1、 Amin Engarnevis is a PhD candidate in the Department of Mechanical Engineering, University of British Columbia, Vancouver, BC. Ryan Huizing is director of research and development at dPoint Technologies, Vancouver, BC. Fouling of Membrane-Based Energy Recovery Ventilators by Aerosols Amin Engarnevi
2、s Ryan Huizing, P.Eng. Ali Vaseghi Student Member ASHRAE Student member ASHRAE Sheldon Green, PhD, P.Eng. Steven Rogak, PhD, P.Eng. Member ASHRAE ABSTRACT Membrane-based Energy Recovery Ventilators (ERV) are an effective means of reducing energy cost and allow for scaling down HVAC equipment. Owing
3、to their compact geometry and rough surfaces of the porous membrane substrate, ERV exchanger cores can be fouled by airborne particulate matter. In this study, the influence of particulate fouling on the membrane-based ERVs was investigated via accelerated material- and core-level fouling experiment
4、s. The core-level experiments, inside an Aerosol Wind Tunnel (AWT), investigated the effect of dust accumulation on the performance of cross-flow cores (including sensible and latent effectiveness, and pressure drop) through comparing pre- and post-fouling performance tests of two core samples. The
5、influence of the membrane surface exposed to particle-laden air, and core face velocity were considered during AWT tests. We have found that for solid particles of 0.310 m, deposition fractions to membrane surfaces range from 0.05 at high air velocities (1 m/s (197 fpm) to as high as 0.2 at lower ai
6、r velocities (0.5 m/s (98 fpm). Nevertheless, accelerated fouling tests using coarse, dry test dust (ISO A3 medium), did not show any significant degradation to the sensible and latent effectiveness of the cores. In the lack of proper filtration, however, this fouling may result in an energy penalty
7、 because of the added pressure drop in the system. Additionally, a mass transport analysis is presented to explain the results of AWT tests. In the material-level experiments, the effects of fouling with nano particles on the membrane material performance (including water vapor transport, gas cross-
8、over, and pressurized air leakage) were examined. Preliminary results show that deposition of non-hygroscopic graphite particles has minimal influence on the membrane, whilst deposition of soluble NaCl particles on the uncoated porous surface of the membrane may result in partial pore blockage, cons
9、equently reducing the water vapor permeation through the membrane up to 15% of the initial value. INTRODUCTION Energy required for air conditioning accounts for about 40% of the primary energy consumption in the building sector. It is thus important to recycle the energy (i.e. heat and moisture reco
10、very) used to condition the indoor air. Counter- and cross-flow air-to-air exchangers for heat and moisture (known as “Energy Recovery Ventilators”, ERV) allow increased fresh air supply with a low energy penalty because the incoming fresh air is partially conditioned by the ERV. It is shown that us
11、ing membrane-based ERVs (mERV) can result in significant energy savings when the latent load constitutes to a large fraction of the total thermal load in the HVAC system, as well as improvement of indoor air quality (IAQ) and comfort in buildings (Zhang 2012). Membranes used for mERVs must be highly
12、 permeable and selective for water vapor over other gases and contaminants. This is of practical importance for mERV exchanger to prevent crossover of undesirable gasses and contaminants from the exhaust air to the supplied fresh air (Huizing et al. 2015), (AHRI 2013). One cost-effective method of f
13、abricating membranes for mERVs is applying a very thin (1m) selective film layer onto a micro-porous substrate layer. The resultant membrane has an asymmetric composite structure. The substrate is made from a low cost, widely available porous polymer providing mechanical integrity and facilitating w
14、ater vapor transport through the membrane; the coating is a dense, hydrophilic film layer that provides the true selective barrier. The membranes have not undergone many years of field testing, yet there are preliminary indications that water vapor transport might degrade after use, possibly as the
15、result of exposure to air pollution, or other environmental stresses (Woods, 2014). Deposition of aerosol particles on membrane surfaces may result in degradation of water vapor permeability due to partial pore blockage of the porous substrate, or failure of the dense layer when aerosols are water s
16、oluble or chemically active. There are many studies of porous membrane fouling from liquids (e.g. reverse osmosis) and gases (e.g. micro- and ultra-filtration), but, we have found only one study for composite membranes in HVAC-relevant conditions. (Charles firstly, experiments comprising deposition
