VROC reads your P&IDs and functional descriptions to learn how your plant is connected and controlled. Then it watches live data and explains every alert with the evidence behind it.
Example alert, for illustration
Each one is built from your plant's own documents and data, and each one shows its working.
Behaviour models for each asset are trained on a period you confirm as healthy, and they understand operating modes. A pump on standby isn't a pump that has failed.
Every alert names the measurements that drove it.Using the plant model from your P&IDs, VROC traces upstream from the symptom and ranks causes by sensor evidence and timing. It writes a summary that cites the FDS clause or reading behind each step.
Grounded in evidence. No unsupported guesses.VROC projects how long a degrading signal has before it reaches the alarm limit in your functional description.
Works from day one, with no trained model.A health-index trend, or a Weibull life model where failure history exists.
Always shown as a range, never a single date.Advisory recommendations from a process model, kept inside FDS limits and the data the model has seen.
Operators approve. It never writes to control.Most predictive maintenance tools start with tags and guess the rest. VROC starts with how your plant is designed.
DEXPI or vector PDF drawings and your functional description. VROC extracts equipment, connections, control loops and alarm limits.
Engineers check the proposed asset graph and tag map before anything goes live.
An edge gateway streams from OPC UA or DNP3 over an outbound-only connection. Historian CSVs backfill the past.
Features for each equipment type, such as pump head and BEP ratio, exchanger UA and compressor efficiency, are scored every ten minutes.
Triage, then raise a work order in Maximo, SAP PM or MaintainX. Completed work comes back to sharpen the models.
Each VROC gateway runs apps from the VROC library, such as pump station monitoring, lighting schedules, generator supervision and camera-based inspection. Operators get local screens on site and the same screens remotely, all through the gateway's outbound connection.
Models, analytics and AI features can't issue commands. Control happens only through an approved app, within its tier.
Gateways dial out. No inbound ports open at your site, and the OPC UA bridge is read-only.
A gateway runs only signed apps. Each app can read and command only what its approved permissions allow, and the gateway enforces this. No app performs a safety function.
Each site has its own time-series database. Every table has row-level security, and access is set for each organisation and site. Sign-in supports passkeys.
Off-BEP running, blocked strainers, cycle drift and stormwater duty.
Engine state and operating modes, with alerts that know when a set is meant to be idle.
Efficiency and fouling trends, plus camera-based defect inspection at the line.
Schedules, calendars and remote screens for distributed assets.
Send us a P&ID and a few weeks of historian data. We'll show you what VROC finds.
