A continuous emissions monitoring system (CEMS) is one of the most expensive and operationally demanding pieces of equipment in an environmental compliance program. A typical CEMS installation costs $100,000–300,000 for the hardware alone, plus ongoing costs for calibration gases, replacement parts, and dedicated technician time. And unlike most process equipment, the CEMS data goes directly to the regulator — every hour, every day, for as long as the permit requires.
I’ve been through CEMS installations at four plants. Here’s what the vendor proposals don’t tell you.
What a CEMS Actually Measures
A CEMS isn’t one instrument. It’s a system of analyzers, sample handling equipment, and data acquisition that continuously measures pollutant concentrations in a stack or duct. A typical CEMS package includes:
Gas analyzers. For criteria pollutants: NOx (chemiluminescence or UV), SO2 (UV fluorescence or NDIR), CO (NDIR or gas filter correlation), CO2 and O2 (for dilution correction). Each analyzer uses a different measurement principle and has different interference sensitivities.
Sample handling system. The sample must be extracted from the stack, filtered, conditioned (heated to prevent condensation), and delivered to the analyzers at controlled flow and pressure. The sample handling system — pumps, filters, heated lines, condensers, flow controllers — causes more CEMS downtime than the analyzers themselves.
Data acquisition system (DAS). Collects the analyzer signals, converts them to concentration values (using calibration curves), applies dilution or moisture corrections, calculates hourly averages, and stores the data. The DAS also flags calibration cycles, maintenance periods, and invalid data.
The Dilution vs. Hot-Wet Decision
The first CEMS design decision is sample extraction method:
Dilution extraction. The sample is diluted with clean, dry air at the probe tip (typically 50:1 to 200:1 dilution ratio). This drops the dew point so the sample can be transported at ambient temperature without condensation. Advantages: simpler sample handling, less maintenance. Disadvantages: the dilution ratio must be precisely controlled and verified; lower pollutant concentration means the analyzers need better sensitivity.
Hot-wet extraction. The sample is extracted undiluted and kept above the dew point (typically 150–180°C) all the way to the analyzers. Advantages: higher concentration for measurement, no dilution uncertainty. Disadvantages: heated sample lines are expensive and can fail; any cold spot causes condensation that scrubs soluble gases (SO2, HCl) from the sample.
For most industrial combustion sources, dilution extraction is the more reliable choice. The simpler sample handling system means less downtime. The exception: if you’re measuring very low concentrations (single-digit ppm) or water-soluble gases (HCl, HF, NH3), hot-wet may be necessary.
Calibration: The Daily Ritual That Proves the System Works
CEMS calibration isn’t optional. It’s required — typically daily for zero and span checks, quarterly for linearity checks (EPA’s 40 CFR Part 60 and Part 75 requirements). Each calibration cycle involves:
Zero check. Zero gas (clean air or nitrogen) is introduced at the probe or as close to it as practical. The analyzers should read zero (or within the zero drift specification, typically ±2% of span). If the zero has drifted, the system needs maintenance — a dirty optical cell, detector aging, or electronic drift.
Span check. A calibration gas with known concentration (typically 80–90% of the analyzer’s full scale range) is introduced. The analyzer must read within the calibration error specification, typically ±2.5% of span for most EPA programs. If it doesn’t, the data from the last valid calibration to the present is potentially invalid.
Calibration gas quality. EPA Protocol gases are certified to ±2% accuracy and must be traceable to NIST standards. They’re expensive — $300–500 per cylinder — and you’ll go through 15–20 cylinders per year for a typical multi-gas CEMS. Store them properly, track expiration dates, and never run out. Running out of calibration gas means you can’t calibrate, which means your data is invalid, which means you’re reporting invalid data to the regulator.
The Most Common CEMS Problems
Sample line blockage. Particulate matter from the stack can plug the sample probe filter or build up in the sample line. A plugged sample line means no sample reaches the analyzers — which looks like zero concentration to the analyzers, potentially triggering a false “low emission” reading. Regular probe filter cleaning (weekly for dirty stacks) and maintaining proper sample line temperature prevents this.
Moisture in the sample system. If a chiller fails or a heated line develops a cold spot, moisture condenses in the sample system. Water dissolves SO2, NO2, and HCl. The result: the analyzer reads lower than the actual stack concentration. This is a serious problem because it’s an under-report error — the regulator will not be sympathetic.
Analyzer drift. All analyzers drift over time. The daily zero/span calibration catches drift, but if the daily calibration fails, you’re into corrective maintenance. The key metric is “calibration error” — the difference between the calibration gas value and what the analyzer reads. If it exceeds the limit (typically ±2.5% of span), you need to troubleshoot and fix the analyzer before resuming monitoring.
Data gaps. CEMS data is reported as hourly averages. If the CEMS is down for calibration, maintenance, or malfunction, those hours are “missing data.” EPA rules require substitution of missing data with conservative values (the 95th percentile of recent data, or the maximum potential concentration). Missing data periods count against your annual uptime requirement (typically 90–95% depending on the rule). Too many data gaps, and you’re out of compliance even if your actual emissions are fine.
The Technician Factor
A CEMS is not a “set it and forget it” system. It needs a dedicated technician — or at minimum, a trained operator with allocated CEMS time every day. The daily routine: check the data from the previous 24 hours, review calibration results, identify any alarms or flags, and log everything in the CEMS maintenance log.
The regulator will ask for the maintenance log during an inspection. A well-kept log that shows consistent attention to the system builds credibility. A sparse or missing log suggests the CEMS data might not be reliable — which can trigger a full audit of your emissions data.
CEMS is expensive, demanding, and unforgiving of neglect. But when properly specified, installed, and maintained, it provides defensible emissions data that protects you in an enforcement situation. The plants I’ve seen with the best CEMS performance share two characteristics: they bought quality equipment up front (not the lowest bid), and they allocated dedicated technician time to daily CEMS care. You can’t cheap out on either one and expect reliable compliance data.