Method of operating wind turbine

Abstract

A method of operating a wind turbine (100) is provided in which the wind turbine (100) is operable in a plurality of different operating modes (41-46) differentiating from at least one of life-time consumption of the wind turbine and energy production of the wind turbine. A sequence of operation modes (305; 306) is determined for a future time period (301). 741-745), in which an optimization parameter (op) is estimated on the basis of the estimated extrinsic parameter (ep). A sequence of operation modes for which the optimization parameters satisfy the optimization objective is selected (305). The method further comprises obtaining a current value (705) of the consumed life of the wind turbine (100) and determining an actual operating mode of the wind turbine (100) for the current point in time (702) taking into account a deviation (720) between the current value (705) of consumed life and the consumed life expected for operation in the selected sequence of operating modes (305). The wind turbine (100) is operated in the determined actual mode of operation.

Description

Technical Field

[0001] The invention relates to a method of operating a wind turbine, wherein the wind turbine is operable in a plurality of different operating modes. The invention also relates to a control system and a wind turbine comprising such a control system, and also to a computer program. Background Art

[0002] The use of wind energy is proliferating. Wind turbines are installed at different locations around the world and are therefore exposed to different environmental conditions. Wind turbines must withstand considerable wind forces acting on the rotor, nacelle and tower of the wind turbine. During its life, the structural components of the wind turbine are subjected to multiple load cycles, which may eventually lead to component failure. Wind turbines can generally be operated in different operating modes, wherein some modes drive the wind turbine more aggressively (generally resulting in higher loads and higher energy production), while other modes drive the wind turbine less aggressively (generally resulting in lower loads and lower energy production).

[0003] Wind turbines may also be exposed to varying environmental conditions, such as wind speed, which may in some cases lead to reduced energy production and in other cases to high loads on the wind turbine and, therefore, high lifetime consumption. It is therefore difficult for the operator to determine the best way to operate the wind turbine. Some optimization methods are known that optimize wind turbine operation to achieve optimization objectives, such as maximizing energy production, minimizing lifetime consumption, maximizing the consistency between energy production and demand, and maximizing revenue. An example of such optimization is described in document WO 2021 / 214152A1. Although this method achieves good optimization for many situations, there may be certain situations where the determined wind turbine operation scheme can still be improved to achieve better results, such as higher energy production or returns, a longer lifetime, or better utilization of the wind turbine during its design lifetime. Summary of the invention

[0004] Therefore, there is a need to improve the operation of wind turbines, and in particular to operate wind turbines in a manner that better achieves operational goals, such as increasing energy production or revenue.

[0005] This need is met by the features of the independent claim. The dependent claims describe embodiments of the invention.

[0006] In one embodiment of the invention, a method of operating a wind turbine is provided, wherein the wind turbine is capable of operating in a plurality of different operating modes, the operating modes being distinguished by at least one of a lifetime consumption of the wind turbine and an energy production of the wind turbine. The method comprises determining a sequence (one or more) of operating modes for a future time period. Determining the sequence comprises estimating an optimization parameter for the future time period and for each of a plurality of different sequences (e.g., candidate sequences) comprising different combinations of operating modes, based at least on an estimated energy production associated with the operation of the wind turbine in the operating mode of the corresponding sequence. The method may further comprise obtaining a current value of at least one parameter, wherein the at least one parameter is an operating parameter and / or an external parameter of the wind turbine, and determining an actual operating mode of the wind turbine at a current point in time, wherein the determination of the actual operating mode is based at least on the determined sequence of operating modes and the obtained current value of the at least one parameter. Further, the wind turbine is operated in the determined actual operating mode.

[0007] The actual operating mode may be allowed to differ from the operating mode prescribed by the operating mode sequence determined for the current point in time, in particular, to differ from the determined sequence (optimal sequence) in which the estimated optimization parameters satisfy the optimization objective.

[0008] Several benefits can be achieved by this method. The determined sequence of operating modes can be or can include a sequence in which the optimization parameter is at its optimum (e.g., at a maximum or minimum value), depending on the type of the optimization parameter. An example is a sequence of operating modes that maximizes energy production or revenue. However, since such a determined sequence may be based on one or more estimated operating parameters and / or estimated external parameters, such a sequence may not always provide the best results in actual operation. In particular, the actual current value of the parameter may be very different from the actual estimate, so that changing the operation to a mode different from the mode specified by the "optimal sequence" may actually improve performance and better achieve the optimization goal. In addition, by determining the actual operating mode based not only on the current value of the parameter but also taking into account the determined (optimal) sequence, the impact of the change in mode on the optimal sequence can be taken into account. This allows the actual operating mode to be selected so that the optimization goal of the entire sequence is better achieved. As an example, an aggressive operation may be selected at present, but an update to the current external parameter (e.g., electricity price) is subsequently received, which is lower than the estimated external parameter, which will produce optimal operation if the curtailment operating mode of the turbine is used, thereby saving life for subsequent use. As another example, if the curtailment operating mode is currently selected due to a low value of an estimated external parameter, but the current value is significantly higher than the estimated value, then changing to an aggressive operating mode of the turbine may now result in more power and revenue being generated at the expense of the lifetime of the turbine. Thus, by determining the actual operating mode of the wind turbine based on both the current value of the parameter and the determined (optimal) operating mode sequence(s), several benefits may be achieved.

[0009] Determining the sequence of operating modes may include determining one or more optimal sequences in which the optimization parameters satisfy the optimization goal. The determined sequence of operating modes may be a sequence of modes in which the wind turbine will operate during the future time period. For example, such an optimal sequence may be used as a prescribed sequence for operating the wind turbine during the future time period, or one or more of such optimal sequences may be used to derive an actual operating mode.

[0010] Optionally, the estimation of the optimization parameter may also be based on an estimation (e.g., an estimated value) of an operating parameter and / or an external parameter for a future time period. The corresponding parameter may be estimated for a time interval of a future time period. The estimated operating parameter may be an estimated life cycle consumption. The estimated external parameter may be an estimated wind condition or an estimated electricity price.

[0011] For example, the estimation of the optimization parameters may also be based on the estimated lifetime consumption. Thus, the end of the life of the wind turbine may be predicted, which may allow a more accurate estimation of the optimization parameters.

[0012] The future (first) time period may include a sequence of time intervals, wherein the sequence of operation modes may be defined by the operation mode for each time interval. The length of the time interval may be, for example, at least one month (e.g., at least 28 days), which may be, for example, 1 month, 2 months, 3 months, 4 months or more months, a quarter, half a year or a whole year. The future (first) time period may start at the current time point.

[0013] The future (first) time period may extend to the end of the lifetime of the wind turbine. The end of lifetime may be a predefined end of lifetime (end of design lifetime, e.g. 20 or 25 years), and / or may be a point in time at which the lifetime of the wind turbine has expired (e.g. as estimated from the corresponding sequence of operating modes). Combinations are also conceivable (e.g. the end of the future time period may correspond to the design lifetime, but may end earlier if the lifetime of the wind turbine expires before reaching the design lifetime).

[0014] The sequence of operating modes may be determined for a first time period in the future, wherein the actual operating mode of the wind turbine may be determined for a second time period in the future, which is shorter than the first time period and at least partially overlaps with the first time period. The first and / or the second time period may include the current point in time. The length of the second time period may correspond to or may be shorter than the length of the time interval of the first time period. A current value of the parameter may be obtained for the current point in time and / or for the second time period, in particular for a second time period shorter than the time interval (which may be an actual (measured, received or obtained) value or a near future estimate).

[0015] In particular, the actual operation mode may be repeatedly determined during a time interval of a future (first) time period (for which only one operation mode is defined for each sequence). The current value of the obtained parameter may be updated accordingly. For example, the determination of the actual operation mode may be performed each time an updated current value of the corresponding parameter is obtained. As an example, the actual operation mode may be regularly determined using a period of 30 days or less, 10 days or less, 2 days or less, 1 day or less, 12 hours or less, 6 hours or less, 1 hour or less or 10 minutes or less. If the parameter is an operating parameter, in particular a consumption life period (e.g., cumulative life period consumption), the period may be between 1 hour and 30 days, such as between 0.5 and 15 days, such as about 1 day. If the parameter is an external parameter, such as an electricity price, the period may be between 10 minutes and 10 days, such as between 1 hour and 5 days. The time interval of the future (first) time period may be at least one month, such as between one month and two years. Thus, optimization of operations may be achieved in both the long term and the short term, without neglecting the impact of short term optimization on long term performance.

[0016] In some exemplary embodiments, determining the sequence of operating modes may include: selecting a sequence from a plurality of different sequences in which the optimization parameters satisfy the optimization objective (e.g., maximization or minimization of the optimization parameters). Such a sequence may be considered to constitute an optimal sequence (for the corresponding optimization parameters). As indicated above, a plurality of such optimal sequences may be determined, e.g., for different sequence groups.

