Prediction of Energy Cost and Specific Energy Consumption in Manufacturing Process with Multiple Linear Regression Equation and Equivalent Unit

Abstract

This research is about how to predict the energy cost and specific energy consumption (SEC) in the manufacturing process with multiple linear regression equations and equivalent units. The case study is the electronic factory that produces 4 types of electronic print circuit boards. To analyze this problem, the equivalent unit is applied to transfer production quantities of each type of electronic print circuit board into the equivalent production quantities. With data of quantities of electricity consumption, LPG consumption, and the equivalent production quantities for 4 types of the electronic print circuit board for 12 months in the past, the multiple linear regression techniques are applied to develop the prediction equations for energy consumption cost of electricity, LPG and total energy as  

1) E_Ele (kWh) = (3.926 x 10⁶) – (9.112 x 10⁵) EQ₁ + (8.287 x 10⁵) EQ₂ - (1.020 x 10⁵) EQ₃+ (5.461 x 10⁵) EQ₄ for energy consumption cost of electricity,

2) E_Ther (Liters) =          (4.239 x 10⁴) - (1.185 x 10⁴) EQ₁ + (2.868 x 10⁴) EQ₂ - (3.023 x 10⁴) EQ₃+ (1.722 x 10⁴) EQ₄ for energy consumption cost of LPG and 3) E_Total (MJ) = (1.526 x 10⁷) - (3.596 x 10⁶) EQ₁ +        (3.747 x 10⁶) EQ₂ - (1.172 x 10⁶) EQ₃+ (2.425 x 10⁶) EQ₄ for energy consumption cost of total energy. Finally, with the additional data of total equivalent production quantities for 4 types of electronic print circuit board, SEC equations of electricity, LPG and total energy are developed as 1) SEC_Ele (kWh/1000 ft²) = 0.007 EQ²_Total – 17.696 EQ_Total + 17,492 for electricity,                         2) SEC_Ther(liters/1000 ft²)  = 0.0005 EQ²_Total - 0.9861 EQ_Total + 549.88 for LPG and 3) SEC_Total (MJ/1000 ft²) = 0.0389 EQ²_Total – 89.953 EQ_Total + 77,610 for total energy.