Utilization of a linguistic response surface methodology to the business strategy of polypropylene in an Indian petrochemical plant

Abstract

Polypropylene is a multi-reason thermoplastic resin with ample of scope for engineering applications. This article presents a very distinct and new methodology called linguistic Response Surface Methodology (RSM) for predicting the quality of polypropylene used in petrochemical industries. This model is framed on the basis of a huge quantity of data obtained from well-known Indian chemical factories. The quality of polypropylene depends on factors such as melt flow index and solubility of the product in xylene. The parameters controlling both the factors are the hydrogen flow, the donor flow, and, the pressure and temperature of the polymerization reactor. Using these four input and output parameters, a response surface method based on the different features of the variable is created. Then the simulation results are compared with the collected plant information and the most appropriate model is selected through a series of fact-finding. This linguistic surface response method, when compared with Type 1, and Type 2 fuzzy logic, was found to provide the maximum value of R. An ANNOVA analysis of the experimental data was done to determine the reliability, the repetition, and the efficiency of the constructed regression models. This method will not only save the expenses but also the time, generally being taken by the other models.

Keywords

• Polypropylene
• Product quality
• Linguistic response surface methodology
• ANNOVA Analysis