Methods and systems for energy prognosis

An energy management system comprising:

• an energy consuming machine for producing a product;
• an energy sensor communicably coupled to the energy consuming machine wherein the energy sensor generates an energy related test signal comprising real power test data and reactive power test data indicative of energy consumed by the energy consuming machine;
• a processor communicably coupled to the energy sensor, wherein the processor executes machine readable instructions to:
• receive the energy related test signal;
• partition the energy related test signal into an energy related test sub-signal;
• apply time domain functions and/or frequency domain functions to the energy related test sub-signal to extract a feature set of data from the energy related test sub-signal;
• transform the feature set of data with a principle component analysis into a reduced test feature set;
• compare the reduced test feature set to training data indicative of a normal operating condition of the energy consuming machine;
• generate a power-based performance index with a performance assessment algorithm, wherein the power-based performance index is based at least in part upon a difference between the reduced test feature set and the training data;
• and
• generate a fault indication based at least in part upon the power-based performance index.

A toolkit for managing an industrial process, wherein the toolkit is stored on a machine readable medium and comprises machine readable instructions to transform an energy related test signal into an energy prognosis, the machine readable instructions comprising:

• a signal loading algorithm to receive the energy related test signal comprising real power test data indicative of electrical energy consumed by an energy consuming machine;
• a partitioning algorithm to transform the energy related test signal into an energy related test sub-signal;
• a feature extraction algorithm to transform the energy related test sub-signal into a feature set of data from the energy related test sub-signal;
• a dimensionality reduction algorithm to transform the feature set of data into a reduced test feature set;
• a performance assessment algorithm for comparing the reduced test feature set to a baseline indicative of a normal operating condition of the energy consuming machine;
• a performance prediction algorithm for generating the energy prognosis that forecasts a future health condition of the energy consuming machine based at least in part upon the reduced test feature set;
• and
• a visualization algorithm for displaying the energy prognosis.