Pyrolytic conversion of thermoplastics into liquid fuels

© 2003 Dr. Su Lyn LowDisposal of plastic waste into landfills has become increasingly problematic due to legislative measures, public pressures and high costs. Growing environmental awareness and reductions in available landfill space have prompted increasing interest in plastics recycling programs. In Australia, attention has focussed on recycling and other alternatives as a means of minimising the amounts of waste being directed to landfills. Pyrolysis is one promising method for the treatment of mixed and contaminated plastic wastes. Plastics can be thermally degraded to produce useful liquid hydrocarbons, which can then either be added to existing fuel or solvent products, or returned to a refinery where they can be added to the oil feedstock. The main objective of this project was to investigate the development of a simple, small-scale waste plastics pyrolysis system suitable for Australian conditions. The pyrolytic system developed in this project differed from existing systems with respect to the scale and simplicity of the equipment employed. The technical feasibility of largescale continuous plants, with catalytic plastics decomposition processes has already been demonstrated. The unique aspect of this project was the design and modelling of a simple, intermittently operated plant to handle small amounts of a waste plastics feedstock. The first stage in the development of such a plant involved the selection of a suitable reactor system. Two pyrolytic reactor systems were examined. A series of experimental investigations was conducted using the six most common polymeric materials found in household waste: linear low and high density polyethylene (LLOPE and HDPE), polypropylene (PP), polystyrene (PS), polyvinyl chloride (PVC) and polyethylene terephthalate (PET). In the first experimental phase, a bench scale spherical reactor was designed and constructed. Experimental results showed that pure samples of LLOPE, HOPE, PP and PS resin can readily be pyrolysed to produce liquid Yields in excess of 70 %. The product yields were shown to be sensitive to operating temperature, heating rate and whether or not a reflux condenser was present. Due to the shortcomings of the first bench scale spherical reactor system, a laboratory scale cylindrical reactor system was designed and constructed. Experiments conducted with the cylindrical reactor system have shown it to be a technically more suitable reactor system. It was shown to be capable of pyrolysing polymeric materials in both batch and semi-continuous modes of operation to give a liquid yield of over 70%. In the third phase of the project, a comprehensive mathematic model, POLYPYROS was developed. Unlike existing pyrolysis models, this unique model encompasses both the polymer pyrolysis and the heat transfer processes occurring within the cylindrical reactor system. The model was created in the form of a user friendly computer program. The model links together three separate modules to describe the overall polymer degradation process including melting, reaction kinetics and the system's heat and mass transfer processes. It is capable of simulating what occurs during both batch and semicontinuous operating modes. Simulation outputs include detailed predictions of temperature profiles and product compositions. The validation of the POLYPYROS model involved comparisons between predicted temperature profiles, product compositions and product yields and the corresponding experimental data. Given the complexity of the processes involved, the model proved able to reproduce the general trends and in particular closely predicted the nature of changes in temperature profiles and product compositions for both the batch and semi-continuous operating modes. The POL YPYROS model was also used to explore the influence of selected operating parameters on liquid compositions and product yields. These studies have shown that the POL YPYROS model could be used to facilitate optimisation of plant design and operating parameters and could reduce the need for extensive pilot plant testing