Unified Aircraft Health Intelligence

Every aircraft
teaches the fleet.

Nadir AI transforms disposable vibration test data into a permanently learning, fleet-wide health intelligence system — ontology-grounded, privacy-preserving, and aligned with FAA and EASA safety requirements.

Request Technical Briefing Explore Fleet Intelligence
INPUT
Raw Vibration
ACES Systems output — unstructured, unlabeled, discarded
L1 — L2
Signal → Semantic
EEMD decomposition, ontology normalization, JSON-LD
L3
Local Diagnosis
Calibrated fault probabilities, RUL estimation, confidence
L4 — L5
Fleet Learning
Federated XGBoost — no raw data shared, DP guaranteed
L6
Decision Orchestration
Predictive maintenance, governed actions, audit trails
The Gap

High-fidelity data.
Zero intelligence extracted.

Current practice processes vibration data only to produce a binary verdict. Once the test concludes, raw signals are discarded. No historical record. No fleet model. No prediction. Nadir AI was built to eliminate this.

Today — The Disposable Model
Every test starts from zero
  • ACES instruments collect high-fidelity vibration data
  • Test produces a binary pass/fail verdict
  • Raw signals are discarded the moment the test concludes
  • No longitudinal health record built across the fleet
  • No cross-platform comparison or degradation trending
  • Maintenance remains reactive — failures discovered, not predicted
  • Each operator learns only from their own narrow data set
With Nadir AI — The Intelligence Model
Every test builds the fleet model
  • Every test event produces a permanent, governed health record
  • Semantic normalization makes records comparable across aircraft
  • Longitudinal trending detects degradation across flight cycles
  • Federated learning builds a global fleet model — no raw data shared
  • Each new aircraft improves predictions for the entire fleet
  • Remaining useful life estimates replace reactive inspection cycles
  • Collective fleet intelligence no single operator could build alone
Seven-Layer Architecture

From raw signal to
fleet-wide intelligence.

The Nadir AI framework is defined as a seven-layer reference architecture. Each layer performs a specific transformation on the input data, preserving full provenance and ontology grounding. The output of each layer is the input to the next — and every transformation is queryable, auditable, and reversible.

L0
Acquisition & Context
INPUT
Raw ACES sensor streams + flight metadata
→ Raw buffer
L1
Signal Transformation
TRANSFORM
EEMD decomposition → intrinsic mode functions + multiscale features
→ IMF set
L2
Semantic Normalization
MAP
ISO 20816 · ATA iSpec 2200 · S1000D · JSON-LD
→ Canonical record
L3
Local Diagnosis & Prognostics
INFER
Calibrated XGBoost ensembles, ECE, trend, RUL with CIs
→ Health evidence
L4
Fleet Learning
FEDERATE
Federated XGBoost histograms + differential privacy (ε, δ)
→ Global model
L5
Governed Adaptation
GOVERN
Shadow validation, approval gates, versioning, rollback
→ Approved update
L6
Decision Orchestration
ACT
Prioritized component actions, inspection types, work scopes
→ Maintenance plan
Data Transformation Pipeline

Raw signal to
fleet-ready record.

Follow a single bearing velocity reading from an ACES Systems test cell through Nadir AI's seven-layer pipeline. What enters as unstructured sensor output exits as a semantically governed, ontology-enriched record ready to contribute to fleet-wide learning.

▼  Raw ACES Output — Test Cell TC04
// TC04_RUN_2026-03-04_14h22.dat

CH_01_V,     0.847,  1709561331
x_acc_mV_raw,  -12.4,  1709561331
0xFFA3,     4096,   1709561331
bearing_1?,    0.312,  1709561331
temp??,       82.3,   1709561331
vib_unscaled_HZ_???, 0.0044

// concept:   UNKNOWN
// units:     UNKNOWN
// ontology:  NONE
// 47 more unlabeled channels
Nadir AI
▲  Governed Health Record — JSON-LD
// Fleet-ready canonical record

{
  "asset_id""TURB-TC04-001",
  "field_id""VIB-RADIAL-INBOARD-001",
  "concept""Radial Shaft Velocity",
  "value":  0.847"unit":"mm/s RMS",
  "iso20816":{"zone":"A","max":2.3},
  "ata":     {"chapter":"72-00-00"},
  "alert":  false,
  "confidence":0.94,
  "provenance":{"source":"ACES-TC04",
              "modelVersion":"v1.2"}
}
L0
Acquisition
Raw signals, tail number, flight phase, sensor health flags
L1
Signal Transform
EEMD decomposes non-stationary signals into multiscale IMFs
L2
Semantic Norm.
ISO 20816, ATA iSpec 2200, S1000D — serialized as JSON-LD
L3
Local Diagnosis
Calibrated ensemble: fault probabilities, RUL, trend scores
L4
Fleet Learning
Federated XGBoost — no raw data shared, DP guaranteed
L5
Governance
Shadow validation, approval gates, versioning, rollback
L6
Decision Orch.
Maintenance actions, inspection types, governed audit trails
L0 · ACQUISITION
Raw sensor signal + operational context

