Engineering journal
Industrial hazard monitoring, term by term
36 terms used across industrial safety, sensing and hazard monitoring. Each one carries a definition you can quote, why it matters operationally, how it actually works, and - where it applies - how it appears in Alfa HazardX.
Monitoring & detection
- Industrial hazardAn industrial hazard is any condition in a working industrial environment with the potential to cause injury, fire, explosion, chemical exposure, equipment damage or unplanned shutdown.
- Hazard monitoringHazard monitoring is the continuous measurement of conditions in an environment over time, so that developing hazardous states are found, recorded and routed to whoever can act on them.
- Hazard detectionHazard detection is the act of determining that a hazardous condition exists now, usually by comparing a sensor reading against a defined limit.
- Threshold detectionThreshold detection classifies a reading by comparing it against one or more fixed limits, producing a discrete state such as safe, warning, danger or critical.
- Risk classificationRisk classification reduces multiple sensor readings to a single ordered state - such as SAFE, MODERATE or CRITICAL - that expresses how dangerous the current situation is.
- Risk state machineA risk state machine models hazard status as a set of discrete states with defined transitions between them, so that events are generated by changes of state rather than by individual readings.
- Local alarmA local alarm is an audible or visual warning produced by the sensing device itself, without depending on a network, server or remote service.
Prediction & analysis
- Time to thresholdTime to threshold is an estimate of how long a monitored channel will take to reach its alarm limit if its current trend continues.
- Baseline deviationBaseline deviation measures how far a current reading has moved from what is normal for that specific sensor in that specific location.
- Z-scoreA z-score expresses how many standard deviations a value sits from the mean of its distribution, making deviations comparable across channels with different units and scales.
- Hazard DNAHazard DNA is Alfa HazardX's behaviour signature: a fixed-length vector computed from a window of sensor readings, built from rates of change rather than raw values, and matched against a library of known hazard patterns.
- Predictive maintenancePredictive maintenance uses condition monitoring to estimate when equipment will need servicing, so that work is scheduled by actual condition rather than by a fixed calendar.
- Hazard genomeA hazard genome is a twelve-element, dimensionless feature vector computed from a rolling window of sensor readings, describing exclusively how those readings are moving rather than what they currently are.
- Sweep prominenceSweep prominence is how far the peak response of a rotational sweep stands above the sweep's mean, expressed relative to the span between that mean and the sweep floor. It is the basis of the confidence attached to a bearing.
- Activity gateAn activity gate is a minimum-magnitude test applied to a behavioural feature vector, below which the window is declared quiet and no signature matching is attempted.
Sensing
- Sensor fusionSensor fusion combines readings from multiple sensing channels into a single assessment, so that patterns spanning several channels become detectable.
- Sensor driftSensor drift is the gradual change in a sensor's output over time for an unchanged input, caused by ageing, contamination or environmental exposure.
- Sensor burn-inBurn-in is an initial powered period a sensing element requires before its readings become stable and trustworthy.
- Cross-sensitivityCross-sensitivity is a sensor's tendency to respond to substances other than the one it is intended to measure.
- Gas detectionGas detection is the measurement of the presence or concentration of combustible or toxic gases in an atmosphere.
- Flame detectionFlame detection identifies the presence of fire directly by sensing the optical radiation a flame emits, rather than by detecting its smoke or heat.
- Machine vibration monitoringMachine vibration monitoring measures mechanical oscillation in equipment to identify wear, imbalance, misalignment or impending failure.
- Industrial air qualityIndustrial air quality describes the concentration of particulates, volatile organic compounds and other contaminants in the air of a working industrial space.
- Rotational sweepA rotational sweep is one complete sampling pass in which a sensing head is rotated through a series of angularly spaced headings, taking a full channel sample at each, to recover the direction of a hazard source from a single fixed installation point.
- Source bearingA source bearing is the compass heading, relative to a sensing node's fixed installation position, of the sector that recorded the strongest response during a rotational sweep.
Systems & architecture
- Fail-safeA fail-safe design ensures that when a component fails, the system moves to a safe state rather than an unsafe or unknown one.
- Industrial IoTIndustrial IoT is the use of networked sensing devices in industrial environments to collect operational data and make it available for monitoring, analysis and control.
- Edge intelligenceEdge intelligence is processing and decision-making performed on the device that collects the data, rather than on a remote server.
- Deterministic analysisDeterministic analysis produces the same output for the same input every time, using explicit rules or arithmetic rather than a learned model.
- Generative AI in safety systemsGenerative AI in a safety system means using a language model to describe, explain or recommend - as an advisory layer over decisions made elsewhere, not as the mechanism that makes them.
- AI gatingAI gating is the practice of invoking an AI service only when defined conditions are met, rather than on every input, to control cost and bound the system's dependence on it.
- Emergency commandEmergency command is the ability to trigger an action on a field device remotely - typically sounding its alarm - from a monitoring interface.
- Digital hazard twinA digital hazard twin is a three-dimensional model of a monitored facility, registered with node positions and a named asset register, onto which a measured hazard source and its forecast spread zones are projected.
- Hazard footprintA hazard footprint is a stored record of one completed localisation - the bearing, the peak magnitude, the estimated source coordinate and the full angular sweep profile - retained so that recurring hotspots at a site can be learned over time.
- Hazard replayHazard replay is an autonomously triggered forensic recording in which a sensing node escalates its reporting cadence on recognising a hazard, preserving per-second frames that carry the system's own contemporaneous reasoning alongside the raw channel readings.
- Burst cadenceBurst cadence is the elevated, incident-triggered data-reporting interval a sensing node adopts in place of its normal baseline cadence, for a bounded duration, when a hazard is recognised.
Where these terms come from
This journal is written from building an industrial hazard monitoring system rather than from surveying the literature, so the entries emphasise the things that turned out to matter in practice - sensor burn-in, cross-sensitivity, drift, the difference between a self-calibration and a traceable one. Where a term describes something Alfa HazardX does not implement, the entry says so.