WO2019066934A1

WO2019066934A1 - Wind turbine performance monitoring

Info

Publication number: WO2019066934A1
Application number: PCT/US2017/054459
Authority: WO
Inventors: Alexis Motto; Cristovian BASDEN; Chao Yuan; Sridharan Palanivelu; Bryan GILLENWATER; Noah SCHELLENBERG; Jennifer ZELMANSKI; Akshay PATWAL; Amit Chakraborty; Michael May; Matthew Evans
Original assignee: Siemens Energy Inc
Current assignee: Siemens Energy Inc
Priority date: 2017-09-29
Filing date: 2017-09-29
Publication date: 2019-04-04
Anticipated expiration: 2020-03-29

Abstract

Examples of techniques for wind turbine performance monitoring are disclosed. In one example implementation according to aspects of the present disclosure, a computer-implemented method for wind turbine performance monitoring includes receiving, by a processing device, wind speed data and power generation data for each of a plurality of wind turbines. The plurality of wind turbines are distributed through a geographic area. The method further includes estimating, by the processing device, a potential power loss for at least one of the plurality of wind turbines based at least in part on the wind speed data and the power generation data. The method further includes implementing a corrective action when it is determined that the potential power loss for the at least one of the plurality of wind turbines is greater than a threshold power loss.

Description

WIND TURBINE PERFORMANCE MONITORING

BACKGROUND

[0001] The present disclosure generally relates to energy generation performance and more particularly to wind turbine performance monitoring.

[0002] A wind turbine converts the kinetic energy of the wind into electrical energy. Large wind turbines (sometimes referred to as utility-level wind turbines) can be used to generate power and supply it to a utility grid. These utility-level wind turbines can be distributed throughout a geographic area (e.g., on-shore or off-shore) to form a group or fleet of wind turbines, also known as a wind farm. A wind farm can be a significant source of renewable energy.

SUMMARY

[0003] According to examples of the present disclosure, techniques including methods, systems, and/or computer program products for wind turbine performance monitoring are provided. A computer-implemented method for energy loss estimation for wind turbine performance monitoring includes receiving, by a processing device, wind speed data and power generation data for each of a plurality of wind turbines. The plurality of wind turbines are distributed through a geographic area. The method further includes estimating, by the processing device, a potential power loss for at least one of the plurality of wind turbines based at least in part on the wind speed data and the power generation data. The method further includes implementing a corrective action when it is determined that the potential power loss for the at least one of the plurality of wind turbines is greater than a threshold power loss.

[0004] A system for wind turbine performance monitoring includes a plurality of wind turbines, a memory having computer readable instructions, and a processing device for executing the computer readable instructions for performing a method. The method includes receiving, by a processing device, wind speed data and power generation data for each of the plurality of wind turbines. The plurality of wind turbines are distributed through a geographic area. The method further includes estimating, by the processing device, a potential power loss for at least one of the plurality of wind turbines based at least in part on the wind speed data and the power generation data. The method further includes implementing a corrective action when it is determined that the potential power loss for the at least one of the plurality of wind turbines is greater than a threshold power loss.

[0005] A computer program product for wind turbine performance monitoring includes a computer-readable storage medium having program instructions embodied therewith, the program instructions being executable by a virtual reality processing system to cause a processing device to perform a method. The method includes receiving, by a processing device, wind speed data and power generation data for each of a plurality of wind turbines. The plurality of wind turbines are distributed through a geographic area. The method further includes estimating, by the processing device, a potential power loss for at least one of the plurality of wind turbines based at least in part on the wind speed data and the power generation data. The method further includes implementing a corrective action when it is determined that the potential power loss for the at least one of the plurality of wind turbines is greater than a threshold power loss.

