Wind turbine maintenance method based on intelligent robot, electronic device and medium
By collecting data using intelligent robots to calculate damage resistance curves and combining this with environmental impact, the final maintenance time for wind turbines can be determined. This solves the problems of downtime losses and delays associated with traditional maintenance methods, enabling accurate and timely maintenance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 深圳市华盛智联科技有限公司
- Filing Date
- 2026-04-29
- Publication Date
- 2026-07-14
AI Technical Summary
Existing wind turbine maintenance methods suffer from significant downtime losses and untimely maintenance. Traditional post-maintenance preparation time is long, and fusion life prediction methods are inaccurate under environmental influences.
By collecting component data through intelligent robots, calculating multiple damage resistance value curves, and combining the types and impact coefficients of environmental impact events, the final maintenance time point for components and wind turbines can be determined, allowing for advance maintenance preparation.
It reduced downtime losses of wind turbine units, improved the accuracy and timeliness of maintenance, and reduced component damage caused by environmental factors.
Smart Images
Figure CN122383618A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of new energy technology, specifically to a wind turbine maintenance method, electronic equipment, and medium based on intelligent robots. Background Technology
[0002] In the field of new energy, wind turbines, as the core equipment for converting wind energy into electricity, are widely used in wind farms. With the continuous growth of global demand for clean energy, the number and scale of wind turbines are constantly expanding, and their stable operation is crucial for ensuring energy supply and reducing carbon emissions. Wind turbine components are prone to wear, aging, and performance degradation, thus requiring regular or irregular maintenance to ensure the normal operation and power generation efficiency of the wind turbines. In existing technologies, there are two main methods for maintaining wind turbines. One is the traditional reactive maintenance method, which typically involves repairing the turbine only after it has failed and is no longer operating normally. The other is a maintenance method that incorporates lifespan prediction. This method primarily uses historical fault information of components to predict their lifespan and then develops corresponding maintenance plans. However, existing technologies have significant drawbacks. Traditional reactive maintenance methods, which only address damage to the wind turbine after it has occurred, require extensive preparation work based on the specific fault condition. This includes preparing necessary spare parts, tools, and personnel, leading to lengthy preparation times. Furthermore, the turbine's inability to generate electricity while awaiting maintenance results in substantial downtime losses. As for maintenance methods incorporating lifespan prediction, environmental events such as typhoons and abnormally high temperatures significantly alter the operating conditions and stress states of components, accelerating aging and damage. This results in inaccurate predictions and prevents timely maintenance. Summary of the Invention
[0003] To address the shortcomings of existing technologies, this application provides a wind turbine maintenance method, electronic equipment, and medium based on intelligent robots. By determining the maintenance time points of components based on the impact types and impact coefficients of all environmental impact events and multiple damage resistance value curves, the final maintenance time point of the wind turbine is determined based on the maintenance time points of all components of the wind turbine. This allows for the determination of the final maintenance time point for environmental impact events that the wind turbine will encounter, thereby enabling maintenance preparations to be carried out in advance before the wind turbine is damaged, and timely maintenance of the wind turbine to reduce downtime losses.
[0004] To address the above problems, the present invention provides the following technical solution: In a first aspect, embodiments of this application provide a wind turbine maintenance method based on an intelligent robot, comprising: acquiring usage status data of multiple components of the wind turbine collected by the intelligent robot; For each component, multiple damage resistance value curves are calculated based on the usage status data corresponding to the component, wherein each damage resistance value curve represents a type of damage resistance. Acquire environmental data within a preset area, including the wind turbine generators; Identify all environmental impact events within the forecast period based on environmental data; The impact type of each environmental impact event is determined based on the pre-defined correspondence information between environmental impact events and impact types. The impact coefficient of each environmental impact event is determined based on the pre-defined correspondence information between environmental impact events and their impact coefficients. The maintenance time point of the component is determined based on the impact type and impact coefficient of all the aforementioned environmental impact events and the multiple damage resistance value curves; The final maintenance time of the wind turbine is determined based on the maintenance time of all components of the wind turbine. The maintenance robot is controlled to perform maintenance on the wind turbine at the final maintenance time.
[0005] In some implementations, calculating multiple damage resistance curves for each component based on its corresponding usage status data includes: For each component, the current wear status parameter of the component is determined based on the usage status data corresponding to the component; Based on the current loss status parameters of the component and the loss model corresponding to the component, multiple damage resistance value curves of the component are calculated.
[0006] In some implementations, calculating multiple damage resistance curves of the component based on the component's current loss state parameters and the corresponding loss model includes: The calculation method for determining the multiple damage resistance value curves corresponding to the component in the loss model based on the current loss state parameters; Based on the calculation method of the multiple damage resistance value curves corresponding to the component, multiple damage resistance value curves are calculated respectively.
[0007] In some embodiments, the multiple damage resistance value curves include a first damage resistance value curve and a second damage resistance value curve; the impact types include cumulative impact types and sudden impact types; and determining the maintenance time point of the component based on the impact types and impact coefficients of all the environmental impact events and the multiple damage resistance value curves includes: When the impact type of all the environmental impact events is not a sudden impact type, the maintenance time point of the component is determined based on the impact coefficient of all the environmental impact events and the first damage resistance value curve of the component; When at least one of the environmental impact events is of the sudden impact type, the maintenance time point of the component is determined based on the second damage resistance value curve, the impact coefficients of all environmental impact events of the sudden impact type, and the predicted time point.
[0008] In some implementations, when the impact type of all the environmental impact events is not a sudden impact type, determining the maintenance time point of the component based on the impact coefficients of all the environmental impact events and the first damage resistance value curve of the component includes: When the impact type of all the environmental impact events is not a sudden impact type, for each environmental impact event, calculate the cumulative impact value of all environmental impact events whose prediction time point is not later than the prediction time point of the environmental impact event, and then obtain multiple cumulative impact values. According to the order of the prediction time points corresponding to each of the plurality of cumulative impact values, it is determined in turn whether each of the cumulative impact values exceeds the first damage resistance value of the first damage resistance value curve at the prediction time point corresponding to the cumulative impact value. When the cumulative impact value exceeds the first damage resistance value of the first damage resistance value curve at the predicted time point corresponding to the cumulative impact value, the predicted time point corresponding to the cumulative impact value is determined as the maintenance time point of the component, and the determination of whether the subsequent cumulative impact value exceeds the first damage resistance value of the first damage resistance value curve at the predicted time point corresponding to the cumulative impact value is stopped.
