How to Validate CEMS Data for Compliance

A CEMS can produce a complete-looking data set while still generating results that are not valid for compliance use. A flatlined signal, an expired calibration gas certificate, a failed daily zero check, or an unrecognized data acquisition issue can turn reported emissions data into a regulatory exposure. Knowing how to validate CEMS data means evaluating the measurement system, not simply accepting the numbers exported from the data acquisition and handling system.

For facility owners, environmental managers, and EHS teams, data validation should be a controlled process with clear acceptance criteria, documented decisions, and defined responsibilities. The objective is to ensure that each reported value is representative, traceable, and acceptable under the permit, applicable rule, or monitoring plan.

Start With the Applicable Monitoring Requirements

Validation criteria are not identical for every CEMS installation. The correct process depends on the pollutants monitored, the source category, the reporting program, the facility permit, and the regulation that governs the unit. A system reporting under 40 CFR Part 60 may have different data availability, quality assurance, and substitution requirements than a unit governed by 40 CFR Part 75 or a site-specific permit condition.

Before reviewing a reporting period, establish the requirements that apply to that monitor. Identify the required calibration frequency, allowable calibration error, drift limits, relative accuracy test audit requirements, missing-data rules, averaging periods, and data availability thresholds. Also confirm the reporting basis: dry or wet, actual or standard conditions, oxygen-corrected or uncorrected, and concentration or mass emissions.

A common validation error is to confirm that an analyzer operated properly while overlooking whether the reported calculation is configured properly. A valid ppm value can become an invalid compliance result if moisture, flow, oxygen, stack temperature, or standard-condition corrections are incorrect.

Confirm the CEMS Was Operating as Intended

The first technical review should focus on the operating status of the complete monitoring system. This includes the analyzer, sample probe, heated sample line, conditioning components, calibration gas delivery system, flow measurement devices, opacity monitor where applicable, and the data acquisition system.

Review maintenance logs alongside operating data. A filter change, probe cleaning, chiller fault, sample pump failure, analyzer service event, or power interruption does not automatically invalidate all data. It does, however, define a period that requires closer review. Determine when the issue began, when the system returned to stable operation, and whether required post-maintenance checks were completed.

Pay particular attention to sample system problems. A heated line below its required temperature, water breakthrough in a conditioning system, plugged probe filter, or uncontrolled sample flow can bias results without producing an obvious analyzer alarm. Compare sample flow, pressure, temperature, and moisture-related indicators against normal operating ranges. Stable emissions values are not necessarily credible if supporting parameters show abnormal behavior.

Review Daily Calibration Checks

Zero and upscale calibration checks are among the most direct indicators of analyzer performance. Compare each result with the applicable acceptance limit and review the trend over time rather than treating each check as an isolated event.

A calibration check that passes but steadily worsens may indicate contamination, component deterioration, leaking fittings, unstable calibration gas delivery, or analyzer drift. If the monitor exceeds its allowable calibration error or drift criterion, determine the exact duration of the failure. Data collected during that period may need to be flagged as invalid, depending on the governing requirements.

Calibration records must also be traceable. Confirm that the gas concentrations are appropriate for the analyzer range, certificates are current, cylinder identification is documented, and calibration gas was introduced through the required point in the system. Introducing gas directly at the analyzer may test the analyzer but not the probe, sample line, or conditioning train.

Check Quality Assurance Test Status

Daily checks do not replace periodic quality assurance testing. Review the status and results of required linearity checks, cylinder gas audits, relative accuracy test audits, relative accuracy test reports, and flow monitor audits. The applicable program determines which tests are required and how test failures affect data validity.

Where a relative accuracy evaluation applies, compare the CEMS output with reference method testing conducted under the appropriate EPA method and approved test plan. A passing relative accuracy result supports confidence in the monitoring system across its operating range. A failed test requires prompt investigation, corrective action, and a documented determination of how the failure affects reported data.

Validate the Data Acquisition and Calculation Logic

A CEMS data acquisition and handling system can introduce errors even when field instrumentation is functioning correctly. Validation must include the calculation path from raw analyzer signal to the final hourly, daily, or reporting-period value.

