Systems and methods for learning data patterns predictive of an outcome

Abstract

Systems and methods for data collection in an industrial environment are disclosed. A plurality of input sensors communicatively coupled to a controller, a data collection circuit structured to collect output data from a plurality of input sensors, and a machine learning data analysis circuit structured to receive the output data and learn received output data patterns indicative of an outcome, wherein the machine learning data analysis circuit is structured to learn received output data patterns by being seeded with a model based on industry-specific feedback.

What is claimed is: 
     
         1 . A system for data collection in an industrial environment, comprising:
 a plurality of input sensors communicatively coupled to a controller;   a data collection circuit structured to collect output data from a plurality of input sensors; and   a machine learning data analysis circuit structured to receive the output data and learn received output data patterns indicative of an outcome, wherein the machine learning data analysis circuit is structured to learn received output data patterns by being seeded with a model based on industry-specific feedback.   
     
     
         2 . The system of  claim 1 , wherein the model is a physical model, an operational model, or a system model. 
     
     
         3 . The system of  claim 1 , wherein the outcome is one of an outcome of a process, an outcome of a calculation, an outcome of an event, an outcome of an activity. 
     
     
         4 . The system of  claim 3 , wherein the outcome is one of a list of outcomes consisting of: a reaction rate, a production volume, a quality performance, a component failure, and required maintenance. 
     
     
         5 . The system of  claim 1 , wherein the industry-specific feedback comprises at least one feedback value selected from a plurality of feedback values consisting of: a utilization measure, an efficiency measure, a measure of success in prediction or anticipation of states, a measure of success in avoidance of faults, a measure of success in mitigation of faults, a productivity measure, a yield measure, and a profit measure. 
     
     
         6 . The system of  claim 1 , wherein the industry-specific feedback comprises one of a power efficiency measure or a financial efficiency measure. 
     
     
         7 . The system of  claim 1 , wherein the machine learning data analysis circuit is further structured to learn received output data patterns based on the outcome. 
     
     
         8 . The system of  claim 1 , wherein the controller keeps or modifies one of operational parameters or equipment of the industrial environment. 
     
     
         9 . The system of  claim 1 , wherein the controller adjusts, based on at least one of: the learned received output data patterns, or the outcome, at least one of a weighting of the machine learning data analysis circuit and a number of data points collected from the plurality of input sensors. 
     
     
         10 . The system of  claim 1 , wherein the controller, based on at least one of: the learned received output data patterns, or the outcome, changes at least one of: a data storage technique for the output data, a data presentation mode and a data presentation manner. 
     
     
         11 . The system of  claim 1 , wherein the controller applies, to the output data, at least one filter selected from a plurality of filters comprising: a low pass filter, a high pass filter, and a band pass filter. 
     
     
         12 . The system of  claim 1 , wherein the controller performs one of removing under-utilized equipment or re-tasking under-utilized equipment based on at least one of: the learned received output data patterns, and the outcome. 
     
     
         13 . The system of  claim 1 , wherein the machine learning data analysis circuit comprises a neural network expert system. 
     
     
         14 . The system of  claim 1 , wherein the machine learning data analysis circuit is structured to learn received output data patterns indicative of one of progress or alignment with at least one of goals or guidelines. 
     
     
         15 . The system of  claim 1 , wherein the machine learning data analysis circuit is further structured to learn received output data patterns indicating at least one of:
 an unknown variable;   a preferred input among available inputs; and   a preferred input data collection band among a plurality of available input data collection bands.   
     
     
         16 . The system of  claim 1 , wherein the industry-specific feedback comprises at least one of: an amount of power generated by a machine about which the plurality of input sensors provide information during operation of the machine; a measure of an output of an assembly line about which the plurality of input sensors provide information; a failure rate of units of product produced by a machine about which the plurality of input sensors provide information; or a fault rate of a machine about which a plurality of input sensors provide information. 
     
     
         17 . The system of  claim 1 , wherein the industry-specific feedback comprises a power utilization efficiency of a machine about which the plurality of input sensors provides information, wherein the machine comprises at least one machine selected from a plurality of machines consisting of: a turbine, a transformer, a generator, a compressor, a machine that stores energy, and at least one power train component. 
     
     
         18 . The system of  claim 17 , wherein the industry-specific feedback comprises at least one of: a rate of extraction of a material by the machine; a rate of production of a gas by the machine; a rate of production of a hydrocarbon product by the machine; or a rate of production of a chemical product by the machine. 
     
     
         19 . A method for data collection in an industrial environment, the method comprising:
 receiving output data from a high number of sensors in an industrial environment;   seeding a machine learning circuit with a model based on performance measures; and   learning output data patterns indicative of an outcome from the received output data.   
     
     
         20 . The method of  claim 19 , further comprising adjusting at least one of a weighting of the model and a number of data points collected from the high number of sensors based on at least one of: the learned output data patterns or the outcome. 
     
     
         21 . The method of  claim 19 , further comprising changing at least one of: a data storage technique for the output data, a data presentation mode and a data presentation manner based on at least one of: the learned output data patterns or the outcome. 
     
     
         22 . The method of  claim 19 , further comprising filtering the output data using at least one filter selected from a plurality of filters consisting of: a low pass filter, a high pass filter, and a band pass filter.