Modeling Hydraulic Accumulators for use in Wind Turbines

Abstract

One of the major challenges in numerical simulation of hydraulic systems, is the long com- putation times of accumulators. In wind power applications, the accumulators must provide necessary hydraulic energy during emergency stops, while the weight of hydraulic equipment must be minimized. Therefore, precise and efficient design tools for accumulators are essen- tial. This paper addresses the issue of understanding the dynamic phenomena in piston type ac- cumulators, and how this leads to improved numerical accumulator models. Developing a numerical model has two challenges. First, the unsteady heat transfer between the accumu- lator gas and the wall has to be described. Secondly, a suitable real gas model has to be identified and employed. To verify the model, a series of experiments were conducted at Fritz Schur Energy in Glostrup, Denmark. The experiments were designed to investigate gas dynamical properties at various precharge pressures, maximum pressures, and ambient temperatures. These parameters were varied to obtain parameter independent conclusions. During controlled piston movements, hydraulic and gas pressures were measured together with the piston position and the gas temperature. It was found that the simple thermal time constant approximation by Rotthäuser was suitable and stable for the application. It was also found that the Soave-Redlich-Kwong equation was overall best suited with experimental data. The Soave-Redlich-Kwong equation is as much as six times faster than the widely used Benedict-Webb-Rubin equation, independent of the ambient temperature, maximum pressure and precharge pressure. This project concludes that the Soave-Redlich-Kwong equation should be used in simulation of piston type accumulators. It is noted that the pressure available after expansion is some 5-10 bar lower than predicted by any model. Therefore, it is suggested that more research is conducted to obtain an improved model for the heat transfer. Finally, measurement qual- ity was confirmed by comparing measured pressure data by pressure calculations based on measured temperature.