Guidance

Refineries: pollution inventory reporting

Updated 21 August 2024

Applies to England

• Publication for Northern Ireland
• Publication for Scotland
• Publication for Wales

If you operate an A1 facility you must submit data to the pollution inventory. The Environment Agency will have sent you a notice explaining this when your environmental permit was issued.

The ‘general guidance’ available in the pollution inventory reporting guidance gives information that applies to all business and industries. It explains what the pollution inventory is and how to report.

This guidance gives information specific to the refineries sector.

If your operations include production of chemicals, waste incineration, or combustion please also read the guidance notes for these activities.

Emissions to air

Relevant pollutants

The most common pollutants released by a refinery, and their main sources, are:

• carbon dioxide (CO₂) – from:
  • process furnaces
  • boilers
  • gas turbines
  • fluidised catalytic cracking regenerators
  • carbon monoxide boilers
  • flare systems
  • incinerators
• carbon monoxide (CO) – from:
  • process furnaces
  • boilers
  • gas turbines
  • fluidised catalytic cracking regenerators
  • CO boilers
  • flare systems
  • incinerators
  • sulfur recovery units
• nitrogen oxides (NOx) – from:
  • process furnaces
  • boilers
  • gas turbines
  • fluidised catalytic cracking regenerators
  • CO boilers
  • flare systems
  • incinerators
  • coke calciners
• particulate matter (including PM2.5 and PM10) – from:
  • process furnaces
  • boilers
  • particularly fluidised catalytic cracking regenerators
  • CO boilers
  • incinerators
  • coke calciners
• sulfur oxides (SOx) – from:
  • process furnaces
  • boilers
  • gas turbines
  • fluidised catalytic cracking regenerators
  • CO boilers
  • flare systems
  • incinerators
  • sulfur recovery units
  • coke calciners
• volatile organic compounds (VOC) – from:
  • storage and handling facilities
  • gas separation units
  • oil to water separation systems
  • fugitive emissions (valves, flanges, and so on)
  • vents
  • flare systems

Use this list as a guide only and check that there are no other pollutants emitted from your process that must be reported.

Specific pollution inventory guidance for refineries from Fuels Industry UK (FIUK) (formerly UK Petroleum Industry Association (UKPIA)) is at the end of this guidance. The guidance covers the major categories of refinery emission sources. The document is split into three categories of pollutant:

• those that will almost certainly be emitted in quantities above the reporting threshold
• those emitted but generally in quantities below the reporting threshold
• those that are unlikely to be emitted by refineries

See the FIUK website for the latest updates and further information. We will update this page as appropriate if required.

Emission sources

Point source emissions

Point source emissions will be the most significant emission sources for combustion activities taking place within refineries. They also come from refinery process sources, such as:

• vacuum distillation
• catalytic cracking
• sulfur recovery

Fugitive emissions

Examples of fugitive emissions from refineries include:

• VOC emissions from piping systems
• wastewater systems
• storage tanks
• loading and unloading systems
• storage and handling

Diffuse VOC emissions sources include:

• pumps
• compressors
• valves
• flanges

Although the release from each individual source may be small, diffuse VOC emission sources may contribute 20% to 50% of the total VOC emissions. This is due to the large number of such sources in a refinery.

We tell you how to estimate fugitive VOC emissions later in this guide.

How to quantify your emissions

In this section, we have given some example sources of emissions relevant to refineries and release estimation techniques (RETs) that you might need.

Take care to make sure the emission concentration and flow rate are compatible. For example, normalised emission concentrations should be multiplied by normalised volumetric flow rates. Actual, measured emission concentrations should be multiplied by actual, measured volumetric flow rates. Normalised emission rates are quoted in terms of a standard oxygen concentration, and are usually dry gas, at a temperature of 273K and a pressure of 101.3kPa.

Check our ‘technical guidance and equations’ for calculations to convert between normalised and actual emission concentrations.

Carbon dioxide emissions from refineries

The UK emissions trading system (UK ETS) allows carbon dioxide (CO₂) emissions to be determined by either calculation or by measurement. Where you use a measurement methodology, you still need to verify the measured emissions by calculation.

In both cases, it is likely that you will have measurements of fuel burned or material flow. You can use these to work out other emission quantities when combined with appropriate emission factors.

