| Summary: | 碩士 === 輔仁大學 === 統計資訊學系應用統計碩士在職專班 === 102 === This study aims to identify the heavy oil refining process five parameters: refining capacity, the reactor temperature, iron content, sodium content, viscosity influence on the life of the catalytic cracker. In research methods, using univariate ARIMA model to predict reactor pressure differential trends, and the various input variables and output variables conversion function ARIMA transfer function model predictive models and causality.
This study by the univariate ARIMA model predictive capability with moderate R2 value of 0.936, can be regarded as an excellent prediction model; each transfer function model has the same reach of the R2 value greater than 0.936. In addition to sodium no significant effect on the amount of differential pressure in the reactor, the remaining four input variables Individually significant impact on output variables, and input variables most influence upon the number of current or delayed.
|
|---|
