The Bacharach Method: A Low-Cost Tool for Small-Scale Combustion Units’ Flue Gas Quality Control

Abstract

Although current EU regulations, such as EU Directive 2015/1189 on the eco-design of solid fuel boilers and Directive 2015/1188, in accordance with the Machinery Directive 2006/42/EC, require manufacturers to meet specific emission requirements for CE marking, the routine and regular onsite testing of household heating appliances is still not mandatory in many EU countries. This research endeavour addressed this gap by evaluating the effectiveness of the Bacharach method as a rapid and cost-effective tool for assessing flue gas quality, particularly in terms of particulate matter mass concentration. This study also compared the results of the Bacharach method with those obtained from two commercially available portable analysers. The research outcomes demonstrate that the Bacharach method, in combination with an innovative evaluation approach, offers a viable solution, enabling the swift and economical assessment of flue gas quality with the primary objective of determining the boiler class according to the limits specified by standard EN 303-5 under operating conditions. The modified Bacharach method for measuring TSP in solid fuel-fired boilers provides qualitatively similar results to the commercially used SM500 and STM225 instruments. The modified Bacharach methodology was primarily developed for comparison to the boiler class 3 limit (i.e., 125 and 150 mg/m³). The study revealed that the modified Bacharach method, when applied to biomass-based boilers, exhibited higher accuracies in the case of classification into classes 3 and 4, whereas fossil fuel-based boilers demonstrated higher accuracy in the case of class 5 limits.

1. Introduction

Air pollution is a persistent and highly emphasized topic every year, especially during the winter season, due to worsened dispersion conditions and the operation of household heating combustion appliances [1]. The sources of pollution that contribute to poor air quality can be divided into three main categories: industry, transport, and local heating. For the first two types of sources, regular monitoring of the operating conditions, coupled with a flue gas analysis, is common and legislatively directed (in the EU), with the length of the period between monitoring varying according to several parameters. In the case of stationary combustion, for solid fuel combustion sources with a thermal input of up to 500 kW (intended primarily for household heating; hereafter referred to as combustion units), no periodic checks on the composition of flue gases are mandated in many European countries. The only mandatory control of the source before its release on the market is a certification process carried out under ideal conditions by an accredited testing laboratory [2]. However, in real operation, several key factors could differ from the testing laboratory situation, such as an alternative fuel usage (usually characterised by different proximate and ultimate analysis results then the standardised ones, often connected to the different characteristic dimensions, which results in a different management of the combustion process in terms of the contact between combustion air and fuel, etc.) [3,4], different chimney draughts (draughts continuously change depending on the flue gas temperature, ambient conditions, and control elements’ setup, and define the fuel/air ratio, which is a crucial characteristic of the combustion process) [5], different control unit setups (these define the fuel/air ratio, which is a crucial characteristic of the combustion process), the operation using a different heat energy output (usually, combustion stationary sources are only certified for operation at a nominal power—100%, while some combustion sources, such as boilers with an automatic fuel delivery, have to be certified for a so-called minimum heat energy output representing 30% of the nominal thermal output) [6,7], and the technical state of the combustion unit (linked with fuel/air ratio by, e.g., the fake air suction and the way in which combustion air is introduced into the combustion chamber by clogging the combustion unit parts) [2], which could significantly change the flue gas composition.

For example, the use of biomass-based residues from agricultural and forestry activities as an alternative fuel for small-scale combustion stationary sources is clearly more environmentally acceptable than open burning (e.g., during wildfires), but this activity must comply with the legislative framework [8]. This approach aligns with studies highlighting the potential of biomass and waste-derived fuels to enhance heating value and reduce air pollutants [9,10,11].

The legally mandated periodic onsite inspections of combustion sources, for example, in the Czech Republic (maximum interval of 3 years), do not include an analysis of the performance of flue gas, only a visual assessment of the condition of the combustion source and related elements [12]. The situation is different in, for example, Germany (Air Protection Act; interval, depending on performance, of 2–4 years) [13], Austria (Air Act; interval, depending on performance, of 2–3 years), [14] or the UK (regional regulations; interval of 1–2 years) [15], where periodic control also includes a flue gas analysis. The results reveal an inadequate flue gas composition, with particulate matter being the most observed pollutant. This may lead to the combustion equipment being decommissioned. National and sub-national laws and regulations are generally supplemented by other legislative documents covering details such as the methodology of sampling procedures.

