engleski

Software analyzer application in refinery hydrodesulfurization unit

Necessity to produce high quality fuels in order to meet environmental regulations requires continuous monitoring, maintaining and product quality improvement. To meet those market needs, refining processes must be equipped with satisfactory number of process analyzers. Over the years, there has been significant number of software analyzers present in process industry, first of all due to their price accessibility. Software analyzers are based on application of a mathematical model in order to continuously, on-line, predict physical and chemical properties or product content. Thus, they can be applied to provide additional analysis to those made by hardware analyzer or laboratory analysis, or can be used as independent analyzer. Their particular importance is in improving process monitoring and control, especially in model predictive control applications. In this paper, practical application of software sensor based on artificial neural networks was explored on exact problem of sulfur content in hydrodesulfurization unit product. Results of a research were applied in practice.

soft sensor, hydrotreatment, neural networks

nije evidentirano