We have given equations to calculate carbon dioxide emissions from processes specific to refineries. Other generic equations are in ‘technical guidance and equations: pollution inventory reporting’.

For catalytic cracker regeneration and other catalyst regeneration, use Equation 1.

Equation 1: ECO₂ = A_e × EF_CO2 × CF

Where:

• ECO₂ = emission of CO₂, te/yr
• A_e = amount of coke burned from the catalyst, te/yr
• EF_CO2 = amount of CO₂ per tonne of coke burned based on carbon content of coke, te/te
• CF = conversion factor (taken as 1)

For refinery hydrogen production, use Equation 2.

Equation 2: ECO₂ = A_e × EF_CO2

Where:

• ECO₂ = emission of CO₂, te/yr
• A_e = amount of hydrocarbon feed processed, te/yr
• EF_CO2 = emission factor (CO₂ per te of feed)

Combustion source factors

For refinery combustion source factors, see FIUK pollution inventory guidance at the end of this guidance. We have given emission factors to help you estimate combustion products. The preferred RET for sulfur dioxide (SO₂), is mass balance based on the sulfur content of the fuel.

Emissions of some trace elements can be estimated using mass balance when fuel composition data are available. If you don’t have the data to do this, you can use default emission factors.

Some sites have oil and gas fired boilers, and furnaces that do not use continuous emissions monitoring systems (CEMS) for annual mass emissions. If this applies to you, you can use the fuel burn and any site-specific nitrogen oxides (NOx) factors which you have previously agreed with us.

Refinery process source factors

For refinery process source factors, see FIUK pollution inventory guidance for refineries at the end of this page.

Fuel analysis and process stream data

The use of fuel analysis and process stream data to determine emissions is like the use of emission factors. Check our ‘technical guidance and equations’ for equations and an example for fuel analysis emission calculations.

Fugitive VOC emissions

For the purposes of this note, we have split fugitive VOC emissions from refineries into four categories:

• process fugitives
• tank farm fugitives
• loading and unloading fugitives
• drainage and effluent system fugitives

Guidance for how to estimate fugitive releases for UK refineries is in two separate protocols produced by the Energy Institute. Note that these have been produced by an external organisation and there is a charge to access the protocol. This has been agreed through the trade body.

The first provides a methodology (speciation protocol), to be used with total volatile organic compound (VOC) releases. This is to work out the fractional speciation of hydrocarbon emissions from oil refineries.

• Protocol for the estimation of petroleum refinery process plant fugitive VOC emissions

The second can be used to estimate annual emissions of non-methane VOCs (NMVOCs), including fugitive VOC releases. Use this when you don’t have enough information to apply the methodology.

• Protocol for the estimation of VOC emissions from petroleum refineries and gasoline marketing operations

Process fugitives

Two methods for calculating process fugitives are presented in the VOC protocol.

The first method uses the USEPA protocol for equipment leak estimates. This protocol assumes knowledge of the number of valves, flanges, and seals on a refinery. It takes a tiered approach to estimating emissions:

1. Tier 1 applies average emission factors based upon the process service

2. Tier 2 applies average emission factors based on leak or no leak criteria

• this requires the use of VOC monitoring equipment to measure threshold VOC concentrations at each fitting

3. Tier 3 applies emission correlations based on actual VOC concentration methods determined at each fitting

The emission factors you need to estimate emissions are given as Table 1. These are for NMVOCs only.

Table 1: Average emission factors for process fugitives

Equipment type	Service type	Emission factor (kg/hr/source)
Connectors	Gas	2.50 x 10-4
Connectors	Light liquid	2.50 x 10-4
Connectors	Heavy liquid	4.34 x 10-5
Flanges	Gas	2.50 x 10-4
Flanges	Light liquid	2.50 x 10-4
Flanges	Heavy liquid	4.68 x 10-5
Compressor seals	Gas	0.636
Pump seals	Light liquid	0.114
Pump seals	Heavy liquid	3.49 x 10-3
Valves	Gas	0.0268
Valves	Light liquid	0.0109
Valves	Heavy liquid	9.87 x 10-5
Open ended lines	All	2.30 x 10-3
Pressure relief valves	Gas	0.16
Sampling connections	All	0.015
Other	Heavy liquid	5.18 x 10-5

Where no screening values are available for your equipment, you should use Tier 1 ‘average emissions factors’. You will need to use Equation 3, which is a generic equation to estimate emissions from all sources in a stream, for a particular equipment type.

