Anda di halaman 1dari 6

Journal of Environmental Management (2002) 65, 7984 doi:10.1006/jema.2001.0530, available online at http://www.idealibrary.

com on

Wastewater treatment with a double-layer submerged biological aerated lter, using waste materials as biolm support
Francisco Osorio* and Ernesto Hontoria
University of Granada, Spain
Received 4 July 2000; accepted 5 November 2001

The aim of this research was to determine the performance of a submerged biological aerated lter, composed of a double-layer bed. This bed is made up of a top layer of ceramic material and a bottom layer of plastic material (both from previously used waste material). Efuent concentrations are presented related to the volumetric and hydraulic loads applied. The results were very satisfactory. If efuent concentrations of under 20 mg TBOD5/L and 25 mg SS/L are to be achieved, 487 kg TBOD5/m3/d and 30 kg SS/m3/d could be applied, respectively. For that maximum TBOD5 volumetric load that can be applied, a very reasonable consumption value of 10 kg O2/kg TBOD5,eliminated was obtained. The counter-current ow system outperformed the co-current ow system with respect to TBOD5 and SS removal. The tests were performed at a pilot plant with full scale height. The inuent used was primary efuent of a conventional treatment plant. A multivariant analysis (ANOVA) was applied to the results. # 2002 Elsevier Science Ltd. All rights reserved.

Keywords: biological aerated lter, waste materials, double-layer bed, biolm reactor, design parameters, pilot plant, wastewater treatment.

Introduction
On the one hand, there are operational problems involved in activated sludge systems (the most commonly used technique for wastewater treatment), including bulking (Blackbeard et al., 1986; Chudoba et al., 1973; Pujol and Canler, 1989; Tomlinson, 1976), while, on the otherhand, such treatment systems require large reactor volumes if ever-more-demanding pollutant elimination targets are to be achieved (Strohmeier and Schroeter, 1993; Smith and Hardy, 1992). These two conditioning factors are the main reason behind the authors' decision to begin the research into submerged

* Corresponding author. Email: fosorio@goliat.ugr.es


0301-4797/02/$ see front matter

biological aerated lters (BAFs) (Grasmick et al., 1979; Rittmann and McCarty, 1980). In recent years a number of BAF systems have been optimized and patented (Lazarova and Manem, 1994), retaining interest at the present time (Mann and Stephenson, 1997; Min Zhang et al., 1998; Kantardjieff and Jones, 1997). The most frequently studied biolm support media have been clay-, schist- or plastic-based ones of various types, such as PE, polyethylene, and PS, polyesterene, (Tschui et al., 1993). The rst BAFs investigation in Spain was initiated with an important French company and U.P.M. (Polytechnic University of Madrid, Spain), with the participation of some members of our Research Group. The Environmental Technology and Environmental Microbiology Groups research into BAFs began in the late 1980s, using waste materials, with resulting
# 2002 Elsevier Science Ltd. All rights reserved.

80

F. Osorio and E. Hontoria

benets for the entire community (Zamorano et al., 1995). Work commenced with a ceramic-based material presenting good adsorption and ltering capacity (Osorio and Hontoria, 1998), and a plastic material presenting low density, which helps to reduce energy consumption during backwashing operation (Zamorano et al., 1996). Several years after study of the system commenced with each material functioning independently, we decided to verify if a system with a single bed composed of two separate layers within a single reactor would be able to improve the results. This system was therefore called a `doublelayer bed'.

