Early warning system for locomotive bearings failures

Abstract

Aspects of the present disclosure involve a system and a method for predicting locomotive engine bearing failures (e.g., main bearing failures). In example embodiments, the method may access locomotive data about a particular locomotive engine, and generate a set of element concentration trajectories therefrom. Each element concentration trajectory includes a measure of element concentrations in the locomotive engine's oil over a period of time. The method further includes determining a health score of the locomotive engine based on a comparison of the set of element concentrations with a set of baseline element concentration trajectories that are representative of element concentrations of a group of locomotive engines' oil over the period of time. The health score provides a measure of the likelihood of bearing failure in the particular locomotive engine. A message regarding the likelihood of bearing failure may then be provided.

What is claimed is: 
     
         1 . A method comprising:
 accessing locomotive data corresponding to a particular locomotive engine, the locomotive data including one or more sets of element concentration measurements associated with the particular locomotive engine;   creating a set of element concentration trajectories for the particular locomotive engine based on the locomotive data, each element concentration trajectory including concentration measurements of a particular element over a period of time;   creating a set of baseline element concentration trajectories using information about a plurality of locomotive engines, each baseline element concentration trajectory providing representative concentration measurements of the plurality of locomotive engines over the period of time;   calculating, using one or more processors of a machine, a health score for the particular locomotive engine based on a comparison of the set of element concentration trajectories of the particular locomotive with the set of baseline element concentration trajectories, the health score indicating a likelihood of a bearing failure in the particular locomotive engine; and   providing a message regarding the likelihood of failure of the bearing of the particular locomotive engine.   
     
     
         2 . The method of  claim 1 , further comprising:
 comparing the health score to a predefined threshold; and   wherein the providing of the message is based on a result of the comparing of the health score to the predefined threshold.   
     
     
         3 . The method of  claim 1 , wherein the creating of each element concentration trajectory includes:
 obtaining a piecewise linear approximation of element concentrations over the period of time by applying linear regression to concentration measurement values included in the locomotive data;   applying smoothing techniques to the piecewise linear approximation to accumulate concentration measurements across oil change intervals; and   filtering accumulated concentration measurements.   
     
     
         4 . The method of  claim 1 , wherein the creating of each baseline element concentration trajectory includes:
 creating an element concentration trajectory for each locomotive engine in the plurality of locomotive engines;   aggregating the created element concentration trajectories to create a population level baseline; and   selecting a representative portion of the population level baseline as the baseline element concentration trajectory.   
     
     
         5 . The method of  claim 1 , wherein the calculating of the health score includes:
 comparing the set of element concentration trajectories of the particular locomotive with the set of baseline element concentration trajectories;   determining a set of comparison features based on the comparing of the set of element concentrations trajectories of the particular locomotive with the set of baseline element concentration trajectories; and   comparing the set of comparison features with a training set of comparison features generated based on a comparison of historical locomotive data of the plurality of locomotives with the set of baseline element concentration trajectories.   
     
     
         6 . The method of  claim 5 , wherein the comparing of the set of comparison features with the training set of comparison features includes using a one class support vector machine (SVM) to determine a similarity between the set of comparison features and the training set of comparison features, wherein the health score is based on the similarity. 
     
     
         7 . The method of  claim 5 , wherein the set of comparison features includes at least one of the group comprising: engine age, instantaneous values, global averages, local averages, trends, and area above baseline trajectory. 
     
     
         8 . The method of  claim 1 , further comprising:
 calculating health scores for the plurality of locomotive engines;   ranking the plurality of locomotive engines based in part on the determined health scores; and   causing display of a list of locomotive engine identifiers corresponding to the plurality of locomotive engines, the list of locomotive engine identifiers being ordered according to the ranking of the locomotive engine.   
     
     
         9 . The method of  claim 1 , further comprising:
 receiving updated locomotive data corresponding to the particular locomotive engine, the updated locomotive data including at least one new element concentration value; and   re-calculating the health score in response to receiving the updated locomotive data.   
     