Vibration signatures vary with load, speed, and flight phase. A bearing reading during cruise means something very different than the same reading on takeoff. L0 captures both the sensor stream and the full operational context that makes downstream analysis possible — assigning a unique identifier to every record and storing the raw data in an append-only buffer.

Signal capture. Raw vibration data is ingested from ACES Systems hardware or equivalent acquisition devices. Channel configuration, sampling rate (typically 10–50 kHz), and calibration factors are recorded per-channel.

Context metadata. Every record is tagged with timestamp, aircraft tail number, sensor location (station coordinates + mount type), operating regime (ground idle, taxi, takeoff, cruise, descent, landing), flight phase, environmental conditions, and recent maintenance history.

Provenance anchoring. A unique record identifier is assigned at capture time and carried through every downstream transformation. This creates the audit chain that makes L5 governance possible — any health decision can be traced back to the originating signal.

Why it matters. Including operating context at capture time is what allows the framework to normalize vibration signatures across flight conditions and to build models that are aware of the envelope in which a measurement was taken.

▸ L0 Output — Raw Record + Context
record_id: 7f3c-a91e
// Raw vibration stream with captured context

{
  "record_id""7f3c-a91e-...",
  "timestamp_utc""2026-03-04T14:22:11Z",
  "tail_number""N748XA",
  "aircraft_type""CH-53K",
  "sensor": {
    "channel""CH_01_V",
    "location""MGB_RADIAL_INB",
    "mount""stud, axial",
    "sample_rate_hz"25600
  },
  "operating_regime""ground_run_100%Nr",
  "flight_phase""maintenance_run",
  "last_maint_flight_hrs"142.6,
  "raw_buffer_uri""s3://...waveform.parquet"
}
L1 · SIGNAL TRANSFORMATION
EEMD decomposition into intrinsic mode functions

Traditional FFT-based vibration analysis assumes stationarity — an assumption that breaks down under transient loads, varying rotor speeds, and real-world operational noise. Nadir AI applies Ensemble Empirical Mode Decomposition (EEMD) to adaptively decompose each vibration signal into a set of intrinsic mode functions (IMFs), each representing oscillations at a different time scale. This allows fault signatures to be isolated to specific frequency bands rather than being masked in the composite signal.

Why EEMD over EMD. Standard EMD suffers from mode mixing — the same physical phenomenon can appear in multiple IMFs across different signals, making comparison unreliable. EEMD injects controlled white noise across multiple trials and averages the resulting decompositions, producing IMFs that are stable, reproducible, and physically meaningful.

Per-IMF feature extraction. From each IMF we extract statistical features (RMS energy, kurtosis, crest factor) and spectral features (peak frequency, bandwidth, spectral centroid). Each feature is stored as a first-class, typed entity in the ontology — which means downstream analytics can query the decomposed components directly.

Reproducibility. The decomposition parameters (sifting iterations, noise amplitude, ensemble trials) are recorded alongside the features, so any result can be independently re-derived from the archived raw buffer.

Operational impact. Bearing defects, gear mesh anomalies, and rotor imbalance each have characteristic signatures in different IMF bands. By making IMFs queryable, maintenance engineers can drill from "something is wrong" down to "the second-stage gear mesh is drifting" without a separate analytical pass.

EEMD Decomposition — CH_01_V · 2-second window
RAW composite IMF 1 high-freq f ≈ 850 Hz IMF 2 mid-freq f ≈ 240 Hz IMF 3 mesh band f ≈ 52 Hz RES trend time (0 → 2.0 s)
ComponentRMSKurtosisPeak f
IMF 1 (high)0.3123.8847 Hz
IMF 2 (mid)0.4814.2241 Hz
IMF 3 (mesh)0.6296.952 Hz
Residual0.0942.1
L2 · SEMANTIC NORMALIZATION
Six ontology dimensions, one canonical record

This is where Nadir AI's interoperability layer lives. Every feature extracted in L1 — plus every metadata tag from L0 — is mapped into a JSON-LD record structured across six orthogonal ontology dimensions. The result is a canonical health record that is queryable, comparable across aircraft and operators, and ready for federated learning.

