Influence of control mechanisms on the performance map of centrifugal compressors

Thermal energy storage (TES) systems convert excess electrical energy into heat typically via electric heaters or heat pumps and store it in the thermal media for later reconversion to electricity via heat to power process. In high performance TES and grid balancing applications, centrifugal(radial) compressors charge the heat pumps, requiring operation across a wide range of mass flow, pressure ratios and power levels. These off-design conditions lead to deviations in compressor efficiency, mass flow, and pressure rise as captures by the compressor performance map. In grid balancing applications, radial compressors drive heat pumps and must operate across a wide range of mass flows, pressure ratios and power levels, leading to off-design performance shifts captured in compressor maps. This work leverages a custom Python tool inverse compressor design tool for the inverse preliminary design of centrifugal compressors. Starting from the specified inlet conditions and a target pressure ratio, the software selects an initial impeller geometry via the Cordier diagram, refines it through one dimensional (mean line) thermodynamic and velocity triangle analysis, and predicts performance. It interfaces with REFPROP for accurate fluid properties and provides modules for geometry generation, efficiency estimation, cost assessment, and performance map construction. Two control mechanisms in which variable rotational speed and recirculation are implemented within this framework to simulate their individual and combined effects on mass flow, pressure ratio, and isentropic efficiency. Parametric studies generate comprehensive compressor maps for baseline, each mechanism alone and their combination. The results quantify how speed variation shifts the map and how bleed mitigates surge, as well as their synergistic influence on the operating envelope. These insights yield design guidelines for matching compressor capabilities to dynamic grid load profiles.