IoT智能防火系統香港｜AI誤報過濾預測性維護 BMS整合｜藍鋼 | IoT Fire Safety Hong Kong | AI False Alarm Filter + BMS

🌡️

Detection TypePhotoelectric smoke detection, with optional fixed-temperature / rate-of-rise heat sensors.
Communication ProtocolLoRaWAN, suitable for low-power, long-range, multi-point building deployment.
Power SupplyCR123A battery with target life of around 5 years; PoE / wired-powered models available for selected locations.
Detection CoverageApprox. 60–80 sqm per sensor, subject to ceiling height and site layout.
CertificationProject-based selection of FSD-approved models where applicable, EN 14604 or UL 268 certified products.
Protection / TemperatureIP42 for indoor models; IP65 available for outdoor or semi-outdoor locations. Operating temperature: -10°C to +55°C.

🤖

AI AlgorithmConvolutional Neural Network (CNN) for visual / sensor feature analysis, combined with time-series analysis for anomaly trends.
False Alarm ReductionTarget reduction of 70–90% for common false alarms; actual figure to be confirmed after site PoC, sample data and tuning.
Recognisable ScenariosCooking steam, incense smoke, dust, humidifiers, temporary moisture and environmental interference.
Learning MethodOn-device edge AI for real-time decision making, with optional cloud-based continuous learning and rule updates.
Decision LatencyTarget preliminary alarm classification within < 3 seconds.
System LinkageIntegration with BMS / control centre via JSON API, Modbus TCP or BACnet/IP.

🔧

Monitoring MetricsBattery voltage, signal strength (RSSI / SNR), sensor sensitivity drift, offline duration and communication stability.
AI PredictionFailure-risk prediction based on historical data, with a target maintenance lead time of 7–30 days.
Maintenance DashboardAutomatic monthly device health reports with real-time alerts for low battery, weak signal, offline status and abnormal drift.
Integration MethodEmail, SMS, WhatsApp notification or API integration with internal work order systems.

🏢

支援平台可按項目對接 Honeywell EBI、Siemens Desigo、Johnson Controls Metasys、Schneider EcoStruxure 等 BMS 平台。
通訊協定BACnet/IP、Modbus TCP、OPC UA、MQTT 及 JSON API。
整合範圍火警、煙感、熱感、CCTV、門禁、公共廣播及中控告警畫面聯動。
自動化情境火警觸發 → 解鎖逃生門 → 啟動廣播 → 切換電梯模式 → CCTV 自動彈出相關畫面。
管理價值將分散訊號集中成可追蹤、可記錄、可回放的應急流程。

Supported PlatformsProject-based integration with BMS platforms such as Honeywell EBI, Siemens Desigo, Johnson Controls Metasys and Schneider EcoStruxure.
ProtocolsBACnet/IP, Modbus TCP, OPC UA, MQTT and JSON API.
Integration ScopeFire alarm, smoke detection, heat detection, CCTV, access control, public address and central alarm display linkage.
Automation ScenarioFire alarm trigger → unlock escape doors → activate public announcement → switch lift mode → auto-pop relevant CCTV views.
Management ValueConverts scattered signals into a traceable, recorded and reviewable emergency workflow.

防火，由被動變主動 From Reactive to Proactive Fire Safety