Asset health identification from multi-modality data analysis
View Patent ↗A first set of data associated with assets can be received. An ontology graph can be constructed based on the first set of data. A second set of data associated with the assets can be received, the second set of data having a first frequency of sampling. Based on the second set of data, nodes of the ontology graph representing the assets can be characterized. A third set of data associated with the assets can be received, the third set of data having a second frequency of sampling. The third set of data can include real time data associated with the assets. Based on the third set of data and information associated with the assets represented by the ontology graph, a deep learning neural network can be trained to predict a future state of at least one asset of the assets and discover dynamic mutual impact of the assets.
1. A system comprising:
a processor;
a memory device coupled with the processor;
the processor configured to at least:
receive a first set of data associated with assets, the assets including machine components and the first set of data including static data from sources that include at least technical manuals and assets' installation location;
construct an ontology graph based on the first set of data, the ontology graph including nodes and edges connecting at least some of the nodes, wherein a node of the ontology graph represents an asset of the assets, and wherein an edge connecting two nodes represents a relationship between two of the assets represented by the two nodes;
receive a second set of data associated with the assets, the second set of data having a first frequency of sampling, the second set of data including periodic maintenance data;
based on the second set of data, characterize the nodes of the ontology graph representing the assets, the characterizing encoding different modes of interaction and dependencies among the machine components represented by the nodes;
receive a third set of data associated with the assets, the third set of data having a second frequency of sampling that has higher frequency of occurrence than the first frequency, wherein the third set of data includes real time data sampled by sensors associated with the assets; and
based on the third set of data and information associated with the assets represented by the ontology graph, train a deep learning neural network to both predict a future state of at least one asset of the assets and generate a dynamic dependency graph that discovers dynamic mutual impact of the assets at a same run,
wherein the deep learning neural network receives multiple input channels, each of the multiple input channels representing the real time data sampled by a sensor associated with an asset of the assets, wherein a sliding time window of the real time data is received for training.
2. The system of claim 1 , wherein the information associated with the assets represented by the ontology graph is represented as a data structure including an adjacency matrix.
3. The system of claim 1 , wherein the deep learning neural network includes a self-attention encoder-decoder.
4. The system of claim 1 , wherein the deep learning neural network takes as input multiple time series data generated from the assets.
5. The system of claim 1 , wherein the information associated with the assets represented by the ontology graph is used to regularize a learning process of the deep learning neural network.
6. The system of claim 1 , wherein the future state is used to prioritize maintenance of the assets.
7. The system of claim 1 , wherein the deep learning neural network is retrained based on receiving updates to the second set of data.
8. The system of claim 1 , wherein the assets include machinery equipment.
9. A computer-implemented method comprising:
receiving a first set of data associated with assets, the assets including machine components and the first set of data including static data from sources that include at least technical manuals and assets' installation location;
constructing ontology graph based on the first set of data, the ontology graph including nodes and edges connecting at least some of the nodes, wherein a node of the ontology graph represents an asset of the assets, and wherein an edge connecting two nodes of the ontology graph represents a relationship between two of the assets represented by the two nodes;
receiving a second set of data associated with the assets, the second set of data having a first frequency of sampling, the second set of data including periodic maintenance data;
based on the second set of data, characterizing the nodes of the ontology graph representing the assets, the characterizing encoding different modes of interaction and dependencies among the machine components represented by the nodes;
receiving a third set of data associated with the assets, the third set of data having a second frequency of sampling that has higher frequency of occurrence than the first frequency, wherein the third set of data includes real time data sampled by sensors associated with the assets; and
based on the third set of data and information associated with the assets represented by the ontology graph, training a deep learning neural network to both predict a future state of at least one asset of the assets and generate a dynamic dependency graph that discovers dynamic mutual impact of the assets at a same run,
wherein the deep learning neural network receives multiple input channels, each of the multiple input channels representing the real time data sampled by a sensor associated with an asset of the assets, wherein a sliding time window of the real time data is received for training.
10. The computer-implemented method of claim 9 , wherein the information associated with the assets represented by the ontology graph is represented as a data structure including an adjacency matrix.
11. The computer-implemented method of claim 9 , wherein the deep learning neural network includes self-attention encoder-decoder.
12. The computer-implemented method of claim 9 , wherein the deep learning neural network takes as input multiple time series data generated from the assets.
13. The computer-implemented method of claim 9 , wherein the information associated with the assets represented by the ontology graph is used to regularize a learning process of the deep learning neural network.
14. The computer-implemented method of claim 9 , wherein the future state is used to prioritize maintenance of the assets.
15. The computer-implemented method of claim 9 , wherein the deep learning neural network is retrained based on receiving updates to the second set of data.
16. The computer-implemented method of claim 9 , wherein the assets include machinery equipment.
17. A computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions readable by a device to cause the device to:
receive a first set of data associated with assets, the assets including machine components and the first set of data including static data from sources that include at least technical manuals and assets' installation location;
construct an ontology graph based on the first set of data, the ontology graph including nodes and edges connecting at least some of the nodes, wherein a node of the ontology graph represents an asset of the assets, and wherein an edge connecting two nodes represents a relationship between two of the assets represented by two nodes;
receive a second set of data associated with the assets, the second set of data having a first frequency of sampling, the second set of data including periodic maintenance data;
based on the second set of data, characterize the nodes of the ontology graph representing the assets, the characterizing encoding different modes of interaction and dependencies among the machine components represented by the nodes;
receive a third set of data associated with the assets, the third set of data having a second frequency of sampling that has higher frequency of occurrence than the first frequency, wherein the third set of data includes real time data sampled by sensors associated with the assets; and
based on the third set of data and information associated with the assets represented by the ontology graph, train a deep learning neural network to both predict a future state of at least one asset of the assets and generate a dynamic dependency graph that discovers dynamic mutual impact of the assets at a same run,
wherein the deep learning neural network receives multiple input channels, each of the multiple input channels representing the real time data sampled by a sensor associated with an asset of the assets, wherein a sliding time window of the real time data is received for training.
18. The computer program product of claim 17 , wherein the information associated with the assets represented by the ontology graph is used to regularize a learning process of the deep learning neural network.