[0009] In some implementations, when at least one of the environmental impact events is of the sudden impact type, determining the maintenance time point of the component based on the second damage resistance value curve, the impact coefficients of all environmental impact events of the sudden impact type, and the predicted time point includes: When at least one of the environmental impact events is of the sudden impact type, the impact coefficient of each environmental impact event is determined sequentially according to the order of the predicted time points of all environmental impact events of the sudden impact type. This determination is made as to whether the impact coefficient of each environmental impact event exceeds the second damage resistance value corresponding to the second damage resistance value curve at the predicted time point of the environmental impact event. When the impact coefficient of the environmental impact event exceeds the second damage resistance value corresponding to the second damage resistance value curve at the predicted time point of the environmental impact event, the predicted time point of the environmental impact event is determined as the maintenance time point of the component, and the determination of whether the impact coefficient of subsequent environmental impact events exceeds the second damage resistance value corresponding to the predicted time point of the environmental impact event is stopped.
[0010] In some implementations, determining the final maintenance time of the wind turbine based on the maintenance time points of all components of the wind turbine includes: A cost model for the wind turbine within a maintenance cycle is constructed based on the unit time maintenance cost, the maintenance material cost and logistics cost corresponding to each component. The availability of the wind turbine corresponding to each maintenance time point is calculated based on the types of all components expected to require maintenance at each maintenance time point; Based on the objective function of the cost model, the final maintenance time point is calculated by determining all components expected to require maintenance at each maintenance time point and the availability corresponding to each maintenance time point, such that the objective function is minimized and the availability is greater than a preset availability threshold.
[0011] In some implementations, calculating the availability of the wind turbine corresponding to each maintenance time point based on the types of all components expected to require maintenance at each maintenance time point includes: The availability of the wind turbine at each maintenance time point is calculated based on the operational dependency information of all components of the wind turbine and the types of all components expected to require maintenance at each maintenance time point. The objective function based on the cost model, the total number of components expected to require repair at each repair time point, and the availability calculation corresponding to each repair time point are used to calculate the final repair time point that minimizes the objective function and ensures that the availability is greater than a preset availability threshold. This includes: The cost model is used to calculate the total maintenance cost for each maintenance time point based on the maintenance costs and maintenance time of all components expected to require maintenance at each maintenance time point; The profit generated by the wind turbine during a preset time period after each maintenance time point is calculated based on the availability of the wind turbine using a preset calculation method. Calculate multiple values of the objective function based on the total maintenance cost and the profit generated from power generation corresponding to each maintenance time point; The final maintenance time point is determined by selecting multiple values of the objective function to minimize the objective function and ensure that the availability is greater than a preset availability threshold.
[0012] Secondly, embodiments of this application provide an electronic device, the electronic device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the intelligent robot-based wind turbine maintenance method as described in the first aspect.
[0013] Thirdly, embodiments of this application provide a computer-readable storage medium storing an executable program, which is executed by a processor to implement the wind turbine maintenance method based on an intelligent robot as described in the first aspect.
[0014] This application provides a wind turbine maintenance method, electronic device, and medium based on intelligent robots. This application determines the maintenance time point of components based on the impact type and impact coefficient of all environmental impact events and multiple damage resistance value curves. Based on the maintenance time points of all components of the wind turbine, the final maintenance time point of the wind turbine is determined. This can determine the final maintenance time point for the environmental impact events that the wind turbine will encounter, thereby enabling maintenance preparation work to be carried out in advance before the wind turbine is damaged, and timely maintenance of the wind turbine to reduce downtime losses. Attached Figure Description
[0015] Figure 1 This is a flowchart illustrating the wind turbine maintenance method based on an intelligent robot provided in an embodiment of this application.
[0016] Figure 2 This is a schematic diagram of multiple damage resistance value curves provided in the embodiments of this application.
[0017] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.
[0018] Figure 4 This is a structural block diagram of a computer-readable storage medium provided in an embodiment of this application. Detailed Implementation
[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.
[0020] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "multiple" means two or more, unless otherwise explicitly specified.
[0021] This application provides a wind turbine maintenance method, electronic device, and medium based on intelligent robots. By determining the maintenance time point of components based on the impact type and impact coefficient of all environmental impact events and multiple damage resistance value curves, the final maintenance time point of the wind turbine is determined based on the maintenance time points of all components of the wind turbine. This allows for the determination of the final maintenance time point for environmental impact events that the wind turbine will encounter, thereby enabling maintenance preparations to be carried out in advance before the wind turbine is damaged, and timely maintenance of the wind turbine to reduce downtime losses.
[0022] This application relates to fault prediction and health management of wind turbine units, which falls under the field of new energy.
[0023] The following section will describe in detail, with reference to the accompanying drawings, the wind turbine maintenance method based on intelligent robots provided in this application.
[0024] Please see Figure 1 , Figure 1 This is a flowchart illustrating the wind turbine maintenance method based on an intelligent robot provided in an embodiment of this application. Figure 1 As shown, the wind turbine maintenance method based on intelligent robots includes steps S100 to S900.
[0025] Step S100: Obtain usage status data of multiple components of the wind turbine collected by the intelligent robot.
[0026] In some implementations, the wind turbine includes multiple components such as a base, generator, rotor, drive system, pitch system, and yaw system. The drive system may include a main shaft and gearbox. The pitch system may include a pitch motor, pitch bearings, and a controller. The yaw system may include a motor and gears.
[0027] In some implementations, the status data used may include vibration parameter data of the component, historical temperature data, wear data, and runtime data.