[0006] Additional features and advantages are realized through the techniques of the present disclosure. Other aspects are described in detail herein and are considered a part of the disclosure. For a better understanding of the present disclosure with the advantages and the features, refer to the following description and to the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] The subject matter which is regarded as the invention is particularly pointed out and distinctly claimed in the claims at the conclusion of the specification. The foregoing and other features, and advantages thereof, are apparent from the following detailed description taken in conjunction with the accompanying drawings in which:

[0008] FIG. 1 depicts a map of a wind farm having a plurality of wind turbines, according to aspects of the present disclosure;

[0009] FIG. 2 depicts a block diagram of a processing system for estimating energy loss for wind turbine performance monitoring, according to aspects of the present disclosure;

[0010] FIG. 3 depicts a flow diagram of a method for estimating energy loss for wind turbine performance monitoring, according to aspects of the present disclosure;

[0011] FIG. 4 depicts a flow diagram of another method for estimating energy loss for wind turbine performance monitoring, according to aspects of the present disclosure;

[0012] FIG. 5 depicts graphs of wind speed weekly time series for the wind turbines of FIG. 1, according to aspects of the present disclosure;

[0013] FIG. 6 depicts graphs of power weekly time series for the wind turbines of FIG. 1 , according to aspects of the present disclosure;

[0014] FIG. 7 depicts graphs of power loss weekly time series for the wind turbines of FIG. 1, according to aspects of the present disclosure; and

[0015] FIG. 8 depicts a processing system for implementing the techniques described herein according to examples of the present disclosure. DETAILED DESCRIPTION

[0016] A fleet of similar or identical utility-scale wind turbines commissioned in a given geographical area converts mechanical energy from wind speed into electrical energy (i.e., power). Barring any binding operational or design constraints, each wind turbine operates insomuch as to efficiently harvest and convert a maximum amount of prevailing mechanical energy into electrical power. Any wind turbine whose operation deviates from the said maximum power-harvesting engineering design principle is said to be under-performing, which results in energy loss (i.e., power loss). By identifying power loss, wind turbine performance can be improved.

[0017] The following parameters are used throughout the disclosure to describe aspects of embodiments of the present disclosure: represents the altitude of turbine i; represents the latitude of turbine i; represents the longitude of turbine i; represents the minimum wind speed at rated real power of turbine i; and represents the restart wind speed of turbine i.

[0018] The following variables are used throughout the disclosure to describe aspects of embodiments of the present disclosure: represents the generator speed of turbine i at time t;

represents the pitch angle of blade b of turbine i at time t; represents the yaw position of turbine i at time t; represents the air density at turbine i at time t; represents the power output of turbine i at time t; represents the ambient temperature at turbine i at time t; represents the energy loss of turbine i at time t; represents the humidity at turbine i at time t; represents the power loss of turbine i at time t; represents the precipitation at turbine i at time t; represents the turbulence at turbine i at time t; represents the wind direction at turbine i at time t; and represents the wind speed at turbine i at time t.

[0019] As used herein, a lowercase letter (e.g. x) denotes a vector, and an uppercase letter (e.g. A) denotes a matrix. A calligraphic symbol denotes a feasible set. For example, the symbol X denotes the feasible set of the variable x. If x is a variable, then x and denote its lower bound and upper bound, respectively. The symbol

denotes the index set of turbines. If x denotes a vector, the symbol n_x denotes the dimension of x. The expression E [x] is used as shorthand for the expected value of x, and the expression E[x|y] is used as as the conditional expectation of x given y.

[0020] A wind turbine operates automatically, self-starting when the wind speed is above a cut-in speed (i.e., the wind speed limit below which the wind turbine does not generate power), denoted f°^r turbine i at time t. During operation below rated

power, denoted for some turbine i at time t, the pitch angle and rotor speed are

continuously adjusted to maximize the aerodynamic efficiency of the wind turbine. Rated power is reached at a wind speed between the cut-in and maximum operational limit values (i.e., the wind spend limit at which the wind turbine can safely operate), denoted π_Γρί, for some turbine i at time t. At wind speeds above

the real power output is regulated at rated power. Speed compliance during power regulation minimizes the dynamic loads on the transmission system. If the average wind speed exceeds the maximum operational limit, here denoted for some wind turbine i at time t, the

wind turbine is shut down by feathering of the blades. When the wind drops back below the restart speed, here denoted for some turbine i at time t, the safety systems reset

automatically.

