GB2586227A

GB2586227A - System and method for controlling shadow flicker from a wind turbine

Info

Publication number: GB2586227A
Application number: GB1911306.7A
Authority: GB
Inventors: Clerc Alex; Cameron Lee; Donnelly Ross
Original assignee: Renewable Energy Systems Ltd
Current assignee: Renewable Energy Systems Ltd
Priority date: 2019-08-07
Filing date: 2019-08-07
Publication date: 2021-02-17
Anticipated expiration: 2039-08-07
Also published as: GB201911306D0; GB2586227B

Abstract

A system for preventing shadow flicker from a wind turbine 2, comprising at least one camera 4 arranged to send images of the turbine’s local environment to a computer 5, wherein the computer analyses the images (preferably through machine learning techniques) to determine if shadow flicker or is present, or is expected to occur. On detection of shadow flicker, the turbine can be shut down by a controller 3, and restart when flicker is no longer detected. The computer can be trained to recognise erroneous readings, and alert a user to this error. A timer can be included so the system only operates at suitable times, to save power.

Description

System and method for controlling shadow flicker from a wind turbine The present disclosure relates to a system and method for controlling shadow flicker from a wind turbine, in particular to such a system and method in which control is automatic and is based on the analysis of images from a camera.

There is a significant desire, the world over, to limit reliance on fossil fuels in the production of electricity. Wind power represents one of the cleanest and most environmentally friendly energy sources presently available. This has resulted in a significant interest in the installation of new wind turbines, and wind farms comprising multiple wind turbines, where possible. Due to limited suitable sites for the installation of wind farms, there has been an inevitable juxtaposition of people and wind turbines.

Control of wind turbines is necessary to limit their impact. For example, noise regulations exist. Under such regulations, the noise emitted from wind turbines that are located close to people must be carefully controlled. Another factor that requires consideration is so-called "shadow flicker". Shadow flicker is the flickering effect caused when rotating wind turbine blades periodically cast shadows, particularly through constrained openings such as the windows of neighbouring properties. Shadow flicker is dependent on a number of environmental factors such as wind speed and direction, the position and point of the sun and cloudiness. Shadow flicker is typically worse when the sun is low in the sky, at sunrise or sunset, or during the winter, which is also when there are commonly faster wind speeds.

Any control for shadow flicker, which typically requires the shutdown of one or more wind turbines, has an impact on a wind farm's yield. Due to the significant capital investment required to install a wind farm, it is critical that any wind farm operates as close as possible to maximum efficiency, whilst avoiding problematic shadow flicker.

Prior art systems exist for controlling shadow flicker. Such systems generally include optical light meters that are used to derive information about the angle and position of the sun. This information is used for monitoring the relative light levels on opposite sides of a wind turbine. Based on the relative light levels, the risk of shadow flicker is calculated and this information is used to control the wind turbine so as to avoid or reduce shadow flicker by the shutdown of the wind turbine.

Such prior art arrangements have not been sufficiently reliable. They have been prone to making wrong decisions that are not representative of the conditions being experienced. Other than general inaccuracies with the systems, there has been a problem with the fact that the sensors are unable to distinguish between errors, caused for example by the blocking of the sensors by dirt, bird waste, or otherwise, and sensed atmospheric conditions. With such inaccuracies, prior art systems are prone to allow a turbine to run when shadow flicker exists (which acts to aggravate the neighbours of wind farms) or to shut down a turbine when shadow flicker does not exist (reducing the efficiency of wind farms with excessive curtailment of the wind turbines).

The present invention arose in a bid to provide a more accurate control system for more effective/efficient control of shadow flicker.

According to the present invention in a first aspect, there is provided a system for controlling shadow flicker from a wind turbine, the system comprising a controller for controlling the operation of the wind turbine, and a camera, wherein a computer is arranged to receive and automatically analyse images received from the camera to determine whether shadow flicker is present and/or whether atmospheric conditions are such that shadow flicker is expected to occur, and the controller is configured to control the operation of the wind turbine in dependence on the results of the analysis.