01 · ASSET STRUCTURE
What is being measured
  • aircraft — tail, model
  • system — rotor, gearbox
  • subsystem — stage, module
  • component — bearing, gear
  • LRU — line replaceable unit
02 · MEASUREMENT CONTEXT
How it was measured
  • sensor_id — channel identifier
  • location — structural station
  • mounting — stud, magnetic
  • unit — mm/s RMS, g peak
  • sampling_rate — Hz
03 · SIGNAL TRANSFORM
What was computed
  • method — EEMD, CEEMDAN
  • imf_index — 1…n
  • feature_name — RMS, kurtosis
  • parameter_set — sifts, noise
  • ensemble_trials — N
04 · HEALTH EVIDENCE
What it means
  • anomaly_score — 0.0–1.0
  • fault_hypothesis — class label
  • confidence — calibrated p
  • trend — slope, window
  • RUL — hours + CI
05 · MAINTENANCE ACTION
What to do about it
  • inspection_type — borescope
  • decision — defer, replace
  • work_scope — tasks
  • interval — flight hours
  • priority — 1–5
06 · PROVENANCE
Where it came from
  • data_source — device id
  • model_version — v1.2.3
  • transform_version — v2.1
  • approval_status — prod, shadow
  • record_id — UUID chain
Standards alignment at the field level

Each canonical record links directly to established aviation and condition-monitoring standards. This is not a mapping layer bolted on top — it is embedded at the JSON-LD @context level so that every field carries its standard identity as it moves through the fleet.

ISO
20816-1
Mechanical Vibration — Measurement and Evaluation

Severity zones A–D, alarm thresholds, and asset class codes applied at the field level in every canonical record. Supersedes ISO 10816.

ATA
iSpec 2200
Aircraft Technical Data Standard

Every classified field receives an ATA chapter code (e.g. 72-00-00 for power plant). Aircraft system hierarchy preserved throughout the ontology and downstream outputs.

S1000D
Technical Publications Specification

Data module codes referenced for any maintenance action emitted by L6. Enables direct linking of diagnostic output to task cards and work packages.

ISO
13374
Condition Monitoring Information Flow (OSA-CBM)

Six-layer information flow (data acquisition through advisory generation) forms the backbone reference model. Nadir AI extends this with federated learning and governance layers.

L3 · LOCAL DIAGNOSIS & PROGNOSTICS
Calibrated probabilities, not overconfident verdicts

Maintenance trust requires calibrated uncertainty. A 90% confidence prediction that's actually right 60% of the time will erode maintenance engineers' trust in the system faster than any false positive. Nadir AI uses XGBoost ensembles with temperature scaling on a held-out calibration set — producing probabilities whose expected calibration error (ECE) meets safety-grade thresholds.

Ensemble architecture. For each asset class (bearing, gearbox stage, shaft), Nadir AI trains a dedicated ensemble of XGBoost classifiers for fault typing and XGBoost regressors for remaining useful life (RUL) estimation. Models consume the full ontology-mapped feature set from L2.

Temperature scaling. Raw model outputs are passed through a single-parameter temperature scaling layer fit on a validation partition. This compresses or stretches the softmax distribution until predicted confidence matches observed accuracy — without changing the ranking of predictions. ECE drops from 0.052 to 0.034 across the full pipeline.

Uncertainty + explanation. Every prediction is packaged with (a) per-class probabilities, (b) expected calibration error for the prediction, (c) a trend score — the slope of the health indicator over the last N flights, and (d) an RUL estimate with confidence interval. An explanation module surfaces the most influential features so maintenance engineers see why a given recommendation was made.

Safety posture. In aviation, false alarms trigger unnecessary inspections that waste time and erode trust. Missed faults have safety consequences. Calibrated probabilities aren't a performance metric — they're a safety requirement.

▸ L3 Output — Diagnostic Record
asset: MGB-STG2-BRG-07
{
  "asset_id""MGB-STG2-BRG-07",
  "fault_probs": {
    "healthy":    0.184,
    "inner_race"0.724,
    "outer_race"0.072,
    "cage":       0.020
  },
  "calibration": {"ece"0.034},
  "trend_7d": {
    "slope"+0.021"dir""degrading"
  },
  "rul": {
    "point_hrs"186,
    "ci_95": [148232]
  },
  "top_features": [
    "imf3.kurtosis""imf2.rms",
    "imf3.peak_f_shift"
  ],
  "model_version""brg-ens-v3.1"
}
FAULT CLASS
72%
INNER RACE
ECE
0.034
CALIBRATED
TREND
DEGRADING
RUL (HRS)
186
± 42 @ 95%
Fleet Health Intelligence

A living AI model of
your entire fleet.