[0021] Existing wind turbine performance monitoring approaches are based on a mapping of wind speed versus real power referred to as a power curve. The International Electrotechnical Commission standardized the estimation of the power curve using a nonparametric approach, known as the "binning method." Given the wind speed and real power observation data, the binning method divides the domain of wind speed into a finite number of intervals or bins. The power curve is then obtained as a mapping of sample mean of wind speed versus sample mean of real power, using all observation data falling within each bin. For a given wind speed, the corresponding real power is the sample conditional expected value of the real power over the period considered (e.g.,

Once constructed, the power curve is then used to predict real power at any

wind speed using mapping or interpolation techniques. This methodology requires prior construction of power curves, which may be meaningful only after observing data over a sufficiently long period of time.

[0022] Another existing approach uses machine learning to create multi-variable models of turbine performance when sufficient information is available. For example, an ensemble of regression trees has been proposed to predict turbine real power as a function of several inputs including shear, wind speed, and hub height turbulence. These existing approaches are instances of the "random forest" approach. Another existing approach uses an additive multi-variate kernel technique that can include the environmental factors listed above as a new power curve model. Note that a power curve, in this case, is a high- dimensional manifold. Multi-variable methods require the prior construction of high- dimensional power curves, which may be meaningful only after observing data over a sufficiently long period of time. In essence, they require more data and are more difficult to understand and to implement than the novel techniques described herein.

[0023] The present techniques overcome the problems of existing approaches for monitoring wind turbine performance by providing a data analytics model for monitoring the performance of utility-scale wind turbines. The present techniques do not require the construction of any actual or statistical performance curves. The present techniques also support model uncertainties, disturbances, soft operational constraints, and hard physical constraints on decision variables. Moreover, the present techniques provide a special- purpose algorithm to exploit the special structure of the data analytics model for monitoring the performance of utility-scale wind turbines.

[0024] In particular, the present techniques enabling estimating energy (i.e., power) loss for a wind turbine using wind speed data and power generation data from other wind turbines in a similar geographic area. FIG. 1 depicts a map 100 of a wind farm having a plurality of wind turbines 101, 102, 103, 104, 105 (collectively "wind turbines 101— 105"), according to aspects of the present disclosure. The wind turbines 101-105 are distributed across a geographic area that experiences similar conditions. Each of the wind turbines 101-105 may experience approximately equivalent environmental conditions at their individual location. For example, the ambient temperature at each of the wind turbines 101-105 may be approximately equal (e.g., with a few degrees of the same temperature). Similarly, the wind speed at each of the wind turbines 101-105 may be approximately equal (e.g., within a few meters per second (m/s) of the same wind speed). Other environmental conditions may also be similar at each of the wind turbines 101-105.

[0025] Because the wind turbines 101-105 are distributed across a geographic area in which the wind turbines 101-105 experience similar conditions, the data collected at each of the wind turbines 101-105 can be compared and used to estimate power loss. The wind turbines 101-105 are shown to be on-shore; however, in some examples, the wind turbines 101-105 can be located on-shore, off-shore, or a combination of on-shore and off-shore. Although five wind turbines are depicted in FIG. 1, other numbers of wind turbines can be implemented according to aspects of the present disclosure.

[0026] Each of the wind turbines 101-105 can be in electronic communication with a processing system to monitor the performance of, and estimate the energy loss for, the wind turbines 101-105. For example, FIG. 2 depicts a block diagram of a processing system 200 for estimating energy loss for wind turbine performance monitoring, according to aspects of the present disclosure. The processing system 200 includes, for example, a processing device 202, a memory 204, a data engine 210, a power loss estimation engine 212, and a corrective action engine 214.

[0027] The various controllers, engines, and/or modules of the processing system 200 can be implemented as instructions stored on a computer-readable storage medium, as hardware modules, as special-purpose hardware (e.g., application specific hardware, application specific integrated circuits (ASICs), as embedded controllers, field- programmable gate arrays (FPGAs), hardwired circuitry, etc.), or as some combination or combinations of these. In examples, the controllers, engines, and/or modules described herein can be a combination of hardware and programming. The programming can be processor executable instructions stored on a tangible memory, and the hardware can include the processing device 202 for executing those instructions. Thus a system memory (e.g., the memory 204) can store program instructions that when executed by the processing device 202 implement the controllers, engines, and/or modules described herein. Other controllers, engines, and/or modules can also be utilized to include other features and functionality described in other examples herein.

