System and method for detecting turbine underperformance and abnormal operation

Abstract

A method of correcting turbine underperformance includes: calculating a power production curve using monitored data; detecting a change between the monitored data and a baseline power production curve; generating an operability curve for pairs of operating variables based on the monitored data; detecting a change between the operability curve and a corresponding baseline operability curve; comparing the change to a corresponding predetermined metric and, if the change exceeds the metric, providing feedback to a turbine control system that identifies at least one of the pairs of operating variables for each pair of variables that exceeds the metric. A system and non-transitory computer-readable medium are also disclosed.

Description

Background Art

[0001] Poor turbine performance is the main contributor to turbine energy production loss. Conventional methods first manually analyze turbine performance data, which can lead to results with large uncertainties. Then, the connection between insufficient energy production and abnormal turbine operation cannot be automatically established, which makes root cause identification difficult. Summary of the invention

[0002] Technical Solution 1. A method for correcting poor turbine performance, the method comprising:

[0003] accessing monitored operational data for the turbine;

[0004] calculating a power generation curve using at least a portion of said monitored operational data;

[0005] predicting a baseline power production curve based on expected performance of the turbine;

[0006] detecting a first set of changes between at least said portion of said monitored operational data and said baseline power production curve;

[0007] generating one or more monitored operability curves based on said at least a portion of said monitored operational data, each of said one or more monitored operability curves describing a relationship between monitored values ​​for a pair of operational variables;

[0008] generating one or more baseline operability curves for the turbine, each of the one or more baseline operability curves describing an expected relationship between the pairs of operating variables;

[0009] detecting a set of changes between the one or more monitored operability curves and a corresponding one of the one or more baseline operability curves;

[0010] comparing one or more of the set of changes to a corresponding predetermined metric for each of the pairs of manipulated variables; and

[0011] Based on determining that one or more members of the set of changes exceed the corresponding predetermined metric, feedback is provided to a turbine control system that identifies at least one of the pairs of manipulated variables corresponding to the member of the set of changes that exceeds the predetermined metric.

[0012] Technical Solution 2. The method according to Technical Solution 1 is characterized by comprising preprocessing the portion of the monitored operation data to exclude at least one of the data representing a period of at least one of turbine shutdown and turbine production curtailment.

[0013] Technical Solution 3. The method according to Technical Solution 1 is characterized in that one or more elements of historical measurement data for the turbine are included in the baseline power production curve.

[0014] Technical Solution 4. The method according to Technical Solution 1 is characterized in that it includes applying a smoothing technique to the expected baseline power generation curve.

[0015] Technical Solution 5. The method according to Technical Solution 1 is characterized in that it includes generating a time series graph of the power residual before detecting the set of changes.

[0016] Technical Solution 6. The method according to Technical Solution 5 is characterized in that it includes applying a change point detection algorithm to detect the set of changes.

[0017] Technical Solution 7. The method according to Technical Solution 1 is characterized in that at least one member of the pair of operating variables identified by the turbine control system has a value that can be controlled by the turbine control system.

[0018] Technical Solution 8. A non-transitory computer readable medium having instructions stored thereon, wherein the instructions, when executed by a control processor, cause the control processor to perform a method of correcting poor turbine performance, the method comprising:

[0019] accessing monitored operational data for the turbine;

[0020] calculating a power generation curve using at least a portion of said monitored operational data;

[0021] predicting a baseline power production curve based on expected performance of the turbine;

[0022] detecting a first set of changes between at least said portion of said monitored operational data and said baseline power production curve;

[0023] generating one or more monitored operability curves based on said at least a portion of said monitored operational data, each of said one or more monitored operability curves describing a relationship between monitored values ​​of a pair of operational variables;

[0024] generating one or more baseline operability curves for the turbine, each of the one or more baseline operability curves describing an expected relationship between the pairs of operating variables;

[0025] detecting a set of changes between the one or more monitored operability curves and a corresponding one of the one or more baseline operability curves;

[0026] comparing one or more of the set of changes to a corresponding predetermined metric for each of the pairs of manipulated variables; and

[0027] Based on determining that one or more members of the set of variations exceed the corresponding predetermined metric, feedback is provided to the turbine control system, which identifies at least one of the pairs of manipulated variables corresponding to the member of the set of variations that exceeds the predetermined metric.

