WO2015079221A2

WO2015079221A2 - Determination of turbulence in a fluid and control of array of energy producing devices

Info

Publication number: WO2015079221A2
Application number: PCT/GB2014/053493
Authority: WO
Inventors: Thomas Clark
Original assignee: Ocean Array Systems Ltd
Priority date: 2013-11-26
Filing date: 2014-11-25
Publication date: 2015-06-04
Also published as: WO2015079221A3

Abstract

In one aspect, the invention provides a unit (1) configured to approximate at least one quantity representative of turbulence in a marine flow by: determining at least one profile of the marine flow; selecting at least one eddy type representative of the turbulence, wherein each eddy type comprises at least one eddy size parameter and an eddy strength parameter; determining a velocity signature and/or a turbulent intensity signature, for each selected eddy type; and using at least one said determined signature to perform a deconvolution of the profile, in order to determine, for each selected eddy type, at least one probability density function as a function of the at least one size parameter and/or the strength parameter.

Description

DETERMINATION OF TURBULENCE IN A FLUID AND CONTROL OF

ARRAY OF ENERGY PRODUCING DEVICES

In a first aspect, the invention relates to systems and methods to determine turbulence in a fluid and its effect on energy generation devices. In a second aspect, the invention relates to systems and methods to control an array of devices adapted to use kinetic energy of a stream of fluid to produce electrical energy.

With respect to the first aspect, it is known that tidal and atmospheric flows comprise boundary layers, i.e. a region of fluid close to a solid surface (or similar wall). Boundary layers in geographic flows of such large scale are almost invariably turbulent. Additional turbulence is generated by features and roughness in the bathymetry or landscape.

Unlike the laboratory, where tools such as Particle Image Velocimetry (PIV), Particle Tracking (PTV) and Dye Visualisation (DV) are readily used to investigate the spatial characteristics of turbulence (e.g. in water or wind tunnels), measurement devices for the marine environment are more limited in capability. Typically, measurement devices are point instruments (e.g. marine Laser/Acoustic Doppler Anemometers (LDAs) and Microstructure Profilers), or have a measurement domain limited to a line (e.g. Acoustic Doppler Current Profilers, ADCP). Many methods are affected by water quality, thus requiring seeding or particulates in the flow to work effectively.

Some attempts have been made to utilise PIV for planar measures in marine and atmospheric flows, but the extent of the plane and the optical constraints significantly restrict the spatial extents over which the technique is available. The quality of the seeding, and therefore the accuracy and robustness of the technique, cannot be easily controlled.

As a result, most or all measurements of turbulence in marine and atmospheric flows have a highly limited spatial extent, i.e. a point or a line, and limited spatial resolution. Many measurements have no directionality or limited dimensionality, i.e. measure a speed or a dissipation rate, rather than a velocity vector, or measure fewer than three components of a velocity vector.

Further problems are caused by the long timescales required to converge on turbulent statistics, especially for large scale turbulence. This is exacerbated by noise, and the time to converge on turbulent quantities greatly exceeds the time to converge on a mean flow distribution. In the case of tidal flows, this is can become terminal, since the mean flow is constantly changing; rapidly enough that the time period required to converge on turbulent statistical quantities for large eddy structures exceeds any time over which the mean flow could be considered constant, and this by orders of magnitude. This results in poor convergence of turbulent spectra at the large scales.

Acoustic Doppler Current Profilers (ADCP) have some ability to measure turbulence. However, their application may be impaired by insufficient spatial resolution, high noise levels, incomplete data and a constantly changing tidal flow (i.e. the measurement time period is implicitly too short to achieve statistical convergence of turbulence metrics, especially high order indices). Noise can be accounted for to some degree (using assumptions of isotropy, as described in the article "Method for Identification of Doppler Noise Levels in Turbulent Flow Measurements Dedicated to Tidal Energy", by Richard J.B. Thomson J., Polagye B. and Bard J. , European Wave and Tidal Energy Conference 2013), but a problem called 'haystacking' (i.e. poor convergence of the spectrum for large and/or slow motions) remains: in reference to ADCP data, bulk tidal motion (considered colloquially to be the 'mean') is incorporated into or otherwise distorts turbulence data.

However, the issue of turbulence, along with wave propagation, meteorological effects and many other (typically unsteady) fluid flow and environmental effects impact the generation of power from wind turbines, tidal stream turbines and wave energy converters.

These issues must all be considered at the assessment, design and analysis stages of building power plants. The presence of unsteadiness in the flow affects mean and instantaneous energy yield, as well as through-life dynamic loading on structural, mechanical and electrical components.

Due to the large range of scales involved (from geographic to sub-mm scales of turbulence), the computational effort required to simulate turbulence in geographic flows is currently prohibitive, as is the effort which would be required to survey (for bathymetry/landscape) and to establish inflow conditions for a computational domain.

Attempts to computationally simulate turbines or other energy generation devices within geographic flows complicates matters further - any attempts to model effects of turbulence must be modified to take into account the energy sink represented by the device as well as accounting for the local change in turbulent characteristics caused by the device. Moreover, to ascertain accurate performance of the device, fine scales of turbulence must be incorporated into the simulation in order that separation characteristics and skin friction quantities are accurately computed. To understand dynamics of the device, the computation must be time resolved and the nature of turbulence at the inlet prohibits the use of symmetry in reduction of computational effort.

In the wind engineering community, attempts to tackle turbulence have relied predominantly on point measurements (at a proposed hub height) or line measurements using masts on which to mount instruments. A series of masts is often used to understand spatial variation of turbulence and wind speed over proposed sites. Such measurements can be used to generate turbulent spectra. Similarly, spectra in the marine environment can be produced from ADCP and other anemometer data. Due to the limitations of instrumentation discussed above, produced spectra may only be partial, i.e. not resolving parts of the spectrum at both upper and lower ends, or suffer from haystacking.

In analysis of turbines, 'low-order' computational techniques such as Blade Element Momentum (BEM) and Actuator Disk (AD) approaches have been used. In some cases, Lifting Line, Lifting Surface, Free Vortex, and Surface Panel methods have been used to provide improved integrity with respect to BEM and AD approaches. These methods provide considerably improved computational speed compared to 'fully viscous' Computational Fluid Dynamics (CFD) approaches, such as Reynolds-Averaged Navier Stokes (RANS) and Large Eddy Simulation (LES). However these approaches are less generalisable in terms of the geometries that can be analysed easily.

Most computational analyses use uniform inflows to simplify the problem. Mean boundary layer profiles are also often used at the inlet of a computational domain to represent a changing mean inflow with height. Fully viscous CFD modeling (RANS, LES and variants) often use simple parameterised models to reflect the effect of turbulence - a typical two-parameter model may assume a representative turbulent lengthscale and intensity. This reflects dissipation of energy due to turbulence, but does not compute unsteady dynamics. With regard to the second aspect, Onshore or offshore power plants comprising farms of devices, comprising e.g., wind turbines, tidal stream turbines and/or wave energy converters, produce energy from a fluid in movement, such as wind or sea currents.

It is known that the issues of turbulence, wave propagation, meteorological effects and many other (typically unsteady) fluid flow and environmental effects impact the generation of power from the devices, and these issues are usually all considered at the design and analysis stages of building power plants.

Attempts have also been made to take into account changes in the fluid flow during operation of the farms.

Figure 8 shows schematically an example of a typical known farm 71 comprising energy producing devices 73, such as wind turbines, tidal stream turbines and wave energy converters. The farm 71 is usually equipped with a Supervisory, Command And Data Acquisition (SCADA) system 72. Such SCADA systems are commonplace for use in control of multiple devices. The SCADA system 72 typically has data input for

local area sensor data (such as coming from a flow determination sensor, usually located upstream at least one of the devices 73); and for

non-real-time processed data (such as coming from a sensor or for run-time data) from respective individual control modules 730 of the individual devices 73 of the farm, as shown in Figure 8.

The farm 71 is thus supervised by the SCADA system 72 in that the SCADA system 72 can send non-realtime operational commands (such as an on/off command or an operation set point) to the individual control modules 730 of the devices 73.

However, as explained above, in conventional farms, power production is not optimal, as the farms only take into account simple measurements about e.g. , the changes in the fluid flow, and are not configured to take into account any physical understanding of the fluid dynamic (unsteady) effects involved, such as environmental and turbine-turbine interference effects.

Aspects of the invention address or at least ameliorate at least one of the above issues. Wth regard to the determination of turbulence in a fluid, in one aspect, the invention provides a method for approximating at least one quantity representative of turbulence in a marine flow, comprising a unit:

determining at least one profile of the marine flow;

selecting at least one eddy type representative of the turbulence, wherein each eddy type comprises at least one eddy size parameter and an eddy strength parameter;

determining a velocity signature and/or a turbulent intensity signature, for each selected eddy type; and using at least one said determined signature to perform a deconvolution of the profile, in order to determine, for each selected eddy type, at least one probability density function as a function of the at least one size parameter and/or the strength parameter.

The profile may be a mean profile.

The method may further comprise the unit performing a convolution of the at least one probability density function with the at least one determined signature for each eddy type, in order to determine turbulent spectra and/or Reynolds stress distributions of the marine flow. The method may further comprise the unit outputting the at least one probability density function and/or the turbulent spectra and/or the Reynolds stress distributions of the marine flow to a control unit configured to control an operation of an array of at least one device adapted to convert kinetic energy of the turbulent marine into electrical energy. The method may further comprise the unit outputting the at least one probability density function and/or the turbulent spectra and/or the Reynolds stress distributions of the marine flow to a design unit configured to perform a design of at least one device adapted to be placed in the marine flow.

The selected eddy type may comprise at least one of the following: type A, and/or type B, and/or type C, and/or a single line element, and/or a ring.

The method may further comprise the unit receiving simulation data and/or measurement data in order to determine the mean profile, and determining the mean profile using at least one of the following: an analytical solution, an analytical fit to unconverged measurement data, and/or a converged average of raw measurement data.

Where a mean flow of the marine flow over a first time window associated with turbulent motions in the liquid is below a predetermined threshold, the method may further comprise the unit applying a filtered value or using data over a second time window narrower than the first time window, so as to generate a mean flow of the marine flow. The method may further comprise the unit performing the approximating one or more times. In another aspect, the invention provides a method for approximating at least one quantity representative of turbulence in a fluid flow interacting with at least one device adapted to convert kinetic energy of the turbulent fluid flow into electrical energy, comprising a unit

determining at least one profile of the fluid flow;

determining a velocity signature and/or a turbulent intensity signature, for each selected eddy type; using at least one said determined signature to perform a deconvolution of the profile, in order to determine, for each selected eddy type, at least one probability density function as a function of the at least one size parameter and/or the strength parameter; and

outputting the determined at least one probability density function to a simulation unit configured to simulate, in a simulation domain, a behaviour of the turbulent fluid flow interacting with at least one device adapted to convert kinetic energy of the turbulent fluid flow into electrical energy for design or operation purposes.

The profile may be a mean profile.

The method may further comprise the unit performing a convolution of the at least one probability density function with the at least one determined signature for each eddy type, in order to determine turbulent spectra and/or Reynolds stress distributions of the fluid flow; and outputting the determined turbulent spectra and/or Reynolds stress distributions of the fluid flow to the simulation unit.

The method may further comprise the unit generating a distribution field of the selected at least one eddy type, oriented in the streamwise direction of the fluid flow; determining a velocity u(x,y,z,t) induced by the generated distribution field, at at least one control point (x,y,z) at an instant t, by implementing a Biot-Savart law; and advecting the generated distribution field through the simulation domain in the flow direction over time.

The method may further comprise the unit outputting the determined probability density function and/or turbulent spectra and/or the Reynolds stress distributions of the fluid flow and/or the induced velocity to the simulation unit configured to use at least one of the following models: a Free Vortex Model, a Lifting Line and/or Surface Analysis, a Panel model, and/or a Blade Element Momentum Model. The method according may further comprise the unit outputting the turbulent spectra and/or the Reynolds stress distributions of the fluid flow to the simulation unit configured to use Synthetic Eddy Methods, for finite volume and particle based Computational Fluid Dynamics analyses, such as Large Eddy Simulation and Smoothed Particle Hydrodynamics.

The method may further comprising the unit outputting the determined probability density function and/or turbulent spectra and/or the Reynolds stress distributions of the fluid flow and/or the induced velocity to the simulation unit forming a control unit configured to control an operation of an array of at least one device adapted to convert kinetic energy of the turbulent fluid flow into electrical energy and/or the simulation unit forming a design unit configured to perform a design of at least one device adapted to be placed in the fluid flow.

