Everything you
need to know.

FSME Logic detects the early signs of structural stress accumulation in equipment — We call this the moment a system "stops telling the truth" about its own health.

Standard monitoring waits for a sensor value to exceed a preset threshold and triggers an alarm. By that point, the failure process is already well underway. FSME Logic identifies the stress signature that precedes the threshold breach — often weeks, months, or in aerospace cases, up to 9 months before standard monitoring would trigger any alert.

The result is a planning window instead of an emergency. Time to schedule maintenance, order parts, and prevent unplanned downtime before it happens.

Three key differences:

Detection timing: Standard predictive maintenance still relies on threshold logic — it just uses more sophisticated thresholds. FSME Logic detects structural degradation before any threshold would ever be triggered.

No training data required: Most predictive maintenance platforms need months or years of historical failure examples before they can operate. FSME Logic deploys cold — on any new asset, from the first day of operation, with no failure history on file.

Fully offline: Most platforms require your operational data to be uploaded to cloud servers for processing. FSME Logic runs entirely on local, air-gapped hardware in Edmonton. Your data is never uploaded to any cloud server or transmitted over any network.

FSME Logic has been validated across aerospace telemetry, commercial turbofan engines, industrial bearings, refrigeration systems, orbital satellite hardware, and commercial vehicle fleets. The detection principle is not specific to any single industry or equipment type.

If your equipment produces sensor data — vibration, temperature, pressure, current draw, RPM, — FSME Logic can audit it. The engine works on the signal record, not on industry-specific failure models.

Current focus areas include commercial and industrial fleets, heavy equipment, refrigeration and HVAC systems, and manufacturing machinery. If you're unsure whether your specific equipment is a good candidate, contact us — we'll tell you honestly.

The Green Dashboard Fallacy is the false sense of security created when all your monitoring indicators show normal — while internal structural degradation is already underway.

A system's dashboard can show green on every metric right up until failure. Temperature normal. Pressure normal. Vibration within spec. Threshold alarms silent. And yet the system is in the early stages of a failure process that will become catastrophic.

This is not a rare edge case — it is the normal behaviour of structural degradation before it becomes symptomatic.

Four steps:

1. Data export. You pull a CSV export from your existing monitoring system. Standard data logger exports, SCADA exports, or BMS exports all work. 30–90 days of readings is the ideal window. No special format required.

2. Secure data transfer. You send us your exported CSV via secure transfer. All processing happens on our local, air-gapped hardware in Edmonton — no internet connection, no cloud upload. No data stored.

3. Forensic report. You receive a comprehensive summary identifying which channels showed early warning indicators, what the stress timeline looks like, and what action is recommended. Written for operations managers, not engineers.

4. Custody receipt. A SHA-256 cryptographic chain of custody document confirming your data was handled securely and never transmitted at any stage. Delivered with your forensic report, if requested.

The offline analysis phase processes 100,000+ sensor rows in under 10 seconds on edge hardware. For most fleet or industrial datasets, the data processing happens quickly.

The full engagement — data intake, analysis, report generation — timeline depends on dataset size and complexity. The forensic PDF report and custody receipt are delivered by email once complete.

No. FSME Logic works with your existing data exports. If your equipment is already generating sensor data — which most modern industrial equipment does — that data is sufficient for an audit.

We bring our own analysis hardware to your site. You don't purchase anything, install anything, or change any existing infrastructure to run a Pilot Integrity Audit.

CSV is the standard format. Most SCADA systems, BMS platforms, fleet telematics systems, and standalone data loggers can export to CSV directly.

Ideally the export includes timestamps and multiple sensor channels — the more channels included, the more complete the health picture. 30–90 days of readings provides a good baseline window. Longer histories are better when available.

If you're unsure whether your system can produce the right export, contact us before booking — we can confirm compatibility in advance.

The report identifies:

Healthy channels: sensor channels showing normal stress patterns — confirmed stable with no early warning indicators detected.

Flagged channels: sensor channels where stress accumulation patterns were detected — with the approximate detection step and a plain-language description of what the pattern indicates.

Recommended action: a clear recommendation based on the findings — whether that's no action required, scheduled inspection, or priority maintenance.

The report is written for operations managers and maintenance supervisors, not engineers or physicists. The findings are presented in plain language with no technical jargon.

There is no cloud upload, no data saved or stored. The analysis hardware has no internet connection during processing. Your operational data is handled with full cryptographic chain of custody documentation.

This is not a policy choice that could change — it is a structural property of how the audit works. Air-gapped, offline processing is how the engine runs.

A SHA-256 chain of custody receipt is a cryptographic document that records the hash value of your dataset at the time of intake and confirms that it was handled securely throughout the audit process.

In plain terms: it's mathematical proof that your data was not modified, copied, or transmitted. The receipt is generated locally during the audit and provided to you at the end of the engagement. You keep it permanently as a verifiable record of secure data handling.

For operations with strict data handling requirements — government contracts, regulated industries, or sensitive operational infrastructure — this receipt provides auditable documentation of your data security posture.

Yes. It is required. We are happy to sign a mutual NDA before any data is exchanged or any audit is conducted. Contact us in advance and we will arrange this before any data transfer takes place.

Every validation result was produced against publicly available datasets — the same datasets used to benchmark the world's leading predictive maintenance algorithms. All datasets are independently verifiable and accessible to any researcher.

The engine was deployed cold on each dataset — no historical failure examples were provided before testing, and no parameters were adjusted after seeing the results. Detection outputs were compared against official ground-truth labels after the audit was complete.

The ESA OPS-SAT and ESA Anomaly Detection Benchmark results were produced as blind tests — no prior access to the dataset before running the audit. Results were compared against ESA ground-truth labels after the fact.

A false positive is when the detection engine flags a channel as showing stress indicators when nothing is actually wrong. High false positive rates are the primary reason industrial operators disable monitoring systems — too many false alarms make the system useless in practice.

Across the full 6-month stable baseline window of the ESA Anomaly Detection Benchmark — the ESA researchers' own highest-priority evaluation criterion — FSME Logic produced zero false positive detections. The engine correctly distinguished normal spacecraft operations from genuine structural degradation throughout the entire baseline period.

The ESA benchmark paper specifically cited false positive rate as more operationally important than detection rate or F-score. Every competing algorithm tested failed on this criterion.

That's a good outcome. A clean audit result means your systems are showing healthy stress patterns with no early warning indicators present — which is exactly what you want to know.

You receive the same forensic report and custody receipt regardless of findings. A confirmed-healthy result gives you a documented baseline — useful for future comparison audits, for compliance documentation, and for operational confidence.

We do not manufacture findings. If the data shows healthy systems, the report says healthy systems.

Use the contact form or email us directly at [email protected]. Include a brief description of your equipment type and approximate number of sensor channels in your data export.

We will confirm compatibility with your data format, arrange an NDA if required, and begin the engagement at a time that works for your operation.

Currently serving Edmonton and the surrounding area for om location audits. Including the greater Edmonton Capital Region. For larger engagements or clients outside this area, contact us to discuss — we will let you know honestly whether we can accommodate your location.

Yes. The Pilot Integrity Audit is the entry point — a single engagement to demonstrate what the engine finds in your data. Clients who want ongoing monitoring can discuss periodic re-audit arrangements or longer-term monitoring configurations after the initial engagement.

Contact us to discuss what ongoing monitoring would look like for your specific operation.