Performance analysis and modelling of diesel engine operational characteristics using pyrolytic oil from scrap tyre

Abstract

In this work, an investigation on the fraction of tyre pyrolysis oil with a similar distillation range to that of automotive diesel (150 – 360 oC) was carried out to determine its suitability as an alternative or additive to petro-diesel fuel. The quality of this oil was evaluated by comparing its key properties to the requirements of South African National Standards for Automotive diesel fuel (SANS-342) and to conventional automotive diesel fuel. The viscosity, density, copper strip corrosion of this fuel were found to be within the acceptable limits set by SANS while sulphur content and flash point were out of their respective set limits. In addition, mixing rule equations for predicting viscosity and density for both pure and blends of the oil as a function of temperature were developed and evaluated. The equations were found to be suitable due to their low Absolute Percentage Deviation. Engine performance tests were carried out with blends of Distilled Tyre Pyrolysis Oil (DTPO) and petro-diesel fuel in a single cylinder air cooled diesel engine. The performance, emission and combustion characteristics of the diesel engine while running on these blends were evaluated and subsequently, a comparative analysis was performed with conventional petro-diesel fuel as the reference fuel. It was found that, the engine could run with up to 60% (DTPO) without any problem. Beyond this level the engine became unstable. The power and torque were similar at low and medium speeds. However, at high speeds, the power dropped with increase in DTPO in the blend. Fuel consumption was very comparable for all the test fuels. Carbon monoxide and unburned hydrocarbons were higher for the blends compared to petro-diesel fuel but oxides of Nitrogen were lower. The peak pressure for petro-diesel fuel was marginally higher than that of the blends. Present results indicate that, petro-diesel fuel can be blended with up to 60% DTPO and produce acceptable performance. Testing the diesel engine under different operating conditions is a time consuming and expensive process that also requires the use of specialised equipment which may not be readily available. An Artificial Neural Network (ANN) model based on a back-propagation learning algorithm was developed to predict engine performance and emissions separately, based on fuel blend and speed. The performance and accuracy of the model were evaluated by comparing experimental and ANN predicted results. The ANN was able to predict both engine performance and emissions with acceptable levels of accuracy. The values of correlation coefficient between experimental and predicted data being greater than 0.99. From this work, it can be implied that engine emission and performance can be predicted using neural network-based mode, consequently, it will be able to do further investigations without running laboratory experiments. Energy recovery from waste is an interesting field for engineers and scientists. It is hoped that this work will prompt new research ideals on how tyre pyrolysis oil can be improved for use as diesel engine fuel and building better models for diesel engine performance and emissions en_US