Self-adaptive regulation and control method and equipment for rotating speed of fan, medium and program product

Abstract

The invention provides a fan rotating speed self-adaptive regulation and control scheme based on virtual back electromotive force and a safety margin factor. A virtual back electromotive force constant reference during stable operation of a fan is established, the dynamic disturbance quantity is calculated in combination with real-time monitoring data, a self-adaptive disturbance threshold value is determined through statistical analysis, and the disturbance degree is accurately quantified. Based on the safety margin factor obtained through interpolation calculation, the system can dynamically optimize PID control parameters, and the limitation of traditional fixed parameter control is reduced. According to the scheme, the technical problems that in the prior art, a draught fan control system cannot accurately recognize the disturbance degree and control parameters are fixed are effectively solved, and the control precision and operation stability of a draught fan in the complex disturbance environment are improved.

Description

Technical Field

[0001] This application relates to the field of wind turbine control technology, and in particular to a method, device, medium, and program product for adaptive speed control of a wind turbine. Background Technology

[0002] In industrial production and environmental control, fans are widely used as key equipment in various scenarios such as ventilation, exhaust, and cooling. With the increasing level of industrial automation, the requirements for precise control and stability of fan operation are also increasing. Especially in complex and variable working environments, the efficient and stable operation of fans is crucial to the performance and safety of the entire system.

[0003] In related technologies, wind turbine control systems typically employ PID control algorithms. These algorithms monitor wind turbine parameters such as speed and power in real time, compare these parameters with target values, and then calculate corresponding adjustment parameters to maintain stable wind turbine operation. When the wind turbine's operating environment changes, the control system automatically adjusts the wind turbine's operating parameters according to preset control logic to ensure that the wind turbine's speed remains consistent with the target speed, thereby meeting the system's operational requirements.

[0004] However, in actual working environments, fans are often subject to various external disturbances, such as inlet blockage, pipeline pressure fluctuations, and ambient temperature changes. The degree, speed, and characteristics (long-term impact) of these disturbances on fan speed vary, and they often exhibit dynamic changes. Existing PID control systems typically employ fixed control gain and response strategies when calculating adjustment parameters, failing to provide differentiated response adjustments based on the actual impact characteristics of the current disturbance. When facing slight, continuous disturbances, the system may apply excessive adjustment force, causing unnecessary fluctuations in fan speed; conversely, when facing severe disturbances with large impacts, the fixed adjustment strategy may result in insufficient response speed, failing to promptly adjust the fan speed to the target value. This makes it difficult for the fan to achieve good control accuracy and stability under different disturbance conditions. Summary of the Invention

[0005] This application provides a method, device, medium, and program product for adaptive speed control of a wind turbine, which can improve the control accuracy and stability of the wind turbine under different disturbance conditions.

[0006] Firstly, this application provides a method for adaptive speed control of a wind turbine. The method includes: when the target wind turbine is running stably, based on monitored stable operating parameters under multiple preset power levels, a preset correlation model is used for calibration to determine the virtual back-EMF constant between the back-EMF and the speed of the target wind turbine during stable operation; the real-time monitored operating parameters are input into the preset correlation model to obtain the real-time virtual back-EMF; the difference between the real-time virtual back-EMF and the virtual back-EMF constant is used as a dynamic disturbance; based on the mean and standard deviation of the dynamic disturbance within a preset time window, the mean is determined as the zero-influence disturbance threshold, and the sum of the mean and three times the standard deviation is determined as the maximum-influence disturbance threshold; by interpolating between the zero-influence disturbance threshold and the maximum-influence disturbance threshold, a safety margin factor corresponding to the dynamic disturbance at the current moment is obtained and output; a wind turbine adjustment strategy for the target wind turbine is selected based on the safety margin factor, and the weight parameters of the PID algorithm are adjusted based on the wind turbine adjustment strategy; the wind turbine adjustment parameters that ensure the speed of the target wind turbine matches the target speed are calculated using the dynamic disturbance and the PID algorithm.

