IoT Smart Hive Monitoring: Sensor Fusion & Anomaly Detection (2D)

Environment

Ambient temperature 22°C

Nectar flow intensity 55%

Sim time speed ×1

Disturbance

Live Sensors

Hive weight

25.00 kg

Weight Δ/hr

0.00 kg/h

Brood-nest temp

34.5°C

Fanning effort

Internal humidity

55%

Sim clock

06:00

Colony strength (post-event)

100%

Anomaly Log

System nominal — monitoring…

How it works

Three coupled sensor models run on a compressed clock (each real second ≈ several simulated minutes):

Weight — a scale under the hive integrates nectar intake: gain rate = flow × forager-activity(hour) × colony strength, plus small periodic jitter from bees walking on/off the platform. Foraging activity follows a bell curve centred on solar noon and is zero at night.
Thermoregulation — a real homeostat: dT/dt = k_env·(T_ambient − T) + activeRate, where activeRate is the bees' corrective power toward the 34.5°C brood-nest setpoint, clamped to a maximum the cluster can produce. When ambient deviation is small the correction easily wins and T tracks the setpoint; past the clamp (fanning effort saturates at 100%) the passive term starts to dominate and internal temperature drifts — exactly the failure mode real apiarists watch for.
Humidity — relaxes toward an ambient-linked baseline but rises during active nectar curing (bees evaporate water out of fresh nectar).
Anomaly detection — rule-based, same logic as a commercial hive-scale alert: a fast negative weight-rate trips a swarm/robbing flag; a temperature deviation while fanning effort is pinned at 100% trips a thermal-stress flag.

dT/dt = k_env(T_amb − T) + clamp(k_bee·(T_set − T), −P_max, P_max)·strength
weight' = flow · activity(hour) · strength · gainMax

Hive weight (kg) Brood temp (°C) Humidity (%RH)