[0028] Technical Solution 9. The non-temporary computer-readable medium according to Technical Solution 8 is characterized in that the instructions are also configured to cause the control processor to execute the method by including preprocessing the portion of the monitored operating data to exclude at least one of the data of the time period representing at least one of turbine shutdown and turbine production reduction.

[0029] Technical Solution 10. The non-transitory computer-readable medium according to Technical Solution 8 is characterized in that the instructions are also configured to cause the control processor to execute the method by including one or more elements of historical measurement data of the turbine in the baseline power production curve.

[0030] Technical Solution 11. The non-transitory computer-readable medium according to Technical Solution 8 is characterized in that the instructions are also configured to cause the control processor to perform the method by including applying a smoothing technique to the expected baseline power generation curve.

[0031] Technical Solution 12. The non-transitory computer-readable medium according to Technical Solution 8 is characterized in that the instructions are also configured to cause the control processor to execute the method by including generating a time series graph of the power residual before detecting the set of changes.

[0032] Technical Solution 13. The non-transitory computer-readable medium according to Technical Solution 12 is characterized in that the instructions are also configured to cause the control processor to execute the method by applying a change point detection algorithm to detect the set of changes.

[0033] Technical Solution 14. The non-transitory computer-readable medium according to Technical Solution 8 is characterized in that the instructions are also configured to cause the control processor to execute the method by including at least one member of the pair of operating variables identified by the turbine control system having a value that can be controlled by the turbine control system.

[0034] Technical Solution 15. A turbine control system for correcting poor turbine performance, the system comprising:

[0035] a turbine having a control processor in communication with the memory unit and data storage;

[0036] The control processor includes a processing unit, the control processor communicating with one or more turbine actuators;

[0037] The memory unit includes a non-transitory computer readable executable medium, the instructions, when executed by the control processor, causing the control processor to perform a method of correcting turbine underperformance, the method comprising:

[0038] accessing monitored operational data for the turbine;

[0039] calculating a power generation curve using at least a portion of said monitored operational data;

[0040] predicting a baseline power production curve based on expected performance of the turbine;

[0041] detecting a first set of changes between at least said portion of said monitored operational data and said baseline power production curve;

[0042] generating one or more monitored operability curves based on said at least a portion of said monitored operational data, each of said one or more monitored operability curves describing a relationship between monitored values ​​of a pair of operational variables;

[0043] generating one or more baseline operability curves for the turbine, each of the one or more baseline operability curves describing an expected relationship between the pairs of operating variables;

[0044] detecting a set of changes between the one or more monitored operability curves and a corresponding one of the one or more baseline operability curves;

[0045] comparing one or more of the set of changes to a corresponding predetermined metric for each of the pairs of manipulated variables; and

[0046] Based on determining that one or more members of the set of changes exceed the corresponding predetermined metric, feedback is provided to control one or more turbine actuators, which affect the value of at least one of the paired operational variables corresponding to the member of the set of changes that exceeds the predetermined metric.

[0047] Technical Solution 16. The system according to Technical Solution 15 is characterized in that the instructions are also configured to cause the control processor to execute the method by including preprocessing the portion of the monitored operating data to exclude at least one of the data of the time period representing at least one of turbine shutdown and turbine production reduction.

[0048] Technical Solution 17. The system according to Technical Solution 15 is characterized in that the instructions are also configured to cause the control processor to execute the method by including one or more elements of historical measurement data of the turbine in the baseline power production curve.

[0049] Technical Solution 18. The system according to Technical Solution 15 is characterized in that the instructions are also configured to cause the control processor to execute the method by including applying a smoothing technique to the expected baseline power generation curve.