[0007] In the above embodiments, by adopting the above technical solution, a benchmark characteristic of the wind turbine's stable operating state is established based on the virtual back EMF constant, and dynamic disturbances are quantified by the difference between the real-time virtual back EMF and this benchmark. This dynamic disturbance is statistically analyzed to generate an adaptive disturbance threshold, and then a continuous safety margin factor is calculated through interpolation. The safety margin factor directly guides the dynamic adjustment of PID control parameters, enabling the control strategy to smoothly transition according to the degree of disturbance. This adaptive control mechanism based on quantitative disturbance analysis achieves high precision and high stability in wind turbine speed control, while avoiding system oscillations caused by sudden changes in control parameters, thus improving the operational reliability of the wind turbine system under complex disturbance environments.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, the step of selecting the target wind turbine based on the safety margin factor and adjusting the weight parameters of the PID algorithm based on the wind turbine adjustment strategy specifically includes: performing modal mapping in a modal strategy library according to the preset numerical range of the safety margin factor to determine the target control mode. The modal strategy library contains at least three preset control modes, and the three preset control modes correspond to different PID parameter configuration sets. After extracting the corresponding target PID parameter configuration set according to the target control mode, the target PID parameter configuration set is corrected based on the relative position of the safety margin factor within the preset numerical range to obtain the weight parameters of the PID algorithm. This correction is calculated between the parameter configurations of adjacent modes using linear interpolation or quadratic interpolation algorithms.

[0009] In the above embodiments, by adopting the above technical solution, a mapping relationship is established between the safety margin factor and preset control modes, covering different operating scenarios through at least three preset control modes. Each mode corresponds to a specific set of PID parameter configurations, and an interpolation algorithm is used to achieve a smooth transition of parameters during mode switching. This mode-based hierarchical control structure enables the system to automatically select the most suitable control strategy according to changes in the safety margin factor and avoids parameter abrupt changes during strategy switching. Through the design of the modal strategy library, refined management and optimized configuration of control strategies are achieved, improving the adaptability of the control system to different types of disturbances, while ensuring the stability and reliability of the control process.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, after the step of using the difference between the real-time virtual back EMF and the virtual back EMF constant as the dynamic disturbance quantity, the method further includes: within a preset number of control cycles, traversing multiple virtual back EMF constants of all wind turbines in the target wind turbine array to which the target wind turbine belongs, and obtaining a reference disturbance value of the target wind turbine array based on statistical analysis; based on the distribution of multiple virtual back EMF constants relative to the reference disturbance value, determining wind turbines exceeding a preset dynamic discrimination threshold as high-load abnormal wind turbines, the preset dynamic discrimination threshold being determined based on the discrete value of the situation; for the cooperating wind turbines that are closest to the physical location of the high-load abnormal wind turbines and whose virtual back EMF constants are not greater than the reference disturbance value, proportionally increasing the target speed of the cooperating wind turbines.

[0011] In the above embodiments, by adopting the above technical solution, a unified evaluation benchmark based on the virtual back electromotive force constant is established at the wind turbine array level, and high-load abnormal wind turbines are identified through statistical analysis. The system can intelligently select wind turbines with similar physical locations and lighter loads as cooperative wind turbines, and share the load by increasing their target speed. This load balancing mechanism based on local cooperation effectively avoids the problem of a single wind turbine operating under overload due to local blockage, and realizes the overall optimized control of the wind turbine array. This solution not only improves the operating efficiency of the system, but also reduces the risk of equipment failure through load distribution, thereby improving the overall reliability and energy utilization efficiency of the wind turbine array.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, before the step of correcting the target PID parameter configuration set based on the relative position of the safety margin factor within a preset numerical range, the method further includes: performing a discrete Fourier transform on the disturbance time series of the dynamic disturbance quantity within a preset time window to obtain spectral data, which consists of a series of frequency points and the energy amplitude corresponding to each frequency point; calculating the sum of all energy amplitudes in the spectral data whose frequencies are lower than a preset gradual change cutoff frequency to obtain the total value of the gradual change disturbance energy, which is determined by a disturbance event caused by a gradual load change or a slow environmental drift; calculating the sum of all energy amplitudes in the spectral data whose frequencies are higher than a preset sudden change start frequency to obtain the total value of the sudden change disturbance energy, which is determined by a disturbance event caused by a sudden impact or a sudden load change; generating a disturbance characteristic correction coefficient based on the proportional relationship between the total value of the gradual change disturbance energy and the total value of the sudden change disturbance energy, and using the disturbance characteristic correction coefficient to proportionally adjust the weight parameters in the target PID parameter configuration set to obtain the adjusted weight parameters.