Verify analyzer range, engineering units, calibration coefficients, span values, and signal scaling. Confirm that the system is applying the correct moisture correction, oxygen correction, diluent calculation, molecular weight, F-factor, stack flow, and standard temperature and pressure conversions where required. Review whether the DAHS uses the correct emissions limits and averaging periods for alarms and excess emissions reporting.

Time synchronization also matters. Analyzer data, unit operating data, calibration events, and DAHS timestamps must use a consistent clock. Misaligned timestamps can associate a calibration failure with the wrong operating period or produce incorrect hourly averages.

For mass emissions systems, reconcile calculated values against independent operating information. A sudden change in reported pounds per hour or tons per day may be legitimate, but it should correspond to a change in load, fuel use, stack flow, concentration, or operating mode. If it does not, investigate the calculation before accepting the result.

Identify Invalid Data and Apply the Correct Code

Data validation is not a process of removing inconvenient values. It is a controlled determination of whether data meet established quality criteria. Each invalid data period should have a reason code supported by calibration records, maintenance documentation, alarms, audit results, or operator logs.

Typical causes of invalid data include failed calibration checks, monitor malfunction, out-of-control conditions, missing required quality assurance tests, loss of sample flow, communications failure, or DAHS calculation errors. The applicable regulation or permit should define whether the data must be excluded, substituted, estimated, reported as monitor downtime, or retained with a specific qualifier.

Do not automatically classify all maintenance periods as invalid. Likewise, do not assume that data are valid immediately after maintenance is complete. Review the required return-to-service checks and the first stable operating data after the repair. The decision must be based on documented criteria, not convenience or production pressure.

Use Trend Review to Find Problems That Alarms Miss

Automated alarms are useful, but they are not a complete validation program. Monthly and quarterly trend reviews often reveal gradual performance changes that individual alarms do not identify.

Review zero and span results, analyzer response time, sample flow, stack flow, oxygen, moisture, pollutant concentration, and data availability. Look for recurring failures after specific operating conditions, such as startup, low-load operation, fuel changes, high-moisture service, or seasonal ambient temperature shifts. Compare emissions patterns to production rates and fuel consumption when those variables should be related.

A useful review also compares redundant or related measurements. For example, a large change in corrected NOx emissions without a corresponding shift in raw NOx, oxygen, flow, or load deserves investigation. The issue may be real process behavior, but it may also point to a failed sensor or calculation input.

Document the Validation Decision

A defensible validation record allows a third party to understand what happened without relying on institutional memory. For each review period, retain the raw data, calibration results, maintenance records, alarm history, QA test documentation, validation flags, calculation configuration, and the final approved data set.

The record should identify who performed the review, when it was completed, what exceptions were found, how each exception was resolved, and whether corrective action is still open. Controlled access to DAHS configuration and an audit trail for manual changes are particularly important. If a value is edited, substituted, or reclassified, the original value and reason for the change should remain recoverable.

For facilities with multiple monitors or complex reporting obligations, a written CEMS data validation procedure helps create consistency across shifts and personnel. It should define review frequency, escalation thresholds, approval authority, retention requirements, and the interface between operations, maintenance, environmental staff, and technical support.

When Independent Technical Review Adds Value

An internal review may be sufficient for stable, well-understood systems with experienced staff. Independent technical review becomes more valuable after repeated calibration failures, an audit failure, a major analyzer replacement, unusual emissions patterns, a permit renewal, or before submitting high-consequence compliance reports.

Air Research Group supports industrial facilities with CEMS-related technical services, calibration support, emissions testing, and compliance-focused data interpretation. The most useful engagement is often not a one-time check of a spreadsheet, but a review that connects instrument performance, reference testing, operating conditions, and reporting requirements.

Reliable CEMS data are built through disciplined daily checks and protected through informed judgment when conditions change. Treat each validation decision as if it may need to be explained years later to a regulator, auditor, or internal reviewer. That standard encourages the documentation and technical care that make emissions reporting defensible.