System and method for monitoring distributed asset data

• US 8,959,065 B2
• Filed: 04/09/2012
• Issued: 02/17/2015
• Est. Priority Date: 04/09/2012
• Status: Active Grant

• Current Assignee: Mitek Analytics LLC
• Original Assignee: Mitek Analytics LLC
• Inventors: Gorinevsky, Dimitry
• Primary Examiner(s): Zen, Wei
• Assistant Examiner(s): HUDA, MOHAMMED NURUL
• Application Number:
  US13/442,817
• Publication Number:
  US 20130268501A1
• Time in Patent Office: 1,044 Days
• Field of Search: 707/693
• US Class Current: 707/693
• CPC Class Codes: G06Q 10/0639 Performance analysis of employees; Performance analysis of enterprise or organisation operations
• CPC Class Codes: G06Q 10/20 Administration of product repair or maintenance

Abstract

A computer-based monitoring system and monitoring method implemented in computer software for detecting, estimating, and reporting the condition states, their changes, and anomalies for many assets. The assets are of same type, are operated over a period of time, and outfitted with data collection systems. The proposed monitoring method accounts for variability of working conditions for each asset by using regression model that characterizes asset performance. The assets are of the same type but not identical. The proposed monitoring method accounts for asset-to-asset variability; it also accounts for drifts and trends in the asset condition and data. The proposed monitoring system can perform distributed processing of massive amounts of historical data without discarding any useful information where moving all the asset data into one central computing system might be infeasible. The overall processing is includes distributed preprocessing data records from each asset to produce compressed data.

Claims

1. A system for monitoring a plurality of assets, the assets comprising engineered systems or machines of the same type, the system comprising:

at least one Local Computing Elements and a Central Computing Element, wherein;

a. each of the plurality of assets generating asset data for monitoring of the asset condition or performance;

b. each of the at least one Local Computing Elements, comprising a processor, is connected to a Local Data Repository receiving and storing the asset data as individual data records, wherein each of the individual data records comprises a plurality of data channels from a corresponding asset of the plurality of assets of the same type collected over a predetermined period of time;

c. each of the at least one Local Computing Elements is connected to the Central Computing Element through a data link;

d. each of the at least one Local Computing Elements pre-processes the individual data records in the Local Data Repository to produce compressed data, wherein the pre-processing of the monitoring data generated by each asset comprises computing time series of regression variable vectors, each regression variable vector comprising one or more dependent variables and one or more regressors variables, wherein the data record pre-processing comprises calculating an empirical covariance matrix for the regression variable vector from the data record and the compressed data include the empirical covariance matrixes, the compressed data being transmitted over the data link to the Central Computing Element;

e. each of the individual data records being pre-processed separately and independently to produce the compressed data;

f. the Central Computing Element collects, stores, and post-processes the compressed data from all Local Computing Elements to obtain monitoring reporting data, the monitoring reporting data being displayed or transmitted by the Central Computing Element as the monitoring system output; and

g. the monitoring reporting data is computed based on a result of an optimal estimation of the model parameters for the plurality of the assets, the model parameters comprising model parameters for each of the assets, the model parameters and the monitoring reporting data computed by the Central Computing Element using compressed data from all the Local Computing Elements being the same as if all the asset monitoring data were simultaneously available and used for the optimal estimation of the model parameters.

2. The system according to claim 1, wherein the regression fit of the asset monitoring data is based on a multi-level regression model with fixed effects describing asset-to-asset variability and data record-to-data record trends.

3. The system according to claim 1, wherein the empirical covariance matrix comprises a sample second moment matrix for the combined regressors and performance variables.

4. The system according to claim 1, wherein the coordinated data processing is performed incrementally, by updating intermediate processing state data and wherein the monitoring reporting data are computed based on the intermediate processing state data.

5. The system according to claim 1, wherein the assets comprise at least one of:

manufacturing tools, machines, power systems, propulsion systems, ground vehicles, aircraft, marine vehicles, weapon systems, structural assets, building assets, production and service plant assets, assets in power, water and waste treatment facilities, distribution network assets, communication network assets, IT system assets and transport system assets.

6. The system according to claim 1, wherein the monitoring reporting data computed by the Central Computing Element comprise the results of at least one of:

anomaly detection, fault isolation, and trend forecasting for the plurality of the monitored assets.