[0013] In the above embodiments, by employing the aforementioned technical solution, the dynamic disturbance quantity is analyzed spectrally using Discrete Fourier Transform to calculate the total energy of both slowly varying and abruptly changing disturbances. Based on the proportional relationship between the two disturbance energies, disturbance characteristic correction coefficients are generated, enabling precise adjustment of PID parameters. This frequency domain analysis-based control parameter optimization mechanism allows the system to accurately identify and distinguish different types of disturbances and adopt corresponding control strategies. Spectral analysis achieves precise quantification of disturbance characteristics, improving the control system's adaptability to disturbances of different frequencies and enhancing the accuracy and stability of wind turbine control.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, after the step of generating a disturbance characteristic correction coefficient based on the ratio of the total value of the gradually varying disturbance energy to the total value of the abrupt disturbance energy, the method further includes: if the proportion of the total value of the gradually varying disturbance energy in the total disturbance energy of the wind turbine exceeds a first preset diagnostic threshold, then generating first diagnostic information; if the proportion of the total value of the abrupt disturbance energy in the total disturbance energy of the wind turbine exceeds a second preset diagnostic threshold, then generating second diagnostic information, which is different from the first diagnostic information.

[0015] In the above embodiments, by adopting the aforementioned technical solution, a dual-threshold diagnostic mechanism is established based on the proportion of slowly varying and abruptly changing disturbance energy in the total energy, enabling the generation of targeted diagnostic information. This energy proportion-based diagnostic method achieves accurate identification and classification of different types of disturbances. Through differentiated diagnostic information output, the system can promptly alert to potential operational anomalies, providing a reliable basis for equipment maintenance and fault prevention, and improving the operational safety and maintainability of the wind turbine system. In some embodiments, in conjunction with the first aspect, the method further includes: if the proportion of the total value of the slowly varying disturbance energy in the total disturbance energy of the wind turbine exceeds a first preset diagnostic threshold, then performing linear regression on the disturbance time series to calculate the predicted disturbance amount for the next control cycle; according to a preset weighting coefficient, weighting the predicted disturbance amount and the dynamic disturbance amount of the current control cycle to obtain a comprehensive disturbance amount, and using the comprehensive disturbance amount to replace the dynamic disturbance amount.

[0016] In the above embodiments, by employing the aforementioned technical solution, upon detecting a gradually changing disturbance, the disturbance amount for the next control cycle is predicted through linear regression, and a weighted average is performed with the current disturbance amount to obtain the comprehensive disturbance amount. This prediction-based feedforward compensation mechanism improves the system's predictability and response speed to gradually changing disturbances. By applying the comprehensive disturbance amount, the control system achieves an early response to the gradual change process, effectively reducing control lag and improving the system's control performance against gradually changing disturbances.

[0017] In conjunction with some embodiments of the first aspect, in some embodiments, the method further includes: when a preset safety margin factor is lower than a preset minimum safety threshold, generating and outputting alarm information, wherein the alarm information is one of a first diagnostic information and a second diagnostic information.

[0018] In the above embodiments, by adopting the above technical solution, an early warning mechanism is established based on continuous monitoring of safety margin factors. When multiple safety margin factors fall below the minimum safety threshold, the system automatically generates alarm information of the corresponding type. This early warning mechanism based on multiple judgments improves the reliability of abnormal state identification. By timely outputting targeted alarm information, the system achieves early warning of potential risks, effectively preventing equipment failures and safety accidents, and improving the operational safety of the wind turbine system. One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. By adopting a disturbance quantization mechanism based on virtual back electromotive force and an adaptive control strategy driven by safety margin factor, the technical problem that the existing wind turbine control system cannot accurately identify the degree of disturbance and the control parameters remain unchanged is effectively solved. This enables high-precision adjustment and smooth transition of wind turbine speed control in complex disturbance environments, thereby improving the operational reliability and control accuracy of the wind turbine system.

[0019] 2. By adopting a multimodal mapping mechanism based on the safety margin factor and a smooth transition strategy between modes, the technical problems of single control strategy and easy system oscillation caused by mode switching in the existing technology are effectively solved. This enables refined management of control strategy and disturbance-free switching, and improves the stability and reliability of the control process.

[0020] 3. By adopting an array-level load assessment mechanism based on virtual back electromotive force constant and a local collaborative control strategy, the technical problems of low system efficiency caused by multiple wind turbine systems operating independently and local blockages in the existing technology are effectively solved. This achieves overall load balancing and energy efficiency optimization of the wind turbine array, improving the overall operating efficiency and reliability of the system. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating a method for adaptive speed control of a fan according to an embodiment of this application; Figure 2 This is another flowchart illustrating a fan speed adaptive control method according to an embodiment of this application; Figure 3 This is a schematic diagram of an exemplary hardware structure of the device in the embodiments of this application. Detailed Implementation

[0022] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0023] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0024] Please see Figure 1 This is a flowchart illustrating a method for adaptive speed control of a fan in an embodiment of this application.