[0050] Technical Solution 19. The system according to Technical Solution 15 is characterized in that the instruction is further configured to cause the control processor to execute the method by including the following steps:

[0051] generating a time series plot of the power residual prior to detecting the set of changes; and

[0052] A change point detection algorithm is applied to detect the set of changes.

[0053] Technical Solution 20. The system according to Technical Solution 15 is characterized in that the instructions are also configured to cause the control processor to execute the method by including at least one member of the pair of operating variables identified by the turbine control system having a value that can be controlled by the turbine control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A system for turbine underperformance detection and root cause identification is depicted according to an embodiment;

[0055] Figure 2 A representation of a time series data graph over time is graphically depicted;

[0056] Figure 3 A process for identifying a root cause of poor turbine performance is depicted according to an embodiment;

[0057] Figure 4A graphically depicts turbine operating performance data for a power curve according to an embodiment;

[0058] Figure 4B Depicted graphically Figure 4A Time series representation of performance data;

[0059] Figure 5A According to an embodiment, a diagram is shown for Figure 4A operational performance data of potential causes of operational performance data;

[0060] Figure 5BDepicted graphically Figure 5A Time series representation of performance data;

[0061] Fig. 6A According to an embodiment, a diagram is shown for Figure 4A operational performance data that are potential causes of the operational performance data; and

[0062] Figure 6B Depicted graphically Fig. 6A A time series representation of performance data. DETAILED DESCRIPTION

[0063] The implemented systems and methods can detect poor performance of a wind turbine power plant and identify the root causes of the poor performance by applying a model of (multiple) operating characteristics of the wind turbine. Embodiments provide automatic detection of poor performance of a turbine and its operational anomalies. Such automatic detection avoids the conventional manual data analysis process, reduces uncertainty, and determines the root cause (multiple) of the poor performance. The root causes of poor performance may include incorrect control parameters, blade misalignment, suboptimal blade pitch control, etc.

[0064] In some implementations, after identifying the root cause(s), embodiments may generate actionable information based on which the turbine's control system may adjust operating parameters of the turbine to increase its energy performance, and / or provide a report indicating component maintenance / replacement recommendations.

[0065] An implemented method can be applied to real-time turbine operating data co-located at both turbines at a specific wind farm and / or turbines at the entire fleet level. Implementation at a broad level can benefit asset performance management strategies and greatly increase energy productivity of facilities across the enterprise. The ability to identify root causes enables more efficient maintenance planning and recovery from lost energy production.

[0066] Figure 1 A system 100 for detecting turbine underperformance and identifying root cause(s) of underperformance is depicted according to an embodiment. The system 100 may be incorporated within a single turbine, implemented locally at a wind turbine farm, or implemented on a remote server. The system 100 may include a control processor 110 (having a processor unit 112) that communicates with other components of the system via a data / control bus 120. The system may communicate with remote servers and other devices over an electronic communications network via a communication interface unit 144.

[0067] The control processor 110 may access executable program instructions 132 which cause the processor to perform implemented operations. The executable instructions may be stored in the memory unit 130 or in the data storage 150. The memory unit 130 may provide a local cache memory 134 for the control processor.

[0068] The data filtering / preprocessing unit 140 receives sensor / monitor data containing information about the turbine operating state (rotation, blade pitch, energy output, etc.) and surrounding environmental conditions (wind speed, wind direction, air density, temperature, humidity, etc.). The operating state and environmental condition data are associated with a timestamp representing the time when the data point was collected. According to an embodiment, the detection of insufficient production due to turbine performance uses full-load turbine operating data. The data filtering / preprocessing unit excludes data representing periods of turbine shutdown and production restriction. Data processing is performed to ensure data quality and data validity, such as air density correction for wind speed measurement. The resulting data is stored in the operating data record 151.