[0028] In some implementations, a control robot periodically or irregularly collects usage status data of multiple components of the wind turbine. Whenever usage status data of multiple components of the wind turbine is collected, steps S200 to S900 are executed to determine or update the final maintenance time of the wind turbine.
[0029] Step S200: For each component, calculate multiple damage resistance value curves for the component based on the corresponding usage status data.
[0030] Each damage resistance value curve represents a type of damage resistance.
[0031] In some implementations, wind turbine components can possess different types of damage resistance. For example, components of offshore wind turbines typically have anti-corrosion coatings, providing high resistance to salt corrosion. However, their resistance is lower in extreme weather conditions such as earthquakes or thunderstorms. Therefore, multiple damage resistance curves can be calculated for components based on their corresponding usage status data.
[0032] In some implementations, the multiple damage resistance value curves include a first damage resistance value curve and a second damage resistance value curve.
[0033] Optionally, the first damage resistance value curve is used to represent the relationship between the wind turbine's ability to withstand cumulative impact type environmental impact events and time, and the second damage resistance value curve is used to represent the relationship between the wind turbine's ability to withstand sudden impact type damage events and time.
[0034] Optionally, the multiple damage resistance value curves may also include other damage resistance value curves.
[0035] Please see Figure 2 , Figure 2 This is a schematic diagram of multiple damage resistance value curves provided in the embodiments of this application. For example... Figure 2 As shown, in some embodiments, the multiple damage resistance value curves include a first damage resistance value curve L1, a second damage resistance value curve L2, and a third damage resistance value curve L3. Figure 2 The horizontal axis T represents time, and the vertical axis Y represents the damage resistance value.
[0036] Optionally, the third damage resistance value curve L3 is used to represent the relationship between the wind turbine's resistance to damage during normal use and time.
[0037] In some implementations, step S200 includes steps S210 to S220.
[0038] Step S210: For each component, determine the current loss status parameters of the component based on the usage status data corresponding to the component.
[0039] In some implementations, the current wear status parameters of the component are calculated based on the component's vibration parameter data, historical temperature data, wear data, and runtime data in the usage status data.
[0040] In some implementations, different components correspond to different calculation methods for the current loss state parameters. The current loss state parameters of a component can be calculated by selecting the calculation method corresponding to the current loss state parameters of the component.
[0041] All parameters used in the calculations in this application have been normalized. For example, the Min-Max normalization algorithm is used to normalize all parameters used in the calculations.
[0042] For example, the formula for calculating the current loss state parameters of the wind turbine is: , in, This indicates the current loss status parameters of the wind turbine. , , , These represent different preset calculation coefficients. Indicates the vibration parameters of the wind turbine. This represents the maximum critical value of the vibration parameter. This represents the historical average temperature of the wind turbine. This indicates the rated temperature of the wind turbine. Indicates the extreme temperature of the wind turbine. This indicates the amount of wear on the wind turbine. This represents the maximum critical value for wear. Indicates runtime. This indicates the design life of the wind turbine.
[0043] Optionally, , , and The current loss state parameters of the wind turbine can be determined by training a formula based on historical data.
[0044] Step S220: Calculate multiple damage resistance value curves of the component based on the current loss state parameters of the component and the loss model corresponding to the component.
[0045] In some implementations, step S220 includes steps S221 to S222.
[0046] Step S221: Determine the calculation method for the multiple damage resistance value curves corresponding to the component in the current loss model based on the current loss state parameters.
[0047] In some implementations, when a component is in different current wear states, its subsequent normal wear rate is different, and its resistance to damage of various types is also different. Therefore, it is necessary to determine the calculation method of multiple resistance to damage value curves based on the component's current wear state parameters and the corresponding wear model of the component.
[0048] Optionally, the loss model corresponding to the component includes the calculation formulas for multiple damage resistance value curves corresponding to the component, and the values of all calculated coefficients in each calculation formula under the condition of the numerical range of each current loss state parameter.
[0049] Optionally, when the current loss state parameter of the component is greater than the first preset threshold corresponding to the component, each calculation coefficient in the calculation method of multiple damage resistance value curves is determined to be the value corresponding to the first value range.
[0050] Optionally, when the current loss state parameter of the component is not greater than the first preset threshold corresponding to the component, but is greater than the second preset threshold corresponding to the component, each calculation coefficient in the calculation method of multiple damage resistance capability value curves is determined to be the value corresponding to the second numerical range. The first preset threshold is greater than the second preset threshold.
[0051] Optionally, when the current loss state parameter of the component is not greater than the second preset threshold corresponding to the component, but is greater than the third preset threshold corresponding to the component, each calculation coefficient in the calculation method of multiple damage resistance capability value curves is determined to be the value corresponding to the third numerical interval. The second preset threshold is greater than the third preset threshold.
[0052] By using the above methods, different damage resistance value curves can be determined for different types of damage resistance under different wear conditions, thereby improving the accuracy of maintenance timing.
[0053] Step S222: Calculate multiple damage resistance value curves based on the calculation method of the multiple damage resistance value curves corresponding to the component.
[0054] As described above, the first damage resistance value curve represents the relationship between the wind turbine's ability to withstand cumulative environmental impact events and time; the second damage resistance value curve represents the relationship between the wind turbine's ability to withstand sudden impact events and time; and the third damage resistance value curve represents the relationship between the wind turbine's ability to withstand damage from normal use and time. It is understandable that, based on actual application scenarios, the second damage resistance value curve decays the fastest, the first damage resistance value curve decays slower than the second, and the third damage resistance value curve decays the slowest.
[0055] Optionally, a pre-trained loss model corresponding to each component is used to determine multiple damage resistance value curves based on the current loss state parameters of the component.
[0056] Optionally, the loss model can be an artificial intelligence model. For example, the loss model can be a neural network model or a Transformer model.
[0057] In some implementations, the calculation method corresponding to the first damage resistance value curve is based on logarithmic decay.
[0058] Optionally, the calculation method corresponding to the first damage resistance value curve is as follows: , in, This represents the curve showing the first damage resistance value. This represents the preset first calculation coefficient. This represents the preset second calculation coefficient. This indicates the current loss status parameter corresponding to the component. Indicates the current time. Indicates no earlier than At any time, is the independent variable. This represents a cumulative environmental impact function. This indicates the preset third calculation coefficient. , and All are greater than 0. , and The value of makes It is never less than 0.
