Hydrocarbon removal with constructed wetlands
Eke, Paul Emeka
Wetlands have long played a significant role as natural purification systems, and have been effectively used to treat domestic, agricultural and industrial wastewater. However, very little is known about the biochemical processes involved, and the use of constructed treatment wetlands in the removal of petroleum aromatic hydrocarbons from produced and/or processed water. Wastewaters from the oil industry contain aromatic hydrocarbons such as benzene, toluene, ethylbenzene and xylene (ortho, meta and para isomers), which are highly soluble, neurotoxic and cause cancer. The components of the hydrocarbon and the processes of its transformation, metabolism and degradation are complex, the mechanisms of treatment within constructed wetlands are not yet entirely known. This has limited the effective application of this sustainable technology in the oil and gas industries. Sound knowledge of hydrocarbon treatment processes in the various constructed wetlands is needed to make guided judgments about the probable effects of a given suite of impacts. Moreover, most of the traditional treatment technologies used by the oil industry such as hydrocyclones, coalescence, flotation, centrifuges and various separators are not efficient concerning the removal of dissolved organic components including aromatics in the dissolved water phase. Twelve experimental wetlands have been designed and constructed at The King’s Buildings campus (The University of Edinburgh, Scotland) using different compositions. Selected wetlands were planted with Phragmites australis (Cav.) Trin. ex Steud (common reeds). The wetlands were operated in batch-flow mode to avoid pumping costs. Six wetlands were located indoors, and six corresponding wetlands were placed outdoors to allow for a direct comparison of controlled and uncontrolled environmental conditions. The experimental wetlands were designed to optimize the chemical, physical and microbiological processes naturally occurring within wetlands. The outdoor rig simulates natural weather conditions while the indoor rig operates under controlled environmental conditions such as regulated temperature, humidity and light. Benzene was used as an example of a low molecular weight petroleum hydrocarbon within the inflow of selected wetlands. This chemical is part of the aromatic hydrocarbon group known as BTEX (acronym for benzene, toluene, ethylbenzene and xylene), and was used as a pollutant together with tap water spiked also with essential nutrients. The study period was from spring 2005 to autumn 2007. The research focused on the advancing of the understanding of biochemical processes and the application of constructed wetlands for hydrocarbon removal. The study investigated the seasonal internal interactions of benzene with other individual water quality variables in the constructed wetlands. Variables and boundary conditions (e.g. temperature, macrophytes and aggregates) impacting on the design, operation and treatment performance; and the efficiency of different wetland set-ups in removing benzene, chemical oxygen demand (COD), five-day @ 20°C N-Allylthiourea biochemical oxygen demand (BOD5) and major nutrients were monitored. Findings indicate that the constructed wetlands successfully remove benzene (inflow concentration of 1 g/l) and other water quality variables from simulated hydrocarbon contaminated wastewater streams with better indoor (controlled environment) than outdoor treatment performances. The benzene removal efficiency was high (97-100%) during the first year of operation and without visible seasonal variations. Seasonal variability in benzene removal was apparent after spring 2006, the highest and lowest benzene removal efficiencies occurred in spring and winter, respectively. In 2006, for example, benzene removal in spring was 44.4% higher than in winter. However, no seasonal variability was detected in the effluent ammonia-nitrogen (NH4-N), nitratenitrogen (NO3-N) and ortho-phosphorus-phosphate (PO4 3--P) concentrations. Their outflow concentrations increased or decreased with corresponding changes of the influent nutrient supply. In addition, benzene treatment led to trends of decreasing effluent pH and redox potential (redox) values but increasing effluent dissolved oxygen (DO) concentrations. Approximately 8 g (added to the influent every second week) of the well balanced slow-releasing N-P-K Miracle-Gro fertilizer was sufficient to treat 1000 mg/l benzene. Results based on linear regression indicated that the seasonal benzene removal efficiency was negatively correlated and closely linked to the seasonal effluent DO and NO3-N concentrations, while positively correlated and closely linked to the seasonal effluent pH and redox values. Temperature, effluent NH4-N and PO4 3--P concentrations were weakly linked to seasonal benzene removal efficiencies. During the entire running period, the seasonal benzene removal efficiency reached up to 90%, while the effluent DO, NO3-N, pH and redox values ranged between 0.8 and 2.3 mg/l, 0.56 and 3.68 mg/l, 7.03 and 7.17, and 178.2 and 268.93 mV, respectively. Novel techniques and tools such as Artificial Neural Network (self-organizing map (SOM)), Multivariable regression and hierarchical cluster analysis were applied to predict benzene, COD and BOD, and to demonstrate an alternative method of analyzing water quality performance indicators. The results suggest that cost-effective and easily to measure online variables such as DO, EC, redox, T and pH efficiently predicted effluent benzene concentrations by applying artificial neural network and multivariable regression model. The performances of these models are encouraging and support their potential for future use as promising tools for real time optimization, monitoring and prediction of benzene removal in constructed wetlands. These also improved understanding of the physical and biochemical processes within vertical-flow constructed wetlands, particularly of the role of the different constituents of the constructed wetlands in removal of hydrocarbon. These techniques also helped to provide answers to original research questions such as: What does the job? Physical design, filter media, macrophytes or micro-organisms? The overall outcome of this research is a significant contribution to the development of constructed wetland technology for petroleum industry and other related industrial application.