[0069] The power production curve baseline model 153 predicts the expected power production of the turbine under different wind speeds and other environmental conditions. To reduce the impact of extreme outliers and generate reliable power curve estimates, the baseline power curve model may include robust smoothing techniques - for example, residual or regression analysis calculations.

[0070] The power curve change detection model 155 quantifies the difference between the power measured at the same wind speed (contained in the operational data record 151) and the corresponding baseline power estimate (from the power curve baseline model 153). Assuming no changes, the power residual time series will follow a common distribution throughout the detection period. The power curve change detection model identifies changes in the power residual time series. The detected changes may correspond to potential power underproduction. The detection model may implement a statistical method that identifies when the time series changes significantly in a particular metric (e.g., mean, variance, etc.). The metric may be predetermined by the user based on field conditions, production requirements, and other factors.

[0071] The operability curve baseline model 157 describes the relationship within one or more pairs of turbine operating variables (e.g., generator torque versus generator speed; blade pitch angle versus wind speed; blade pitch angle versus power; etc.) For each type of operability curve, a robust baseline model is provided to represent the normal turbine operating mode.

[0072] The operability curve change detection model 159 defines a vertical residual for any particular type of operability baseline curve. The vertical residual may be the difference between the actual value on the y-axis and the associated baseline estimate corresponding to the same value on the x-axis. The operability curve change detection model identifies significant changes in the residual time series (based on a predetermined metric), and any detected changes may correspond to potential abnormal turbine operation.

[0073] The baseline model may implement robust regression technique(s) to accurately capture the baseline turbine performance, and may use a change point detection algorithm to identify the most significant changes from the baseline.

[0074] The underperformance detection and root cause identification unit 142 (referred to herein as the "identification unit") monitors the output from the power curve change detection model. If the monitored power curve changes exceed a predetermined metric (threshold and / or magnitude), the identification unit analyzes one or more of the paired turbine operating variables from the operability curve change detection model to identify the root cause of the power curve change. In some implementations, the power curve change detection model may compare the power curve change to the predetermined metric and send a signal of an out-of-tolerance condition to the identification unit.

[0075] For example, Figure 2 A time series data graph 200 is graphically depicted for a plurality of data points 205. Data graph region 210 precedes time t1 and includes data points having a data point mean 220 of approximately zero. Data graph region 212 extends between time t1-t2 and includes data points having a data point mean 222 of magnitude of approximately Y1. Data graph region 214 extends after time t2 and includes data points having a data point mean 224 of magnitude of approximately Y2. Time series data graph 200 may represent residual time series data generated by comparing a measured operability curve to an operability curve baseline model 157 for one of a pair of turbine operating variables.

[0076] According to an embodiment, the identification unit 142 may evaluate the residuals for one or more time periods of a time series data graph (e.g., graph 200). The evaluation may be a comparison of the magnitude of the residuals (multiple) with a predetermined metric. Based on the results of the evaluation, the identification unit may determine whether the pair of turbine operating variables that produced the particular data is the root cause of the change in the generated power curve for the turbine.

[0077] The implemented systems and methods provide automatic detection of turbine underperformance and operational anomalies. The implemented methods can be applied to real-time turbine operational data for one or more turbines at a wind farm and / or fleet level simultaneously, which facilitates the development of improved asset performance management strategies. The improved strategies can increase power productivity. Additionally, the ability to identify root causes can enable more efficient maintenance planning and reduce energy production losses.

[0078] In some implementations, a user-defined baseline performance model may be used, such as using a user-specified engineering power curve in a power curve detection model. In addition, the method may be applied to various turbine operating conditions and / or parameters (low or high wind speed, specified humidity range, atmospheric particulate suspension, etc.) to detect condition-specific root causes. In some implementations, these user-specified conditions and / or parameters may be used during design specification development to predict the expected production capacity of turbine, wind farm, and / or fleet operations.