[0059] Optionally, the loss model corresponding to the component also includes a calculation formula for the cumulative environmental impact function under the condition of the numerical range of each current loss state parameter.
[0060] For example, the formula for calculating the cumulative environmental impact function is obtained by fitting a big data analysis of the historical usage data of the component.
[0061] It is understandable that the formula for calculating the environmental impact function is a theoretical formula, and it is unrelated to the impact coefficient of actual environmental impact events or whether they occur.
[0062] Optionally, the specific calculation formula for the cumulative environmental impact function is determined based on the preset loss model corresponding to the component and the current loss state parameters of the component, and the method is the same as the description of determining the calculation coefficients above.
[0063] In some implementations, the second damage resistance value curve is calculated using an exponential decay-based method.
[0064] Optionally, the calculation method corresponding to the second damage resistance value curve is as follows: , in, This represents the value of the first damage resistance curve. This represents the preset fourth calculation coefficient. This represents the preset fifth calculation coefficient. This represents the preset sixth calculation coefficient. This represents a function representing sudden environmental impacts. , and All are greater than 0.
[0065] Optionally, the specific calculation formula for the sudden environmental impact function is determined based on the preset loss model corresponding to the component and the current loss state parameters of the component, and the method is described above.
[0066] For example, the formula for calculating the sudden environmental impact function is obtained by fitting the historical usage data of the component through big data analysis.
[0067] In some implementations, the calculation method corresponding to the third damage resistance value curve is based on the inverse proportional function decay calculation method.
[0068] Optionally, the calculation method corresponding to the third damage resistance value curve is as follows: , in, This represents the value of the third damage resistance curve. This represents the preset seventh calculation coefficient. This represents the loss function of a component under normal operating conditions.
[0069] Optionally, the specific calculation formula of the loss function can be determined based on the preset loss model corresponding to the component and the current loss state parameters of the component, and the method is described above.
[0070] For example, the formula for calculating the loss function is obtained by fitting a big data analysis of the historical usage data of the component.
[0071] Step S300: Obtain environmental data within a preset area including the wind turbine.
[0072] Step S400: Identify all environmental impact events within the forecast period based on environmental data.
[0073] In some implementations, environmental data include temperature, humidity, wind speed, rainfall, sunshine duration, and radiation intensity.
[0074] In some implementations, environmental data is uploaded to a weather forecasting center, and then all environmental impact events within the forecast period are obtained from the weather forecasting center.
[0075] In some implementations, a pre-trained artificial intelligence model is used to identify all environmental impact events within the predicted time period based on environmental data.
[0076] For example, environmental impact events include tides, light rain, earthquakes, typhoons, electromagnetic radiation, strong gusts, high temperatures, and sea fog.
[0077] For example, the forecast period includes the current time point and 1 day, 3 days, 7 days or 30 days after the current time point.
[0078] Step S500: Determine the impact type of each environmental impact event based on the preset correspondence information between environmental impact events and impact types.
[0079] In some implementations, the impact type includes cumulative impact type and sudden impact type.
[0080] For example, the impact types of tides and light rain are both cumulative impact types, while the impact types of earthquakes, typhoons, and strong gusts are all sudden impact types.
[0081] Optionally, the impact types also include conventional impact types.
[0082] For example, the impact type of sea fog is a conventional impact type.
[0083] Optionally, the types of influence may also include other types.
[0084] Step S600: Determine the impact coefficient of each environmental impact event based on the preset correspondence information between environmental impact events and impact coefficients.
[0085] In some implementations, the impact coefficient of each environmental impact event is determined at once based on a preset correspondence between environmental impact events and impact coefficients.
[0086] In other implementations, an initial impact coefficient for each environmental impact event is determined based on a pre-defined correspondence between environmental impact events and their impact coefficients. The probability of occurrence for each environmental impact event is then obtained, and the final impact coefficient is obtained by multiplying the initial impact coefficient by the corresponding probability. In this way, the impact coefficient reflects the probability of occurrence of environmental impact events, making the calculation process more accurate.
[0087] Step S700: Determine the maintenance time point for the component based on the impact type and impact coefficient of all environmental impact events and multiple damage resistance value curves.
[0088] In some implementations, step S700 includes steps S710 to S720.
[0089] Step S710: When the impact type of all environmental impact events is not a sudden impact type, determine the maintenance time point of the component based on the impact coefficient of all environmental impact events and the first damage resistance value curve of the component.
[0090] In some implementations, step S710 includes steps S711 to S713.
[0091] Step S711: When the impact type of all environmental impact events is not a sudden impact type, for each environmental impact event, calculate the cumulative impact value of all environmental impact events whose predicted time point is not later than the predicted time point of the environmental impact event, and then obtain multiple cumulative impact values.
[0092] Each prediction time point corresponds to a cumulative impact value.
[0093] In some implementations, the maintenance time point of a component is determined based on the impact coefficients of all environmental impact events of the cumulative impact type and the component's first damage resistance value curve.
[0094] In some implementations, the cumulative impact value is the sum of the impact coefficients of all environmental impact events whose prediction time is no later than the prediction time of an environmental impact event.
[0095] In some implementations, for the current predicted time point of an environmental impact event, the further the predicted time point is from the current predicted time point, the greater the impact coefficient of the environmental impact event on the calculated cumulative impact value. This is because when a wind turbine is damaged due to an environmental impact event, the degree of damage generally increases during subsequent operation.
[0096] In some implementations, a time-impact function is used to calculate the cumulative impact value of all environmental impact events whose predicted time points are no later than the predicted time point of the environmental impact event. This method allows for accurate calculation of the subsequent impact of environmental impact events.