[0017] For example, the optimization parameter may be at least one of energy production, satisfaction of electrical energy demand, or revenue. For example, revenue may be estimated by utilizing an external parameter in the form of an electricity price. The electricity price generally depends on the electrical power demand, in particular the extent to which the demand can be satisfied by the available electrical power (if the available electrical power is less than the demand, the electricity price will generally rise, and vice versa). Therefore, feeding electrical energy into the grid to satisfy the demand may be achieved by feeding electrical energy into the grid when the electricity price is high. Feeding electrical energy into the grid when demand is high (and not feeding energy into the grid when demand is low) stabilizes the power supply of the grid and stabilizes the grid itself; for example, frequency and / or voltage fluctuations caused by a mismatch between the available power and the power demand may be reduced. Therefore, feeding energy into the grid in order to maximize revenue may also reduce grid fluctuations, stabilize the grid, and ensure a stable power supply of the grid. Maximizing revenue may particularly correspond to maximizing power feed under high power demand.

[0018] The optimization parameter may specifically correspond to revenue. The optimization parameter may be a net present value (NPV), which may correspond to the cumulative revenue from the production of electrical energy of the wind turbine determined from a current point in time (e.g. until the estimated end of the life of the wind turbine). The NPV may be determined by summing the revenue estimated for time intervals for future time periods. The revenue may be a discounted revenue, which may be discounted depending on how far in the future the time interval is, e.g. to take into account the risk of not realizing the revenue (e.g. for time intervals further in the future the security of the predicted revenue is lower and the risk of wind turbine failure etc. may increase). By such an optimization parameter the reliability of the optimization may be increased.

[0019] The optimization objective may be, for example, maximization of an optimization parameter in the form of energy production, satisfaction of power demand (e.g. how well the estimated power output corresponds to the power demand estimated based on external parameters) or revenue, or may be minimization of an optimization parameter in the form of consumption life.

[0020] The multiple different sequences may be candidate sequences. For each possible combination of operating modes for a future time period, a candidate sequence may be provided. Optimization parameters may be estimated for each candidate sequence. Thus, all possible combinations of operating modes may be considered when determining one or more optimal sequences.

[0021] For different candidate sequences, the future time period may have different lengths, for example, the life span of the wind turbines of different candidate sequences may be estimated to end at different future time points, ie, the end of the life of the wind turbines may be different.

[0022] Optimization parameters of the sequence of operating modes may be estimated by estimating wind conditions (particularly wind speed) for a future time period, estimating a lifetime consumption of the wind turbine when operating with the sequence of operating modes under the estimated wind conditions, and estimating an energy production of the wind turbine for operation according to the sequence of operating modes under the estimated wind conditions in the future time period. The optimization parameters may be determined based on the estimated energy production for the future time period and optionally under consideration of an estimate of external parameters. Estimating the optimization parameters of the sequence may further comprise estimating an end of life of the wind turbine based on the estimated lifetime consumption of the respective sequence, which may correspond to an end of the future time period of the respective sequence. After reaching the end of life of the wind turbine, the wind turbine will likely no longer generate energy, which may therefore affect the total energy production achieved with the respective sequence of operating modes and thus the optimization parameters.

[0023] The determination of the optimization parameters may be performed for each of a plurality of time intervals in the future time period. For example, the determination of the optimization parameters may include: estimating the energy production of the wind turbine for each of a plurality of time intervals in the future time period based on the operation mode of the sequence corresponding to the respective time interval, estimating the external parameters for each time interval, calculating the optimization parameters for each time interval based on the energy production and the external parameters estimated for the respective time interval, and summing the optimization parameters determined for the time intervals (until the end of life is reached) to obtain the optimization parameters for the sequence.

[0024] Optionally, the optimization parameters may be discounted for the risk or uncertainty of achieving the optimization parameters in at least some time intervals. Preferably, the discount is higher for time intervals further into the future. Thus, higher uncertainties associated with achieving estimated optimization parameters in the distant future may be taken into account.

[0025] The external parameter may be at least one of wind speed, electrical power demand and electricity price. The electricity price may also be referred to as electrical power price or electrical energy price, and may be measured in Euros / MWh or equivalent units.

[0026] The plurality of different operating modes may include one, two or more degrade modes and one, two or more upgrade modes. The plurality of different operating modes may include at least two, three, four or five of the following modes: a stop mode, in which energy production of the wind turbine is stopped; a life enhancement (LE) operation mode, which reduces the electrical power output of the wind turbine for predetermined operating conditions (e.g., if turbulence above a threshold is determined at the wind turbine, wherein the turbulence may be determined, for example, based on a measured wind speed and load model); a baseline operation mode, in which the wind turbine is operated conventionally (in particular, no upgrade or degrade features are activated); a first performance enhancement (PE) mode, in which the power output of the wind turbine is increased by a first amount (e.g., above rated power, such as by increasing the power limit of the wind turbine under predetermined wind conditions); and a second performance enhancement (PE) mode, in which the power output of the wind turbine is increased by a second amount greater than the first amount.

[0027] The stop mode and the LE mode may be degraded modes, in which the wind turbine is operated less aggressively (i.e., lower lifetime consumption). The first and second PE modes may be upgrade modes, in which the wind turbine may be operated more aggressively (providing a higher power output at the expense of lifetime). The first PE mode may, for example, be a PE5 mode, in which the power output (in the corresponding part of the operating characteristic) is increased by 5%, and the second PE mode may, for example, be a PE10 mode, in which the power output is increased by 10%. These are only example values, and other increases may also be used.

[0028] According to the first aspect of the present disclosure, determining the sequence may also include: selecting an operating mode sequence (which may be referred to as a selected (optimal) sequence) whose optimization parameters satisfy the optimization target. The optimization parameters and the optimization target may have any configuration disclosed herein. In particular, the initial optimization of the selected operating mode sequence will not be repeated for determining the actual operating mode, but the (repeated) determination of the actual operating mode is based on the same initially selected (optimal) sequence. Therefore, the short-term optimization by selecting the actual operating mode can be performed separately from the long-term optimization of the selected (optimal) operating mode sequence. Since there is no need to repeat the long-term optimization, the control can be made more efficient. In addition, the optimization of operation can be achieved in both the long-term and short-term, which can increase the flexibility of mode selection and can lead to an improvement in overall performance, such as improving annual energy generation and / or net present value.

[0029] Furthermore, obtaining the current value of the at least one parameter may include obtaining a current value of a consumed lifetime of the wind turbine. For example, the lifetime consumed by the wind turbine during actual operation up to the current point in time (e.g., cumulative lifetime consumption) may be determined or estimated based on, for example, sensor data, statistical data, etc.

[0030] Determining the actual operating mode may include: taking into account the deviation between the current value of the consumption life period and the consumption life period expected for operation in the selected operating mode sequence (for example, a target value associated with the consumption life period specified for the selected sequence), selecting the actual operating mode from the multiple different operating modes, wherein the selection is performed so as to keep the current consumption life period within the range of consumption life periods expected for the selected (optimal) operating mode sequence.

[0031] In this way, it is ensured that the objectives of the selected (optimal) sequence of operating modes are achieved. The determination of the sequence of operating modes may be based on the expectation of the lifetime consumption during future time periods. When the wind turbine is operated according to the selected sequence and the actual lifetime consumption deviates, the optimization objectives may no longer be achieved. For example, if the lifetime consumption increases during certain time periods, the end of the life of the wind turbine may occur prematurely, and the energy production in the corresponding time range before the expected end of the life may be lost. This negative impact may significantly affect the overall performance of the selected sequence, for example, if good wind conditions and / or high electricity prices are expected (forecasted) for the last time range. By making the selection of the actual mode based on the deviation between the actual consumed lifetime and the expectation of the consumed lifetime, and by allowing the actual mode to be different from the mode of the selected (optimal) sequence, it is ensured that the wind turbine reaches the end of its lifetime envisaged by the selected (optimal) sequence, thereby avoiding the corresponding negative impact. On the other hand, if the consumed lifetime is less than what is actually expected according to the selected sequence of operating modes, selecting the mode according to the deviation may allow the wind turbine operation to make better use of the available resources, for example by selecting an actual operating mode that consumes more lifetime and provides more energy production. Thus, allowing deviations from the selected sequence of operating modes and selecting the actual mode according to the deviations may provide several benefits and in particular may ensure that optimization goals are achieved and the wind turbine is utilized optimally.

[0032] Keeping the current consumption life within the expected consumption life range according to the selected operating mode sequence may include or be performed in the following manner, namely: maintaining, reducing or minimizing the deviation at least within multiple cycles of selecting the actual operating mode. It should be clear that the actual operating mode can be repeatedly selected during operation, for example, repeatedly selected in cycles as described above. Not every operation in the actual operating mode can make the actual consumption life closer to the expected consumption life, but some selections may result in large deviations, which can then be compensated again in the next (subsequent) selection step for the actual operating mode. This may be due to the nature of the process, because it is usually impossible to stay exactly on the estimated life curve of the selected (optimal) sequence. In addition, operating in a mode that will make the current value of the consumption life closer to the expected consumption life may not be beneficial at every moment, for example at high winds and high energy prices. However, the selection is performed so that there is a tendency for the current consumption life to remain close to the expected consumption life.

