WO2019158171A1

WO2019158171A1 - Systems and vehicles for managing wind turbine systems

Info

Publication number: WO2019158171A1
Application number: PCT/DK2019/050035
Authority: WO
Inventors: Ingemann Hvas Sandvad; Bjarne IVERSEN
Original assignee: Vestas Wind Systems A/S
Priority date: 2018-02-13
Filing date: 2019-01-29
Publication date: 2019-08-22

Abstract

Systems and vehicles (312, 314) for managing wind turbine systems (300) are disclosed, in particular for managing the aerial inspection of a wind park. A system comprises a plurality of drones (312, 314) or unmanned aerial vehicles (UAVs) and a drone control computer (322). The drone control computer is configured to determine (600) inspection missions for each of the plurality of drones, to provide (602) mission data corresponding to those inspection missions to respective drones and to dispatch (604) said drones on said inspection missions. At least one of the plurality of drones may be configured to implement (606) autonomous flight routines so as to rendezvous with a respective wind turbine (204, 206) without operator (320) guidance.

Description

SYSTEMS AND VEHICLES FOR MANAGING WIND TURBINE SYSTEMS

FIELD OF THE INVENTION

This invention is directed to systems and vehicles for managing interactions between unmanned aircraft systems (UAS), using unmanned aerial vehicles (UAV) or drones, and wind turbine systems, and inspection of such wind turbine systems, in particular in wind parks.

BACKGROUND OF THE INVENTION

Wind turbines for power generation are well known in the art. In a common arrangement, a nacelle is mounted on a tower, with a rotor and blades being mounted on the nacelle. The rotor is mounted on a rotor shaft which is supported in the nacelle by a shaft housing.

Methods for conducting routine maintenance and inspection of wind turbines, wind turbine systems and wind turbine parks or arrays are known. In addition, other types of interaction with such turbines and their components, such as observing their function at close quarters, replacing parts, connecting

communications infrastructure, photographing the turbines or arrays and the like are also known.

Previously considered methods for conducting such inspections, maintenance and other such engagements usually involve manually interacting with the wind turbine, usually using a technician to conduct the inspection or maintenance or the like. These methods can require safety measures and precautions for the technician, additional equipment, and may require certain weather conditions in order to undertake maintenance. Unmanned aircraft systems (UAS) and unmanned aerial vehicles are also known to the art. However, these systems are typically ill-equipped for use in interacting in a wind turbine system environment.

The present invention aims to address these problems and provide

improvements upon the known devices and methods.

STATEMENT OF INVENTION

Aspects and embodiments of the invention are set out in the accompanying claims.

In general terms, one embodiment of a first aspect of the invention can provide a system for managing the aerial inspection of a wind park, the system comprising: a plurality of drones; and a drone control computer, wherein the drone control computer is configured to determine inspection missions for each of the plurality of drones, to provide mission data corresponding to those inspection missions to respective drones and to dispatch said drones on said inspection missions.

In this manner, individual drones or UAVs from a fleet can be given separate, specific missions or tasks, and dispatched separately for these missions. This provides simple management of inspection particularly of large wind farms or parks, by use of a fleet of drones with missions distributed amongst the fleet. In addition, each individual drone can have specific, separate instructions for each mission or task, relevant to the site or component that the drone is visiting (rather than a standard set of tasks for each drone).

For example, the drone control computer may be configured to determine a first inspection mission and a second inspection for a wind park, and assign the first mission to a first drone, and the second to a second drone. A mission may also be shared between two drones (second and third drones), or completed by a single drone in separate sorties. The mission data may comprise first and second mission data sets, each applied to the respective mission. The mission data may also comprise data common to the missions, such as a standard inspection procedure.

Optionally, at least one of the plurality of drones is configured to implement autonomous flight routines so as to rendezvous with a respective wind turbine without operator guidance.

Suitably, mission data for a respective inspection mission for a respective drone comprises an autonomous flight routine, and wherein the autonomous flight routine comprises an identification attribute of a component of a respective wind turbine.

In embodiment, the identification attribute comprises a visually readable identification means and/ or a radio frequency identification means.

In some embodiments, the radio frequency identification means (such as an RFID or NFC device) may have the advantage that it can be embedded in the component of the wind turbine and thus will not be subject to wear or dirt building up on the component.

Suitably, the visually readable identification means comprises a bar code or QR code. Suitably, upon a rendezvous with a wind turbine, the respective drones are configured to use an on-board sensor to capture data associated with the identification attribute of the component of the wind turbine, and to process the captured identification attribute to identify the wind turbine component.

Optionally, the on-board sensor comprises an image capture device, configured to capture image data of the identification attribute. In embodiments, the drone is configured to process the captured identification attribute data using a pattern recognition algorithm to identify the component.

In embodiments, the on-board sensor comprises a radio frequency device, configured to interact with a radio frequency identification means of the

identification attribute. The radio frequency identification means may comprise a radio frequency identification device (RFID), or a near-field communication (NFC) device.

Suitably, the autonomous flight routine comprises a plurality of identification attributes for respective components of the wind turbine, and the on-board sensor is configured to capture data associated with a selected one or more of the component identification attributes.

In embodiments, the drone control computer is configured to manage a set of inspection tasks for the wind park, the set comprising respective inspection tasks for respective wind turbines of the park, and the drone control computer is configured to use the set of inspection tasks to determine the inspection missions for each of the plurality of drones, and to provide mission data corresponding to those inspection missions to the respective drones.

Optionally, the drone control computer is configured to update the set of inspection tasks with data captured by one or more of the plurality of drones.