[0079] Figure 3 A process 300 for identifying the root cause of poor turbine performance is depicted according to an embodiment. At step 305, operational data containing information about the turbine operating state (rotation, blade pitch, energy output, etc.) and the turbine's ambient environmental conditions (wind speed, wind direction, air density, temperature, humidity, etc.) is accessed. The operational data is acquired in real time (i.e., during turbine operation) by sensors and monitors.

[0080] At step 310, the data of the turbine output power readings is filtered to remove data indicating periods of turbine shutdown and curtailment. Data preprocessing associates environmental conditions with the same timestamp as the output power reading data. The resulting filtered and processed data is stored in the operational data record 151.

[0081] At step 320, a baseline model of the turbine's power curve is generated. The baseline model represents an expected power production curve for the turbine under various wind speeds and other conditions. The expected power production may be based on the manufacturer's specifications, which may be supplemented with historical (individualized or fleet-level) measurement data for the turbine.

[0082] Figure 4A Turbine operating performance data 400 is graphically depicted according to an embodiment. The performance data is for turbine output power (over time) at multiple wind speeds recorded simultaneously with sensed output power. Baseline curve 410 represents the expected output power versus wind speed. Many data samples of power output within region 420 do not increase with increasing wind speed. Such stagnation in power output may indicate underproduction.

[0083] Changes between the turbine's power production curve (eg, from operational data log 151) and a baseline model of the turbine's power production curve are detected at step 322. These changes can be identified by examining the power residual time series, where negative changes can correspond to potential power production shortfalls.

[0084] Figure 4B Depicted graphically Figure 4A A time series representation of the performance data of 450. Prior to t1, the time series data has a mean residual of about zero, indicating that the measured power production curve of the turbine matches its baseline power production curve model. Between time t1-t2, the time series data has a negative mean residual. After time t2, the residual has a positive bias.

[0085] At step 324, one or more baseline models of operability curves for the turbine are generated. These baseline model operability curves represent the relationship between one or more pairs of turbine operating variables (e.g., generator torque versus generator speed; blade pitch angle versus wind speed; blade pitch angle versus power; etc.). These operability curves may be based on manufacturer specifications, which may be supplemented with historical (individualized or fleet-level) data for the turbine.

[0086] For example, Figure 5A According to an embodiment, operational performance data 500 is graphically depicted for pairs of operational variables blade pitch angle versus wind speed, which may be Figure 4A Curve 510 represents the baseline operability curve generated at step 324 for this pair of variables.

[0087] Fig. 6A According to an embodiment, operating performance data 600 is graphically depicted for pairs of operating variables generator torque and generator speed, which may be Figure 4A Curve 610 represents the baseline operability curve generated at step 324 for this pair of variables.

[0088] At step 326, vertical residual changes in the residual time series data between the operability curve baseline model and the actual operability curve of the turbine are detected. The actual operability curve may be generated from the time stamped paired operational variable sensor data 161 for various parameter pairs of paired operational variables of the turbine (generator torque vs. generator speed; blade pitch angle vs. wind speed; blade pitch angle vs. power; etc.). The magnitude of the detected change in the residual time series (based on a predetermined metric) may correspond to potential abnormal turbine operation.

[0089] For example, Figure 5B Depicted graphically Figure 5A A time series representation of the operational performance data of 550. Prior to t1, the time series data has a mean residual of about zero, indicating that the measured operation of these paired variables matches its baseline operability curve. Between time t1-t2, the time series data has a negative mean residual. After time t2, the residual has a positive bias.

[0090] Figure 6B Depicted graphically Fig. 6A A time series representation of the performance data is shown at 650. The time series data has a mean residual of approximately zero over the entire monitoring time, which indicates that the measured operability of these paired variables matches their baseline operability curves.

[0091] For one or more root causes, at step 330, a vertical average residual change in the residual time series data of the operability curve at a time corresponding to a change in the power production curve is quantified by comparison to a predetermined metric.

