3.2.1. Dynamic Filtration Resistance
In the dynamic filtration experiment, the residual resistance drop was considered to be an important index of the filtration performance. The variation in the residual resistance drop depends on the amount of dust deposition on the surface or inside the filter media. Based on the test requirements of ASTM-D6830 standard, the dynamic filtration test was divided into three phases: aging phase, recovery phase, and measurement phase. Regardless of a variety of testing times and the filtration cycles of a filter media, consistency was only observed in the aging phase (cleaning was conducted once every 5 s for a total of 10,000 filtration cycles). Therefore, the increase rate of the residual resistance drop in the aging phase can reflect the rate of dust deposition on the surface or inside the different filter media. During the aging phase, there were 10,000 filtration cycles of each filter media. To facilitate lawful analysis, we divided 10,000 filtration cycles into 20 intervals of the aging phase that were associated with each filter media. Each interval contained the data of 500 residual resistance drops. The average value of these 500 residual resistance drops represented the residual resistance drop of the interval.
Figure 7a illustrates that the average residual resistance drops of the three filter media increased exponentially when the filtration cycle was increased for the depth filtration media (PSF, PNF, and APNF). This indicates that in the dynamic filtration process, the filter media could be cleaned and regenerated; however, a foundation structure of the stable dust cake could still be formed inside the depth filtration media. With a continuous process of filtration, the structure of the dust cake could be increased continuously inside the filter media. This led to a continuous increase in the residual resistance drop. In the three depth filtration media, the average residual resistance drops increased in the following order: APNF > PSF > PNF. The initial filtration resistance drop of the three filter media were equivalent. At the end of the first 500 filtration cycles, the average residual resistance drops of the three filter media were as follows: 54.7 Pa (PSF), 48.2 Pa (PNF), and 57.9 Pa (APNF). At the end of the aging phase, the final average residual resistance drops of three filter media were as follows: 273.6 Pa (PSF), 216.7Pa (PNF), and 324.2 Pa (APNF). The difference between average residual resistance drops of the three depth filtration media reflected the different dust deposition conditions in the three depth filtration media. In this experiment, both the PSF filter media and PNF filter media had the same fiber diameter, but the average pore size of PNF filter media was greater than that of PSF filter media. For PNF filter media, the increase in the rate of average residual resistance drop during the aging phase was less than that of PSF filter media. This indicates that when the fiber diameter was constant, smaller the average pore size of the depth filtration media, the faster would be the dust deposition rate of depth filtration media, and greater would be the amount of dust deposited in the depth filtration media. The average pore size of PNF filter media was 22.49 μm, while the average pore size of APNF filter media was 22.52 μm. This indicates that the average pore size of APNF filter media was similar to that of PNF filter media; however, the increasing rate of the average residual resistance drop of APNF filter media was the highest among the three depth filtration media. This may be attributed to the fact that the fiber diameter of APNF filter media was larger. Larger the fiber diameter, greater would be the chances and the probability of dust deposition on the surface of the fiber. These conditions were more favorable for the deposition of dust in the depth filtration media, and it led to the formation of a dust cake.
Unlike the depth filtration media, it was difficult for the dust to pass into the surface filtration media and form a dust cake. Most of the dust particles could only deposit on the surface of the membrane. For the surface filtration media, the fiber diameter was not the main factor that affected the deposition of dust. Moreover, the surface of the membrane-coated filter media was smooth and could be cleaned easily; therefore, it was difficult to form a stable layer of dust on the surface of the membrane-coated filter media during the process of dynamic filtration. The dynamic filtration resistance of the surface filtration media was mainly governed by the size and distribution of membrane pores.