The control of shadow flicker will require some curtailment of the wind turbine's operation. For such purposes, it is preferable that the controller is arranged to shut down the wind turbine upon determination of the presence of shadow flicker and is arranged to restart the wind turbine upon determination that atmospheric conditions are such that shadow flicker will not occur. The controller is further arranged to report detected errors when a condition is detected that does not meet expectations. This may be the case if the view of the camera is obstructed somehow, such as by dirt on the lens or otherwise.

The camera may comprise any suitable digital camera and most preferably comprises a digital video camera. The computer preferably analyses a live video feed in real time. Accordingly, the system may respond without delay in the event shadow flicker occurs and, moreover, may minimise any down time whilst eliminating shadow flicker, since the wind turbine may be restarted following a shut down as soon as it is determined that shadow flicker will not occur.

The computer preferably implements machine learning techniques to analyse the images received from the camera.

The machine learning analysis may be provided on behalf of any number of machine learning algorithms and trained models, including but not limited to deep learning models (also known as deep machine learning, or hierarchical models) that have been trained to perform image recognition for shadow flicker or for visually identifiable atmospheric conditions giving rise to shadow flicker from training images. As used herein, the term "machine learning" is used to refer to the various classes of artificial intelligence algorithms and algorithm-driven approaches that are capable of performing machine-driven (e.g., computer-aided) identification of trained image characteristics, with the term "deep learning" referring to a multiple-level operation of such machine learning algorithms using multiple levels of representation and abstraction. However, it will be apparent that the role of the machine learning algorithms that are applied, used, and configured in the presently described system may be supplemented or substituted by any number of other algorithm-based approaches, including variations of artificial neural networks, learning-capable algorithms, trainable object classifications, and other artificial intelligence processing techniques, as will be appreciated by those skilled in the art.

In accordance with preferred arrangements, the computer is arranged to run a machine learning model, the machine learning model being trained to evaluate image characteristics associated with shadow flicker; the computer determining a state of shadow flicker, a state of potential shadow flicker, or an error state, from the received image data, using the machine learning model.

A state of shadow flicker will be determined when the received images correlate with those training images in which shadow flicker is determined to be present. A state of potential shadow flicker will be determined when the received images correlate with those training images in which the turbine is not running but it is determined that shadow flicker would be present were the turbine running. An error state will be determined when the images do not correlate with any of the training images.

With machine learning techniques implemented a high degree of accuracy can be achieved. This is especially so, since the phenomenon of shadow flicker is typically at least partly subjective. Moreover, errors may be identified.

A training data set of received images may be classified (as either having shadow flicker or not) manually by an expert taking into consideration any specific sensitivities of local residents to shadow flicker. The machine learning model is then trained using this training data set so that the machine learning model can classify new images in an analogous way to the human expert. The machine learning model subsequently determines shadow flicker from received images in the live system. As above, the machine learning model may further detect errors, wherein if the camera is obstructed the images will not correspond to the images of the training data set. An alert may be sent to an operator to allow for a suitable fix. The alert may be sent by the computer in the form of an SMS message, an email, or in any other suitable form, as will be readily appreciated by those skilled in the art. The obstruction will not be misclassified as in the case of the prior art light sensors.

The system may comprise a timer for restricting operation of the system to predetermined time periods. The timer may, for example, be set to allow operation of the system only within predetermined periods each day, which periods are determined to be the only periods in which shadow flicker may occur, such as a predetermined period during which sunrise will occur and a predetermined period during which sunset will occur. Outside these predetermined periods, the system may enter a low power or "sleep" state, or otherwise. The predetermined periods and their timings may alter through the year.

According to the present invention in a further aspect, there is provided a method for controlling shadow flicker from a wind turbine, the method comprising receiving and automatically analysing images received from a camera to determine whether shadow flicker is present and/or whether atmospheric conditions are such that shadow flicker is expected to occur, and controlling the operation of the wind turbine in dependence on the results of the analysis.

Non-limiting embodiments of the present invention will now be described, by way of example only, with reference to the accompanying drawings, in which: Figure 1 is a simplified schematic view of an embodiment of the present zo invention; and Figure 2 is a flowchart showing the operation of a system according to an embodiment of the present invention.