Nadir AI doesn't just analyze individual aircraft. It builds a continuously evolving, privacy-preserving AI model of fleet health — learning from every test, every flight cycle, every maintenance outcome across every platform under contract.

From disposable test results to permanent fleet intelligence

Current vibration monitoring produces a pass/fail verdict and then discards the underlying data. Nadir AI transforms that same data into a permanent, comparable, longitudinally trackable health record — and feeds it into a global AI model that improves with every new data point across the fleet.

The result: bearing degradation curves built from real operational data. Vibration signature drift detected across flight cycles and platforms. Cross-platform anomaly detection that surfaces failure modes before any single operator would see them. Component life prediction grounded in actual usage patterns, not theoretical schedules.

Each operator contributes to the global model — and each operator benefits from everyone else's learning. The fleet becomes smarter with every aircraft, every test, every flight hour.

"To transition from reactive to predictive maintenance, data must be transformed into actionable, shareable intelligence." — IATA Aircraft Health Management White Paper, 2023
SCENARIO 1 — GROUND
Ground-Based Streaming at Landing

Aircraft lands. ACES device connects. Instead of a pass/fail that disappears, vibration data streams into Nadir AI's pipeline — enriched, normalized, and delivered to your cloud destination within seconds. Every test becomes a permanent, queryable fleet intelligence asset.

SCENARIO 2 — IN-FLIGHT
In-Flight Live Streaming

Nadir AI's edge-deployed classification normalizes and prioritizes high-value channels for in-flight transmission. Full semantic enrichment completes on ground receipt. Maintenance crews receive actionable intelligence before the aircraft touches down — anomalies caught in the air.

SCENARIO 3 — FLEET AI
Fleet Health AI Model

All aircraft. All platforms. All test cycles. Contributing to a single federated intelligence model that predicts failures, surfaces degradation patterns, and generates maintenance intelligence that no individual operator's data could produce alone. This is the game.

L4 · Federated Fleet Learning

Collective intelligence.
Zero raw data shared.

The most powerful capability Nadir AI delivers — and its most critical design constraint. Every operator contributes to the global fleet model without ever exposing raw vibration data. Formal differential privacy guarantees are built in at the mathematical level.

How federated learning works in Nadir AI

At each training round, each operator site computes per-feature histograms of gradients and Hessians from their local dataset — clipped to a maximum L2 norm as a sensitivity bound. The central server aggregates these histograms, adds calibrated Gaussian noise for differential privacy, and selects optimal model splits using the XGBoost gain formula.

No raw vibration records ever leave the operator's environment. The only data transmitted is statistical gradient summaries — which contain no individual data points and provide formally quantifiable privacy guarantees expressed as a privacy budget ε.

The result: each operator's local diagnostic model benefits from the collective learning of every aircraft in the federated fleet, while their sensitive operational data stays within their own infrastructure. Operators who would never share raw data with competitors can still contribute to — and benefit from — shared fleet intelligence.

Validated: Federated learning achieves 95% of the performance of a fully centralized model while reducing communication bandwidth 73.3%. Privacy budget ε = 0.68 at δ = 10⁻⁵ with less than 2% utility loss versus the non-private model, accounted for using the moments accountant.

Federated Server — Global Model θ
Aggregates clipped histograms + Gaussian noise · Selects optimal splits
↕     ↕     ↕
Operator 1
Local data only
gradient histograms
Operator 2
Local data only
gradient histograms
Operator K
Local data only
gradient histograms
Raw vibration data never leaves operator environment
Differential Privacy Guarantee
ε = 0.68  ·  δ = 10−5  ·  < 2% utility loss
BANDWIDTH REDUCTION
73.3% vs. raw
VS. CENTRALIZED MODEL
98% performance
LOCAL-ONLY AUROC
0.88 baseline
FEDERATED AUROC
0.93 fleet model
L5 · Governance Control Plane

Every update validated.
Every change reversible.

Aviation demands more than AI that works — it demands AI that is traceable, controllable, and safe. Nadir AI's governance control plane runs cross-cutting across all pipeline layers, ensuring every model update is shadow-validated before deployment and unsafe changes automatically roll back.

🔬
Shadow Mode
Proposed changes run in parallel with production — predictions logged but not acted upon
📊
Validation Metrics
Accuracy, calibration, false positive rate monitored against pre-defined thresholds
Approval Gate
Human approver grants permission. Zero false promotions recorded during evaluation period
🚀
Fleet Propagation
Approved update distributed to all sites via federated mechanism with full versioning
↩️
Auto Rollback
Performance degradation triggers revert within 1 hour. Full version history preserved
FAA
AC 43-218
Integrated Aircraft Health Management Systems

Governed adaptation model, shadow validation, and approval workflows aligned with FAA operational authorization requirements. Every change to diagnostic logic is traceable and reversible.

