| Summary: | The Revolving Vane (RV) compressor mechanism was introduced in 2006. Since then, the mechanism has been studied, developed and improved. However, it has never been investigated for expander applications. It is the main objective of this project to develop an expander with the mechanism.
In order to achieve the aforementioned objective, the RV expander was first analysed by developing a comprehensive mathematical model. The model consists of two main parts, namely the working fluid processes and the solid components dynamics. The first part of the model includes the thermodynamics, the fluid dynamics and the heat transfer of the working fluid. The internal leakage models were also included. The second part includes the geometrical, the kinematics and the dynamics models of the main components. The endface effects and the journal dynamics models were also integrated. Computer simulation codes were then developed to numerically simulate the working principles of the expander. It was found that, among others, the output torque of the expander is a function of the inertia of the driven component, the pressure differences across the vane and the frictional losses. In the range of operating conditions tested, the friction losses were mainly dominated by the losses at the vane side (70%) and the shaft bearings (21%). It was also found that the expander internal leakage is mainly through the radial clearance (79%).
The RV mechanism has at least five design variants. Mathematical models for the other four variants have also been developed in this project. Comparisons between the variants suggest that the design where the vane is attached to the rotor and the rotor is used as the driving component (the RV-I design) is the most desirable design. It has the highest mechanical efficiency of up to 96.5% and the output torque profile is the most stable. Based on this finding, an RV-I prototype was then designed and manufactured. This is believed to be the first RV-I expander ever built in the world. The suction valve of the prototype was designed to function with air at a suction pressure of 3 barg and a discharge pressure of 0 barg. Experimental studies were then carried out at suction pressures of up to 6 barg and operational speeds of up to 600 rpm. The prototype was operated without any significant problem. It was also found that the isentropic efficiency is up to 32%. The volumetric efficiency is only up to 26% when operated in the design condition due to the larger than expected manufacturing clearances. Verifications of the theoretical models show that the theoretical model can simulate the average torque data with a standard deviation of mostly within 8%.
Another prototype was also developed based on the RV-0 design, where the vane was allowed to swivel in the cylinder and to slide relative to the rotor. This is believed to be the first RV-0 expander prototype in the world. It was found that it can successfully operate without any significant operational problem in the range of the operational pressures (suction pressures of up to 5 barg and a fixed discharge pressure of 0 barg) and speeds (up to 900 rev/min) tested. The expander exhibited a volumetric efficiency of up to 77%. From the static leakage test, the main leakage paths were found to be through the housing and the radial clearances, accounting for around 50% and 40% of the total leakage, respectively. The test also showed that the leakage through the radial clearance accounted for 67% to 93% of the total internal leakage of the expander, depending on the operational pressure. The isentropic efficiency of the prototype was found to be up to 17.5%. Comparisons between the theoretical predictions and the experimental data show that the mathematical models can predict the average torque and flow rate within average standard deviations of 20% and 11%, respectively.
Finally, an improved dynamics model was proposed. This model is based on the dynamic interactions of the forces and the torques, allowing the modelling of the dynamics and impacts between the vane and the vane slot as the expander operates. It allows the observation of previously unobservable phenomena. These include, among others, the interaction between the vane and the vane slot and its effect on the thermodynamic processes of the working fluids and to the cylinder dynamics. It was also found that the selection of the impact time is crucial to the accuracy of the model.
|
|---|