[0028] The processing system 200 is now described with reference to the method 300 of FIG. 3. In particular, FIG. 3 depicts a flow diagram of a method 300 for estimating energy loss for wind turbine performance monitoring, according to aspects of the present disclosure. The method 300 can be performed by any suitable processing system, such as the processing system 200, the processing system 800 of FIG. 8, suitable combinations thereof, and/or another suitable processing system.

[0029] At block 302, the data engine 210 can receive data from the wind turbines 101-105. For example, the wind turbines 101-105 can be equipped with a sensor, sensors, and/or sensor arrays to collect data about the wind turbines 101-105 and the environmental conditions at the wind turbines 101-105. The sensors associated with the wind turbines can collect wind speed data, ambient temperate data, atmospheric pressure data, and other data about the environmental conditions around the wind turbines 101— 105. The sensors can also collect data about the wind turbines 101-105, such as real power output of each wind turbine. According to an example, one or more of the wind turbines 101-105 can be equipped with a temperature sensor to collect temperature data, a wind sensor (e.g., an anemometer) to collect wind speed data, an atmospheric pressure sensor (e.g., a barometer) to collect atmospheric pressure data, a power sensor (e.g., a power meter) to collect power generation (i.e., real power output) data, and/or other sensors to collect other data. In other examples, other sensors can be used to collect other data.

[0030] Once the data are received from the wind turbines 101-105, such as by the data engine 210, potential power loss can be estimated. For example, at block 304, the power loss estimation engine 212 estimates a potential power loss for one or more of the wind turbines 101-105 using the wind speed data and the power generation data. According to aspects of the present disclosure, the power loss estimation engine 212 estimates a potential power loss by comparing wind speed data and power generation data for one of the wind turbines (e.g., wind turbine 103) to wind speed data and power generation data for the other of the wind turbines (e.g., the wind turbines 101, 102, 104, 105).

[0031] Potential power loss ("power loss" or "PL") is defined as the energy per unit time below a threshold at a given time t > 0. Power loss can be defined mathematically using the following formula:

where y_thit is an auxiliary variable, used here for notational convenience, and defined as follows:

By definition, Integrating over the period [0, t]

yields the corresponding performance energy index as follows:

[0032] For computational efficiency, the indices can be recast in vector form for each of the turbines 101-105. Accordingly, let

denote a permutation matrix such that the following condition holds true:

where x denotes an n-dimensional vector whose components are x_lt x_n.

[0033] The condition (4) states that applying the permutation matrix P_wst to the wind speed vector, observed at time t, for all

yields a new vector of the observed wind speed value-sorted in ascending order. Using the permutation matrix, the performance power index can be recast in a compact vector form as follows:

[0034] Integrating component-wise over the period [0, T] yields the associated performance energy index in vector form:

[0035] Although this approach can produce estimates of power loss that are within acceptable engineering tolerances for most practical cases, it is possible to improve the estimation accuracy of the power loss estimation described herein by defining small classes of turbines based on prevailing environmental factors (e.g., ambient temperature, atmospheric pressure, etc.), excluding the wind speed, are sufficiently small. This approach implements elements of stochastic differential calculus using the following equations:

where

and

[0036] Once the power loss is estimated, a corrective action can be implemented at block 306 (e.g., by the corrective action engine 214) when it is determined that the potential power loss for one of the wind turbines is greater than a threshold power loss. Examples of corrective actions can be issuing an alert to an operator, dispatching a technician to perform maintenance (e.g., troubleshoot, repair, etc.) on the wind turbine, disabling the wind turbine to prevent damage or in anticipation of maintenance, etc. [0037] FIG. 4 depicts a flow diagram of another method 400 for estimating energy loss for wind turbine performance monitoring, according to aspects of the present disclosure. The method 400 can be performed by any suitable processing system, such as the processing system 200, the processing system 800 of FIG. 8, suitable combinations thereof, and/or another suitable processing system. In particular, the method 400 identifies a wind turbine of interest, determines a subset of wind turbines have less wind speed than the wind turbine of interest, and compares the power generated by the wind turbine of interest to the wind turbines with the maximum power generation of the subset. The power loss for the turbine of interest can then be determined.

