Membrane bioreactors for advanced treatment of wastewater from a large petrochemical industrial area

The protection of sensitive water bodies is today claiming for stricter legislation to protect the environment from discharge of conventional and priority pollutants present in industrial and urban wastewaters. The use of the Best Available Techniques (BAT) is advised by national and international regulation: at this regard the knowledge of the limits achievable by the different technologies is of great importance especially where high quality of effluent wastewaters is required, including concentrations of discharged nonconventional pollutants in the order of some μg/L or pg/L. To date membrane bioreactors (MBR) have been presented as one of the BAT in the treatment of certain industrial discharges (allowing also for water reuse) and have known a widespread applications in the last decade. A relatively new area for large MBRs concerns the treatment of the wastewaters from oil refineries and petrochemical industries. This work presents the treatment of real petrochemical wastewater taking into account both pilot scale experimental research and full scale investigations. In particular, the wastewater came from a large petrochemical area located in the Venice area, one of the world's most sensitive water bodies. Even in accordance to the special laws for the protection of the Lagoon, the membrane bioreactor is called for the best removal of conventional and nonconventional pollutants, such as heavy metals, cyanides and PCB. In addition, over the last decades the MBR is subject to low, but very changeable, loading related to the crisis and unsteadiness of the chemical productions. The main experimental work was carried out by a pilot MBR with reaction volume of 4.2 m3, operating in parallel with the full scale MBR. The first objective of the experimental work was to study the effect of low and changeable loadings on a membrane bioreactor, adequate for hydraulic loadings but very oversized for mass loadings. In particular, the consequences of the slower metabolic activity of the biomass were evaluated, even to find the best reactor configuration adequate to the real influent loadings. For this purpose, once validated the pilot MBR as representative for the full scale plant, four experimental runs were carried out aiming at: increase the fraction of active biomass within the activated sludge (Run II and Run III); better use the excess bioreaction volume (Run IV and Run V) by finally creating an hybrid MBR, which realizes the sequencing enrichment of nitrifying biomass in the continuous WWTP. In addition, specific studies concerned the inhibitory effect of petrochemical (i.e. caustic spent) fluxes on nitrifying activity and on feasible solution to increase the nitrification potential of the plant. As far as nonconventional pollutants are concerned, the work focused on cyanide and PCB compounds and the maximal treatment potentials, intended as the actual limit achievable by the MBR as a BAT for petrochemical wastewater treatment. Besides the occurrence and chemical-physical forms in the petrochemical wastewater; the mechanisms and removal efficiencies of the initial clariflocculation and membrane bioreactor were studied. As far as concern the pilot plant, the effects of different operating conditions to enhance the mechanism of biodegradation and bioadsorption were assessed and the feasibility at full scale was studied. Results from experimental pilot plant activity showed the effect of the proper organic loading on the fraction of the active biomass. Therefore, the study was addressed towards the study of the best feasible approaches to increase the F/M . In particular, the best use of surplus volumes of the biological unit (run IV), then the doubling of the load applied at the biological reactor (run V) (that corresponds to the elimination of one line of treatment in the full scale plant). In this last case the nitrification activity was also doubled, so as to demonstrate that the autotrophic activity is currentlysubstrate-limited and not influenced by inhibition. However in order to cope with nitrogen variability, possible increase of inhibiting fluxes (i.e. spent caustic that reveals to produce nitrification inhibition up to 60%) or collection of new fluxes (i.e. refinery wastewater), a more feasible approach was the enrichment of the reactor with nitrifiers obtained in the batch growth. An increase of 0.1 gN/kgVSS*h in the nitrification potential of the overall biomass in the reactor could be achieved within 10-15 days. Sludge enriched with nitrifiers, moved to the continuously fed MBR, demonstrated an increase of nitrification potential up to a NLR of 0.022 kgN/m3*d (versus a NLR of 0.008 kgN/m3*d applied in normal operation), increased proportionally with the added nitrifiers. Removal of free cyanide in pilot and full scale MBR showed that free cyanides are biodegraded up to the the limit of quantification (2 μg/L). On the other hand strong complex cyanide (probably in the form of ferricyanide) are inert to the biodegration and undergo event to low adsorption (10%). These results were confirmed by specific lab-scale batch tests.. However mass balance in the pilot MBR shows rates of removal of free cyanides one order of magnitude lower of the one found in the batch test (6-11 μgCN/gVSS/h). This can be explained with the lower concentrations influent to the bioreactor that influenced the biodegradation kinetics (being close to the semisaturational constrant) estimated in the around of 0.65 μgCN/gVSS/h. On the other side the dosage of organic substrate to the aerobic compartment during the experimental activity improved the removal of total cyanide probably due to increased activity of biomass (biodegradation and bioadsorption potential) and related co-metabolism processes. As far as concern treatment of PCB good efficiency of removal are obtained in both primary (60-70%) and secondary treatment (50-60%) obtaining an overall efficiency of about 80%. In particular an efficiency of removal in clariflocculation depends on the particulate fraction in the influent wastewater, while a linear removal in the biological treatment is obtained in relation to the influent concentrations at the reactor. In addition, the effluent concentration was always within a stable range of concentration that varies depending on the single congener and that can be assumed as the lower limit value achievable by the MBR. This shows that the membrane bioreactor have an additional effect on the removal of PCB up to a certain concentration which can be assumed as actual limit of the BAT-MBR. Finally the role of the clogging sludge of a membrane in the removal of metals was investigated, and compared with the removal that occurs in the activated sludge of the biological plant. Possible enhancement of bioaccumulation-bioprecipitation mechanisms in this layer was proposed. In particular heavy metals more accumulated in the clogging sludge are in the order As>Ni>Cd>Fe. The release of the metals present in the clogging sludge during acid and oxidazing chemical cleaning was evaluated trough batch test: results show that release of the metals is usually in the range of 5-10% of the metal content of the clogging sludge, but it is not directly related with the biosorption in/on the clogging sludge.