Equation 3: E = F × WF × N

Where:

• E = emission rate (kg/hr) of VOC from all sources grouped in a particular type and service equipment (for example, valves in light liquid service)
• F = average emission factor for the equipment type (from Table 1)
• WF = the average weight fraction of VOC in the stream
• N = the number of pieces of equipment grouped according to equipment type, service, and weight fraction of VOC category

The average emission factors are in terms of VOCs, but you still need to input the weight fraction of VOC in the process stream (‘WF’ in Equation 3). This is to account for any non-organic compounds. For example, if the stream contains water vapour, you will need to account for this in your calculations. Example 1 shows how to use this methodology.

For the Tier 1 method, you must first work out the number and service type of each equipment type in the refinery (Step 1). Service types are:

• gas and vapour: material in a gaseous state under operating conditions
• light liquid: material in a liquid state where the sum of the concentrations of individual constituents with a vapour pressure over 0.3kPa at 20°C ≥ 20weight percent (wt%)
• heavy liquid: material that does not fall under either of the above two definitions

Next, group the inventory into ‘streams’ (Step 2). To simplify calculations, we recommend you group the service mode combinations identified in Step 1 (for example, valves in gas service) into streams. You can do this according to the approximate weight fraction of VOCs (WF) in each stream.

Another simplification might be to group areas of the refinery according to the ‘average’ weight fraction of VOCs in the process streams. You will need to take account of the various service modes for each equipment type (for example, gas, light liquid, heavy liquid) contained within that area.

If in doubt, you could take the conservative assumption that all streams are approximately 100% VOCs, thereby making WF = 1.

When you have grouped your ‘streams’, use emission factors to estimate emission rates (Step 3).

Use the relevant emission factors and Equation 3 to calculate the emissions from each equipment type. Add these emissions together to derive a total emission rate for all equipment pieces quantified using this methodology. Finally, for the specific equipment category defined by the above three steps, you need to estimate the number of operational hours (Step 4).

Example 1 (using Equation 3)

1. A particular section of the refinery has 300 valves (Step 1)
2. 200 of these are in gas service (Step 1).
3. Within this smaller group of valves in gas service, it is ascertained that 100 valves are, on average, 80 weight percent NMVOCs, 10 % methane, and 10% water vapour (Step 2).
4. The appropriate emission factor for valves in gas service is 0.0268 kg/hr/source (from Table 1) (Step 3). Emissions from this group of valves are thus estimated by using the following parameters: F = 0.0268, WF = 0.8, N = 100.
5. It is estimated that this group of valves operates for 5,500 hours per year (Step 4).
6. The final emission estimate for the group of 100 valves specified above is approximately 11,792 kg NMVOCs per year (kg/hour/source multiplied by annual hours of operation).

The above steps would then be repeated for the remaining 200 valves that were not included in the above estimate for that section of the refinery. Similarly, emissions need to be calculated from other potential fugitive emission sources in that section, followed by the next refinery ‘section’ and so on until fugitive emissions from the entire refinery have been quantified.

For Tier 2 (leak or no leak methodology), you need to conduct screening using a portable monitoring device.

First, you need to measure leaks from fugitive sources (Step 1).

A leak is typically defined and recorded if a screening value of >10,000 ppmv is returned by an appropriate calibrated monitoring instrument. The emission factor you choose from Table 2 will therefore depend on whether the component tested returns a pass (a reading ≥10,000 ppmv) or fail (a reading <10,000 ppmv).

Next, estimate the VOC emission rate for your equipment type (Step 2).

You can use Equation 4 to estimate emissions for each of the equipment types listed.

Equation 4: E = (FG × NG) + (FL × NL)

Where:

• E = VOC emission rate for the equipment type (kg/hr)
• FG = emission factor for sources with screening values ≥10,000 ppmv (kg/hr/source)
• NG = for the equipment type of concern, the number of sources with screening values ≥10,000 ppmv
• FL = emission factor for sources with screening values <10,000 ppmv (kg/hr/source)
• NL = for the equipment type of concern, the number of sources with screening values ≥10,000 ppmv

When you have completed Step 2, you can determine total VOC emissions (Step 3).