[0025] S101. When the target wind turbine is running stably, based on the stable operating parameters monitored under multiple preset power levels, the virtual back EMF constant between the back EMF and the rotational speed is determined through calibration using a preset correlation model.

[0026] Among them, the target fan refers to a specific fan device that requires adaptive speed control, which can be a single fan or a fan unit in a fan array; stable operation refers to the working state in which the key parameters of the fan, such as speed, power, and current, remain within a certain range with minimal fluctuations under conditions of no significant external interference; preset power refers to several different power points determined in advance, usually selected from several typical power values ​​covering the working range of the fan, such as 30%, 50%, 70%, and 90% of the rated power; stable operation parameters refer to various operating data of the fan under stable operating conditions, including but not limited to speed, current, voltage, power factor, and temperature; back electromotive force refers to the induced electromotive force generated when the motor rotates due to the conductor cutting magnetic lines of force, which is proportional to the speed; virtual back electromotive force constant represents the proportional relationship coefficient between back electromotive force and speed under specific operating conditions, and is an important indicator of the fan's electrical characteristics.

[0027] During the initial installation, commissioning, or periodic maintenance and calibration phase of the wind turbine, and while ensuring the target wind turbine is in a stable operating state, the wind turbine runs continuously for a period of time (e.g., 15-30 minutes) with all parameters fluctuating within allowable ranges (e.g., speed fluctuation not exceeding ±1%). Following a pre-set test plan, the wind turbine is sequentially adjusted to multiple different preset power points (typically 4-8 power points, covering the wind turbine's common operating range), and maintained in stable operation at each power point for a period of time (e.g., 3-5 minutes). During the stable operation at each power point, the equipment continuously collects the wind turbine's operating parameters, including but not limited to voltage, current, power factor, speed, output power, and bearing temperature, with a sampling frequency typically 1-10 times per second. For each power point, the equipment calculates the average value of each parameter during stable operation, forming a characteristic parameter set for that power point. The equipment inputs the characteristic parameter sets of all power points into a preset correlation model. This model, based on motor theory, describes the relationship between back electromotive force and parameters such as speed, current, and voltage. Using the least squares method or other optimization algorithms, the equipment fits the model parameters to obtain the best-fit virtual back electromotive force constant. This constant reflects the inherent proportional relationship between the back electromotive force and the rotational speed of the target wind turbine under ideal steady-state conditions, and is an important benchmark value for subsequently judging the degree of deviation of the wind turbine's operating state. The equipment stores the calculated virtual back electromotive force constant in the system database as a characteristic parameter of the wind turbine for subsequent real-time monitoring and control calculations.

[0028] S102. Input the real-time monitored operating parameters into the preset correlation model to obtain the real-time virtual back electromotive force.

[0029] During normal operation of the wind turbine, the equipment continuously collects the turbine's operating parameters using various sensors installed on the turbine system, including voltage sensors, current sensors, speed sensors, and temperature sensors, at a preset sampling frequency (typically 5-20 times per second, adjustable depending on control accuracy requirements). The collected raw data is first amplified and filtered by a signal conditioning circuit to eliminate potential electrical noise and interference. Then, the equipment performs preliminary verification of the processed data, checking for any obvious outliers (such as data exceeding the physically possible range). Outliers are processed using interpolation or replacement with the most recent valid value. Next, the equipment inputs the verified operating parameters into the preset correlation model established and calibrated in step S101. This model may be based on mathematical equations of motor theory (such as an equivalent circuit model considering resistance, inductance, magnetic flux, etc.) or an empirical model obtained through data fitting (such as a multinomial regression model or a neural network model). The equipment calculates through the model, converting parameters such as voltage, current, power factor, and speed into virtual back electromotive force values ​​under the current operating condition. During the calculation process, the equipment considers the impact of temperature on electrical parameters and makes necessary parameter corrections based on the measured temperature. Finally, the equipment temporarily stores the calculated real-time virtual back electromotive force value in the system cache for subsequent disturbance analysis and control calculations.

