Methods and systems for inferred information propagation for aircraft prognostics

Abstract

Methods and systems are provided for inferred information propagation for aircraft prognostics. The method includes receiving, by a processor, an original time-series of data points for a component as an input; preprocessing the input to divide the original time-series of data into subsets of data by applying a time-window over the original time-series of data points; and computing, by the processor, a Mutual Information (MI) value for each pair of variables within each subset of data. The method also includes constructing, by the processor, a sequence of relationship graphs using the computed MI values; clustering, by the processor, each relationship graph; and analyzing, by the processor, the time-ordered sequence of clustered relationship graphs to identify features in the component.

What is claimed is: 
     
         1 . A method of inferred information propagation for aircraft prognostics, said method comprising:
 receiving, by a processor, an original time-series of data points for a component as an input;   preprocessing, by the processor, the input to divide the original time-series of data into subsets of data by applying a time-window over the original time-series of data points;   computing, by the processor, a Mutual Information (MI) value for each pair of variables within each subset of data;   constructing, by the processor, a sequence of relationship graphs using the computed MI values;   clustering, by the processor, each relationship graph; and   analyzing, by the processor, the time-ordered sequence of clustered relationship graphs to identify features in the component.   
     
     
         2 . The method in accordance with  claim 1 , wherein preprocessing the input further comprises preprocessing the input to divide the original time-series of data into subsets of data by applying a time-window of predefined width over the original time-series of data points using a pre-defined step-size. 
     
     
         3 . The method in accordance with  claim 2 , further comprising:
 placing the time-window at a first time to define a first subset; and   moving the time-window forward to a second time by the pre-defined step-size to define a second subset.   
     
     
         4 . The method in accordance with  claim 3 , further comprising incrementing the time-window over the entire time-series of data to produce the time-ordered sequence of data subsets that are used to compute the MI values. 
     
     
         5 . The method in accordance with  claim 1 , wherein preprocessing the input further comprises converting the original time-series to a time-series in which all variables are sampled at the same rate when the variables are measured using different sampling rates. 
     
     
         6 . The method in accordance with  claim 5 , wherein converting the original time-series further comprises at least one of down-sampling the measurements for variables with sampling rates that are too high and interpolating the measurements of variables with sampling rates that are too low. 
     
     
         7 . The method in accordance with  claim 1 , wherein constructing a sequence of relationship graphs further comprises:
 interpreting each variable as a node in the graph;   placing an undirected edge between each node; and   setting the weights on the edges to the corresponding MI values.   
     
     
         8 . The method in accordance with  claim 1 , wherein analyzing the time-ordered sequence of clustered relationship graphs further comprises searching for changes in the component structure and characteristics of clusters over time as indicators of shifts in the operating state of an underlying subsystem. 
     
     
         9 . The method in accordance with  claim 8 , wherein searching for changes further comprises searching for at least one of cluster formation, annihilation, division, and merging of the clusters. 
     
     
         10 . A prognosis system for inferred information propagation for aircraft prognostics, said prognosis system comprising:
 a memory for storing data; and   a processor in communication with said memory, said processor programmed to:
 receive an original time-series of data points for a component as an input; 
 preprocess the input to divide the original time-series of data into subsets of data by applying a time-window over the original time-series of data points; 
 compute a Mutual Information (MI) value for each pair of variables within each subset of data; 
 construct a sequence of relationship graphs using the computed MI values; 
 cluster each relationship graph; and 
 analyze the time-ordered sequence of clustered relationship graphs to identify features in the component. 
   
     
     
         11 . The prognosis system in accordance with  claim 10 , wherein to preprocess the input, said processor is further programmed to preprocess the input to divide the original time-series of data into subsets of data by applying a time-window of predefined width over the original time-series of data points using a pre-defined step-size. 
     
     
         12 . The prognosis system in accordance with  claim 11 , wherein said processor is further programmed to:
 place the time-window at a first time to define a first subset; and   move the time-window forward to a second time by the pre-defined step-size to define a second subset.   
     
     
         13 . The prognosis system in accordance with  claim 12 , wherein said processor is further programmed to increment the time-window over the entire time-series of data to produce the time-ordered sequence of data subsets that are used to compute the MI values. 
     
     
         14 . The prognosis system in accordance with  claim 10 , wherein said processor is further programmed to wherein to preprocess the input, said processor is further programmed to convert the original time-series to a time-series in which all variables are sampled at the same rate when the variables are measured using different sampling rates. 
     
     
         15 . The prognosis system in accordance with  claim 14 , wherein to convert the original time-series, said processor is further programmed to at least one of down-sample the measurements for variables with sampling rates that are too high and interpolate the measurements of variables with sampling rates that are too low. 
     
     
         16 . The prognosis system in accordance with  claim 10 , wherein to construct a sequence of relationship graphs, said processor is further programmed to:
 interpret each variable as a node in the graph;   place an undirected edge between each node; and   set the weights on the edges to the corresponding MI values.   
     
     
         17 . The prognosis system in accordance with  claim 10 , wherein to analyze the time-ordered sequence of clustered relationship graphs, said processor is further programmed to search for changes in the component structure and characteristics of clusters over time as indicators of shifts in the operating state of an underlying subsystem. 
     
     
         18 . One or more non-transitory computer-readable storage media having computer-readable instructions encoded thereon, wherein when executed by a processor, said computer-readable instructions cause the processor to:
 receive an original time-series of data points for a component as an input;   preprocess the input to divide the original time-series of data into subsets of data by applying a time-window over the original time-series of data points;   compute a Mutual Information (MI) value for each pair of variables within each subset of data;   construct a sequence of relationship graphs using the computed MI values;   cluster each relationship graph; and   analyze the time-ordered sequence of clustered relationship graphs to identify features in the component.   
     
     
         19 . The one or more non-transitory computer-readable storage media in accordance with  claim 18 , wherein the instructions further cause the processor to preprocess the input to divide the original time-series of data into subsets of data by applying a time-window of predefined width over the original time-series of data points using a pre-defined step-size 
     
     
         20 . The one or more non-transitory computer-readable storage media in accordance with  claim 18 , wherein to analyze the time-ordered sequence of clustered relationship graphs, the instructions further cause the processor to search for changes in the component structure and characteristics of clusters over time as indicators of shifts in the operating state of an underlying subsystem.