The mass concentration of particulate matter in flue gas or other process gases (e.g., gasification or pyrolysis [16,17,18]) could have a positive impact on soil properties [19]; however, it is also one of the most important components from the perspective of the negative impact on human health [20,21]. The mass concentration in the flue gas during the certification process of combustion units is determined by a gravimetric method, as stipulated by the EN 303-5 standard, which designates this method as appropriate for this measurement. This process involves comparing the mass of the filter before and after exposure relative to the measured volume of flue gas passing through the filter. The EN 303-5 standard also provides a detailed scheme and methodology (measuring apparatus) for accurately measuring flue gas dust in solid fuel heating boilers reaching up to 500 kW. During the on-site testing of domestic combustion units (periodical testing, for example, in the above-mentioned countries), it is a common practice to use simpler (case) instruments, which often allow for the analysis of several flue gas components simultaneously, for example, SM 500 (Wöhler; Bad Wünnenberg, Germany) [22,23] or STM 225 (Afriso; Güglingen, Germany) [24]. The first mentioned device determines the mass concentration of particulate matter based on the direct semi-gravimetric method, while the second one determines the mass concentration of particulate matter using the optical method. The complexity of the instruments and their low volume and weight requirements, combined with the type of sensors used and the requirement for instantaneous evaluation, significantly affect their acquisition and service costs. Other disadvantages of these instruments include their limited resistance to significantly polluted flue gases (which can even cause irreversible damage to the measuring sensors) and limited measurement accuracy.

The complete opposite of portable instruments in terms of the applicability of analysis and the equipment requirements is the so-called Ringelmann method, which compares the darkness of the flue gases at the chimney outlet with the corresponding degree of the Ringelmann darkness scale using human vision [25]. Ringelman’s method can only be performed during daylight hours, there must be a prescribed distance between the chimney and the observer, and the readings must be taken against a clear sky, without clouds. These limitations often make measurement impossible, especially during the winter (heating) season. This method is listed in the Communication of the Czech Ministry of the Environment as a method for controlling the operation of the combustion units. The accuracy of the method in its original form is limited by the introduction of the human factor in evaluation and is not very robust because of the need for specified lighting conditions and a specified position of the observer. However, the cost of the acquiring scale is negligible [26].

Between the portable analysers and the Ringelmann method stands the Bacharach method. This was developed by the American company Bacharach, Inc. (New Kensington, PA, USA), which has been a leader in combustion and environmental measurement since the early 20th century. The measurement, according to the methodology, is taken by sucking the flue gas right from the flue gas duct, using a manual pump with a known volume, and through the filter; then, the filter shade is evaluated.

The use of the Bacharach method in the context of liquid fuel burning combustion units was mentioned in a study by Leary et al. [27], where a correlation between the obtained soot number and the mass concentration of organic gaseous compounds was observed but not demonstrated. McDow et al. [28] described the relationship between the mass concentration of polycyclic aromatic hydrocarbons in the flue gas from a liquid fuel burning combustion unit as linear, and the straight line characterized the trend very reliably (R² = 0.96). Jiménez et al. [25] described a strong correlation between soot mass concentration and soot number when burning gaseous and liquid fuels in an experimentally designed source. The Bacharach method was also used to determine differences in the flue gas of ten liquid fuels in the study of Lee [29], where the experiments were performed by using a small-scale boiler for residential heating. The relationship between the flame temperature and carbonaceous particulate emissions during liquid fuel combustion was observed in a study by Belyea and Holland [30], where one of the methods used was the Bacharach method. The Bacharach method was also used in the field of solid fuel combustion in small-scale appliances, such as in the study of Yoshida et al. [31], where the Bacharach method was used for the determination of differences in the flue gas composition during the combustion of standard (wood pellets) and alternative fuel (pellets from torrefied Japanese cedar) in a pellet stove.