If all process fugitive sources have been screened, you can calculate total VOC emissions from all sources by adding emission rates from each individual equipment component.

The emission factors you need to estimate emissions using the steps discussed above are presented as Table 2.

Table 2: Leak and no leak emission factors for process fugitives

Equipment type	Service	Leak (≥10,000 ppmv) emission factor (kg/hr)	No leak (<10,000 ppmv) emission factor (kg/hr)
Connectors and flanges	All	0.0375	0.00006
Valves	Gas	0.2626	0.0006
Valves	Light liquid	0.0852	0.0017
Valves	Heavy liquid	0.00023	0.00023
Pump seals	Light liquid	0.437	0.012
Pump seals	Heavy liquid	0.3885	0.0135
Compressors	Gas	1.608	0.0894
Pressure relief valves	Gas	1.691	0.0447
Open ended lines	All	0.01195	0.0015

Finally, for the specific equipment pieces screened, you need to estimate the annual number of operational hours (Step 4). This is required to derive annual emissions based on the hourly emission rates.

You can only use the Tier 3 (correlation equations) method if you get screening values (ppmv) through a fugitive leak screening programme. You must collect the required screening value (SV) data using an appropriate calibrated monitoring instrument.

The following points are important to note when using this methodology:

• emission estimates are for ‘total organic compounds’ (TOC) and, therefore, a correction must be made to convert the estimates to NMVOCs (to exclude methane)
• emission factors are on a ‘per source’ basis
• each individual screening value must be entered into the correlation equation to predict emissions for an equipment piece - it is important not to average screening values and then enter the average value into the correlation equation to estimate emissions

To determine fugitive emissions using the correlation equation approach, you first need to measure leaks from fugitive sources.

For each piece of equipment tested, the screening value will fall into one of three categories. You must use the correct estimation methodology for each category, as follows:

• for ‘zero’ readings (no emission detected)
  • if no emissions are detected (below detection limit), use the ‘default zero emission rate’ emission factors (see Table 3)
  • if the lower detection limit of the monitoring device is >1 ppmv, use half the detection limit.
• for screening values (SVs) between the lower and upper detection limits of the monitoring device
  • if SVs are determined through testing (measured level is between the lower and upper detection limits), use the ‘correlation equations’ in Table 3 to determine the leak from each relevant component tested
• for values greater than the upper detection limit of the monitoring device (a ‘pegged’ emission reading)
  • use the ‘pegged emission rate’ emission factors presented in Table 3

You can group pieces of equipment if they share the same process stream, and so have similar VOC or TOC ratios. TOC emissions can be added for each equipment group at this stage. This reduces the number of calculations you need to do.

Once you have estimated emissions from each source, you need to convert the emissions from TOCs to VOCs. To do this, you need additional information on the approximate weight % of VOCs and TOCs in the process streams from which the emissions originate. Combine this with the emission estimate for each equipment component using Equation 5.

Equation 5: E_VOC = E_TOC × (WP_VOC / WP_TOC)

Where:

• E_VOC = the VOC emission rate from the equipment (kg/hr)
• E_TOC = the TOC emission rate from the equipment (kg/hr) calculated using the emission factors for correlations from Table 3
• WP_VOC = the concentration of VOC in the equipment in weight %
• WP_TOC = the concentration of TOC in the equipment in weight %

For the specific equipment pieces tested, you should estimate the annual number of operational hours. This information is required to derive annual emissions based on the hourly emission rates.

If all process fugitive sources have been tested, you can determine total VOC emissions from all sources by adding the emissions from each individual equipment component.

Table 3: Correlation equations for process fugitive emissions

Equipment type	Default zero emission rate (kg/hr)	Pegged emission rate (kg/hr): 10,000 ppmv	Pegged emission rate (kg/hr): 100,000 ppmv	Correlation equation (kg/hr)
Connector	7.5 × 10-6	0.028	0.028	leak = 1.53 × 10-6 (SV) 0.735
Flange	3.1 × 10-7	0.085	0.084	leak = 4.61 × 10-6 (SV) 0.703
Valve	7.8 × 10-6	0.064	0.14	leak = 2.29 × 10-6 (SV) 0.746
Open-ended line	2.0 × 10-6	0.03	0.079	leak = 2.20 × 10-6 (SV) 0.704
Pump seal	2.4 × 10-5	0.074	0.16	leak = 5.03 × 10-5 (SV) 0.610
Other	4.0 × 10-6	0.073	0.11	leak = 1.36 × 10-5 (SV) 0.589

The second method presented in the VOC protocol provides a very crude estimation of your emissions. This method should only be used if you don’t have data on the number of valves, flanges, pump seals, and so forth, for the refinery under consideration.

