Sensorless Resonance Tracking and Stroke Control of a Linear Vapor Compressor Via Nonlinear Observers

Linear compressors have become more common in refrigeration applications due to their high efficiency. However, in order to achieve efficient and reliable operation of the machine, the system controller must be capable of dealing with two main issues. First, the nonlinear force of gas compression acting against the mass spring piston system causes the frequency at which system efficiency is maximized, i.e., the resonant frequency, to vary with the load. Second, the unconstrained piston stroke allows for the possibility of collision with the head, which may reduce reliability. Accounting for these issues requires knowledge of the piston dynamics; however, for reliability and cost effectiveness it is preferred to operate the system without the use of internal sensors. In this work a novel system level controller is designed to simultaneously achieve resonance-tracking and stroke control objectives utilizing a pair of nonlinear observers to obtain the necessary information sensorlessly. This design is validated under experimental testing, which demonstrates both the accuracy of the observed signals and the stability and convergence of the controller.