Detecting anomalies in a plurality of showcases

1. A method for failure prevention and mitigation of a plurality of showcases in a structure based on detecting anomalies in the plurality of showcases, comprising:

receiving a data stream from a sensor network including a plurality of sensors embedded in the plurality of showcases and a plurality of external environmental sensor disposed on heating, ventilation, and air conditioning (HVAC) system components in the structure and obtaining a corresponding table between each of the plurality of showcases, each of the HVAC components, and at least one corresponding sensor selected from the plurality of sensors, and at least one external environmental sensor;

obtaining information for showcase and HVAC system component clustering;

determining, by a processor device, at least one cluster of showcases and HVAC system components based on the information for showcase and HVAC system component clustering and the corresponding table between each of the plurality of showcases, each of the HVAC components, the at least one corresponding sensor selected from the plurality of sensors, and the at least one external environmental sensor;

building at least one model for each of the at least one cluster of showcases and HVAC system components based on historical performance data of the plurality of showcases, performance data for showcases from at least one different location, expected customer behavior, and additional data related to the structure, the additional data including a time of day at which the data stream is measured and seasonal location information; and

detecting at least one anomaly based on data from the at least one cluster of showcases and HVAC system components and the at least one model; and

automatically mitigating faults identified by the at least one anomaly by redirecting refrigeration or shutting down one or more devices in the at least one cluster associated with the at least one anomaly upon detection of the at least one anomaly.

2. The method as recited in claim 1 , wherein detecting the at least one anomaly further comprises:

determining whether the at least one anomaly is found in data from the at least one cluster using reconstruction error of the at least one model.

3. The method as recited in claim 1 , further comprising:

generating an alert in response to detecting the at least one anomaly.

4. The method as recited in claim 1 , further comprising:

implementing an action to correct the at least one anomaly in response to detecting the at least one anomaly.

5. The method as recited in claim 1 , wherein the information for showcase and HVAC system component clustering implies at least one of similarity or dependency between each showcase or HVAC system component in the at least one cluster of showcases and HVAC system components.

6. The method as recited in claim 1 , wherein building the at least one model further comprises:

building the at least one model to model dependency between attributes of the at least one cluster of showcases and HVAC system components.

7. The method as recited in claim 1 , wherein detecting the at least one anomaly further comprises:

monitoring the at least one cluster of showcases and HVAC system components based on the at least one model; and

generating ranking scores for the plurality of sensors embedded in the plurality of showcases and the plurality of external environmental sensors disposed on HVAC system components, the ranking scores indicating which of the plurality of sensors embedded in the plurality of showcases and the plurality of external environmental sensors disposed on HVAC system components provide a most relevant information regarding the faults.

8. The method as recited in claim 1 , wherein the information for showcase and HVAC system components clustering comprises at least one of:

data driven clustering information;

topology driven clustering information; or

product similarity-based clustering information.

9. The method as recited in claim 1 , wherein building the at least one model for each of the at least one cluster of showcases and HVAC system components further comprises applying at least one of:

a long short-term memory (LSTM) auto-encoder,

invariant relationships, and

a principal component analysis (PCA) based method.

10. The method as recited in claim 1 , wherein the data from the at least one cluster of showcases and HVAC system components further comprises:

multi-variate time series data.

11. The method as recited in claim 1 , further comprising:

sharing at least one non-unique attribute among the at least one model.

12. A computer system for failure prevention and mitigation of a plurality of showcases in a structure based on detecting anomalies in the plurality of showcases, comprising:

a processor device operatively coupled to a memory device, the processor device being configured to:

receive a data stream from a sensor network including a plurality of sensors embedded in the plurality of showcases and a plurality of external environmental sensors disposed on heating, ventilation, and air conditioning (HVAC) system components in the structure and obtain a corresponding table between each of the plurality of showcases, each of the HVAC components, at least one corresponding sensor selected from the plurality of sensors, and at least one external environmental sensor;

obtain information for showcase and HVAC system component clustering;

determine at least one cluster of showcases and HVAC system components based on the information for showcase and HVAC system component clustering and the corresponding table between each of the plurality of showcases, each of the HVAC components, the at least one corresponding sensor selected from the plurality of sensors, and the at least one external environmental sensor;

build at least one model for each of the at least one cluster of showcases and HVAC system components; and

detect at least one anomaly based on data from the at least one cluster of showcases and HVAC system components and the at least one model based on historical performance data of the plurality of showcases, performance data for showcases from at least one different location, expected customer behavior, and additional data related to the structure, the additional data including a time of day at which the data stream is measured and seasonal location information; and

automatically mitigate faults identified by the at least one anomaly by redirecting refrigeration or shutting down one or more devices in the at least one cluster associated with the at least one anomaly upon detection of the at least one anomaly.

13. The system as recited in claim 12 , wherein, when detecting the at least one anomaly, the processor device is further configured to:

determine whether the at least one anomaly is found in data from the at least one cluster using reconstruction error of the at least one model.

14. The system as recited in claim 12 , wherein the processor device is further configured to:

generate an alert in response to detecting the at least one anomaly.

15. The system as recited in claim 12 , wherein the processor device is further configured to:

implement an action to correct the at least one anomaly in response to detecting the at least one anomaly.

16. The system as recited in claim 12 , wherein the information for showcase and HVAC system component clustering implies at least one of similarity or dependency between each showcase or HVAC system component in the at least one cluster of showcases and HVAC system components.

17. The system as recited in claim 12 , wherein, when building the at least one model, the processor device is further configured to:

build the at least one model to model dependency between attributes of the at least one cluster of showcases and HVAC system components.

18. The system as recited in claim 12 , wherein, when detecting the at least one anomaly, the processor device is further configured to:

monitor the at least one cluster of showcases and HVAC system components based on the at least one model; and

generating ranking scores for the plurality of sensors embedded in the plurality of showcases and the plurality of external environmental sensors disposed on HVAC system components, the ranking scores indicating which of the plurality of sensors embedded in the plurality of showcases and the plurality of external environmental sensors disposed on HVAC system components provide a most relevant information regarding the faults.

19. The system as recited in claim 12 , wherein the information for showcase and HVAC system components clustering comprises at least one of:

data driven clustering information;

topology driven clustering information; or

product similarity-based clustering information.

20. A computer program product for failure prevention and mitigation of a plurality of showcases in a structure based on detecting anomalies in the plurality of showcases, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing device to cause the computing device to perform the method comprising:

receiving a data stream from a sensor network including a plurality of sensors embedded in the plurality of showcases and a plurality of external environmental sensors disposed on heating, ventilation, and air conditioning (HVAC) system components in the structure and obtaining a corresponding table between each of the plurality of showcases, each of the HVAC components, at least one corresponding sensor selected from the plurality of sensors, and at least one external environmental sensor;

obtaining information for showcase and HVAC system component clustering;

determining, by a processor device, at least one cluster of showcases and HVAC system components based on the information for showcase and HVAC system component clustering and the corresponding table between each of the plurality of showcases, each of the HVAC components, the at least one corresponding sensor selected from the plurality of sensors, and the at least one external environmental sensor;

building at least one model for each of the at least one cluster of showcases and HVAC system components based on historical performance data of the plurality of showcases, performance data for showcases from at least one different location, expected customer behavior, and additional data related to the structure, the additional data including a time of day at which the data stream is measured and seasonal location information;

detecting at least one anomaly based on data from the at least one cluster of showcases and HVAC system components and the at least one model; and

automatically mitigating faults identified by the at least one anomaly by shutting down one or more devices in the at least one cluster associated with the at least one anomaly upon detection of the at least one anomaly.