This method assumes that fugitive VOC emissions from plant and pipework represent 0.03wt% of the actual material processed in the refinery. This includes all thermally processed material but not product blend stocks.

Thus, annual process fugitives = 0.0003 × annual refinery throughput (tonnes).

Tank farm fugitives

For a detailed coverage of estimation of emissions from both floating and fixed roof tanks refer to section 2 of the VOC protocol.

Loading and unloading fugitives

For a detailed coverage of estimation of emissions from both controlled and uncontrolled loading operations refer to section 3 of the VOC protocol.

Drainage and effluent system fugitives

For a more detailed coverage of estimation of emissions from drains, separators, air flotation units and biological treatment facilities refer to section 1.9 of the VOC protocol.

Speciation of VOC emissions

The VOC speciation protocol provides a methodology for determining the fractional speciation of hydrocarbon emissions from oil refineries. The results are reported as the mass fraction of the annual hydrocarbon emissions. This provides individual annual mass release data for the following pollution inventory substances:

• benzene
• toluene
• xylene (all isomers)
• 1,3-butadiene
• 1,2-dichloroethane
• 1,1,2,2-tetrachloroethane

Emissions to water

Emissions of substances to water can be either direct to controlled waters or indirect, following transfer to off-site effluent treatment plant.

Check the ‘general guidance’ in the pollution inventory reporting: guidance notes for what constitutes an emission or a transfer.

Relevant pollutants

Water discharges from refinery processes can be contaminated by a variety of substances. The most common discharges, and their main sources, are:

• phenols – from:
  • distillation units
  • visbreaker
  • catalytic cracking
  • ballast water
• organic chemicals (TOC) – from:
  • distillation units
  • hydro treatment
  • visbreaker
  • catalytic cracking
  • hydro cracking
  • spent caustic
  • storage tanks water heel drainings
• benzene, toluene, and xylene – from:
  • distillation units
  • hydro treatment
  • catalytic cracking
  • visbreaker
  • storage tanks water heel drainings

Use this list table as a guide only, and check that there are no other pollutants emitted from your processes.

The following determinands are generally measured directly from the effluent discharge unless otherwise indicated:

• arsenic (As)
• benzene (API factor of 6µg/l applied to process water)
• cadmium (Cd)
• chromium (Cr)
• copper (Cu)
• ethyl benzene
• lead (Pb)
• mercury (Hg)
• nickel (Ni)
• nitrogen – total (N)
• phenols
• toluene (API factor of 1 µg/l applied to process water)
• total organic carbon (TOC)
• xylene (API factor of 0.25 µg/l applied to process water)
• zinc (Zn)

Emission sources

More generally, emissions to water arise from the following sources:

• process water: steam and wash water - these waters become contaminated with process fluids, dissolved gases and, apart from oil, will also have taken up hydrogen sulfide, ammonia, and phenols; the process water is treated in several steps before discharge to the environment
• cooling water: once-through or circulating systems - this stream is theoretically free of oil, however, leakage into once-through systems, even at low concentrations, can result in significant mass losses because of the large volume of water involved
• rainwater from process areas - this type of water has not been in contact with the process fluids, but it comes from rainfall on surfaces which are possibly oil contaminated; this water is typically treated prior to discharge to the environment

The resulting discharges of these substances depend on the ‘in process’ preventative measures (such as good housekeeping and re-use) and the presence and technical standards of wastewater treatment facilities.

Off-site waste transfers

You must classify wastes using the European Waste Catalogue 6-digit codes and the relevant Waste Framework Directive disposal or recovery codes. Check the ‘reporting codes list’ in the pollution inventory reporting: guidance notes.

