Deriving Waste Incineration BATAELS Under 21st Century Legislation

The second in a series of WMW articles on the Waste Incineration Best Available Techniques Reference Document tackles the specific features of the sector when it comes to deriving new BATAELs during the review of the WI-BREF and specifically its BAT-Conclusions. by Hubert de Chefdebien and Guillaume Perron-Piché Because of the heterogeneous composition of waste, the generally very low emissions from Waste Incineration plants (WI-plant) can temporarily peak. In addition, for Waste Incineration only, compliance must be assessed in Other Than Normal Operating Conditions (OTNOC). The challenge is integrating these aspects in the new Best Available Techniques Associated Emission Levels (BATAELs) which will be translated into legally-binding Emission Limit Values (ELVs) in future permits of plants, according to the new rules set up by the Industrial Emission Directive (IED). What's so special about waste to energy? Given the heterogeneous nature of waste, incineration plants are monitored for a larger array of substances than any other sector. However, it is important to note that Waste Incineration is not the only sector to emit these pollutants, but in most cases it is the only one to have strict limits on them. (See Table 1) Further, high performance flue gas cleaning systems are integrated in all European WI-plants, enabling the sector to comply with lower Emission Limit Values (ELVs) than any other combustion industry. Emissions are so low that they are close to the detection limit of the measurement instruments: in terms of environment and health protection as a whole, emissions from WI-plants do not represent a concern. According to the IED's Annex VI, Part 3, point 1.2, compliance with ½-hourly Emission Limit Values for Waste Incineration can be assessed in two ways: either 97% of the ½-hourly average values over the year do not exceed any of the ELVs set out in column B of the table given under point 1.2 none of the ½-hourly average values exceeds any of the ELVs set out in column A of the same table. Table 1: Daily ELVs to air according to IED for incineration of waste, co-incineration and LCP. For LCP, the values given in ref. to 6% O2 are corrected in brackets and in italics to 11% O2 as for incineration. Waste incineration must monitor more substance than co-incineration and LCP and meet lower ELVs The so-called 'Column B' conditions, which are applied in a few European Member States (MSs), are more stringent than the conditions applied for short-term ELVs for Large Combustion Plants (LCPs). The latter can discard up to 5% of the measured values (instead of 3% for WI-plants) and the short-term ELVs are set at 200% of the daily average ELVs (instead of being equal to them in the case of WI-plants). The so-called 'Column A' conditions, which are applied in the majority of MSs, refer to higher ELVs than 'Column B', but conditions are much more demanding since none of the values can be discarded. As shown e.g. in Figure 4, nearly all ½-hr average values achieved by WI-plants are noticeably low and only a few are higher. In passing, it can be noted that two different sets of BATAELs will be needed in the revised WI-BREF to match conditions of column A and column B of the IED, while there is only one set of ½-hourly average BATAELs in the current WI-BREF. This is an example of the impact of the new meaning of BATAELs introduced in the IED: according to previous legislation (IPPC Directive), BATAELs were typical average values achieved when implementing the Best Available Techniques whereas, according to the IED, they are capping values to be used when setting up the ELVs in the permits. The standard IED requirement for all industries is that compliance with ELVs is requested only in Normal Operating Conditions (NOC). Only for WI-plants, the IED sets up much stricter requirements: ELVs compliance is demanded in Effective Operating Time (EOT), which includes also some OTNOC (Other Than Normal Operating Conditions), such as start-up and shut-down periods when waste is combusted and some breakdown situations. Despite the fact that WI-plants are designed and operated so as to comply with the ELVs during the 8760 hours in a year, a few exceedances of the short-term average values are accepted by the IED reflecting the reality of industrial processes. WI-plants are only granted up to four uninterrupted hours of exceedance and a total of 60 hours per year. As a comparison, for LCPs the allowance is 24 hours uninterrupted and 120 hours per year. (See in Table 2) First conclusion In the light of the above considerations, a specific method to derive BATAELs from operational data is needed for WI-plants since the ones used for other sectors are not replicable. For example in the data collection developed for the review of the LCP BREF, high values, such as outliers, were discarded. For Waste Incineration plants this would imply a mistake, because as stated before, in conditions of column A, compliance must be continuously assessed at these facilities. Why must we refer to the raw data of numerous and identified plants? As outlined, the EU sets strict requirements on Waste Incineration. Table 2: Summary of the main differences in ELVs and the assessment of compliance with the ELVs for incineration of waste, co-incineration and LCP according to the IED. The rules for incineration are the most stringent However, it leaves leeway for competent permitting authorities (MSs and sometimes local authorities) when it comes to details. In addition to the already mentioned 'Column A' and 'Column B' conditions, the definitions of EOT, OTNOC and of the 60 hours allowance are not uniform within the EU and in some cases there is no definition at all. In the same way practices in respect of the treatment of data (instrument calibration, value