The results of these studies, obtained by analysing the flue gas produced during the combustion of liquid, gaseous, and even solid fuels, confirmed the suitability of this method for the fast screening of flue gas quality in terms of particulate matter mass concentration.

This present research aims to modify and evaluate the suitability of the Bacharach method for the determination of the permissible darkness of flue gas as an indicative, quick, and cheap alternative to portable instruments, with the possibility of classifying the source according to the EN 303-5 standard, with a particular focus on the class 3 borders that figures in some national laws. The presented Bacharach method focuses on different field of utilisation from TSP measurements using the gravimetric method, which will always be the most accurate but, simultaneously, the most demanding to perform.

The Czech Air Protection Law is a critical division of equipment into operable and inoperable [12,32]. Another aim was to compare the results obtained by the Bacharach method with the portable analyser’s results related to the reference obtained by the gravimetric method. A significant novelty lies in the use of this Bacharach method, which was designed to measure liquid fuel combustion appliances and has been upgraded by an innovative evaluation process that enables the measurement of solid fuel boilers with high accuracy.

2. Materials and Methods

The Bacharach method for the assessment of the darkness of flue gas shares fundamental principles with the Ringelmann method. However, there is a crucial distinction: instead of measuring flue gas darkness at the chimney outlet, the Bacharach method evaluates it directly within the flue gas duct downstream of the emission source using the in-situ method. This assessment was performed by examining the darkness of the filter area through which a known volume of flue gas sample had passed. This approach eliminated inaccuracies caused by adverse weather conditions. Both methods use the basic version, based on humans’ assignment of smoke darkness according to a comparison with a reference scale, which may result in increased measurement uncertainties.

The Bacharach method is, according to the recently renewed standard DIN 51402-1:2020-09, the standard for the visual and photometric determination of the smoke darkness of oil burning systems. A unified manual pump with a known volume (0.163 dm³) and a filter holder was used for fixing the proper filter. The Bacharach set (MRU GmbH; Neckarsulm, Germany) also included a Bacharach scale consisting of ten coloured intermediate rings with shades of grey numbered from 0 to 9, representing the resulting value (hereafter referred to as the soot number), while soot number 0 represents pure white. In the case of oil-burning appliances, the standard sampling procedure consisted of ten full pump stretches (10 × 0.163 dm³ = 1.63 dm³) according to DIN 51402 [33]. In the case of combustion sources using liquid fuel, soot numbers 0 to 1 represented the standard combustion conditions, while soot numbers 2 and above represented poor combustion conditions [34]. Such a simple division cannot be applied in the case of solid fuel combustion due to the broader spectrum of combustion parameters that must be taken into consideration during the evaluation, and also due to the pollution contained in the flue gas originating from solid fuel combustion, which is several times higher than that from the combustion of liquid fuels [35,36].

The measurement procedure used in the following experiments partially followed the standard (original) method, except that instead of ten pump volumes, only one pump volume was sucked in. Seven flue gas samples were taken over 15 min (with 2 min intervals between the sampling). On each filter, two samples were taken immediately after each other, while only the second sample was considered. The point was to fill the Bacharach sampling line with the flue gas for the second measurement and to have (in the case of necessity) a backup shade spot (the first one) for cases of failure of the second sampling; these spots were never used during the measurements. During this period, sampling was also carried out using the reference device SM 96 (Wöhler, Germany), which may be used for the certification of combustion sources in an accredited testing laboratory. Particulate matter sampling was conducted on a cellulose filter placed in a heated probe in the immediate vicinity of the flue gas duct. The instrument was equipped with a cooling device, pump, and flow meter to determine the volume of the sample. The results obtained from the SM 96 device were evaluated ex-post under laboratory conditions using the gravimetric method.

During the same period, a 15 min sampling procedure was conducted using SM 500 (Wöhler; Bad Wünnenberg, Germany) and STM 225 (Afriso; Güglingen, Germany) simultaneously, enabling the comparison of the resulting data from each measuring device. The data were compared after manual recalculation of the resulting values with the reference oxygen volume fraction in the flue gas. The volume fraction of oxygen in the flue gas was measured using an AO2020 analyser (ABB; Heidelberg, Germany) via the paramagnetic method to avoid measurement uncertainties caused by the use of less accurate electrochemical sensors in portable analysers. In the portable SM 500 analyser, the raw value of the mass fraction of the particulate matter in the flue gas was taken into consideration.