[0033] In particular, obtaining the current value, determining the actual operating mode and operating the wind turbine in the determined actual operating mode may be performed repeatedly using the same selected operating mode sequence. Thus, a plurality of selection cycles may be performed, for example within the time interval mentioned above. Preferably, the selection of the actual operating mode is performed so as to drive the current value of the consumption lifetime on average over a plurality of repeated selections towards the consumption lifetime expected according to the selected operating mode sequence. Thus, deviations from the selected (optimal) sequence are allowed, for example for exploiting periods of high energy demand etc., while, on average, ensuring that the actual consumption lifetime remains close to the expected lifetime consumption specified by the selected (optimal) sequence. Thus, the selection of the actual operating mode may occur such that, on average, the current lifetime consumption follows the lifetime consumption expected according to the initially selected (optimal) operating mode sequence.

[0034] Thus, in some embodiments, the range may not be a predetermined range, but rather a range around the expected consumption lifespan, which is still acceptable and consistent with a trend to keep the current and expected consumption lifespans consistent. In other embodiments, the range may have predetermined boundaries that may be used in the selection process for the current operating mode. The range may, for example, include lower and upper boundaries that are respectively below and above the expected consumption lifespan, and may be selected so that the current value of the lifespan consumption remains within the lower and upper boundaries.

[0035] The expected consumption life span according to the selected sequence of operating modes may be used in different ways for determining the deviation, for example, it may correspond to a required remaining life span, such as the life span required to reach the end of life defined by the selected sequence (e.g., the end of a future time period). Such a required remaining life span may be used to estimate a target consumption life span at the current point in time, or an estimated end of life when starting from the actual current consumption life span, etc.

[0036] In one example, the expected consumed life span for the selected operation mode sequence may be determined by determining a target value for the consumed life span of the wind turbine at the current point in time based on the selected operation mode sequence, wherein the deviation is the deviation between the target value for the consumed life span at the current point in time and the current value of the consumed life span. Thus, it may be determined whether the consumed life span is more or less than expected, and the deviation may be used to make the selection of the actual operation mode.

[0037] As another example, an estimated current end of life of the wind turbine may be determined based on the current value of the consumed life and the expected life consumption of operations in the selected sequence of operating modes, thereby determining an expected consumed life for the selected sequence of operating modes, wherein the end of the future time period may correspond to a target end of life, wherein the deviation may be a deviation between the target end of life and the estimated current end of life. The target end of life may be estimated based on the sequence of operating modes, for example when selecting the sequence, for example when performing an initial optimization based on an optimization objective. The target end of life may be an end of life to be reached for the selected sequence.

[0038] Determining a target value for the consumed lifetime of the wind turbine (for the current point in time) may comprise: calculating a cumulative lifetime consumption associated with operation of the wind turbine in different operating modes of the selected sequence, wherein the lifetime consumption is calculated backwards from the end of a future time period (according to the end of the lifetime of the sequence) to the current point in time to determine a required remaining lifetime of the wind turbine, i.e. the remaining lifetime required to operate the wind turbine according to the selected sequence at the current point in time until the end of the future time period. Such a calculation may provide a reliable reference, which provides a good indication of deviations, and may therefore be used as a basis for selecting an actual operating mode.

[0039] Determining the estimated current end of life of the wind turbine may include adding the lifetime consumption associated with operation of the wind turbine in the selected sequence of operating modes to the current value of the consumed lifetime (e.g., by adding the lifetime consumption for each time interval of the future time period of the corresponding operating mode to the current consumed lifetime). Thus, based on the current consumed lifetime, an end of life may be estimated (referred to herein as the "estimated current end of life"). This may then be compared to the target end of life, i.e., the total lifetime expected when operating in the selected sequence. Thus, it may be quickly and efficiently determined whether the current consumed lifetime is higher or lower than expected.

[0040] In some embodiments, the actual operating mode is selected from a plurality of different operating modes by adopting a random method that takes into account the corresponding deviations. The random method can be configured to reduce the deviations, for example, by selecting the actual operating mode multiple times. Such a random method can provide a fast and efficient selection means, and in particular can allow the actual consumption life period to be consistent with the consumption life period expected according to the sequence within multiple selection cycles.

[0041] The plurality of different operating modes may include at least one or more first operating modes and one or more second operating modes, wherein the one or more first operating modes may have a lower life cycle consumption than the one or more second operating modes. The method may include adjusting the probability of selecting the first operating mode or the second operating mode as the actual operating mode based on a deviation (e.g., a deviation between a target consumption life cycle and an actual consumption life cycle or between a target end of life and an estimated current end of life). Adjusting the probability of selection helps to take into account changes in the deviation when selecting the actual operating mode. The first mode may be a degraded (less aggressive) mode, resulting in lower energy production, and the second mode may be an upgraded (more aggressive) mode, resulting in higher energy production of the wind turbine.

[0042] The method may include assigning a selected probability to one or more first operating modes and one or more second operating modes, wherein the probability depends on the deviation. For example, if the actual current consumption life period is higher than expected according to the selected sequence, a higher probability may be assigned to the one or more first operating modes (and vice versa). Optionally, the probability may be further assigned according to an external parameter. The external parameter may be at least one of wind speed and electricity price.

[0043] For example, the probability of selection may be assigned to two or more first operating modes and / or the probability of selection may be assigned to two or more second operating modes. Thus, more than one mode may be used to drive the current consumption life back to the expected value.

[0044] The probabilities may be distributed between the respective modes depending on the deviation and / or an external parameter (e.g. electricity price). For example, depending on the deviation, a first probability may be assigned to the one or more first modes and a second probability may be assigned to the one or more second modes. The first probability may then be distributed between the one or more first operating modes depending on the current value of the external parameter and / or the second probability may be distributed between the one or more second operating modes depending on the current value of the external parameter. For large values ​​of the external parameter, a higher probability may be assigned to a mode with higher energy production (a more aggressive mode) and vice versa. Thus, on the one hand it is ensured that the actual consumption lifespan is consistent with the consumption lifespan expected according to the selected (optimal) sequence, while on the other hand it is ensured that the prevailing conditions are optimally exploited (e.g. for currently high electricity prices a highly overrated mode is selected with a high probability and for very low electricity prices the wind turbine is stopped with a high probability).

[0045] In an exemplary embodiment, selecting an actual operating mode from a plurality of different operating modes taking into account the deviation may include: assigning a first portion (e.g., a portion, a share, or a small portion) of a value range to one or more first operating modes, and assigning a second different portion of the value range to one or more second operating modes, wherein the sizes of the first and second portions are selected based on the deviation; generating a random number within the value range; and selecting the operating mode associated with the portion of the value range in which the random number is located as the actual operating mode. The first small portion and the second small portion may span the value range; they may not overlap; the one or more first operating modes may have a lower lifespan consumption than the one or more second operating modes (as also mentioned above). The boundaries of the value range may be selected to correspond to the boundaries of a range of expected consumption lifespans of the selected sequence, wherein the actual current consumption lifespan will be kept within this range. By adjusting the portions of the value range, the probability of selecting one of the first operating modes or one of the second operating modes may be adjusted in a simple and efficient manner based on the deviation.

[0046] The value range may be configured to extend equally around a value representing an expected consumption life period (e.g., a target consumption life period or a target end of life period) for the selected sequence of operating modes. The position of a boundary (e.g., epsilon, ε) between a first portion and a second portion of the value range may be determined based on the deviation. The value of the boundary may correspond to a current value of the consumption life period. The first range may extend from a first boundary of the value range to the boundary, and the second portion may extend from the boundary to a second boundary of the value range.

[0047] If the deviation indicates that the current value of the consumption lifetime is higher than the consumption lifetime expected according to the selected sequence of operating modes, the boundary between the first part and the second part can be set so that the first part (corresponding to the mode with less lifetime consumption) spans a larger part of the value range than the second part (and vice versa). Thus, it is made more likely that a mode with low lifetime consumption, such as the LE mode or the stop mode, will be selected. Thus, different selection probabilities can be effectively assigned to the respective operating modes according to the deviation.

[0048] As a simple example, if the current value of the consumption life span is to be maintained between +10% and -10% of the expected consumption life span, the boundaries of the value range from 0 to 1 may be set to correspond to the corresponding values. The boundary between the first portion and the second portion may then be set to a value within the range corresponding to the current value of the consumption life span, e.g., 0.5 if there is no deviation, 0.75 if the deviation is +5%, and 0.25 if the deviation is -5%. Thus, the sizes of the first and second portions, and therefore the probability of a random number falling within the portions (and therefore the probability of selecting the corresponding mode), may be effectively adjusted based on the deviation.