7. A method for monitoring a plurality of assets, the assets comprising engineered systems or machines of the same type, the method being performed in a system comprising at least one Local Computing Elements and a Central Computing Element, the method comprising:

a. generating, by each of the plurality of assets, asset data for monitoring of the asset condition or performance;

b. receiving and storing the asset data by a Local Data Repository as individual data records, wherein each of the individual data records comprises a plurality of data channels from a corresponding asset of the plurality of assets of the same type collected over a predetermined period of time, wherein each of the at least one Local Computing Elements is connected to the Local Data Repository and wherein each of the at least one Local Computing Elements is connected to the Central Computing Element through a data link;

c. pre-processing by each of the at least one Local Computing Elements the individual data records in the Local Data Repository to produce compressed data, wherein the pre-processing of the monitoring data generated by each asset comprises computing time series of regression variable vectors, each regression variable vector comprising one or more dependent variables and one or more regressors variables, wherein the data record pre-processing comprises calculating an empirical covariance matrix for the regression variable vector from the data record and the compressed data include the empirical covariance matrixes, wherein each of the individual data records are being pre-processed separately and independently to produce the compressed data;

d. transmitting the compressed data over the data link to the Central Computing Element;

e. collecting, storing, and post-processing by the Central Computing Element the compressed data from all Local Computing Elements to obtain monitoring reporting data, the monitoring reporting data being displayed or transmitted by the Central Computing Element as the monitoring system output; and

f. computing the monitoring reporting data based on a result of an optimal estimation of the model parameters for the plurality of the assets, the said model parameters comprising model parameters for each of the assets, the model parameters and the monitoring reporting data computed by the Central Computing Element using compressed data from all the Local Computing Elements being the same as if all the asset monitoring data were simultaneously available and used for the optimal estimation.

8. The method according to claim 7, wherein the regression fit of the asset monitoring data is based on a multi-level regression model with fixed effects describing asset-to-asset variability and data record-to-data record trends.

9. The method according to claim 7, wherein the empirical covariance matrix comprises a sample second moment matrix for the combined regressors and performance variables.

10. The method according to claim 7, wherein the coordinated data processing is performed incrementally, by updating intermediate processing state data and wherein the monitoring reporting data are computed based on the intermediate processing state data.

11. The method according to claim 7, wherein the assets comprise at least one of:

manufacturing tools, machines, power systems, propulsion systems, ground vehicles, aircraft, marine vehicles, weapon systems, structural assets, building assets, production and service plant assets, assets in power, water and waste treatment facilities, distribution network assets, communication network assets, IT system assets and transport system assets.

12. The method according to claim 7, wherein the monitoring reporting data computed by the Central Computing Element comprise the results of at least one of:

anomaly detection, fault isolation, and trend forecasting for the plurality of the monitored assets.

13. A non-transitory computer-readable medium embodying a set of instructions, which, when executed by one or more processors of a system comprising at least one Local Computing Elements and a Central Computing Element, cause the one or more processors to perform a method for monitoring a plurality of assets, the assets comprising engineered systems or machines of the same type, the method comprising:

a. generating, by each of the plurality of assets, asset data for monitoring of the asset condition or performance;

b. receiving and storing the asset data by a Local Data Repository as individual data records, wherein each of the individual data records comprises a plurality of data channels from a corresponding asset of the plurality of assets of the same type collected over a predetermined period of time, wherein each of the at least one Local Computing Elements is connected to the Local Data Repository and wherein each of the at least one Local Computing Elements is connected to the Central Computing Element through a data link;

c. pre-processing by each of the at least one Local Computing Elements the individual data records in the Local Data Repository to produce compressed data, wherein the pre-processing of the monitoring data generated by each asset comprises computing time series of regression variable vectors, each regression variable vector comprising one or more dependent variables and one or more regressors variables, wherein the data record pre-processing comprises calculating an empirical covariance matrix for the regression variable vector from the data record and the compressed data include the empirical covariance matrixes, wherein each of the individual data records are being pre-processed separately and independently to produce the compressed data;

d. transmitting the compressed data over the data link to the Central Computing Element;

e. collecting, storing, and post-processing by the Central Computing Element the compressed data from all Local Computing Elements to obtain monitoring reporting data, the monitoring reporting data being displayed or transmitted by the Central Computing Element as the monitoring system output; and

f. computing the monitoring reporting data based on a result of an optimal estimation of the model parameters for the plurality of the assets, the said model parameters comprising model parameters for each of the assets, the model parameters and the monitoring reporting data computed by the Central Computing Element using compressed data from all the Local Computing Elements being the same as if all the asset monitoring data were simultaneously available and used for the optimal estimation.

14. The non-transitory computer-readable medium according to claim 13, wherein the regression fit of the asset monitoring data is based on a multi-level regression model with fixed effects describing asset-to-asset variability and data record-to-data record trends.

15. The non-transitory computer-readable medium according to claim 13, wherein the empirical covariance matrix comprises a sample second moment matrix for the combined regressors and performance variables.

16. The non-transitory computer-readable medium according to claim 13, wherein the coordinated data processing is performed incrementally, by updating intermediate processing state data and wherein the monitoring reporting data are computed based on the intermediate processing state data.

17. The non-transitory computer-readable medium according to claim 13, wherein the assets comprise at least one of:

manufacturing tools, machines, power systems, propulsion systems, ground vehicles, aircraft, marine vehicles, weapon systems, structural assets, building assets, production and service plant assets, assets in power, water and waste treatment facilities, distribution network assets, communication network assets, IT system assets and transport system assets.