Development of combined fouling model in a membrane bioreactor

Abstract

Membrane bioreactor (MBR) system has recently become very important in wastewater treatment. However, membrane capacity during the filtration process is reduced by fouling, which is the deposition or adsorption of particles on the membrane surface or inside of the pores. For an efficient operation of MBR system, prediction of MBR fouling progress and determination of proper membrane chemical cleaning interval is necessary. Until now, four blocking models, which are complete blocking, intermediate blocking, cake filtration, and standard blocking, have typically been used to describe and predict the MBR fouling. In this study, a new combined membrane fouling model, based on the four constant flowrate blocking models, is developed to explain the combined effects of the different fouling mechanisms. This model is assumed to be more accurate in determining the chemical cleaning interval by considering variation of dominant fouling mechanisms in the filtration process. The combined fouling model is applied to data sets from a pilot‐scale MBR for monitoring the variations of dominant fouling stage. By checking temporal variations of root mean square error (RMSE) values from fitting trans‐membrane pressure data to four constant flowrate blocking models, the dominant fouling mechanism by each fouling stage is determined. In particular, the fouling prediction result by the new combined model of three fouling mechanisms, which are complete blocking, cake filtration, and intermediate blocking, excluding the standard blocking model has better predictive accuracy than the combined model of four mechanisms, which remarkably reduced the RMSE from 10.48 to 1.73. On the basis of the improvement, the new combined model with three mechanisms has the potential to predict MBR fouling process and to determine the optimum interval for membrane chemical cleaning. Our results would contribute the economic operation of MBR plants by suggesting the appropriate membrane cleaning interval with the filtration capacity maintained at optimal level. Copyright © 2011 Curtin University of Technology and John Wiley & Sons, Ltd.

Experimental study of colloidal flow in porous media at high velocities

“…For this purpose, the HPLC pump delivered the background solution (which later was used for the preparation of colloidal suspension) prepared from MilliQ deionized water adjusted to 0.1 M NaCl and pH = 3.5. Liquid permeabilities of the filter at various fluid velocities were calculated according to Eqn (1). Low range flow rates were as follows: 0.1, 0.2, 0.4, 0.6, 0.8, 1.0, 2.0, 4.0, 6.0, 8.0, and 10 mL/min.…”

“…Such processes include fouling/clogging of membranes and filters, [1,2] transport of contaminants in groundwater, [3] permeability reduction during aquifer remediation, [4] and injectivity decline of oil reservoirs because of particle invasion. [5] Among main problems during oil production by waterflooding of oilfields are well injectivity with fines lifting and straining due to high velocity flow or water salinity alteration during seawater flooding, re-injection of produced water, or disposal of produced water in aquifers.…”

Photocatalytic membrane filtration and its advantages over conventional approaches in the treatment of oily wastewater: A review

“…Fouling is the result of contaminant accumulation on the membrane surface and in the pores during the filtration. It is caused by organic colloids, organic macromolecules (organic fouling), inorganic suspended solids (inorganic fouling), soluble inorganic compounds (scaling), and by living/growing microorganisms (biofouling) 6,46,83,88 . In the case of oil‐contaminated waters, scaling is caused by the precipitation of salts and hydroxides, and the blockage of the surface and the fouling of the pores are caused mainly by oil droplets 46 .…”

A Simple Method to Identify the Dominant Fouling Mechanisms during Membrane Filtration Based on Piecewise Multiple Linear Regression

“…Second, the main mechanism may vary largely during the filtration process. However, a large number of researchers have used a fixed model to fit the whole process [ 23 , 24 ], and only a few have conducted piecewise fitting using multiple models [ 12 , 25 , 26 ]. The unreasonable setting of the time range or the number of time segments in the model fitting inevitably lead to deviations, such as the resultant autocorrelation of time series that has seldom been mentioned in previous fouling studies.…”