Methodology
Description of the pilot plant
The pilot-scale BAF system consisted of a 30 cm (inner diameter) 6 37 m methacrylate column that was lled with the support material to form the submerged bed. The bed consisted of 0.6 m of plastic material (described below) overlaid by 07 m of ceramic material (described below) for a total depth of 13 m. The principal reasons to put the ceramic layer on top were to facilitate the operation, especially during backwashing, due to the low density of the plastic material. That is a oating media. We had noticed an elevated concentration of biomass on top, operating with the ceramic material alone and down-ow, and a very low concentration of biomass on the bottom. Thus the idea of the double layer appeared to substitute at the bottom of the bed. We had serious doubts as to whether the ceramic layer was going to be able to improve the results of the plastic layer, operating with downow and putting the plastic layer on top. In relation to bed material height, among the conclusions of Francisco Osorio's Doctoral Thesis (1998) was the nding that the most suitable height for a ceramic layer alone was 1215 m. The separation between materials was carried out by means of a wire netting plate. This plate was fastened by means of a central vertical rail. Thus this plate was always on horizontal position, but the vertical movement was allowed during the backwashing. The plant's computer equipment was congured to make real-time measurements of the parameter values and variables being studied with a remote-control operation capability. Basically, the water ow, temperature, and pressures in both columns were monitored, as well as the state of
Figure 1. Schematic diagram of the pilot plant.

the mechanical equipment and alarms system management. The pilot plant ow schematic is shown in Figure 1.

Materials used for beds


It should be stressed that all the bed materials used were waste products, thereby maximising the use of discarded material. First, a support generated by the ceramics industry was used, obtained by grinding the waste material to make it suitable for the intended new use. The particle size was 25 mm, and the real relative density of the material was 218 g/cm3. The other bed material used was a plastic (polyethylene), generated by recycling the plastic sheeting that is used for the intensive greenhouse-based agriculture now commonly practised in southern Spain. Basically, the plastic is cut, washed, compacted and extruded. The real relative density of this material was 092 g/cm3, and the particle size was standardised at 5 mm.

Raw waters used


The inuent that fed the pilot plant was primary efuent, produced by the Granada Wastewater Treatment Plant (Spain). The volumetric loads applied were gradually increased over time by increasing the volume of wastewater owing into

Wastewater treatment

81

Table 1. Pilot plant inuent wastewater characteristics, during the periods of the study with different ow types, in the double-layer biolter Parameter Minimum Cod (mg/L) TBOD5 (mg/L) SS (mg/L) 390 180 96 Counter-current ow Maximum 598 340 148 Arithmetical average 507 255 116 Minimum 297 130 60 Co-current ow Maximum 609 310 208 Arithmetical average 488 243 106

the system. The characteristics of the inuent and loads applied corresponding to the operation phases are shown in Table 1.

Statistical methods
The software used for statistic analysis of the obtained results was Statgraphics Plus for Windows. The tools used to examine and interpret the data consisted of simple regression analysis and multifactor variance analysis (ANOVA). The efuent concentrations of various wastewater pollutants were analysed, co-varying with their volumetric loads applied. The factor included in the study was the ow type. The main aim was to determine if the covariable and the factor had a statistically signicant effect on the efuent concentrations, and especially to obtain the same comparative information for different levels of the factor (i.e., to determine whether there was any statistically signicant difference between them, to a condence level of 95%). The method used to discriminate between the media was Fisher's `least signicant difference' (LSD) procedure. The specic model used is known as the `covariance analysis model' (ANCOVA). This model is similar to ANOVA, with the sole difference being that with ANCOVA only one factor is studied, whereas several are introduced in ANOVA.

Experimental procedure
For the current study, the variable that was used in order to vary the conditions of the different tests was the water ow (m3/h), and one variable whose value was xed at the start was the air process ow. Aiming for a maximum consumption value of 12 kg O2/kg TBOD5,eliminated (for intermediate loads between those applied), the air volume supplied was xed at 013 kg O2/h (726 Nm3/m2/h of air). The volumetric gas ow was measured with a rotameter and the air was supplied at a pressure of 2 bar. For this study, the measured variables were as follows: Water ow; Air process ow; Total BOD5 (mg/L); SS (mg/L); COD (mg/L); pH. The calculated parameters when operating with the values adopted by the variables or other parameters were as follows kg O2/kg TBOD5,eliminated; Hydraulic load (m3/m2/h); Organic load or TBOD5 volumetric load (kg TBOD5/ m3/d); SS volumetric load (kg SS/m3/d). The sampling took place on a daily basis and the next parametersTotal BOD5 (mg/L); SS (mg/L); COD (mg/L); pHwere measured following the Standard Methods for the Examination of Water and Waste Water, 17th Edn. (American Public Health Association, 1989). Two different operational phases were studied, with counter-current ow and co-current ow systems, respectively. During backwashing, air and water ows were up-ow. Firstly, we carried out backwashing when a preset head pressure loss was reached. It was usually reached after 24 h for carbonaceous matter removal. Thus, we changed the method and the frequency of backwashing was xed every 24 h. On the basis of our experience, we carried out intense and short backwashing. We applied 055 m3/m2/min of water and 075 m3/m2/ min of air.