Relevant wastes

The amount of waste generated by refineries is small when compared to the volume of raw materials and products that you process. Oil refinery waste normally covers three categories of materials:

Oily and non-oily sludges

Oily and non-oily sludges originate from sources such as:

• alkylation units
• biotreaters
• boiler feed water preparation
• cleaning of heat exchanger bundles and equipment
• crude and product tank bottoms
• desalters
• oil spills
• soil remediation

Oily sludges represent the largest waste category from refineries. Bio-sludge production only takes place if a refinery operates a bio-treater.

Other refining wastes

This category includes other wastes produced from:

• refining processes
• petroleum handling operations
• wastewater treatment

Both hazardous and non-hazardous wastes are generated. Spent catalysts originate from:

• reformers
• catalytic crackers
• hydro crackers
• hydrodesulfurisation
• hydro treating units

Non-refining wastes

These would be, for example, domestic-type waste, demolition, and construction wastes.

Do not include waste from outside your installation boundary, such as office and canteen waste.

Transboundary shipments of hazardous waste

In your pollution inventory return, you must report the annual quantities of any transboundary hazardous waste shipments taking place.

FIUK pollution inventory guidance for refineries

This guidance was produced by the Refinery Emissions Working Group of UKPIA (now FIUK). It was developed through a review of relevant industry guidance. In particular it used information from the US Environment Protection Agency, the American Petroleum Institute and the Canadian Petroleum Products Institute. In this section, we have given some key information from the guidance for your reference.

This guidance is designed to be an aid for compiling pollution inventory submissions from petroleum refineries. It covers the major categories of emission sources:

• process emissions
• combustion emissions
• wastewater
• other

The following tables are taken from the UKPIA guidance. They include all the pollutants which we consider are most likely to be emitted from a typical oil refinery. For pollutants not contained within the tables, you can enter ‘n/a’ (not applicable) in most cases.

Entering ‘n/a’ indicates that this pollutant is not knowingly discharged by your site at all. It is your responsibility to check that there are no other pollutants emitted from your process that must be reported.

In method column:

• M = measurement
• C = calculation
• E = estimation (using engineering judgement).

Other abbreviations used in the table are:

• AP42 – US EPA emission factor guidance AP-42: Compilation of Air Emissions Factors
• API – American Petroleum Institute
• BRT – below reporting threshold
• CPPI - Canadian Petroleum Products Institute
• EEA – European Environment Agency
• EMEP - European Monitoring and Evaluation Programme
• FCC[U] – fluidised catalytic cracking [unit]
• HF – hydrogen fluroide
• HHV – higher heating value
• NG – natural gas
• RFG – refinery fuel gas
• RFO - refinery fuel oil
• SRU – sulfur recovery unit
• SWS – sour water system
• UK ETS – UK Emissions Trading Scheme

Emissions to air

Category 1 emissions are almost certainly produced by refineries in volumes above the reporting threshold (returns are expected for these substances).

