This article explains what a CEMS measures, how it is built, which checks keep its data valid and how its upkeep should be organised. It is aimed at owners, environmental managers and operations leads at campuses that run turbines for prime or bridge power, where emissions data is reviewed by regulators and, increasingly, by the campus’s own sustainability team.
Why a turbine plant needs continuous monitoring
In the US, stationary combustion turbines fall under federal performance standards such as 40 CFR Part 60 Subpart KKKK, and larger units that sell power may also fall under the Part 75 monitoring rules. State permits frequently add their own limits and averaging periods. Where the permit sets an hourly or short rolling average limit, periodic stack tests are not enough, and a CEMS is the practical way to show compliance.
Outside the US the legal basis differs, but the logic is the same: an emission limit is only credible if it is measured. Planning the CEMS early, while the air permit application for the campus is being prepared, avoids promising averaging periods or detection levels that the monitoring equipment cannot support.
What the system measures
A typical gas turbine CEMS measures NOx, a diluent gas (oxygen or carbon dioxide) and often CO. Where an SCR system is fitted, ammonia slip may also be monitored. Stack flow may be measured directly or calculated from fuel flow, depending on the rule that applies.
Turbine limits are usually expressed in parts per million corrected to 15 percent oxygen on a dry basis, so the diluent measurement matters as much as the pollutant reading. The correction is simple arithmetic: the measured concentration is multiplied by (20.9 minus 15) divided by (20.9 minus the measured oxygen percentage). A drifting oxygen analyser therefore distorts every corrected NOx value, even if the NOx analyser is perfect.
How the sample reaches the analysers
Most turbine installations use extractive sampling. A probe in the stack draws gas through a heated sample line to a conditioning system that removes moisture, then to analysers housed in a climate-controlled shelter. Dilution extractive systems mix the sample with clean air at the probe, which reduces moisture problems and line size. In-situ analysers measure inside the stack itself and avoid long sample lines, but are harder to reach for calibration.
Probe location follows reference methods that require straight duct lengths upstream and downstream of disturbances, so the gas is well mixed where it is sampled. On compact modular plants with short stacks this needs attention at layout stage. Stratification tests during commissioning confirm whether a single sample point is representative.
Shelter design is part of the same layout exercise. Analysers need stable temperature, clean power, certified calibration gas cylinders stored safely nearby and enough space for technicians to work. Placing the shelter close to the stack shortens heated sample lines, which reduces response time and the number of places where condensation or leaks can corrupt the sample.
The data acquisition and handling system
Analyser readings flow into a data acquisition and handling system, usually called the DAHS. It applies the corrections, calculates the averages required by the permit, flags exceedances, records calibration results and produces the reports submitted to the regulator. It also applies missing data procedures when an analyser is out of service, which can substitute conservative values that make the plant look worse than it is.
Because of that, availability of the CEMS itself is a performance indicator. The DAHS should feed alarms to the plant control room so operators see a rising NOx trend or a failed calibration in real time, alongside the other operator alarm screens for the plant. Its network connections should also be covered by the site’s OT security arrangements.
Quality assurance: daily, quarterly and annual checks
- Daily calibration drift checks, where zero and span gases are injected and the analyser response is compared with the known value.
- Quarterly cylinder gas audits or linearity checks at several concentrations, using certified gases.
- An annual relative accuracy test audit (RATA), where an independent test team measures the stack with reference methods and the CEMS readings are compared against them.
- Routine maintenance of probes, filters, pumps, chillers and heated lines, logged in a written quality assurance plan.
Failing a check does not just mean recalibrating. It can invalidate data back to the last good check, triggering missing data substitution for those hours. A good QA plan keeps spare parts, certified gas cylinders and trained technicians on hand so a failed check is corrected within hours, not days.
Who runs the CEMS
Some owners keep CEMS upkeep in house, others place it with a specialist under a service contract, and many include it in the plant’s operations scope. Whichever route is chosen, responsibilities for daily checks, audits, RATA scheduling, report submission and record retention should be written down, with named people and backups.
When the CEMS sits inside a wider operations and maintenance contract, the contract should state the target data availability and how missing data periods are handled. Linking that target to the turbine maintenance plan matters too: combustor tuning and catalyst condition both show up first in the emissions data, which makes the CEMS an early warning tool as well as a compliance record.
Data center operators often want emissions figures for their own reporting as well as for the regulator. The DAHS can export validated hourly data for that purpose, but the permit record should remain the single source, so sustainability reports and compliance reports never quote different numbers for the same hour.
About the authors
This guide comes from the USP&E environmental compliance group that keeps stack monitoring valid on running plants. USP&E is an EPC and O&M contractor, not a broker, with 150+ power projects in 45+ countries, 25,000 MW contracted since 2002 and 24+ years in the field. Our overview of turbine power plants for data centers explains how emissions control fits the wider plant design.
Offices: Colorado Springs (1755 Telstar Dr, Suite 300), Knoxville (800 S Gay St, Suite 700), London (162-168 Regent Street), Johannesburg (22 Kildoon Rd, Bryanston), Cape Town (31 Brickfield Road, Woodstock) and China (No. 1 Ningzhou Road, Wuxi). Regions: USA | AFRICA | UK | UAE | CHINA. ISO 9001 and ISO 45001 certified.
CEMS questions for data center turbine plants
What does CEMS stand for?
Continuous emissions monitoring system: analysers, sampling equipment and software that measure stack emissions continuously and record them for compliance.
Which gases does a gas turbine CEMS measure?
Usually NOx and oxygen or carbon dioxide, often CO, and sometimes ammonia slip where an SCR system is installed.
Why are readings corrected to 15 percent oxygen?
Turbine exhaust contains a lot of excess air. Correcting to a reference oxygen level stops dilution from making emissions look lower than they are.
What is a RATA?
A relative accuracy test audit, usually annual, in which an independent team measures the stack with reference methods and the CEMS results are compared with theirs.
What happens when the CEMS is out of service?
The data system applies missing data procedures, which can substitute conservative values, so quick repair and good spares matter.
Can CEMS data help maintenance?
Yes. Rising NOx or CO often signals combustor tuning drift or catalyst wear before other symptoms appear.
Plan monitoring with the plant
Share your permit conditions, stack layout and operating profile, and our team will review the CEMS scope, sampling location and quality assurance plan for your campus turbines.
If a document set is easier than a conversation, send it to info@uspeglobal.com or upload a summary through the plant operations enquiry; to talk it over, dial +27 10 822 2324. Turbine packages available for data center plants are listed in our natural gas turbine listings.