[0097] Optionally, the formula for calculating the cumulative impact value is: , in, Indicates the first The cumulative impact value corresponding to the predicted time point of an environmental impact event. This means that all predicted time points are no later than the [number]th [time point]. The total number of environmental impact events at the predicted time point for each environmental impact event. Indicates the first Predicted time points for environmental impact events Indicates the first Predicted time points for environmental impact events Indicates the first The impact coefficient of an environmental impact event This represents the time-effect function. This indicates the preset time interval.
[0098] Using the above method, the shorter the prediction time interval between two adjacent environmental impact events, the greater the cumulative impact value, reflecting the characteristics of continuous impact.
[0099] Optionally, the formula for calculating the time influence function is: , in, This indicates the calculation coefficients that are affected by the preset time.
[0100] Optionally, the formula for calculating the time influence function is: , in, This indicates the preset time interval.
[0101] Using the above method, the shorter the prediction time interval between two adjacent environmental impact events, the greater the cumulative impact value, reflecting the characteristics of continuous impact.
[0102] Step S712: According to the order of the prediction time points corresponding to each cumulative impact value among the multiple cumulative impact values, determine in turn whether each cumulative impact value exceeds the first damage resistance value of the first damage resistance value curve at the prediction time point corresponding to the cumulative impact value.
[0103] Step S713: When the cumulative impact value exceeds the first damage resistance value of the first damage resistance value curve at the predicted time point corresponding to the cumulative impact value, the predicted time point corresponding to the cumulative impact value is determined as the maintenance time point of the component, and the determination of whether the subsequent cumulative impact value exceeds the first damage resistance value of the first damage resistance value curve at the predicted time point corresponding to the cumulative impact value is stopped.
[0104] For example, such as Figure 2 As shown, when there are multiple prediction time points, including a first prediction time point T1 and a second prediction time point T2, and the second prediction time point T2 is earlier than the first prediction time point T1, it is first determined whether the cumulative impact value corresponding to the second prediction time point T2 exceeds the first damage resistance value Y2 of the first damage resistance value curve L1 at the second prediction time point T2. If it exceeds, the second prediction time point T2 is determined as the maintenance time point of the component, and no further calculation is needed. If it does not exceed, it is determined whether the cumulative impact value corresponding to the first prediction time point T1 exceeds the first damage resistance value Y1 of the first damage resistance value curve L1 at the first prediction time point T1. If it exceeds, the first prediction time point T1 is determined as the maintenance time point of the component.
[0105] Step S720: When there is at least one environmental impact event of the sudden impact type, determine the maintenance time point of the component based on the second damage resistance value curve, the impact coefficients of all environmental impact events of the sudden impact type and the predicted time point.
[0106] In some implementations, step S720 includes steps S721 to S722.
[0107] Step S721: When there is at least one environmental impact event whose impact type is sudden impact type, according to the order of the predicted time points of all environmental impact events of sudden impact type, determine in turn whether the impact coefficient of each environmental impact event exceeds the second damage resistance value corresponding to the second damage resistance value curve at the predicted time point of the environmental impact event.
[0108] Step S722: When the impact coefficient of an environmental impact event exceeds the second damage resistance value corresponding to the predicted time point of the environmental impact event on the second damage resistance value curve, the predicted time point of the environmental impact event is determined as the maintenance time point of the component, and the determination of whether the impact coefficient of subsequent environmental impact events exceeds the second damage resistance value corresponding to the predicted time point of the environmental impact event on the second damage resistance value curve is stopped.
[0109] In some implementations, when the impact coefficient of an environmental impact event of the sudden impact type exceeds the second damage resistance value corresponding to the second damage resistance value curve at the predicted time point of the environmental impact event, the predicted time point of the environmental impact event is determined as the maintenance time point of the component. In this way, the maintenance time point can be accurately determined when an environmental impact event of the sudden impact type occurs. Furthermore, since there is still some time between the current time and the predicted maintenance time point, maintenance preparations can be carried out in advance.
[0110] In some implementations, when there is no cumulative impact value exceeding the first damage resistance value curve at the predicted time point corresponding to the cumulative impact value, and / or when there is no impact coefficient of the environmental impact event exceeding the second damage resistance value curve at the predicted time point corresponding to the environmental impact event, it is determined that the component does not need to be repaired.
[0111] In some implementations, when there is no cumulative impact value exceeding the first damage resistance value of the first damage resistance value curve at the predicted time point corresponding to the cumulative impact value, and / or when there is no impact coefficient of the environmental impact event exceeding the second damage resistance value of the second damage resistance value curve at the predicted time point corresponding to the environmental impact event, the component is determined to be in normal wear and tear. The time point when the third damage resistance value is lower than the preset damage threshold corresponding to the component is calculated based on the third damage resistance value curve, and this time point is determined as the maintenance time point of the component.
[0112] In some implementations, step S710 can be performed when all environmental impact events are of the conventional impact type.
[0113] In some other implementations, when all environmental impact events are of the conventional impact type, step 700 further includes step S730.
[0114] Step S730: Determine the maintenance time point of the component based on the third damage resistance value curve, the impact coefficients of all environmental impact events of the conventional impact type, and the predicted time point.
[0115] In some implementations, step S730 includes steps S731 to S733.
[0116] Step S731: For each environmental impact event, calculate the conventional impact value of all environmental impact events whose prediction time point is no later than the prediction time point of the environmental impact event, and thus obtain multiple conventional impact values.
[0117] Each prediction time point corresponds to a conventional impact value.
[0118] In some implementations, the conventional impact value is the positive square root of the sum of the impact coefficients of all environmental impact events with a prediction time no later than the prediction time of an environmental impact event, or the result of taking the natural logarithm of the sum of the impact coefficients.
[0119] Step S732: According to the order of the predicted time points corresponding to each of the multiple conventional influence values, determine in turn whether the sum of each conventional influence value and the preset damage threshold corresponding to the component exceeds the third damage resistance value of the third damage resistance value curve at the predicted time point corresponding to the conventional influence value.
[0120] In this way, it is possible to determine the repair time point when the component's third resistance to damage falls below the damage threshold under environmental influences.