[0038] At block 402, wind speed is measured (e.g., using an anemometer) at each of the wind turbines (e.g., the wind turbines 101-105) to generate wind speed data. At block 404, power generation is measured (e.g., using a power meter) at each of the wind turbines (e.g., the wind turbines 101-105) to generate power generation data. Blocks 402 and 404 occur concurrently such that the wind speed data and the power generation data are correlated with respect to time (e.g., a wind speed data point correlates to a power generation data point at the same time t). The wind speed data and the power generation data are sent to the data engine 210. The wind speed data is depicted in FIG. 5, while the power generation data are depicted in FIG. 6

[0039] Regarding the wind speed data, FIG. 5 depicts graphs 501, 502, 503, 504, 505 of wind speed weekly time series (collectively "wind graphs 501-505") for the wind turbines 101-105 of FIG. 1, according to aspects of the present disclosure. The wind speed is expressed in meters per second (m/s or ms^"1) over time for a period of two years (e.g., from 01/01/2013 through 12/31/2014).

[0040] Regarding the power generation data, FIG. 6 depicts graphs 601, 602, 603, 604, 605 of power weekly time series (collectively "power graphs 601-605") for the wind turbines 101-105 of FIG. 1 , according to aspects of the present disclosure. The power graphs 601-605 illustrate the normalized power of each of the wind turbines 101- 105 over a period of two years (e.g., from 01/01/2013 through 12/31/2014). In this example, the graph 601 corresponds to the wind turbine 101, the graph 602 corresponds to the wind turbine 102, etc. According to aspects of the present disclosure, the power generation is expressed as a ratio of the capacity of the wind turbine generator. For example, the ratio of capacity is expressed as the actual real power of a wind turbine generator divided by the capacity of the wind turbine generator. This provides a normalized power measure, which is depicted in FIG. 6.

[0041] It should be appreciated that the wind speed of a wind turbine as depicted in FIG. 5 has a positive correlation to the power for the turbine depicted in FIG. 6. It should be expected that as wind speed increases at a wind turbine, the power generated by a wind turbine also increases.

[0042] With continued reference to FIG. 4, at block 405, i is set to 1, where i represents a wind turbine of interest. Blocks 406, 408, 410, 412, 414, and 416 are now described, which can be performed by the power loss estimation engine 212. At block 406, a wind turbine (TB) is designated as a turbine of interest from the group of wind turbines. For example, wind turbine 101 is designated as a turbine of interest from the group of wind turbines 101-105. At block 408, a subset of the group of wind turbines is identified. The members of the subset are wind turbines with a wind speed less than the wind speed of the wind turbine of interest. For example, the wind turbines 102, 104, and 105 may have wind speeds less than the wind turbine 101 (i.e., the turbine of interest). At block 410, one of the wind turbines of the subset is identified as having the greatest power generation. For example, of the wind turbines 102, 104, and 105 of the subset, the wind turbine 105 may be determined to have the greatest power generation. At block 412, the power loss for the wind turbine of interest is determined by comparing the power generated by the wind turbine of interest against the power generated by the wind turbine of the subset with the greatest power generation. For example, the power generated by the wind turbine 101 is compared to the power generated by the wind turbine 105. To the extent that the power generated by the wind turbine of interest is less than the power generated by the wind turbine of the subset with the greatest power generation, a power loss is determined to exist.

[0043] At decision block 414 of FIG. 4, it is determined whether the value /^' for the turbine of interest is equal to the total number n of wind turbines in the geographic area. If not, the method 400 continues to block 416 and the value /^' is incremented such that a next wind turbine is designated as the wind turbine of interest at block 406. In this way, each wind turbine of the group of wind turbines is designated as the wind turbine of interest iteratively until the power lost for each wind turbine is determined at block 412. Once the value /^' is equal to the number n at decision block 414, a power loss graph can be displayed (see, e.g., FIG. 7).