taken when above the range of the instrument or below the detection limit, confidence interval etc.) can vary significantly between MSs. Representativeness of the data As explained above, under the IED the BATAELs are now ceiling values for the ELVs to be set up in the permits. For this reason, we shall have to consider the maximum values which were obtained by the “well performing” techniques candidate to the BAT status (and not the minimum or average values). Two questions then arise: Which maximum value is to be considered? (The maximum maximorum or the 2nd maximum, the 10th...?) How to be sure that the maximum chosen among the data of a single reference year is representative of the maxima observed on the same plant over many years? Figure 1: Set of data showing the ½-hourly averages of HCl for an individual line over one year. For this line, the candidate BATAEL values are respectively 40 mg/Nm3 and 17 mg/Nm3 (upper and lower ends of the range in column A conditions, 1 hour and 17 hours of exceedance in dark and pale blue) and 6 mg/Nm3 (in column B conditions, 3% of the data being discarded in purple). Note that all these values are below the current ELVs given in the IED which are set up at 60 mg/Nm3 in conditions A and 10 mg/Nm3 in condition B Second set of conclusions Due to the compliance assessment rules being different for the various plants, it would be difficult to choose in advance a handful of supposedly “well-performing” plants. In order to receive operational data that can be analysed and used in a standardised manner by the members of the WI BREF Technical Working Group (TWG) they have to be delivered as raw data, with only primary corrections made. And all the average values of the reference year must be provided (i.e. for each continuously monitored substance, 17520 ½- hourly and 365 daily average values). It will be difficult to collect data over many years from the same plants to secure a significant set of data. Therefore the correctness of the results can be achieved by maximising the sample of the collection, i.e. the number of plants participating. Figure 2: (Theoretical) distribution of the BATAEL candidate values provided by each 'well performing' line for each continuously monitored substance and corresponding to each ELV requested by the IED. As commonly accepted in statistics, our draft proposal is that the retained BATAEL value is the 95th percentile of all the candidate values To ensure their representativeness, data for substances subject to periodic measurements must be collected over at least a period of 5 years. And in order to give the opportunity to the TWG to crosscheck the data, neither anonymous nor aggregated data can be accepted. Figure 3: The BATAEL range resulting from the assessment method can be either below, overlapping or above the IED ELV. In the 1st case, the proposed BATAEL is the same as the resulting one; in the 2nd case, the proposed BATAEL is limited on top to the IED ELV; in the 3rd case, the proposed BATAEL is the IED ELV A race to the lowest? One may be tempted to say that the best plant is the one with the lowest emissions and declare these emissions to be the benchmark. However the need to account for other environmental criteria associated with specific emissions performance – an integrated approach – is a cornerstone of the IED which defines 'best' as the “most effective in achieving a high general level of protection of the environment as a whole”. It has to be considered that lowering air emissions usually has a negative influence on some of the other criteria to be taken into account according to the IED (Annex III). Some examples of said impacts are: potential increase in the emission of other substances, increased resources consumption, higher production of residues, production of liquid effluent, etc. The question of costs must not be overlooked either. Taken that, as said in the first article, the impact of the emissions of WI-plants is negligible, what is the benefit to spend resources and money to reduce again the emissions when this money could be spent elsewhere to improve the environment as a whole? Therefore, is a “well-performing plant” a plant with minimised emissions to air or one with overall optimisation of integrated criteria, e.g. minimised consumption of scarce resources such as water? An integrated approach allows local conditions to be considered: MSs with no lack of water may propose plants with high water consumption because they show excellent emission results, while others may choose one with little water consumption and zero water discharge. Deriving Air BATAELS from plants implementing BATS Based on the information given above a draft proposal could be the following method: 1) Among the sets of data provided by the plant operators, identify the lines with a well-performing flue gas cleaning system Our proposal is that “well-performing” Waste Incineration flue gas cleaning systems are the ones implementing BATs and having not exceeded during the reference year the ½-hourly Column A ELVs in EOT for more than 60 hours for all continuously monitored pollutants as provided for in the IED. 2) Setting up the ½-hour average 'Column A' BATAELs Our proposal is to use again the 60 hours counter from the IED as a tool to set up the ½-hourly average 'Column A' BATAELs. The principle is to calculate a set of BATAEL candidate values for each WI line which provided data and was recognised as 'well performing'. Then among these candidate values to set up the BATAEL values by a classical statistical approach which is to retain the 95th percentile among the candidates. The lower end of the ½-hr average 'Column A' BATAELs range Given that the IED requires 7 substances to be continuously monitored in WI-plants and assuming that the 60 hours are equally shared between them, each one is allowed a little bit more than 8 