[0049] In one embodiment, the actual operation mode is selected from the plurality of operation modes by an epsilon greedy algorithm, wherein the position of the epsilon parameter within the value range of the epsilon greedy algorithm is determined by the deviation.

[0050] According to a second aspect of the present disclosure, the estimation of the optimization parameter may also optionally be based on at least one (e.g., 1, 2, 3 or more) estimated external parameter (in particular, the estimated value of the external parameter disclosed above). It should be clear that multiple external parameters may be used. Obtaining the current value of at least one parameter may include obtaining the current value of the at least one external parameter.

[0051] Determining the actual operation mode may include estimating an adjusted optimization parameter for the determined sequence (in particular for one or more of the determined sequences) and for the current value of the at least one external parameter. The adjusted optimization parameter may be estimated for each different operation mode at the current point in time. Based on the estimated adjusted optimization parameter, one of the different operation modes may be selected as the actual operation mode.

[0052] By this method, it is possible to estimate how operation in different operating modes at the current point in time affects the optimization parameters, taking into account the current values ​​of the external parameters. In particular, if the current values ​​of the external parameters deviate significantly from the estimated values ​​on which the sequence is based, it may be beneficial to select a mode different from the mode prescribed by the optimal sequence. In addition, since the adjusted optimization parameters are determined for the current values ​​of the external parameters and the different modes at the current point in time, it is possible to estimate how changing the mode at the current point in time affects the optimization parameters in the long term, which may allow an optimal choice to be made for the actual operating mode. For example, if demand and / or electricity prices are very high, driving the wind turbine in the most aggressive mode may not always be optimal, considering the increased life cycle consumption and associated production losses due to the premature end of the life of the wind turbine.

[0053] Selecting one of the different operation modes as the actual operation mode based on the estimated adjusted optimization parameters may include: selecting the operation mode in which the operation at the current time point produces the adjusted optimization parameters that meet the optimization goal. The optimization goal may be any optimization goal disclosed herein, for example, it may be the maximization of the adjusted optimization parameters, i.e., selecting the mode with the highest value of the adjusted optimization parameters for the current operation.

[0054] Estimating the adjusted optimization parameters for different operation modes at the current time point may include: for different operation mode sequences distinguished at least by their operation modes at the current time point (e.g., the current time interval), estimating the dependence of the optimization parameters obtained for the corresponding sequences on the value of the at least one external parameter, and estimating the adjusted optimization parameters for each of the sequences for the current value of the at least one external parameter obtained. Thus, a simple method may be provided that can determine how the current value of the at least one external parameter affects the value of the optimization parameter without having to perform a complete re-estimation of the optimization parameter throughout the future time period when the current value changes.

[0055] The plurality of different operation mode sequences may include a plurality of groups, and each group may include a plurality of different operation mode candidate sequences. Sequences within the same group may include the same operation mode for the current time interval of the future time period (i.e., they may define the same mode for the current time interval). Sequences of different groups may include different operation modes for the current time interval (i.e., the group may differ due to the mode in which the sequence starts within the current time interval). Optimization parameters may be estimated for each different candidate sequence within each group. Thus, information may be obtained about how operations in different modes at the current time point affect optimization parameters.

[0056] Determining the sequence of operating modes for the future time period may include determining an optimal sequence for each different group by selecting candidate sequences from the corresponding group whose optimization parameters satisfy the optimization objectives (e.g., any objectives mentioned herein), wherein the adjusted optimization parameters are estimated for each optimal sequence of the groups. Considering these optimal sequences helps to estimate how changing the operating mode at the current point in time affects the optimization parameters (e.g., NPV) over the lifetime of the wind turbine.

[0057] For each set of optimal sequences, a dependency of the optimization parameter on the value of the at least one external parameter may be estimated to determine an adjusted optimization parameter for the respective optimal sequence. This may allow consideration of how a change in the operating mode may change the overall outcome of the wind turbine operation, taking into account the current value of the at least one external parameter.

[0058] Estimating the adjusted optimization parameter may include: employing a linear dependency or a nonlinear dependency (e.g., a polynomial) between the value of the optimization parameter and the value of the at least one external parameter. This may allow for fast and efficient calculation of the adjusted optimization parameter for each optimal sequence even if the current value of the at least one external parameter is frequently updated.

[0059] The slope of the dependency or another parameter may be determined so that the adjusted optimization parameter is estimated for the operating mode depending on the operating parameter estimated for the operating mode at the current point in time (e.g. for the current time interval). The operating parameter may be, for example, the energy production estimated for the respective operating mode. For example, the revenue or NPV may increase linearly with increasing electricity prices, wherein if the energy production of the wind turbine is higher, the growth rate is also higher, which may be taken into account by this method.

[0060] Estimating the adjusted optimization parameters for the operating mode at the current point in time may include obtaining optimization parameters estimated for a sequence of corresponding operating modes including the current time interval (e.g., for a group optimal sequence). The estimation may be based on at least one external parameter estimated for the current time interval (e.g., the initial optimization may be based on the estimation, as indicated above). In addition, the estimated optimization parameters may be adjusted using the current value of the at least one external parameter and the dependency between the optimization parameter estimated for the current time interval and the at least one external parameter (e.g., using the above-mentioned linear relationship) to determine the adjusted optimization parameters. For example, this may be done for each group optimal sequence to obtain adjusted optimization parameters for each of these group optimal sequences.

[0061] The method may further comprise: estimating energy production of the wind turbine during a current time interval when operating in the respective operating mode (e.g. as part of an initial optimization of a first time period), and estimating a dependency between the optimization parameter and the at least one external parameter based on the energy production during the current time interval. Thereby, an efficient and accurate estimation of the optimization parameter adjusted for the current value of the at least one external parameter may be achieved.

[0062] The determined actual operating mode may be set for the current time interval for a future time period, or may be set for a second shorter time period.If a new current value becomes available, it may be updated.

[0063] It should be clear that the optimal first and second aspects of the present disclosure can be combined. For example, the wind turbine can be operated according to the first aspect and its embodiments. Therefore, an actual operating mode can be selected that makes the current consumption life period consistent with the target consumption life period expected according to the selected (optimal) operating mode sequence. If the life period consumed by the operation is within a predetermined (second) range of the target consumption life period (for example, very close to the target life period), the operation can be changed to a method according to an embodiment of the second aspect of the present disclosure, according to which an actual operating mode can be selected, which provides an overall optimal result for the optimization parameters according to the current values ​​of the external parameters.

[0064] Therefore, in one embodiment, the wind turbine may be operated in a first short-term optimization mode, in which the wind turbine is operated by the method according to any embodiment of the first aspect disclosed herein, and when the deviation between the actual current consumption life span and the target consumption life span of the wind turbine at the current time point reaches or exceeds a threshold, the operation of the wind turbine may be switched to a second short-term optimization mode, in which the wind turbine is operated by the method according to any embodiment of the second aspect disclosed herein. Therefore, benefits of two aspects may be achieved.

[0065] According to another embodiment of the invention, a control system for operating a wind turbine is provided, wherein the control system is configured to operate the wind turbine in a plurality of different operating modes, the operating modes being distinguished by at least one of lifetime consumption of the wind turbine and energy production of the wind turbine. The control system is configured to perform any of the methods disclosed herein. The control system may in particular comprise a processing unit and a memory, wherein the memory stores control instructions which, when executed by the processing unit, cause the processing unit to perform any of the methods disclosed herein.

[0066] According to another embodiment of the present invention, a wind turbine is provided. The wind turbine comprises a control system, wherein the control system is configured to control the operation of the wind turbine in a plurality of different operation modes. The control system may have any configuration disclosed herein, and in particular may be configured to perform any method disclosed herein.

[0067] According to another embodiment of the invention, there is provided a computer program for controlling the operation of a wind turbine, wherein the computer program comprises control instructions which, when executed by a processing unit of a control system controlling the operation of the wind turbine, cause the processing unit to perform any of the methods disclosed herein. The computer program may be provided on a volatile or non-volatile data carrier or storage device and may further be provided via a network connection. The computer program may be operable on any of the control systems disclosed herein.

[0068] Additionally and / or alternatively, another aspect and / or an embodiment in particular being one of the aforementioned aspects and / or embodiments is disclosed, namely a method for operating a wind turbine, in particular an offshore wind turbine, further in particular a floating offshore wind turbine, and / or a method for feeding electrical energy and / or chemical energy to an electrical grid and / or a chemical network, the method comprising the following steps:

[0069] performing the steps of an embodiment or a combination of embodiments of a method as disclosed above, and / or operating an embodiment of a wind turbine as disclosed above,

[0070] generating electrical power and / or electrical energy by means of a wind turbine and in particular generating chemical energy in the form of a gas or liquid, in particular in the form of hydrogen, at least partially from the electrical energy, further in particular by converting the electrical energy into hydrogen by means of electrolysis,

[0071] at least partially transmitting electrical power, electrical energy and / or chemical energy to an energy receiving arrangement, in particular wherein the receiving arrangement is not located in international waters, on shore and / or within a 12 mile zone of a respective sovereign state over which the respective state has jurisdiction, and

[0072] The electrical power, electrical energy and / or chemical energy is at least partially supplied to an electrical and / or chemical utility network, in particular to an onshore electrical and / or chemical utility network and / or to a chemical / electrical storage device.

