Skip to content
Ordanis
PT

Ordanis Reliability · first commercial product

Know why the line stops — and prove what changed.

Ordanis Reliability reads your machine telemetry without touching operations, classifies downtime and microstops, detects deviations and links every finding to an action, an outcome and the value it produced — all auditable.

Request Pilot How the pilot works

PCK-003 · Industrial Packaging Machine

Demo
Temperature
91.7 °C+14%
Vibration
6.3 mm/s+40%
Microstops (24 h)
47 min
Spare motor
SHP-4821 · +2 daysdelayed
Downtime avoided
4.2 h
Estimated event value
~R$ 182k
Illustrative scenario · fictional data, not a customer result.

The problem

  • 01

    Stops without a clear cause

    The stop shows up in the shift report, but with no reason, no reliable duration and no link to maintenance.

  • 02

    Invisible microstops

    Stops of a few seconds add up to hours per month and never reach any system.

  • 03

    Metrics without a method

    MTBF and MTTR computed in different spreadsheets with different definitions — impossible to audit.

  • 04

    Value nobody validates

    Improvements are announced, but nobody can show the baseline, the action and the outcome.

What Ordanis Reliability does

Every number has a versioned definition, a source and lineage. When the data is not enough, the answer is INSUFFICIENT_DATA — never invented precision.

  • 01

    Machine state

    Derived from telemetry with per-asset rules, processed continuously.

  • 02

    Downtime and microstops

    Seven classes (planned, unplanned, microstop, changeover, starved, blocked, unknown); configurable microstop threshold.

  • 03

    MTBF, MTTR and availability

    Failures from telemetry, CMMS import and manual records, with a versioned method.

  • 04

    Baselines and detection

    Baselines persisted per signal and asset; every detection shows the trace behind it and can be replayed.

  • 05

    Data quality

    Explainable score per signal: availability, freshness, completeness, confirmed unit and mapping.

  • 06

    Maintenance history

    Work orders and asset register imported from your CMMS with preview, idempotency and per-row lineage.

From finding to validated value

The whole loop is recorded. Author and approver are always different people.

  1. Incident

    Opened from a detection or by a person.

  2. Recommendation

    With evidence and sources; it never executes anything on its own.

  3. Approval

    A person with the right role approves or rejects.

  4. Work order

    Native or in your CMMS, with an owner and a due date.

  5. Outcome

    Confirmed by whoever did the work and checked against telemetry.

  6. Lesson and value

    Value is VALIDATED only with a confirmed outcome and plant manager approval.

Safe for the plant

  • 01

    Read-only

    OPC UA in read mode. No code path writes to a PLC.

  • 02

    Outbound-only

    The Ordanis Edge opens the connection; no port is opened on the OT network.

  • 03

    Audit

    Every change has an author, a time, a before and an after — in an append-only log.

Safe for the plant →

Pilot KPIs

Every KPI has a documented definition. Targets are set only after the baseline on your data — we never promise a fixed reduction up front.

  • Unplanned downtime
  • Microstops
  • MTBF
  • MTTR
  • Availability
  • OEE (when data exists)
  • Failure detection lead time
  • Maintenance response time
  • Validated value

Honest status

Software implemented and tested in our repository, against an OPC UA simulator. Not yet validated on a physical PLC or with customer data — that is exactly what the first pilot proves.

What it is not

  • It does not replace your CMMS: it integrates (CONNECT) or selectively provides work orders (NATIVE)
  • Not "black-box" predictive maintenance: every finding shows its evidence
  • It does not control machines
Request Pilot