The method may further comprise the unit distributing the at least one eddy type randomly in the streamwise direction and in the cross-stream direction; and/or according to a predetermined distribution function. The method may further comprise the unit maintaining the eddy type constant over time, or modifying the eddy type as a result of a self-influence and/or an influence of at least one device placed in the turbulent fluid flow and/or at least a wake.

The method may further comprise:

at least one module of the unit performing the deconvolution;

at least one module of the unit determining at least one quantity representative of the turbulence in the flow; and

a reconciliation module reconciling outputs of the modules.

The method may further comprise an optimisation module implementing Machine Learning and/or Artificial Intelligence to the reconciliation module. The reconciliation module may implement a Projection Onto Convex Sets, POCS, algorithm.

The unit may perform the approximating one or more times.

In another aspect, the invention provides a unit for approximating at least one quantity representative of turbulence in a marine flow, configured to:

determine at least one profile of the marine flow;

select at least one eddy type representative of the turbulence, wherein each eddy type comprises at least one eddy size parameter and an eddy strength parameter; determine a velocity signature and/or a turbulent intensity signature, for each selected eddy type; and use at least one said determined signature to perform a deconvolution of the profile, in order to determine, for each selected eddy type, at least one probability density function as a function of the at least one size parameter and/or the strength parameter.

The profile may be a mean profile.

The unit may be configured to perform a convolution of the at least one probability density function with the at least one determined signature for each eddy type, in order to determine turbulent spectra and/or Reynolds stress distributions of the marine flow. The unit may be configured to output the at least one probability density function and/or the turbulent spectra and/or the Reynolds stress distributions of the marine flow to a control unit configured to control an operation of an array of at least one device adapted to convert kinetic energy of the turbulent marine into electrical energy. The unit may be configured to output the at least one probability density function and/or the turbulent spectra and/or the Reynolds stress distributions of the marine flow to a design unit configured to perform a design of at least one device adapted to be placed in the marine flow. The selected eddy type comprises at least one of the following: type A, and/or type B, and/or type C, and/or a single line element, and/or a ring.

The unit may be configured to:

receive simulation data and/or measurement data in order to determine the mean profile, and determine the mean profile using at least one of the following: an analytical solution, an analytical fit to unconverged measurement data, and/or a converged average of raw measurement data.

Where a mean flow of the marine flow over a first time window associated with turbulent motions in the liquid is below a predetermined threshold, the unit may be configured to apply a filtered value or use data over a second time window narrower than the first time window, so as to generate a mean flow of the marine flow. The unit may be configured to perform the approximating one or more times. In another aspect, the invention provides a unit for approximating at least one quantity representative of turbulence in a fluid flow interacting with at least one device adapted to convert kinetic energy of the turbulent fluid flow into electrical energy, configured to:

determine at least one profile of the fluid flow;

select at least one eddy type representative of the turbulence, wherein each eddy type comprises at least one eddy size parameter and an eddy strength parameter;

determine a velocity signature and/or a turbulent intensity signature, for each selected eddy type; use at least one said determined signature to perform a deconvolution of the profile, in order to determine, for each selected eddy type, at least one probability density function as a function of the at least one size parameter and/or the strength parameter; and

output the determined at least one probability density function to a simulation unit configured to simulate, in a simulation domain, a behaviour of the turbulent fluid flow interacting with at least one device adapted to convert kinetic energy of the turbulent fluid flow into electrical energy for design or operation purposes.

The profile may be a mean profile. The unit may be configured to: perform a convolution of the at least one probability density function with the at least one determined signature for each eddy type, in order to determine turbulent spectra and/or Reynolds stress distributions of the fluid flow; and output the determined turbulent spectra and/or Reynolds stress distributions of the fluid flow to the simulation unit. The unit may be configured to: generate a distribution field of the selected at least one eddy type, oriented in the streamwise direction of the fluid flow; determine a velocity u(x,y,z,t) induced by the generated distribution field, at at least one control point (x,y,z) at an instant t, by implementing a Biot-Savart law; and advect the generated distribution field through the simulation domain in the flow direction over time. The unit may be configured to output the determined probability density function and/or turbulent spectra and/or the Reynolds stress distributions of the fluid flow and/or the induced velocity to the simulation unit configured to use at least one of the following models: a Free Vortex Model, a Lifting Line and/or Surface Analysis, a Panel model, and/or a Blade Element Momentum Model. The unit may be configured to output the turbulent spectra and/or the Reynolds stress distributions of the fluid flow to the simulation unit configured to use Synthetic Eddy Methods, for finite volume and particle based Computational Fluid Dynamics analyses, such as Large Eddy Simulation and Smoothed Particle Hydrodynamics.

The unit may be configured to output the determined probability density function and/or turbulent spectra and/or the Reynolds stress distributions of the fluid flow and/or the induced velocity to the simulation unit forming a control unit configured to control an operation of an array of at least one device adapted to convert kinetic energy of the turbulent fluid flow into electrical energy and/or the simulation unit forming a design unit configured to perform a design of at least one device adapted to be placed in the fluid flow.

The unit may be configured to distribute the at least one eddy type randomly in the streamwise direction and in the cross-stream direction; and/or according to a predetermined distribution function.

The unit may be configured to maintain the eddy type constant over time, or modify the eddy type as a result of a self-influence and/or an influence of at least one device placed in the turbulent fluid flow and/or at least a wake.

The unit may further comprise:

at least one module configured to perform the deconvolution;

at least one module configured to determine at least one quantity representative of the turbulence in the flow; and

a reconciliation module configured to reconcile outputs of the modules.

The unit may further comprise an optimisation module configured to implement Machine Learning and/or Artificial Intelligence to the reconciliation module. The reconciliation module may be configured to implement a Projection Onto Convex Sets, POCS, algorithm.

The unit may be configured to perform the approximating one or more times.

The unit may be implemented at least partially as software or firmware and/or at least partially in a physical casing.

With regard to the control of array of energy producing devices, in one aspect, the invention provides a control unit configured to control an operation of an array comprising at least two devices adapted to use kinetic energy of a stream of fluid to produce electrical energy, the unit being configured to:

determine a state of the array using real-time data relating to an operation of at least one of the devices of the array; modify the operation of at least one device of the array, as a function of the determination.

The real-time data may refer to at least one of the following parameters: torque, thrust, rotations per minute, shaft strain and/or shaft stress. The unit may further be adapted to use external data, such as environmental data and/or the economic, operational and logistic data. The unit may further be configured to: output a real- time control command for the at least one device, as a function of real-time data relating to an operation of at least two devices in the array.

In another aspect, the invention provides a control unit configured to control an operation of an array of at least one device adapted to convert kinetic energy of a fluid into electrical energy, comprising:

at least two algorithm modules, each module being configured to generate at least one real-time control response for at least one of the devices, by implementing, a respective control algorithm; and

a reconciliation module configured to reconcile the real-time control responses in order to output the real-time control command.

The unit according may be configured to:

generate, as a function of real-time data relating to the operation of all the devices in operation in the array, at least one real-time control response for all the devices in operation in the array; and

output a real-time control command for all the devices in operation in the array.

The algorithm module may be configured to implement at least one of the following control algorithms: Array- As-A-Sensor, Array-As-A-Device, Operational response, and/or Real-time feature recognition response, Wave propagation, Thrust/mass flux distribution response, Unsteady inflow vs. yield response, Unsteady flow vs. peak load response, Unsteady flow vs. dynamic range response, Fluid feature and debris advection, Awareness model. The awareness model may include at least one of the following: Set point response, Peak allowable load awareness model, Lifetime awareness model, Economic awareness model, Operational and/or logistical awareness model, and/or Environmental data response.

at least algorithm module configured to implement at least one of the following control algorithms: Array-As-A-Sensor,

Array-As-A-Device,

Operational response, and/or

Real-time feature recognition response.

The algorithm module may be configured to generate at least one real-time control response for at least one of the devices, by implementing the at least one control algorithm. The algorithm module may further be configured to implement at least one of the following control algorithms: Wave propagation, Thrust/mass flux distribution response, Unsteady inflow vs. yield response, Unsteady flow vs. peak load response, Unsteady flow vs. dynamic range response, Fluid feature and debris advection, Awareness model. The awareness model may include at least one of the following: Set point response, Peak allowable load awareness model, Lifetime awareness model, Economic awareness model, Operational and/or logistical awareness model, and/or Environmental data response. The unit may further comprise at least one optimisation module configured to implement Machine Learning and/or Artificial Intelligence to the algorithm module. The unit may further comprise an optimisation module configured to implement Machine Learning and/or Artificial Intelligence to the reconciliation module.

The unit may be implemented at least partially as software or firmware, and/or implemented at least partially in a physical casing, separate from the devices.

In another aspect, the invention provides a system comprising:

an array of at least two devices adapted to convert kinetic energy of a fluid into electrical energy; and at least one control unit according to any one of aspects of the invention.

The system may comprise at least two redundant control units. The device may comprise a wind turbine, a tidal stream turbine or a wave energy converter.

In another aspect, the invention provides a method for controlling an operation of an array of at least two devices adapted to use kinetic energy of a stream of fluid to produce electrical energy, comprising a control unit:

determining a state of the array using real-time data relating to an operation of at least one of the devices of the array;

modifying the operation of at least one device of the array, as a function of the determination.

The real-time data may refer to at least one of the following parameters: torque, thrust, rotations per minute, shaft strain and/or shaft stress. The unit may use external data, such as environmental data and/or the economic, operational and logistic data. The unit may output a real-time control command for the at least one device, as a function of real-time data relating to an operation of at least two devices in the array.

In another aspect, the invention provides a method for controlling an operation of an array of devices adapted to convert kinetic energy of a fluid into electrical energy, comprising:

at least two algorithm modules generating at least one real-time control response for at least one of the devices, by implementing, a respective control algorithm; and

a reconciliation module reconciling the real-time control responses in order to output the real-time control command.

The unit may:

The algorithm may be at least one of the following control algorithms: Array-As-A-Sensor, Array-As-A-Device, Operational response, and/or Real-time feature recognition response, Wave propagation, Thrust/mass flux distribution response, Unsteady inflow vs. yield response, Unsteady flow vs. peak load response, Unsteady flow vs. dynamic range response, Fluid feature and debris advection, Awareness model. The awareness model may include at least one of the following: Set point response, Peak allowable load awareness model, Lifetime awareness model, Economic awareness model, Operational and/or logistical awareness model, and/or Environmental data response. In another aspect, the invention provides a method for controlling an operation of an array of devices adapted to convert kinetic energy of a fluid into electrical energy, comprising:

at least algorithm module implementing at least one of the following control algorithms:

Array-As-A-Sensor,

Array-As-A-Device,

Operational response, and/or

Real-time feature recognition response.

The algorithm module may generate at least one real-time control response for at least one of the devices, by implementing the at least one control algorithm. The algorithm module may implement at least one of the following control algorithms: Wave propagation, Thrust/mass flux distribution response, Unsteady inflow vs. yield response, Unsteady flow vs. peak load response, Unsteady flow vs. dynamic range response, Fluid feature and debris advection, Awareness model. The awareness model may include at least one of the following: Set point response, Peak allowable load awareness model, Lifetime awareness model, Economic awareness model, Operational and/or logistical awareness model, and/or Environmental data response. At least one optimisation module may implement Machine Learning and/or Artificial Intelligence to the algorithm module. An optimisation module may implement Machine Learning and/or Artificial Intelligence to the reconciliation module.

Aspects of the invention extend to computer program products such as computer readable storage media having instructions stored thereon which are operable to program a programmable processor to carry out a method as described in the aspects and possibilities set out above or recited in the claims and/or to program a suitably adapted computer to provide the system recited in any of the claims.

The invention has numerous advantages over the prior art.

The invention may be applied to both tidal and atmospheric turbulent flows comprising boundary layers of large scale (e.g. , geographic).

The invention does not need although may beneficially utilise input data from measurement tools such as Particle Image Velocimetry (PIV), Particle Tracking (PTV) and Dye Visualisation (DV), and may use measurement devices, such as point instruments (e.g. marine Laser/Acoustic Doppler Anemometers (LDAs) and Microstructure Profilers) or instruments having a measurement domain limited to a line (e.g. Acoustic Doppler Current Profilers, ADCP) and other anemometer data.