[0073] The term "chemical energy" includes ways of storing energy in chemical substances, such as hydrogen. For example, the chemical energy network can be a gas (hydrogen) network and / or a storage device.

[0074] It is to be understood that the features mentioned above and the features to be explained below can be used not only in the respective combinations indicated, but also in other combinations or alone, without departing from the scope of the present invention. In particular, the features of different aspects and embodiments of the present invention can be combined with each other, unless otherwise indicated. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The foregoing and other features and advantages of the present invention will become further apparent from the following detailed description read in conjunction with the accompanying drawings, in which like reference numerals represent like elements.

[0076] Figure 1 is a schematic diagram showing a wind turbine including a control system according to an embodiment.

[0077] Figure 2 is a schematic diagram showing functional blocks of a control system according to an embodiment.

[0078] Figure 3 is a flow chart illustrating a method of operating a wind turbine according to an embodiment, which includes determining a sequence of operating modes (as part of long-term optimization).

[0079] Figure 4 Shown is a schematic diagram illustrating different input parameters for an optimization and the progress of the optimization according to an embodiment.

[0080] Figure 5 is a schematic diagram graphically illustrating different operation mode candidate sequences sorted according to the values ​​of estimated optimization parameters.

[0081] Figure 6 is a flow chart illustrating a method of determining an actual operation mode of a wind turbine based on a deviation between a current consumption life period and a target consumption life period according to an embodiment.

[0082] Figure 7 The diagram shows Figure 6 The method obtains a schematic diagram of the target consumption life period from the selected operation mode sequence.

[0083] Figure 8 It is a diagram showing the Figure 6 The method uses a random method to select the schematic diagram of the actual operation mode.

[0084] Fig. 9 It is a diagram showing the Figure 6 The method uses a random method to select the schematic diagram of the actual operation mode.

[0085] Fig.10 is a schematic diagram illustrating the allocation of operation modes according to external parameters, which can be used according to Figure 8 and Fig. 9 The actual operation mode is selected by one of the random methods. DETAILED DESCRIPTION

[0086] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. It is to be understood that the following description of the embodiments is given for the purpose of illustration only and should not be understood in a restrictive sense. It should be noted that the accompanying drawings should be considered as schematic representations only, and the elements in the accompanying drawings are not necessarily drawn to scale with each other. On the contrary, the representations of various elements are selected so that their functions and general purposes become apparent to those skilled in the art. As used herein, the singular forms "one", "one" and "a kind of" are intended to also include plural forms, unless the context clearly indicates otherwise. Unless otherwise stated, the terms "include", "have", "comprise" and "contain" will be interpreted as open terms (i.e., meaning "including but not limited to").

[0087] Figure 1 A wind turbine 100 is schematically shown and comprises a rotor 101 with rotor blades 102. A nacelle 103 of the wind turbine 100 is mounted to a wind turbine tower 104 and is rotatable by means of a yaw drive 105. The wind turbine 100 may also comprise a generator 106, an optional gearbox (not shown) and a converter 107. In operation, wind energy is converted into rotational mechanical energy by the blades 102 of the rotor, wherein the rotation of the rotor 101 turns the generator 106 which converts the mechanical energy into electrical energy. The converter 107 may provide an AC voltage at a desired AC frequency. Although Figure 1 A full converter (FC) topology is shown, but different topologies, such as a DFIG topology, may also be used. Of course, there may also be a transformer receiving electrical power from the converter 107 and other components common to such wind turbines (not shown). The wind turbine may be coupled to a grid to provide electrical power to the grid, for example via a collection network of a wind farm or more directly.

[0088] The wind turbine is controlled by a control system 10 comprising a processing unit 11 and a memory 12. The processing unit 11 may be any type of processor, such as a microprocessor, an application specific integrated circuit, a digital signal processor, etc. The memory 12 may comprise volatile and non-volatile memories, in particular RAM, ROM, EEPROM, FLASH memory, a hard disk drive, etc. The control system 10 may further comprise a communication interface 15 for communicating with an external data processing entity, via which the wind turbine may receive external data, in particular data values ​​of one or more external parameters, such as environmental conditions (e.g. wind conditions, wind speed), energy demand, electricity prices, grid-related information, statistical and / or historical information, etc. The control system 10 may further comprise a user interface (not shown) and other communication interfaces, such as a data acquisition interface, for example for receiving data from one or more sensors 20 of the wind turbine. Such sensors may be configured to measure environmental parameters (e.g. wind conditions, such as speed, direction), wind turbine loads (e.g. strain, bending moment, etc.) and other information useful for operating the wind turbine.

[0089] The control system 10 is configured to control the wind turbine according to a plurality of different operating modes, which differ in at least one of the energy production and lifetime consumption of the wind turbine, and preferably in both. To operate the wind turbine in these different modes, the control system 10 may, for example, control at least one, preferably all, of the pitch drives 109 (pitch angle control) of the blades, the yaw drives 105, the converters 107 and the generators 106 (e.g., via the converters), and other controllable components of the wind turbine.

[0090] The control system 10 may operate the wind turbine in one or more first operating modes that are degraded modes and preserve the life of the wind turbine, for example by rotating the rotor 101 at a lower speed (e.g., pitching the blades 102 outward). The first operating modes may include a stop mode (at least stopping the energy production of the wind turbine for output purposes) and an LE mode. In the LE mode, the control system may be configured to detect turbulence at the wind turbine and reduce the output power of the wind turbine if turbulence above a predetermined threshold is detected. Thus, in the LE mode, fatigue loads on the wind turbine may be reduced at the expense of energy production. Optionally, the first operating modes may include: a simple curtailment mode, in which the output power of the wind turbine is curtailed (e.g., by pitching the rotor blades outward, which reduces stress on the rotor and the wind turbine); and / or a HWRT (high wind ride-through) mode. In the HWRT mode, the control system 10 may be configured to perform a load-based reduction in output power above a predetermined wind speed. For example, if the wind speed exceeds a predetermined value (e.g., 23 m / s) for a certain period of time, the rotational speed of the rotor and the power output of the wind turbine may be gradually reduced. The first operating mode may also include a turbine load control (TLC) mode. The control system 10 may implement a turbine load control system that continuously monitors the structural loads on the wind turbine, and if the measured loads exceed normal operating values, the control system may automatically adjust the operation of the wind turbine to bring the loads back into the design range, such as by reducing the rotational speed of the rotor (pitching the rotor blades outward).

[0091] The control system 10 may also operate the wind turbine in one or more second operating modes, which are upgrade modes and increase the energy production of the wind turbine, which is usually at the expense of the life of the wind turbine (increasing life consumption), for example by allowing a rotor speed higher than the rated rotor speed. The first operating mode may include a first and a second performance enhancement (PE) mode. The PE mode may increase the energy production of the wind turbine by increasing the power limit under predetermined conditions. At high wind speeds, the wind turbine may operate at the limit of the output power (e.g., rated power), but the operation may occur below the load limit of the structural loads of the wind turbine. In the PE mode, the control system increases the rotational speed of the rotor in proportion to the increase in output power. For example, in the corresponding PE mode, the net power generation may increase by up to 5% (PE5) or 10% (PE10).

[0092] The control system 10 may implement other operating modes than the first and second operating modes, such as a baseline mode (i.e., normal operation) in which the wind turbine operates at nominal parameters. Preferably, the control system 10 is configured to operate the wind turbine in at least a baseline mode, a stop mode, a LE mode, and two different PE modes, but two modes (stop and baseline) may be sufficient.

[0093] The control system 10 may be implemented by one or more controllers; it may be implemented by a wind turbine controller of the wind turbine 100, a wind farm controller, an external server, or a combination thereof. Some portions of the disclosed method (e.g., overall optimization) may be performed by one controller (e.g., a wind farm controller) and other portions may be performed by another controller (e.g., a wind turbine controller), or the method may be performed by a single controller.

[0094] The control system 10 is configured to perform a long-term (LT) optimization, in which it determines one or more optimal operating mode sequences, and a short-term (ST) optimization, in which, based on the results of the LT optimization, the actual operating mode at the current point in time is selected and used to operate the wind turbine. Two different methods of performing short-term optimization are disclosed herein, which can be used in isolation or can be combined to synergistically improve the performance of the wind turbine. First, the long-term optimization is described.

