by Maria E. Barrera-Medrano
Centrifugal compressors are widely used in industrial applications. Although industrial compressors are always equipped with state of the art filtration systems, due to their industrial location and large mass flows, these machines can end up ingesting contaminants in the forms of dust, dirt, soot, salt and other types of particulate matter. After a prolonged time of operation, these contaminants end up fouling the blades, ducts and internal walls of the compressor, resulting in a loss of performance. Compressor performance is also dependent on inlet ambient conditions such as pressure and temperature, while fouling is a slow growing, non-uniform and non-linear mechanism, so finding the source of performance reduction is a challenging task.
Due to the lack of information about the fouling phenomenon affecting these machines, the aim of this research is to characterize the fouling grade existing in the compressor and its effect on the compressor performance parameters. To reach this goal, three parallel ways will be used in order to get a correlation that describes the centrifugal compressor status comparing the performance parameters variation over different fouling grades.
The methodology proposed combines three different paths of data obtaining, which validate one to each other: compressor performance map generation, which is able to evaluate graphically the parameters variation under different fouling grades; 1D mean-line development, in order to obtain the performance parameters of the compressor; and centrifugal compressor testing, which is the source of experimental data obtaining under different deterioration levels.
This research is part of the Energy SmartOps FP7 European Community program, composed of 13 early stage researchers, with the main aim of reducing the power consumption and enhances energy savings in large scale industrial and process plants. This project will contribute to the overall energy reduction by being able detect and evaluate when fouling is present and its effects on the compressor performance parameters, defining when the compressor should be cleaned in order to make it operate at its maximum efficiency. The performance prediction ability of the model will be used by other researchers in the program to optimize and control the operations of compressor networks in a plant.