In combination with e.g. , ADCPs, the invention may have the advantage of deriving turbulent quantities from the profile, e.g. , the mean profile, mean flow profiles converging more rapidly than turbulence metrics. The invention may thus overcome the haystacking problem of the prior art, may improve overall convergence and may be far less vulnerable to noise in the measurement data.

The invention may enable simulation of turbulence in geographic flows, with an acceptable computational effort. The invention may provide convergence of turbulent spectra at the large scales, and is not affected by haystacking. The invention may be used in combination with computational techniques such as Blade Element Momentum (BEM), Actuator Disk (AD), Lifting Line, Lifting Surface, Free Vortex, Surface Panel, Computational Fluid Dynamics (CFD), such as Reynolds-Averaged Navier Stokes (RANS) and Large Eddy Simulation (LES). The invention may thus be used to aid selection of an inlet condition for a higher order CFD process, in a smaller number of high order simulations, where it is computationally expensive but allows independent validation of some of the scenario cases run to improve confidence in the lower order simulation technique.

The invention may thus enable taking into account the issue of turbulence, along with wave propagation, meteorological effects and many other (typically unsteady) fluid flow and environmental effects, e.g. , the generation of power from wind turbines, tidal stream turbines and wave energy converters. As a result, the invention may be applied in the fields of wind and tidal power engineering for assessment of available resource, energy yield and structural loading characteristics.

The invention may also thus enable taking into account the issue of turbulence, along with wave propagation, meteorological effects and many other (typically unsteady) fluid flow and environmental effects, at the assessment, design and analysis stages of building power plants. As a result, the invention may be applied to computational simulation of aerodynamic and/or hydrodynamic loading on structures and/or devices present in a turbulent flow.

The invention may thus enable taking into account the presence of unsteadiness in the flow and how it affects mean and instantaneous energy yield, as well as through-life dynamic loading on structural, mechanical and electrical components.

The invention may thus have the advantages of Coherent Structural Modeling (CSM) in marine turbulence. The invention may be integrated to many measurement data, in order to provide net improvement in accuracy, robustness or representation of physical behaviour in post-processed results.

The invention may reveal dynamics of the fluid flow for a period of time, but may also be applied to multiple or many scenarios in order to ascertain statistics related to hydrodynamic loads, load distributions, generated power and other performance variables, thanks to its computational efficiency. This 'many-run' use can therefore be used to ascertain peak loadings in normal usage and under extreme environmental loading scenarios, useful in certification processes and specification of component strength and design. The many-run use can also be used to ascertain lifetimes of componentry (knowledge of load spectra allows lifetime and fatigue studies to be carried out), useful in financial modeling as well as cost/value engineering of devices. The many-run use can also be used to provide predictions of the energy yield of potential and existing turbine arrays which take site turbulence into account.

The invention may enable taking into account physical understanding of the fluid dynamic (unsteady) effects involved, such as environmental and turbine-turbine interference effects. Thus the invention may enable the farm to handle the unsteady fluctuations in load and power generation, as well as spatial variations between different devices in the farm, and this during operation of the farm.

The invention may allow integration of all the data within a control approach in a farm, or some of the data may be used in tandem with the existing measurements, in order to further improve performance and robustness of a device and/or an array of devices.

The invention may provide a real-time (or nearly real-time) centralized control of the devices. Where multiple rotors are used, e.g., in an array comprising either several single-rotor devices, or at least one device comprising multiple rotors, the invention may enable a wider array-level control approach, in order to address the effects of the fluid dynamic and environmental and turbine-turbine interference, thus improving the performance of the individual devices as well as the net performance and operability of the array.

A unit according to the disclosure may have:

the ability to make informed selection of control algorithm used (either pre-set or in real time); and/or the ability to incorporate the power of Machine Learning and/or Artificial Intelligence at more than one level, whilst retaining awareness of the physics at play; and/or

the ability to handle high-dimensional problems, multiple solution problems and multiple objective problems.

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

Figure 1 schematically illustrates an example method according to the disclosure;

Figure 2 schematically illustrates an example unit implementing a method according to the disclosure;

Figure 3 schematically illustrates an example boundary layer with wall and a mean flow profile;

Figure 4 schematically illustrates fields of realistically shaped and representative eddies;

Figure 5 schematically illustrates an example of a probability density function, as a function of the strength parameter and/or the size parameter;

Figure 6 schematically illustrates an example representative hairpin vortex with strength and size parameters;

Figure 7 schematically illustrates a field of eddies advecting through a turbine disc, showing induction of velocity at control points on a turbine blade;

Figure 8, already discussed, schematically illustrates a typical layout of a known farm having a SCADA system;

Figure 9 schematically illustrates a farm having a Control, Command And Data Acquisition (CCADA) system; Figure 10 schematically illustrates an exemplary implementation of external data feeds into the CCADA system of a farm comprising redundancy;

Figure 1 1 schematically illustrates an exemplary implementation of redundancy in an array of a farm comprising redundancy;

Figure 12 schematically illustrates an exemplary implementation of a CCADA system using multi-layer control system;

Figure 13 schematically illustrates an exemplary implementation of a CCADA system using multi-layer control system, optionally but advantageously implementing Machine Learning/Artificial Intelligence (ML/AI);

Figure 14A schematically illustrates the principle of an implementation of device-device interaction in an array of devices, or thrust vs. mass flux;

Figure 14B schematically illustrates the principle of an implementation of an array thrust control in a case of an outage coverage, in an array of devices; Figure 15 schematically illustrates the principle of an implementation of response control for spatial variation in the environment in an array of devices;

Figure 16 schematically illustrates an exemplary implementation of an external data feed in a CCADA system using multi-layer control system; and

Figures 17A and 17B schematically illustrate curves representing the behaviour of a device without real-time feature recognition (dotted lines) and the control of the device with real-time feature recognition (solid line). In all of the Figures, similar parts are referred to by like numerical references.

Overview of a unit and a method

Figures 1 and 2 schematically illustrate a unit 1 configured to approximate at least one quantity representative of turbulence in a fluid, e.g. , a liquid such as a marine flow in one non-limiting example. As described in more detail below, the quantity may be a shape, and/or a size, and/or a strength and/or a distribution of the turbulence, represented by at least one type of coherent structure, referred to below as an eddy type.

The unit 1 is thus mainly configured to:

determine, in S10, a profile, preferably the mean profile, of flow quantities in the fluid, especially the mean velocity profile;

select, in S1 1 , at least one eddy type, preferably two eddy types, representative of the turbulence, utilizing experience and available validation data such as visualisations or measurements to select the shape types;

determine, in S12, a turbulent quantity signature, such as a velocity signature (such as a velocity defect signature) and/or a turbulent intensity signature (usually both), for each selected eddy type; and

perform, in S13, a deconvolution of the determined signatures from the profile(s), preferably the mean profile(s), determined in S10, in order to determine, for each selected eddy type, at least one probability density function 41 as a function of the strength parameter and/or the size parameters. In S13 the unit uses at least one determined signature to perform the deconvolution of the profile, e.g. a mean profile, such as the mean velocity profile.

Figure 3 schematically illustrates a boundary layer determined in S10.

As known to the skilled in the art, the boundary layer is a region of fluid flow adjacent to a solid (or relatively solid) surface. As explained in the introductory part, atmospheric and tidal flows contain large scale boundary layers, with land (or the sea surface) acting as the boundary (in atmospheric flows) and a seabed 1 1 acting as the boundary in marine flows. In many flows, especially geographic ones, the Reynolds Numbers of boundary layers are large enough for the boundary layer to be turbulent - i.e. be comprised of a profile, e.g., a mean flow profile, 12 varying with distance from the boundary and a superimposed component that fluctuates in time due to turbulent eddies.

In S10, in order to determine e.g., the mean velocity profile, the unit 1 may be configured to receive simulation data or measurement data. As explained in greater detail below, the measurement data may be provided by measurement devices, such as point instruments (e.g. marine Laser/Acoustic Doppler Anemometers (LDAs) and Microstructure Profilers) or instruments having a measurement domain limited to a line (e.g. Acoustic Doppler Current Profilers, ADCP) and other anemometer data.

In order to determine e.g., the mean velocity profile from the simulation data or measurement data, the unit 1 may be further configured to use at least one of the following: an analytical solution, an analytical fit to un- converged (noisy) measurement data, and/or a converged average of raw measurement data.

In S10, in embodiments where a mean flow of the fluid over a first time window (e.g. the time span of measurement data over many tidal cycles) is not representative of the mean flow (e.g. for marine use where mean flow over a long period of time is close to zero), the unit 1 may be further configured to apply a filtered value (such as a windowed average or value filtered using a Kalman filter) in order to use data over a second time window narrower than the first to be representative of a mean flow. This is the case e.g. for marine use in order to perform decomposition for turbulence at a given point in the tidal cycle. For example, a windowed average may be applied to measurement data over a timescale at least as long as that associated with the turbulent motions (over the second window) but shorter than the a recorded full tidal cycle (the first window) to get a representative mean flow as a function of time throughout the tidal cycle, in the absence of short timescale turbulent motion. As already stated, alternatively data must be taken over a sufficiently large number of tidal cycles that average flows for an entire cycle can be produced.

In S1 1 , the unit 1 selects at least one eddy type representative of the turbulence, utilizing experience and available validation data such as visualisations or measurements including the mean profile to select the shape types.

The invention takes advantage of the fact that turbulence in boundary layers 12, as shown in Figure 3, comprises coherent structures, as shown in Figures 4 and 6, especially 'hairpin vortices' 33 (so called due to the appearance of a common eddy type). Figure 4 shows a field of turbulent eddies 33 having complex shapes in a flow. The field of coherent structures 33 may preferably be simplified by a field of analogs or representative structures 32, It is appreciated that the simplified representative structures 32 may be used for computational efficiency by the unit 1 , compared to more complex structures 33.

To represent the coherent structures present in turbulence, at least one type of representative structure ('eddy type') is used. Each eddy type comprises:

at least one straight or arcuate line vortex element 321 , 322;

a strength parameter 23 (i.e. K, representative of the circulation as known to the skilled in the art); and

at least one geometric parameter 22 determining the size, orientation to the flow direction, orientation to the boundary, and location in the wall-normal direction of the line vortex element(s).

Figure 6 shows an example eddy type known as a hairpin or delta vortex comprising:

four line vortex elements 321 , 322 each having strength parameter 23 (i.e. circulation K);

a single geometric parameter 22; two parallel straight lines 321 , separated by a distance 21 related to size parameter 22 (e.g. 1 .6 times distance parameter h), and oriented in the mean flow direction 31 ; and

two inclined straight lines 322, from an end of the lines 321 and joining at a tip 323. The tip 323 is located at a height and/or size parameter 22 (i.e. height h) from a plane formed by the two parallel straight 5 lines 321 , and at a downstream distance equal to the size parameter 22 from the end of the lines 321 .

Figure 4 shows a field of representative eddies 32 in a flow, all eddies having the same type with a distribution 41 of sizes 22 and strengths 23.

10 The use of the complex coherent structures 33, and/or their representative structures 32, by the unit 1 to perform analyses of turbulence may be referred to as Coherent Structural Modeling (CSM), and the selection of the eddy types by the unit 1 in S1 1 may use observation, characterisation of structures from experimental data and/or use a technique known by the skilled in the art from the articles:

"On the mechanism of wall turbulence", by Perry A.E. and Chong M. S., J. Fluid Mech. (1982), Vol. 15 1 19, which illustrates the concept of the horse-shoe, hairpin or 'A' vortex 34, and shows that these models give a connection between the mean velocity distribution, the broadband turbulence intensity distributions and the turbulence spectra; and

"A wall-wake model for the turbulence structure of boundary layers, Part 1 : Extension of the attached eddy hypothesis" and "A wall-wake model for the turbulence structure of boundary layers, Part 2: Further

20 experimental support", by Perry A. E. and Marusic I. , J. Fluid Mech. (1995), Vol. 298, which illustrate the eddy hypothesis wherein two eddy types, representative of the turbulence, enable determination of all the components of the Reynolds stresses. The first type may be referred to as type-A and shown in figure 6 may be interpreted as giving a 'wall structure'. The second type may be referred to as type-B and may give a 'wake structure'. If the above mean velocity formulation is accepted, once the eddy geometries are fixed for the two

25 eddy types, all Reynolds stresses and associated spectra contributed from the attached eddies can be computed without any further empirical constants. This is done by using the momentum equation and certain convolution integrals (or the signature of individual vortices).