[0095] The control system 10 may implement an optimizer 200 that takes into account the operating history of the turbine (i.e., lifetime consumption), future lifetime consumption depending on the operating mode of the turbine and environmental conditions, statistical and short-term predictions of environmental conditions, and economic values. The long-term optimization determines a sequence of operating modes for future time periods (which may include sequences of fixed time intervals, such as years or months) until the end of the life of the wind turbine, for which the optimization parameters estimated for the sequence achieve the optimization goal. For example, the optimization parameter is energy production, satisfaction of energy demand, or revenue, in particular reflecting discounted revenue and net present value (NPV) in time intervals for future time periods.

[0096] Figure 3An embodiment of a long-term optimization method that may be performed by the optimizer 200 is illustrated. In step S11, measured data of the operation of the wind turbine in the past is obtained. In step S12, the past lifetime consumption of the wind turbine is estimated for the current point in time based on the obtained data and / or statistical data. These steps may be performed by a lifetime estimator 210, which, for example, obtains long-term wind turbine data 211 (e.g., wind speed (ws), turbulence intensity (ti), generated power, rotor speed, nacelle acceleration, etc.). The lifetime estimator 210 may include a load simulation block 215 (which may use load simulation and / or an aeroelastic model of the wind turbine), which is configured to determine the lifetime consumed by the wind turbine based on historical data and predict future lifetime consumption for different operating modes of the wind turbine. Through the long-term wind turbine data 211 and the simulation block 215, historical damage 212 may be obtained and historical lifetime consumption 213 may be estimated ( Figure 3 Step S12). Therefore, the consumption life span at the current time point can be obtained.

[0097] Furthermore, based on the load simulation block 215, the dependence of damage or lifetime consumption on wind speed (and / or other environmental conditions, such as turbulence intensity, temperature, wind shear, wind direction, etc.) and on the operating mode can be obtained, which is provided in block 216 (step S13). Furthermore, sensor data of the turbine, such as rotor speed, generated power, strain, etc., can be used in a model for damage, which model can determine in an accurate manner the past damage suffered by the turbine. The optimizer 200 also includes an environmental condition estimator 220, which uses external data 221 (which may include site selection studies, site selection reports, meteorological simulations, meteorological mast data, mesoscale data, etc.) and further uses site-specific information 222 (which may include site-specific Weibull distributions, wind rose diagrams, etc.) to estimate future wind conditions, such as wind speed. Thus, the wind distribution of the site of the wind turbine can be analyzed and an estimation can be made for the future.

[0098] The optimizer 200 also includes a parameter estimator 230, which is configured to estimate optimization parameters according to the operating mode and environmental conditions. As an example, future income can be estimated by the estimator 230, but other optimization parameters, such as energy production or the satisfaction of energy demand, can also be used. The power curve block 231 provides a power curve of the wind turbine for each operating mode. The energy production estimator 232 estimates the energy production of the wind turbine from the power curve according to the operating mode. The estimation can be performed for a time period corresponding to the time interval of the future time period (for example, one year), so that the annual energy production (AEP) of each operating mode can be estimated for each time interval (see also step S13). The estimation can also use environmental data for the corresponding time interval from the estimator 220, in particular the wind distribution. The estimation block 233 can estimate the optimization parameters, for example, for a time interval depending on the corresponding operating mode, for which additional data can be used. An external parameter forecasting unit 234 can be provided, which estimates the external parameters for each time interval of the future time period (see also step S14). If revenue is the optimization parameter, the forecast unit 234 may, for example, provide a forecast of electricity prices, for example from customer expectations of future prices. The additional data block 235 may provide other relevant information for estimating the optimization parameters, such as the internal rate of return when optimizing for revenue. Thus, accurate estimates of the optimization parameters for different operating modes and different time intervals may be obtained by the estimator 230.

[0099] The optimization unit 201 can now determine one or more, preferably a plurality of, sequences of operation modes of the wind turbine for the future time period. As an example, the estimation can be performed for all time intervals and all combinations of operation modes for the future time period. For example, if the life of the wind turbine has been consumed, the estimation may end for a particular sequence. The optimization unit 201 can compile different candidate sequences, each candidate defining an operation mode for each time interval until the end of the life of the wind turbine (step S15).

[0100] The lifetime estimator 210 may include a lifetime consumption block 217 that determines the lifetime consumption (or damage) of the wind turbine based on the estimated environmental conditions (from the estimator 220, such as wind speed, such as a statistical wind speed distribution or mesoscale data obtained from the site) and the corresponding operating mode. The lifetime consumption may refer to the rate at which the lifetime is consumed in the corresponding mode, so that different operating modes may have different slopes in the lifetime graph, and the unit 217 may determine this rate or slope for each mode (e.g., driving the wind turbine harder or in high wind conditions will result in a higher slope of lifetime consumption within the time interval). Thus, for a particular candidate sequence, the lifetime consumption within each time interval may be estimated. The end of life estimator 218 determines the maximum lifetime of the wind turbine for the corresponding sequence of operating modes (and therefore determines the end of life), i.e. determines the time interval over which the lifetime of the wind turbine has been consumed.

[0101] For each operating mode candidate sequence, the value of the optimization parameter is determined at the end of the lifetime of the wind turbine (i.e. at the end of a future time period for the respective wind turbine). For example, the annual discounted revenue is calculated and the net present value (NPV) can be derived from this. The candidate sequences can then be sorted in a list according to the value of the optimization parameter (e.g. if the optimization goal is to maximize the optimization parameter, such as revenue, the highest value can be ranked first). The result is Figure 5 In the figure, Figure 5 A future time period 301 is shown, which has five time intervals 302, each corresponding to one year. The horizontal axis is the time axis (a) in years, whereby the years are counted from the start of operation of the wind turbine. The vertical axis represents the value of the optimization parameter op in arbitrary units (au), which may be, for example, revenue. At the top, the operating mode candidate sequences 305 that produce the highest values ​​of the estimated optimization parameter are shown. Note that for some sequences (e.g., 306, 307), the lifetime of the wind turbine has ended before the 20th year, where Figure 5 The hatching indicated by reference numeral 43 in FIG. 3 corresponds to a wind turbine "death". Mode 41 is LE mode, mode 42 is baseline mode, mode 44 is PE5 mode, mode 45 is PE10 mode, and mode 46 is stop mode. As can be seen, the sequence 305 that produces the highest values ​​stops the wind turbine in years 16, 17, and 18, and operates in baseline and LE modes in years 19 and 20. Note that Figure 5 Only a subset of the candidate sequences are shown.

[0102] Figure 4A diagram illustrating different input parameters of the optimization and the progress of the optimization is shown. The upper diagram 401 shows an estimate of an external parameter ep for a future time period provided by block 234, such as the price of electricity per MWh in Euro (expected future price assumption). Diagram 402 shows the damage rate (or lifetime consumption) Dr as a function of wind speed for different operating modes, in damage values / year. As can be seen, at higher wind speeds, the upgrade modes 44, 45 cause higher damage to the wind turbine, while the stop mode 46 results in almost zero damage.

[0103] Figure 4 The last graph 403 of illustrates the accumulated damage D or consumed life of a wind turbine over time t (in years) (a) until the current time point 405, and further illustrates the estimated future damage (lifetime consumption) for different operating modes, showing different slopes for each mode. The value range is selected so that D=0 corresponds to no damage and D=1 corresponds to lifetime consumption (i.e., end of life). The consumed lifetime can be a health indicator of a wind turbine, which represents the health state as a numerical value, for example, as a value between 0 and 1, where 1 means that the wind turbine is damaged to the extent that it can no longer operate. The thick solid line 404 indicates the previous damage accumulated over the life of the turbine until the current time point 405. The diagonal dashed line 409 indicates the design case. It is assumed that once the accumulated damage (vertical axis) reaches a value of 1, the wind turbine stops operating. In this example, the current value of the consumed lifetime is higher than the design case. This indicates that the wind turbine has been operating more aggressively in its past than expected in the design case. If the wind turbine continues to operate in this manner, it will not reach its design life span, which is shown here as a black vertical line 410. Different operating modes result in different slopes in graph 403, as calculated by block 217. Graph 403 graphically indicates whether the wind turbine will reach its design life span 410 when operated in a particular operating mode.

[0104] In this example, the current year is year 16, leaving the wind turbine in operation for another four years. Due to the price assumptions shown in graph 401 (low prices for the first 18 years, and high prices for the remaining years until the end of life), the optimization unit estimates the highest values ​​of the optimization parameters for a sequence that stops the turbine in years 16, 17, and 18, and operates the turbine in a less aggressive mode in the last two years when prices are high. Figure 5 It can be seen that when operated in a more aggressive mode in the 19th year, the wind turbine reaches the end of its life and fails in the 20th year.

[0105] It should be clear that, in addition to calculating the optimized parameters for all possible combinations of patterns in the sequence, an optimization algorithm, such as a search algorithm or an algorithm from the field of reinforcement learning, can also be used to find the sequence of patterns that produces the highest optimized parameters. The above explanations apply accordingly to this method.

