Development model of small orifice flow using simple linear and multilinear regression

Abstract. The concept of "small orifice flow" describes the movement of a fluid through a tiny hole or orifice. This is a typical occurrence in engineering and fluid mechanics. In machine learning, linear regression is a fundamental and extensively used technique for modelling the connection. In this research, an analysis of the influence of the variables H1, H2 and Q was carried out on the output value of the discharge coefficient (Cd). Model evaluation is done by calculating calibration parameters like the coefficient of determination. R square (R2), also known as the coefficient of determination, describes the extent to which dependent data can be described by independent data. There are a total of 47 datasets that will be included in the development of linear and multilinear models. In this study, 90% (42) of the training data and 10% (5) of the total test data were used. Based on simple linear model, Q parameter having a stronger linear relationship with the Cd parameter compared to the H1 and H2 parameters with a strong multiple R value of 0,754 and moderate R2 correlation value of 0.569. There is a significant influence between the variables Q and Cd based on significant F value. The multilinear regression method was chosen because of its statistical data handling capabilities and ease of creation, implementation, and operation. the multilinear regression model with H1, H2 and Q values as independent variables can explain the Cd value as the dependent variable very well with the higher value of multiple R and R2, and small error value.