Suitably, the drone control computer is configured to modify an inspection mission for a given drone in response to an update of the set of inspection tasks. For example, the given drone may be given an updated mission with different parameters, or an additional mission, or a new or different mission. A different drone may be substituted for the given drone, for example if the given drone is malfunctioning, or if the updated task is only manageable by the replacement drone. If a task takes longer than predicted, a following mission for a given drone may be scrubbed, and given to another drone.

In embodiments, at least one of the plurality of drones is configured to monitor environmental conditions when in proximity to a respective wind turbine for an interaction therewith, and the drone is further configured to modify one or more flight parameters based on changes in the monitored environmental conditions.

Optionally, the drone is configured to compare an instant value for the monitored environmental condition with a threshold for the environmental condition; and activate modification if the value exceeds the threshold.

Suitably, at least one of the plurality of drones is configured to record

environmental condition data associated with the monitored environmental condition.

In embodiments, upon a rendezvous with a wind turbine, the respective drones are configured to use an on-board sensor to capture audio data associated with the component of the wind turbine.

Suitably, the drone control computer is configured to use respective unique identifiers for each of the plurality of drones, to dispatch the respective drones on the respective inspection missions.

In embodiments, the drone control computer is configured to, following dispatch of the plurality of drones, display to an operator a position for each of the plurality of drones.

Suitably the system comprises a plurality of relay stations, said relay stations configured to: communicate with one or more of the plurality of drones; and provide handoff between relay stations during a communication with a given drone.

One embodiment of another aspect of the invention can provide a wind turbine for a wind park for aerial inspection managed by a system according to any of the above described embodiments. Optionally, the wind turbine comprises at least one said component comprising an identification attribute.

One embodiment of another aspect of the invention can provide an unmanned aerial vehicle (UAV) for interacting with a wind turbine system, the vehicle comprising: a powered flight system for generating lift; at least one sensor; a memory; and a processor, wherein the vehicle is configured to: by the flight system, carry out an autonomous flight routine determined by a drone control computer; rendezvous with a wind turbine for an interaction; during the interaction, by the at least one sensor, capture data associated with an identification attribute of a component of the wind turbine; and by the processor, process the captured identification attribute to identify the wind turbine

component.

Suitably, the at least one sensor comprises an image capture device, configured to capture image data associated with the identification attribute.

In embodiments, the at least one sensor comprises a radio frequency device, configured to interact with a radio frequency identification means associated with the identification attribute. The at least one sensor may be an ultrasonic device.

One embodiment of another aspect of the invention can provide a drone control computer for a system according to any of the above described embodiments. One embodiment of another aspect of the invention can provide a method of managing an interaction between an unmanned aircraft system (UAS) and a wind turbine system, comprising: initiating flight of an unmanned aerial vehicle (UAV) to rendezvous with a wind turbine for an interaction; during the interaction, monitoring at least one environmental condition; and on variation of the at least one environmental condition relative to a threshold, modifying a flight parameter for the UAV.

The interaction may be an engagement of, exchange of information with, or inspection of the turbine system. The interaction could be close-range or at a given distance from the turbine. The monitor may use a sensor device. The condition may be near the turbine, or may be a general condition. The flight parameter may be an attitude, lift, direction, translation or other like parameter.

Optionally, the step of monitoring the environmental condition comprises detecting a weather condition. Suitably the step of monitoring the environmental condition comprises detecting a change in a wind parameter measured by a wind detection component of the UAV. In an embodiment, the wind detection component of the UAV comprises an accelerometer.

Suitably, the step of monitoring the environmental condition comprises using a light sensor to monitor a light level.

In embodiments the step of modifying comprises: comparing an instant value for the monitored environmental condition with a threshold for the environmental condition; and activating modification if the value exceeds the threshold.

Suitably, the step of modifying the flight parameter comprises instructing an alteration to a current flight characteristic of a powered flight system of the UAV.

Optionally, the step of instructing the alteration comprises temporarily increasing lift generated by the powered flight system. Suitably, the method comprises, during the interaction, detecting a distance between the UAV and the wind turbine, for example by range-sensing device such as a radar or ultrasound device.

In embodiments, the method comprises, during the interaction, recording environmental condition data associated with the monitored environmental condition.

One embodiment of another aspect of the invention can provide a method of managing an interaction between an unmanned aircraft system (UAS) and a wind turbine system, comprising: initiating flight of an unmanned aerial vehicle (UAV) to rendezvous with a wind turbine for an interaction; during the interaction, monitoring at least one environmental condition; and recording environmental condition data associated with the monitored environmental condition.

Optionally, the method comprises, on variation of the at least one environmental condition relative to a threshold, modifying a flight parameter for the UAV.

Suitably, the method comprises, recording flight parameter modification data associated with the modification of the flight parameter for the UAV. In this manner the modifications may be used for machine learning or for iterative adaptation of the interactions between the drones and the turbine systems, for instance to learn conditions which are common, or learn paths or manoeuvres which may be different from those programmed, but which are commonly required to be altered during inspections, or certain types of inspection.

One embodiment of another aspect of the invention can provide a method of managing an interaction between an unmanned aircraft system (UAS) and a wind turbine system, comprising: initiating flight of an unmanned aerial vehicle (UAV) to rendezvous with a wind turbine for an interaction; using a sensor device of the UAV to capture data associated with an identification attribute of a component of the wind turbine; and processing the captured identification attribute data to identify the wind turbine component. The attribute may be a marking, image or code associated with the component.

Suitably, the identification attribute comprises a visually readable identification means and/ or a radio frequency identification means.

Suitably, the visually readable identification means comprises a bar code or QR code. In embodiments, the step of using a sensor device comprises using an image capture device of the UAV to capture image data of the identification attribute.

Optionally, the step of processing the captured identification attribute data comprises using a pattern recognition algorithm to identify the component.