17、fraction measurements, loading tests, and initial and post-fouling performance measurements were conducted to investigate the fouling problem in the core level. The second step was focused on material-level examination of membrane fouling with nano particles and changes to physical properties of mem
18、brane material fouled by particulate matter. EXPERIMENTAL METHODOLOGY Core-Level Experiments Test Apparatus. An Aerosol Wind Tunnel (AWT) was developed based on the requirements of ASHRAE standard 52.2 (ASHRAE 2012) to study particulate fouling behavior and performance variations of full ERV cores u
19、nder accelerated fouling tests. Figure 1 (a) shows a schematic of the test apparatus. (a) Schematic illustration of Aerosol Wind Tunnel (AWT) (b) Photograph of custom-built aerosol generator Figure 1 Experimental Test Rig Inlet airflow is passed through a HEPA filter to remove pre-existing particles
20、 from supply air. Particles of known properties are introduced into the wind tunnel, and passed through a special test section that holds ERV cores. This test section is designed to work with two airstreams to simulate the real operating conditions of ERVs in HVAC systems (ASHRAE Standard 84 (ASHRAE
21、 2013). Primary (S,i S,o) and secondary (E,i E,o) airstreams are considered as the representatives of fresh air from outdoor and stale air from indoor space, respectively. The flow rate of both airstreams and the temperature of the primary air stream can be controlled between 0.03-0.3 m3/s (65-600 c
22、fm), and 25-80C (77-177F), respectively. Poly-disperse particles are generated by means of a fluidized-bed aerosol generator (TSI model 3400A) via dispersing dry powders of standard test dusts. After the flow has fully developed and uniformly mixed with particles, size-resolved concentrations in the
23、 air up- and down-stream of the test section are determined using a TSI Optical Particle Sizer (OPS 3330). In order to simulate the accelerated fouling conditions, very high concentrations of particles (up to 150 mg/m3 at 0.094 m3/s (200 cfm) can be injected into the primary airstream using a custom
24、-built aerosol generator (Figure 1 (b). Measurement of high concentrations is conducted using a TSI DUSTTRAK DRX. Fouling Test Procedures. The test procedures for deposition fraction measurements and dust loading tests were similar to that described in ASHRAE 52.2 standard for testing the efficiency
25、 of air cleaners. A series of fouling tests including four dust loading steps and five deposition fraction measurements were conducted for three mean air velocities (0.5, 0.8, 1 m/s (98, 157, 197 fpm) under dry working conditions (50C (122F) and 6% RH). Detailed test procedures are explained elsewhe
26、re (Engarnevis et al. 2014). Performance Measurement of ERV Cores. Initial and post-fouling performance measurements (including sensible, and latent effectiveness, and pressure drop) were conducted according to the procedures described in AHRI standard 1060 (AHRI 2013). The performance parameters ar
27、e determined by measuring the inlet/outlet temperature, humidity, and flow rate of the air streams across the ERV core: = ( )(,)( )(,)(1) = ( )(,)( )(,)(2) Material-Level Experiments Aerosol Particulate Loading. For accelerated loading of membrane samples, a test setup is developed (Figure 2) to sim
28、ulate the fouling conditions. An atomizer (TSI Aerosol Generator Model 3076) and a PALAS GFG1000 are used for generating test aerosols of salt (dg=88nm), and a soot-like carbon aerosol called here “graphite” (dg=82nm), respectively. These generators inject high concentrations of aerosol nanoparticle
29、s (up to 100 mg/m3 at 10-4 m3/s (0.212 cfm) into the primary airstream. This enables one to perform 8hr loading experiments that are the equivalent to one year of working in the field (i.e. having same accumulated mass concentration of incident particles). Size-resolved concentration of particles ar
30、e measured by a TSI SMPS 3080 spectrometer from up- and down-stream locations (alternating back and forth), and are used to calculate the fraction of particles deposited onto the membrane samples. Figure 2 Schematic illustration of nano aerosol loading test setup Membrane Material. The MX4 membrane
31、used for this study has a porous substrate made from silica-loaded polyethylene, and a water vapor selective coating made of a polyether-polyurethane copolymer. The membrane samples have a thickness between 105 to 115m. Performance Testing of Membranes. Material-level experiments will determine the
32、particle deposition fraction, and initial and post-fouling performance of membrane samples. The performance metrics include membrane water vapor flux, pressurized air crossover leak rate (at 6894 pa (1 PSI), and EATR rate (oxygen crossover). Measurement methodology is explained elsewhere (Huizing et
33、 al. 2014). RESULTS AND DISCUSSION Dust Loading of ERV cores In the AWT, according to the procedures described in section 2.1.2, two cross-flow ERV sample cores A and B were loaded with ISO 12103-1, A3 medium test dust (Powder Technology Inc.) on the coated and uncoated sides, respectively. Fouling
