Drag Performance Degradation Due to Icing of eVTOL (Paper 1202)

Abstract

An empirical model was developed to estimate the liquid water content (LWC) and mean volume diameter (MVD) of super-cooled water droplets of an icing cloud based on the torque variation of a single rotor. The model was developed for eVTOL configurations that can actively monitor on-board torque via an electric speed controller (ESC). Experimental data was collected in the Penn State Adverse Environment Rotor Test Stand chamber under various icing conditions. From experimental data, a relationship between torque degradation and LWC was developed. This LWCis coupled to the FAR Appendix C Icing Envelopes to determine an MVD range of possible values. The range discrepancy is resolved by running Blade Element Momentum Theory (BEMT) and the Han-Palacios Correlation (HPC) for a given set of condition and finding the results that best match with the recorded torque. This process is repeated over the flying duration, using the actively incoming torque information from the ESC. Through the coupling of BEMT and HPC, torque degradation due to blade icing can be modeled and predicted. Being able to predict torque degradation means that the time until a critical torque is reached can also be estimated. This estimated time is then relayed to the vehicle as the available flying time in the current cloud environment. The proposed model was verified by comparing experimental data to the predicted values. The BEMT prediction deviated from the constant measured torque value by 6.0%. For HPC verification, 11 experimental cases were compared to estimated results. On average HPC deviated from the experimental cases by 14.8% at 20 MVD and 16.6% at 40 MVD. With the model verified, a blind study was performed using this method. A case that emulates possible flying conditions in a realistic environment was chosen. Torque data was acquired during the icing event. Within 90 seconds of data available during testing, the MVD and LWC were predicted within 1.39% and 20% respectively. Every 10 seconds the prediction for 10 more seconds into the future was compared to the 'blind' experimental data. On average, the prediction overestimated the torque by 6.5%. Based on the final calculations it was determined that 5.25 minutes of flying time was available to this particular rotor operating within this cloud environment. The presented effort introduces and verifies the capability to determine icing cloud conditions and available flight time available based on the proposed approach.