Suitably, the wind turbine comprises a plurality of components each having different identification attributes, and wherein the step of using the sensor device comprises capturing data associated with a selected one or more of the component identification attributes. The data may be capture for the selected component(s) at the same time, or in turn.

One embodiment of another aspect of the invention can provide a wind turbine for implementing a method of managing an interaction with an unmanned aircraft system according to any of the above described embodiments, the turbine comprising at least one said component comprising an identification attribute. One embodiment of another aspect of the invention can provide an unmanned aerial vehicle (UAV) for interacting with a wind turbine system, the vehicle comprising: a powered flight system for generating lift; at least one environmental monitoring sensor; a memory; and a processor, wherein the vehicle is configured to: by the flight system, rendezvous with a wind turbine for an interaction; during the interaction, monitor by the at least one environmental monitoring sensor at least one environmental condition; and (a) by the memory, record environmental condition data associated with the monitored environmental condition; or (b) by the processor, on variation of the at least one environmental condition relative to a threshold, modify a flight parameter for the flight system of the UAV.

In one embodiment of the drone control computer according to the invention a drone automatically via the central drone controller and a SCADA system in the wind farm can stop and start the wind turbines before and after missions.

In this embodiment the drone may be able to configure the wind turbine in a safe mode for performing the inspection. For example such that the rotor is stationary.

In one embodiment of the drone control computer according to the invention a drone can be dispatched automatically on an inspection mission of a wind turbine by the drone control computer when said wind turbine is being shut down.

By dispatching a drone to do an inspection automatically the time while the wind turbine is shut down has a number of advantages. The first being that a scheduled inspection action at a fixed time causing the wind turbine to shut down and thereby loose production may be avoided if the inspection can replace the scheduled inspection. Additionally if the turbine shut down is due to an error on the turbine which can be detected by a drone, will enable the detected

information to be forwarded to a remote wind part operator and prevent restart of a faulty turbine. For example in case of shut downs due to vibration caused by blade damage or fire in the turbine. Further aspects of the invention comprise computer programs which, when loaded into or run on a computer, cause the computer to become apparatus, or to carry out methods, according to the aspects described above.

Processor and/or controllers may comprise one or more computational processors, and/or control elements having one or more electronic processors. Uses of the term“processor” or“controller” herein should therefore be considered to refer either to a single processor, controller or control element, or to pluralities of the same; which pluralities may operate in concert to provide the functions described. Furthermore, individual and/or separate functions of the processor(s) or controller(s) may be hosted by or undertaken in different control units, processors or controllers.

To configure a processor or controller, a suitable set of instructions may be provided which, when executed, cause said control unit or computational device to implement the techniques specified herein. The set of instructions may suitably be embedded in said one or more electronic processors. Alternatively, the set of instructions may be provided as software to be executed on said computational device.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention will now be described by way of example with reference to the accompanying drawings, in which:

Figure 1 is a diagram illustrating a typical wind turbine arrangement;

Figure 2 is a diagram illustrating components of an unmanned aerial system and a wind turbine generator array according to an embodiment of the invention; Figure 3 is a diagram illustrating further components of an unmanned aerial system and a wind turbine generator array according to an embodiment of the invention;

Figure 4 is a diagram illustrating components, structure and functionality of a computer management system or on-board computer system, according to embodiments of the invention;

Figure 5 is a diagram illustrating steps of a method according to an embodiment of the invention; and

Figure 6 is a diagram illustrating steps of a method according to an embodiment of the invention.

DETAILED DESCRIPTION OF EMBODIMENTS

Embodiments of the invention provide unmanned aircraft systems (UAS) and unmanned aerial vehicles (UAV) or drones, in order to provide a means to interact with wind turbines, whether individually or in wind farms, parks and arrays. In particular embodiments, fleets of such drones are used to provide the interactions, such as maintenance inspections. Embodiments can in particular provide simple management of inspection particularly of large wind farms or parks, by use of a fleet of drones with missions distributed amongst the fleet.

This can simplify procedures which would otherwise require technicians, or technicians using drones, and use of a fleet of drones with a specifically assigned series of tasks or missions can reduce the time incurred in inspections.

In addition, each individual drone can have specific, separate instructions for each mission or task, relevant to the site or component that the drone is visiting (rather than a standard set of tasks for each drone). This can provide more flexibility in completing the overall fleet’s inspection event or round; for example, the separate instructions can if necessary be altered on-the-fly, without affecting other drone instructions and/or with minimal disruption of the overall fleet inspection.

In contrast with previously considered methods in which UAVs merely captured image data of or around a wind turbine, embodiments of the invention determine and record environmental conditions during UAV missions or sorties, such as precise local aerodynamic conditions around the turbines and their blades, allowing greater safety for the UAV and the turbine. These conditions can then be learned to provide these safety factors for future interactions.

Furthermore, embodiments of the invention provide identification attributes on components of the wind turbines, which can be read by the UAVs. This provides a much more efficient and reliable means of selecting components of the turbines for inspection, particularly components where access for a technician would otherwise be difficult.

Embodiments of the invention can assist the human-controlled supervision and maintenance of a wind farm with automated drone inspections. Drones or UAVs can access parts of the turbines faster than service technicians can, especially offshore, and with fewer necessary safety precautions, safety equipment, and in some cases in more inclement weather conditions. For example, a central service office in a wind farm can send out one or more drones, for example a swarm or flock, to do specified tasks, primarily video inspections, on specific wind turbines in the wind farm.

