Efficiency optimization and anomaly detection in hvac shelters for wireless communication facilities

A non-transitory machine-readable medium, comprising executable instructions that, when executed by a processing system including a processor, facilitate performance of operations, the operations comprising:

• retrieving an input dataset comprising energy consumption or cost data and categorical attributes related to a plurality of heating, ventilation and air conditioning (HVAC) shelters for wireless communication facilities of a wireless communication network;
• building a regression model using a robust model builder algorithm configured to generate predicted energy consumption or cost values representative of HVAC shelters maintained in good technical condition;
• applying the regression model to the input dataset to calculate, for each HVAC shelter, a predicted energy consumption or cost value and to compute a residual value for each HVAC shelter, forming a set of residual values;
• identifying anomalous values among the set of residual values, wherein the anomalous values indicate an abnormal energy consumption or cost profile;
• and
• generating an alert for a selected HVAC shelter based on the anomalous values.

A method, comprising:

• retrieving, by a processing system including a processor, an input dataset comprising energy consumption or cost data and categorical attributes for a plurality of heating, ventilation and air conditioning (HVAC) shelters of infrastructure equipment configured to provide wireless communication service on a wireless communication system;
• applying, by the processing system, a regression model to the input dataset to calculate predicted energy consumption or cost values for each HVAC shelter of the plurality of HVAC shelters;
• computing, by the processing system, a set of residual values, each respective residual value corresponding to a variance between a predicted energy consumption or cost value for a respective HVAC shelter, as calculated by the regression model, and a respective energy consumption or cost data value of the input dataset;
• identifying, by the processing system, anomalies among the plurality of HVAC shelters, wherein the identifying the anomalies is based on the set of residual values;
• and
• generating, by the processing system, an alert for HVAC shelters or groups of HVAC shelters corresponding to the anomalies.