDER BY flights DESC
LIMIT 10

Flight Delays

Trace delays for a specific aircraft, including the cause and airport:

MATCH (a:Aircraft {tail_number: 'N95040A'})-[:OPERATES_FLIGHT]->(f:Flight)
OPTIONAL MATCH (f)-[:HAS_DELAY]->(d:Delay)
OPTIONAL MATCH (f)-[:DEPARTS_FROM]->(dep:Airport)
OPTIONAL MATCH (f)-[:ARRIVES_AT]->(arr:Airport)
RETURN a.tail_number, f.flight_number, dep.iata AS origin, arr.iata AS dest,
       d.cause, d.minutes
ORDER BY d.minutes DESC
LIMIT 10

Maintenance Events

Recent maintenance events with fault codes, severity, and affected systems:

MATCH (me:MaintenanceEvent)-[:AFFECTS_AIRCRAFT]->(a:Aircraft)
OPTIONAL MATCH (me)-[:AFFECTS_SYSTEM]->(s:System)
RETURN a.tail_number AS aircraft, me.event_id AS event,
       me.reported_at AS date, me.severity AS severity, me.fault AS fault,
       s.name AS system
ORDER BY me.reported_at DESC
LIMIT 10

Common fault types include: vibration exceedance, overheat, sensor drift, contamination, bearing wear, fuel starvation, leak, and electrical fault.

Maintenance Event Traversal

Trace the full path from aircraft through system and component to a critical maintenance event:

MATCH (a:Aircraft)-[:HAS_SYSTEM]->(s:System)-[:HAS_COMPONENT]->(c:Component)
      -[:HAS_EVENT]->(m:MaintenanceEvent {severity: 'CRITICAL'})
RETURN a.tail_number, s.name AS system, c.name AS component, m.fault, m.reported_at
ORDER BY m.reported_at DESC
LIMIT 10

Component Removals

Investigate removal records with time-since-new (TSN) and cycles-since-new (CSN) data:

MATCH (a:Aircraft)-[:HAS_REMOVAL]->(r:Removal)
OPTIONAL MATCH (r)-[:REMOVED_COMPONENT]->(c:Component)
RETURN a.tail_number, c.name AS component, r.reason,
       r.tsn AS time_since_new, r.csn AS cycles_since_new
ORDER BY r.tsn DESC
LIMIT 10

Document-Chunk Structure

The enrich command loaded three maintenance manuals as Document nodes, split them into Chunks with OpenAI embeddings, and linked them with NEXT_CHUNK chains:

MATCH (d:Document)
OPTIONAL MATCH (d)<-[:FROM_DOCUMENT]-(c:Chunk)
WITH d, count(c) AS chunks,
     sum(CASE WHEN c.embedding IS NOT NULL THEN 1 ELSE 0 END) AS embedded
RETURN d.documentId AS doc_id, d.aircraftType AS aircraft,
       d.title AS title, chunks, embedded
ORDER BY d.documentId

Three manuals: AMM-A320-2024-001 (~43 chunks), AMM-A321neo-2024-001 (~58 chunks), AMM-B737-2024-001 (~53 chunks).

Preview the chunk chain for a document:

MATCH (c:Chunk)-[:FROM_DOCUMENT]->(d:Document {documentId: 'AMM-A320-2024-001'})
WITH c ORDER BY c.index
OPTIONAL MATCH (c)-[:NEXT_CHUNK]->(next:Chunk)
RETURN c.index AS idx, substring(c.text, 0, 80) AS preview, next.index AS next_idx
LIMIT 10

Extracted Entities — OperatingLimits

The enrich command used LLM entity extraction to create OperatingLimit nodes from the maintenance manuals. Each limit is qualified by aircraft type:

MATCH (ol:OperatingLimit)
RETURN ol.name AS name, ol.parameterName AS param,
       ol.aircraftType AS aircraft, ol.unit AS unit,
       ol.regime AS regime, ol.minValue AS min, ol.maxValue AS max
ORDER BY ol.aircraftType, ol.parameterName

Examples: "EGT - A320-200", "Vibration - B737-800", "N1Speed - A321neo", "FuelFlow - B737-800".

Cross-Links: Knowledge Graph to Operational Graph

The enrichment pipeline created cross-links between the extracted knowledge and the operational graph:

Document → Aircraft — Each maintenance manual is linked to the fleet aircraft of that model:

MATCH (d:Document)-[:APPLIES_TO]->(a:Aircraft)
RETURN d.title AS manual, a.tail_number AS aircraft, a.model
ORDER BY d.title, a.tail_number

Sensor → OperatingLimit — Sensors are matched to extracted operating limits by parameter name and aircraft type:

MATCH (s:Sensor)-[:HAS_LIMIT]->(ol:OperatingLimit)
RETURN s.sensor_id AS sensor, s.type AS sensor_type, ol.name AS operating_limit
LIMIT 10

Provenance chain — Trace an OperatingLimit back through its source chunk and document to the fleet aircraft it applies to:

MATCH (ol:OperatingLimit)-[:FROM_CHUNK]->(c:Chunk)-[:FROM_DOCUMENT]->(d:Document)
      -[:APPLIES_TO]->(a:Aircraft)
RETURN ol.name AS limit, substring(c.text, 0, 60) AS source_chunk,
       d.title AS manual, a.tail_number AS aircraft
LIMIT 10

Graph Statistics

MATCH (n)
WITH labels(n)[0] AS label
RETURN label, count(*) AS nodeCount
ORDER BY nodeCount DESC