CEMS: continuous emission monitoring on a medical waste incinerator

An emission limit is a figure in a permit until something measures it. Continuous emission monitoring is what turns that figure into a record, and the parameter list is what decides how much of a medical waste incinerator is instrumented and how much is sampled by hand.

A two-chamber solid waste incinerator with its control panel and combustion air fan
Incineration & flue gas

What CEMS is and what it is not

CEMS stands for continuous emission monitoring system, and it is three things assembled: analysers that measure the gas, a sampling system that brings a representative sample of it to those analysers, and a data system that averages, corrects and stores the result. Continuous means it measures while the plant runs, so what it produces is a mean over a period rather than a spot reading, and it is the mean that the permit is written against.

It is worth separating CEMS from the idea of a stack test. A stack test is a campaign: a team samples the gas for a defined period with a reference method, and the result describes that period. A CEMS is an instrument that is present all the time, which is what makes it useful for operation as well as for compliance — it shows what the plant was doing at the moment a figure moved.

Continuous emission monitoring for medical waste incinerators

What has to be monitored is not chosen by the operator. It follows from the emission limit that applies to the site, and that limit is set by the regulation in force there. For dedicated medical-waste incineration the national standard that the emissions chart here draws on is GB 39707-2020, whose limits carry the same numeric values as those GB 18484-2020 sets for hazardous-waste incineration; the two differ in scope and in how the residual metals are grouped, since GB 18484-2020 excludes dedicated medical-waste facilities.

Two rules follow from that. A provincial standard may be stricter than the national one, and a project may also carry an environmental impact assessment limit, so the parameter list is read from the permit that applies rather than from a general table. And the list covers the plant as a whole: some parameters are measured continuously and some are sampled over a period, and that difference is a property of the measurement rather than a statement about importance.

Parameters monitored in a medical waste incinerator

The list below is the limit set in the emissions chart here, which draws for dedicated medical-waste incineration on GB 39707-2020 Table 4 and for hazardous-waste incineration on GB 18484-2020 Table 3. Values are mg/m³ unless stated, referred to dry flue gas at 11 % oxygen under standard conditions.

  • Particulates — 30 as a one-hour mean and 20 as a 24-hour mean. Measured continuously, most often by a light-scattering or opacity instrument rather than by weighing.
  • Carbon monoxide — 100 as a one-hour mean and 80 as a 24-hour mean. A combustion indicator rather than a cleaning-stage one, and the fastest warning of incomplete combustion the plant has.
  • Nitrogen oxides — 300 as a one-hour mean and 250 as a 24-hour mean. The figure any SNCR or SCR stage downstream is installed to meet.
  • Sulphur dioxide — 100 as a one-hour mean and 80 as a 24-hour mean. It follows the waste and the reagent, and moves with the scrubber or sorbent dosing.
  • Hydrogen chloride — 60 as a one-hour mean and 50 as a 24-hour mean. The parameter that tracks the chlorine in the feed and the performance of the acid-gas stage.
  • Hydrogen fluoride — 4.0 as a one-hour mean and 2.0 as a 24-hour mean. Listed for medical and hazardous waste, and not listed for municipal solid waste.
  • Mercury and its compounds — 0.05 as a measurement mean. Sampled over a period rather than measured continuously, and the parameter that most often decides whether carbon has to be injected.
  • Cadmium and thallium — 0.05 each as a measurement mean and listed separately, where municipal solid waste sets one combined figure of 0.1.
  • Other metals — lead, arsenic and chromium at 0.5 each, with a combined group at 2.0, all as a measurement mean. Sampled, because these travel on the dust and are captured with it.
  • Dioxins and furans — 0.5 ng TEQ/m³ as a measurement mean. There is no continuous analyser for them, so the figure comes from periodic sampling and is expressed as a toxic equivalent over a defined period.

Oxygen is measured too, and not because it is a pollutant. Every limit in that list is referred to 11 % oxygen in dry gas, so a concentration measured in the duct means nothing until the oxygen and the moisture at that moment are known. Temperature, flow and pressure are measured for the same reason: they turn a concentration into a mass emission rate and let the reading be corrected onto the basis the limit is written on.

Continuous emission monitoring of incinerator flue gas: how the measurement is made

Two families of instrument cover the list. An extractive system draws gas out through a probe and a heated line to an analyser, which is the usual arrangement for measurements that need a cell or a spectrometer — oxygen, carbon monoxide, nitrogen oxides, sulphur dioxide. An in-situ instrument measures across the duct or at the wall, which suits a particulate or opacity monitor and does away with the sampling line altogether. Each has a weakness: an extractive line deposits salt and has to be heated, and an optical instrument reads a trend rather than a mass.