Table 4: Category 1 emissions and relevant information

Determinand	Method	Emission source	Reference	Factors, guidance and notes
Ammonia	C	FCCU Regen	AP42	FCCU (uncontrolled): 0.155kg/1,000L When CO boiler is present NH3 emissions from FCCU are considered negligible
Ammonia	C	SRU to stack	Industry experience	1 ppm in SRU flue gas SWS overhead gas to the sulfur recovery unit is included in this calculation
Ammonia	C	SWS to flare	Industry experience	Flared SWS gas × 0.002
Carbon dioxide (CO2)	C	Site combustion and process	UK ETS	As part of UK ETS
Carbon monoxide (CO)	C	Combustion	AP42	RFG: 1.34 E0 kg/1,000m3 (0.082 lb/MBTU) RFO: 6.00 E-1 kg/1,000L (0.033 lb/MBTU)
Carbon monoxide (CO)	M	FCCU Regen	Onsite analyser	Measurement
Chlorine	E	Reformers	Industry experience	‘brt’ from catalyst regeneration: chlorine used in regeneration of catalyst (reactivation of active sites). Losses of free chlorine from cooling towers (used as biocide) considered negligible on basis that chlorine will be dissolved in water droplets.
Dinitrogen oxide (nitrous oxide)	C	Combustion	EMEP and EEA air pollutant emission inventory guidebook	RFG: 2.16 E-2 kg/1,000m3 (22 g/million grams)
Dinitrogen oxide (nitrous oxide)	C	Combustion	AP42	NG (uncontrolled): 3.52   E-2 kg/1,000m3 NG (low-NOx-burner): 1.02   E-2 kg/1,000m3 RFO: 1.32 E-2 kg/1,000L
Dinitrogen oxide (nitrous oxide)	E	Process	Industry experience	n/a, for non-combustion plant
1,2 Di-chloroethane	M	Site	IP Speciation Protocol	Based on speciation surveys
Hydrogen sulfide (not stated on pollution inventory list of pollutants, but reportable as ‘Other individual acid forming gases’)	C	SRU to flare	CPPI	0.5% of measured H2S content in gas flared (98% combusted, 1.5% reacted with CO). For process units which are subject to strict fugitive emission control, for example HF, the emissions may be considered negligible and assumed to be zero.
Methane (CH4)	C	Combustion	AP42	RFO (industrial boiler): 1.20 E-1 kg/1,000L NG: 3.68 E-2 kg/1,000m3 Emission source is unburnt residuals in heater flue gases.
Methane (CH4)	C	Fugitive	Industry experience	RFG flow × Me content × 0.03%
Methane (CH4)	C	FCCU regen	Industry experience	4.9 E-3 kg/1,000m3 exhaust gas flow
Methane (CH4)	C	Incomplete flare combustion	Industry experience	n/a
Nitrogen oxides (NOx)	C	Combustion	Environment Agency	Site specific factors agreed with Environment Agency for combustion plant
Nitrogen oxides (NOx)	M or C	FCCU regen	Environment Agency	Site specific Environment Agency agreed factors or emission analyser
Nitrogen oxides (NOx)	C	FCCU regen	AP42	2.04 E-1 kg/1,000L cracker feed
Nitrogen oxides (NOx)	C	Flare	Industry experience	0.068 lbs/MMBtu gas flared (HHV)
Nitrogen oxides (NOx)	C	SWS to flare (stoichiometric conversion of combusted ammonia to NOx)	Industry experience	n/a
Particulate matter (total)	C	Combustion	AP42	RFO: 1.1S + 0.39 kg/1,000L (where   S = %sulfur in fuel) RFG/NG: 1.22 E-1 kg/1,000m3
Particulate matter (total)	M	FCCU regen	BS 3504: 1983	Continuous   analyser or spot analysis. PM emissions from catalytic cracking units mostly come from catalyst fines entrained in the exhaust gas from the catalyst regenerator.
PM10	C	Combustion	AP42	RFO:  86% of total particulate RFG/NG: 100% of total particulate
PM10	C	FCCU	CPPI	PM10: 70% of total particulate
PM2.5	E	Site-wide	Industry experience	Worst case conservative estimate PM2.5 = PM10 (see above)
PM2.5	M or C	FCCU, oil fired units, calciners	Industry experience	Where data exists, a lower value for PM2.5 may be reported
Sulfur oxides (SOx)	C	Combustion	Industry experience	Fuel sulfur content × flow rate
Sulfur oxides (SOx)	M	FCCU regen	Environment Agency	Continuous   analyser   or   spot analysis
Sulfur oxides (SOx)	C	FCCU regen	AP42	1.41 E0 kg/1,000L of FCC feed
Sulfur oxides (SOx)	C	SRU to stack	Industry experience	Based upon sulfur unit efficiency (% sulfur not removed)
Sulfur oxides (SOx)	C	Flare	Industry experience	Flared gas sulfur content × flow rate
Toluene	M or C	Site and fugitive	IP speciation protocol	In-house data based on site speciation protocol
Non-methane volatile organic compounds (NMVOCs)	M or C	Site or fugitive	IP estimation protocol	Modified by measured data as it becomes available from LDAR programmes.
Poly aromatic hydrocarbons (PAHs) - total	C	Combustion	API	RFO:  Sum of Borneff Six: 1.39 E-6 kg/1,000L NG: Sum of Borneff Six: 1.10 E-7 kg/1,000m3
Poly aromatic hydrocarbons (PAHs) - total	C	FCCU	CARB	9.3 E-8 kg/1,000L fresh feed (6.71 E-6 lbs/MBar)
Benzo(a)pyrene	C	Combustion	AP42	RFO: 1.05 E-7 kg/1,000L NG: 9.60 E-9 kg/1,000m3
Benzo(a)pyrene	C	FCCU	CARB	4.0 E-9 kg/1,000L fresh feed (2.60 E-7 lbs/MBar)
Benzene	M or C	Combustion	Industry experience	n/a
Nickel	C	Combustion	AP42	RFO: 1.01 E-2 kg/1,000L
Nickel	M or C	FCCU	Industry experience	Catalytic cracker particulates × 584 ppm
Vanadium	C	Combustion	AP42	RFO: 3.82 E-3 kg/1,000L
Vanadium	M or C	FCCU	Industry experience	Catalytic cracker particulates × 1,053 ppm
Zinc	C	Combustion	AP42	RFO: 3.49 E-3 kg/1,000L