hours of exceedance (7 x 8 = 56). Our proposal is therefore to consider the 17th maximum value (16 x ½-hours + 1) of each line. This being the minimum acceptable limit corresponding to the set of data when using the 60 hours allowance. (See Figure 1) The upper end of the ½-hr average 'Column A' BATAELs range In principle, for each substance, the upper end of the BATAEL candidate value of each 'well-performing' line should be the maximum ½-hour average of the year. However, in order not to get a bias from a few possibly abnormal values, we propose to discard the 2 first maxima of the year and to take only into account the 3rd maximum (See as well Figure 1). Then, in order not to get a bias from a few 'well performing' lines with abnormal values, the proposal is to take among these candidate values the 95th percentile of all these 17th and 3rd maximum values as the lower and upper end of the BATAEL range (See Figure 2). 3) Setting up the ½-hour average 'Column B' and daily average BATAELs The proposal for ½-hour average 'Column B' and daily average BATAELs is also to take the 95th percentile of the candidate values of all 'well performing' lines corresponding to 'Column B' conditions (i.e. the maximum value remaining after having discarded 3% of the data) (See Figure 1). And for the daily average BATAELs, the proposal is as well to take the 95th percentile of all the maximum daily averages of the 'well performing' lines. In summary: 1) For each of these lines, identify the candidate values, i.e.: a. For each substance which is continuously monitored: - The 3rd and the 17th maximum ½-hourly values of the year. - The 97th percentile ½-hourly value of the year. - The maximum daily value over the year. b. For each substance which is periodically monitored, the maximum value over 5 years. 2) Derive BATAELs a. For each continuously monitored substance, set the BATAEL values for: - ½-hourly BATAELs range for Column A ELV (100% of values): at the 95th percentile of the candidate values (i.e. all 3rd and 17th maximum ½-hourly values of validated lines) - ½-hourly BATAELs for Column B ELV (97% of values): at the 95th percentile of all the candidate values (i.e. all 97% ½-hourly values of validated lines) - Daily BATAEL: at the 95th percentile of all the candidates (i.e. the maximum daily values of the validated lines). b. For each periodically monitored substance: at the 95th percentile of all maximum values of the validated lines. Testing the method In order to test this method, a collaboration between various organisations under the leadership of CEWEP and ESWET enabled the gathering of operational data from 32 WI lines of various capacities and Flue Gas Cleaning types in six European countries. The result is that the daily BATAELs are mostly in line with the current Emission Limit Values. Derived ½-hourly BATAELs, calculated based on Column B (97%) would in general be slightly below the current ELVs. And the data shows that the ½-hourly BATAELs based on Column A (100%) would in general be at the current ELVs. (See the 3 typical cases on Figure 3) Conclusion “When the facts change, I change my conclusions!” This famous quote is well-suited to the IED and the need to update the BAT Conclusions and their BATAELs. What previously (under IPPC) was an indicative figure based on aggregated averages will now be a capping limit value since the IED states that the ELVs to set up in the permits must not exceed the future BATAELs. In normal operation a WI-plant records very low emissions for most of the time intervals, but some recorded values are higher (peaks) because of the changing pollutant load in the incoming waste or for operational reasons. The method presented here, and on which we still work, aims at deriving BATAELS which actually reflect the fluctuating emission levels recorded when incinerating waste in a 'well performing' plant in the Effective Operating Time (Normal Operating Conditions plus some Other Than Normal Operating Conditions) when ELVs must be complied with. The first step is to clearly define this Effective Operating Time. The second one is to collect a statistically significant amount of operation data covering various local conditions on this full period of time. The 3rd one is to apply a non-biased statistical calculation to obtain BATAEL values actually corresponding to the performances of the 'well operating plants' and achievable in the reality of industrial processes. It is not surprising that the updated BATAEL values will be higher than the current ones - based on the former IPPC Directive - which were typical average ranges of recorded emission levels. This is not because Waste Incineration is backtracking on environmental protection, but simply because the meaning of the values has changed. Nevertheless, the new BATAELs will remain below or equal to the IED ELVs and the actual emissions from the WI-plants will remain at scrupulously low level. Hubert de Chefdebien is the chairman of ESWET's Technical Committee and a member of various European Commission's BREFs Technical Working Group including the one on the revision of the WI-BREF; Guillaume Perron-Piché is the policy officer at ESWET. Glossary BAT: Best Available Technique BATAEL: Best Available Techniques Associated Emission Level BREF: BAT Reference Document EIPPCB: European IPPC Bureau (in Seville) ELV: Emission Limit Value EOT: Effective Operating Time EU: European Union IED: Industrial Emissions Directive (2010/75/EU) IPPC: Integrated Pollution Prevention and Control Directive (96/61/EC) LCP: Large Combustion Plants MSs: Member States NGO: Non-Governmental Organisation NOC: Normal Operating Conditions OTNOC: Other Than Normal Operating Conditions TWG: Technical Working Group (set up by the EIPPCB) WI-BREF: Waste Incineration BREF WID: Waste Incineration Directive (2000/76/EC) WI-plant: Waste Incineration plant