The eddy types may be of type A, B and C as known to those skilled in the art or other types according to 30 requirements for a particular flow (e.g. horizontal line vortices oriented in the cross stream direction to account for shear layers in the flow). Eddy types may be 'attached' to the boundary wall 1 1 (i.e. have one or more ends of the vortex elements e.g. 321 , 322 touching the boundary) or some distance from it ('unattached'). Distance of unattached eddies from the wall can vary with the eddy size parameter(s) which is typically the case with type A eddies 34 and Type B eddies as known to those skilled in the art, but may vary with an additional 35 geometric parameter for other eddy types.

In some embodiments, the unit may be configured to select at least one of the following eddy types: A, B, C, single line elements, or rings in order to account for features not necessarily appearing in laboratory or analytical studies of turbulence e.g. , shear layers caused by thermoclines or density gradients in marine flows or additional structural content resulting from wave breaking.

Preferably, the unit 1 is configured to determine at least the strength and/or size distributions 41 for fields 32 of each eddy type such as 'Type A' 34. Thus in S12, the unit determines a velocity defect signature and a turbulent intensity signature, for each selected eddy shape type.

In S12 the determination of the velocity signature (e.g., the velocity defect signature) and/or the turbulent intensity signature is performed with respect to a wall 1 1 normal distance, non-dimensionalised by an eddy characteristic size, for each eddy type selected. Eddy signatures (i.e. deficit functions and/or turbulent intensity functions) describe (for an eddy of particular type having unit characteristic size and strength) the contribution of an individual eddy structure to turbulent intensity, spectra and velocity deficit distributions of a flow containing that single eddy at unit size.

In addition to the presence of a boundary layer wall 1 1 (e.g. , the seabed), the free surface also requires consideration when computing velocity and/or intensity signatures for marine applications.

In S13, the unit 1 performs a deconvolution of the determined signatures, in order to determine, for each selected eddy type, at least one probability density function 41 as a function of the strength parameter and/or the geometric parameter(s) e.g. size 22, as shown in Figure 5. In S14, the unit 1 may perform convolution of at least one probability density function 41 with the eddy signatures for eddies of unit parameter value. This convolution of individual signatures with PDFs 41 of at least one parameter for an entire field of structures 32 allows determination of full turbulent spectra and Reynolds stress distributions in a fluid containing the field of structures 32 whose parameters are distributed according to the PDFs 41 .

In some embodiments, for validation purposes, turbulent spectra derived using this technique can be validated against turbulent spectra directly calculated from unsteady flow measurements.

Unit

Figure 1 also schematically illustrates an exemplary unit 1 further comprising at least one module 120 configured to determine at least one quantity representative of a turbulent stream of fluid.

To advantage, the unit 1 further comprises a reconciliation module 5 configured to reconcile the at least one determination of the at least one module 120 and the at least one approximation of the at least one unit 1 according to the disclosure, e.g. performed by at least one dedicated module 1 10.

The unit 1 , and particularly the at least one module 120, may thus be used to advantage where data (such as measurement data) integrity is sufficient to be reliable for parts of the spectrum (i.e. a range of scales) or for parts of the spatial domain for which data is required, or where data contains aspects of behaviour of the fluid not encompassed by the CSM model (e.g. internal wave breaking) implemented e.g., by the module 1 10 of the unit 1 . This part of the data from the module 120 is thus taken into account by in the unit 1 , and the CSM model may be applied by the module 1 10 of the unit 1 for other parts, and then taken into account by in the unit 1 .

An optimisation module 51 may be configured to implement Machine Learning and/or Artificial Intelligence to the reconciliation module 5. The module 5 may implement reconciliation or a weighting framework, such as Projection Onto Convex Sets (POCS).

It will be appreciated that one possible advantage of data reconciliation is to provide data conditioning and/or post-processing rooted in physical knowledge of the fluid flow characteristics. This is equivalent to denoising or filtering results, but with a filter selected on the basis of direct observations of the turbulent structure at play in the measurement, as opposed to an arbitrarily selected filter (e.g. gaussian smoothing). Within this framework, results can be based strongly on measurement data by the unit 1 (such as coming from at least one module 120 and/or sensor 20) where it is valid, whilst results depending on data outside a valid measurement range (e.g. outside the band limit or spatial resolution of the measurement instrument or sensor 20, or in locations such as close to the free surface where data integrity is frequently poor due to wave interaction) to be ascertained by the unit 1 using the more robust and less noisy mean flow data from the module 1 10 of the unit 1 . Similarly results can be based strongly on measurement data from at least one module 120 and/or sensor 20 in regions where the CSM module 1 10 of the unit 1 does not appropriately model the physics in play, such as part of a domain subject to internal wave breaking or a region of the spectrum where a particular bathymetric feature causes a spike - but more strongly on CSM behaviour from the module 1 10 of the unit 1 elsewhere where the model is more valid.

In embodiments involving at least two modules 1 10 or at least two modules 120, the module 5 allows reconciliation of multiple instruments operating in adjacent, overlapping or separate parts of the spectrum or domain. Data from each instrument can be reconciled into a single results set valid across a wider range, based on CSM adjusted and weighted according to confidences by the module 5.

It will be appreciated that the at least one module 120 and/or the at least one module 1 10 and/or the reconciliation module 5 may be implemented at least partially as software or firmware. Some functionalities of the at least one module 120 and/or the at least one module 1 10 and/or the reconciliation module 5 may thus be performed interchangeably, or at least partially merged. Additionally or alternatively the at least one module 120 and/or the at least one unit 1 10 and/or the reconciliation module 5 may be implemented at least partially in a physical casing.

Overview of applications

As shown also in Figure 1 , the unit 1 may output the results to a simulation unit, such as:

a control unit 300 configured to control an operation of an array of at least one device 3 adapted to convert kinetic energy of the turbulent fluid into electrical energy (such as wind turbines, tidal stream turbines and wave energy converters); and/or

a design unit 101 configured to perform a design of at least one device 3 adapted to be placed in the turbulent fluid (such as wind turbines, tidal stream turbines and wave energy converters); a simulation unit 102 configured to simulate a behaviour of the turbulent fluid.

It is thus appreciated that in examples and as explained below, a representation of turbulence in the fluid (e.g. , comprising coherent structures 34 as shown in Figure 6) whose strength, distribution, shape and size is determined by the unit 1 , using very preferably a CSM, can be used in a hydrodynamic or aerodynamic model, such as computational simulation (e.g. in calibration of turbulence models for Computational Fluid Dynamics packages) and/or experimental verification of aero/hydrodynamic performance of energy generation devices such as wind and tidal turbines. The strength, distribution, shape and size of the coherent structures 34 may comprise the turbulent spectra and the Reynolds Stresses.

The unit 1 may be configured to output the turbulent spectra and the Reynolds stress distributions of the stream to a simulation unit 102 configured to use at least one of the following models known to those skilled in the art: a Free Vortex Model, a Lifting Line and/or Surface Analysis, a Panel model, and/or a Blade Element Momentum Model, and/or Synthetic Eddy Methods (as described e.g., in the article "A New Divergence Free Synthetic Eddy Method for the Reproduction of Inlet Flow Conditions for LES", by Poletto R. , Craft T. and Revell A., Flow Turbulence Combust (2013) 91 :519-539), to produce inlet conditions representing the actual measured turbulence at a site, suitable for finite volume and particle based Computational Fluid Dynamics analyses, such as Large Eddy Simulation (LES) and Smoothed Particle Hydrodynamics.

As shown in Figures 2, 4 and 6, in examples in S15 the unit 1 may be configured to:

generate a distribution field of the selected at least two eddy types (e.g. Type A eddies 34), oriented in the streamwise direction 31 of the fluid;

determine a velocity u(x,y,z,t) 51 induced by the generated distribution field, at at least one control point (x,y,z) 53 at an instant t, by implementing a Biot-Savart law as known to those skilled in the art; and advect the generated distribution field through the simulation domain in the flow direction 31 over time.

The unit 1 may be configured to distribute the at least two eddy types randomly in the streamwise direction 31 and in the cross-stream direction and/or according to a predetermined distribution function.

In examples where the unit 101 and/or 102 is configured to using a Blade Element Momentum (BEM) Model or a Free Vortex Model (FVM), the control points 53 may be located on a blade 52 of a device 3, such as a turbine, or elsewhere (e.g. control points of a wake sheet or other elements in the simulation), at a point in time t.

The unit 1 may maintain the eddy types 32 and/or 33 constant over time, or modify the eddy types as a result of a self-influence and/or an influence of at least one device 3 placed in the turbulent fluid and/or wakes. Examples of applications

It is known that ADCPs are capable in measuring mean flow profiles (hence 'Current Profilers') and an ACP output dataset typically consists of a set of mean flow profiles, as expected at different points throughout the tidal cycle.

In an example of an application of a unit according to the disclosure, and as shown in Figure 1 , an ADCP unit 20 may thus be deployed gathering high-resolution time resolved data for one or more lunar cycles. A microstructure profiler 200 may also be deployed, measuring small scale (e.g. sub 1 m) scales.

In S10 the unit 1 , e.g. the modules 1 10 and/or 120, implements conventional post-processing to the data (windowed averaging or other filtering), producing mean profiles at different points in the tidal cycle. Analysis and treatment of the high resolution ADCP data from the unit 20 produces turbulent spectra, but they are usually highly noisy, band limited and suffer from haystacking. Analysis of the data from the Microstructure Profiler 200 produces a partial turbulent spectrum at the small scales (outside the band of measurement of the ADCP) with good confidence but at a single point in the water column.

Therefore, preferably, the unit 1 uses data from the module 1 10 where a CSM according to the disclosure is preferably applied to the mean flow profiles from the ADCPs 20 to ascertain PDFs 41 of structural content, and their evolution throughout the tidal cycle. Smooth turbulent spectra across the entire relevant bandwidth are produced for each profile.

These turbulent spectra produced by the module 1 10 are compared by the module 5 to the ADCP data from the unit 20 and may be found to be within the measurement error. However if at the small scales they do not match with data from the location of the Microstructure Profiler 200, the module 5 implements a reconciliation algorithm, and thus reapplies the CSM according to the disclosure using the module 1 10 and good quality data from the Microstructure Profiler 200, in order that the physical model applied better represents motion at the dissipative scales for which reliable data is available.

Using the updated CSM data, PDFs 41 of turbulent structural content (size and strength distributions as a function of time) are used to create a field of eddies and calculate fluctuations in velocity components in the flow as a function of time.

These components are superimposed onto velocity components from other effects (e.g. self influence or influence of other turbines, influence from waves on the free surface and bulk motion) to calculate the velocity field at the control points of a Free Vortex Model. The simulation is time-stepped, allowing computation of loads and energy yield as a function of time, including the dynamics associated with turbulence.

Using this example analysis, dynamic loadings and turbine behaviour as a result of turbulence and as a function of time can be calculated in a low order (computationally efficient) manner. Wave loadings and pressure gradients (e.g. due to bathymetric effects) may also be included through superposition of appropriate potential flow fields. The approach described above reveals dynamics of the fluid flow for a period of time; however due to its computational efficiency, the approach may be applied to multiple or many scenarios in order to ascertain statistics related to hydrodynamic loads, load distributions, generated power and other performance variables.

This 'many-run' use can therefore be used to ascertain peak loadings in normal usage and under extreme 5 environmental loading scenarios, useful in certification processes and specification of component strength and design.

The many-run use can also be used to ascertain lifetimes of componentry (knowledge of load spectra allows lifetime and fatigue studies to be carried out), useful in financial modeling as well as cost/value engineering of devices.

10 The many-run use can also be used to provide predictions of the energy yield of potential and existing turbine arrays which take site turbulence into account.

Use of CSM to aid selection of an inlet condition for a higher order CFD process can be used in a smaller number of high order simulations - computationally expensive but allowing independent validation of some of the scenario cases run to improve confidence in the lower order simulation technique.

15

Overview

Figure 9 schematically illustrates a farm 71 comprising at least an array comprising at least one device 73 adapted to produce energy from the movement of a fluid, i.e. using kinetic energy of a stream or flow of fluid to produce electrical energy. The fluid may be a gas (such as air) or a liquid (such as sea water), and the device 20 73 may thus comprise at least one wind turbine, tidal stream turbine or wave energy converter.

The array may be multiple-rotor, i.e. the array may be formed of at least two rotors of one of the devices 73, or the array comprises at least two single-rotor devices 73.

