Sensor-to-work-order in one system. Edge inference keeps the data on your floor; our ML models handle the diagnosis.
Three layers: physical sensing, edge compute, and maintenance system integration.
Clip-on tri-axial MEMS accelerometers (±16g, 10kHz sampling), PT100 temperature probes, and Rogowski current transducers attach without modifying the asset. IP67-rated enclosures, 4–20mA and IO-Link connectivity, and a 5-year battery backup for remote installations.
The EdgeRun EG-1 gateway runs an ARM Cortex-A55 processor at 1.8GHz with 4GB RAM. Inference is optimized using TensorRT and runs at sub-50ms latency. Up to 64 sensors per gateway. Connectivity: Ethernet, Wi-Fi 6, LTE Cat-M1, and OPC-UA for DCS integration.
Aggregated health scores and trending plots are visible in the EdgeRun web dashboard. Alert events push directly to your CMMS as structured work order objects — no manual ticket creation, no missed shift handoffs.
Capability descriptions written for maintenance engineers, not marketers.
The platform computes real-time Fast Fourier Transform on raw accelerometer data, decomposing the vibration signal into frequency bands associated with specific failure modes.
Fault frequency libraries for over 4,800 bearing models (SKF, NSK, Timken, FAG) ship pre-loaded. Gear mesh frequency, blade pass frequency, and shaft harmonics are tracked independently per asset.
EdgeRun does not use fixed alarm thresholds. Each asset builds a probabilistic baseline model that accounts for production speed, load profile, and ambient temperature.
A stamping press running at 40% capacity has a different normal signature than the same press at full load. EdgeRun's load-conditioned baseline reduces false positive rates by 61% compared to threshold-based systems.
After an anomaly is detected, the RUL engine projects failure timing using a degradation curve fitted to the asset's historical fault progression data.
Output: an estimated hours-to-failure range with 80% confidence interval. Maintenance planners use the RUL estimate to decide whether to repair at the next scheduled stop or expedite immediately. Accuracy validated at ±14% across a test population of 312 bearing failures.
EdgeRun writes structured alert data directly into your CMMS using native API connectors, not generic webhooks.
Supported systems: IBM Maximo 7.6+, SAP PM/EAM, Fiix (REST API), Limble CMMS, and eMaint. Each work order includes: asset ID, fault mode classification, anomaly score trend, sensor readings, and recommended action from the fault library. Connector configuration takes under 2 hours for a standard Maximo install.
EdgeRun gateways expose an OPC-UA server, allowing existing SCADA and DCS systems (Siemens WinCC, Rockwell FactoryTalk, Inductive Automation Ignition) to subscribe to asset health tags.
Plant operators see EdgeRun health status overlaid on their existing HMI screens. No separate monitoring interface for operators to learn. Integration tested with Siemens S7-1500 and Allen-Bradley ControlLogix PLCs.
All sensor data is encrypted in transit using TLS 1.3. At rest, edge storage uses AES-256. The platform supports role-based access control mapped to your Active Directory groups.
For customers who cannot allow any cloud connectivity, EdgeRun operates fully air-gapped. Model updates are delivered as cryptographically signed firmware packages via USB or isolated transfer server. Audit logs are immutable and stored locally for 24 months.
| Parameter | Specification |
|---|---|
| Processor | ARM Cortex-A55, quad-core 1.8GHz |
| Memory | 4GB LPDDR4 RAM, 64GB eMMC storage |
| Inference latency | <50ms per sensor per cycle |
| Max sensors per gateway | 64 (vibration + temp + current) |
| Connectivity | Ethernet (RJ45), Wi-Fi 6 (802.11ax), LTE Cat-M1, RS-485 Modbus |
| Industrial protocol | OPC-UA server, IO-Link master (4 ports) |
| Operating temperature | -20°C to +70°C |
| Ingress protection | IP65 enclosure |
| Certifications | CE, FCC, UL, IEC 61010-1, IECEx (ATEX Zone 2 on EG-1X variant) |
| Power supply | 24V DC (IEC 60715 DIN rail), UPS input supported |
| Firmware update | OTA (TLS), signed USB, isolated TFTP server |
| Model update cycle | Quarterly baseline model refresh, critical patches within 7 days |
Remote kickoff plus an on-site walkthrough. We identify priority assets, confirm sensor mounting points, and document existing CMMS configuration.
Sensors and gateway deployed. Network connectivity tested. OPC-UA or CMMS connector configured and validated in a staging environment before production cutover.
14-day calibration window. EdgeRun collects operating data across multiple production schedules. Baseline model reviewed and signed off by our application engineer.
Real-time anomaly detection active. Your maintenance team receives the first alert within normal CMMS workflows. Weekly review calls for the first month included.
Book a technical session with our application engineers. We'll scope a pilot for one asset class in your facility.