The drones can return pictures/video of the observed items back to a database for later or real-time inspection and decisions on actions by the wind farm operator. Alternatively, the drones can stream data directly and receive commands for new task on-the-fly. The tasks may be occasional, i.e. looking after lightning hits on blades after a thunderstorm, or regular such as

investigating rust/cracks in welding on the towers every second year. To navigate within the wind farm, the drones or UAVs can use conventional GPS technology.

In particular embodiments, the control software algorithm is adaptive and can ‘learn’ the 3D structure of the type of wind turbine, so it knows how to manoeuvre around the turbine structure. In addition, the UAV can measure and/or learn specific aerodynamic variances that can be present close to the turbine blades.

In embodiments, labels on the turbine and its components indicate identification attributes of these items, e.g. by means of QR-codes, so the drone can orientate itself, for instance to find a specific blade.

Referring to Figure 1 , a wind turbine 1 according to an embodiment of the invention comprises a tower 2, a nacelle 4 rotatably coupled to the top of the tower 2 by a yaw system 6, a rotating hub 8 mounted to the nacelle 4 and a plurality of wind turbine rotor blades 10 coupled to the hub 8. The nacelle 4 and rotor blades 10 are turned and directed into the wind direction by the yaw system 6. The nacelle 4 may house generating components of the wind turbine, including the generator, gearbox, drive train and brake assembly, as well as convertor equipment for converting the mechanical energy of the wind into electrical energy for provision to the grid. It may be noted that“direct drive” wind turbines that do not use gearboxes are also known; a gearbox may therefore be optional. Identification attributes (102) such as QR-codes or RFIDs may be affixed to or incorporated in areas or components of the turbine, such as on the tower or on the blades.

Figures 2 and 3 are diagrams illustrating components of an unmanned aircraft system (UAS) and a wind turbine generator array according to an embodiment of the invention. A UAV (202) is in embodiments one of a number of a fleet which can be deployed to inspect and interact with wind turbines (206). Figure 3 illustrates two such UAVs 312 and 314, each interacting with different turbines (206) of a wind turbine park or array 300.

The UAVs can be deployed relatively autonomously, in that they are able to follow pre-programmed schedules and find turbine locations, but can also be controlled on-the-fly by operators (320) at a base station or central drone control station, using a suitable user terminal or interface 322.

The communications between the base station control and the UAVs are in embodiments propagated by relay stations 204. The communication can be by any suitable wireless or telecommunications network, such as a wide-area wireless network for a small array, or a WCDMA, WiMax, LTE or similar telecommunications protocol for a large array. This may be particularly suitable for off-shore arrays. The communication between relay stations 204 can be by a wired protocol as shown here, or wireless in situations in which this is not practical.

Referring briefly to Figure 5, this is a diagram illustrating steps of a method according to an embodiment of the invention. Initially, a wind turbine from an array may be selected for interaction, although in an embodiment, an interaction may be carried out using a single UAV with a single wind turbine. Once the wind turbine is selected, the flight of the UAV to rendezvous with the turbine for interaction is initiated (500). The drone can for example follow a pre-determined or self-determined flight path, by any known navigational system (such as a GNSS system, e.g. GPS), from a base station (or other position in the farm or array at which the drone is presently located) to the selected wind turbine, which may for example be identified by a waypoint. The drones comprise navigational devices which along with their powered lift systems (such as rotors) can be used to undertake the flight. The base station can provide real-time navigational guidance for each drone in the swarm based on flight data it receives from it. Alternatively, it is envisaged that the base station may provide each drone with mission data indicating which wind turbine is to be inspected, and what sort of inspection is needed. The mission data may be augmented with GPS coordinates of the wind turbine to be inspected, although that information may already be available to the drone in an on-board map of the wind park so all that it needs is the identification number of the wind turbine.

The drones are suitably equipped for autonomous flight and so will have a suitable on-board software platform for this purpose. Autonomous flight systems are generally known in the art and so will not be discussed in detail here.

Briefly, however, each drone will comprise suitable sensing systems (332) to provide it with flight data relating to its position, orientation, velocity, angular velocity, and acceleration, for example. Such data may be derived from a state estimator coupled to sensing systems such as GPS, LiDAR, optical imaging systems, inertial measurement units (IMUs) and the like. In such autonomous systems, and as is known, sensor data is fed into a state estimator such as an Extended Kalman Filter (EKF) from which the navigation system of the drone can derive localisation and mapping information, as discussed above. This is sometimes referred to as Simultaneous Localisation and Mapping (SLAM) in the art.

There is also a fully described 3D model of each wind turbine available to each drone on each mission, for example as would be stored in on-board memory; these images may have been captured on previous missions and stored for use during interactions. In alternatives embodiments, the 3D model can be generated independently of such image captures, for example from the 3D dimensions established during manufacturing or installation of the WTGs or the wind farm. Wind farms vary in the types of turbines installed and configuration of the turbines, for example by tower height, blade length, and the like. Some components can be site specific. Therefore one or either of these methods, or a combination, will be of use in assessing the layout of the farm and the

dimensions and configuration of each turbine. Even so, some features will be shared across turbines and farms, for example common components such as sensors, relays and site infrastructure which will be common to most farms or parks of a given type. These components can therefore be pre-programmed for a template 3D model.

Once the drone has arrived, the interaction can get underway. Typically this will involve the UAV approaching the selected wind turbine, or in embodiments a component thereof. The interaction may be to drop off a maintenance item, to communicate with a control unit of the turbine, or any other such interaction; in this example, the drone inspects the turbine component, for example using an image capture device. The image capture device may, in principle, be any suitable imaging system suitable for drone use in outdoor environments, and would be known to the skilled person.

During the interaction, an environmental condition is monitored (502). This may be done at the base station - for example, there may be a weather alert, or the light condition may be fading. In this embodiment, the condition is monitored using a sensor on the UAV, for example a motion sensor or accelerometer.