The farm 71 comprises a Farm Control Unit (FCU) 300 configured to form a Control, Command And Data

Acquisition (CCADA) system. The CCADA is configured to form a controller for the array of at least one device 25 73. The unit 7300 is thus configured to:

determine a state of the array using real-time data relating to the operation of at least one of the devices 73 of the array; and

modify the operation of at least one device 73 of the array, as a function of the determination.

The state of the array refers to the operation status of all the devices 73 linked to the unit 7300. The unit 7300 30 is preferably linked to more than two devices 73. The status of one of the devices may, e.g., be "on" operation

(under different possible regimes of operation, as it will be apparent from the present specification) or "off" operation (deliberate shut down or breakage).

As non-limiting examples, real-time data may refer to at least one of the following parameters of the device 73: torque, thrust, rotations per minute, shaft strain and/or shaft stress.

35 It will be appreciated that, in the specification, the term "real-time" encompasses "near real-time", i.e. the only time delay introduced between:

the occurrence of an event during the operation of at least one of the devices 73 of the array and the transmission of the data to and/or the receipt of the data by the unit 7300

is introduced by the basic processing of raw data by the device 73 and/or the data transmission time from the 40 device 73 to the unit 7300. There are thus no significant delays. It is appreciated that in the farm of Figure 8, the SCADA system 72 does not process real-time data or near real-time data, as the data is processed first at least by the controls 730. The controls 730 then transmit delayed feedback of processed data to the SCADA system 72.

It will be appreciated that the modification of the operation of the at least one device 73 of the array may also preferably be real-time (and thus near real-time), i.e. the only time delay introduced between:

the determination of a state of the array and

the modification of the operation of at least one device 73 of the array

is introduced by the basic processing of raw data by the unit 7300 and/or the command transmission time from the unit 7300 to the device 73. There are thus no significant delays.

It is appreciated that in the farm of Figure 8, the SCADA system 72 does not modify the operation of the device in real-time, as the command is processed first at least by the controls 730.

The unit 7300 may further be adapted to use data external to the devices 73, such as:

environmental data, e.g. local area sensor data 76; and/or

economic, operational and logistic data, e.g. external and/or Wide Area Network data 77 (the feed for the economic and operational data may be indirect, as explained in greater detail below).

In order to control the array of devices 73, the control unit 7300 is also thus configured to:

generate, as a function of real-time data relating to an operation of at least two devices 73 in the array, at least one real-time control response for at least one of the devices 73, by implementing, when generating each control response, a respective control algorithm; and

output a real-time control command for the at least one device 73, as a function of the generated real-time control responses.

Preferably, the unit 7300 is configured to generate at least one real-time control response for all the devices 73 in operation in the array, as a function of real-time data relating to the operation of all the devices 73 in operation in the array. The unit 7300 is then preferably configured to output a real-time control command for all the devices 73 in operation in the array.

As shown in Figure 9, control lines 731 , or loops, interconnect the control unit 7300 and at least two, preferably each, of the devices 73.

The farm 71 may also comprise a conventional Supervisory, Command And Data Acquisition (SCADA) system 72.

The SCADA system 2 may typically have data input 721 for non-real-time processed data from the unit 7300 (such as coming from a sensor of a device 73 or for run-time data), as shown in Figure 9. The SCADA system 72 is also preferably configured to send, via an output 722, non-real-time operational commands (such as an on/off command or an operation set point) to the unit 7300. The SCADA system 72 may also have a data output 723 to, e.g. , a shoreside server facility 74.

The facility 74 is linked, via command and/or data links 741 and 742, to an online interface and/or dashboard 75 operated by a human and/or automated operator of the farm 71 . The operator may then feed the server facility 74 with commands and/or data via the interface and/or dashboard 75.

Therefore, the control lines 731 are preferably able to transmit both commands and data, such that the system 72 may be operated as a conventional farm comprising only a SCADA system 72 as shown in Figure 8, e.g. , in the event of a failure or a maintenance operation of the unit 7300. Furthermore, command and supervisory functions may appear identical to a typical SCADA from the perspective of an operator, which may also facilitate retro-fitting on existing farms. The server facility 74 may also have an input 743 for an external and/or Wide Area Network (WAN) data 77 feed, feeding data such as environmental data and/or economic and operational data as explained in greater detail below.

Preferably, communication down control lines 731 , and to shore, may be made via fibre optic cables, as the latter have a high bandwidth, but other forms of communication, such as TCP/IP, are also possible.

It will be appreciated that Figure 9 is only schematic, and that the unit 7300 may be physically located at a shore station, in a subsea hub or similar subsea container, on a service platform (at or above sea level, floating or fixed) or within the devices (e.g. inside nacelles or platforms) in the array. It will also be appreciated that the unit 7300 may be implemented at least partly as software or firmware, i.e. relying on parallel or distributed computing capability (e.g. distributed to a cloud server, or to local processors in other control systems within the array).

Similarly, the SCADA system 72 may be implemented at least partly as software or firmware, and be a part of the unit 7300 and/or be a part of the server facility 74, and/or be located in a physical casing, separate from the unit 7300 and/or the server facility 74.

Optionally, each device 73 may have an onboard backup controller 730. However, in an advantageous normal way of operation, control functions are undertaken centrally by the unit 7300.

As known to those skilled in the art, registration signals may be periodically sent via the same communication infrastructure to and from the FCU 7300 (e.g. by multiplexing a periodic registration signal down the same lines as the control and data signals) allowing the unit 7300 to identify and safely manage communication interruptions (e.g. broken cable or similar fault states or planned interruptions and outages).

As shown in Figures 10 and 1 1 , in order to mitigate the risk of failure of an FCU 7300 or of communication lines to the unit 7300, multiple physical FCUs 7300 may be incorporated as redundant units into an array. In this case, the FCUs 7300 may be incorporated into some device nacelles, making these devices (referred to as 73') key devices, and other devices (referred to as 73) non-key devices. Figure 10 shows the diagrammatic layout of a turbine farm, excluding the cabling, hubs and shoreside features. The non-key turbines 73 may house onboard safety layer with basic backup control system (shown as 730 in Figure 9) which can be activated under a variety of conditions e.g. fault state of the remote controller. The Farm Control Units 7300 may reside in subsea hubs, platforms, non-locally or in the key turbine 73'.

It will be appreciated that array maintenance operations preferably take into account the requirement for minimum redundancy when removing key devices 73' or e.g., subsea hubs containing FCUs 7300. Redundancy (multiple FCUs, one or more active at a given time) allows key turbines 73' or FCUs 7300 to be disconnected for maintenance. The degree of redundancy (ratio of key to non-key turbines in the illustrated case) preferably allows for both the likelihood of fault states and the likelihood of key turbine 73' removal (maintenance, etc.). In an extremely redundant scenario, an FCU 300 may be incorporated into each device 3. Figure 1 1 shows that which FCU is active at any given time may be controlled via operational input or via fault state logic. Whilst the system is foreseen to be run with a sole FCU active at a time in order to prevent conflicting behaviour, modes in which multiple FCUs run at once may be envisaged.

The currently operating FCU 7300 in a farm with redundancy can be switched manually by an operator (locally or remotely), and/or automatically on occurrence of a fault state or communication interruption.

Farm Central Unit As already mentioned, the unit 7300 is configured to generate at least one real-time control response for at least one of the devices 73 by implementing at least one respective control algorithm. The unit 7300 thus enables the use of a plurality of models and control algorithms, for real-time (and near-real-time) control of the devices 73 of the array. The control is said to be real-time (or near-real-time as explained above) because it may use real-time data, e.g., relating to an operation of at least one device (preferably at least two devices, and very preferably all the devices, in operation in the array), i.e. knowledge of the current operating state of the devices, the said device being taken either from the point of view of a device or from the point of view of a sensor as explained in greater detail below. The unit may use real-time data external to the devices, such as environmental data and/or economic and operational data.

At least one sensor in the device 73 is configured to measure the real-time state of the device (such as data relating to torque, thrust, rotations per minute, shaft strain and/or shaft stress) thus determining the state of the array. The unit 7300 is in turn configured to receive the data (e.g., the measurements of the sensor) via the control lines 731 forming a feedback loop for each of the devices 73. The unit 7300 is furthermore configured to apply at least one control algorithm to correct for any deviation between the desired and actual states, i.e. modify the operation at least one device 73 of the array or output a real-time control command for the at least one device 73.

The desired state can be set in advance (it is the case e.g. , for a basic set operating point) or dynamically updated based on user commands or automated processes, such as machine learning to improve performance over time as described in more detail below. It is appreciated that a single desired state (i.e. one value for each of the independent variables) must be chosen (either by the operator or through the automated process). For example, a standalone turbine with fixed pitch blades may have as little as one independent variable (e.g. terminal voltage varied in order to control rotational speed and shaft torque for optimal power generation). However, even for a single dependent variable (as in the example above: maximising power generation) it is appreciated that there may be multiple methods for determining what state will maximise the desired property. For the simple example above, lookup tables of known performance (embodying engineer's experience), a machine learning and/or artificial intelligence (ML/AI) algorithm, or an optimisation algorithm could be used, and different methods would give different responses to the input state.

Figures 12 and 13 show examples of a unit 7300 comprising at least one algorithm module 7303 configured to implement at least one control algorithm. The unit 7300 thus forms a Multi-Layered Control System (MLCS). The unit 7300 may thus take advantage of the fact that, in some circumstances, one algorithm may yield a more desirable response than another. In other cases, several algorithms may each have an associated inaccuracy; so blending of several algorithms can improve the overall accuracy in prediction of the most favourable response. In still other cases, some models might capture important aspects of physics at play, offering a better insight into the likely dynamics of the response than those which do not.

The algorithm module 7303 may be configured to implement at least one of the following control algorithms and/or models, described in greater detail below:

Array-As-A-Sensor;

Array-As-A- De vi ce ;

Operational response; and/or

Real-time feature recognition response; and/or

Wave propagation;

Thrust/mass flux distribution response; Unsteady inflow vs. yield response;

Unsteady flow vs. peak load response;

Unsteady flow vs. dynamic range response;

Fluid feature and debris advection; and/or

Awareness models, including at least one of the following:

Set point response;

Peak allowable load awareness model;

Lifetime awareness model;

Economic awareness model;

Operational and/or logistical awareness model; and/or

Environmental data response.

It will be appreciated that in a device array, there are many independent control variables, many sensors and highly complex interactions occurring between devices and the environment. Many different additional methods for selection of the best operating point may also be used.

It will be also appreciated that each different response model provides a different, possibly conflicting, recommendation for the command to the devices. The unit thus comprises a reconciliation module 7301 configured to reconcile the real-time control responses for the devices 73 from the modules 7303, in order to output the real-time control command to at least one, preferably all, of the devices 73. The MLCS thus offers ability to make informed selection of the control algorithm used, either statically (e.g. by defining preset conditions under which each method is used, with blending or weighting between them) or dynamically (e.g. using Machine Learning (ML) and/or Artificial Intelligence (Al), or similar, to determine under which conditions each method performs best, then select and blend accordingly).

As known by those skilled in the art, each artificial intelligence and/or machine learning module can be viewed as a 'black box' which receives raw sensor data and outputs the best response, having received some initial training (the latter based usually on engineer's insight). However the integrity and rate of learning decrease with the number of degrees of freedom, and modules implementing ML/AI are not 'physics aware' which means that their learning is based on minimising the difference between expected and actual behaviour based on correlation between key variables. Therefore the capacity of artificial intelligence and/or machine learning modules to identify and respond to unusual events is limited, as is awareness of why and when key variables are correlated. Figure 13 shows an example using ML/AI in a more powerful way by compartmentalising the tasks required.

To advantage, the unit 7300 thus further may comprise an optimisation module 7302 configured to implement Machine Learning and/or Artificial Intelligence to the reconciliation module 7301 .

Similarly, the unit 7300 may further comprise at least one optimisation module 7304 configured to implement Machine Learning and/or Artificial Intelligence to at least one of the algorithm module 7303.

It is thus appreciated that by implementing the optimisation module 7302 directly and only to the reconciliation module 7301 and/or by implementing an optimisation module 7304 directly and only to one of the algorithm modules 7303 (thus compartmentalising the tasks required), the integrity and rate of learning stays high, because the number of degrees of freedom is locally limited, and some algorithm modules 7303 may be 'physics aware', with better capacity to identify and respond to unusual events and with awareness of why and when key variables are correlated. For example, in lower stages (such as in feature recognition module 7303), the ML/AI module 7304 provides fine tuning to individual models, improving accuracy of prediction. In higher stages (such as in array-as-a-sensor), the ML/AI module 7304 can be used in a 'physics aware' mode, accepting inputs relating to physical metrics or model outputs (rather than just raw sensor signals) to assist in the reconciliation process in the reconciliation module 7301 .