[0121] Step S733: When the sum of the conventional impact value and the preset damage threshold corresponding to the component exceeds the third damage resistance value of the third damage resistance value curve at the predicted time point corresponding to the conventional impact value, the predicted time point corresponding to the conventional impact value is determined as the maintenance time point of the component, and the determination of whether the subsequent conventional impact value exceeds the third damage resistance value of the third damage resistance value curve at the predicted time point corresponding to the conventional impact value is stopped.
[0122] This application does not limit the number of damage resistance value curves or the number of environmental impact event impact types. A damage resistance value curve can be selected based on the impact coefficients of multiple types of environmental impact events to calculate and determine the maintenance time point.
[0123] Step S800: Determine the final maintenance time of the wind turbine based on the maintenance time points of all components of the wind turbine.
[0124] In some implementations, step S800 includes steps S810 to S830.
[0125] Step S810: Construct a cost model for the wind turbine within a maintenance cycle based on the unit time maintenance cost, the maintenance material cost for each component, and the logistics cost.
[0126] In some implementations, the cost model is calculated using the following formula: , in, This represents the total maintenance cost of a wind turbine within a maintenance cycle. This indicates the set of parts that need repair. Indicates the first The cost of repair materials for each component. Indicates the first Logistics costs for each component Indicates the first Repair time corresponding to each component This indicates the cost of maintenance per unit of time.
[0127] Step S820: Calculate the availability of the wind turbine corresponding to each maintenance time point based on the types of all components expected to require maintenance at each maintenance time point.
[0128] In some implementations, the availability of the wind turbine at a maintenance time point is calculated based on the operational dependency information of all components of the wind turbine and the types of all components expected to require maintenance at each maintenance time point.
[0129] In some implementations, step S820 includes steps S821 to S822.
[0130] Step S821: Based on the operational dependency information of all components of the wind turbine and the types of all components expected to require maintenance at each maintenance time point, determine the probability of normal operation of each component at each maintenance time point.
[0131] In some implementations, the operational dependency information includes the probability that each component will function normally when it needs maintenance, and the probability that other components will function normally when one or more components fail.
[0132] Optionally, work dependency information can be determined in advance based on historical maintenance data.
[0133] In some implementations, the probability of each component operating normally at each maintenance time point is determined based on the operational dependency information of all components of the wind turbine and the types of all components expected to require maintenance at each maintenance time point.
[0134] Optionally, based on the operational dependency information of all components of the wind turbine, the probability of a component requiring maintenance operating normally is determined as the probability of normal operation for that component. For example, when the wind turbine needs maintenance, the probability of normal operation for that wind turbine is 15%.
[0135] For example, when the wind turbine needs maintenance, the probability of the generator operating normally is 5%. When the wind turbine base needs maintenance, the probability of the wind turbine operating normally is 70%.
[0136] Step S822: Calculate the availability of the wind turbine corresponding to each maintenance time point based on the probability that each component is working normally at each maintenance time point.
[0137] In some implementations, the formula for calculating the availability of a wind turbine at a maintenance point in time is: , in, Indicates the availability of wind turbine units. This indicates the number of all components that need to be repaired at that specific maintenance point in time. Indicates the time of maintenance. The probability of each component working properly.
[0138] The calculation formula is based on the assumption that components work independently. It assumes that the overall availability of the wind turbine is conditional on all components that need maintenance being able to work normally. Therefore, availability is equal to the product of the probabilities of each component working normally.
[0139] In some implementations, considering the complex dependencies between components, a Bayesian network can be constructed based on the operational dependency information of all components of the wind turbine, and the availability of the wind turbine can be calculated using a method that calculates through joint probability distribution.
[0140] Step S830: Based on the objective function of the cost model, the total number of parts expected to be repaired at each repair time point, and the availability corresponding to each repair time point, calculate the final repair time point that minimizes the objective function and has an availability greater than the preset availability threshold.
[0141] In some implementations, step S830 includes steps S831 to S834.
[0142] Step S831: Using a cost model, calculate the total maintenance cost for each maintenance point based on the maintenance costs and maintenance time corresponding to all components expected to require maintenance at each maintenance point.
[0143] Step S832: Calculate the profit generated by the wind turbine during a preset time period after each maintenance time point, based on the availability of the wind turbine and using a preset calculation method.
[0144] In some implementations, step S832 includes steps S8321 to S8323.
[0145] Step S8321: Calculate the power generation function of the wind turbine in a preset time period after each maintenance time point, based on the availability of the wind turbine.
[0146] In some implementations, the availability of the wind turbine is multiplied by the rated power and the power time variation function of the wind turbine to obtain the power generation function of the wind turbine in a preset time period after each maintenance time point without maintenance.
[0147] Optionally, the power time-varying function differs depending on availability. A range of availability values can correspond to a single power time-varying function.
[0148] Optionally, a pre-trained artificial intelligence model can be used to determine the power time variation function based on availability.
[0149] Step S8322: Obtain the predicted electricity price change function.
[0150] Step S8323: Calculate the integral of the function obtained by multiplying the power generation function and the electricity price change function over a preset time period to obtain the profit generated by the wind turbine in the preset time period after each maintenance time point without maintenance.
[0151] Optionally, the preset time period is 3 days, 5 days, or 10 days, etc.
[0152] Step S833: Calculate multiple values of the objective function based on the total maintenance cost and the profit generated from power generation at each maintenance time point.
[0153] In some implementations, the formula for calculating the objective function corresponding to a maintenance time point is as follows: , in, The time value factor representing the value of money within a preset time period. This represents the profit generated from electricity generation within a preset time period following the current maintenance time.
[0154] Optionally, These are preset values or values calculated by a prediction model.
[0155] Step S834: The final maintenance time point based on the selection of multiple values of the objective function that minimizes the objective function and makes the availability greater than the preset availability threshold.
[0156] Step S900: Control the intelligent maintenance robot to perform maintenance on the wind turbine at the final maintenance time.
[0157] In some implementations, intelligent maintenance robots include drones used for maintenance.
[0158] Alternatively, the drone may be an amphibious drone.