The environmental condition is monitored to determine whether it is changing, particularly relative to a threshold (504). In close proximity to a turbine, a drone is susceptible to small variations in wind and aerodynamic conditions around the turbine and the blades, which may endanger the drone, or possibly damage the turbine or blades.

On detecting the variation, a flight parameter for the UAV is modified (506). The variation of the condition can be tracked, so that if too severe, the drone can stop the interaction and vacate the area if necessary. The flight parameter may be a speed of the UAV along a flight path, a rotor speed or the like; for example, rotor speed may be temporarily increased to create lift, so that the drone flies away from the turbine. A distance measurement device, such as a radar or

ultrasounds device, can be included on the UAV in order to monitor the distance between the UAV and the component inspected.

Adaptive learning can be provided for the UAS, either on-board or at base. The systems can record the environmental conditions during missions, and also record instances where thresholds were exceeded. The response of the drone to the threshold breach can also be recorded. This provides additional information to the drones, making each mission increasingly efficient; the drones will be ‘aware’ that in certain conditions for example, the approach distance will have to be increased if the threshold is not to be repeatedly exceeded, and the drone instructed to back away constantly.

Referring in addition to Figure 6, a general method for inspection of a wind park or farm may proceed as follows. Initially, a central drone control computer or system determines (600) inspection missions for each of a plurality of drones. This may be done by a user selecting missions from a set of missions, or creating missions from a set of inspection tasks (and associated flight paths for the drones), or in certain cases may be determined automatically. For example, if an fleet inspection round or event is to proceed as per a previous fleet inspection round, the system can be used to automatically select appropriate or available drones for the tasks set out, which may be a standard set of tasks, and compile these into missions for the drones.

Mission data corresponding to the inspections missions for the respective drones is then provided (602) to the drones. For example, the data may comprise flight paths, instructions for inspection, typical and inclement conditions, and the like. The drones are then dispatched (604) on the missions. The drones may be dispatched simultaneously, or in a staggered release depending on their respective missions; some may have been assigned shorter missions than others (for example if some are more complex and/or are less predicable in timescale for achievement). Autonomous flight routines will typically be implemented (606)

5 for the drones to rendezvous with the respective turbines, as described herein.

A specific embodiment of the invention will now be described. An operator of the wind farm observes the performance of the wind turbines and can supervise when scheduled and un-scheduled service activities can be executed with least 0 loss of produced energy and to lowest possible costs.

On the monitoring system, there is an application in which the observer can operate the automatic drone inspection system. The drone control application is executed in a so-called drone control unit. The operator can choose specific 5 tasks for the drone inspection as below:

The types of tasks are predefined, for example a full body inspection is a photo session of all blades, nacelle roof, yaw from outside and tower welding, taken in 0 a given sequence and perhaps with pre-defined text descriptions added to the photos.

The items subject for potential inspection are fully described, so the operator can select very specific positions on a given wind turbine to be inspected. The operator defines whether the result of the inspection should be a photo session or video, and where the data should be presented. He can also choose to control the drone manually by orders. In a specific interface, he can order the drone to stay at a given distance to the tower in a given height and compass direction. Then he can order it to lower and elevate a given distance. In such a session, the operator can order the drone to stream live video directly to his monitor. He can take pictures for storage.

Before a given mission, the operator brings the turbine in question to an operational stage, which is safe for the drone depending on how close the drone needs to approach the turbine and components.

The drone is controlled remotely by orders from the operator’s application.

The command to the drone is given with a drone ID, so more than one drone can operate.

In the wind farm there are positioned the necessary number of relay stations that can ensure a continuously communication with the drones. The relay stations are typically positioned at the WTG-pads so the wired communication between the relay stations can follow the communication installations as used for a Supervisory Control and Data Acquisition (SCADA) control system on board the WTG, potentially using the same fibre net. If needed there is relay station in between WTG-pads. The protocol in the relay stations ensures that when a drone flies from one relay station to another, the communication is continued.

The drones can transmit pictures and video back to the drone control unit and the operator. Continuously, the drone transmits its 3D position when it operates in the wind-farm. A specific application shows the immediate position of all the active drones in the wind farm to the operator. This is both for optimizing the operations, for safety and for finding drones that have made emergency landing. The bandwidth in the network is dimensioned to the amount of data to and from the fleet of drones.

Depending on the definition of the task, data collected and handled by the system are stored in specific predefined folders or displayed to the operator on her request. The operator can analyse the data and decide what action may be needed for the further physical maintenance for the particular turbine in question. Back-office functions and the turbine owner can have access to the database for analysis as well. The site management can implement routines for housekeeping the relatively high amount of data. Pictures and movies without relevance can be deleted accordingly.

The drone is equipped with a sensor system and camera that can deliver a number of data for the operator, but also for its own safety and navigations. As shown in Figure 2, these can include systems that measure height and distance to components. These include devices for reading QR codes and other marks, and for pattern recognition, for GPS positioning, monitoring XYZ translational movements, time stamping, light sensors (for daylight intensity), image capture devices, telecommunications capabilities, and systems for determining remaining flight capability, such as battery monitoring.

In this embodiment for navigation, a conventional global GPS-system is used. Each drone has its own individual positioning system. The drone is continuously communicating its 3D position to the central drone control unit that makes the data available for the operator and data storage. The control unit also ensures by orders to the particular drone that it will not navigate outside of the reach of the relay stations. The drone has an on-board application that also ensures this. If a drone loses communication with the relay-station, it performs a landing nearest to the last communicated position or at pre-defined spots, so the operators can find it again. Specifically in offshore wind farms it is defined that the drones can only land at the heli-hoist platforms on the roof of the wind turbines. When the communication from the relay station is re-established, the drone position is again present as data in the drone control unit and thus for the operator.