The unit may thus enable the operation of systems with high-degree of freedom having multiple solutions for possible commands, because the reconciliation process in the module 7301 has the inherent capacity to handle multi-objective problems. An example would be in an array of for example four turbines, attempting to generate a maximum power. With any two turbines operating off-design (i.e. at a lower power), local acceleration of flow through the other two turbines can result in a local maximum in the hyperspace that consists of the independent control variables and the power. However, elsewhere in the hyperspace, a different solution with greater or equal power (such as all turbines operating uniformly) can exist. By making the reconciliation module 7301 aware of different aspects of the physics and/or economics in play (through use of multiple models in the algorithm modules 7303), the MLCS has the ability to identify and suppress the effect of multiple solution problems, since a different version of the hyperspace is used in each model, and the control algorithms do not get caught in local minima or oscillate between different solutions.

Control algorithms and models

As shown in Figure 13, the unit 7300 may use different control and/or management models, including at least one of the following models or algorithms:

Array-As-A-Sensor, e.g. , implemented in the module 73031 ;

Array-As-A-Device, e.g. , implemented in the module 73032;

Operational response, e.g. , implemented in the module 73033.

Arrav-As-A-Sensor (AAAS) model

In the AAAS model, the module 73031 is configured to observe the state of other devices (and wider sensor data 77) to predict impending events and/or input state changes elsewhere in the array. The AAAS module 73031 may preferably incorporate the principle of the device as an upstream sensor. The module 73031 allows the ultimate device and array response to be updated in real or near-real time as a response to events happening elsewhere within the network.

Arrav-As-A-Device (AAAD) model

In the AAAD model, the module 73032 is configured to manage the state of some, or all, of the devices in the array, in order to improve performance of under-performing individual devices within the array, as well as the net array performance. For example, thrust distribution throughout an array can be managed to optimise net array yield, as explained in greater detail below. The AAAD module 73032 may advantageously incorporate a module 73034 implementing a device-device-environment interaction model.

Operational response model

In the Operational response model, the module 73033 is configured to combine engineering metrics, operational and economic cost functions and/or constraints to provide optimal trade-off between maximisation of revenue and operational costs, including device lifetime management and assisting in decision making between lost revenue and unplanned or more frequent maintenance or shorter lifetimes. The model may incorporate Economic and Logistic metrics, as described in more detail below.

The above-mentioned models or algorithms may use at least one of the following sub-models.

Real-time feature recognition (RTFR) response;

Wave propagation;

Thrust/mass flux distribution response;

Unsteady inflow vs. yield response;

Unsteady flow vs. peak load response;

Unsteady flow vs. dynamic range response;

Fluid feature and debris advection; and/or

Awareness models, including at least one of the following:

Set point response;

Peak allowable load awareness model;

Lifetime awareness model;

Economic awareness model

Operational and/or logistical awareness model, and/or

Environmental data response. Real-time feature recognition response

The real-time feature recognition response, described in more detail below, allows disaster mitigation, peak load relief and unsteady yield optimisation in wind, wave and tidal industries, as it allows a signal or group of signals to be monitored in real time (or near-real time), and features within the signal identified (e.g. blade break, impact, impingement of a particular type of turbulent flow structure). The real-time feature recognition response may take into account at least one signature data of upstream events.

As shown in Figure 13, at least one of the RTFR modules 7303 may utilise an optimisation module 7304 implementing a learning pack. The learning pack comprises instructions configured to teach the module 7303 to recognise events. For example, impact of marine debris on a blade can be simulated in a laboratory or virtual environment and the results used to create a learning pack for identification of such events. Learning packs may include but are not limited to: marine mammal proximity, marine mammal strike, bird proximity, bird strike, fish proximity, fish strike, impact of debris, impingement of turbulent flow structures of different types, sizes or strengths, blade breakage, gearbox breakage, short circuit or other electrical power system events, vessel proximity, UAV/ROV proximity, plane or helicopter proximity. Wave propagation

Using knowledge of wave parameters (e.g. height) and direction at one position, a predictive model can be used in at least one of the modules 7303 to ascertain the passage of waves in real time from that position to another in the same locale. For example, a wave passing over a wave height sensor (this is an example of sensors integrated at a farm level rather than at a device level) with known direction and known bathymetry will propagate some distance DX in time DT. For example, as waves pass over turbines 73 in an array, fluctuations in the inflow velocity affect the turbine. Using a wave propagation model, measurement of waves at the edge of an array allows prediction of the time at which waves arrive at turbine locations within the array, as well as predicting the wave characteristics (e.g. amplitude, wavelength) at the turbine locations. The wave propagation model thus allows fluctuations in load and power throughout the array caused by a wave to be predicted and responded to if necessary.

The measurement of waves at the edge of an array can be done directly, e.g. using sensors measuring wave 5 height and direction (or similar parameters), or can be done indirectly, e.g. by recognition of the signature of a wave affecting a turbine at the edge of an array.

As explained above the wave propagation model can be refined (or established entirely, in the lack of an analytical, computational or experimental model as a basis) using machine learning or artificial intelligence.

10 Thrust vs. mass flux distribution response

For a given device installation, the thrust applied to the flow by the device, e.g., a turbine, is related strongly (and nonlinearly) to the mass flow through the device. It is this relationship that gives rise to various models (such as the Betz Limit) for optimal power output of a turbine, where Thrust and Mass flux are optimally balanced for greatest power capture. Even for a single turbine, increasing the thrust coefficient decreases

15 mass flow through the device, and as known by those skilled in the art, there exists an optimum thrust (which in some cases can be analytically derived, e.g. , by the Betz' Theorem) at which power is maximised.

An important aspect, as shown in Figures 14A and 14B, is that altering thrust at one turbine affects thrust, mass flow and power output at all other turbines in an array (to varying degrees depending on flow conditions and relative locations and taking into account aspects such as viscous performance, bathymetry, surface

20 effects and interactions between multiple turbines). Figures 14A shows that for arrays, turbines interfere with one another. In Figure 14A, changing thrust coefficient at turbine 73 alters mass flux not only through turbine 73 but also through turbines 73₂ and 73₃. There is an optimal thrust distribution (balance of thrust between turbines) as well as net thrust. Depending on the spatial arrangement of the array layout, there may be multiple optimal distributions and local optima close to arbitrary thrust distributions. Figure 14A shows four (of many)

25 possible configurations in which device-device interaction can take place.

The net thrust has a primary effect on net yield. However it is known that there is a secondary effect on net yield and that the distribution of thrust affects individual turbine yield, loading and wear rates. Thus loading, wear rates and yield can be traded off between individual units in the array for best economic benefit.

In Figure 14A(1 ), low thrust at turbine 73 leads to higher mass flux through turbine 73₂ in its wake. In Figure

30 14A(2), high thrust at turbine 73 leads to low mass flux through turbine 73₂ in its wake. In Figure 14A(3), high thrust at turbine 73 relative to turbines 73₂ and 73₃ leads to higher mass flux through turbines 73₂ and 73₃, outside but adjacent to the turbine 73 stream tube: net thrust affects the net mass flux through and yield of the array. Figure 14A(4) shows that decreasing thrust at turbine 73 relative to turbines 73₂ and 73₃ leads to reduced mass flux through turbines 73₂ and 73₃, outside but adjacent to the turbine 71 stream tube: again, net

35 thrust affects the net mass flux through and yield of the array.

More generally, numerical or experimental models of device (e.g., turbine) arrays (or even careful variation within full scale arrays) may be used to ascertain the relationship between thrust (at each device - i.e. the thrust distribution in the array) and the mass flow and power output of each device (and therefore the entire array). This relationship is the thrust/mass flux response model.

40 Thrust exerted by an individual turbine can be controlled using various means; designing blades to have a particular flexibility or aerodynamic characteristic, regulating voltage or current at the generator using the electrical power system, feathering blades (in a controllable pitch system), altering gearing in the drivetrain or actively altering aero/hydrodynamic characteristics or shape.

Thus, with a thrust/mass flux response model, together with the ability to control turbine thrust, the individual turbine loads and outputs across the array as well as the net array thrust and output can be varied in order to achieve a desirable state.

Figure 14B shows an example of control of the thrust distribution (per 'Device-Device Interaction') in case of a failure of a turbine, in which array yield (and therefore lost revenue) may be de-sensitised to prolonged outage or downtime, by taking into account aspects of operations and maintenance costs, such as including vessel availability, weather windows and lost revenue. In the example of Figure 14B, marine operations in case of a failure of a turbine can be conducted with increased flexibility and less weather constraints (thus in safer and more cost effective conditions). Figure 14B(1) shows normal operation of an array forming a tidal fence. Figure 14B(2) shows that a failure (represented by a missing turbine) causes change in thrust distribution. Mass flux thus increases disproportionally through the failed turbine, and the remainder of the array suffers decreased yield in addition to the lost yield form the failed device. Figure 14B(3) shows that in a thrust control scenario, devices are operated at a point associated with increased thrust, and distribution of thrust within the array in this way is used to alter (e.g. increase) mass flow through the rest of the array, allowing the net yield to be less sensitive to the outage.

As explained above, the model can be refined (or established entirely, in the lack of an analytical, computational or experimental model as a basis) using machine learning or artificial intelligence.

Unsteady inflow vs. yield response

Whether wave impingement, environmental turbulence from the marine boundary layer and bathymetry, atmospheric turbulence in the wind, or turbulence created by upstream structures or devices, unsteady flow features impinging on a device (as well as in the region of, but not directly impinging on a device) has an effect on energy yield of devices. For example, aero/hydrodynamic performance of blades is affected by the change in stall characteristics of 2D sections resulting from the presence of viscous scale turbulence - while blade sections are also affected by larger scale turbulent motions, affecting the angles of inflow to blade sections in an unsteady way. Thus presence of turbulence and more general unsteady flow effects alter both mean and instantaneous power output from a device. Intensity and scale length spectra of inlet turbulence can be inferred or measured either a priori (i.e. before turbine installation) or during operation. Computational modelling, engineering assumption, correlations from lab and full scale data and machine learning and/or artificial intelligence can all be used to ascertain the effect of turbulence on yield both in the mean and in realtime or near-real time. Similar procedures can be carried out for other unsteady flow effects such as waves, internal waves and advection of thermoclines, etc. The classification of the resultant change in energy yield due to these effects are taken into the unsteady flow vs. yield response model, and includes both the effect of a particular turbulent spectrum on the mean yield, and the effect of individual (or groups of) turbulent structures on the power output of a device as they advect through it.

Having established this response, it can be used to inform siting of turbines as well as improve future energy yield estimates before turbine installation and during operation. Applied in real time or near-real time, the model informs the control unit of the effect of impending unsteady flow features on energy yield, allowing the device response to be dynamically tuned for optimal energy yield through unsteady events. As explained above, the model can be refined (or established entirely, in the lack of an analytical, computational or experimental model as a basis) using machine learning or artificial intelligence.

Unsteady flow vs. peak load response

Similar to the unsteady flow vs. yield response model, the unsteady flow vs. peak load response model allows peak loading on a device to be predicted as a function of an unsteady flow characterisation, either a priori or in real time.

For example, the peak load caused by an identified unsteady flow feature (such as a wave or turbulent structure) about to impinge on a device can be computed from the response model in real or near-real time. The effect of changing turbine control parameters on the peak load caused by that event can be ascertained. Thus using the unsteady flow vs. peak load response model in real time, the control unit can alter control parameters to reduce the magnitude of a load condition about to occur or in progress.

The model can be refined (or established entirely, in the lack of an analytical, computational or experimental model as a basis) using machine learning or artificial intelligence.

Unsteady flow vs. dynamic range response

Similar to the unsteady flow vs. yield response model, the unsteady flow vs. dynamic range response model allows the likely dynamic range (which can be expressed either as an absolute range or as a normalised value relative to the mean) of loading (or other parameters) on a device to be predicted as a function of an unsteady flow characterisation, either a priori or in real time.