Figure 7b illustrates that the average residual resistance drops of the two surface filtration media (APMCNF and PMCNF) increased exponentially with an increase in filtration cycles. For APMCNF filter media, the increasing rate of average residual resistance drop was greater than that of PMCNF filter media. At the end of the first 500 filtration cycles, the average residual resistance drops of the two filter media were as follows: 306.7 Pa (APMCNF) and 218.1 Pa (PMCNF). At the end of the aging phase, the final average residual resistance drops of the two surface filtration media were 583.5 Pa (APMCNF) and 382.3 Pa (PMCNF). The difference between the average residual resistance drops of the two filter media increased with the filtration process. The average pore sizes of the two filter media were as follows: 0.42 μm (APMCNF) and 1.05 μm (PMCNF). This indicates that smaller the average pore size of the surface filtration media, greater would be the dynamic filtration resistance.
Figure 8 illustrates that the average residual resistance drop of the surface filtration media increased when the dust blocked the pores of the filter membrane. The experimental results indicate that dust can easily block the micropores of the surface membrane when the average pore size of the surface filtration media is smaller. Consequently, the average residual resistance drop of the surface filtration media increased rapidly.
Figure 9 illustrates the initial filtration resistances, the final residual resistance drops, and the increments in the residual resistance drop of the five filter media in the aging phase. For the two surface filtration media, the initial filtration resistances and the final residual resistance drops were greater than those in the remaining three depth filtration media. Out of the five filter media, APMCNF filter media showed following characteristics: the highest initial filtration resistance, final residual resistance drop, and residual resistance drop increment. For PMCNF filter media, the increment in the residual resistance drop was less than that of APNF and PSF filter media. These results indicate that smaller the average pore size of the filter media, faster would be the increasing rate of the residual resistance drop of the filter media. In particular, when dust entered and deposited inside the depth filter media, the fiber diameter of the depth filter media further affected the dynamic filtration resistance of the filter media by influencing the deposition rate of dust.
Figure 10 illustrates the mass gain of filter media after the aging phase. The mass gains of various filter media were as follows: APNF > PSF > PNF > PMCNF > APMCNF. The mass gain of the depth filtration media was significantly greater than that of the surface filtration media. Therefore, the amount of dust deposited in the depth filtration media was much greater than that deposited on the surface of the membrane-coated filter media. For the depth filtration media, the experimental results indicate that the dynamic filtration resistance increases mainly because of the following reasons: (i) the deposition of dust inside the filter media and (ii) the formation of a dust cake. The experimental results further indicate that smaller the average pore size of the depth filtration media, greater would be the deposition rate of dust in the depth filtration media. A larger fiber diameter was more conducive for the deposition of dust and the formation of a dust cake inside the depth filtration media. For the surface filtration media, the experimental results indicate that the mass gain of APMCNF filter media was the least; however, the initial filtration resistance, the final residual resistance drop, and the increment in the residual resistance drop of APMCNF filter media were greater than those of PMCNF filter media. Based on the experimental results of the mass gain for the surface filtration media in the aging phase, it was further proved that the increase in the residual resistance drop of the surface filtration media was due to the blockage of dust in the pores of the filter membrane. Smaller the average pore size of the surface filtration media, easier would be the irreversible blockage of dust in the micropores of the surface filtration media and faster would be the increasing rate of dynamic filtration resistance of the surface filtration media. Moreover, smaller the average pore size of the surface filtration media, smoother would be the surface filtration media. Furthermore, the dust was stripped very thoroughly under the action of a cleaning airflow, and the mass gain of the surface filtration media was smaller in magnitude.
Figure 11 illustrates the residual resistance drop that occurred in the five kinds of filter media during the recovery phase.
