Why it matters
Without AI, cleaning tasks rely on fixed schedules or manual observations, risking missed debris, inconsistent window visibility, uneven substrate cleanliness, and unbalanced diver workloads that can affect safety and exhibit quality.
From design to safety
Design
AI informs in-tank task workflow
Define data flows from substrate sensors, window cleanliness sensors, and diver time logs; establish safety thresholds and integration with dive planning software and display dashboards; set standardized reporting formats for handoffs; incorporate data governance, access controls, and audit trails for chain-of-custody of sensor readings.
- substrate siphoning system
- window cleaning equipment
- rockwork cleaning tools
Build
Systems integration and commissioning
Install sensors on siphons, window cleaning tools, and rockwork brushes; configure task queues, calibrate dive timers, and validate data streams to the central control system; perform tolerance tests; verify compatibility with existing SCADA and sound level controls.
- sensor nodes on siphon
- window scraper sensor
- rockwork brush mounting
Operation
Real-time task negotiation
AI interprets sensor readings, predicts debris accumulation, and assigns cleaning segments and dive time windows; operators confirm or adjust via the task log, with anomaly alerts; integrate with live-water quality alarms and diver sign-off.
- dive time recorder
- task log tablet
- communication headset
Maintenance
Analytics-driven upkeep
Track tool wear, sensor drift, and battery health; schedule calibration and replacement, and adjust AI thresholds based on trend data; generate maintenance work orders; link to parts inventory and vendor service windows.
- substrate siphon motor
- window cleaning motor
- rockwork cleaner motor
Safety
Safety-first scheduling
AI flags unsafe diving conditions, potential entanglements, and water quality risks tied to cleaning events; requires supervisor approval before execution; maintains incident logging; supports regulatory reporting and drill readiness.
- dive time monitor
- emergency stop relay
- breathing air pressure gauge
Part by part
- 1
substrate siphon nozzle
- Watch for
- clogging, uneven substrate suction, abrasion wear on nozzle tip
- AI action
- read siphon flow rate and turbidity, trigger nozzle maintenance or cleaning queue
- 2
window cleaning blade
- Watch for
- blade wear, streaking, loss of suction in automated sweep
- AI action
- monitor blade health and surface cleanliness, schedule blade replacement or recalibration
- 3
rockwork cleaning brush head
- Watch for
- abrasion of rockwork, brush clogging, reduced debris removal
- AI action
- assess debris load, advise brush cycle length changes or tool swap
- 4
dive timer console
- Watch for
- incomplete dive logs, overruns, unsafe dive durations
- AI action
- authorize or shorten dive windows, log as safety event if thresholds breached
- 5
task log tablet
- Watch for
- missing entries, inconsistent task completion records
- AI action
- prompt operator to confirm tasks, auto-summarize daily activity
- 6
dive safety radio
- Watch for
- communication gaps, lost signal
- AI action
- alert supervisor, switch to manual coordination if needed
Worked example
Preventing window haze during a multi-tank cleaning day
Illustrative scenario with sample data.
A routine cleaning day requires synchronized substrate siphoning and window cleaning across three exhibits while divers manage rockwork debris without impacting show schedules. The plan includes contingency for sudden visibility drops and a late diver
Signal in
waterTemp=24.6C, substrateFlow=38L/min, windowSmearIndex=62, rockworkLoad=14%, debrisLoad exhibitA=7, exhibitB=5, scheduleSlots=[08:20-08:40,09:00-09:20], lastService=2026-10-15, sensorStatus={siphon:OK, windowBlade:OK, rockbrush:OK}
System out
taskLogGenerated: summary of cleaning blocks, assigned tasks, divers allocated, and automated alerts; maintenance queue created for siphon nozzle and window blade replacements; next cycle adjusted for debris load; compliance check passed
Human sign-off: Dive supervisor reviews the AI plan, validates diver assignments, approves the daily plan, and authorizes the maintenance work orders; facilities manager confirms inventory and spare parts readiness.
How to roll it out
Phase 1
Weeks 1 to 2
Baseline data collection from sensors, establish clear task queues, and define safety overrides.
Phase 2
Weeks 3 to 6
Pilot AI task scheduling in two exhibits, refine thresholds, and implement dive time integration with the dive log.
Phase 3
Weeks 7 to 12
Full deployment across all exhibits, optimize task routing, and formalize reporting to aquarists and facilities team.
Where people stay in charge
- All scheduling decisions that impact diver safety must be approved by the Dive Safety Officer.
- Critical water quality or safety thresholds require sign-off by the Life Support Supervisor and Aquatic Veterinarian where applicable.
- Human-on-loop prerequisite ensures final confirmation of any automated maintenance work before execution.
Questions owners and crews ask
What data streams drive AI in-tank cleaning decisions?
The AI consumes substrate siphoning flow, turbidity, window cleanliness metrics, rockwork debris load, dive time, and task log entries to plan and adjust cleaning actions. It also references sensor health indicators and recent maintenance notes to avoid cascading failures.
Who must approve changes to dive schedules created by AI?
The Dive Safety Officer must approve any AI-generated dive window changes that affect diver safety or exhibit access; any conflict with show timing requires escalation to the exhibits manager.
How does AI handle tool wear and maintenance needs?
The system tracks wear indicators on siphon nozzles, window blades, and rockwork brushes, and schedules replacements before performance degrades; it logs part lifetimes and triggers reorder alerts when stock falls below thresholds.
What records are produced for accreditation and audits?
The AI generates task logs, dive time records, maintenance work orders, sensor calibration logs, safety incident notes, and drill reports for audits; it exports to standardized formats for regulator reviews and accreditation files.
Sources and further reading
- HSE: Diving at WorkDiving at Work Regulations and approved codes covering unsafe diving conditions and dive team safety.
- OSHA: Commercial DivingUS commercial diving standards and diver safety rules relevant to dive approval decisions.
- Aquatic Animal Life Support Operators (AALSO)Life support operator guidance on water quality alarms, operation, training, and certification for operators.
- EAZA: AccreditationAccreditation standards and screening that define the records zoo and aquarium audits expect.
- GOV.UK: Keeping Zoo AnimalsZoo licence requirements, inspections, and obligations that shape reporting for aquariums open to the public.
- MBARI: TechnologiesOcean sensors and underwater vehicles showing how sensors are mounted and validated in marine science.
- International Association for Aquatic Animal MedicineAquatic veterinary medicine network and conferences covering aquarium vet sign-off and clinical roles.
- GOV.UK: Standards of Modern Zoo PracticeStatutory zoo licensing standards setting expectations for calibration records and safety documentation.