3.1. Combustion Performance
Concerning the temperature evolution during this experiment, there was a constant increase in flue gas temperature during the three first measurements, corresponding to the ignition stage. During the operating stage, the comparison between flue gas temperature and set temperature (30 °C) was essential to understand the different working regimes in this stage. Thus, once the flue gas temperature was higher than the set temperature, the rotation rate of the screw conveyor was different, as explained in previous sections, affecting the flue gas temperature, especially after 45 min. Finally, during the phasing-out stage, a constant decrease in temperature was found due to the fact that the stove is off and some of its components are gradually disconnected, as previously explained.
It should be noted that flue gas temperature values were relatively low due to the fact that the pipe that connects the stove with the chimney is considerably long (around 6 m) and metallic, absorbing heat and, consequently, decreasing the temperature value at the point where the temperature probe is placed.
Regarding flue gas composition, the following figures (
Figure 13,
Figure 14,
Figure 15 and
Figure 16) show its evolution over time, being influenced by the working stage. Thus, in the case of carbon monoxide evolution (
Figure 13), there was an increase from 0 to approximately 1600 ppm, corresponding to the operating stage. This peak corresponded to the maximum combustion intensity, as observed in oxygen depletion and temperature increase [
30]. Afterwards, a constant decrease in CO content was found, including the phasing-out stage, where CO concentration is below 50 ppm after 65 min of experience.
A similar trend was observed in the case of carbon dioxide (see
Figure 14), where the peak concentration (exceeding 3.75%) was observed at 40 min in the operating stage to drastically decrease from then on, with negligible values from 60 min to the end of the experiment.
Regarding NO content (
Figure 15), the highest concentrations (at around 120 ppm) were found for at least 15 min in the operating stage, followed by a considerable decrease up to 65 min, where the concentration was nearly negligible.
Finally, concerning oxygen evolution (
Figure 16), there were two clear levels, with a decrease in O
2 content during the ignition stage, keeping the lowest concentration constant (at around 15%) practically during the whole operating stage to present an increase in the phasing-out stage, where atmospheric levels were found after 60 min of experience. These values were within the range found in the literature for the flue gas composition in a biomass pellet stove [
22]. Also, compared to the previous facility where these improvements were not considered (included in previous studies [
15]), lower CO emissions were obtained in this current study, proving the higher efficiency of the improvements incorporated into this facility.
As observed in the previous figures, the different operational stages have an influence on combustion and the final outcome of each experience. Depending on the final users’ habits (mainly on their demand), different scenarios can be considered for a certain use of this pellet stove: first, frequent switch on/off of the system, that would show lower yields in terms of combustion due to the higher prevalence of phasing-out stage; second, less frequent switch on/off, which improves the yield of the combustion system as the phasing-out stage would be reduced on the whole. This fact is in accordance with previous studies where the burn rate had a strong influence on combustion efficiency, also depending on different burning phases. Thus, low and high burn rates implied lower combustion efficiencies [
31]. In this sense, as observed in the literature, short working regimes in pellet stoves could increase the emissions of pollutants due to the lower combustion efficiency of the ignition stage [
9].
Compared to commercial pellet stoves without these improvements, it could be claimed that manufacturers are currently designing these devices under safety standards, based on a realistic performance for different kinds of biomass. Nevertheless, in this work, the redefinition of these facilities and determining the margins to perform new parameters and outcome parameters have been carried out, obtaining better results in general compared to commercial equipment. Thus, some of the results observed in this work could be implemented as potential improvements in new versions of the setting of current products, which could have a significant impact with low cost.
3.2. Biomass Characterization
The different biomass samples studied in this work should comply, at least partially, with the following requirements: high value of HHV, low moisture, low ash content, high fixed carbon percentage, and low sulfur content. As observed in
Table 4, the selected products had a high HHV, which makes them suitable for energy purposes. In that sense, poplar pellets presented the highest HHV (exceeding 17,000 kJ·kg
−1), whereas the lowest value was found for holm oak (slightly below 16,000 kJ·kg
−1). Another interesting property is the ash content, which is undesirable as its generation during combustion could imply technical problems. In that sense, pear tree and poplar pellets offered the best results, with values below 2.5%, whereas sugarcane bagasse clearly exceeded these values, with 3.87%. Regarding moisture levels, which should be as low as possible to avoid the impact of unnecessary pre-treatments to improve the combustion performance as well as to improve the efficiency of transportation, two clear groups were found: on the one hand, holm oak and pear tree presented the highest moisture levels, exceeding 7%, whereas poplar and sugarcane bagasse had values of around 3%. When volatile matter values were compared, poplar pellets showed the highest value, exceeding 80%, whereas holm oak pellets had the lowest volatile matter. Concerning C and H content, there were slight differences, ranging from 45.9 to 47.4 for the former and from 5.77 to 5.99 for the latter. Other parameters, such as N and S content, have a strong influence on some components in flue gas emissions, such as NO or sulfur compounds, respectively. In that sense, low values are desired, making the use of pear tree and holm oak pellets suitable for this purpose regarding N and S content, respectively. This way, ultimate analysis values were similar to those included in the literature [
32]. Finally, bulk density, which is an essential parameter for the correct combustion performance in a boiler (as it determines the fuel flow, combined with the feeding rate of the screw conveyor), presented different values depending on the kind of biomass, with differences exceeding 15% between the lowest value and the highest value (corresponding to holm oak and sugarcane bagasse pellets, respectively). It should be noted that the adjustment of the domestic stove to these circumstances will be essential to contributing to better combustion performance. In general, most of these parameters are within the range observed for typical pellets, as observed in the literature [
30,
32], which makes them suitable for combustion processes.