[0092] At step 340, a determination is made as to whether one or more operability curves include a change that exceeds their respective predetermined amounts. If the change is less than the predetermined metric, process 300 returns to step 305. If the change exceeds their predetermined metric, then at step 350, feedback information regarding the root cause of the underproduction is provided to the turbine control system. The turbine control system may then adjust one or more turbine actuators to affect the value of a corresponding one of the operating variables to increase power production. Process 300 may then return to step 305 for continued underperformance detection and root cause identification.

[0093] According to some embodiments, a computer program application stored in a non-volatile memory or computer-readable medium (e.g., register memory, processor cache, RAM, ROM, hard drive, flash memory, CD ROM, magnetic media, etc.) may include code or executable program instructions that, when executed, may instruct and / or cause a controller or processor to perform the methods discussed herein, such as the method of detecting turbine production deficiencies and identifying root causes, as disclosed above.

[0094] The computer readable medium may be a non-transitory computer readable medium, which includes all forms and types of memory and all computer readable media except transient propagating signals. In one implementation, the non-volatile memory or computer readable medium may be an external memory.

[0095] Although specific hardware and methods have been described herein, it is noted that any number of other configurations may be provided in accordance with embodiments of the present invention. Therefore, although the basic novel features of the present invention have been shown, described, and pointed out, it should be understood that various omissions, substitutions, and changes may be made to the form and details of the illustrated embodiments and their operation by those skilled in the art without departing from the spirit and scope of the present invention. The substitution of elements from one embodiment to another is also fully anticipated and contemplated. The present invention is limited only with respect to the appended claims and the equivalents listed therein.

Claims

1. A method for correcting poor turbine performance, the method comprising: accessing monitored operational data for the turbine; calculating a power generation curve using at least a portion of said monitored operational data; predicting a baseline power production curve based on expected performance of the turbine; detecting a first set of changes between at least said portion of said monitored operational data and said baseline power production curve; generating one or more monitored operability curves based on said at least a portion of said monitored operational data, each of said one or more monitored operability curves describing a relationship between monitored values ​​for a pair of operational variables; generating one or more baseline operability curves for the turbine, each of the one or more baseline operability curves describing an expected relationship between the pairs of operating variables; detecting a set of changes between the one or more monitored operability curves and a corresponding one of the one or more baseline operability curves; comparing one or more of the set of changes to a corresponding predetermined metric for each of the pairs of manipulated variables; and Based on determining that one or more members of the set of changes exceed the corresponding predetermined metric, feedback is provided to a turbine control system that identifies at least one of the pairs of manipulated variables corresponding to the member of the set of changes that exceeds the predetermined metric.

2. The method according to claim 1, characterized in that Included is pre-processing the portion of the monitored operational data to exclude at least one of data indicative of a period of at least one of turbine shutdown and turbine curtailment.

3. The method according to claim 1, characterized in that One or more elements of historical measurement data for the turbine are included in the baseline power production curve.

4. The method according to claim 1, characterized in that This includes applying a smoothing technique to the expected baseline power production curve.

5. The method according to claim 1, characterized in that This includes generating a time series graph of the power residual prior to detecting the set of changes.

6. The method according to claim 5, characterized in that This includes applying a change point detection algorithm to detect the set of changes.

7. The method according to claim 1, characterized in that At least one member of the pair of manipulated variables identified by the turbine control system has a value that can be controlled by the turbine control system.

8. A non-transitory computer readable medium having stored thereon instructions which, when executed by a control processor, cause the control processor to perform a method of correcting turbine underperformance, the method comprising: accessing monitored operational data for the turbine; calculating a power generation curve using at least a portion of said monitored operational data; predicting a baseline power production curve based on expected performance of the turbine; detecting a first set of changes between at least said portion of said monitored operational data and said baseline power production curve; generating one or more monitored operability curves based on said at least a portion of said monitored operational data, each of said one or more monitored operability curves describing a relationship between monitored values ​​of a pair of operational variables; generating one or more baseline operability curves for the turbine, each of the one or more baseline operability curves describing an expected relationship between the pairs of operating variables; detecting a set of changes between the one or more monitored operability curves and a corresponding one of the one or more baseline operability curves; comparing one or more of the set of changes to a corresponding predetermined metric for each of the pairs of manipulated variables; and Based on determining that one or more members of the set of variations exceed the corresponding predetermined metric, feedback is provided to the turbine control system, which identifies at least one of the pairs of manipulated variables corresponding to the member of the set of variations that exceeds the predetermined metric.

