Bioreactor landfills: experimental simulations, full scale monitoring and fuzzy modelling

In the perspective of a sustainable waste management, the amount of biodegradable municipal solid waste (MSW) destined to landfilling should be reduced. In such context, new technologies were developed in recent years with the aim of a more rapid stabilization of the waste, such as pretreatments and bioreactor landfills (BRLs). In the Italian waste management scenario, although solution have been adopted towards waste recycling and recovery, landfilling is still playing an important role. The case study of Cerro Tanaro (CT) landfill depicts a typical situation of an average Italian district without incineration facilities. The MSW disposed in CT landfill is the residual fraction, pretreated through aerobic mechanical-biological treatment (MBT) in order to reduce the biodegradability prior to landfilling. This disposal site, originally built as a conventional landfill, was equipped with a leachate recirculation system. Although the benefit of moisture increase had been previously demonstrated, most of the studies in literature tested the effects on raw MSW, with relatively high organic contents, above 40%. Moreover, the use of existing models for the simulation of landfill behaviour are not suitable for unconventional landfill technologies, unless very high uncertainties are introduced. Based on the case study of CT landfill, the need of novel approaches for the study and the simulation of emerging landfilling solution had been identified. The main aim of this thesis was to offer a novel and simple tool for the prediction of landfill behaviour when unconventional management practises are introduced, regarding both the quality of the MSW landfilled and the operational conditions. Therefore, experimental tests were conducted at lab-scale, the full-scale case study of CT landfill was monitored and a fuzzy-logic (FL) based model was developed for the prediction of landfill gas production. The experimental tests at lab-scale demonstrated that coupling MBT with leachate recirculation could reactivate the biodegradation processes even for low biodegradable waste (LBW), thanks to moisture increase. Although it was difficult to establish a stable methanogenic phase, CH4 production reached 28 NL kg-1 after 442 days of experimentation, that is 85% of its bio-methane potential (BMP). Also leachate quality presented reduced pollution strength with low COD and NH4+ concentrations. The results highlighted the differences between the tested LBW and fresh not pretreated MSW and between the optimized lab-scale and the heterogeneities of the full-scale landfill. Two deterministic models were tested for the estimation of CH4 production from LBW under leachate recirculation: Gompertz kinetic model and BIO-5 model. The production curves obtained by the two models confirmed the limits of deterministic methods and underlined the need of different approaches, able to deal with the uncertainties typical of landfill gas modelling. A FL-based model to predict methane generation in BRLs was proposed. Eleven deterministic inputs (pH, ORP, COD, VFA, NH4+ content, age of the waste, temperature, moisture content, organic fraction, particle size and recirculation flow rate) were identified as antecedent variables. Two outputs, or consequents, were chosen: methane production rate and methane fraction. The fuzzy model was built and tested on the lab-scale experimentation of LBW under leachate recirculation. Additionally, the data of other six lab-scale studies from literature were used to widen the applicability of the proposed model. The proposed fuzzy model was then applied on the full-scale case study of CT landfill. Although this case study was characterized by lack of information from the previous literature, fuzzy modelling represented a valid tool which could be easily adapted to the specific system under study.Another aspect concerning modern landfills has been treated, that is the presence of emerging contaminants among MSW. A FL-based model was developed to evaluate biogas and methane production from BRLs in presence of ZnO nanoparticles (NPs). The fuzzy model was tested on the data of a lab-scale study simulating a bioreactor landfill with 100 mg of added nano-ZnO/kg of dry waste, conducted in the Institute of Environmental Sciences (Boğaziçi University – Istanbul). By comparing the results of the proposed FL-based models with the deterministic models, the FL approach showed better performances for both lab-scale and full- scale methane predictions, confirming its high potential in the modelling of landfill environments under different emerging scenarios. The proposed FL-based models represent a basis to describe methane production in such systems, which, thanks to its learning structure, can be easily upgraded with additional parameters and information coming from future findings in this topic.