Particulates need a word of their own. A continuous particulate monitor responds to light scattered by the dust, so its output is proportional to something rather than being a weight; the relation to mg/m³ is established by comparing it against a gravimetric reference method, and it holds only while the dust keeps its character. A change in the dust can move the instrument without moving the emission.

Sample conditioning is where a good deal of CEMS trouble lives. Cooling and drying a sample to protect an analyser also removes the compounds that dissolve in water, and hydrogen chloride and sulphur dioxide are two of them. Where those limits are strict, the measurement is often made hot and wet for exactly this reason.

Flue gas monitoring for waste incineration beyond the stack

The stack instruments say whether the plant complies. The instruments on the plant say why, and the combustion parameters come first. GB 39707-2020 Table 1, as listed in the emissions chart here, sets the medical-waste incinerator indicators as a high-temperature-section temperature of at least 850 °C, a flue-gas residence time of at least 2.0 seconds and a combustion efficiency of at least 99.9 %; for hazardous waste the temperature indicator rises to at least 1,100 °C.

  • Secondary chamber temperature — measured continuously, because the destruction of the organic compounds depends on it.
  • Oxygen at the chamber outlet — part of the combustion efficiency figure and the reference for everything measured at the stack.
  • Carbon monoxide — a fast signal of incomplete combustion, and the one that moves long before a dioxin sampling run would show anything.
  • Quench outlet temperature — the figure the cooling section is controlled on, and the one that decides what the bag material has to tolerate.
  • Filter inlet temperature and compartment differential pressure — the condition of the cake that the adsorption stage depends on.
  • Reagent and sorbent dosing rates — the scrubber, the sorbent injection and the carbon feed, each with its own alarm.

A dioxin figure is a slow measurement taken over a period, so it cannot be used to steer a plant. The fast parameters are what an operator actually runs on, and the periodic sampling is what confirms that the fast parameters were the right proxies.

What the emission monitoring data for incinerator compliance is used for

  • Demonstrating that the limit was held over the averaging period the permit names, which is what emission monitoring data for incinerator compliance exists for.
  • Telling an instrument fault from a plant fault. A step change that keeps its shape across several parameters usually belongs to the analyser or the sampling system rather than to the furnace.
  • Showing which stage is failing — particulate points at the filter, hydrogen chloride at the acid-gas stage, nitrogen oxides at the reagent dosing, carbon monoxide at the combustion.
  • Reading the combustion while there is still time to act on it, which is the use the fast parameters are for.
  • Scheduling maintenance from a trend instead of from a breakdown, which is where the operating cost of a plant is decided.
  • Reporting — to the authority where the permit requires it, and to the operating record of the plant in every case.

All of it rests on the instrument being looked after. Zero and span checks, calibration against the reference method, and the record of how much of the time the system was available are not paperwork around the measurement; on many sites they are the part of the submission that decides whether the rest of it is accepted. Which intervals and which reference methods apply is set by the permit, and the analyser arrangement that fits a particular unit is confirmed with the factory for your waste composition.

Common mistakes

  • Buying a CEMS against a parameter list that was copied from another project rather than read from the permit the site holds.
  • Measuring the pollutants and forgetting oxygen, moisture and flow, without which no reading can be referred to the basis the limit is written on.
  • Reading a continuous particulate instrument as though it were a weight. It is a trend that has to be tied to a gravimetric method.
  • Cooling and drying a sample that contains hydrogen chloride, and then wondering where the hydrogen chloride went.
  • Treating a dioxin result as an operating figure. It is a mean over a sampling period, and it describes the plant rather than the minute.
  • Running the instrument without a calibration and availability record, which leaves numbers and no way to defend them.

From a monitored parameter to a compliant plant

Continuous emission monitoring is what turns a limit into a record, and the record is only as good as the sampling system and the calibration behind it. The limits it has to show are set out in the incinerator emissions chart, the stages that have to be held inside those limits are in the article on flue gas treatment in an incinerator, and the combustion conditions they rest on are in the article on why incinerators run at 850 °C for two seconds. The parameters to monitor and the analyser arrangement for your plant are confirmed with the factory for your waste composition and the permit that applies to your site.

Tell us what the site has to treat.