Adaptive system monitoring using incremental regression model development

A regression based system monitoring processor, comprising:

• a processor;
• a memory communicatively coupled to the processor;
• a received data processor that, when operating, is configured to:
• accumulate, over a time duration, a first set of data comprising a set of output power values of an electrical generation system and associated environmental data values for the electrical generation system;
• determine at least one statistical relationship between output power values within the set of output power values and environmental data values within the associated environmental data values;
• determine, based on the at least one statistical relationship, outlying data within the first set of data;
• create a selected set of data by removing the outlying data from the first set of data;
• a regression model monitoring processor that, when operating, is configured to:
• develop, based on the selected set of data, a regression model mapping predicted inverter output power values to values of environmental data;
• receive, over a monitoring time duration that is subsequent to the time duration, monitored data comprising monitored inverter output power values of the electrical generation system and monitored associated environmental data for the electrical generation system;
• determine, based on the regression model and the monitored associated environmental data, predicted output power values of the electrical generation system corresponding to the monitored associated environmental data;
• identify an output power discrepancy based on a comparison of the predicted output power values to the monitored output power values;
• and
• providing a notification of an anomaly based on identification of the output power discrepancy.

A computer program product for monitoring an operational system, the computer program product comprising:

• a computer readable storage medium having computer readable program code embodied therewith, the computer readable program code comprising instructions for:
• accumulating, over a time duration, a first set of data comprising a set of inverter output power values of an electrical generation system and associated environmental data values for the electrical generation system;
• determining at least one statistical relationship between output power values within the set of output power values and environmental data values within the associated environmental data values;
• determining, based on the at least one statistical relationship, outlying data within the first set of data;
• creating a selected set of data by removing the outlying data from the first set of data;
• developing, based on the selected set of data, a regression model mapping predicted output power values to values of environmental data;
• receiving, over a monitoring time duration that is subsequent to the time duration, monitored data comprising monitored output power values of the electrical generation system and monitored associated environmental data for the electrical generation system;
• determining, based on the regression model and the monitored associated environmental data, predicted inverter output power values of the electrical generation system corresponding to the monitored associated environmental data;
• identifying an output power discrepancy based on comparing the predicted output power values to the monitored output power values;
• and
• providing a notification of an anomaly based on identification of the output power discrepancy.