Inside the high‑stakes war on Legionella in power‑plant cooling towers

Cooling towers run on physics that also favor bacteria: warm (20–50 °C) recirculating water, misting aerosols, biofilms, and organic nutrients create ideal conditions for Legionella growth (beta.co.id) (pmc.ncbi.nlm.nih.gov). The stakes are high: a 2015 Bronx outbreak linked to a cooling tower caused 138 cases and 16 deaths (pmc.ncbi.nlm.nih.gov), while U.S. incidence has risen ~5.6×, from 0.48 to 2.71 cases per 100,000 between 2002–2018 (pmc.ncbi.nlm.nih.gov), with ~95% hospitalization and ~10% mortality (pmc.ncbi.nlm.nih.gov).

Power plants in tropical Indonesia face especially high ambient temperatures, which exacerbate the risk (beta.co.id). In response, operators are converging on a three‑pillar plan: a robust biocide program, regular cleaning and disinfection, and routine Legionella testing — each tied to quantified outcomes, from disinfectant residuals to log‑scale bacteria reductions.

Risk profile and operating context

The mechanics are straightforward: evaporative towers generate [aerosols that can carry Legionella](https://beta.co.id/en/blog/hospitals-hidden-aerosol-risk-a-practical-plan-to-keep-legionella-out-of-cooling-towers), and biofilm/scale layers shelter microbes from chemicals (pmc.ncbi.nlm.nih.gov). For combined‑cycle plants, that means controlling water chemistry and fouling is non‑negotiable. Ancillary chemistry to limit deposition supports this goal; for example, plants commonly deploy dispersant chemicals to keep particles from agglomerating, improving system efficiency and reducing fouling.

Biocide program: shock and residual control

A continuous, automated regimen pairing oxidizing biocides (e.g., sodium hypochlorite or chlorine dioxide) and non‑oxidizing agents (e.g., glutaraldehyde or isothiazolines) is standard because oxidizers act quickly while non‑oxidizers extend control; “dual‑biocide” programs are common (chemaqua.com) (chemaqua.com). Italian guidance details a 2‑hour chlorine shock at ~50 mg/L (mdpi.com) followed by continuous chlorine at ~2–3 mg/L (mdpi.com) (mdpi.com); in one study, that shock‑and‑maintain scheme drove L. pneumophila from ~10^5 CFU/L to ~10^1.8 CFU/L (>99.9% reduction) (mdpi.com). Not all shocks perform: an H₂O₂‑silver treatment failed and counts rose, underscoring the need to validate protocols in situ (mdpi.com).

Residual control is the day‑to‑day anchor. A common target is ≈0.5 mg/L free chlorine (adjusted for pH and temperature). Risk rises sharply when residuals fall: residual <0.2 mg/L raised the odds of L. pneumophila by ~8.5× in one analysis (researchgate.net). Automated feed with feedback control is strongly advised (cdc.gov), and towers maintaining >0.4 ppm free chlorine showed Legionella detection <10% of the time versus >80% when below 0.2 ppm (researchgate.net). For bromine/biguanide, similar ppm guidelines apply. Plants increasingly deploy dosing pumps to maintain precise chemical feed.

Monitoring must be daily when the tower runs: log dose rates and residuals, track pH 7–9, conductivity, hardness, and Heterotrophic Plate Count (HPC — a general microbial indicator) with verification after shocks or process upsets; NYC law requires ≥1 dosing per day and weekly measurement is a minimum (nyc.gov) (nyc.gov). As Chem-Aqua puts it, no biocide program works in a dirty system (chemaqua.com). When residuals remain high on low feed, operators also watch for corrosion risk or dosing errors; corrosion inhibitors are often part of the same treatment envelope. Many plants procure an integrated cooling‑tower chemical program to consolidate oxidizers, non‑oxidizers, and monitoring.

Cleaning and disinfection scheduling

Chemistry alone cannot overcome fouling. Full mechanical cleaning of the basin, eliminators, fill, and nozzles at least twice per year is widely cited: Indonesian health guidance advises cleaning 2×/year with periodic chlorination (infeksiemerging.kemkes.go.id); Victorian regulations mandate cleaning and disinfection of all wetted surfaces every six months (health.vic.gov.au), and NYC codes recognize biannual cleaning as a baseline (nyc.gov) (health.vic.gov.au). For large or heavily loaded power‑plant towers, quarterly cleaning is often considered.

End‑to‑end cleaning removes biofilm, scale, sludge, and debris with mechanical brushing and high‑pressure rinsing, followed immediately by disinfection. CDC classifies scale/sediment removal as “critical” maintenance (cdc.gov). After cleaning, a start‑up shock (e.g., ~50 mg/L NaClO for 2 hours) is applied before returning to service (mdpi.com) (nyc.gov). Plants document what was cleaned, methods used, and microbial checks (residual, HPC) before restart. Professional outsourcing is common at scale; many operators book a cooling‑tower cleaning service to meet schedule and access constraints.