Category 2 emissions may be produced by refineries but are generally below the reporting threshold. These are given as Table 5.

Table 5: Category 2 emissions and relevant information

Determinand	Method	Emission source	Reference	Factors, guidance and notes
Arsenic	C	Combustion	AP42	RFO: 1.58 E-4 kg/1,000L
Hydrogen chloride	E	Reformer	Industry experience	‘brt’ from catalyst regeneration: chlorine used in regeneration of catalyst (reactivation of active sites)
Hydrogen chloride	E	Combustion	Industry experience	‘brt’ for process units which are subject to strict fugitive emission control, for example HF, the emissions may be considered negligible and assumed to be zero
Mercury	C	Combustion	AP42	RFO: 1.36 E-5 kg/1,000L
Tetrachloroethylene	E	Site or fugitive	Industry experience	Industry consensus ‘brt’ or n/a
Trichloroethylene	E	Site or fugitive	Industry experience	Industry consensus ‘brt’ or n/a

Category 3 substances are unlikely to be produced by refineries (no returns are expected for these substances) unless caused by specific unit operations.

Fluorine and inorganic compounds as HF may arise where HF alkylation is carried out. HF detection equipment should be used, and ‘brt’ should be reported where HF emissions are not detected.

For dioxins from combustion, use method E. Industry consensus indicates refineries are not a source of dioxin emissions, so assume ‘brt’ unless you have evidence otherwise. For more information, see Barnes, ‘UK Oil Refining and Atmospheric Emissions of Dioxins’ (2004) and ENTEC, ’Development of UK Cost Curves for Abatement of Dioxin Emissions to Air’ (2003).

Substances listed in the pollution inventory Schedule but not included in these tables are unlikely to apply to refinery emissions. Other substances should be entered as ‘n/a’ in your pollution inventory return unless you have evidence otherwise.

If you carry out non-refinery activities, you might release other substances which need to be reported.

Releases to controlled waters and transfers in wastewater

Wherever possible data should be net of incoming cooling water quality. Where this would result in a negative return, for example for removal of pollutants from incoming waters, write ‘brt’.

Category 1 emissions are almost certainly produced by refineries in volumes above the reporting threshold. As such, returns are expected for these substances.

Emissions from effluent treatment facilities should be measured using effluent quality data. Several of these emissions may be below the reporting threshold for PI. If so, you can enter ‘brt’ in your return, but we encourage you to provide an actual value if this is available. Substances with measured quantities are expected to include:

• arsenic (As)
• benzene (can also be calculated) – use effluent quality measured data or ‘brt’ if bio treated
• cadmium (Cd)
• chromium (Cm)
• copper (Cu)
• lead (Pb)
• mercury (Hg)
• nickel (Ni)
• nitrogen (N) – use 14/17 since ammoniacal
• phenols – use 72/94 for ‘as C’
• toluene (can also be calculated)
• total organic carbon (TOC)
• xylene (can also be calculated)
• zinc (Zn)

Emissions from process water should be calculated dependent on volume. Factors are derived from American Petroleum Institute (API) guidance. Substances are expected to be:

• benzene – use 6 µg/litre process water
• toluene – use 1 µg/litre process water
• xylene – use 0.25 µg/litre process water

Substances listed in the pollution inventory but not mentioned here are considered to not apply to refinery emissions. You can record them as ‘n/a’ in the pollution inventory return, unless site specific information is available.

It is your responsibility to check that there are no other pollutants emitted from your process that must be reported.