Chapter 1 — System Components

This chapter describes the complete component architecture of a surface water environmental monitoring system, covering the three deployment zones from in-water hardware through station enclosure to the central cloud platform. It details the responsibilities, inputs, outputs, and key performance indicators of each component, along with common mismatch risks and the working principles governing startup, normal operation, and exception handling.

1.1 System Architecture

The system is organized into three deployment zones that reflect the physical and logical boundaries of the monitoring infrastructure. Zone A encompasses all in-water and bank-side hardware that interfaces directly with the water body. Zone B covers the station enclosure where pre-treatment, measurement, control, and communications equipment are housed. Zone C represents the cloud or central platform where data is ingested, validated, stored, analyzed, and reported. Understanding these boundaries is essential for assigning responsibilities, designing maintenance workflows, and managing cybersecurity perimeters.

Core modules — intake assembly, pre-treatment, water quality probes, edge controller/gateway, communications, and central platform ingestion/storage/QA/QC/alarm/reporting — are mandatory in every deployment. Optional modules such as auto-sampler, refrigerated storage, advanced analyzers (ammonia, nitrate, COD/TOC proxy), hydrology instruments (ADCP, flow radar), CCTV, and satellite communications are selected based on site requirements and budget. Supporting systems including UPS/solar, lightning protection, physical security, and fire safety (where a shelter exists) are treated as core engineering requirements rather than optional add-ons.

System Architecture Deployment Boundary Diagram

Figure 1.1: System architecture — three-zone deployment boundary diagram showing in-water hardware (Zone A), station enclosure (Zone B), and cloud/center platform (Zone C)

1.2 Components and Responsibilities

Each component in the system has a defined primary responsibility, a set of inputs and outputs, key performance indicators (KPIs), and characteristic mismatch risks that arise when the component is incorrectly specified or installed. The table below provides a structured reference for system designers and procurement engineers.

Component Inventory with Inputs, Outputs and KPIs

Figure 1.2: Component inventory diagram — field components, station control, and central platform with inputs, outputs, and key KPIs

Component Primary Responsibility Inputs Outputs Key KPIs (typical) Common Mismatch Risk
Intake head + strainer Deliver representative raw water River/lake water Raw water stream Anti-clog interval ≥ 7–30 days Wrong depth causes bias (surface algae / bottom sediment)
Pre-treatment module Debubble, reduce sediment load, stabilize flow Raw water Conditioned flow to sensors Flow stability ±10%; backflush success ≥ 95% Under-sized filter clogs daily in high turbidity
Flow cell manifold Provide stable measurement chamber Conditioned flow Flow-through for probes Leak rate 0; bubble rate low Poor venting causes DO/pH spikes
Multi-parameter probe Measure pH/DO/EC/Temp/ORP Water in flow cell Digital readings Drift within spec; uptime ≥ 95% Wrong materials corrode in brackish water
Turbidity sensor Measure turbidity / suspended solids proxy Water NTU readings Optical window fouling rate No wiper leads to persistent high bias
Algae sensor (optional) Chlorophyll-a / phycocyanin measurement Water µg/L or RFU Bloom detection sensitivity Sunlight intrusion causes measurement artifacts
Water level sensor Stage monitoring Water level Meters (m) Accuracy ±1 cm (site dependent) Wrong venting for vented pressure sensor
Rain/wind/air temp Meteorological context for events Atmosphere mm, m/s, °C Uptime ≥ 98% Poor siting (wind shadow, solar radiation bias)
Auto-sampler (optional) Grab samples for lab confirmation Trigger / schedule Bottled samples + log Sample integrity; temp 0–4 °C if required No refrigeration breaks chain-of-custody validity
RTU/PLC + IO Station logic, interlocks, control Sensor signals Commands + logs Deterministic control; watchdog Insufficient IO isolation introduces noise
Edge gateway Protocol conversion, local buffering RTU data MQTT/HTTPS payload Buffer ≥ 7–30 days Storage too small loses outage data
Industrial router Secure uplink to central platform SIM/Ethernet VPN tunnel MTBF, reconnect time NAT/firewall misconfig blocks inbound management
Power system Stable DC power to all station components AC/solar DC bus Autonomy hours Battery undersized causes brownouts
Central ingestion Validate and store incoming data Encrypted payload Time-series records Latency, loss rate Schema mismatch silently drops fields
QA/QC engine Apply validity flags to data Data + metadata Flags + annotations False reject rate Rules ignore seasonal effects
Alarm engine Notify and escalate threshold events Thresholds/events SMS/email/webhook Time-to-notify Too many alarms causes operator fatigue
O&M system Maintenance traceability and scheduling Tickets, logs Audit trail Closure SLA Lack of discipline breaks traceability

1.3 Working Principles

Startup Sequence

On power-on, the station controller performs a self-test of all IO channels and verifies valve positions. The intake pump starts and the pre-treatment line is flushed for a configurable period (typically 3–10 minutes) to clear stagnant water. The flow cell stabilizes, sensors warm up, and initial readings are flagged "stabilizing." The first valid sample is recorded only after stability criteria are met — specifically, when the rate-of-change of key parameters falls below the configured limit and flow rate is within the target band. This prevents erroneous early readings from entering the data record.

Normal Operating Cycle

During normal operation, the station executes a repeating cycle that includes: (1) continuous measurement at the configured sampling interval, (2) periodic automatic cleaning via wiper actuation, air purge, or chemical injection on a schedule adapted to fouling conditions, (3) scheduled calibration checks against reference solutions, (4) edge buffering and encrypted uplink to the central platform, and (5) platform-side QA/QC tagging and alarm evaluation. The edge gateway maintains a local time-series buffer sufficient to survive the worst expected network outage without data loss.

Exception Chains

The system is designed to handle predictable failure modes through defined exception chains. Three representative chains are described below, illustrating how the system detects, responds to, and recovers from common failure scenarios.

Exception Chain A — High Turbidity / Filter Clog: A post-storm turbidity surge exceeds the pre-treatment filter capacity. Flow rate drops below the minimum threshold. The RTU detects low flow and triggers an automatic backflush sequence. If flow is not restored within the configured timeout, the station enters "low-flow" mode: measurements continue but are flagged "suspect — low flow." An O&M ticket is created automatically. The platform suppresses quality alarms during this period to avoid false positives.
Exception Chain B — Algae Bloom / Optical Fouling: An algal bloom causes progressive fouling of the turbidity and chlorophyll-a optical windows. Wiper frequency increases automatically. If drift persists despite cleaning and the trend diverges from expected hydrology, the QA/QC engine flags readings as "fouling suspected." The system generates a maintenance request and cross-checks against any available grab sample data. Trend analysis is suspended for the affected parameters until a site visit confirms cleaning.
Exception Chain C — 4G Outage / Timestamp Drift: A cellular network outage causes the station to buffer data locally. During the outage, the edge gateway relies on its internal GNSS/NTP time source. On reconnection, the platform checks timestamp monotonicity and applies correction rules. Data outside the acceptable time uncertainty tolerance receives a "time uncertainty" flag. The backfill process is logged with the reconnection event, providing a complete audit trail of the outage period.