# US 2018/0320658 A1 · Computer System & Method for Predicting an Abnormal Event at a Wind Turbine in a Cluster > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2018/0320658 A1 - Document ID: US20180320658A1 - Status: Published application - Kind: A1 - Application number: 15/585,940 - Filed: 2017-05-03 - Published: 2018-11-08 - Earliest claimed: 2017-05-03 - Assignees: Uptake Technologies, Inc. - Applicants: Uptake Technologies, Inc. - Inventors: James Herzog; Benedict Augustine - Web page: https://patentgrep.com/patent/US20180320658A1 ## CPC - G05B 19/0426: {Programming the control sequence} - F03D 7/045: {with model-based controls} - F03D 7/048: {controlling wind farms} - G05B 23/024: {Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks} - F03D 17/00: Monitoring or testing of wind motors, e.g. diagnostics (testing during commissioning of wind motors F03D13/30) - F05B 2270/404: active, predictive, or anticipative - G05B 2219/2619: Wind turbines - Y02E 10/72: Wind turbines with rotation axis in wind direction - G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G) - F03D: WIND MOTORS - F05B: INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G - Y02E: REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION ## Abstract The example systems, methods, and devices disclosed herein generally relate to performing predictive analytics on behalf of wind turbines. In some instances, a data-analytics platform defines and executes a predictive model for a specific wind turbine. The predictive model may be defined and executed based on operating data for the specific wind turbine and for other wind turbines that experience similar environmental conditions as the specific wind turbine and that are operating in an expected operational state. In response to executing the predictive model, the data-analytics platform may cause an action to occur at the specific wind turbine or cause a user interface to display a representation of the output of the executed model, among other possibilities. ## Claims ### Claim 1 A non-transitory computer-readable medium having program instructions stored thereon that are executable to cause a computing system to: for each given wind turbine of a plurality of wind turbines, defining a predictive model related to the operation of the given wind turbine, wherein defining the predictive model comprises: identifying a historical-time-varying wind-turbine cluster that the given wind turbine was a member of during a period of time in the past, wherein the historical-time-varying wind-turbine cluster comprises a plurality of past-active wind turbines that were active during at least a portion of the period of time; and based at least on (i) historical operating data for each of the plurality of past-active wind turbines and (ii) historical operating data for the given wind turbine, defining the predictive model related to the operation of the given wind turbine; for each given wind turbine of the plurality of wind turbines, executing the predictive model related to the operation of the given wind turbine, wherein executing the predictive model comprises: identifying a wind-turbine cluster that the given wind turbine is a member of, wherein the wind-turbine cluster comprises a plurality of active wind turbines including the given wind turbine; based at least on (i) recent operating data for at least one of the plurality of active wind turbines other than the given wind turbine and (ii) recent operating data for the given wind turbine, executing the predictive model to output a prediction related to the operation of the given wind turbine; and causing a graphical user interface that is communicatively coupled to the computing system to display a representation of the output of the executed predictive model. ### Claim 2 (depends on 1) The non-transitory computer-readable medium of claim 1, wherein identifying the historical-time-varying wind-turbine cluster is based at least on historical environmental data of the given wind turbine from the period of time in the past, and wherein identifying the wind-turbine cluster is based at least on recent environmental data of the given wind turbine. ### Claim 3 (depends on 2) The non-transitory computer-readable medium of claim 2, wherein the historical environmental data indicates a measure of one or more historical environmental conditions that the given wind turbine experienced during the period of time in the past, and wherein the recent environmental data indicates a measure of one or more environmental conditions that the given wind recently experienced. ### Claim 4 (depends on 1) The non-transitory computer-readable medium of claim 1, wherein identifying the historical-time-varying wind-turbine cluster comprises: identifying a first set of wind turbines that experienced at least one environmental condition similarly as the given wind turbine during the period of time in the past; and identifying, from the first set of wind turbines, a second set of wind turbines that were active when the at least one environmental condition was experienced, wherein the plurality of past-active wind turbines is the second set of wind turbines. ### Claim 5 (depends on 4) The non-transitory computer-readable medium of claim 4, wherein identifying the second set of wind turbines is based at least on one or both of historical maintenance data and historical electricity-production data. ### Claim 6 (depends on 4) The non-transitory computer-readable medium of claim 4, wherein identifying the second set of wind turbines is based at least on an observed power-curve and one or both of historical electricity-production data and historical environmental data corresponding to the historical electricity-production data. ### Claim 7 (depends on 1) The non-transitory computer-readable medium of claim 1, wherein identifying the wind-turbine cluster comprises: identifying a first set of wind turbines that are determined to be experiencing at least one environmental condition similarly as the given wind turbine; and identifying, from the first set of wind turbines, a second set of wind turbines that are determined to be active when the at least one environmental condition was experienced, wherein the identified wind-turbine cluster is the second set of wind turbines. ### Claim 8 (depends on 7) The non-transitory computer-readable medium of claim 7, wherein identifying the second set of wind turbines is based at least on one or both of maintenance data and present electricity-production data. ### Claim 9 (depends on 7) The non-transitory computer-readable medium of claim 7, wherein identifying the second set of wind turbines is based at least on an observed power-curve and one or both of recent electricity-production data and recent environmental data corresponding to the recent electricity-production data. ### Claim 10 (depends on 1) The non-transitory computer-readable medium of claim 1, wherein the