Machine Learning Approach to Design of Biodiesel Production Extraction Equipment from Tamanu Seed Oil

Abstract

Tamanu oil is one of the sources of vegetable oil needed on an industrial scale, so it is widely developed for its production. In designing essential oil processing on an in-dustrial scale requires mathematical calculations in measuring parameters in the re-actor, where this requires quite a long time in manual calculations. Utilization of ar-tificial intelligence technology, especially Machine Learning, That is very useful in the calculation process in the Tamanu oil extraction reactor. Machine Learning method used in this research is linear regression for data prediction and mean absolute error is used to measure absolute error. The model is used to predict the results of the outer diameter of the connecting pipe between tools with the parameters used, namely the target output in tons per year from cooking oil products made from tamanu seeds. R2 value of the graph of the CCIO Training Nominal Size data; Cooking oil; Methanol; Petroleum Ether is close to a value of 1, which means that the Input Feed value required to achieve the desired Output is close to 100% accuracy. Meanwhile, the Mean Absolute Error (MAE) value is the result of determining the Nominal Size with the Output value = 5000; 55000; and 150000 kg/hour shows a small difference in error values so that the design is acceptable.