For example, typical dynamic ranges caused by a particular sea state likely to occur in the region of a device (e.g. an impending storm event) can be computed from the response model in real or near-real time. The effect of changing turbine control parameters on the dynamic range during the event is also calculable from the response model. Thus using the unsteady flow vs. dynamic range response model in real time, the control system can alter control parameters during the storm event to improve the utility factor and/or lifetime of the device (both of which are dependent on dynamic range, see 'Dynamic range vs. Lifetime Awareness model'). The model can be refined (or established entirely, in the lack of an analytical, computational or experimental model as a basis) using machine learning or artificial intelligence. Figure 15 shows an example where different turbines 73 (or turbine groups) within the same array are subject to different mean flow speeds and different turbulent flow characteristics. In this figure, a group of three turbines experiences a highly turbulent environmental flow while a neighbouring pair of turbines experiences less intense turbulence. More severe turbulence (higher intensity at the important scales) can increase the dynamic range of the loadings experienced by devices. To compensate, a more conservative control strategy is put in place to relieve cyclic/unsteady loadings and extend component lifetimes. Less turbulent inflow on this side leads to lower fatigue rates and peak loading on devices. A less conservative control strategy can be applied to these devices, favouring yield over fatigue life. This maximises yield of the array and keeps lifetimes and wear rates consistent between all devices. Fluid feature and debris advection model

As explained above, the wave propagation model ascertains, for a given wave event at the boundary of an array or some distance from a device, the time taken for the wave to propagate to the device.

In a very similar way, the fluid feature and debris advection model relates a fluid event at one turbine or sensor (the passage of a patch of turbulence or item of debris) and uses an advection model to ascertain the later time at which the fluid event (or an evolved version of it) advects through another device location.

The advection model can be based on computational simulations (e.g. from software such as of the marks Telemac or Mike21 ) by observing flow through the field then computing Lagrangian trajectories originating at all device or sensor locations. Timescales can be derived from the Lagrangian trajectories between the originating points and the closest passing point to relevant locations downstream.

One possible method of determining relevant locations is to take a Lagrangian trajectory and use a cone whose angle is governed by the turbulent viscosity of the fluid (i.e. the rate at which turbulent diffusion occurs at the appropriate Reynolds number). Any devices residing in the downstream cone may be considered to be at risk of debris impact, or affected by turbulent features, where debris or turbulent features pass through the originating location of that trajectory. The cone may have a non-singular radius at the originating location governed by a characteristic scale length of the debris or turbulent feature. In an example implementation, an item of debris is identified (or impacts) at a device location on the upstream boundary of an array. The debris continues to advect through the array with the flow. Using the fluid feature and debris advection model, the devices downstream which may be affected can be identified, and the time at which they are likely to be affected estimated, in real or near-real time. Control parameters for these devices can be adjusted to respond appropriately to the impending event.

The model can be refined (or established entirely, in the lack of an analytical, computational or experimental model as a basis) using machine learning or artificial intelligence. Awareness models

The awareness models apply desirability or constraint functions to core engineering metrics, allowing the implications of a changing metric to be evaluated. These awareness models are often computed a priori to device deployment but may be updated through-life. They can be used in real time in conjunction with the above models to help inform the control system of the most desirable response to impending and current events.

As shown in Figure 9, external data 7 may be fed into the CCADA control system 7300, possibly via the facility 74. Environmental data (such as tidal chart data, or weather forecast) are typically provided via live feed, or a periodically updated database and/or lookup table. Models for costs and constraints of logistics, operations and power production can be uploaded from 'head office' or a designated authority, allowing the cost/energy ratio value to be optimised by the FCU 7300. The device, e.g. , turbine, model provides the expected performance responses per designer's analyses, maintenance requirements and expected lifetimes (such as Failure Mode Effects Analysis (FMEA) data).

The awareness models may include at least of the following.

Peak allowable load awareness model

The peak allowable load awareness model is informed by engineering simulations and calculations of the structural and electrical properties of the device (and network). As named, it is used to define the allowable peak load for a given structure (beyond which failure occurs or insufficient safety factor is maintained). Set point response

The set point response may take into account models of steady state individual device performance and/or maps of safety or operational constraints.

Lifetime awareness model

The lifetime awareness model is informed by engineering simulations and calculations of the structural and electrical properties of the device (and network). It is used to define the relationship between engineering metrics (such as time histories of loading, impulsive loading events, peak load values, mean load values, and dynamic ranges) and the expected product lifetime and maintenance intervals.

Economic awareness model

The economic awareness model uses economic analysis (incorporating for example revenue from sale of power, subsidies, contractual penalties, and cost of capital) and considerations of operational constraints (such as vessel cost and availability, weather windows and port operations) in order to relate engineering performance metrics to costs and financial returns associated with operating an array.

Operational and/or logistical awareness model

The operational and/or logistical awareness model uses operational costs and constraints (such as vessel cost and availability, weather windows, port operations, knowledge of scheduled maintenance operations) in order to ascertain cost and ability to undertake marine operations at given times.

Environmental data response

As shown in Figures 9 and 16, the environmental data response may take into account wide area network data or external environmental data 77, such as tidal chart and/or weather forecast and/or sea state forecast.

As shown in Figures 9 and 16, the CCADA control system or unit 7300 may also take into account data 76 from at least one local area sensor, such as coming from e.g., an Acoustic Doppler Current Profilers (ADCP) (far upstream) and/or wave height and direction sensors (far field).

It is known that Acoustic Doppler Current Profilers (ADCP) have some ability to measure turbulence, and they can be used in combination with a unit according to the disclosure in order to derive turbulent quantities from a mean profile provided by the ADCP, mean flow profiles converging more rapidly than turbulence metrics. The ADCP may thus enable taking into account the issue of turbulence, along with wave propagation, meteorological effects and many other (typically unsteady) fluid flow and environmental effects, in the generation of power from wind turbines, tidal stream turbines and wave energy converters. As a result, the disclosure may be applied in the fields of wind and tidal power engineering for assessment of available resource, energy yield and structural loading characteristics.

All awareness models can be simplistic (e.g. take a single parameter of maximum peak load allowed) or advanced in nature (e.g. adjust the peak load allowable depending on the time history of the loading and other metrics from elsewhere in the network, or a combination of variables).

In an example application of the awareness models, collection of engineering metrics from an operating device takes place in real (or near-real) time by the device's control unit. The awareness models are used (again in real or near-real time) to convert between these metrics and time-to-maintenance and/or time-to-failure estimates, which can in turn be used in calculation of cost metrics (which may include cash flow parameters, cost of energy, ROI metrics, etc.). The control unit compares these estimates to target values, and uses the response models to update control parameters (i.e. provide a control response) in order to optimise the output of the farm to meet the operator's requirements.

Figures 17A and 17B illustrate exemplary applications of the models for tidal array control.

The example of Figure 17A relates to a debris impact. In that example, an item of debris impacts a device 73 (e.g., "Turbine 1 ") at t1 , causing blade breakage (see the interrupted line in Figure 17A). The module 73031 implementing the Array-As-A-Sensor model uses real time feature recognition implemented by module 7303 on Turbine 1 , in order to identify the event. The module 7303 implementing the fluid feature and debris advection model in the module 73031 identifies which other devices (e.g., "Turbine 2") in the array may be affected by the debris advecting through the site, and identifies a danger period ΔΤ from the likely time (given by At) at which debris will reach each device location (i.e. Turbine 2). The module 73033 implementing the Operational model for example may clearly indicate that potential damage is to be avoided due to the cost of maintenance far outweighing lost revenue for the danger period (likely to be a constraint hard coded within the Operational model module 73033 and not requiring any financial calculation in this clear-cut case). The module 73032 implementing the Array-As-A-Device model is temporarily overridden by the module 7301 implementing the reconciliation algorithm in the MLCS unit 7300, since the Operational model applies a constraint. The unit 7300 thus outputs a command in order to perform a managed shut down of affected devices (i.e. Turbine 2) for their danger periods before bringing them back to full power (see solid line in Figure 17A). The dotted lines show that without the disclosure, Turbine 2 continues to operate and is at risk of similar damage from a convecting object (i.e. dotted line similar to interrupted line).

Once devices which remained intact are back online, a module 73034 implementing a thrust/mass flux distribution response model, along with allowable peak load and dynamic range awareness models, in the Array-As-A-Device module 73032 may utilise to redistribute thrust in the array in order to maximise yield without violating engineering constraints. The Operational module 73033 may interact with the Array-As-A- Device module 73032 to determine whether to maximise yield using thrust redistribution as shown in Figure 14B (penalising lifetimes of the remaining devices) until the next planned maintenance period, or schedule unplanned maintenance to recover lost revenue (and if so indicate the window in which it must be scheduled), or whether to sacrifice yield in order to meet desired lifetimes and maintenance intervals.

The example of Figure 17B relates to an energetic gust event. In that example, an energetic gust impinges on Turbine 1 at t1 (see interrupted line). The load spike can be similar in magnitude to the impact of Figure 17A, but the signature of the feature is different. The module 73031 implementing the Array-As-A-Sensor model uses real time feature recognition implemented by module 7303 on Turbine 1 , in order to identify the event. The module 7303 implementing the fluid feature and debris advection model in the module 73031 identifies which other devices (e.g., "Turbine 2") in the array may be impinged by the gust through the site, and identifies an energetic period ΔΤ from the likely time (given by At) at which the gust will reach each device location (i.e. Turbine 2). The unit 7300 may then output a command in order to dynamically tune Turbine 2 to the gust as it arrives (see solid line), capturing more energy from the gust and reducing dynamic load range. The dotted lines show that without the invention, Turbine 2 continues to operate has a similar response to Turbine 1 . In another example (not shown in the Figures) a large wave propagates. In that example, a wave is recorded by a sensor buoy at the edge of an array. The Array-As-A-Sensor module 73031 applies Real Time Feature Recognition to the sensor buoy data, ascertaining wave amplitude, period and direction. The wave propagation model (with knowledge of the bathymetry) ascertains the times at which that wave will arrive at each device in the array. The Array-As-A-Device module 73032 uses the unsteady flow response models to estimate the effect the impending wave will have on the operational state of each device in terms of peak load, energy yield, etc. The effect is further estimated for altered device control responses to determine how altering the device response affects peak load, energy yield, etc., given that the event is about to occur. The Operational model module 73033 indicates whether priority for each device is presently on lifetime preservation or power generation. If reaching the end of a 4-hour energy sale block contract with the energy quota not yet fulfilled, priority will be to maximise yield. If located in a particularly turbulent zone compared to other turbines, and needing to extend the lifetime to meet planned maintenance operations, priority will be to extend lifetime. The reconciliation module 7301 may take account of AAAD, Operational model and AAAS inputs to adjust the devices responses as recommended by the AAAD, noting the weighting preference between yield and lifetime indicated by the Operational model, over the time periods indicated by AAAS.

The above embodiments are to be understood as illustrative examples of the invention.

Further embodiments of the invention are envisaged. It is to be understood that any feature described in relation to any one embodiment may be used alone, or in combination with other features described, and may also be used in combination with one or more features of any other of the embodiments, or any combination of any other of the embodiments. Furthermore, equivalents and modifications not described above may also be employed without departing from the scope of the invention, which is defined in the accompanying claims.

Claims

1 . A method for approximating at least one quantity representative of turbulence in a marine flow, comprising a unit (1 ):

determining at least one mean profile of the marine flow;

determining a velocity signature and/or a turbulent intensity signature, for each selected eddy type; and

using at least one said determined signature to perform a deconvolution of the mean profile, in order to determine, for each selected eddy type, at least one probability density function (41) as a function of the at least one size parameter and/or the strength parameter.

2. The method according to claim 1 , further comprising the unit (1):

performing a convolution of the at least one probability density function (41) with the at least one determined signature for each eddy type, in order to determine turbulent spectra and/or Reynolds stress distributions of the marine flow.

3. The method (1 ) according to any one of claims 1 or 2, further comprising the unit (1 ):

outputting the at least one probability density function (41 ) and/or the turbulent spectra and/or the

Reynolds stress distributions of the marine flow to a control unit (300) configured to control an operation of an array of at least one device (3) adapted to convert kinetic energy of the turbulent marine into electrical energy.

4. The method (1 ) according to any one of claims 1 or 2, further comprising the unit (1 ):

Reynolds stress distributions of the marine flow to a design unit (101 ) configured to perform a design of at least one device (3) adapted to be placed in the marine flow.

5. The method according to any one of claims 1 to 4, wherein the selected eddy type comprises at least one of the following: type A, and/or type B, and/or type C, and/or a single line element, and/or a ring.

6. The method according to any one of claims 1 to 5, further comprising the unit (1 ):

receiving simulation data and/or measurement data in order to determine the mean profile, and: determining the mean profile using at least one of the following: an analytical solution, an analytical fit to unconverged measurement data, and/or a converged average of raw measurement data.