[0106] The results of the optimization unit 201, such as the determined one or more operating mode sequences and their associated estimated optimization parameters, can be selected and set for operating the wind turbine (for example, a sequence that meets the optimization goal, such as the sequence that generates the highest revenue, can be selected as the optimal sequence for operation), and / or the results provided by the optimization unit 201 can be used in further short-term optimization.

[0107] Based on one or more determined sequences of operating modes, the method may further determine an actual operating mode of the wind turbine and operate the wind turbine in this actual operating mode, wherein the actual operating mode is allowed to deviate from the operating mode specified, for example, by the global optimal sequence (i.e. the sequence that produces the highest / best estimated values ​​of the optimization parameters). By means of such a short-term optimization, it may be determined whether there is another operating mode, different from the operating mode determined by the long-term optimization, which operating mode produces better optimized parameter values ​​(e.g. higher revenue) taking into account current conditions, e.g. current values ​​of external parameters such as wind speed or electricity price, and / or current values ​​of operating parameters such as the current consumption lifespan.

[0108] According to a first aspect, current values ​​of operating parameters may be used for short-term optimization. Such current values ​​may be current values ​​of the consumed life span, which may be estimated, for example, by block 213 of the optimizer 200, which obtains historical life span consumption from actual turbine data 211 and a wind turbine model of block 215, from which the current consumed life span (current cumulative damage) may be determined.

[0109] Figure 6 A flow chart illustrating an example of a corresponding method for determining an actual operating mode is shown. Figure 3-5 The method explained above determines a part of the sequence of operating modes, and the method may include selecting a candidate sequence whose optimization parameters satisfy the optimization target as an optimal sequence (step S61), according to which the wind turbine will operate. Figure 5In the example of , this may be the sequence 305 for which the optimization parameter (revenue, e.g. NPV) has the highest value. The wind turbine may then be operated according to the selected optimal sequence. However, since the lifetime consumption of the sequence depends on an estimated parameter (e.g. estimated wind speed) which may differ from the actual value, and during the course of operation the operating mode may change (e.g. by operating the wind turbine according to the second aspect part of the time, e.g. more aggressively), the actual current consumption lifetime of the wind turbine may deviate from the consumption lifetime expected for the current point in time when the optimal sequence of operating modes was derived. As a result, the optimization goal may not be achieved, e.g. due to the wind turbine reaching the end of its lifetime too early. Therefore, the method may determine an actual operating mode which allows deviations from the mode prescribed by the optimal sequence in such a way that the deviation between the current consumption lifetime and the consumption lifetime expected according to the selected (optimal) sequence is kept within a certain range, e.g. kept constant or reduced (at least within a number of cycles of selecting the actual operating mode). It should be clear that the range in which the deviation is to be maintained (e.g. the maximum permissible deviation) may be determined by the operator.

[0110] In step S62, the current value of the operating parameter, in particular the current value of the consumed life span, is obtained. The current value of the consumed life span 705 in particular refers to the accumulated life span consumption from the start of the operation 703 of the wind turbine to the current time point 705, such as Figure 7 Thus, it may also be referred to as current cumulative damage or current cumulative lifetime consumption. The control system 10 may continuously determine the current value of the consumed lifetime 705 since the start of wind turbine operation, or may receive historical data 211 to make such a determination, for example using block 213 described above.

[0111] In step S63 , the end of lifetime 710 determined for the selected optimal sequence is obtained, for example by estimating the long-term optimization of the end of lifetime 710 of the selected sequence in block 218 .

[0112] In order to determine the deviation 720, the expected consumption life 730 of the selected optimal sequence at the current time point 702 can be determined in step S64. In the selected sequence, each operating mode is associated with a life consumption rate and therefore with a different slope in the life graph 702, as determined in block 217. Therefore, starting from the end of life 710 and using the slope of the operating mode specified for the selected sequence for the corresponding time interval 302, the consumption life can be calculated backwards until the current time point 702 is reached, so that the expected consumption life 730 at the current time point is obtained. This value is used as the target consumption life at the current time point. As in Figure 7As can be seen in FIG. 7 , there is a deviation 720 between the current consumption life period 705 and the target consumption life period 730. The deviation 720 can be used in the selection of the actual operation mode at the current time point, in particular to keep the deviation within a certain range, such as keeping it constant or reducing it. Figure 7 The dashed line 740 in illustratively illustrates this design case.

[0113] Instead of this backward calculation, the consumption life span can also be calculated in the forward direction, starting with the current value of the life span consumption and adding the life span consumption for each time interval according to the slope of the pattern specified for the selected sequence of the corresponding time intervals. This can be done until the expected end of life 710 is reached, wherein the deviation will appear as a value of the consumption life span CLT different from 1.0, for example larger (if the current consumption life span is too high) or smaller (if the current consumption life span is too low). Alternatively, the end of life can be estimated by this forward calculation when starting from the current value of the consumption life span to obtain an "estimated current end of life", which can then be compared with the expected end of life 710 of the optimal sequence to determine the deviation (this estimate may assume operation beyond the end of life 710 in the pattern used in the last time interval 740). Therefore, there are different possibilities for detecting deviations between the actual consumption life span and the expected consumption life span.

[0114] There are several possibilities for selecting the actual operating mode based on this deviation. In step S65, the actual operating mode is selected from the plurality of available operating modes using a random method. The random method may be configured to keep the current consumption life span within a target consumption life span range expected for the selected (optimal) sequence of operating modes. This range may have a predetermined limit on the current value of the life span consumption.

[0115] In step S66, the wind turbine is operated in the selected actual operating mode, which is allowed to be different from the mode specified by the optimal operating mode sequence.

[0116] As an example, a random method in the form of an ε-greedy algorithm can be used for selection, which Figure 8 Schematically illustrated in . The maximum allowable deviation (positive and negative) of the current consumption life period 705 from the target consumption life period 730 can be set by the operator, or can be set to a value that provides optimized beneficial results. A value range 800 corresponding to the target consumption life period and the maximum allowable deviation range can be selected, for example by setting a first boundary 801 to correspond to the maximum negative deviation, setting a second boundary 802 to correspond to the maximum positive deviation, and setting a target value 803 that can correspond to the target consumption life period 730. For the maximum allowable deviation that is symmetric about the target life period consumption, the target value 803 can be in the middle of the value range 800. Figure 8In the example of , the value range 800 spans from 0 to 1, but other value ranges may of course be used. The position of the epsilon value ε within the value range corresponds to the current consumption life period 705 (in other words, the deviation 720 corresponds to the distance between the target value 803 and ε).

[0117] One or more first operating modes may be assigned to a first portion 811 of the value range 800, and one or more second operating modes may be assigned to a second portion 812 of the value range 800. A boundary 804 between the first portion 811 and the second portion 812 is at a value ε. The one or more first operating modes are associated with a lower lifetime consumption than the one or more second operating modes, i.e., the first operating mode may be a degraded mode, and the second operating mode may be an upgraded mode.

[0118] The control system may generate a random number from within the value range 800. The control system may then select as the actual operating mode the operating mode associated with the portion of the value range within which the random number falls. Figure 8 In the example of , due to the deviation 720, the value of ε and therefore the limit 804 moves towards the higher boundary 802, so that the portion 812 spans a smaller range of values ​​than the portion 811. Therefore, the probability of selecting the more aggressive second operating mode is lower than the probability of selecting the less aggressive first operating mode. Therefore, when such selection of the actual operating mode is repeatedly performed and the wind turbine is operated in the selected actual operating mode, the current consumption life will be driven towards the expected target consumption life, and it will not cross the maximum allowable deviation. As the deviation 720 increases, ε moves further towards one of the boundaries 801, 802, so that the probability of selecting a mode that re-aligns the current consumption life with the target consumption life increases towards 1 (i.e., towards a 100% probability of selection). Therefore, based on the deviation 720, the probability of selecting one of the first or second modes can be adjusted automatically and in a manner that keeps the deviation within a range around the target consumption life. For example, if ε is at a value of 0.7, the probability of selecting the first operating mode is 70%, and the probability of selecting the second operating mode is 30%.

[0119] Fig. 9 Another representation of a corresponding selection method is shown. A random number r is generated 901 from the value range 800. If r<ε, a first (less aggressive) operating mode is selected 902. If r>=ε, a second (more aggressive) operating mode is selected 903. Thus, the actual consumed lifetime is kept within the lifetime consumption range expected according to the optimal operating mode sequence.

[0120] If the deviation 720 is obtained in a different manner, such as in any of the above-mentioned manners, such a random method may be used accordingly.

[0121] Since there may be a plurality of first operating modes and / or a plurality of second operating modes, a portion 811 of the value range may be allocated to one, two or more first operating modes, and / or a portion 812 of the value range may be allocated to one, two or more second operating modes. This distribution may occur depending on the current value of an external parameter, which may be obtained as described above, such as the current value of the wind speed or the electricity price.

