Methods of Experimental Design in Modeling and Control Tasks Illustrated by Fuel and Energy Complex Facilities

Objective . To carry out experimental design for twoand three-factor cases using facilities of the fuel and energy complex as examples. Methods . An orthogonal central composite design was applied to improve model accuracy and account for nonlinear relationships among parameters. Results . The design parameters were calculated, planning matrices were constructed, and regression equations were obtained. A two-factor experiment was implemented, yielding adequate quadratic regression models for predicting flow rate and pressure in a liquid-transport system. A three-factor experiment was conducted based on a sectional model of primary oil refining. The factors affecting the criterion were: superheated steam flow rate, reflux ratio, and potential content of light fractions in the feed. MatLab and Python software packages were used. The adequacy of the obtained quadratic regression relationships to the original data was evaluated by Fisher’s F-test. Conclusions . The problem of maximizing liquid flow in the transport system was solved. The study of air injection into a circulating viscous liquid can serve as a convenient model for two-phase flow investigation. The application of the statistical model to oil rectification units contributes to improved operational efficiency of the installations.