With reference to the figures, embodiments of the present invention will now be discussed.

In accordance with a preferred arrangement, there is provided a system 1 for controlling shadow flicker from a wind turbine 2. The system 1 comprises a controller 3 for controlling the operation of the wind turbine 2, and a camera 4. A computer 5, which in the present arrangement comprises the controller but in other arrangements could be provided separately, is arranged to receive and automatically analyse images received from the camera to determine whether shadow flicker is present and/or whether atmospheric conditions are such that shadow flicker is expected to occur. The controller is configured to control the operation of the wind turbine in dependence on the results of the analysis.

Control of the operation of the wind turbine will comprise curtailment of the wind turbine. The wind turbine 2 will generally be shut down upon determination of the presence of shadow flicker under control of the controller 3. The wind turbine 2 will generally be restarted under control of the controller 3 upon determination that atmospheric conditions are such that shadow flicker will not occur.

The camera preferably comprises a video camera and may comprise any suitable form of digital video camera. It will generally comprise an IP (Internet Protocol) video camera that supplies a live video feed to the computer 5 as required.

The video camera will be provided at a suitable measuring point. Various measuring points will be possible. The camera will be located where it can see an open view of the relevant part of the sky for shadow flicker (the pad of the sky where the sun will be when shadow flicker is possible to occur). The camera may be mounted to the wind turbine or elsewhere within a wind farm comprising the wind turbine, such as at a wind farm substation (if the substation building has good views of the sky).

It is to be noted that whilst the present arrangement comprises a single wind turbine and a single camera, there may be provided multiple wind turbines and/or multiple cameras. The system may be scaled as required. There may in some arrangements be a plurality of cameras for each wind turbine or alternatively, a single camera for a plurality of wind turbines.

In some arrangements, the computer and/or controller may receive additional inputs from a Supervisory Control and Data Acquisition (SCADA) system (7, Figure 2) of the wind turbine. The SCADA system may monitor meteorological conditions, including but not limited to wind direction and speed, and may comprise suitable sensors for such purposes. Such information could, however, alternatively be gathered by suitable sensors associated with the controller 3 rather than the SCADA system 7.

Figure 2 shows an operational flowchart of an arrangement of the system that is used for control of a wind farm comprising a plurality of wind turbines. The system comprises a plurality of cameras.

Claims

Claims 1. A system for controlling shadow flicker from a wind turbine, the system comprising a controller for controlling the operation of the wind turbine, and a camera, wherein a computer is arranged to receive and automatically analyse images received from the video camera to determine whether shadow flicker is present and/or whether atmospheric conditions are such that shadow flicker is expected to occur, and the controller is configured to control the operation of the wind turbine in dependence on the results of the analysis.
2. A system as claimed in Claim 1, wherein the computer analyses a live video feed from the camera in real time.
3. A system as claimed in Claim 1 or 2, wherein the controller is arranged to shut down the wind turbine upon determination of the presence of shadow flicker.
4. A system as claimed in Claim 3, wherein the controller is arranged to restart the wind turbine upon determination that atmospheric conditions are such that shadow flicker will not occur.
5. A system as claimed in any preceding claim, wherein the computer implements machine learning techniques to analyse the images.
6. A system as claimed in any preceding claim, wherein the computer is arranged to run a machine learning model, the machine learning model being trained to evaluate image characteristics associated with shadow flicker; the computer determining a state of shadow flicker, a state of potential shadow flicker, or an error state from the received image data, using the machine learning model.
7. A system as claimed in Claim 6, wherein upon determination of an error state, the computer is arranged to send an alert to a user.
8. A system as claimed in any preceding claim, which further comprises a timer, which is arranged to limit operation of the system to predetermined time periods.
9. A method for controlling shadow flicker from a wind turbine, the method comprising receiving and automatically analysing images received from a camera to determine whether shadow flicker is present and/or whether atmospheric conditions are such that shadow flicker is expected to occur, and controlling the operation of the wind turbine in dependence on the results of the analysis.