MULTIVARIABLE CONTROL OF VAPOR COMPRESSION SYSTEMS

This paper presents the results of a study of multi-input multi-output (MIMO) control of vapor compression cycles that have multiple actuato rs and sensors for regulating multiple outputs, e.g. superheat and eva porating temperature. The conventional single-input single-output (SIS O) control was shown to have very limited performance. A low order lum ped-parameter model was developed to describe the significant dynamics of vapor compression cycles. Dynamic modes were analyzed based on the low order model to provide physical insight of system dynamic behavio r. To synthesize a MIMO control system, the Linear-Quadratic Gaussian (LQG) technique was applied to coordinate compressor speed and expansi on valve opening with guaranteed stability robustness in the design. F urthermore, to control a vapor compression cycle over a wide range of operating conditions where system nonlinearities become evident, a gai n scheduling scheme was used so that the MIMO controller could adapt t o changing operating conditions. Both analytical studies and experimen tal tests showed that the MIMO control could significantly improve the transient behavior of vapor compression cycles compared to the conven tional SISO control scheme. The MIMO control proposed in this paper co uld be extended to the control of vapor compression cycles in a variet y of HVAC and refrigeration applications to improve system performance and energy efficiency.