7. The method according to any one of claims 1 to 6, where a mean flow of the marine flow over a first time window associated with turbulent motions in the liquid is below a predetermined threshold, further comprising the unit (1 ):

applying a filtered value or using data over a second time window narrower than the first time window, so as to generate a mean flow of the marine flow.

8. The method (1 ) according to any one of claims 1 to 7, further comprising the unit (1 ):

performing the approximating one or more times.

9. A method for approximating at least one quantity representative of turbulence in a fluid flow interacting with at least one device (3) adapted to convert kinetic energy of the turbulent fluid flow into electrical energy, comprising a unit (1 ):

determining at least one mean profile of the fluid flow;

determining a velocity signature and/or a turbulent intensity signature, for each selected eddy type; using at least one said determined signature to perform a deconvolution of the mean profile, in order to determine, for each selected eddy type, at least one probability density function (41) as a function of the at least one size parameter and/or the strength parameter; and

outputting the determined at least one probability density function (41 ) to a simulation unit (101 , 300) configured to simulate, in a simulation domain, a behaviour of the turbulent fluid flow interacting with at least one device (3) adapted to convert kinetic energy of the turbulent fluid flow into electrical energy for design or operation purposes.

10. The method according to claim 9, further comprising the unit (1 ):

performing a convolution of the at least one probability density function (41) with the at least one determined signature for each eddy type, in order to determine turbulent spectra and/or Reynolds stress distributions of the fluid flow; and

outputting the determined turbulent spectra and/or Reynolds stress distributions of the fluid flow to the simulation unit (101 , 300).

1 1 . The method according to claim 9, further comprising the unit (1 ):

generating a distribution field of the selected at least one eddy type, oriented in the streamwise direction (31) of the fluid flow;

determining a velocity u(x,y,z,t) induced by the generated distribution field, at at least one control point (x,y,z) at an instant t, by implementing a Biot-Savart law; and

advecting the generated distribution field through the simulation domain in the flow direction over time.

12. The method according to any one of claims 9 to 1 1 , further comprising the unit (1):

outputting the determined probability density function (41 ) and/or turbulent spectra and/or the

Reynolds stress distributions of the fluid flow and/or the induced velocity to the simulation unit (101 , 300) configured to use at least one of the following models: a Free Vortex Model, a Lifting Line and/or Surface Analysis, a Panel model, and/or a Blade Element Momentum Model.

13. The method according to claim 10, further comprising the unit (1 ):

outputting the turbulent spectra and/or the Reynolds stress distributions of the fluid flow to the simulation unit (101 , 300) configured to use Synthetic Eddy Methods, for finite volume and particle based Computational Fluid Dynamics analyses, such as Large Eddy Simulation and Smoothed Particle 5 Hydrodynamics.

14. The method according to any one of claims 12 or 13, further comprising the unit (1 ) outputting the determined probability density function (41 ) and/or turbulent spectra and/or the Reynolds stress distributions of the fluid flow and/or the induced velocity to

10 the simulation unit (300) forming a control unit (300) configured to control an operation of an array of at least one device (3) adapted to convert kinetic energy of the turbulent fluid flow into electrical energy and/or the simulation unit (101) forming a design unit (101 ) configured to perform a design of at least one device (3) adapted to be placed in the fluid flow.

15 15. The method according to any one of claims 1 1 to 14, further comprising the unit (1 ) distributing the at least one eddy type:

randomly in the streamwise direction and in the cross-stream direction; and/or

according to a predetermined distribution function.

20 16. The method according to any one of claims 9 to 15, further comprising the unit (1):

maintaining the eddy type constant over time, or

modifying the eddy type as a result of a self-influence and/or an influence of at least one device (3) placed in the turbulent fluid flow and/or at least a wake.

25 17. The method according to any one of claims 1 to 16, further comprising:

at least one module (1 10) of the unit (1) performing the deconvolution;

at least one module (120) of the unit (1 ) determining at least one quantity representative of the turbulence in the flow; and

a reconciliation module (5) reconciling outputs of the modules (1 10, 120).

30

18. The method according to claim 17, further comprising an optimisation module (51 ) implementing Machine Learning and/or Artificial Intelligence to the reconciliation module (5).

19. The method according to any one of claims 17 or 18, wherein the reconciliation module (5) implements a 35 Projection Onto Convex Sets, POCS, algorithm.

20. The method (1 ) according to any one of claims 9 to 19, further comprising the unit (1 ):

performing the approximating one or more times.

40

21 . A unit (1) for approximating at least one quantity representative of turbulence in a marine flow, configured to:

determine at least one profile of the marine flow;

select at least one eddy type representative of the turbulence, wherein each eddy type comprises at 5 least one eddy size parameter and an eddy strength parameter;

determine a velocity signature and/or a turbulent intensity signature, for each selected eddy type; and use at least one said determined signature to perform a deconvolution of the profile, in order to determine, for each selected eddy type, at least one probability density function (41 ) as a function of the at least one size parameter and/or the strength parameter.

10

22. A unit for approximating at least one quantity representative of turbulence in a fluid flow interacting with at least one device (3) adapted to convert kinetic energy of the turbulent fluid flow into electrical energy, configured to:

determine at least one mean profile of the fluid flow;

15 select at least one eddy type representative of the turbulence, wherein each eddy type comprises at least one eddy size parameter and an eddy strength parameter;

determine a velocity signature and/or a turbulent intensity signature, for each selected eddy type; use at least one said determined signature to perform a deconvolution of the mean profile, in order to determine, for each selected eddy type, at least one probability density function (41 ) as a function of the at 20 least one size parameter and/or the strength parameter; and

output the determined at least one probability density function (41 ) to a simulation unit (101 , 300) configured to simulate, in a simulation domain, a behaviour of the turbulent fluid flow interacting with at least one device (3) adapted to convert kinetic energy of the turbulent fluid flow into electrical energy for design or operation purposes.

25

23. The unit according to any one of claims 21 or 22, implemented at least partially as software or firmware.

24. The unit according to any one of claims 21 to 23, implemented at least partially in a physical casing.

30 25. A method for approximating at least one quantity representative of turbulence in a marine flow, comprising a unit (1 ):

determining at least one profile of the marine flow;

35 determining a velocity signature and/or a turbulent intensity signature, for each selected eddy type; and

using at least one said determined signature to perform a deconvolution of the profile, in order to determine, for each selected eddy type, at least one probability density function (41 ) as a function of the at least one size parameter and/or the strength parameter.

40

26. A method for approximating at least one quantity representative of turbulence in a fluid flow interacting with at least one device (3) adapted to convert kinetic energy of the turbulent fluid flow into electrical energy, comprising a unit (1 ):

determining at least one profile of the fluid flow;

5 selecting at least one eddy type representative of the turbulence, wherein each eddy type comprises at least one eddy size parameter and an eddy strength parameter;

determining a velocity signature and/or a turbulent intensity signature, for each selected eddy type; using at least one said determined signature to perform a deconvolution of the profile, in order to determine, for each selected eddy type, at least one probability density function (41 ) as a function of the at 10 least one size parameter and/or the strength parameter; and

15

27. Computer readable storage media having instructions stored thereon which are operable to program a programmable processor to carry out a method according to any one of claims 1 to 20 or 25 or 26 and/or to program a suitably adapted computer to provide a unit according to any one of claims 21 to 24.

20 28. A control unit (7300) configured to control an operation of an array comprising at least two devices (73) adapted to use kinetic energy of a stream of fluid to produce electrical energy, the unit being configured to: determine a state of the array using real-time data relating to an operation of at least one of the devices of the array;

modify the operation of at least one device (73) of the array, as a function of the determination.

25

29. The unit (7300) of claim 28, wherein the real-time data refer to at least one of the following parameters: torque, thrust, rotations per minute, shaft strain and/or shaft stress.

30. The unit according to any one of claims 28 or 29, further adapted to use external data, such as 30 environmental data and/or the economic, operational and logistic data.

31 . The unit according to any one of claims 28 to 30, further configured to:

output a real-time control command for the at least one device (73), as a function of real-time data relating to an operation of at least two devices (73) in the array.

35

32. A control unit (7300) configured to control an operation of an array of at least one device (73) adapted to convert kinetic energy of a fluid into electrical energy, comprising:

at least two algorithm modules (7303), each module (7303) being configured to generate at least one real-time control response for at least one of the devices (73), by implementing, a respective control algorithm; 40 and

a reconciliation module (7301 ) configured to reconcile the real-time control responses in order to output the real-time control command.

33. The unit according to claim 32, configured to:

generate, as a function of real-time data relating to the operation of all the devices (73) in operation in the array, at least one real-time control response for all the devices (73) in operation in the array; and 5 output a real-time control command for all the devices (73) in operation in the array.

34. The unit according to any one of claims 32 or 33, wherein the algorithm module (7303) is configured to implement at least one of the following control algorithms: Array-As-A-Sensor, Array-As-A-Device, Operational response, and/or Real-time feature recognition response, Wave propagation, Thrust/mass flux distribution

10 response, Unsteady inflow vs. yield response, Unsteady flow vs. peak load response, Unsteady flow vs. dynamic range response, Fluid feature and debris advection, Awareness model.

35. The unit according to claim 34, wherein the awareness model includes at least one of the following: Set point response, Peak allowable load awareness model, Lifetime awareness model, Economic awareness

15 model, Operational and/or logistical awareness model, and/or Environmental data response.

36. A control unit (7300) configured to control an operation of an array of at least one device (73) adapted to convert kinetic energy of a fluid into electrical energy, comprising:

at least algorithm module (7303) configured to implement at least one of the following control 20 algorithms:

Array-As-A-Sensor,

Array-As-A- De vi ce ,

Operational response, and/or

Real-time feature recognition response.

25

37. The unit according to claim 36, wherein the algorithm module (7303) is configured to generate at least one real-time control response for at least one of the devices (73), by implementing the at least one control algorithm.

30 38. The unit according to any one of claims 36 or 37, wherein the algorithm module (7303) is further configured to implement at least one of the following control algorithms: Wave propagation, Thrust/mass flux distribution response, Unsteady inflow vs. yield response, Unsteady flow vs. peak load response, Unsteady flow vs. dynamic range response, Fluid feature and debris advection, Awareness model.

35 39. The unit according to claim 38, wherein the awareness model includes at least one of the following: Set point response, Peak allowable load awareness model, Lifetime awareness model, Economic awareness model, Operational and/or logistical awareness model, and/or Environmental data response.

40. The unit according to any one of claims 32 to 39, further comprising at least one optimisation module 40 (7304) configured to implement Machine Learning and/or Artificial Intelligence to the algorithm module (7303).

41 . The unit according to any one of claim 31 to 25, further comprising an optimisation module (7302) configured to implement Machine Learning and/or Artificial Intelligence to the reconciliation module (7301 ).

42. The unit according to any one of claims 28 to 41 , implemented at least partially as software or firmware.

5

43. The unit according to any one of claims 28 to 42, implemented at least partially in a physical casing, separate from the devices.

44. A system (71) comprising:

10 an array of at least two devices (73) adapted to convert kinetic energy of a fluid into electrical energy; and

at least one control unit (7300) according to any one of claims 28 to 43.

45. The system according to claim 44, comprising at least two redundant control units (7300).

15

46. The system according to any one of claims 44 or 45, wherein the device comprises a wind turbine, a tidal stream turbine or a wave energy converter.

47. A method for controlling an operation of an array of at least two devices (73) adapted to use kinetic energy 20 of a stream of fluid to produce electrical energy, comprising a control unit (7300):

modifying the operation of at least one device (73) of the array, as a function of the determination.

25 48. A method for controlling an operation of an array of devices (73) adapted to convert kinetic energy of a fluid into electrical energy, comprising:

at least two algorithm modules (7303) generating at least one real-time control response for at least one of the devices (73), by implementing, a respective control algorithm; and

a reconciliation module (7301 ) reconciling the real-time control responses in order to output the real- 30 time control command.

49. A method for controlling an operation of an array of devices (73) adapted to convert kinetic energy of a fluid into electrical energy, comprising:

at least algorithm module (7303) implementing at least one of the following control algorithms:

35 Array-As-A-Sensor,

Array-As-A- De vi ce ,

Operational response, and/or

Real-time feature recognition response.

40 50. Computer readable storage media having instructions stored thereon which are operable to program a programmable processor to carry out a method according to any one of claims 47 to 49 and/or to program a suitably adapted computer to provide a unit according to any one of claims 28 to 43.