predictive model comprises a failure model, and wherein the output prediction comprises a likelihood that the given wind turbine will experience a failure within a certain amount of time in the future. ### Claim 11 (depends on 1) The non-transitory computer-readable medium of claim 1, wherein the predictive model comprises an anomaly detection model, and wherein the output prediction comprises an indication of whether the given wind turbine is experiencing an anomalous operating condition. ### Claim 12 (depends on 1) The non-transitory computer-readable medium of claim 1, wherein executing the predictive model comprises: determining one or more operating-data differentials based at least on (i) recent operating data for at least one of the plurality of active wind turbines other than the given wind turbine and (ii) recent operating data for the given wind turbine; and executing the predictive model based at least on the one or more operating-data differentials. ### Claim 13 (depends on 1) The non-transitory computer-readable medium of claim 1, wherein the program instructions are further executable to cause the computing system to: after executing the predictive model, in response to the output prediction, cause the given wind turbine to change an operational state of the given wind turbine. ### Claim 14 A computing system configured to perform data analytics for a plurality of wind turbines, the computing system comprising: a network interface configured to facilitate communications over a communication network with one or more wind-turbine data sources; 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: for each given wind turbine of the plurality of wind turbines, define a predictive model related to the operation of the given wind turbine, wherein defining the predictive model comprises: identifying a historical-time-varying wind-turbine cluster that the given wind turbine was a member of during a period of time in the past, wherein the historical-time-varying wind-turbine cluster comprises a plurality of past-active wind turbines that were active during at least a portion of the period of time; and based at least on (i) historical operating data for each of the plurality of past-active wind turbines and (ii) historical operating data for the given wind turbine, defining the predictive model related to the operation of the given wind turbine; for each given wind turbine of the plurality of wind turbines, execute the predictive model related to the operation of the given wind turbine, wherein executing the predictive model comprises: identifying a wind-turbine cluster that the given wind turbine is a member of, wherein the wind-turbine cluster comprises a plurality of active wind turbines including the given wind turbine; based at least on (i) recent operating data for at least one of the plurality of active wind turbines other than the given wind turbine and (ii) recent operating data for the given wind turbine, executing the predictive model to output a prediction related to the operation of the given wind turbine; and causing a graphical user interface that is communicatively coupled to the computing system to display a representation of the output of the executed predictive model. ### Claim 15 (depends on 14) The computing system of claim 14, wherein the predictive model comprises a failure model, and wherein the output prediction comprises a likelihood that the given wind turbine will experience a failure within a certain amount of time in the future. ### Claim 16 (depends on 14) The computing system of claim 14, wherein the predictive model comprises an anomaly detection model, and wherein the output prediction comprises an indication of whether the given wind turbine is experiencing an anomalous operating condition. ### Claim 17 (depends on 14) The computing system of claim 14, wherein the program instructions are further executable to cause the computing system to: after executing the predictive model, in response to the output prediction, cause the given wind turbine to change an operational state of the given wind turbine. ### Claim 18 (depends on 14) The computing system of claim 14, wherein identifying the historical-time-varying wind-turbine cluster comprises: identifying a first set of wind turbines that experienced at least one environmental condition similarly as the given wind turbine during the period of time in the past; and identifying, from the first set of wind turbines, a second set of wind turbines that were active when the at least one environmental condition was experienced, wherein the plurality of past-active wind turbines is the second set of wind turbines. ### Claim 19 (depends on 14) The computing system of claim 14, wherein identifying the wind-turbine cluster comprises: identifying a first set of wind turbines that are determined to be experiencing at least one environmental condition similarly as the given wind turbine; and identifying, from the first set of wind turbines, a second set of wind turbines that are determined to be active when the at least one environmental condition was experienced, wherein the identified wind-turbine cluster is the second set of wind turbines. ### Claim 20 A method performed by a computing system configured to perform data analytics for a plurality of wind turbines, the method comprising: for each given wind turbine of a plurality of wind turbines, defining a predictive model related to the operation of the given wind turbine, wherein defining the predictive model comprises: identifying a historical-time-varying wind-turbine cluster that the given wind turbine was a member of during a period of time in the past, wherein the historical-time-varying wind-turbine cluster comprises a plurality of past-active wind turbines that were active during at least a portion of the period of time; and based at least on (i) historical operating data for each of the plurality of past-active wind turbines and (ii) historical operating data for the given wind turbine, defining the predictive model related to the operation of the given wind turbine; for each given wind turbine of the plurality of wind turbines, executing the predictive model related to the operation of the given wind turbine, wherein executing the predictive model comprises: identifying a wind-turbine cluster that the given wind turbine is a member of, wherein the wind-turbine cluster comprises a plurality of active wind turbines including the given wind turbine; based at least on (i) recent operating data for at least one of the plurality of active wind turbines other than the given wind turbine and (ii) recent operating data for the given wind turbine, executing the predictive model to output a prediction related to the operation of the given wind turbine; and causing a graphical user interface that is communicatively coupled to the computing system to display a representation of the output of the executed predictive model.