Figure 11a illustrates that the residual resistance drop of the three kinds of depth filtration media increased with an increase in the filtration cycles that occurred during the recovery phase. In the 30 normal filtration cycles, the residual resistance drops of the three depth filtration media increased in the following order: APMCNF > PSF > PNF. In the recovery phase, the increasing rates of residual resistance drop of the three depth filtration media were obviously lower than those in the aging phase. This indicates that after providing aging treatment to the depth filtration media, the foundation of the dust cake improved gradually inside the depth filtration media. As a result, the formation rate of a dust cake became slower, and the deposition rate of dust decreased in the depth filtration media. Therefore, the residual resistance drops of the three depth filtration media increased slowly. During the recovery phase, the increments in the residual resistance drop of the three depth filtration media were as follows: 12.1 Pa (PSF), 10.3 Pa (PNF), and 17.3 Pa (APNF). The increments in the residual resistance drop and the increase rate of dynamic filtration resistance of the three depth filtration media were similar and much smaller than those in the aging phase. This indicates that for the depth filtration media, the average pore size and the fiber diameter greatly influenced the dust deposition rate at the initial stage of a dynamic filtration process. With the progress in dynamic filtration, the size and the distribution of filter pores in the depth filtration media changed with the deposition of dust inside the filter media. This weakened the influence of the average pore size and the fiber diameter of depth filtration media on the dust deposition rate.
As shown in
Figure 11b, the residual resistance drops of the two surface filtration media were always observed at a relatively stable level in 30 normal filtration cycles, which was consistent with the experimental results in reference [
27]. In contrast, the residual resistance drop of APMCNF and PMCNF filter media were about 600 Pa and 400 Pa, respectively. This indicates that a dynamic equilibrium was established between the deposition and the stripping of dust, which accumulates on the surface of the two membrane-coated filter media. After the aging treatment, fine particles were deposited on the micropores of the surface membrane; these particles were steadily removed by the cleaning airflow. Moreover, residual resistance drop of surface filtration media was also stabilized.
Figure 12 shows the changes in the residual resistance drop of the five filter media during the measurement phase: the residual resistance drops of the three depth filtration media increased continuously when the filtration cycles were elongated after the aging and recovery phases. The residual resistance drops of the two surface filtration media remained relatively stable during the recovery and measurement phases, which occurred after the aging phase. The results of the entire dynamic filtration test of the five filter media showed that in the initial filtration stage (
Figure 7), the increasing rate of the residual resistance drop was the fastest when a clean filter media was transformed into a dust-containing filter media (this conclusion can be obtained by comparing
Figure 7,
Figure 11 and
Figure 12.). With the progress of the dynamic filtration process, there was a gradual improvement in the structure of a dust cake that was formed inside the depth filtration media. Moreover, a dynamic equilibrium was reached between the following two processes: (i) the blockage of membrane pores by dust and (ii) the removal of dust layer formed on the membrane of the surface filtration media. Consequently, the increasing rate of the residual resistance drop of the surface filtration media remained relatively stable. In the entire dynamic filtration test, the residual resistance drop of the depth filtration media increased steadily. At the end of the measurement phase, the final residual resistance drop of APNF filter media was greater than 400 Pa, indicating that it was greater than the final residual resistance drop of PMCNF filter media. It can be inferred that when the depth filtration media was used for long-term dynamic filtration, the residual resistance drop would eventually surpass that of the surface filtration media; moreover, the dynamic filtration resistance of the depth filtration media would increase at a rate faster than that of surface filtration media.
3.2.2. Dynamic Filtration Efficiency
As shown in
Figure 13,
Figure 14,
Figure 15 and
Figure 16, the dynamic filtration efficiencies of the five filter media were determined by performing dynamic filtration tests.
Figure 13 and
Figure 14 showed the dynamic filtration efficiencies of the five filter media in the aging and recovery phases, respectively.
Figure 15 and
Figure 16 illustrate the dynamic filtration efficiencies of PM
2.5 and TOT for the five filter media during the measurement phase. The experimental results indicate that dynamic filtration efficiencies of the five filter media increased with the progression of the dynamic filtration process. In the entire dynamic filtration tests, the dynamic filtration efficiencies of the five filter media were greater than 99.96%. Compared to the static filtration efficiencies of the five filter media (
Figure 6), the filtration efficiencies of the clean filter media were the lowest for a life cycle of filtration. Compared to the static filtration tests, the differences between the filtration efficiencies of the depth filtration media and the surface filtration media were greatly reduced during the dynamic filtration test. This indicates that dust deposition and dust cake formation were essential to improve the filtration efficiency of depth filtration media. At different phases of a dynamic filtration process, the dynamic filtration efficiencies of the five filter media were as follows: APMCNF > PMCNF > PSF > PNF > APNF. The order of dynamic filtration efficiencies of the five filter media was opposite to that of their average pore size, that is, smaller the average pore size of the filter media, greater was the dynamic filtration efficiency of the filter media.
