Assessment and Evaluation of Pongamia Pinnata Oil as an Alternative Fuel for Mine Equipment

Abstract

The growing population has demanded to need more power generation. The present era relies mainly on fossil fuels as solid and liquid power sources. The demand for more power can lead to the exhaustion of fossil fuels. This also has given rise to a search for an alternative fuel source. Liquid fuels can readily be supplemented using other alternative fuel sources, including ethanol, natural gas, electricity, hydrogen, propane, methanol, P-series fuels, and biodiesel. Biodiesel can be used as a supplement to diesel in the form of blends. Pongamia pinnata is one of the most popular sources of biodiesel in South India. In the present study, Pongamia pinnata is introduced as a biodiesel source in different forms, namely raw form, ester form, and combined form. i.e., Raw pongamia oil (RPO), methyl ester of Pongamia oil (MPO), a combinational form of methyl esters of Pongamia and waste cooking oil (MPWO), methyl esters of Pongamia pinnata and waste cooking oil (MWO). Samples were prepared with blends varying from 0% to 30% in increments of 5% volumetrically, and these samples were referred to as base biodiesel blends. A similar set of samples were prepared with an additional 10% volume of ethanol and acetone to the base biodiesel blends by reducing the volumetric contribution of diesel. The total sample size was 75, including diesel. For each prepared sample, property tests were performed to identify the density, kinematic viscosity, and calorific value. A deviation was computed with reference to diesel. It was found that the RPO showed the largest deviation as far as the properties were concerned, which were found to be 1.66% lower, 13.24% lower, and 1.81% for density, Kinematic viscosity, and calorific value, respectively. Further, the engine test was carried out on a single-cylinder four-stroke diesel engine. The engine speed was kept constant (1500 rpm), and the engine load varied from 5% to 100% in increments of 25% of the engine load. i.e., the engine speed was maintained constant for each engine load, and the same was followed for all 75 samples. The time taken for 10cc fuel consumption and the corresponding emissions at each engine load was recorded. Four of the emission parameters were observed, namely carbon monoxide (CO), Carbon dioxide (CO2), unburnt hydrocarbons (UHC), and oxides of Nitrogen (NOx). The mass of fuel consumed vi(MF), brake thermal efficiency (BTE), and brake-specific energy consumption (BSEC) were calculated. The effects of blending on the performance and emissions of the engine were evaluated by plotting a scatter graph. The average deviations of performance and emission parameters were found. The average deviations were measured by referencing each base biodiesel blend and its ethanol and acetone additions. BTE and BSEC of diesel-ethanol (DE) were found to be near to the BTE and BSEC of diesel, with deviations (higher) of 2.59% and 2.33%, respectively. DE is considered the best alternative for diesel in terms of efficiency and energy consumption. RPO is considered the poorest source regarding efficiency and energy consumption, with deviations (Lower) of 32.73% and 4l.4%, respectively. All biodiesel blends showed lower CO and UHC emissions, exhibiting higher CO2 and NOx. RPE (raw Pongamia-ethanol) blends showed better CO and UHC with deviations (Lower) of 57.47% and 51.1% and poorer CO2 and NOx with deviations (higher) of 49.16% and 172.67%. Statistical analysis was performed to predict the performance and emissions of the engine. Regression modelling (multivariate) and ANOVA analysis were used to develop regression equations and identify the significant parameters. The parameters blend, load, density, kinematic viscosity, calorific value, and mass of the fuel consumed were considered input parameters. The output parameters include BTE, BSEC, CO, CO2, UHC, and NOx. It was observed that the regression models showed a performance of more than 70% R-squared values at a confidence interval of 95% in predicting the performance and emissions of the engine. Blend and load on the engine were considered the most significant parameters. Prediction studies were also performed using the ANN technique. The number of neurons varied from 4 -10, and a two-layered perceptron neural network was chosen for the analysis. TRAINLM and LEARNGDM were used as training and learning algorithms. The Transfer function is TRANSIG. Each neuron's Root mean squared error (RMSE) value was computed. The best model was evaluated to be the one with the least RMSE. With 375 data sets, 263 were used for training the model, and the remaining 112 were used for testing and validating the model. 50% of the remaining data is shared equally for each testing and validation. An MLPNN network, the optimised model for predicting BTE, BSEC, CO, and NOx, was found to have 6 viineurons. The best model for predicting the UHC and CO2 is the one with 5 neurons with R2 value of 0.99 for all training, testing, and validation. Field studies were conducted in one of the esteemed Underground metal mines in southern India. A mine Tipper was selected to conduct the test with 20% of the blends of base biodiesel. Two conditions, idle and high idle, were implemented to study the variations. The four emission parameters studied were NO, NO2, NOx, and CO. The observations were similar to experimentation, and the deviations are estimated. RPO reduced CO emissions of the engine. Alternatively, increasing nitrogen emissions. The deviations of RPO compared to diesel were 29.5% and 50.74%, with diesel at both idle and high idle conditions for CO, respectively. The deviations measured are 29.5% and 33% at both idle and high idle conditions for NOx, respectively.