During recent years, microfluidics based microelectromechanical systems (MEMSs) have found multiple applications in biomedical engineering. One of their most important implementations is fluid transfer in microliter and nanoliter scales. Nowadays, micropumps are extensively used in various medical applications such as drug delivery. In this study, the performance of a piezoelectric micropump is investigated and optimized. This micropump consists of a pump chamber and three deformable walls in a nozzle-diffuser shape, which are used to create pressure gradient between the inlet and outlet. The performance of the micropump is evaluated by transient Computational Fluid Dynamics (CFD) simulation using dynamic mesh. Then its performance is optimized using the Design of Experiment (DOE) method based on mean net outlet mass flow rate and flow reversibility at the pump outlet. The results indicate an improvement of 34.5% in mean net outlet mass flow rate and a significant decrease in reversibility. The maximum mean net outlet mass flow rate and the minimum reversibility corresponding to the optimum geometries are 95.82 mL/min and 0.05%, respectively.

In centrifugal compressors, variable inlet guide vanes (IGVs) are used to control the mass flow rate with negligible change in pressure ratio and shaft speed. The efficient operation of IGVs is limited to the range of aerodynamic stability of their vane profiles. Thin symmetric profiles, which are usually used, tend to exhibit flow separation already at low setting angles, resulting in a rapid decrease of the global efficiency. Three different guide vane profiles including a symmetric reference, a two-piece tandem profile with variable angle, and an s-cambered, together with two auxiliary profiles were considered. These variants were numerically and experimentally investigated at different setting angles, covering negative and positive inlet swirl and at different operating points. The results show that tandem and s-cambered vanes have superior aerodynamic performance in comparison with the symmetric reference profile and extend the operating range of the compressor over which the efficiency remains high. The tandem profile shows better aerodynamic performance than the s-cambered profile at negative inlet swirl, while the asymmetric s-cambered profile offers an extended range of high efficiency at positive preswirl. Considering manufacturing requirements and cost, the s-cambered profile is preferred for applications with positive preswirl.

Supercritical carbon dioxide (SCO2) cycle has drawn attention for extracting energy and generating power from low grade heat resources due to its higher efficiency, lower costs and compactness. Combining the SCO2 cycle with a gas turbine is a promising field of research that has applications in industries as waste heat recovery systems. In this research, three different configurations of power generation SCO2 cycles includes Recuperation, Precompression, and Reheating were thermodynamically designed and modeled. Extracting the main parameters of cycles, and appropriate objective which is cycle thermal efficiency, up to 80,000 data constructed. Random forest (RF) and support vector regression (SVR), Machine Learning techniques, are used for performance prediction of cycles. Results show the strength of RF method to predict the cycle efficiency with up to 3% error and the SVR method with the precision more than 90%. Particle Swarm Optimization algorithm is employed for optimizing the Recuperation cycle. The outcomes of the optimization model based on the prediction models as the proxy have nearly close results compared with traditional optimization, which is based on thermodynamic model. In conclusion, the ML models can be used for prediction and maintenance since the methods are up to 10 times faster than thermodynamic models.