In the depth filtration media, the “Dust filter Dust” filtration mechanism of was implemented. According to this mechanism, greater the deposition of dust in the filter media, better would be the structure of a dust cake and greater would be the dynamic filtration efficiency. In contrast, the dynamic filtration resistance of a filter media increases with the deposition of dust inthe filter media. Based on the above inferences, we propose the following conclusion: under the same experimental conditions, faster the increasing rate of dynamic filtration resistance of the depth filtration media, faster would be the deposition rate of dust inthe filter media, and higher would be the dynamic filtration efficiency of the filter media. In the dynamic filtration experiments, we found that the increasing rate of the residual resistance drop was the fastest for the APNF filter media, followed by the PSF filter media, and the PNF filter media. Therefore, the deposition rate of dust was the highest for the APNF filter media, followed by the PSF filter media and the PNF filter media. This inference has been confirmed in the mass gain of the filter media, which occurred during the aging phase. The magnitude of dynamic filtration efficiencies of the three filter media were in the following order: APNF > PSF > PNF; this finding was not compliant with the actual situation of the test (PSF > PNF > APNF). The reason for the above discrepancy was the fact that in the dynamic filtration process, the structure of the dust cake formed inside the depth filtration media was destroyed by the compressed air, enabling dust to penetrate through the depth filtration media. Some research studies [
28] have shown that in a filtration cycle, dust leakage occurs mainly in a short period of time after cleaning the filter media; moreover, dust penetration rate decreases rapidly as the dust layer is repaired on the surface of the filter media. After cleaning the depth filtration media, the surface and the fabric structure of the depth filtration media were exposed to an airflow of dust. Larger the average pore size of the depth filtration media, easier was it for the dust to penetrate into the depth filtration media; moreover, larger the fiber diameter, lower was the filtration efficiency of the fabric structure of the depth filtration media. Therefore, the APNF filtration media has the highest increase rate of residual resistance drop; however, its dynamic filtration efficiency was the lowest.
In the surface filtration media, the microporous membrane was used as a barrier for filtering the particles (sieving effect). The size and the distribution of membrane pores in the filter media were used to determine the performance of filtration. For the surface filtration media, the experimental results of the dynamic filtration efficiency were compliant with the laws used to perform experiments of dynamic resistance. These experimental results were different from those obtained for the depth filtration media. Smaller the average pore size of the surface filtration media, higher would be the dynamic filtration resistance and the dynamic filtration efficiency.
In the aging phase, dynamic filtration efficiencies of the three depth filtration media were in the range of 99.96–99.98%. For the two surface filtration media, the dynamic filtration efficiencies were greater than 99.99% in the aging phase. During the recovery phase, the dynamic filtration efficiencies of all the five filter media were greater than 99.99%. In the measurement phase, the dynamic filtration efficiencies of the five filter media were greater than 99.997% for PM2.5, and the dynamic filtration efficiencies of the five filter media were greater than 99.998% for TOT. In the entire dynamic filtration tests, the dynamic filtration efficiencies of the two surface filtration media were always greater than those of the three depth filtration media. This implies that dynamic filtration efficiencies of the surface filtration media were more stable in magnitude. In the entire dynamic filtration process, the standard deviations of the dynamic filtration efficiencies of the three depth filtration media were always greater than those of the two surface filtration media. For the dynamic filtration efficiencies of dust, the standard deviations of the five filter media were as follows: APNF > PNF> PSF > PMCNF > APMCNF, which were in the same order as the average pore size of five filter media. It can be concluded that smaller the average pore size of the filter media, smaller would be the standard deviation in the dynamic filtration efficiency; therefore, the dynamic filtration efficiency of the filter media would be more stable.