9. The non-transitory computer-readable medium of claim 8, wherein: The instructions are further configured to cause the control processor to perform the method by at least one of pre-processing the portion of the monitored operational data to exclude data indicative of periods of at least one of turbine shutdown and turbine curtailment.

10. The non-transitory computer readable medium of claim 8, wherein: The instructions are further configured to cause the control processor to perform the method by including one or more elements of historical measurement data of the turbine in the baseline power production curve.

11. The non-transitory computer-readable medium of claim 8, wherein: The instructions are also configured to cause the control processor to perform the method including applying a smoothing technique to the expected baseline power production curve.

12. The non-transitory computer readable medium of claim 8, wherein: The instructions are further configured to cause the control processor to perform the method including generating a time series graph of power residuals prior to detecting the set of changes.

13. The non-transitory computer readable medium of claim 12, wherein: The instructions are further configured to cause the control processor to perform the method including applying a change point detection algorithm to detect the set of changes.

14. The non-transitory computer readable medium of claim 8, wherein: The instructions are further configured to cause the control processor to execute the method by including at least one member of the pair of operating variables identified by the turbine control system having a value that is controllable by the turbine control system.

15. A turbine control system for correcting turbine underperformance, the system comprising: a turbine having a control processor in communication with the memory unit and data storage; The control processor includes a processing unit, the control processor communicating with one or more turbine actuators; The memory unit includes non-transitory computer-readable executable instructions that, when executed by the control processor, cause the control processor to perform a method of correcting turbine underperformance, the method comprising: accessing monitored operational data for the turbine; calculating a power generation curve using at least a portion of said monitored operational data; predicting a baseline power production curve based on expected performance of the turbine; detecting a first set of changes between at least said portion of said monitored operational data and said baseline power production curve; generating one or more monitored operability curves based on said at least a portion of said monitored operational data, each of said one or more monitored operability curves describing a relationship between monitored values ​​of a pair of operational variables; generating one or more baseline operability curves for the turbine, each of the one or more baseline operability curves describing an expected relationship between the pairs of operating variables; detecting a set of changes between the one or more monitored operability curves and a corresponding one of the one or more baseline operability curves; comparing one or more of the set of changes to a corresponding predetermined metric for each of the pairs of manipulated variables; and Based on determining that one or more members of the set of changes exceed the corresponding predetermined metric, feedback is provided to control one or more turbine actuators, which affect the value of at least one of the paired operational variables corresponding to the member of the set of changes that exceeds the predetermined metric.

16. The system according to claim 15, characterized in that The instructions are further configured to cause the control processor to perform the method by at least one of pre-processing the portion of the monitored operational data to exclude data indicative of periods of at least one of turbine shutdown and turbine curtailment.

17. The system according to claim 15, characterized in that The instructions are further configured to cause the control processor to perform the method by including one or more elements of historical measurement data of the turbine in the baseline power production curve.

18. The system according to claim 15, characterized in that The instructions are also configured to cause the control processor to perform the method including applying a smoothing technique to the expected baseline power production curve.

19. The system according to claim 15, characterized in that The instructions are further configured to cause the control processor to perform the method by comprising the steps of: generating a time series plot of the power residual prior to detecting the set of changes; and A change point detection algorithm is applied to detect the set of changes.

20. The system according to claim 15, characterized in that The instructions are further configured to cause the control processor to execute the method by including at least one member of the pair of operating variables identified by the turbine control system having a value that is controllable by the turbine control system.

Images