[0159] In summary, the wind turbine maintenance method based on intelligent robots provided in this application has the following advantages: 1. By determining the maintenance time point of components based on the impact type and impact coefficient of all environmental impact events and multiple damage resistance value curves, and by determining the final maintenance time point of the wind turbine based on the maintenance time points of all components of the wind turbine, it is possible to determine the final maintenance time point for the environmental impact events that the wind turbine will encounter. This allows for advance preparation for maintenance work before the wind turbine is damaged, timely maintenance of the wind turbine, and reduction of downtime losses.
[0160] 2. By determining the calculation method of multiple damage resistance value curves corresponding to the component in the current damage model based on the current damage state parameters, different damage resistance value curves can be determined for different types of damage resistance under different damage conditions, thereby improving the accuracy of maintenance time points.
[0161] 3. By multiplying the initial impact coefficient of each environmental impact event by its corresponding probability of occurrence to obtain the final impact coefficient of the environmental impact event, the impact coefficient can reflect the probability of occurrence of the environmental impact event, thus making the calculation process more accurate.
[0162] 4. By using the time impact function to calculate the cumulative impact value of all environmental impact events whose predicted time points are no later than the predicted time point of the environmental impact event, the subsequent impact degree of the environmental impact event can be accurately calculated.
[0163] Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. For example... Figure 3 As shown, the electronic device 400 includes: one or more processors 410 and a memory 420. Figure 3 Take a processor 410 as an example.
[0164] In some implementations, the processor 410 and the memory 420 may be connected via a bus or other means. Figure 3 Taking the example of a connection between China and Israel via a bus.
[0165] In some implementations, the processor 410 is configured to acquire usage status data of multiple components of the wind turbine collected by an intelligent robot; for each component, calculate multiple damage resistance value curves based on the corresponding usage status data, wherein each damage resistance value curve represents a type of damage resistance; acquire environmental data within a preset area including the wind turbine; determine all environmental impact events within a predicted time period based on the environmental data; determine the impact type of each environmental impact event based on preset correspondence information between environmental impact events and impact types; determine the impact coefficient of each environmental impact event based on preset correspondence information between environmental impact events and impact coefficients; determine the maintenance time point of the component based on the impact types and impact coefficients of all environmental impact events and the multiple damage resistance value curves; determine the final maintenance time point of the wind turbine based on the maintenance time points of all components of the wind turbine; and control the intelligent maintenance robot to perform maintenance on the wind turbine at the final maintenance time point.
[0166] In some embodiments, memory 420 serves as a non-volatile computer-readable storage medium, used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules of the intelligent robot-based wind turbine maintenance method in the embodiments of this application. Processor 410 executes various functional applications and data processing of electronic device 400 by running the non-volatile software programs, instructions, and modules stored in memory 420, thereby implementing the intelligent robot-based wind turbine maintenance method described in the above embodiments.
[0167] In some embodiments, memory 420 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of electronic device 400, etc. Furthermore, memory 420 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 420 may optionally include memory remotely located relative to processor 410, and this remote memory may be connected to the controller via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.
[0168] In some implementations, one or more modules are stored in memory 420 and, when executed by one or more processors 410, perform the intelligent robot-based wind turbine maintenance method in any of the above method embodiments, for example, performing the above-described... Figure 1 The method steps S100 to S900.
[0169] Please refer to Figure 4 , Figure 4 This is a structural block diagram of a computer-readable storage medium provided in an embodiment of this application. The computer-readable storage medium 500 stores program code 510, which can be called by a processor to execute the wind turbine maintenance method based on an intelligent robot described in the above method embodiments.
[0170] The computer-readable storage medium 500 may be an electronic storage device such as flash memory, electrically erasable programmable read-only memory (EEPROM), hard disk, or read-only memory (ROM). Optionally, the computer-readable storage medium includes a non-volatile computer-readable medium. The computer-readable storage medium 500 has storage space for program code that performs any of the method steps of the above-described intelligent robot-based wind turbine maintenance method. This program code can be read from or written to one or more computer program products. The program code may, for example, be compressed in a suitable form.
[0171] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described intelligent robot-based wind turbine maintenance method.
[0172] In summary, this application provides a wind turbine maintenance method, electronic device, and medium based on an intelligent robot. The wind turbine maintenance method includes: acquiring usage status data of multiple components of the wind turbine collected by the intelligent robot; for each component, calculating multiple damage resistance value curves based on the corresponding usage status data, wherein each damage resistance value curve represents a type of damage resistance; acquiring environmental data within a preset area including the wind turbine; determining all environmental impact events within a predicted time period based on the environmental data; determining the impact type of each environmental impact event based on preset correspondence information between environmental impact events and impact types; determining the impact coefficient of each environmental impact event based on preset correspondence information between environmental impact events and impact coefficients; determining the maintenance time point of the components based on the impact types and impact coefficients of all environmental impact events and the multiple damage resistance value curves; determining the final maintenance time point of the wind turbine based on the maintenance time points of all components of the wind turbine; and controlling the intelligent maintenance robot to perform maintenance on the wind turbine at the final maintenance time point. This application determines the maintenance time point of components based on the impact type and impact coefficient of all environmental impact events and multiple damage resistance value curves. Based on the maintenance time points of all components of the wind turbine, it determines the final maintenance time point of the wind turbine. This allows for the determination of the final maintenance time point for environmental impact events that the wind turbine will encounter, thereby enabling maintenance preparations to be carried out in advance before the wind turbine is damaged, and timely maintenance of the wind turbine to reduce downtime losses.
[0173] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.
Claims
1. A wind turbine maintenance method based on intelligent robots, characterized in that, include: Acquire usage status data of multiple components of the wind turbine generator, collected by an intelligent robot; For each component, multiple damage resistance value curves are calculated based on the usage status data corresponding to the component, wherein each damage resistance value curve represents a type of damage resistance. Acquire environmental data within a preset area, including the wind turbine generators; Identify all environmental impact events within the forecast period based on environmental data; The impact type of each environmental impact event is determined based on the pre-defined correspondence information between environmental impact events and impact types. The impact coefficient of each environmental impact event is determined based on the pre-defined correspondence information between environmental impact events and their impact coefficients. The maintenance time point of the component is determined based on the impact type and impact coefficient of all the aforementioned environmental impact events and the multiple damage resistance value curves; The final maintenance time of the wind turbine is determined based on the maintenance time of all components of the wind turbine. The maintenance robot is controlled to perform maintenance on the wind turbine at the final maintenance time.
