Development And Application Of A Vertical High Speed Motor-Compressor Simulator For Rotor Drop Onto Auxiliary Bearings.

During the conceptual development of a subsea motor-compressor prototype running on active magnetic bearings, the emergency landing on the catcher bearings was identified as a potential technology gap. This is because the rotor in object is vertical and supercritical. A subscale simulator was then developed with the twofold aim to collect useful experimental data for data matching with a numerical simulation (also developed for this special application) and to perform endurance testing of the catcher bearings to understand their mechanical limits. The selected scale factor was one-third in order to utilize standard size active magnetic bearings and meet the project schedule. The subscale rotor design was aimed to respect the dynamic similitude with the real machine having the same separation between the rotor critical speeds and the operating speed range. Maximum test rig rotor speed is 30,000 rpm. Moreover the test rig casing design simulated also the real casing dynamic behavior. The test campaign was then divided in two steps. During the first step a test matrix was defined to address all the potential important parameters in the landing dynamics: unbalance magnitude, unbalance distribution (to excite different modes), rotation speed at drop start, duration of the delevitation, type of axes to be delevitated and active magnetic bearings (AMBs) cooling medium. Overall 19 tests were performed. The main outcomes from the test were: • The rotor always showed a forward cylindrical subsynchronous whirl. • The whirl behavior was not affected by unbalance level and distribution, landing duration, or drop speed. • The whirl frequency is not related to any rotor/casing natural frequencies. The numerical simulation was developed in a numerical computing environment software specifically for the test rig system and was tuned on the basis of the experimental results. A key factor in model tuning was the introduction of a cross coupled term to allow the rotor to follow a forward whirling motion since the beginning of the drop. The physical nature of this cross-coupled force is still to be exactly identified. The final simulation results well matched with the experimental data both in whirling direction and frequency. Afterward the simulation was extended to the full-scale machine to predict landing behavior during the future full load testing and operation. Simulation shows the machine is stable for nominal conditions and the stability margin is greater than two. Moreover catcher bearing loads are within allowable limits. Finally a sensitivity study was performed showing that decreasing catcher bearings clearances and reducing seal entry swirl will lead to improved stability margin and dynamic loads. A second step of testing was the endurance test of the catcher bearings to determine the maximum number of landing. For this purpose the same drop speed (30 krpm) and landing duration (10 sec) was used for all the test runs. Several catcher bearings damage indicators were experienced; finally it was evident that the most reliable indications came from the accelerometers mounted on the bearing housing. The endurance test led to the conclusion that the current design of catcher bearings for this rotor configuration allows at least seven safe drops from full speed in conditions, which can be considered similar or even more severe than the real machine scenario.