Empirical modelling for sound absorption coefficient of Kenaf Fibre by using response surface method

Abstract

This thesis discusses the use of Box Behnken Design (BBD) approach to plan experiments to find the Sound Absorption Coefficient (SAC) for Kenaf/Polypropylene (PP) and Kenaf/Epoxy sandwich sample with an overall objective of optimising the density (0.25g/cm3 - 0.65g/cm3), thickness (10mm - 50mm), pressure (1000kg - 5000kg), and frequency of sound (125Hz - 5000Hz). Based on the literature review, research that identifies the significance of independent variables of sound absorber capability using particular methodology remain scarce. Besides that, there was no research that adopted Design of Experiment (DOE) to be used before experimental work on acoustic areas related to natural fibre, specifically on Kenaf Fibre (KF). All research on identification of SAC was done by analysing previous research, by using calculation and experimental work and analysis. It was found that all research on KF and its sound characteristics such as SAC did not produce any empirical equation which could give benefits to other researchers on their study on the same characteristics. The researcher managed to achieve all the objectives of this research. The first objective of this study is to identify the significance of factors for a mixture of the composition between KF with PP and a mixture of the composition between KF with Epoxy to get the SAC by using the RSM as the DOE. The second objective is to compare the SAC between the simulation result and the experimental result based on the suggested test sample by the RSM and tested in an impedance test tube. The third objective is to develop and compare the empirical equation for the SAC from simulation result and experimental result by using the density, thickness, compression pressure, and frequency as the significant factors. RSM was adopted to validate the output parameters (responses) which were decided by the input process parameters. RSM also quantified the relationship between the variable input parameters and the corresponding output parameters and validated by Analysis of Variance (ANOVA) before experiments on 24 sets of sample that proposed by BBD were performed. Value of R2, Adjusted R2 and Predicted R2 with >95%, and p-value with <0.005 were referred. RSM estimated the interaction and the linear effects. A regression model was developed and its adequacy was verified to predict the output values at nearly all conditions. The output parameters measured through experiments (actual) were in good match with the predicted values using the model developed by BBD with Average of Residuals value < -0.051235. Using Design-Expert software, 2D and 3D plots were generated and explicitly give an idea of the dominating process of variables over others and the order of dominance (Density, Thickness, Frequency and Pressure). This research is significant because it produced minimum number of experiments, a validated empirical equation model and conducted an optimised test. Therefore, in conclusion, the SAC and Empirical Equation output from the RSM is reliable, valid and compatible with the experimental results having minimum residuals. As such, a mixture of KF with Epoxy (!"# = 0.2045 − (0.27081 × 1234567) + (0.002305 × ;ℎ5=>3244) + (0.00014 × ?@2AB23=7) + (0.000009583325 × D@244B@2)) is more reliable as a sound absorber if compared to KF with PP.