EASA
2024/009/R
Vibration Health Monitoring for Rotorcraft

EASA 2024 rulemaking on certifiable fidelity for VHM systems. Provenance embedding and approval gates meet continuing airworthiness integration requirements.

NASA
TM-2023
Barriers to Predictive Maintenance in Aviation

Directly addresses the five barriers NASA identified: data availability, validation, regulatory acceptance, cost, and decision value — through governed adaptation and provenance.

IATA
AHM 2023
Aircraft Health Monitoring to Management

IATA's framework for transitioning from reactive to predictive maintenance. Nadir AI operationalizes the "actionable, shareable intelligence" principle at the architecture level.

L6 · DECISION ORCHESTRATION
Analytics become work packages.

The final layer turns diagnostic output into maintenance action. For each aircraft under monitoring, Nadir AI produces a prioritized component list, recommended inspection type, suggested interval, confidence level, and a concise explanation — packaged for ingestion into the operator's maintenance planning system. The loop closes when outcome feedback flows back into the learning pipeline.

From probability to plan. A calibrated fault probability of 0.72 on an inner race bearing with a degrading 7-day trend and an RUL of 186 flight hours doesn't just become an alert — it becomes a specific, scoped work recommendation: borescope inspection at next scheduled check, with defer authority up to N flight hours.

Structured output. Recommendations carry the inspection type (borescope, oil analysis, vibration re-test), decision class (defer, replace, monitor), work scope tasks, interval target, and a priority score. All fields are S1000D-compatible so they can be linked directly to task cards and work packages.

Audit trail. Every recommendation carries its full provenance chain — from raw signal through L1 IMFs through the exact model version that produced it. Ops authorities can trace any maintenance action back to the originating data point.

Closed-loop learning. When a recommendation is followed, the outcome (confirmed fault, no fault found, different fault discovered) feeds back into L3 local diagnosis and L4 fleet learning. The fleet becomes smarter with every dispatch.

▸ L6 Output — Maintenance Recommendation
aircraft: N748XA
{
  "aircraft""N748XA",
  "priority_actions": [{
    "asset_id""MGB-STG2-BRG-07",
    "hypothesis""inner race fault",
    "inspection""borescope + oil debris",
    "decision""schedule_defer_180fh",
    "priority"2,
    "s1000d_dm""DMC-...-BRG-MGB-A-...-01",
    "confidence"0.72,
    "explanation": [
      "IMF3 kurtosis +38% over 7d",
      "peak f shift toward BPFI",
      "fleet prior: 0.68 → inner race"
    ]
  }],
  "audit_chain""record_7f3c→imf→diag→rec"
}
Validated Performance

Peer-reviewed results.
Real datasets.

Validated on NASA IMS bearing run-to-failure datasets, simulated 50-aircraft × 100-cycle fleets, and controlled fault injection. Five-fold cross-validation, chronological splits, paired t-tests at p < 0.01. Effect sizes above Cohen's d = 0.8 across all primary metrics. Full methodology in Dr. Reza Nehzati's peer-reviewed paper.

94.7%
Classification Accuracy

Full framework vs. 87.3% raw-features baseline. 7.4pp improvement, statistically significant at p < 0.01.

34.6%
Calibration Error ↓

ECE reduced from 0.052 to 0.034. Predicted confidences now match observed accuracy — safety-grade.

96%
Semantic Mapping Precision

500-file corpus across three systems with inconsistent schemas, units, and naming. Field mapping recall: 94%.

73.3%
Bandwidth Reduction

Federated histogram aggregation vs. transmitting raw data. 98% of centralized model performance retained.

Framework Configuration — Ablation Study
Accuracy
Calibration Error (ECE)
Baseline — raw features only
0.873
0.052
+ EEMD signal transformation
0.912
0.041
+ Semantic normalization (L2)
0.918
0.040
+ Calibrated diagnosis (temperature scaling)
0.922
0.037
+ Federated fleet learning (L4)
0.935
0.036
+ Governance control plane — Full Framework
0.947
0.034
Get Started

The data is already being collected.
It just needs somewhere to go.

Request a technical briefing with the Nadir AI team. We'll map your current ACES output against the full pipeline and show you exactly what fleet health intelligence looks like on the other side.

Request Technical Briefing Download Architecture Paper