System and method for controlling a vapor compression system with safe actuator changes

A method for controlling an operation of a vapor compression system (VCS), the method comprising:

• collecting data points indicative of control of the operation of the VCS with different combinations of setpoints for different actuators of the VCS and corresponding costs of operation of the VCS for each of the different combinations of setpoints;
• computing, using a Local Search Region Bayesian optimization (LSR-BO) of the combinations of setpoints and their corresponding costs of operation, a probabilistic surrogate model providing a probabilistic mapping between various combinations of setpoints for different actuators of the VCS and their corresponding costs of operation, wherein the probabilistic surrogate model defines at least first two order moments of the cost of operation in the probabilistic mapping;
• selecting from the probabilistic surrogate model a current combination of setpoints improving the cost of operation with respect to a previous combination of setpoints used to control the actuators of the VCS at a previous control step, according to an acquisition function of the first two order moments of the cost of operation subject to a LRS constraint restricting a distance between the current combination of setpoints and the previous combination of setpoints while allowing an incremental search for a global minimum at the probabilistic surrogate model;
• submitting the current combination of setpoints to a feedback controller to control the actuators of the VCS at a current control step according to their corresponding setpoints in the current combination of setpoints to collect a current cost of operation;
• and
• updating the collected data points with the current combination of setpoints and the corresponding current cost of operation.

A non-transitory computer readable storage medium embodied thereon a program executable by a processor for performing a method for controlling an operation of a vapor compression system (VCS), the method comprising:

• collecting data points indicative of control of the operation of the VCS with different combinations of setpoints for different actuators of the VCS and corresponding costs of operation of the VCS for each of the different combinations of setpoints;
• computing, using a Local Search Region Bayesian optimization (LSR-BO) of the combinations of setpoints and their corresponding costs of operation, a probabilistic surrogate model providing a probabilistic mapping between various combinations of setpoints for different actuators of the VCS and their corresponding costs of operation, wherein the probabilistic surrogate model defines at least first two order moments of the cost of operation in the probabilistic mapping;
• selecting from the probabilistic surrogate model a current combination of setpoints improving the cost of operation with respect to a previous combination of setpoints used to control the actuators of the VCS at a previous control step, according to an acquisition function of the first two order moments of the cost of operation subject to a LRS constraint restricting a distance between the current combination of setpoints and the previous combination of setpoints while allowing an incremental search for a global minimum at the probabilistic surrogate model;
• submit the current combination of setpoints to a feedback controller to control the actuators of the VCS at a current control step according to their corresponding setpoints in the current combination of setpoints to collect a current cost of operation;
• and
• updating the collected data points with the current combination of setpoints and the corresponding current cost of operation.