Demand defrost

1 . A controller configured to be operable for:

triggering a defrost cycle when a gap between Control Temperature (CT) and Saturated Suction Temperature (SST) exceeds a defined threshold of an expected/learned gap between the Control Temperature (CT) and Saturated Suction Temperature (SST);

following the defrost cycle, executing a machine learning process for a determined amount of time during which the controller learns a current expected/learned gap between the Control Temperature (CT) and Saturated Suction Temperature (SST); and

averaging the current expected/learned gap with a previous expected/learned gap, which was machine learned by the controller following a previous defrost cycle, to thereby provide an average expected/learned gap usable by the controller as the previous expected/learned gap following the next defrost cycle.

2 . The controller of claim 1 , wherein the controller is configured to be operable for smoothing values of the Control Temperature (CT) and Saturated Suction Temperature (SST) on a provisional basis when the controller is machine learning the current expected/learned gap between the Control Temperature (CT) and Saturated Suction Temperature (SST).

3 . The controller of claim 1 , wherein the controller is configured to be operable for:

checking guardrails for a violation if the gap between the Control Temperature (CT) and Saturated Suction Temperature (SST) does not exceed the defined threshold of the expected/learned gap between the Control Temperature (CT) and Saturated Suction Temperature (SST); and

triggering a Safeguard Defrost if the guardrails are in violation.

4 . The controller of claim 1 , wherein the controller is configured to be operable with a dynamic pulldown threshold, and to adjust a maximum allowed time between defrosts based on the dynamic pulldown threshold.

5 . The controller of claim 4 , wherein the controller is configured to determine the maximum allowed time between defrosts by dividing twenty-four hours by a scheduled number of defrosts and then subtracting the amount of time already spent in pulldown.

6 . The controller of claim 1 , wherein the controller is configured to fallback on an original predetermined scheduled frequency of defrost cycles when one or more defined criteria are satisfied.

7 . The controller of claim 1 , wherein the controller is configured to be operable to fallback on an original predetermined scheduled frequency of defrost cycles based on an evaluation that indicates ineffective defrosting.

8 . The controller of claim 1 , wherein the controller is configured to be operable for establishing, via machine learning, the expected/learned gap between the Control Temperature (CT) and Saturated Suction Temperature (SST).

9 . The controller of claim 1 , wherein:

the controller is configured to be operable for defining, via machine learning, the defined threshold of the expected/learned gap between the Control Temperature (CT) and Saturated Suction Temperature (SST); and/or

the controller is configured to be operable for defining, via data analysis, an optimal threshold as the defined threshold of the expected/learned gap between the Control Temperature (CT) and Saturated Suction Temperature (SST); and/or

the controller is configured to be operable for allowing a user input of the defined threshold of the expected/learned gap between the Control Temperature (CT) and Saturated Suction Temperature (SST).

10 . The controller of claim 1 , wherein the controller is configured to be operable for:

monitoring the Control Temperature (CT) and Saturated Suction Temperature (SST);

determining the actual gap in real time between the monitored Control Temperature (CT) and Saturated Suction Temperature (SST); and

triggering a defrost cycle when the actual gap in real time between the monitored Control Temperature (CT) and Saturated Suction Temperature (SST) exceeds the threshold of the expected/learned gap between the Control Temperature (CT) and Saturated Suction Temperature (SST) for a predetermined duration.

11 . The controller of claim 1 , wherein the Control Temperature (CT) is Return Air Temperature (RAT).

12 . The controller of claim 1 , wherein the controller is configured to be operable for determining a hierarchal order of which of a plurality of cases on a rack that cannot all defrost simultaneously are allowed to defrost.

13 . The controller of claim 1 , wherein:

the controller is a microprocessor-based controller usable in controlling temperature, superheat, and defrost in a refrigerated fixture, walk-in box, walk-in freezer, walk-in cooler, or other refrigerated space, and/or

the controller is configured to be operable for controlling initiation of a defrost cycle of an evaporator heat exchanger of a refrigeration system operatively associated with a refrigerated space.

14 . A method comprising:

triggering a defrost cycle when a gap between Control Temperature (CT) and Saturated Suction Temperature (SST) exceeds a defined threshold of an expected/learned gap between the Control Temperature (CT) and Saturated Suction Temperature (SST);

following the defrost cycle, executing a machine learning process for a determined amount of time during which the method learns a current expected/learned gap between the Control Temperature (CT) and Saturated Suction Temperature (SST); and

averaging the current expected/learned gap with a previous expected/learned gap, which was machine learned by the method following a previous defrost cycle, to thereby provide an average expected/learned gap usable by the method as the previous expected/learned gap following the next defrost cycle.