Mobile relay stations can be installed in, for example, service cars for usage of drones in areas without coverage by fixed relay stations. The mobile relay station communicates via a safe medium in the particular wind farm, such as a mobile telephony system.

For weather monitoring, the central drone control unit receives data from the SCADA systems in the wind farm about the weather condition, both the instant/actual weather and the forecast. For the specific drones and specific types of missions, it is predefined which weather conditions are acceptable for missions. Parameters like wind speed, wind gust, precipitation are relevant.

Some missions can accept harder conditions compared to others; it depends on how close the drone must approach the wind turbine.

These data are available for the operator when she is planning the missions. When she requests a specific mission in a specific timeslot, the control unit checks up if the mission can be performed in the expected weather conditions or not, and the control unit can suggest when the next possible weather window will be available.

During a specific mission, the drone communicates weather observations with the central control unit. From the unit it will receive actual data and short-term (10 min and 30 min) forecasts, from which it can reconsider if it can complete the actual mission or finish it. It will also continuously observe the current conditions in the air during the operation. If pre-defined limits are exceeded, it can interrupt the mission, land and wait for new orders from the operator. On light levels, there are limits for how little light the drone can operate within including reading ID-labels and identify turbine components. The drone continuously measures the light intensity and can interrupt a mission, if it gets too dark to complete the task. It can also report this back to the operator, so he can decide actions. In the planning tool for flight operations, every requested mission is evaluated to determine if it can be done within required daylight based on general data of daylight on the specific site and time of the year and day.

It should be noted here that some operations possible outside daylight hours, for example by means of illumination devices such as headlights incorporated into the drones, or by illumination devices at the turbine sites which may be activated remotely, or automatically on approach of a drone (either by detecting the drone by means such as RF communication, or by the drone transmitting an approach signal to a receiver on the site. This can be of use for example in emergency situations where immediate investigation outside daylight hours may be required.

Regarding drone navigation at the turbine(s), the controller on board the drone is equipped with a system that can read identification labels at a certain distance e.g. QR-codes or bar-codes. Labels with ID-code for a turbine in a wind farm are placed in specific positions at each turbine. It can be on the roof, above the door and other well-defined positions that are easy for the drone to identify. The label can for instance have a red circle around the ID-code or similar so the drone has a specific pattern to recognize.

In general the shape of a wind turbine seen from different angles and heights is placed in the drone controller, so when it approaches the wind turbine it can recognize it and orientate itself accordingly.

Together with the recognition system, the drone is also equipped with a system for measuring distance to nearby components. It can be a radar, ultrasonic, or laser-based system, for example a Light Direction and Ranging (LiDAR) system. The measurement of distance is spherical around the drone. It is used for landing purposes and to keep specified safety distances to the turbine components.

The drone can typically only operate close to the wind turbine, when the turbine rotor is stopped and parked (brake applied). The drone then by sampling various pictures of the turbine observes if the wind turbine rotor is rotating. The drone communicates this information back to the operator and steps into a waiting position (landing) until the turbine is stopped.

It can be considered if the drone automatically via the central drone controller and the SCADA system in the wind farm can be authorized to stop and start the wind turbines before and after missions.

For identifying components and positions, components like the blades, cooler top, wind sensors, hub-cone, tower sections are subject for inspection by the drones. The shapes of these components from different angles are placed in the memory of the drone controller for pattern recognition and following orientation. Based on measurement of actual distance to the component, the drone controller can identify exactly which position of the component it is observing.

Visually readable identification means, such as labels with QR-codes

representing ID (item number, serial number, etc.) are placed on specific spots on these components which the drone can read. In embodiments, the

identification attribute may use a radio frequency identification means (such as an RFID or NFC device), which can be embedded in the component of the wind turbine rather than being applied to the surface.

For instance it may be required that the drone can identify which blade (A, B or C) it is operating at. Via communication with the central drone controller the drone can have such information to link to or insert into the pictures it will produce from the mission. The central drone controller has access to a database with all basic data of the components in the wind farm.

When the drone operates close to the turbine components, it locks its navigation to the component by means of pattern recognition and it continuously measures the distance to the component for safety reasons. The drone keeps a pre-defined safety distance to the component.

To make orientation easier, the components are equipped with labels indicating positions, for instance towers can have labels with information of height and compass direction for every 5 metres; blades can have information on distance from the hub and respective side of the blade. The drone can search for these indicators for locating a given position on the specific component.

Around a wind turbine, there are typically aerodynamic phenomena that can influence the flying capabilities of the drone. Sudden wind gusts can push the drone into the turbine. The wind can be very different around tower sections, blades and other components. The drone observes the aerodynamic conditions through its movements when it approaches a component and corrects its operation accordingly. It includes updating the safety distance limits to the components.

If the drone experiences aerodynamic influence above pre-defined limits its pre defined escape-flying algorithm takes over the control and brings the drone out of the danger, for example by elevating or backing off. The drone controller has a learning program that can analyse and include these behaviours into the flight operation algorithms. The incidents and corresponding actions are also stored into the central drone control unit, so they can be shared with the other drones on site and even other sites. They may be different from turbine to turbine and from site to site; consequently, they should be stored together with the specific turbine data. On finding specific positions of components, the purpose for inspecting the wind turbine components is typically to find failures or irregularities. The drone operating from the root end to the blade tip, capturing images, may inspect a blade for impact from lightning strokes. These pictures are compared to normal looking images and to images from impacted blade surfaces. Based on that, the control system decides whether a certain issue is reported to the operator and stored. Information of blade number, date, exact position on the blade etc. are stored together with the pictures.