Due to the nature of the instruments used and due to the specific character of the onsite measurement, including the later analysis of the fuel, the variable oxygen volume fraction in the flue gas, etc., the sampling was not performed in an isokinetic way. Sampling using the Bacharach equipment took approximately 1.5 s each spot. The position of the sampling probes for each device is shown in Figure 1.

2.1. Soot Number and Mass Fraction of Particulate Matter

The method used to evaluate the results in the form of grey spots (the places through which the flue gases were sucked) on the filters was designed to eliminate human factors and to allow the results to be compared with reference values for the mass concentration of particulate matter in the flue gas obtained by gravimetrical methodology. All seven exposed filters from each fifteen-minute measurement were placed in the designated positions (always in the same arrangement) and placed in a 240 × 240 × 240 mm photo studio equipped with two LED strips, which were mounted on the ceiling and emitted daylight (5700 K). The studio’s ceiling was equipped with an opening to capture photos from above; the photos were taken using an iPhone XR mobile phone (Apple, Cupertino, CA, USA), which was always placed in the same position. Photographs were taken in flash and greyscale mode. The filters are presented in their picture-taking positions in the light tent in Figure 2. An example photo taken by the resulting filters is shown in Figure 3.

The resulting photographs were transferred to a computer and opened in the Paint program, where an area of at least 30 × 30 pixels was selected from the exposed area of each filter. The chosen area was extracted, and its average brightness was determined using the same program (range 0–240; 0—black; 240—white). Brightness was not expressed in units such as candela per square meter (cd/m²), as in professionally calibrated monitors or in the technical specifications of imaging equipment, but was a dimensionless quantity representing the intensity of the colour at the pixel level. Brightness was assigned a soot number using a function (1) obtained during the calibration of the evaluation method.

Calibration was performed by placing the original Bacharach scale in the photo studio at four different positions (always rotated 90°). Four areas of at least 30 × 30 pixels were cut off from each position photo from each of the intermediate grayscale circles characterizing the soot number and were assigned an average brightness, as described above. Thus, sixteen brightness values were always assigned to one soot number value (see Table 1). The graph obtained from the light tent calibration, from which the recalculation of Equation (1) was retrieved, is presented in Figure 4.

$$SN=-0.0463B+9.9239$$

(1)

SN—soot number [-];

B—average value of the brightness of the 16 areas from original Bacharach grayscale. [-]

The soot number was determined by the abovementioned method for each individual filter (with each of the seven belonging to one 15 min sampling), both separately and then as the average of all seven filters.

The correlation between soot number and particulate matter mass concentration at the current O₂ volume fraction is shown in Figure 5. The calculated regression line from the pairwise comparisons provided a directive of about 37 (36.98). However, a directive of magnitude 20 (red line in the Figure 5) was chosen for the recalculation, as it corresponded better to the majority of the measured points, especially in the area with medium and higher soot numbers, which were significant for the following purposes. The changing multiplication factor is a step towards the improvement of the methodology. The weight of the measured values in the areas around the TSP mass concentration limits (125 and 150 mg/m³) was increased, roughly corresponding to soot numbers in the range of 3 to 7 with respect to the actual O₂ concentration in the flue gas.

The soot number was converted to the particulate matter mass concentration using Equation (2): The equation was determined considering the relation between soot number and particulate matter mass concentration, determined according to the reference gravimetrical method.

$${\rho }_{B\_O2M}=20·SN$$

(2)

ρ_{B_O2M}—mass concentration of particulate matter in the flue gas determined by the Bacharach method at the measured volume fraction of oxygen [mg·m⁻³].

Experimentally determined equations for the same purpose, such as the one in the study of Jiménez et al. [25], were also taken into consideration; however, they had a significantly lower accuracy (in the final evaluation) mainly due to their scope, which is preferable for liquid fuel combustion.