[0122] Fig.10 An example is shown in , which shows a density diagram of the density d of the electricity price ep (arbitrary units) and the thresholds of different operating modes 46 (stop mode), 41 (LE mode), 42 (baseline mode), 44 (PE5 mode) and 45 (PE10 mode). Depending on the current value 910 of the electricity price, the corresponding mode can be associated with the first or second part 811, 812. At the shown high value of the electricity price, the part 812 can be associated with the mode 45 (PE10) and the first part 811 can be associated with one of the degraded modes 902 that still provides sufficient energy production, such as the LE mode 41 or the baseline mode 42. It is to be noted that the baseline mode 42 can be grouped into the first or second operating mode 902, 903. If the price is very low, for example, the stop mode 46 can be assigned to the first part 811. Thus, the current consumption life period can be made consistent with the target consumption life period, while still optimizing the optimization parameters, such as income.

[0123] In an even further embodiment, the respective portions 811, 812 may be subdivided into sub-portions, and each sub-portion may be assigned to one of the operating modes 41-46. The relative sizes of the sub-portions may then be adjusted according to the current value of the external parameter, for example corresponding to the process described above. For high electricity prices, the size of the sub-portion associated with the mode PE10 45 may be increased, and for lower electricity prices, the size of the sub-portion associated with the PE5 mode 44 within the second portion 812 may be increased. Thus, the probability of selecting a particular one of the first or second operating modes may be effectively adjusted according to the current value of the external parameter.

[0124] It should be clear that Figure 6 Steps S62-S66 of the method may be repeatedly performed during operation, for example, in the cycles mentioned above (eg, daily, weekly or monthly). It should also be clear that Figure 3 The long-term optimization and selection step S61 may be repeated at certain time intervals, for example when the estimate on which the long-term optimization is based deviates too much from the actual current conditions.

[0125] Additionally or alternatively, the method may further comprise determining the actual operating mode using the method as described above according to the second aspect. The determination of the actual operating mode may for example comprise determining the operating mode according to the first aspect (e.g., as described with respect to Figure 6-9 The invention relates to a method for controlling the wind turbine in a manner that is consistent with the present invention and that determines the operating mode according to the second aspect when the deviation between the current value of the consumption life span and the target consumption life span falls below a predetermined (static or dynamic) value. In this way, it can be ensured that the life span of the wind turbine is not excessively consumed while making optimal use of external conditions, for example by driving the wind turbine in an aggressive mode when the electricity price is high. When the deviation exceeds a second predetermined (dynamic or static) value, the operation can be continued according to the first aspect, for example to bring the actual consumption life span back into line with the target.

[0126] Although specific embodiments are disclosed herein, various changes and modifications may be made without departing from the scope of the invention. The present embodiments should be considered in all aspects as illustrative and not restrictive, and all changes coming within the meaning and equivalent range of the appended claims are intended to be included therein.

Claims

1. A method of operating a wind turbine, wherein: The wind turbine (100) is operable in a plurality of different operating modes (41-46), the plurality of different operating modes differing in at least one of lifetime consumption of the wind turbine and energy production of the wind turbine, wherein the method comprises: determining a sequence (305) of the operating modes for a future time period, wherein determining the sequence (305) comprises: estimating an optimization parameter for the first time period (750) and for each of a plurality of different sequences comprising different combinations of operating modes, based at least on an estimated lifetime consumption and an estimated energy generation associated with operation in the operating mode of the corresponding sequence, wherein a sequence of operating modes for which the optimization parameter satisfies an optimization objective is selected from the plurality of different sequences; Obtaining a current value (705) of a consumed life span of the wind turbine (100); determining an actual operating mode of the wind turbine (100) at a current point in time (702), wherein determining the actual operating mode comprises: selecting the actual operating mode from the plurality of different operating modes taking into account a deviation (720) between the current value (705) of the consumption life span and a consumption life span expected for operation in the selected sequence of operating modes (305), wherein the selecting is performed so as to keep the current value (305) of the consumption life span within a range of consumption life span expected for the selected sequence of operating modes (305); and operating the wind turbine (100) in the determined actual operating mode, The actual operation mode is allowed to be different from the operation mode specified by the operation mode sequence (305) determined for the current time point (702).

2. The method according to claim 1, wherein: The expected consumption life for the selected sequence of operating modes is determined by: determining a target value (730) for the consumed life span of the wind turbine at the current point in time based on the selected sequence of operating modes (305), wherein the deviation (720) is a deviation between the target value (730) for the consumed life span at the current point in time (702) and the current value (705) of the consumed life span, or An estimated current end of life of the wind turbine is determined based on a current value (705) of the consumed life and an expected life consumption of operations in the selected sequence of operating modes (305), wherein the end (710) of the future time period corresponds to a target end of life, and wherein the deviation is a deviation between the target end of life and the estimated current end of life.

3. The method according to claim 2, wherein: Determining a target value (730) for the consumed lifetime of the wind turbine comprises calculating a cumulative lifetime consumption associated with operation of the wind turbine in different operating modes of the selected sequence (305), wherein the lifetime consumption is calculated backwards from the end (710) of the future time period to the current point in time (702) to determine a required remaining lifetime of the wind turbine required at the current point in time (702) to operate the wind turbine according to the selected sequence (305) until the end (710) of the future time period.

4. The method according to claim 2 or 3, wherein: Determining an estimated current end of life of the wind turbine includes adding a lifetime consumption associated with operation of the wind turbine in the selected sequence (305) of operating modes to a current value of consumed lifetime to estimate the current end of life.

5. A method according to any one of the preceding claims, wherein: The obtaining of the current value, the determination of the actual operating mode and the operation of the wind turbine in the determined actual operating mode are performed repeatedly using the same selected operating mode sequence (305), wherein the selection of the actual operating mode is performed so as to drive the current value of the depletion lifetime towards the depletion lifetime expected according to the selected operating mode sequence (305) on average over a plurality of repeated selections.

6. A method according to any one of the preceding claims, wherein: The multiple different operating modes include at least one or more first operating modes (902) and one or more second operating modes (903), wherein the one or more first operating modes (902) have a lower life cycle consumption than the one or more second operating modes (903), and wherein the method includes adjusting the probability of selecting one of the first operating modes (902) or one of the second operating modes (903) as the actual operating mode based on the deviation (720).

7. The method according to claim 6, wherein: The method comprises assigning selection probabilities to the one or more first operating modes (902) and the one or more second operating modes (903), wherein the probabilities depend on the deviation (720) and optionally on external parameters.

8. The method according to claim 7, wherein: Based on the deviation (720), a first probability is assigned to one or more first modes (902) and a second probability is assigned to one or more second modes (903), wherein the first probability is distributed between the one or more first operating modes based on the current value of the external parameter, and / or wherein the second probability is distributed between the one or more second modes based on the current value of the external parameter.

9. A method according to any one of the preceding claims, wherein: Selecting the actual operating mode from the plurality of different operating modes taking into account the deviation (720) includes: a first portion (811) of the value range (800) is assigned to one or more first operating modes (902), and a second, different portion (812) of the value range (800) is assigned to one or more second operating modes (903), wherein the size of the first portion (811) and the second portion (812) is selected in dependence on the deviation (720), generating a random number (r) within said range of values ​​(800); and The operating mode associated with the portion (811, 812) of the value range where the random number is located is selected as the actual operating mode.

10. The method according to claim 9, wherein: The value range (800) is configured to extend around a value (803), which represents an expected value (730) of the consumption life period expected based on the selected sequence (305), and wherein a boundary (804) between the first portion (811) and the second portion (812) within the value range is determined based on the deviation (720).

11. The method according to claim 9 or 10, wherein: The one or more first operating modes (902) have a lower life cycle consumption than the one or more second operating modes (903), wherein, if the deviation indicates that the current value (705) of the life cycle consumption is higher than the life cycle consumption (730) expected based on the selected operating mode sequence, the boundary (804) between the first part (811) and the second part (812) is set so that the first part (811) spans a larger part of the value range (800) than the second part (812).

12. A method according to any one of the preceding claims, wherein: The actual operation mode is selected from a plurality of operation modes by an ε-greedy algorithm, wherein the position of an epsilon parameter (ε) within a value range (800) of the ε-greedy algorithm is determined by the deviation (720).

13. A control system for operating a wind turbine, wherein: The control system (10) is configured to operate the wind turbine (100) in a plurality of different operating modes (41-46), the plurality of different operating modes differing in at least one of lifetime consumption of the wind turbine and energy production of the wind turbine, wherein the control system (10) is configured to perform a method according to any of the preceding claims.

14. A wind turbine, wherein: The wind turbine (100) comprises a control system (10) according to claim 13, wherein the control system (10) is configured to control the operation of the wind turbine in a plurality of different operating modes (41-46).

15. A computer program for controlling the operation of a wind turbine (100), wherein: The computer program comprises control instructions which, when executed by a processing unit (11) of a control system (10) controlling the operation of the wind turbine (100), cause the processing unit (11) to perform a method according to any one of claims 1 to 12.

Images