Abstract: Infrared images of a fluid facility are captured to detect, localize, and quantify fluid leaks with confidence at the fluid facility. Changes in pixel values for a fluid leak, which reflect changes in temperature due to the fluid leak, are used to quantify the fluid leak. The location of the fluid leak is identified using the setup of the infrared camera and the fluid leak captured within the infrared images. A dispersion of the fluid leak is calculated and presented to show the path of the fluid leak.

Abstract: Fluid leak observations made by different types of sensors are combined to detect fluid leaks at a fluid facility. Separate fluid leak observations made by different types of sensors are reconciled using a Bayesian model. The Bayesian model outputs likelihoods of different fluid leak probabilities, and the likelihoods of different fluid leak probabilities are used to facilitate operations at the fluid facility.

Abstract: Differential equations defining physics of a reservoir are modeled as a neural network. Measured data for the reservoir is used as boundary condition to calculate the different equation parameters. The result is a neural ordinary differential equation network that models reservoir characteristics (e.g., inter-well connectivities, response times for injection wells and production wells) using physics that are encoded into the network. The neural ordinary differential equation network provides a solution for the reservoir that is constrained by the physics of the reservoir.

Abstract: Words in a requirement may be decomposed into sub-parts to dynamically generate a requirement graph model. The requirement graph model may include nodes for the sub-parts of the requirement. The requirement graph model may be connected to a constraint graph model based on one or more nodes of the requirement graph model. The connection of the requirement graph model to the constraint graph model may form a combined graph model, and analysis may be performed based on the combined graph model.

Abstract: A computer implemented method for designing chemical reactions for catalyst construction is described. The method includes extracting historical data including historic chemical reaction data and historic catalyst construction yield data and converting the historic chemical reaction data into graph models to represent molecular structure data. The method also includes incorporating the graph models into a chemical reaction algorithm and training a vectorized cognitive deep learning network of the chemical reaction algorithm by using the graph models and a property of the historic chemical reaction data to produce a catalyst chemical reaction model. Further, the method includes validating the catalyst chemical reaction model by inputting the historic chemical reaction data and comparing a generated property corresponding to the catalyst chemical reaction model to the property of the historic chemical reaction data. Lastly, the method includes updating the training of the catalyst chemical reaction model.

Abstract: Differential equations defining physics of a reservoir are modeled as a neural network. Measured data for the reservoir is used as boundary condition to calculate the different equation parameters. The result is a neural ordinary differential equation network that models reservoir characteristics (e.g., inter-well connectivities, response times for injection wells and production wells) using physics that are encoded into the network. The neural ordinary differential equation network provides a solution for the reservoir that is constrained by the physics of the reservoir.

Abstract: A requirements knowledge graph may include nodes for different elements of a requirement's relationship to physical characteristics, such as the requirement and the associated unsafe conditions, control objects, and hazards. A connection between nodes may define a relationship between the corresponding elements. The requirement, that is decomposed into a subject, an object, a predicate, and a context, may define a rule for the control object in the unsafe condition to avoid the hazard. A model for a plan (e.g., drilling plan for a well, design for a drilling tool) may be constructed based on the requirements knowledge graph. The plan may be generated based on the model.

Abstract: Keywords obtained from a user and/or extracted from uploaded document(s) may be used to generate potential keywords. Documents may be identified based on the keywords and the potential keywords accepted by the user. A knowledge graph model representing the identified documents may be generated. The knowledge graph model may include document nodes representing the identified document and a search node representing the keywords. The relative position of the document nodes with respect to the search node may represent similarity between the corresponding documents and the keywords.

Abstract: Multiple sets of documents for different domains may be used to train multiple domain-specific models. A graph model may be generated to include nodes representing concepts included within the domain-specific models. A white space not including any nodes within the graph model may be identified. Analysis of the white space may be performed based on two or more nodes at periphery of the white space. Words/documents that cover the white space may be generated. Novelty of concepts may be readily assessed using the graph model/white space.

Abstract: Document information may define words, key groups of words, and sets of context words within a document. Word feature scores for words within the document may be generated. Key group feature scores for individual key groups of words may be generated based on aggregation of word feature scores the words within the individual key groups of words and word feature scores for words within corresponding sets of context words. A document feature score for the document may be generated based on aggregation of word feature scores for words within the document. The key group feature scores and the document feature score may enable context-sensitive searching of words/word vectors in the document.

Abstract: Embodiments of labeling an unlabeled dataset are provided. One embodiment comprises (a) obtaining a labeled dataset comprising a first plurality of data inputs and corresponding labels; (b) training a classification model using the labeled dataset; (c) obtaining the unlabeled dataset comprising a second plurality of data inputs without labels; (d) applying the classification model to the unlabeled dataset to generate a predicted label for each data input of the unlabeled dataset; (e) determining a verification quantity of the predicted labels to be verified by a user; (f) obtaining a verification dataset for the verification quantity of the predicted labels verified by the user; (g) updating the classification model using the verification dataset; and (h) applying the updated classification model to the remaining predicted labels that did not undergo verification and updating in response to the updated classification model. The verification dataset comprises an update to at least one predicted label.

Abstract: A requirements knowledge graph may include nodes for different elements of a requirement's relationship to physical characteristics, such as the requirement and the associated unsafe conditions, control objects, and hazards. A connection between nodes may define a relationship between the corresponding elements. The requirement, that is decomposed into a subject, an object, a predicate, and a context, may define a rule for the control object in the unsafe condition to avoid the hazard. A model for a plan (e.g., drilling plan for a well, design for a drilling tool) may be constructed based on the requirements knowledge graph. The plan may be generated based on the model.

Abstract: Document information may define words, key groups of words, and sets of context words within a document. Word feature scores for words within the document may be generated. Key group feature scores for individual key groups of words may be generated based on aggregation of word feature scores the words within the individual key groups of words and word feature scores for words within corresponding sets of context words. A document feature score for the document may be generated based on aggregation of word feature scores for words within the document. The key group feature scores and the document feature score may enable context-sensitive searching of words/word vectors in the document.

Abstract: Methods and systems for predicting part reliability in energy infrastructure are disclosed. One such method includes collecting a time-series history of failures for each failure mode of a plurality of failure modes occurring in one or more energy infrastructure assets of an energy infrastructure. The method also includes fitting a failure model to each failure mode within the plurality of failure modes, and calculating a mean time between failures and a variance associated with failure time as a function of time. The method further includes calculating times of changes in each of a plurality of preventative and mitigative barriers and interrelationships among barriers to determine a schedule of inspection, maintenance, and repair activities, based on a cost of the associated activity and the failure model for each failure mode as a function of time when time series data sets representative of the associated barrier are included as process inputs.