The mass concentration of particulate matter according to the Bacharach method at the measured oxygen volume fraction was consequently recalculated to the reference oxygen volume fraction in the flue gas according to Equation (3).

$${\rho }_B={\rho }_{B\_O2M}·\frac{21-{\phi }_{O2\_ref}}{21-{\phi }_{O2\_m}}$$

(3)

ρ_B—mass concentration of particulate matter in the flue gas determined by the Bacharach method at the measured volume fraction of oxygen [mg·m⁻³].

φ_{O2_ref}—reference volume fraction of oxygen in the flue gas for the observed category of combustion equipment according to the standard EN 303-5 [32]; φ_{O2_ref} = 10%.

φ_{O2_m}—measured volume fraction of oxygen in the flue gas [%].

The equivalent of Equation (3) was also used for the recalculation of the results obtained by the reference gravimetric method, while in both cases, φ_{O2_m} acquired the same values—the average of the 15 min sampling.

2.2. Set of Samples

In total, 121 small-scale combustion units for solid fuel combustion for all basic types of construction (over-fire, down-draft, gasification, and automatic) were tested. In this group, units using a wide range of fuels, including both biogenic and fossil fuels (firewood, lignite, charcoal, wood pellets, and wood briquettes), were included. The range of basic parameters of the fuels used for the experiments is presented in Table A1. Combustion units were almost exclusively operated by their operators (owners), not by trained technicians during the certification process. The owners/operators of the boilers followed the rules in the manual to the best of their knowledge and conscience regarding the procedure for adding used fuel, etc. The heat energy output was not set as the nominal one during the experiments, but depended on the outdoor conditions (the actual heat energy loss of the building) and the kind of installation (the presence of an accumulation tank, etc.). The abovementioned details represent the differences between the laboratory experiments during the certification and the onsite measurements.

This ensured the procurement of a wide range of results representing the factual situation in the field. Overall, 355 separate 15 min intervals were observed (usually three samples from each device).

Compliance and non-compliance with the limits for each class (3, 4, and 5) according to the EN 303-5 standard were observed. The limit values for each combustion unit and the used fuel are presented in Table 2.

3. Results and Discussions

A graph representing the relationship between the mass concentration of particulate matter obtained using the Bacharach method at the reference volume fraction of oxygen and the mass concentration of particulate matter obtained using the gravimetric method at the reference volume fraction of oxygen is shown in Figure 6. It is obvious that there is a dependency between the two values; however, their dispersion suggests a non-negligible degree of measurement inaccuracy. However, precise results were not expected to be obtained from this simple and affordable method. The most important output was the successful classification of sources into individual classes according to EN 303-5. The approach of classifying small-scale boilers into categories from different perspectives was previously used in several studies, such as the studies by Rabbat et al. [37], Verma et al. [38] or Lasek et al. [39]. The overall classification of the success rate of the Bacharach method is presented in Table 3, and additional details (regarding the boiler construction type or fuel used) are presented in Table A2, Table A3 and Table A4. The tables were divided into two primary columns (“in compliance” and “not in compliance”, using the Bacharach method) with two sub-columns (“correctly classified” and “incorrectly classified”). The “in compliance” column with the sub-column “correctly classified” represents the number of experiments where the particulate matter of the flue gas was in compliance with the limit value. “Incorrectly classified” under “in compliance” represents the numerous experiments which the Bacharach method was identified to be “in compliance” but the results from the gravimetrical method showed that the results were not in accordance with the standard limit. The column “not in compliance” can be viewed in the same way.