[0044] For example, FIG. 7 depicts graphs of power loss weekly time series 701, 702, 703, 704, 705 (collectively "power loss graphs 701-705") for the wind turbines 101-105 of FIG. 1, according to aspects of the present disclosure. The power loss graphs 701-705 plot a performance index as power loss per kilowatt generated over time. It should be expected that under normal operating conditions, the power loss for a wind turbine at a particular time is approximately zero. When the power loss rises, such as above a particular threshold, a problem with the wind turbine may exist. For example, the wind turbine may be experiencing a mechanical problem, may not be manipulating its blades to account for changes in wind speed/direction, or may be otherwise performing sub- optimally. Power loss may also be caused, for example, by a wind turbine being temporarily shut down, such as for maintenance, which may be the cause for the extreme power loss spike in power loss graph 702 around August 2014. Generally, the peaks of the power loss graphs 701-705 represent more significant power loses, while the valleys (i.e., plots around zero) represent less (or no) power loss.

[0045] By identifying instances of power loss in a wind turbine, problems with the wind turbine can be addressed quickly and efficiently in order to increase the operating efficiency of the wind turbine to improve power generation. By displaying the power loss graph at block 418, an operation/technician may be alerted to the power loss problem.

[0046] At decision block 420, it is determined whether power loss exists. This may include comparing the power loss for each wind turbine to an adjustable power loss threshold to determine whether the power loss is problematic. According to examples of the present disclosure, it can be determined whether the power loss exceeds the adjustable power loss threshold. For example, if a power loss threshold is set to 0.05%, any power loss below this power loss threshold may be considered normal, while power loss in excess of this power loss threshold may be considered problematic. It should be appreciated that other power loss thresholds can be implemented and that the power loss threshold can be adjustable for different geographic environments, operating conditions, etc. and can even be adjusted on a turbine-by-turbine basis.

[0047] If at decision block 420 it is determined that a power loss has occurred, an action can be implemented. Examples of corrective actions can be issuing an alert to an operator, dispatching a technician to perform maintenance (e.g., troubleshoot, repair, etc.) on the wind turbine, disabling the wind turbine to prevent damage or in anticipation of maintenance, etc. If at decision block 420 it is determined that no power loss has occurred and/or upon the action being implemented at block 422, the method 400 can repeat or end.

[0048] The present techniques provide several benefits. For example, by determining potential power loss of a wind turbine, problems with the wind turbine can be identified and corrected, thereby increasing the efficiency of the wind turbine to generate power. The present techniques utilize fewer data and are easier to understand and implement that existing wind turbine performance monitoring techniques. For example, no power curve needs to be constructed to predict real power as in existing techniques. [0049] Example embodiments of the disclosure include or yield various technical features, technical effects, and/or improvements to technology. Example embodiments of the disclosure provide for wind turbine performance monitoring by using wind speed data and power generation data to estimate a potential power loss for a wind turbine, such as by comparing the wind speed data and the power generation data for the wind turbines to the wind speed data and the power generation data for the other wind turbines in the same geographic area, without the need to construct a power curve (e.g., a statistical or actual performance curve). These aspects of the disclosure constitute technical features that yield the technical effect of improving power generation efficiency in wind turbines by estimating power loss and using the power loss to implement actions to correct a problem with the wind turbine. As a result of these technical features and technical effects, a wind turbine performance monitoring system in accordance with example embodiments of the disclosure represents an improvement to wind turbine performance monitoring techniques. Accordingly, wind turbines can generate power more efficiently. It should be appreciated that the above examples of technical features, technical effects, and improvements to the technology of example embodiments of the disclosure are merely illustrative and not exhaustive. These and other benefits will be apparent as described herein.

[0050] It is understood that embodiments of the present invention are capable of being implemented in conjunction with any other suitable type of computing environment now known or later developed. For example, FIG. 8 illustrates a block diagram of a processing system 800 for implementing the techniques described herein. In examples, processing system 800 has one or more central processing units (processors) 821a, 821b, 821c, etc. (collectively or generically referred to as processor(s) 821 and/or as processing device(s)). In aspects of the present disclosure, each processor 821 may include a reduced instruction set computer (RISC) microprocessor. Processors 821 are coupled to system memory (e.g., random access memory (RAM) 824) and various other components via a system bus 833. Read only memory (ROM) 822 is coupled to system bus 833 and may include a basic input/output system (BIOS), which controls certain basic functions of processing system 800.

