Computer architecture and method for modifying intake data rate based on a predictive model

Patent metadata

Field	Value
Publication number	US-10025653-B2
Application number	US-201514963212-A
Country	US
Kind code	B2
Filing date	Dec 8, 2015
Priority date	Dec 1, 2014
Publication date	Jul 17, 2018
Grant date	Jul 17, 2018

Abstract

Disclosed herein is a computer architecture and software that is configured to modify data intake operation at an asset-monitoring system based on a predictive model. In accordance with the present disclosure, the asset-monitoring system may execute a predictive model that outputs an indicator of whether at least one event from a group of events (e.g., a failure event) is likely to occur at a given asset within a given period of time in the future. Based on the output of this predictive model, the asset-monitoring system may modify one or more operating parameters for ingesting data from the given asset, such as a storage location for the ingested data, a set of data variables from the asset that are ingested, and/or a rate at which data from the asset is ingested.

First claim

The invention claimed is:

1. A computing system comprising:
a network interface configured to receive data from a plurality of assets;
a data intake system configured to ingest data received from the plurality of assets;
at least one processor;
a non-transitory computer-readable medium; and
program instructions stored on the non-transitory computer-readable medium that are executable by the at least one processor to cause the computing system to:
based on historical operating data for a plurality of assets, define a predictive model that is configured to (a) receive sensor data for an asset as input, (b) for each of at least two failure types from a group of failure types related to mechanical operation, make a respective prediction of whether the failure type is likely to occur at the asset in the future, and (c) based on the respective predictions, determine and output a health metric indicating whether at least one failure type from the group of failure types related to mechanical operation is predicted to occur at the asset in the future, wherein the historical operating data comprises (i) historical abnormal-condition data for the plurality of assets that indicates past occurrences of abnormal conditions that are associated with the group of failure types and (ii) historical sensor data for the plurality of assets that indicates sensor measurements associated with the past occurrences of abnormal conditions;
operate in a first mode in which the data intake system ingests operating data received from a given asset of the plurality of assets at a first ingestion rate, wherein the operating data comprises data related to the mechanical operation of the given asset that includes abnormal-condition data and sensor data;
while operating in the first mode, (a) receive operating data from the given asset; and (b) ingest at least a portion of the received operating data at the first ingestion rate, wherein the ingested portion of the received operating data includes ingested sensor data for the given asset;
apply the predictive model to the ingested sensor data and thereby determine a health metric for the given asset that indicates whether at least one failure type from the group of failure types related to mechanical operation is predicted to occur at the given asset in the future;
compare the health metric for the given asset to a threshold condition that defines whether an asset is considered to be in a state of impending failure and thereby make a determination that the health metric satisfies the threshold condition such that the given asset is considered to be in a state of impending failure;
in response to the determination, transition from operating in the first mode to operating in a second mode in which the data intake system ingests operating data from the given asset at a second ingestion rate that is higher than the first ingestion rate; and
while operating in the second mode, (a) receive operating data from the given asset and (b) ingest at least a portion of the received operating data at the second ingestion rate, wherein the ingested operating data comprises data related to the mechanical operation of the given asset that includes abnormal-condition data and sensor data.

2. The computing system of claim 1 , wherein the health metric indicating whether at least one failure type from the group of failure types related to mechanical operation is predicted to occur at the given asset in the future comprises a probability that no failure type from the group of failure types related to mechanical operation is predicted to occur at the given asset in the future, and wherein the determination that the indicator satisfies the threshold condition comprises a determination that the probability is at or below a threshold value.

3. The computing system of claim 1 , wherein the health metric indicating whether at least one failure type from the group of failure types related to mechanical operation of the given asset is predicted to occur at the given asset in the future comprises a probability that at least one f failure type from the group of failure types related to mechanical operation of an asset is predicted to occur at the given asset in the future, and wherein the determination that the indicator satisfies the threshold condition comprises a determination that the probability is at or above a threshold value.

4. The computing system of claim 1 , wherein the second ingestion rate comprises a variable rate that is determined based on the health score.

5. The computing system of claim 1 , wherein the given asset comprises a transportation machine, an industrial machine, or a utility machine.

6. The computing system of claim 1 , wherein the second ingestion rate comprises a variable rate that is determined based on the comparison of the health metric to the threshold condition.

7. The computing system of claim 1 , wherein the second ingestion rate comprises a variable rate that is determined based on one or more characteristics of the given asset.

8. A non-transitory computer-readable medium having instructions stored thereon that are executable to cause a computing system to:
based on historical operating data for a plurality of assets, define a predictive model that is configured to (a) receive sensor data for an asset as input, (b) for each of at least two failure types from a group of failure types related to mechanical operation, make a respective prediction of whether the failure type is likely to occur at the asset in the future, and (c) based on the respective predictions, determine and output a health metric indicating whether at least one failure type from the group of failure types related to mechanical operation is predicted to occur at the asset in the future, wherein the historical operating data comprises (i) historical abnormal-condition data for the plurality of assets that indicates past occurrences of abnormal conditions that are associated with the group of failure types and (ii) historical sensor data for the plurality of assets that indicates sensor measurements associated with the past occurrences of abnormal conditions;
operate in a first mode in which the computing system ingests operating data received from a given asset of the plurality of assets at a first ingestion rate, wherein the operating data comprises data related to the mechanical operation of the given asset that includes abnormal-condition data and sensor data;
while operating in the first mode, (a) receive operating data from the given asset, and (b) ingest at least a portion of the received operating data at the first ingestion rate, wherein the ingested portion of the received operating data includes ingested sensor data for the given asset;
apply the predictive model to the ingested sensor data and thereby determine a health metric for the given asset that indicates whether at least one failure type from the group of failure types related to mechanical operation is predicted to occur at the given asset in the future;
compare the health metric for the given asset to a threshold condition that defines whether an asset is considered to be in a state of impending failure and thereby make a determination that the health metric satisfies the threshold condition such that the given asset is considered to be in a state of impending failure;
in response to the determination, transition from operating in the first mode to operating in a second mode in which the computing system ingests operating data from the given asset at a second ingestion rate that is higher than the first ingestion rate; and
while operating in the second mode, (a) receive operating data from the given asset and (b) ingest at least a portion of the received operating data at the second ingestion ra

Assignees

• Uptake Tech Inc

Inventors

• Goldstein Michael
• Ravensberg Tom
• Hansmann Will

Classifications

• G06Q10/20 Primary

  Administration of product repair or maintenance · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06Q10/0633

  Workflow analysis · CPC title

• G06F11/0754

  by exceeding limits · CPC title

• G05B23/0254

  based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks · CPC title