The algorithm for this part is typically placed in the central drone controller. Similar algorithms for other standard inspections are pre-programmed to the system.

The operator can also define a specific position, that she wants the drone to inspect, given by specific coordinates on the turbine. The drone is then sent for the specific mission based on the data. An example could be taking a picture of Blade B trailing edge, 35.5 meters from the root. The drone will navigate to this position based on the above described system of IDs.

The operator can also remotely operate the drone, when it is at a turbine site.

It can be operated by giving commands of positions or simply controlled by up, down, turn left-right commands and the like. The drone is in any case streaming live pictures to the operator and she can decide to store these and make snap- shots. Furthermore, the live or real-time images can also be provided to other users, such as design engineers; these may be able to make real-time maintenance or analysis decisions for feeding back to the drones.

The drones transmit pictures from operations to the central drone controller. Besides the picture itself, transmission can contain data such as:

• Time stamp • Turbine ID (turbine number)

• Component ID (serial number)

• Position on the component (i.e. on blades, distance from root-end and which side)

• GPS-position

In addition, the drones may incorporate an acoustic sensor device such as a microphone in order to be able to detect sounds from the inspection. For example, it may be that certain sounds are known to be indicative of particular faults, malfunctions or inefficient operation modes. The drone may therefore capture audio data which can also be transmitted back to the central control system.

The drone itself when it approaches and works around the turbine creates these data. The images are stored in the controller in pre-defined folders for the specific turbine in question.

At the central drone control station the drones can dock in and be charged with power for the next missions. These installations can be present at all or some of the relay stations as well. A controller system in each drone supervises the remaining power capacity in the batteries on-board the drone and ensures it returns to a docking station for recharging before the power capacity runs out. The capacity for a given mission is continuously reported to the central drone control unit and displayed for the operator, so he is informed what operational capability the fleet of drones represent at a given time.

Drones require regular maintenance to keep them available for missions.

The intervals for maintenance are defined by the manufacturers relatively to the operational conditions that the drones meet. The control system or software in every drone continuously calculates the remaining flight capacity before next scheduled maintenance and communicates these data to the central drone unit. The operator can, based on these data, schedule and book maintenance for the individual drones in the fleet. Accordingly he can include the drone service into the overall planning for the missions, i.e. plan the maintenance in timeslots where missions to the turbines are not possible like strong wind, darkness, heavy precipitation.

In embodiments of the invention, drones employed in the manner described herein can also be used to support wind farm management in a safety

contingency or emergency scheme for a wind farm. For example, if a serious incident should happen to a wind turbine or a technician present on the farm, one or more drones such as those described herein can be deployed to quickly obtain information which would help an emergency response. For instance, drones may be immediately deployed on receipt of an alert, to the site of the alert, and can capture sensor information as described above at or near the identified turbine or turbine component. Such information may provide real-time imaging of the incident, or size of an area affected, or the like.

In other embodiments, the same capabilities of the drones and the control system can be used for example for security for a wind park. For instance, drones may be deployed in similar manner as described, but in addition to or instead of performing inspection for maintenance, may inspect the turbines or the park for security, for example to check that no intruding persons or vehicles are present.

Figure 4 is a diagram illustrating the components, structure and functionality of an on-board UAV computer processing and management system (400) for a UAV (202) according to an embodiment of the invention.

The computer system 400 comprises a processing environment 420 with processor 421 and memory 422, with associated communications functionality 423. The communications functionality typically includes a networking capability allowing communication with a network, or directly with another computer system or computer device. For example, in the above embodiments, this may be communication from the drone to the relay station(s), to the base station or central drone controller, or even to other drones.

The memory 422 may store readable instructions to instruct the processor to perform the functions of the on-board UAV computer system. For example, instructions to provide the functions of monitoring the environmental conditions may be stored. The processor 421 is a representation of processing capability and may in practice be provided by several processors.

A database 410 is provided, storing data as applicable. For the computer system (400), this database can provide the storage for any pre-programmed flight paths, locations of WTGs, 3D models of WTGs and the like.

Elements shown within the processing environment 420 use the processor 421 and the memory 422 to deliver functionality; for example, these elements can provide steps of embodiments of the invention such as on variation of the at least one environmental condition relative to a threshold, modifying a flight parameter for the UAV.

An optional management system (430) module can be located within the processing environment 420, to provide the management functions for the on board computer system. The management system may also comprise functions of other parts of the system, such as the processor 421 , the memory 422 and the database 410 itself.

In embodiments, similar features may be provided in a computer UAS

management system (400), for example such a system hosted at the central drone control site, according to an embodiment of the invention, which can provide the features of managing the UAVs, relay stations, wind turbine sites and the like described above. The computing devices noted above with reference to Figure 4 may include one or more of logic arrays, memories, analogue circuits, digital circuits, software, firmware and processors. The hardware and firmware components of the computing devices may include various specialized units, circuits, software and interfaces for providing the functionality and features described herein. The processor(s) may be or include one or more microprocessors, application specific integrated circuits (ASICs), programmable logic devices (PLDs) and

programmable logic arrays (PLAs).

Processors and/or controllers may comprise one or more computational processors, and/or control elements having one or more electronic processors. Uses of the term“processor” or“controller” herein should therefore be considered to refer either to a single processor, controller or control element, or to pluralities of the same; which pluralities may operate in concert to provide the functions described. Furthermore, individual and/or separate functions of the processor(s) or controller(s) may be hosted by or undertaken in different control units, processors or controllers.

To configure a processor or controller, a suitable set of instructions may be provided which, when executed, cause the control unit, computer system, computer device or the like to implement the techniques described herein. The set of instructions may suitably be embedded in the one or more electronic processors. Alternatively, the set of instructions may be provided as software to be executed on the computational device.