The total success rate for the determination of the combustion equipment class was slightly below 80% for Classes 3, 4, and 5. Due to the character of the tested combustion units and the attained mass concentrations of pollutants, the highest partial success rate in the case of the Class 3 limits was considered as “correctly classified” units “in compliance” with the limit values, while in the case of the limits of Classes 4 and 5, the highest success rates were accomplished in the category of “correctly classified” units as “not in compliance” with the limits. Taking a closer look at the success rate, divided into categories according to the fuel or boiler construction type, it is obvious that the success rate was the highest in the case of biomass fuel combustion for Classes 3 and 4, at almost 83 and 80%, respectively, whereas in Class 5, the success rate was the highest for fossil fuels, at ranges of higher than 85%. The maximum success rate for fossil fuel combustion and the Class 5 limit value was 90.2% according to the combustion unit type, obtained when the fuel type was automatically fed boilers. This could be due to the poor results obtained in this particular category of combustion equipment when no boiler in this category met the limits of Class 5 under the observed real conditions. In contrast, the categories with the lowest success rates were automatically fed boilers for fossil fuel combustion in the case of Class 3, while the success rate was only 64.7%. A total of 25.5% of incorrectly classified results were labelled as compliant with the Class 3 limit.

When evaluating the appropriateness of implementing this method and deciding whether it fulfilled the limits set by the standard (and, in some countries, by the law), the most sensitive category was “incorrectly classified” as “not in compliance”. The owners of sources categorised as such would be directly harmed. This category was represented by values 5.1, 2.0%, and 1.4% for Classes 3, 4, and 5, respectively, indicating a very low rate of error, which decreased with increasing demands. The category of automatic units that combusted fossil fuels had the highest value for “not in compliance” and “incorrectly classified” categories, reaching 9.8% (for Class 3). Meanwhile, the entire category of boilers combusted with fossil fuels reached 7.0%. When applying this methodology in practice, the category of sources with the highest value should be observed as closely as possible, implementing solutions such as increasing the number of samples or incorporating a higher level of measurement uncertainty.

It was also possible to examine this issue separately, where only sources identified as not in compliance were assessed. The success rates of correctly and incorrectly classified noncompliance units according to the Bacharach method for Classes 3, 4, and 5 are shown in Table A5. Similarly, with increasing boiler class (higher requirements), the number of incorrectly classified boilers decreased (17.1, 3.7, and 2.3% for Classes 3, 4, and 5, respectively). The highest absolute value was reached for the category “Class 3 Biomass Automatic” with a value of 50%; however, this anomaly was probably due to the low number of sources that fit into this category. The second-riskiest category from this perspective for Class 3 was “Fossil Automatic”, which reached a value of 23.5%. The value reached for this category was also the highest for Class 4 (20%); however, three categories were entirely correctly classified. The same categories were classified correctly in the case of Class 5, while the highest error was reached in the case of the combustion unit category “Biomass Automatic”, even though the error was only 5.9%.

3.1. Bacharach Method vs. Portable Instruments

The results obtained from portable analysers SM 500 and STM 225 were compared with those obtained using the Bacharach method. For SM 500, the raw values measured by the instrument were used for evaluation, as displayed in Figure 7. SM 500 corrected the raw measured values downwards by 43.68% in the standard evaluation (Figure 7 includes corrected and uncorrected values). As the basic method of comparison, the value of the standard deviation, expressed as a percentage relative to the reference value of the measured mass concentration of particulate matter determined by the SM 96 instrument using the gravimetric method, was chosen.

The total dispersion of values, as well as the interquartile range, was the smallest in the case of the Bacharach method, followed by STM 225. The interquartile obtained by SM 500 was widely dispersed in the range of 6–220%, which formed the most diffuse interquartile of those compared. Considering that the median (59%) and average (171%) values were significantly above “0”, the SM 500 analyser may be significantly higher than the real state. This was in accordance with the findings of Oischinger et al. [23], who found that, in the raw flue gas, the SM 500 analyser obtained higher values compared to the other tested analysers. Contemplating the corrected values of the analyser, the dispersion range would be significantly smaller than in the case of uncorrected values (−39 ÷ 81%) but remains the principal one in comparison with the rest of the analysers. In this case of data evaluation, the median value reached −9%, which represented a very close value to the zero line; however, the average value of 53% represented the furthest value from the zero-line compared to the average values of the other instruments.

The median (−22, −33%) and average (8, −22%) values of the STM 225 device and the Bacharach method were very close to zero, while the boxplot layouts commonly meant that, in general, the resulting values were frequently lower than the reality. In the case of the STM 225 device, there are probably some corrections that could be made to adjust the resulting values; however, these are not known to the operator. Therefore, comparisons of uncorrected and corrected values are not available.