Results and discussion


Study of ow in double-layer bed
We concluded that the elimination rate of the double-layer bed system was optimal with a downow rather than an up-ow system. This is demonstrated by the graphs shown at Figures 2 and 3, for the TBOD5 and SS parameters, obtained from an ANCOVA analysis. These graphs conrm the tendency of BAFs to perform better as secondary treatment with counter-current ow. That afrmation was previously reported by other authors (Canler and Perret, 1994), using other types of support material.

82

F. Osorio and E. Hontoria

Figure 2. ANCOVA analysis results: variation in efuent TBOD5 concentration with different ow types, in doublelayer biolter.

Figure 4. Relation of double-biolter effuent TBOD5 concentrationTBOD5 volumetric load applied.

Figure 3. ANCOVA analysis results: variation in efuent SS concentration with different ow types, in double-layer biolter.

Figure 5. Relation of double-layer biolter efuent SS concentrationSS volumetric load applied.

For the relationship between the covariable (TBOD5 load applied) and the dependent variable (efuent TBOD5 concentration), the P value was almost zero, which is consistent with a strong link. For the TBOD5 parameter, the functional differences between the two ow types were statistically signicant (the difference in mean LS between down-ow and up-ow was 373743, with a P value of 00096). For the SS parameter, however, there were no signicant differences between the ow types (the P value in this case was 01035). The link between the covariable (SS load applied) and dependent variable (efuent SS concentration), as we expected, remained very strong (P 00001).

Figure 6. Variation in efuent TBOD5 concentration with double-layer biolter hydraulic load applied.

Design parameters for double-layer bed


Hereafter, as the pollutant elimination rates were lower with up-ow than with down-ow, only the latter results are presented, for the phase of the study with counter-current ow. Figure 4 shows

the TBOD5 efuent concentration results obtained once the water had passed through the BAF, related to the TBOD5 volumetric load applied. The SS efuent concentrations are presented in relation to the SS volumetric load applied, at Figure 5. Results for the hydraulic load applied, for the TBOD5 parameter, are shown in Figure 6. The

Wastewater treatment

83

Figure 7. Relation of oxygen consumption (with regard to eliminated TBOD5)double-layer biolter TBOD5 volumetric load applied.

Figure 8. Relation of ceramic-based material bed biolter efuent TBOD5 concentrationvolumetric load applied TBOD 5.

variation in oxygen consumption (with regard to eliminated TBOD5) in relation to the TBOD5 volumetric load applied is shown at Figure 7.

Dual media system in comparison with ceramic-based material system


Several years after beginning to study Submerged Biological Aerated Filters, the Environmental Technology and Environmental Microbiology Research Group (University of Granada, Spain) have optimised the performance of this system, using a ceramic-based material bed. A graph including ceramic-based material system results is shown below (Figures 8 and 9). A comparison with the combined system (Figures 4 and 5) indicates the benets of the dual media system.

Figure 9. Relation of ceramic-based material bed biolter efuent SS concentrationSS volumetric load applied.

reasonable consumption value of 10 kg O2/kg TBOD5,eliminated was obtained. As a result, there is a wide range of potential applications for the system proposed, due to the wide range of admissible loads which give adequate elimination rates, along with a reasonable oxygen consumption.