[0051] Further illustrated are an input/output (I/O) adapter 827 and a communications adapter 826 coupled to system bus 833. I/O adapter 827 may be a small computer system interface (SCSI) adapter that communicates with a hard disk 823 and/or a tape storage drive 825 or any other similar component. I/O adapter 827, hard disk 823, and tape storage device 825 are collectively referred to herein as mass storage 834. Operating system 840 for execution on processing system 800 may be stored in mass storage 834. A network adapter 826 interconnects system bus 833 with an outside network 836 enabling processing system 800 to communicate with other such systems.

[0052] A display (e.g., a display monitor) 835 is connected to system bus 833 by display adaptor 832, which may include a graphics adapter to improve the performance of graphics intensive applications and a video controller. In one aspect of the present disclosure, adapters 826, 827, and/or 832 may be connected to one or more I O busses that are connected to system bus 833 via an intermediate bus bridge (not shown). Suitable I/O buses for connecting peripheral devices such as hard disk controllers, network adapters, and graphics adapters typically include common protocols, such as the Peripheral Component Interconnect (PCI). Additional input/output devices are shown as connected to system bus 833 via user interface adapter 828 and display adapter 832. A keyboard 829, mouse 830, and speaker 831 may be interconnected to system bus 833 via user interface adapter 828, which may include, for example, a Super I/O chip integrating multiple device adapters into a single integrated circuit.

[0053] In some aspects of the present disclosure, processing system 800 includes a graphics processing unit 837. Graphics processing unit 837 is a specialized electronic circuit designed to manipulate and alter memory to accelerate the creation of images in a frame buffer intended for output to a display. In general, graphics processing unit 837 is very efficient at manipulating computer graphics and image processing, and has a highly parallel structure that makes it more effective than general-purpose CPUs for algorithms where processing of large blocks of data is done in parallel.

[0054] Thus, as configured herein, processing system 800 includes processing capability in the form of processors 821, storage capability including system memory (e.g., RAM 824), and mass storage 834, input means such as keyboard 829 and mouse 830, and output capability including speaker 831 and display 835. In some aspects of the present disclosure, a portion of system memory (e.g., RAM 824) and mass storage 834 collectively store an operating system 840 to coordinate the functions of the various components shown in processing system 800.

[0055] The descriptions of the various examples of the present disclosure have been presented for purposes of illustration, but are not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described techniques. The terminology used herein was chosen to best explain the principles of the present techniques, the practical application or technical improvement over technologies found in the marketplace, or to enable others of ordinary skill in the art to understand the techniques disclosed herein.

Claims

CLAIMS What is claimed is:

1. A computer-implemented method for wind turbine performance monitoring, the method comprising: receiving, by a processing device, wind speed data and power generation data for each of a plurality of wind turbines, wherein the plurality of wind turbines are distributed through a geographic area; estimating, by the processing device, a potential power loss for at least one of the plurality of wind turbines based at least in part on the wind speed data and the power generation data; and implementing a corrective action when it is determined that the potential power loss for the at least one of the plurality of wind turbines is greater than a threshold power loss.

2. The computer-implemented method of claim 1, wherein the plurality of wind turbines distributed throughout the geographic area experience similar environmental conditions.

3. The computer-implemented method of claim 1, wherein estimating the potential power loss for at least one of the plurality of wind turbines further comprises comparing the wind speed data and the power generation data for the at least one plurality of wind turbines to the wind speed data and the power generation data for the other of the plurality of wind turbines.

4. The computer-implemented method of claim 1 , further comprising collecting the wind speed data and the power generation data from sensors associated with the plurality of wind turbines.

