Equipment repair management and execution

Abstract

In some examples, a computer system may receive historical repair data for first equipment, and may extract features from the historical repair data for the first equipment as training data including one or more of: free-text variables associated with comments related to the first equipment; usage attributes associated with the first equipment; equipment attributes associated with the first equipment; sensor data associated with the first equipment; or event data associated with the first equipment. The system may determine a repair hierarchy including a plurality of repair levels for the equipment. The system may use the training data to train a machine learning model as a multilayer model trained to perform multiple tasks for predicting individual levels of the repair hierarchy. The system may receive a repair request associated with second equipment and uses the machine learning model to determine at least one repair action based on the received repair request.

What is claimed: 
     
         1 . A system comprising:
 one or more processors; and   one or more non-transitory computer-readable media maintaining executable instructions, which, when executed by the one or more processors, configure the one or more processors to perform operations comprising:
 receiving historical repair data for first equipment; 
 extracting features from the historical repair data for the first equipment as training data including one or more of: free-text variables associated with comments related to the first equipment; usage attributes associated with the first equipment; 
   equipment attributes associated with the first equipment; sensor data associated with the first equipment; or event data associated with the first equipment;
 determining, from the historical repair data, a repair hierarchy including a plurality of repair levels for the first equipment; 
 training, using the training data, a deep learning model as a multilayer model trained to perform multiple tasks for predicting individual levels of the repair hierarchy; 
 receiving a repair request associated with second equipment; and 
 using the deep learning model to determine at least one repair action based on the received repair request. 
   
     
     
         2 . The system as recited in  claim 1 , the operations further comprising:
 extracting features from data associated with the second equipment including one or more of: free-text variables associated with comments related to the second equipment; usage attributes associated with the second equipment; equipment attributes associated with the second equipment; sensor data associated with the second equipment; or event data associated with the second equipment; and   inputting the extracted features associated with the second equipment into the deep learning model to determine the at least one repair action.   
     
     
         3 . The system as recited in  claim 2 , the operations further comprising:
 extracting features from the event data associated with the second equipment as a plurality of event data feature vectors;   determining a context vector based on the features extracted from at least one of the free-text variables, the usage attributes, or the equipment attributes associated with the second equipment; and   weighting the event data feature vectors based at least in part on the context vector.   
     
     
         4 . The system as recited in  claim 1 , the operations further comprising, during the training of the deep learning model, employing a backpropagation algorithm to calculate gradients with respect to each learnable variable in the deep learning model using a loss function. 
     
     
         5 . The system as recited in  claim 1 , wherein the deep learning model is one of a deep learning neural network or a recurrent neural network. 
     
     
         6 . The system as recited in  claim 1 , the operations further comprising determining a repair plan based on the one or more repair actions and a respective probability of success determined for each repair action. 
     
     
         7 . The system as recited in  claim 6 , the operations further comprising, based on the repair plan performing at least one of:
 sending an order for a part for a repair;   sending a communication to assign labor to perform the repair;   sending a communication to schedule a repair time for the repair; or   remotely initiating a procedure on the equipment to effectuate, at least partially, the repair.   
     
     
         8 . A method comprising:
 receiving, by a first computing device, historical repair data for the first equipment;   extracting features from the historical repair data for the first equipment as training data including one or more of: free-text variables associated with comments related to the first equipment; usage attributes associated with the first equipment; equipment attributes associated with the first equipment; sensor data associated with the first equipment; or event data associated with the first equipment;   determining a repair hierarchy including a plurality of repair levels for the first equipment;   training, using the training data, a machine learning model as a multilayer model trained to perform multiple tasks for predicting individual levels of the repair hierarchy;   receiving a repair request associated with second equipment; and   using the machine learning model to determine at least one repair action based on the received repair request.   
     
     
         9 . The method as recited in  claim 8 , further comprising:
 extracting features from data associated with the second equipment including one or more of: free-text variables associated with comments related to the second equipment; usage attributes associated with the second equipment; equipment attributes associated with the second equipment; sensor data associated with the second equipment; or event data associated with the second equipment; and   inputting the extracted features associated with the second equipment into the machine learning model to determine the at least one repair action.   
     
     
         10 . The method as recited in  claim 9 , further comprising:
 extracting features from the event data associated with the second equipment as a plurality of event data feature vectors;   determining a context vector based on the features extracted from at least one of the free-text variables, the usage attributes, or the equipment attributes associated with the second equipment; and   weighting the event data feature vectors based at least in part on the context vector.   
     
     
         11 . The method as recited in  claim 8 , further comprising, during the training of the machine learning model, employing a backpropagation algorithm to calculate gradients with respect to each learnable variable in the machine learning model using a loss function. 
     
     
         12 . The method as recited in  claim 8 , wherein the machine learning model is one of a deep learning neural network or a recurrent neural network. 
     
     
         13 . The method as recited in  claim 8 , further comprising determining a repair plan based on the one or more repair actions and a respective probability of success determined for each repair action. 
     
     
         14 . The method as recited in  claim 13 , further comprising, based on the repair plan performing at least one of:
 sending an order for a part for a repair;   sending a communication to assign labor to perform the repair;   sending a communication to schedule a repair time for the repair; or   remotely initiating a procedure on the equipment to effectuate, at least partially, the repair.   
     
     
         15 . One or more non-transitory computer-readable media storing instructions that, when executed by one or more processors, program the one or more processors to perform operations comprising:
 receiving historical repair data for first equipment;   extracting features from the historical repair data for the first equipment as training data including one or more of: free-text variables associated with comments related to the first equipment; usage attributes associated with the first equipment; equipment attributes associated with the first equipment; sensor data associated with the first equipment; or event data associated with the first equipment;   determining a repair hierarchy including a plurality of repair levels for the first equipment;   training, using the training data, a deep learning model as a multilayer model trained to perform multiple tasks for predicting individual levels of the repair hierarchy;   receiving a repair request associated with second equipment; and   using the deep learning model to determine at least one repair action based on the received repair request.   
     
     
         16 . The method as recited in  claim 15 , the operations further comprising:
 extracting features from data associated with the second equipment including one or more of: free-text variables associated with comments related to the second equipment; usage attributes associated with the second equipment; equipment attributes associated with the second equipment; sensor data associated with the second equipment; or event data associated with the second equipment; and   inputting the extracted features associated with the second equipment into the deep learning model to determine the at least one repair action.   
     
     
         17 . The method as recited in  claim 16 , wherein extracting features from the event data associated with the second equipment as a plurality of event data feature vectors;
 determining a context vector based on the features extracted from at least one of the free-text variables, the usage attributes, or the equipment attributes associated with the second equipment; and   weighting the event data feature vectors based at least in part on the context vector.   
     
     
         18 . The method as recited in  claim 15 , the operations further comprising, during the training of the deep learning model, employing a backpropagation algorithm to calculate gradients with respect to each learnable variable in the deep learning model using a loss function. 
     
     
         19 . The method as recited in  claim 15 , wherein the deep learning model is one of a deep learning neural network or a recurrent neural network. 
     
     
         20 . The method as recited in  claim 15 , further comprising:
 determining a repair plan based on the one or more repair actions and a respective probability of success determined for each repair action; and   based on the repair plan performing at least one of:
 sending an order for a part for a repair; 
 sending a communication to assign labor to perform the repair; 
 sending a communication to schedule a repair time for the repair; or 
 remotely initiating a procedure on the equipment to effectuate, at least partially, the repair.