Singapore Cooling Tower Compliance: Using Smart Sensors and AI-Assisted Workflows to Manage Legionella Risk

A practical framework for monitoring cooling-tower conditions, coordinating maintenance and maintaining clearer compliance evidence.

Cooling towers support air-conditioning, process cooling and industrial operations across Singapore. They also generate aerosols, which means poor water management or inadequate maintenance can create a workplace and public-health concern. For building owners, facility managers, warehouse operators and SMEs, the challenge is not simply installing more sensors. It is creating a dependable operating process that connects measurements, maintenance actions, laboratory results, contractor responses and compliance records.

Smart sensors and AI-assisted workflows can improve visibility and response time. However, they do not replace the owner’s statutory duties, competent technical decisions or mandatory laboratory testing. The most effective approach is to use digital tools as an operational layer around an established water-management and maintenance programme.

Understand the compliance responsibility first

Singapore’s National Environment Agency identifies cooling towers as aerosol-generating systems. Owners are responsible for managing the system appropriately, including registration or other applicable administrative requirements, water-quality control, testing, reporting and temporary-shutdown procedures. Operators should refer directly to the latest NEA requirements and guidance for the specific system and situation.

A particularly important planning point is a temporary shutdown lasting more than five consecutive days. Based on NEA guidance, such a shutdown requires an application and the cooling tower must be cleaned and tested. The system must be within the applicable regulatory limits before it resumes operation. This creates a clear need for a controlled shutdown, cleaning, testing and restart workflow rather than an informal decision to switch equipment back on.

Digital monitoring can support this process, but it is not itself proof of compliance. A temperature trend, conductivity reading or AI-generated alert cannot substitute for required laboratory analysis, formal reporting or professional judgement.

What should a practical sensor strategy monitor?

The right architecture depends on the cooling-tower design, operating conditions and existing building-management systems. A useful starting point is to map each measurement to a decision or action.

• Temperature: Supply and return-water temperatures can help teams identify abnormal thermal performance, unexpected load changes or conditions requiring investigation.
• Water quality: Conductivity, pH, oxidation-reduction potential, turbidity or other suitable measurements may provide early indication of changing water conditions. Sensor suitability, calibration and installation location matter.
• Make-up and blowdown: Flow data can help identify unusual consumption, dosing imbalance, blocked components or possible leaks.
• Chemical dosing: Monitoring pump status, tank levels, dosing events and fault signals can make treatment activity easier to verify and escalate.
• Vibration and equipment condition: Fan, pump and motor vibration data can support early investigation of imbalance, bearing wear or other mechanical issues that may affect reliable operation.
• Maintenance status: Work orders, cleaning dates, calibration records, contractor attendance, test results and photographs should be connected to the relevant asset and date.

Not every site needs every sensor. Start with the risks that are difficult to observe manually, the data already available from the control system and the actions that operators can realistically take.

Use AI to prioritise work, not to make unsupported compliance decisions

AI-assisted monitoring is most useful when it helps people interpret multiple signals and follow a consistent process. For example, a workflow can compare current readings with site-configured operating ranges, recent trends and equipment status. It can then classify an issue as an observation, an investigation or an urgent escalation.

Useful AI-assisted functions may include:

• summarising daily cooling-tower conditions for a facility supervisor;
• identifying combinations such as rising conductivity, abnormal blowdown and a dosing fault;
• flagging missed inspections, overdue calibration or incomplete maintenance records;
• routing alerts to the correct internal owner or service contractor;
• drafting an incident summary from time-stamped readings, work orders and photographs; and
• checking whether required shutdown or restart steps have been completed before a human authorises the next stage.

The system should not present a prediction as a confirmed Legionella result. It should not automatically declare a tower safe to operate, replace a required laboratory test or bypass an approval step. AI outputs should be explainable, time-stamped and reviewed by an appropriately responsible person.

Build a digital evidence trail

Many operational problems occur because information is split between spreadsheets, email, paper forms, contractor messages and different building systems. A central digital record can make oversight more reliable.

For each cooling tower, consider maintaining an asset profile containing the equipment identifier, location, responsible personnel, sensor list, treatment arrangement, inspection frequency and relevant documents. Each event should record who performed the task, when it was performed, what was observed, what readings were captured and what action followed.

Evidence may include laboratory reports, calibration certificates, chemical-delivery or dosing records, cleaning reports, photographs, alarm acknowledgements, contractor attendance and approval records. Access controls and change history are important so that the record remains credible and useful during internal reviews, contractor meetings or regulatory engagement.

Plan shutdown and restart as a controlled workflow

Temporary shutdowns deserve special attention because they often involve several parties and a narrow restart window. A practical workflow can include the following stages:

1. Plan: Confirm the reason, expected duration, affected operations, responsible persons and applicable NEA requirements.
2. Isolate and record: Capture the final operating condition, isolate equipment safely and record the shutdown time.
3. Clean and inspect: Complete the required cleaning and inspection activities, documenting findings and corrective work.
4. Test: Arrange the required water-quality testing through the appropriate process. Do not rely only on live sensor readings.
5. Review: Check laboratory results, maintenance records and outstanding defects against the restart checklist.
6. Authorise and monitor: Obtain the required human approval, restart under controlled conditions and monitor the system closely.

Automated reminders can reduce missed deadlines. A workflow can also prevent a restart task from being marked complete when a required document or result is missing. The exact checklist should be developed with the responsible engineering, facility-management and compliance teams.

Connect compliance with reliability and energy management

Cooling-tower hygiene should remain the primary objective, but the same data can support wider operational improvements. Abnormal temperatures, excessive make-up water, irregular fan operation or repeated pump alarms may indicate opportunities to improve reliability and reduce avoidable operating waste.

Singapore’s broader smart-facilities direction also supports more connected monitoring, predictive analytics and technology-enabled workplace-safety practices. These benefits should be treated as supporting outcomes. Energy optimisation must not lead to reduced water-treatment control, deferred cleaning or the removal of necessary human checks.

A sensible implementation path for Singapore sites

Start with a site survey and compliance-process review. Identify the cooling towers, existing controllers, water-treatment equipment, test points, maintenance contracts and current records. Next, define alert ownership and escalation rules before installing technology. An alert without a named responder and response time is unlikely to improve risk control.

Use a pilot on one system where the data and maintenance process can be reviewed closely. Validate sensor quality, network reliability, calibration arrangements and the usefulness of the workflow. Then refine the operating ranges, dashboard views and approval steps before expanding to other towers or facilities.

ISS can support businesses reviewing engineering operations, facility-management processes and AI automation opportunities. Contact ISS to discuss a practical approach to cooling-tower monitoring, digital maintenance workflows and connected facility operations. Any implementation should be aligned with the latest NEA requirements and the site’s responsible technical and compliance personnel.