The information included in the previous table, along with the working regimes of the combustion stove, is important to calculate the corresponding thermal capacity of this facility. As observed in previous studies, the impact of the nature of biomass, along with other parameters such as pellet length, percentage of fines, or moisture, could play an important role in the combustion performance of residential pellet stoves or pellet boilers, especially concerning emissions (which are related to the ultimate and proximate analyses) [
33]. That is the reason the characterization of the pellets is important in any case.
Figure 17 shows the different feeding calibrations for the different samples considered in this study.
As explained in the literature, a continuous and stable pellet supply is important to make combustion as efficient as possible with low CO emissions and temperature fluctuations [
24].
Thus, for each case, a specific linear equation was found (with high determination coefficients, from 0.9905 for sugarcane bagasse to 0.9988 for poplar pellets), which can be used to obtain the thermal capacity for the different biomass samples. This way, following the example of poplar pellets (as the rest of the samples would follow the same reasoning), Equation (1) was obtained:
where “
P” is the weight of biomass (expressed in kg) and “
t” is the time, in min or h. Considering this equation for 1 min,
P = 54.401 g·min
−1 = 9.06·10
−4 kg·s
−1. According to the working times of the screw conveyor established for the operating stage, 8.57 s per minute, the mass flow rate (expressed in kg·min
−1) can be obtained (Equation (2)):
Finally, the thermal capacity (expressed in kW) of the stove is given by Equation (3):
Thus, a thermal capacity of 2.22 kW was obtained for poplar pellets. According to this reasoning,
Table 5 shows the main results for all the pellets included in this work.
The thermal capacity of the stove for the different biomass samples is included in
Table 6.
On the other hand, one of the ways to increase the yield in a domestic stove is the improvement of its combustibility. For this purpose, the percentage of oxygen supplied for combustion should be modified, which can be easily carried out by changing the rotation speed of the flue gas fan. In this fan, a potentiometer was installed in order to vary the power of this device. Thus, according to the different rotation speeds, flue gas composition was analyzed, representing these results in the Ostwald diagram, whose interpretation would give useful steps to improve the combustion performance in this case.
Regarding the different rotation speeds of the flue gas fan, the main results are included in
Figure 18. Thus, according to O
2 evolution with rotation speed, its percentage decreased with the corresponding decrease in rotation speed (finding oxygen content, similar to other studies [
23]), whereas the opposite was observed in the case of CO
2 percentage, with a slight increasing trend when the speed was reduced. Finally, CO and NO content did not show a clear trend with fan speed (although a slight increase was observed at lower rpm), which could be due to the low concentrations found for each compound near the quantification limit of the analyzer. The presence of compounds due to incomplete combustion, even at low concentrations and with airflows that are enough to satisfy complete combustion, could be on account of short retention durations of these products in the stove [
22]. It should be noted that, for these experiments, the flue gas temperature was relatively similar regardless of the rotation speed, which is due to the fact that the chimney was long (6 m) and metallic, which could have contributed to heat absorption from the flue gas. In any case, these values ranged from above 30 °C (for 2000 and 320 rpm) to above 35 °C in the other cases.
Concerning the Ostwald diagram obtained from experimental data,
Figure 19 shows the main results found for this work at different rotation speeds of the flue gas fan. The aim was to be as close as possible to the optimum point at which all the carbon are completely oxidized without excess air. Thus, by observing this figure, it can be noted that as the rotation speed was decreased (with the subsequent decrease in O
2 percentage), the corresponding points get closer to the optimum point, which is desired to obtain better combustion performance.
These results are in accordance with the literature, where a pellet stove calibration model was proposed, showing that a decrease in excess air ratio implied a decrease in flue gas temperature, improving combustion efficiency [
23].