International Business Machines Corporation

Tiago Bertoni Scarton

Thiago Bianchi

Joao Souto Maior

Tarcisio Pereira

Thiago Roberto Fontes Iunes

Marlon Nascimento Vicente

Eduardo Dias Felicio Junior

Gary A Cooper

Dashiell Lucena Gonzaga Borges

Suraj Shirodkar Singh

A method for predicting a time to a failure condition of an asset includes a first model, trained to reconstruct input image pairs to resemble a healthy condition asset. Similarity coefficients are generated for respective historical image pairs that include a visible-light and infrared light image by use of the reconstructed image pair by the first model as a similarity base and the respective historical image pairs include a timestamp of image capture. A second model is trained to predict a time to a failure condition of the asset based on similarity coefficients and timestamps of real-time image pair capture, timestamps of an asset failure condition, and similarity coefficients of the respective historical images. Responsive to receipt of a first real-time image pair, the method computes a predicted time to the failure condition of the asset.

Asset maintenance prediction using infrared and regular images

A method for predicting a time to a failure condition of an asset receives image pairs of a healthy condition asset and historical image pairs of the asset from inspections and at a failure condition of the asset. A first model is trained to rebuild input images to resemble a healthy condition asset. Similarity coefficients are generated for the historical image pairs of a regular and infrared image by use of the rebuilding of images by the first model as a similarity base and historical image pairs include a timestamp of image capture. A second model is trained to predict a time to an asset failure condition based on timestamps of image capture, timestamps of an asset failure condition, and similarity coefficients of the historical images. Responsive to receipt of a real-time image pair, the method computes a predicted time to a failure condition of the asset.

Weather data, asset variable values, and failure data may be received. A correlation factor may be generated by comparing the weather data and the asset variable values. One or more weather variables may be selected based on the correlation factor. A predicted asset variable value may be determined based on the weather data and the selected weather variables. The weather data and the predicted asset variable values may be compared to predict a failure probability for a first asset segment.

Risk failure prediction for line assets

A method for asset maintenance includes determining a location of an asset, receiving meteorological data corresponding to the location, determining forecasted meteorological conditions for the asset from the meteorological data, determining a forecasted probability of degradation or failure of the asset based, at least in part, on the forecasted meteorological conditions, and dispatching asset maintenance services for the asset responsive to determining that the forecasted probability of degradation or failure of the asset is above a selected threshold. The method may also include collecting sensor data corresponding to the asset, and determining the forecasted probability of degradation or failure of the asset based, at least in part, on the forecasted meteorological conditions and the sensor data. A corresponding apparatus, computer program product, and computer system are also disclosed herein.

Predictive asset maintenance

Described are techniques for identifying anomalies in log files. The techniques including a method encompassing mapping a plurality of log entries into a bag-of-words matrix and determining respective cosine similarities between consecutive log entries in the bag-of-words matrix. The method further includes converting respective cosine similarities to respective states and determining respective state transitions between consecutive log entries. The method further includes generating a stochastic matrix based on the respective state transitions and identifying an anomalous log entry based on an anomalous state transition. The anomalous state transition has a probability below a probability threshold, and the probability is based on the stochastic matrix.

Anomaly identification in log files

A linear asset processing and managing system and method may include receiving a plurality of metrics associated with a linear asset segment from an asset management system over a network, in response to a request by the computing system to the asset management system, calculating an intermediate point on the geospatial context using a geometric analytic tool that analyzes the plurality of metrics received from the asset management system, determining a geospatial length of the linear asset segment, using the intermediate point and at least one metric of the plurality of metrics, and displaying the linear asset segment of the linear asset on the geospatial context.

Visualizing linear assets using client-side processing

Processing location sensor data to detect location anomalies by defining a multi-cell overlay grid for a defined asset tracking area, determining a current location on the overlay grid for an asset according to location data, determining that a probability of transitioning from a previous location to the current location is below a threshold, registering the current location as an anomaly and updating the last known location of the asset to the current location.

Growing community of inventors

Sao Paulo, Brazil

Tiago Bertoni Scarton