Press machine process improvement by using DOE method

Abstract

The purpose of this study is to improve the efficiency of press machine for producing finish good product thru design of experiments application. Press process is significant to improve since the output of the palm oil product will be divided into two which are palm oil and palm nuts. The objective of this study is to identify the best parameter setting for reduce broken nuts during press process thru adoption of the design of experiments method. A full factorial design with single replication has been used to study the effects of important parameters which are temperature, cone angle, cone pressure and percentage of water. The measurement of output response is identified as percentage of broken nut. Referring to the company specification, the percentage of broken nuts must not exceed than 7 %. Randomize experiments was conducted based on table generated thru Minitab software. Normal probability test was carried out using Anderson Darling Test and show that the P-value is 0.325. Thus, the data is normal since there is no significance different between the actual data with the ideal data. Referring to effect estimate value for each parameter, factor B (Cone Angle) was not significant. Thus the experiments were converted from 24 to 23 with two replications. Referring to the ANOVA, all of the factors are significant except the AC interaction since the P -value for each of parameters less than 0.05. From the main plot and interaction plot, the recommended setting for each of parameters were suggested as low level for temperature, low level for cone pressure and high percentage of water. The prediction model was developed thru regression in order to measure effect of output response for any changes on parameters setting. In the future, the experiments can be enhanced using two replications of experiment and Taguchi methods in order to do verification of result.