Sensorless Position Estimator for Reciprocating Linear Actuator Hardware

This Phase I project aims to demonstrate the effectiveness of a new sensorless piston position estimation technique for reciprocating linear actuators. During this six-month SBIR Phase I project, we propose constructing a model-based dynamic piston position observer in software and validating its performance using moving-coil linear actuator hardware. Our sensorless position estimator approach holds the potential to enhance the reliability and performance of NASA Radioisotope Power Systems (RPS) missions by supporting controllers that: 1) adjust the piston amplitude in response to radioisotope fuel decay or when individual redundant multi-converter units are activated or deactivated 2) identify and correct for piston mean position offset (“drift”) 3) coordinate the piston phase and amplitude of dual-opposed converters to minimize net vibration 4) manage single converter piston motion for optimal use of active/ passive vibration absorbers We understand that many of the convertor controllers developed by NASA already monitor coil current and terminal voltage. Therefore, controllers that contain microprocessors could, in principle, be upgraded to host our position estimation algorithms without requiring additional sensors or current injection hardware, thus enhancing overall system reliability. Furthermore, utilizing information about piston position (including amplitude, offset, and phase) could improve specific power density by allowing the controller to operate safely closer to the physical limits of piston travel while synchronizing with other converters and vibration absorbers. It may be feasible to implement our sensorless position observer for the linear actuator of an active (driven) vibration absorber, potentially eliminating the need for a separate accelerometer. Our piston position observer is suitable for moving-magnet-based converters and the low-power, low-inductance moving coil alternator-based converter developed by NASA-GRC.