VERAXIS
Condition Monitoring · Alpha

Machinery health · sensed at the source

Rotating machinery reports its own condition — before it fails.

Veraxis is the engine that makes the call. It reads a running machine at the source and returns a deterministic verdict with supporting evidence — ahead of failure. It runs on the operator's own hardware, takes its input from whatever acquisition is already in place, and nothing leaves the site.

Function
Detect · Identify · Grade · Predict
Deployment
Self-contained · operator-installed
Data
Sovereign · stays on-site
Connectivity
Wired or wireless

Operational context

The warning of a failing machine is present in the signal long before the machine stops.

The early indication of a developing bearing, winding, or rotor fault is already carried in the machine's current and vibration. Existing approaches miss it — at a cost measured in unplanned downtime.

Periodic inspection misses onset

A route walked every few weeks cannot observe a fault that develops between visits. The available lead time is gone before anyone reaches the asset.

Fixed monitoring is heavy and exposed

Permanent online systems require per-asset installation, integration, and in most cases transmit operational data to an external network the operator does not control.

Threshold alarms lack discrimination

A bare limit trips on ambient disturbance as readily as a genuine defect. Operators learn to disregard it — and miss the indication that matters.

Capability

Four determinations from a single measurement core.

Veraxis establishes each machine's own healthy reference, then continuously measures departure from it. The same core operates across signal types and fault classes.

01 / Detect

Condition has changed

The instrument registers departure from the machine's established healthy reference. Common-mode and ambient content is rejected against that reference.

02 / Identify

The fault is classified

The indication is resolved to a fault class — bearing race, rolling element, rotor, or winding — by where the energy concentrates relative to the machine's geometry.

03 / Grade

Severity is quantified

The magnitude of departure tracks fault severity, distinguishing an incipient fault from a developed one.

04 / Predict

Lead time is established

The trend rises ahead of failure. On run-to-failure data, onset is observable well before the machine reaches the end of service — the interval in which action can be taken.

What is delivered

The deliverable is the determination engine. Everything around it is yours.

Veraxis is the element that makes the call — it reads the signal and returns the verdict. Acquisition, storage, transport, and display are open at both boundaries: feed it from whatever you already run, and route the determination into whatever you already use. What you procure is the judgment, not the pipeline.

Open at both boundaries
Defined inputs in, defined verdict out. Operators feed the engine from existing acquisition and consume the result into existing systems. No proprietary ingest, no proprietary output, no sensor lock-in. Where transducers are supplied, inputs are landed by the operator's own engineers.
Delivery scaled to existing hardware
Where no instrumentation exists, the engine ships with a self-contained sensing and compute package. Where the platform already senses and computes — as fielded systems do — the engine is delivered as software onto that existing hardware. No box is added that the platform does not need.
Sovereign by design
The engine runs entirely on the operator's hardware, on the operator's network. Nothing is transmitted off-site. The determination, and all data behind it, remains under the operator's control.
Inspectable, on your cadence
Capability is delivered as a discrete artifact the operator inspects, validates, and loads through its own process — up to and including push onto fielded hardware through an existing update path. No standing connection to the supplier.
Open boundary
Operator acquisition
existing sensors · existing pipeline · operator's choice
▼
Veraxis engine — the determination
reads the signal · returns the verdict · runs on operator hardware
▼
Operator systems
existing readout · existing health management · operator's choice

Validation

Capability demonstrated against established public fault datasets.

Performance has been confirmed across the principal rotating-machinery fault classes, using independently published, labelled reference data. Detection and classification are evaluated against ground truth — not asserted.