The findings mentioned earlier are further corroborated by Figure 8, which illustrates the prevalence of measured values from SM 500, STM 225, and the Bacharach method in relation to the reference method. The points on the full compliance line indicate complete alignment between the measuring devices (SM 500, STM 225, or the Bacharach method) and the reference device (SM 96). As the perpendicular distance between individual points on this line increased, the inaccuracy of the measuring device also escalated.

Specific results are as follows:

• SM 500 results predominantly occurred above the curve and were associated with measurements of higher mass concentrations of particulate matter in the flue gas.
• Bacharach method results exhibited dispersion both above and below the line, particularly in the range of 25–125 mg·m³N (measured by the reference method). This indicates a lower particulate matter mass concentration compared to the reference.
• STM 225 measurements were widely scattered both below and above the full compliance line, with most points falling below that line.

A correlation table presented in Table 4 provides an overview of the relationships between different pairs of measuring devices for measuring the mass concentration of particulate matter. This table provides a convenient overview of the relationships between different instruments when measuring a specified phenomenon, which can be utilised for analysing and comparing their performance and reliability. The values in the table represent correlation coefficients that measure the strength and direction of the relationship between the two instruments.

A strong positive correlation of 0.807 was observed between SM 96 and SM 500, indicating that if the values measured by the two instruments increases, it is likely that both are increasing. A similarly strong positive correlation was observed between the SM96 and Bacharach instruments, with a value of 0.748.

A positive correlation was also observed between the SM 96 and STM 225 instruments, but it was slightly weaker than that of the previous pair, with a value of 0.671. This suggests a degree of correlation between these instruments, although it is not as pronounced as for the SM 96, SM 500 and Bacharach instruments.

Similarly, a moderate positive correlation was found between SM 500 and STM 225 (0.556) and between SM 500 and Bacharach (0.517). The weakest correlation was observed between the STM 225 and the Bacharach instruments, with a value of 0.348. This suggests a certain degree of association between these instruments, but is less pronounced than for the other pairs of instruments.

The on-site testing of 121 combustion devices also revealed some advantages and disadvantages of the selected devices. The Bacharach method and STM 225 analyser required supplementation with a second analyser for the determination of the volume fraction of oxygen in the flue gas, and the resulting value of the mass concentration of pollutants needed to be recalculated. The recalculation of the results is the only step desired in an evaluation of the STM 225 results; however, in the case of the Bacharach method, a post-processing including an evaluation of soot number and recalculation of this number to the particulate matter mass concentration was necessary. This process took approximately 10 min, which is comparable to the time required to prepare the SM 500 and STM 225 analysers (unpacking, preparation, filter placement, stabilization, cleaning, and packing). The most significant advantage of the SM 500 and STM 225 analysers was the real-time display of the mass concentration of flue gas or the increment in the log weight on the filter. The long-term continuous measurements for fifteen minutes can also be considered an advantage of the portable analysers, compared to seven separate samplings lasting only a few seconds with an interval of two minutes. Predominantly, for automatic boilers, the period over which the fuel was batched could indicate a fluctuation in the mass concentration of PM (mainly caused by a sudden rearrangement of the fuel in the burner, releasing the ash or fuel particulates into the flue gas flow) in the flue gas appeared to be important, but the measured experimental data could not confirm this undesirable effect [40].

The foremost disadvantage of the SM 500 and STM 225 devices was their sensitivity to the highly polluted flue gas that originated during the operation of older devices or even newer ones which were improperly operated (incorrect fuel–air ratio setting, improper fuel usage, incorrect placement of fuel in the combustion chamber–vault formation, etc.) or not well maintained. The highly polluted flue gas emissions caused early termination of the sampling (without the recorded result) and, in the long-term perspective, highly polluted flue gas had a destructive influence on the sensors inside the analysers. This outcome was in accordance with a previous study of Ryšavý et al. [3] focusing on alternative biomass-based fuel combustion, where the SM 500 analyser was used (it was also necessary to terminate the experiments early due to the higher mass concentration of particulate matter).

