1.1 System Architecture

The smart campus environmental monitoring system is structured around three clearly delineated deployment boundaries: the field layer, the zone edge layer, and the central layer. Each boundary has distinct responsibilities, hardware requirements, and failure modes. Understanding these boundaries is essential for correct system sizing, redundancy planning, and O&M staffing.

The field boundary encompasses all stations and probes installed on poles, rooftops, discharge manholes, pump rooms, boundary fences, and loading bays. These devices are exposed to the full range of environmental conditions and require robust enclosures, surge protection, and accessible mounting for maintenance. The zone edge boundary consists of aggregation cabinets or telecom rooms housing gateways, PoE switches, surge protection devices, and optional UPS units. The central boundary is either an on-premises server room or a cloud virtual private cloud containing platform services, storage systems, and integration endpoints.

System Architecture with Core, Optional, and Support Boundaries

Figure 1.1: Layered Component Architecture — Core, Optional, and Support Boundary Zones

Deployment Boundaries

Boundary Location Key Hardware Primary Function
Field Poles, rooftops, manholes, fence lines, loading bays Sensor stations, probes, enclosures, SPDs Measurement, basic self-diagnostics
Zone Edge Aggregation cabinet or telecom room per zone Gateway, PoE switch, UPS (optional), surge protection Protocol mediation, buffering, local rules, uplink management
Central On-prem server room or cloud VPC Application servers, TSDB, object storage, GIS, firewall Governance, analytics, reporting, audit, integration

Data & Control Flow

Data flows from sensors through protocol adapters to the edge buffer, then via secure MQTT/HTTPS transport to the central broker, through stream processing to storage, and finally to API/UI/reporting endpoints. Control flows in the opposite direction: the platform pushes configuration updates including sampling intervals, threshold values, and OTA firmware to the edge, which applies them and logs the changes before the station updates its behavior.

1.2 Components & Functions

The following inventory diagram and component table provide a complete engineering-level description of every major component in the system. Each component entry includes its primary responsibility, input and output specifications, key performance indicators, and the most common mismatch risks encountered during deployment and commissioning.

Component Inventory Bill of Blocks

Figure 1.2: Component Inventory Diagram — Sensing Devices, Connectivity, and Platform & Integration Blocks

Component Primary Responsibility Inputs Outputs Key KPIs (Typical) Common Mismatch Risk
Ambient Air Station (PM/NO₂/O₃/CO/SO₂) Boundary and roadside ambient monitoring Sensor signals, meteorological data 1–5 min averages, QA flags Data completeness ≥98%, clock drift <2s/day Wrong inlet height; poor airflow; condensation
Dust/PM Hotspot Node Construction and yard dust event detection PM sensor, wind data Event spikes and trend data Event detection latency <60s Sensor saturates; wrong filter/heater configuration
Odor/VOC Node (PID/metal-oxide) Odor complaint correlation TVOC/odor index readings Rate-of-change alarms Drift control; false alarm rate KPI Cross-sensitivity to interferents; no calibration plan
Noise Monitor Boundary noise compliance measurement SPL microphone signal Leq/Lmax, spectral data Class compliance per applicable standard Windscreen missing; acoustic reflections
Weather Station Dispersion context and safety support Wind speed/direction, rain, T/RH 1-min meteorological feed Wind accuracy per specification Poor siting; tower turbulence effects
Water Discharge Monitor (pH/ORP/EC/Turbidity/COD) Effluent excursion detection Probe readings from flow cell Alarms and compliance reports Probe health score; cleaning cycle compliance Fouling; air bubbles; wrong flow cell sizing
Energy Meter (3-phase) Energy monitoring and correlation CT/PT signals kWh, demand, power quality data Accuracy class per specification; time sync CT ratio mismatch; wiring polarity errors
Edge Gateway Normalize, buffer, local rule execution Multi-protocol device data MQTT/HTTPS uplink, local alarms Buffer ≥7 days; uplink failover <30s Under-sized CPU/storage for device count
Central Platform Store, analyze, visualize, workflow management Data streams, configuration events UI, alarms, reports, APIs Ingest capacity, retention SLA, RBAC enforcement Poor schema governance; retention policy gaps

1.3 Working Principles

Startup Sequence

System startup follows a structured commissioning sequence designed to ensure data quality from the first measurement. The process begins with station commissioning, during which device IDs, geographic coordinates, and channel mappings are registered in the asset inventory. The edge gateway then registers the device, validates time synchronization against the NTP server, and applies the assigned sampling profile. Finally, the platform verifies data continuity and QA flag behavior, enabling alarm rules only after a stabilization window of 30 to 120 minutes depending on the sensor technology — electrochemical sensors require longer warm-up than optical sensors.

Normal Operation

During normal operation, sensors sample at their configured intervals, the edge gateway applies timestamps and plausibility validation, compresses and forwards data via secure transport, the platform stores and evaluates the data against rule sets, alarms notify operators and create work tickets, and O&M staff close tickets with on-site verification evidence. This closed loop ensures that every anomaly is tracked from detection to resolution.

Exception Chains

Exception Chain A — Network Outage → Alarm Integrity Maintained: When uplink loss exceeds 60 seconds, the edge gateway buffers all data locally and continues executing local rules to generate alarms via buzzer, relay, or onsite messaging. When the link is restored, the edge backfills data with original timestamps and integrity checksums, and the platform marks the backfilled segments for audit purposes.
Exception Chain B — Sensor Drift/Failure → Quality Flags Prevent Wrong Decisions: When a plausibility check fails — for example, PM constant at zero for six hours, VOC stuck at a fixed value, or pH out of physical range — the edge sets a QA=Suspect flag. The platform suppresses compliance score calculations for that channel and automatically creates a calibration ticket. Recovery requires probe replacement, a post-calibration burn-in period, and comparison with a reference instrument.
Exception Chain C — Power Event at Cabinet → Safe Shutdown and Evidence Preserved: When AC power fails and the UPS activates, the platform receives an "on-battery" event. The edge reduces sampling to essential channels only and logs the power state. Upon power restoration, the system ensures time synchronization, verifies storage integrity, and generates an outage report documenting the affected measurement period.