Rolling-element bearingCWRU reference
Inner race, outer race, and rolling-element faults detected and correctly classified against geometry-predicted signatures.
Confirmed
Induction motor — rotorBroken-bar reference
Broken rotor bar detected from the current channel across the full load range, including light load where conventional sideband methods lose the indication.
Confirmed
Stator windingInter-turn reference
Incipient inter-turn faults detected and graded at low fault percentage, where the indication is otherwise weak.
Confirmed
PrognosisRun-to-failure reference
Fault onset identified well ahead of end-of-service on full run-to-failure data, establishing actionable lead time.
Confirmed
Envelope spectra — healthy reference compared with a developed bearing fault

Rolling-element bearing — envelope spectra. Healthy reference (upper) compared with a developed outer-race fault (lower); energy concentrates on the geometry-predicted defect harmonics. CWRU public dataset.

Broken rotor bar current-channel result across load

Induction motor, broken rotor bar — current-channel result. The indication holds across the load range, including light load where conventional sideband methods lose it. Public broken-bar dataset.

Stator inter-turn fault graded response

Stator winding, inter-turn fault — graded response at low fault percentage. Public inter-turn dataset.

Run-to-failure prognosis trend

Prognosis — run-to-failure trend. Fault onset is observable well ahead of end-of-service, establishing actionable lead time. NASA/IMS public run-to-failure dataset.

Quantified results, evaluation methodology, and complete validation records are provided to qualified evaluation partners under non-disclosure.

Class of system

Continuous-monitoring capability without the installation, exposure, or recurring dependency.

Veraxis is not a portable analyzer carried between machines, and it is not a fixed appliance that owns your pipeline. It is the determination engine — the always-on layer that makes the call — delivered to run on the hardware you already field, with acquisition and output left open to what you already use.

Veraxis
  • +The determination engine — you procure the judgment, not the pipeline
  • +Runs on operator hardware; data stays on-site
  • +Open acquisition and output — no lock-in either end
  • +Detect · identify · grade · predict
  • +Verdicts delivered with supporting evidence
  • +Operator integrates and certifies; supplier never touches the platform
Fixed online monitoring
  • −Per-asset wired install and integration program
  • −Operational data typically transmitted off-site
  • −Vendor personnel on site; clearance burden
  • −Recurring service / analyst relationship
  • −Proprietary hardware locked to the supplier

Defense & platform integration

You integrate it. You certify it. We never touch the platform.

For fielded systems, the engine is validated as far as it can be validated independently — and the system it runs on is its own certification. That boundary is the point. It is what lets the capability cross onto a platform without the supplier crossing with it.

Compartmentalized by construction
The supplier delivers the engine and a defined interface — and nothing else. The platform's architecture, data, and operation are never exposed to the supplier. No read-in, no site access, no embedded personnel, no clearance burden on the supplier side.
Last-mile integration stays with the operator
The operator's own cleared personnel integrate the engine and certify it on the target system, behind the operator's walls, under the operator's authority. The capability that arrives is independent of any single platform; certifying it on a given platform is the operator's to perform.
Validated here is not certified there — by design
Independent validation establishes that the engine works. Each system remains its own certification. The supplier does not, and cannot, assert performance on a platform it has never seen — and is not asked to. The operator proves it on the target, with what the supplier provides to do so.
Delivered as an inspectable update
Where the platform already senses and computes, the engine is delivered as software the operator inspects and pushes onto fielded hardware through an existing update path. No new hardware, no supplier connection, no standing dependency.
The boundary
Supplier provides
the engine + a defined interface · nothing more
▼
— compartment boundary —
supplier learns nothing of the platform · operator learns nothing of the method beyond the interface
▼
Operator performs
integration · certification · operation — on their system, behind their walls

Qualified evaluation partners

The measurement method, quantified performance, and complete validation record are held as proprietary and disclosed to qualified partners under NDA. A test unit is made available to evaluation partners at hardware cost, with full margin visibility from the outset.

Request evaluation

Alpha · evaluation partners

Bring a machine. Confirm it on your own hardware.

Veraxis is a working prototype seeking qualified partners to confirm performance in the field. A test unit is provided at hardware cost. Where it earns its place, it proceeds to production jointly — with full visibility into the program.

Request evaluation Technical brief (NDA)