3.2. Economic Evaluation

The main advantage, besides the simplicity of the measurement, is its cost; therefore, an economic evaluation is appropriate.

An integral part of particulate matter measurement in solid fuel combustion sources is the measurement of O₂ in the range of 0–21%_vol. for a recalculation of the resulting values to the reference volume fraction (10%_vol. for boilers; 13%_vol. for stoves), and CO in the minimum range of 0–10,000 ppm (minimum 0–0.1%_vol.) as an indicator of the combustion quality. For the economic evaluation, sets that measure all these variables were compared.

The Bacharach set is available at a price of approximately 200 EUR (excluding VAT); however, an additional device for measuring O₂ and CO must be used in this case. The Wöhler A550 seems to be a suitable device for measuring O₂ (0–21%_vol.) and CO (0–100,000 ppm), at a price of around 3800. The total price of the kit is approximately 4000 EUR. Consumable costs are in the order of tens of cents per measurement.

SM 500, which is equipped with particulate matter and O₂ (0–21%_vol.) and CO (0–100,000 ppm) measurements, is available at a price of around 12,000 EUR (excluding VAT). For each measurement, special filter cartridges are needed, at a cost of approximately 250 EIR for 10 pieces, or the recycling of 10 filter cartridges for 90 EUR.

The AFRISO STM 225 used for particulate matter measurement must be supplemented with the Multilyzer STx for O₂ (0–21%_vol.) and CO (0–10,000 ppm) measurements and can be purchased at a price of approximately 14,500 EUR (excl. VAT).

Comparing the above purchase costs, the Bacharach kit with an O₂ and CO analyser is about 8000 EUR cheaper than the Wöhler SM 500 and about 10,500 EUR cheaper than the STM 225 with Multilyzer STx.

4. Conclusions

This research investigation has highlighted the potential of the Bacharach method as a practical tool for assessing flue gas quality, particularly in the field of household heating combustion units. Despite its simplicity and low cost, this method demonstrated a promising success rate in classifying combustion units according to the limit values of each class mentioned in the related standard. Furthermore, comparisons with the established analyser SM 96 revealed a reasonable correlation.

The range of the standard deviation of the measured values from the reference values indicates a high degree of usability for the Bacharach method, comparable to selected portable analysers. Notably, the method proved highly resistant to highly polluted flue gas without the risk of damaging sensitive and expensive sensors, unlike portable analysers. This robustness suggests that the Bacharach method could serve as a premeasuring tool to protect sensitive sensors in portable analysers from highly polluted flue gases. Therefore, the application potential of the Bacharach method could be also extended to its use as a pre-measuring method before the portable analyser is used to protect sensitive sensors from highly polluted flue gas.

The overall success rate for determining combustion equipment classes using the Bacharach method was slightly below 80% for Classes 3, 4, and 5. The highest success rates were achieved for biomass fuel combustion in Classes 3 and 4 (almost 83% and 80%, respectively) and for fossil fuel combustion in Class 5 (above 85%). The highest observed success rate was 90.2% for boilers automatically fed using fossil fuels in Class 5.

Economically, the Bacharach method is significantly cheaper than other portable analysers like the SM 500 and STM 225. The total cost of a complete Bacharach kit, including additional devices for measuring O₂ and CO, is approximately 4000 EUR, making it a cost-effective solution for periodic flue gas analysis.

Our findings suggest the necessity of research focused on improving the correlation between dust mass, dust colour, and dust’s chemical composition using a refined 10-degree scale, which could improve the method’s reliability, particularly in highly polluted environments where protecting sensitive sensors is crucial.

Further future research will be focussed on the easier determination of the soot number, without the use of a light tent, based on an innovative mobile phone application in connection with a uniquely designed reference data sheet for filter placement, including reference shade areas, to eliminate different light conditions. It also has potential as an inexpensive way to measure O₂ concentration using a lambda probe, which could complement the inexpensive Bacharach method. Further research and the further development of this method may allow for the more accurate measurement of the mass concentration (mg/m³) of particulate matter in exhaust gases in future.