Summary of conclusions
 The elimination rate of the double-layer bed system was optimal with a down-ow rather than an up-ow system.  The elimination rates of the system are more than enough for effective removal of TBOD5 and SS at relatively high loading rates. The maximum TBOD5 and SS volumetric loads that can be applied, if efuent concentrations of under 20 mg TBOD5/L and 25 mg SS/L are to be achieved, correspond to 487 kg TBOD5/ m3/d and 30 kg SS/m3/d, respectively.  For the maximum TBOD5 volumetric load that can be applied, 487 kg TBOD5/m3/d, a very

References
Blackbeard, J. R., Ekama, G. A. and Marais, G. R. (1986). A survey of lamentous bulking and foaming in activated sludge plants in South Africa. Water Pollution Control 1, 90100. Canler, J. P. and Perret, J. M. (1994). Biological aerated lters: assesment of the process based on 12 sewage treatment plants. Water Science Technology 29, 1322.

84

F. Osorio and E. Hontoria Pujol, R. and Canler, J. P. (1989). Activated sludge performance: situation of the problem in France (Le foisonnement des boues activees, situation du pro bleme en France). T.S.M. 1, 1924. Rittman, B. E. and McCarty, P. L. (1980). Design of xedlm processes with steady-state biolm model. Biotechnology and Bioengineering 22, 23432357. Smith, A. T. and Hardy, J. P. (1992). High rate sewage treatment using biological aerated lters. JIWEM 6, April. Strohmeier, A. and Schroeter, I. (1993). Experiences with biological ltration in advanced wastewater treatment. Proceedings of European Water Filtration Congress, Ostend, 12 pp. Tomlinson, E. J. (1976). BulkingA survey of activated sludge plants. TR 35, WRC, England. Tschui, M., Boller, M. and Gujer, W. (1993). Tertiary nitrication in aerated biolter reactors. Proceedings Europe water Filter Congress, March, Ostende, Belgium, Vol. 17, 2.252.38. Zamorano, M., Gomez, M.A., Hontoria, E. and Gonzalez Lopez, J. (1995). Study of different materials used as support beds for the purication of wastewater with biolter. Biolm Structure, growth and dynamics Congress, Noordwijkerhout, The Netherlands, pp. 343348. Zamorano, M., Menendez, A., Osorio, F., Perez, J. I., Hontoria, E. and Gonzalez-Lopez, J. (1996). Use of recycled plastic material as biolm support in waste water biological treatment (Utilizacion de plastico reciclado como soporte para la depuracion biologica de aguas residuales). IV Siaga, Almera, Spain, II: 361370.

Chudoba, J., Ottava, V., Grau, P., Blaha, J. and Madera, V. (1973). Control of activated sludge bulking. Water Research 7/8, 11631237. Grasmick, A., Elmaeh, S. and Ben Aim (1979). Experimental study of submerged biological lters (Etude experimentale de la ltration biologique immergee). Water Research 14, 11371147. Kantardjieff, A. and Jones, J. P. (1997). Practical experiences with aerobic biolters in TMP (thermomechanical pulping), sulte and ne paper mills in Canada. Water Science Technology 35, 227234. Lazarova, V. and Manem, J. (1994). Advances in biolm aerobic reactors ensuring effective biolm activity control. Water Science Technology 29, 319327. Mann, A. T. and Stephenson, T. (1997). Modelling biological aerated lters for wastewater treatment. Water Research 31, 24432448. Min Zhang, Joo Hwa Tay, Yi Qian and Xia Sheng Gu (1998). Coke plant wastewater treatment by xed biolm system for COD and NH3-N removal. Water Research 32, 519527. Osorio, F. (1998). Optimization of recycled residual materials as biolm support in submerged biological aerated lters (Optimizacion de lechos inundados con soporte de residuos reciclados). Department of Civil Engineering, University of Granada (Unpublished). Osorio, F. and Hontoria, E. (1998). Use of waste materials in wastewater biological treatment (Aprovechamiento de residuos para la depuracion biologica de aguas residuales). Environmental Meeting of Almera, Spain, pp. 7685.