Cleanings are also event‑driven: after high Legionella results, after downtime >2 weeks to avoid stagnation, or following major repairs; national and local guidance recommends disinfection after each routine clean or extended shutdown (mdpi.com). Routine semi‑annual cleaning has been correlated with halving long‑term Legionella‑positive prevalence in district data (mdpi.com). Victorian guidance notes that cleaning more often “helps to control nutrient growth” (health.vic.gov.au), whereas dirty systems see rapid microbial rebound regardless of chemical dose (chemaqua.com). Plants often support cleaning with scale inhibitors to reduce deposits that shelter biofilms.

Routine testing and measurable thresholds

Periodic Legionella cultures (ISO methods, reported as CFU/mL — colony‑forming units per milliliter) are central. New York State requires owners to test towers to assess maintenance effectiveness (health.ny.gov). While no Indonesian law sets a frequency, best practice is at least quarterly tower sampling for Legionella spp., and weekly sampling during an outbreak investigation (pmc.ncbi.nlm.nih.gov). Action thresholds are explicit in U.S. codes: >1,000 CFU/mL triggers corrective measures and notifications (health.ny.gov). Over time, many programs aim for <10 CFU/mL (essentially undetectable).

Operators complement cultures with routine microbial indicators. Weekly HPC (heterotrophic plate count) on basin water offers early biofilm warning; many aim to keep HPC <500 CFU/mL (some target <100) as a rough benchmark. If HPC spikes, it is a call‑to‑action to adjust dosing or inspect equipment. Plants that choose to augment disinfection with hardware sometimes add ultraviolet systems to reduce pathogens without chemicals.

Disinfectant residual and core chemistry are logged at least once daily in operation — NYC mandates ≥1/day and water quality checks 3×/week (nyc.gov) (nyc.gov). Logs typically include free chlorine (or equivalent oxidizer), pH, temperature, and conductivity, which also support cycles‑of‑concentration and scale/corrosion risk assessments. Automated sensors and trend software help trigger alarms if residuals drift low; studies show risk grows rapidly when residual <0.2 ppm, with ~90% chance of contamination in that zone (researchgate.net). Monitoring is treated as crucial process control, not optional.

Training, oversight, and plan governance

Only trained personnel — “competent persons” — should clean or sample towers, with appropriate PPE for aerosol exposure (health.vic.gov.au). Plans are formalized under ASHRAE 188 or local guidance, with checklists for weekly and monthly tasks, quarterly data reviews (HPC trends, residual logs, Legionella results), and change control when the data indicate — for example, increasing shock frequency or upgrading equipment.

Recordkeeping and compliance mirror leading jurisdictions. New York requires owners to register towers, file testing results, and report Legionella exceedances (health.ny.gov). In the absence of local regulation, operators adopt similarly rigorous standards — immediate notification on exceedance, corrective cleaning/biocide application, and weekly re‑testing until safe. Many standardize procurement around proven biocides to ensure consistent kill performance and documentation.

Clear KPIs tighten execution: maintain Legionella <100 CFU/mL in >95% of samples, HPC consistently <200 CFU/mL, pH in range, and zero system failures. Plants also track the “risk level” (non‑conformances per year) with a goal of near zero; if frequent positives occur, the program is elevated — stronger chemicals, more frequent cleaning, or hardware upgrades like UV or secondary filters (cdc.gov). Where solids control is a bottleneck, some operators add side‑measures alongside chemistry using cooling‑tower chemical regimens to stabilize performance between cleans.

Business outcomes and measurable impact

The ROI is unambiguous. A 2005 UK study found outbreak costs were dominated by medical care (£392k) versus £64k for public‑health response (pmc.ncbi.nlm.nih.gov). By contrast, routine treatment chemicals and scheduled maintenance are a small fraction of operating expense. International experience shows that after stringent controls — daily dosing, quarterly testing, semi‑annual cleaning — Legionella detections drop sharply (often to zero), and no community cases recur (mdpi.com) (pmc.ncbi.nlm.nih.gov). Complying with, or exceeding, CDC, ASHRAE 188, and NY codes helps manage public‑health risk and legal liability (cdc.gov) (health.ny.gov).

Actionable summary and targets

The water‑management plan rests on three measurable pillars. First, a managed biocide regimen — shock plus continuous dosing with a maintained residual (commonly ≈0.5 mg/L free chlorine) — validated by culture log‑reductions and supported by automation; deploying dosing pumps to sustain residuals closes the loop (mdpi.com) (cdc.gov). Second, a strict cleaning schedule — at least 2×/year and event‑driven — with immediate post‑clean shock and restart only after residual/HPC checks (infeksiemerging.kemkes.go.id) (nyc.gov). Third, routine testing — weekly chemistry/HPC and at least quarterly Legionella cultures — with thresholds and corrective actions (>1,000 CFU/mL action; aim for <10 CFU/mL) (health.ny.gov). The objective metrics are clear: residual ppm, CFU counts, and documented log‑reductions — and none of it works in a dirty system (chemaqua.com).