It will be appreciated by those skilled in the art that the invention has been described by way of example only, and that a variety of alternative approaches may be adopted without departing from the scope of the invention, as defined by the appended claims.

Claims

1. A system (300) for managing the aerial inspection of a wind park, the system comprising:

a plurality of drones (312, 314); and

a drone control computer (322),

wherein the drone control computer is configured to determine (600) inspection missions for each of the plurality of drones, to provide (602) mission data corresponding to those inspection missions to respective drones and to dispatch (604) said drones on said inspection missions.

2. A system according to Claim 1 , wherein at least one of the plurality of drones is configured to implement (606) autonomous flight routines so as to rendezvous with a respective wind turbine (206) without operator (320) guidance.

3. A system according to Claim 2, wherein mission data for a respective inspection mission for a respective drone comprises an autonomous flight routine, and wherein the autonomous flight routine comprises an identification attribute (102) of a component of a respective wind turbine.

4. A system according to Claim 3, wherein the identification attribute comprises a visually readable identification means and/ or a radio frequency identification means.

5. A system according to Claim 4, wherein the visually readable identification means comprises a bar code or QR code.

6. A system according to any of the Claims 3 to 5, wherein upon a rendezvous with a wind turbine, the respective drones are configured to use an on-board sensor (332) to capture data associated with the identification attribute of the component of the wind turbine, and to process (421 ) the captured identification attribute to identify the wind turbine component.

7. A system according Claim 6, wherein the on-board sensor comprises an image capture device, configured to capture image data of the identification attribute.

8. A system according to Claim 6 or Claim 7, wherein the drone is configured to process the captured identification attribute data using a pattern recognition algorithm to identify the component.

9. A system according to any of the Claims 6 to 8, wherein the on board sensor comprises a radio frequency device, configured to interact with a radio frequency identification means of the identification attribute.

10. A system according to any of the Claims 6 to 9, wherein the autonomous flight routine comprises a plurality of identification attributes for respective components of the wind turbine, and wherein the on-board sensor is configured to capture data associated with a selected one or more of the component identification attributes.

1 1. A system according to any of the above claims, wherein the drone control computer is configured to manage a set of inspection tasks for the wind park, the set comprising respective inspection tasks for respective wind turbines of the park,

and wherein the drone control computer is configured to use the set of inspection tasks to determine the inspection missions for each of the plurality of drones, and to provide mission data corresponding to those inspection missions to the respective drones.

12. A system according to Claim 1 1 , wherein the drone control computer is configured to update the set of inspection tasks with data captured by one or more of the plurality of drones.

13. A system according to Claim 12, wherein the drone control computer is configured to modify an inspection mission for a given drone in response to an update of the set of inspection tasks.

14. A system according to any preceding claim, wherein at least one of the plurality of drones is configured to monitor environmental conditions when in proximity to a respective wind turbine for an interaction therewith, and wherein the drone is further configured to modify one or more flight parameters based on changes in the monitored environmental conditions.

15. A system according to Claim 14, wherein the drone is configured to compare an instant value for the monitored environmental condition with a threshold for the environmental condition; and activate modification if the value exceeds the threshold.

16. A system according to Claim 14 or Claim 15, wherein at least one of the plurality of drones is configured to record environmental condition data associated with the monitored environmental condition.

17. A system according to any of the above claims, wherein upon a rendezvous with a wind turbine, the respective drones are configured to use an on-board sensor to capture audio data associated with the component of the wind turbine.

18. A system according to any of the above claims, wherein the drone control computer is configured to use respective unique identifiers for each of the plurality of drones, to dispatch the respective drones on the respective inspection missions.

19. A system according to any of the above claims, wherein the drone control computer is configured to, following dispatch of the plurality of drones, display to an operator a position for each of the plurality of drones.

20. A system according to any of the above claims, comprising a plurality of relay stations (204), said relay stations configured to:

communicate with one or more of the plurality of drones; and provide handoff between relay stations during a communication with a given drone.

21. A wind turbine (1 ) for a wind park for aerial inspection managed by a system according to any of the above claims.

22. A wind turbine according to Claim 21 in respect of a system according to any of the Claims 3 to 20, comprising at least one said component comprising an identification attribute (102).

23. An unmanned aerial vehicle (UAV) (312, 314) for interacting with a wind turbine system (300), the vehicle comprising:

a powered flight system for generating lift;

at least one sensor (332);

a memory (422); and

a processor (421 ),

wherein the vehicle is configured to:

by the flight system, carry out an autonomous flight routine determined by a drone control computer (322); rendezvous with a wind turbine(206) for an interaction;

during the interaction, by the at least one sensor, capture data associated with an identification attribute (102) of a component of the wind turbine; and

by the processor, process the captured identification attribute to identify the wind turbine component.

24. A UAV according to Claim 23, wherein the at least one sensor comprises an image capture device, configured to capture image data associated with the identification attribute.

25. A UAV according to Claim 23 or Claim 24, wherein the at least one sensor comprises a radio frequency device, configured to interact with a radio frequency identification means associated with the identification attribute.

26. A drone control computer (322) for a system (300) according to any of the above Claims 1 to 20.

27. A drone control computer according to claim 26, wherein a drone automatically via the central drone controller and a SCADA system in the wind farm can stop and start the wind turbines before and after missions.

28. A drone control computer according to claim 26 or 27, wherein a drone can be dispatched automatically on an inspection mission of a wind turbine by the drone control computer when said wind turbine is being shut down.

29. A computer program application comprising computer program code adapted, when loaded into or run on a computer or processor, to cause the computer or processor to become a drone control computer according to Claim 26.