5. The computer-implemented method of claim 1, wherein estimating the potential power loss comprises: designating, by the processing device, one of the plurality of wind turbines as a wind turbine of interest; identifying, by the processing device, a subset of the plurality of wind turbines excluding the turbine of interest as having a wind speed less than a wind speed of the wind turbine of interest; identifying, by the processing device, a wind turbine of the subset that has a greatest power generation; and determining, by the processing device, the potential power loss for the wind turbine of interest based at least in part on the wind turbine that has the greatest power generation.

6. The computer-implemented method of claim 1 , wherein the potential power loss is expressed as a performance index.

7. The computer-implemented method of claim 1, wherein the corrective action comprises dispatching a technician to repair the at least one of the plurality of wind turbines.

8. The computer-implemented method of claim 1 , wherein the corrective action comprises temporarily disabling the at least one of the plurality of wind turbines.

9. The computer-implemented method of claim 1, further comprising plotting, by the processing device, a power loss graph for the at least one wind turbine based at least in part on the potential power loss for the at least one wind turbine.

10. A system for wind turbine performance monitoring, the system comprising: a plurality of wind turbines; a memory comprising computer readable instructions; and a processing device for executing the computer readable instructions for performing a method, the method comprising: receiving, by the processing device, wind speed data and power generation data for each of the plurality of wind turbines, wherein the plurality of wind turbines are distributed through a geographic area; estimating, by the processing device, a potential power loss for at least one of the plurality of wind turbines based at least in part on the wind speed data and the power generation data; and implementing a corrective action when it is determined that the potential power loss for the at least one of the plurality of wind turbines is greater than a threshold power loss.

11. The system of claim 10, wherein the plurality of wind turbines distributed throughout the geographic area experience similar environmental conditions.

12. The system of claim 10, wherein estimating the potential power loss for at least one of the plurality of wind turbines further comprises comparing the wind speed data and the power generation data for the at least one plurality of wind turbines to the wind speed data and the power generation data for the other of the plurality of wind turbines.

13. The system of claim 10, wherein the method further comprises collecting the wind speed data and the power generation data from sensors associated with the plurality of wind turbines.

14. The system of claim 10, wherein estimating the potential power loss comprises: designating, by the processing device, one of the plurality of wind turbines as a wind turbine of interest; identifying, by the processing device, a subset of the plurality of wind turbines excluding the turbine of interest as having a wind speed less than a wind speed of the wind turbine of interest; identifying, by the processing device, a wind turbine of the subset that has a greatest power generation; and determining, by the processing device, the potential power loss for the wind turbine of interest based at least in part on the wind turbine that has the greatest power generation

15. The system of claim 10, wherein the potential power loss is expressed as a performance index.

16. The system of claim 10, wherein the corrective action comprises dispatching a technician to repair the at least one of the plurality of wind turbines.

17. The system of claim 10, wherein the corrective action comprises temporarily disabling the at least one of the plurality of wind turbines.

18. The system of claim 10, wherein the method further comprises plotting, by the processing device, a power loss graph for the at least one wind turbine based at least in part on the potential power loss for the at least one wind turbine

19. A computer program product for wind turbine performance monitoring, the computer program product comprising: a computer readable storage medium having program instructions embodied therewith, wherein the computer readable storage medium is not a transitory signal per se, the program instructions executable by a processing device to cause the processing device to perform a method comprising: receiving, by the processing device, wind speed data and power generation data for each of a plurality of wind turbines, wherein the plurality of wind turbines are distributed through a geographic area; estimating, by the processing device, a potential power loss for at least one of the plurality of wind turbines based at least in part on the wind speed data and the power generation data; and implementing a corrective action when it is determined that the potential power loss for the at least one of the plurality of wind turbines is greater than a threshold power loss.

20. The computer program product of claim 19, wherein estimating the potential power loss comprises: designating, by the processing device, one of the plurality of wind turbines as a wind turbine of interest; identifying, by the processing device, a subset of the plurality of wind turbines excluding the turbine of interest as having a wind speed less than a wind speed of the wind turbine of interest; identifying, by the processing device, a wind turbine of the subset that has a greatest power generation; and determining, by the processing device, the potential power loss for the wind turbine of interest based at least in part on the wind turbine that has the greatest power generation