MODELLING AND SIMULATION OF PERFORMANCE MAP FOR VARIABLE GEOMETRY AXIAL COMPRESSOR

One of the equipment used in oil and gas processing is gas turbine. In Malaysia, an estimated 45% of the electric power is generated by using gas turbines. A gas turbine has three main components: axial compressor, combustion chamber and turbine. The objective of this project is to develop regression models for performance maps of axial compressors. Support vector regression method available in MATLAB is applied to develop the models. Three multi-stage axial compressors are used for the case study: (i) three stage transonic compressor, (ii) five stage low aspect ratio compressor, and (iii) ACP 400 compressor. In all the studied cases, the SVM models demonstrated low root mean squared errors. The project also included a discussion on the effect of variable geometry and fouling to the performance maps. The developed models are useful for further investigation of gas turbines performance. One limitation of the present work, however, is that it needs further refinement to account for low speed behavior and the effect of bleed flow.