34、behavior (including particle deposition fraction and dust loading pattern on the membrane surfaces) for three different mean air velocities, as well as the impact of dust loading on the thermal and hydraulic performance of the cores are discussed in the following sections. Deposition Fractions. Size
35、-resolved deposition fractions of the ERV core A (loaded on coated membrane side) for three air velocities are plotted in Figure 3(a). For comparison, the results of the deposition fraction at air velocity of 0.5m/s (100fpm) for the ERV cores A and B (loaded on the uncoated substrate side) are plott
36、ed in Figure3(b). (a) (b) Figure 3 (a) Deposition fraction of ISO A3 medium dust particles at three mean air velocities on the coated membrane side, (b) comparison between deposition fraction of the coated and uncoated membrane sides at 0.5 m/s (100 fpm). Deposition fractions were found to be a stro
37、ng function of air velocity and a weaker function of particle size. As shown in Figure 3 (a), deposition fractions for sub-micron particles vary within the range of 0.05 to 0.15. In the absence of thermophoresis under isothermal conditions, small particles are most likely to be transported to membra
38、ne surfaces by Brownian diffusion (Hinds 1998). Given the relatively large depth of membrane channels, lower mean air velocities result in higher residence times for fine particles with high diffusion velocities (Epstein 1997). As can be seen in Figure 3 (a), this leads to significantly higher depos
39、ition to the membrane surfaces at lower air velocities. For coarse particles (1-10 m size range), deposition in compact exchanger geometries similar to ERVs is dominated by inertial impaction on leading edges, air turbulence impaction to walls, and gravitational settling inside narrow channels (Sieg
40、el the equivalent resistance of 22 s/m (6.71 s/ft) would be significant. However, since there is a large area that is not covered in dust, deposit layer is modelled as an additional porous layer with a parallel resistance of large voids (discontinuties). Although the dust deposit contributes signifi
41、cant resistance, the discontinuities provide a low resistance pathway which is negliable compared to other series resistances in the membrane. Therefore, one can expect minimal impact on total vapor transport across the membrane due to the formation of dust layer. 203040506070809010030 50 70 90Sensi
42、bleEffectiveness (%)Flow Rate (SCFM)Core A - CleanCore B - CleanCore A - Loaded on coated sideCore B - Loaded on uncoated side010203040506070809010030 50 70 90Latent Effectiveness (%)Flow Rate (SCFM)Core A - CleanCore B - CleanCore A - Loaded on coated sideCore B - Loaded on uncoated side0.000.040.0
43、80.120.160.2030 50 70 90Pressure Drop(inchWC)Flow Rate (SCFM)Core A - CleanCore B - CleanCore A - Loaded on coated sideCore B - Loaded on uncoated side(a) (b) Figure 5 (a) Dust deposit on the uncoated substrate side of membrane, (b) schematic of the dust loaded membrane and mass transfer resistances
44、 to vapor transport through the membrane. Nanoparticle Loading of Dense Polymer Membranes Membranes samples of MX4 were loaded with NaCl and graphite nanoparticles using the apparatus described in section 2.2.1. All experiments were conducted isothermally at 21 C (70F) with aerosol stream inlet cond
45、itions of TA=21C (70F), RHA=60%, QA=10-4 m3/s (0.212 cfm), and sweep stream inlet conditions of TC=21C (70F), RHC=0%, QB=10-4 m3/s (0.212 cfm). Loading tests were completed in a counter-flow test module (explained in (Huizing et al. 2014) for about eight hours. This is equivalent to one year of oper
46、ation in a highly polluted atmosphere with PM10 aerosol concentration of 130 g/m3. Deposition Fractions. Deposition fraction measurements were completed (as described in section 2.2.1) for each membrane sample at the beginning of the loading process. Figure 6 shows typical deposition curves of NaCl
47、and soot particles loaded on the uncoated side of membrane samples. Figure 6 Deposition efficiency of NaCl and graphite particles on the uncoated side of MX4 membrane The trend of deposition fractions in figure 7 are consistent with those of the sub-micron segment of deposition curves for dust loadi
48、ng experiments. Since the SMPS uses electrical mobility to measure the particle size, it allows for resolving the deposition curve for ultrafine particle sizes (as low as 16 nm). Given the geometry and the flow conditions of the membrane test module, for ultrafine particles; Brownian diffusion resul
49、ts in deposition fractions between 0.1 (at 100nm) to 0.6 (at 16nm). Considering the pore size distribution of membrane porous substrate (median pore 70 nm), it would be possible for a particle to enter pores. The Palas soot has an open fractal structure which would presumably offer little resistance to water vapor transport, but the NaCl particles are not porous. Membrane P
