Parameter Estimation and a Series of Nonlinear Observers for the System Dynamics of a Linear Vapor Compressor

Linear compressors are becoming more widely utilized in refrigeration for their high efficiency in comparison with traditional rotary devices. The nature of the free piston in a linear compressor requires additional control which necessitates knowledge of the mechanical dynamics of the system. Sensorless control of the system is preferred for cost effectiveness. To this end, a pair of nonlinear observers have been developed to observe signals within the system dynamics and reduce the need for costly sensors. The first utilizes a current and voltage measurements to arrive at an observed velocity. The second utilizes a velocity measurement to arrive at observations of piston acceleration as well as a lumped uncertainty term in the mechanical dynamics. A stability analysis is presented to prove boundedness and convergence for these nonlinear observers with experimental results providing further validation. As these observers rely on accurate knowledge of the model parameters, to this end, an estimator has been designed to estimate the electrical parameters of the system. Experimental results validate this method as well.