2. The wind turbine maintenance method based on intelligent robots according to claim 1, characterized in that, For each component, multiple damage resistance curves are calculated based on the corresponding usage status data of the component, including: For each component, the current wear status parameter of the component is determined based on the usage status data corresponding to the component; Based on the current loss status parameters of the component and the loss model corresponding to the component, multiple damage resistance value curves of the component are calculated.
3. The wind turbine maintenance method based on intelligent robots according to claim 2, characterized in that, The calculation of multiple damage resistance curves for the component based on its current loss state parameters and the corresponding loss model includes: The calculation method for determining the multiple damage resistance value curves corresponding to the component in the loss model based on the current loss state parameters; Based on the calculation method of the multiple damage resistance value curves corresponding to the component, multiple damage resistance value curves are calculated respectively.
4. The wind turbine maintenance method based on intelligent robots according to claim 1, characterized in that, The multiple damage resistance capability value curves include a first damage resistance capability value curve and a second damage resistance capability value curve. The impact types include cumulative impact types and sudden impact types. Determining the maintenance time point of the component based on the impact types and impact coefficients of all the environmental impact events and the multiple damage resistance capability value curves includes: When the impact type of all the environmental impact events is not a sudden impact type, the maintenance time point of the component is determined based on the impact coefficient of all the environmental impact events and the first damage resistance value curve of the component; When at least one of the environmental impact events is of the sudden impact type, the maintenance time point of the component is determined based on the second damage resistance value curve, the impact coefficients of all environmental impact events of the sudden impact type, and the predicted time point.
5. The wind turbine maintenance method based on intelligent robots according to claim 4, characterized in that, When the impact type of all the environmental impact events is not a sudden impact type, the maintenance time point of the component is determined based on the impact coefficients of all the environmental impact events and the first damage resistance value curve of the component, including: When the impact type of all the environmental impact events is not a sudden impact type, for each environmental impact event, calculate the cumulative impact value of all environmental impact events whose prediction time point is not later than the prediction time point of the environmental impact event, and then obtain multiple cumulative impact values. According to the order of the prediction time points corresponding to each of the plurality of cumulative impact values, it is determined in turn whether each of the cumulative impact values exceeds the first damage resistance value of the first damage resistance value curve at the prediction time point corresponding to the cumulative impact value. When the cumulative impact value exceeds the first damage resistance value of the first damage resistance value curve at the predicted time point corresponding to the cumulative impact value, the predicted time point corresponding to the cumulative impact value is determined as the maintenance time point of the component, and the determination of whether the subsequent cumulative impact value exceeds the first damage resistance value of the first damage resistance value curve at the predicted time point corresponding to the cumulative impact value is stopped.
6. The wind turbine maintenance method based on intelligent robots according to claim 4, characterized in that, When at least one of the environmental impact events is of the sudden impact type, the maintenance time point of the component is determined based on the second damage resistance value curve, the impact coefficients of all environmental impact events of the sudden impact type, and the predicted time point, including: When at least one of the environmental impact events is of the sudden impact type, the impact coefficient of each environmental impact event is determined sequentially according to the order of the predicted time points of all environmental impact events of the sudden impact type. This determination is made as to whether the impact coefficient of each environmental impact event exceeds the second damage resistance value corresponding to the second damage resistance value curve at the predicted time point of the environmental impact event. When the impact coefficient of the environmental impact event exceeds the second damage resistance value corresponding to the second damage resistance value curve at the predicted time point of the environmental impact event, the predicted time point of the environmental impact event is determined as the maintenance time point of the component, and the determination of whether the impact coefficient of subsequent environmental impact events exceeds the second damage resistance value corresponding to the predicted time point of the environmental impact event is stopped.
7. The wind turbine maintenance method based on intelligent robots according to claim 1, characterized in that, Determining the final maintenance time of the wind turbine based on the maintenance time points of all components of the wind turbine includes: A cost model for the wind turbine within a maintenance cycle is constructed based on the unit time maintenance cost, the maintenance material cost and logistics cost corresponding to each component. The availability of the wind turbine corresponding to each maintenance time point is calculated based on the types of all components expected to require maintenance at each maintenance time point; Based on the objective function of the cost model, the final maintenance time point is calculated by determining all components expected to require maintenance at each maintenance time point and the availability corresponding to each maintenance time point, such that the objective function is minimized and the availability is greater than a preset availability threshold.
8. The wind turbine maintenance method based on intelligent robots according to claim 7, characterized in that, The calculation of the availability of the wind turbine corresponding to each maintenance time point based on the types of all components expected to require maintenance at each maintenance time point includes: The availability of the wind turbine at each maintenance time point is calculated based on the operational dependency information of all components of the wind turbine and the types of all components expected to require maintenance at each maintenance time point. The objective function based on the cost model, the total number of components expected to require repair at each repair time point, and the availability calculation corresponding to each repair time point are used to calculate the final repair time point that minimizes the objective function and ensures that the availability is greater than a preset availability threshold. This includes: The cost model is used to calculate the total maintenance cost for each maintenance time point based on the maintenance costs and maintenance time of all components expected to require maintenance at each maintenance time point; The profit generated by the wind turbine during a preset time period after each maintenance time point is calculated based on the availability of the wind turbine using a preset calculation method. Calculate multiple values of the objective function based on the total maintenance cost and the profit generated from power generation corresponding to each maintenance time point; The final maintenance time point is determined by selecting multiple values of the objective function to minimize the objective function and ensure that the availability is greater than a preset availability threshold.
9. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the wind turbine maintenance method based on an intelligent robot as described in any one of claims 1 to 8.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an executable program, which is executed by a processor to implement the wind turbine maintenance method based on an intelligent robot as described in any one of claims 1 to 8.