     
         10 . The method of  claim 1 , wherein the locomotive data includes at least one of oil change events, maintenance events, bearing failure events, and bearing change-out events. 
     
     
         11 . The method of  claim 1 , wherein the providing of the message includes transmitting at least one of an electronic mail message, text message, or notification to a computer device. 
     
     
         12 . The method of  claim 1 , wherein the providing of the message includes causing display of a user interface, the user interface presenting the message. 
     
     
         13 . The method of  claim 12 , wherein the user interface further includes a graphical representation of the comparison of the set of element concentration trajectories of the particular locomotive with the set of baseline element concentration trajectories. 
     
     
         14 . A system comprising:
 one or more processors of a machine; and   a machine-readable storage medium storing instructions that, when executed by the one or more processors, cause the machine to perform operations comprising:   accessing locomotive data corresponding to a particular locomotive engine, the locomotive data including one or more sets of element concentration measurements associated with the particular locomotive engine;   creating a set of element concentration trajectories for the particular locomotive engine based on the locomotive data, each element concentration trajectory including concentration measurements of a particular element over a period of time;   creating a set of baseline element concentration trajectories using information about a plurality of locomotive engines, each baseline element concentration trajectory providing representative concentration measurements of the plurality of locomotive engines over the period of time;   calculating, using one or more processors of a machine, a health score for the particular locomotive engine based on a comparison of the set of element concentration trajectories of the particular locomotive with the set of baseline element concentration trajectories, the health score indicating a likelihood of a bearing failure in the particular locomotive engine; and   providing a message regarding the likelihood of failure of the bearing of the particular locomotive engine.   
     
     
         15 . The system of  claim 14 , wherein the calculating of the health score includes:
 comparing the set of element concentration trajectories of the particular locomotive with the set of baseline element concentration trajectories;   determining a set of comparison features based on the comparing of the set of element concentrations trajectories of the particular locomotive with the set of baseline element concentration trajectories; and   comparing the set of comparison features with a training set of comparison features generated based on a comparison of historical locomotive data of the plurality of locomotives with the set of baseline element concentration trajectories.   
     
     
         16 . The system of  claim 15 , wherein the set of comparison features includes at least one of the group comprising: engine age, instantaneous values, global averages, local averages, trends, and area above baseline trajectory. 
     
     
         17 . The system of  claim 14 , wherein the operations further comprise:
 calculating health scores for the plurality of locomotive engines;   ranking the plurality of locomotive engines based in part on the determined health scores; and   causing display of a list of locomotive engine identifiers corresponding to the plurality of locomotive engines, the list of locomotive engine identifiers being ordered according to the ranking of the locomotive engine.   
     
     
         18 . The system of  claim 14 , wherein the providing of the message includes causing display of a user interface, the user interface including the message and a graphical representation of the comparison of the set of element concentration trajectories of the particular locomotive with the set of baseline element concentration trajectories. 
     
     
         19 . The system of  claim 14 , wherein the operations further comprise:
 receiving updated locomotive data corresponding to the particular locomotive engine, the updated locomotive data including at least one new value when compared to the historical locomotive data; and   re-calculating the health score for the bearing based on the updated locomotive data.   
     
     
         20 . A non-transitory machine-readable storage medium embodying instructions that, when executed by at least one processor of a machine, cause the machine to perform operations comprising:
 accessing locomotive data corresponding to a particular locomotive engine, the locomotive data including one or more sets of element concentration measurements associated with the particular locomotive engine;   creating a set of element concentration trajectories for the particular locomotive engine based on the locomotive data, each element concentration trajectory including concentration measurements of a particular element over a period of time;   creating a set of baseline element concentration trajectories using information about a plurality of locomotive engines, each baseline element concentration trajectory providing representative concentration measurements of the plurality of locomotive engines over the period of time;   calculating, using one or more processors of a machine, a health score for the particular locomotive engine based on a comparison of the set of element concentration trajectories of the particular locomotive with the set of baseline element concentration trajectories, the health score indicating a likelihood of a bearing failure in the particular locomotive engine; and   providing a message regarding the likelihood of failure of the bearing of the particular locomotive engine.