[0030] In some embodiments, the calculation of real-time virtual back EMF can be achieved in several ways: Optionally, the device can use a method based on the equivalent circuit of the motor to calculate the real-time virtual back EMF. Specifically, the device first collects the phase voltage, phase current, power factor, and speed data of the wind turbine motor. Then, based on the equivalent circuit equation of the motor, considering factors such as stator resistance, stator inductance, and iron losses, the device obtains the amplitude and phase of the back EMF through vector calculation. Next, the device applies temperature compensation to the calculation results, adjusting the resistance parameters according to the measured motor temperature to improve the calculation accuracy. Finally, the device outputs the compensated back EMF value as the real-time virtual back EMF. Optionally, the device can also use a data-driven real-time calculation method. Specifically, the device first organizes the collected multi-dimensional operating parameters into feature vectors. Then, the device inputs the feature vectors into a pre-trained machine learning model (such as a support vector regression or random forest model), which can directly map the input features to the back EMF value. Next, the device smooths the model output to reduce the impact of random fluctuations on subsequent control. Finally, the device uses the smoothed value as the real-time virtual back EMF. It is understandable that other methods can be used to calculate the real-time virtual back EMF, such as state observer-based methods or hybrid model methods, which are not limited here.

[0031] S103. The difference between the real-time virtual back EMF and the virtual back EMF constant is used as the dynamic disturbance quantity.

[0032] In each control cycle, the equipment reads the virtual back EMF constant calibrated and stored in step S101 from the system database. This constant represents the characteristic benchmark of the fan under ideal steady-state conditions. Simultaneously, the equipment acquires the real-time virtual back EMF value calculated in step S102, which reflects the actual characteristics of the fan under its current operating state. Then, the equipment calculates the difference between these two values. This can be done using an absolute difference calculation method: Dynamic Disturbance = Real-time Virtual Back EMF - Virtual Back EMF Constant; or a relative difference calculation method: Dynamic Disturbance = (Real-time Virtual Back EMF - Virtual Back EMF Constant) / Virtual Back EMF Constant × 100%. The latter, expressed as a percentage, is more convenient for comparing different fan specifications. The calculated dynamic disturbance may be positive, indicating that the current disturbance to the fan causes the back EMF to be higher than the ideal state, usually meaning a reduced fan load or improved airflow; it may also be negative, indicating that the current back EMF is lower than the ideal state, usually meaning an increased fan load or obstructed airflow. The device temporarily stores the calculated dynamic disturbance values ​​in the system cache and simultaneously adds them to the historical data queue. This queue stores dynamic disturbance value data for a recent period (e.g., the last 10 minutes or longer) for subsequent statistical analysis and threshold calculation. Furthermore, the device performs preliminary smoothing of the dynamic disturbance values, such as using moving averages or exponential smoothing methods, to reduce the impact of random noise on subsequent judgments. The entire calculation process is completed within milliseconds, ensuring that the control system can promptly capture changes in the wind turbine's operating status.

[0033] In some embodiments, the calculation and processing of dynamic disturbance quantities can be implemented in multiple ways: Optionally, the device can employ a multi-timescale difference calculation method. Specifically, the device first calculates the difference between the real-time average virtual back EMF and the virtual back EMF constant over a short timescale (e.g., within 1 second) as the fast-response disturbance quantity. Then, the device calculates the difference between the real-time average virtual back EMF and the virtual back EMF constant over a medium timescale (e.g., within 10 seconds) as the medium-term disturbance quantity. Next, the device calculates the difference between the real-time average virtual back EMF and the virtual back EMF constant over a long timescale (e.g., within 1 minute) as the persistent disturbance quantity. Finally, the device comprehensively judges the nature and severity of the current disturbance based on the changing characteristics of the disturbance quantities at different timescales. Optionally, the device can also use a frequency domain analysis method to process dynamic disturbance quantities. Specifically, the device first converts the time series of dynamic disturbance quantities over a period of time (e.g., the most recent 30 seconds) to the frequency domain using a fast Fourier transform. Then, the device analyzes the energy distribution of the disturbance signal in different frequency bands to identify the main disturbance frequencies. Next, the equipment classifies disturbances into low-frequency continuous disturbances (such as changes in ambient temperature), mid-frequency periodic disturbances (such as changes in system load cycles), and high-frequency transient disturbances (such as sudden impacts) based on frequency characteristics. Finally, the equipment employs corresponding filtering strategies for different types of disturbances to extract effective disturbance information. It is understood that other methods can also be used to calculate and process dynamic disturbances, such as time-frequency joint analysis methods based on wavelet analysis or adaptive filtering methods; these are not limited here. Furthermore, it should be noted that when calculating dynamic disturbances, the equipment also considers the transient characteristics during the fan startup and shutdown processes, using different calculation parameters or temporarily shielding disturbance judgment during these special phases to avoid misjudging normal operation processes as abnormal disturbances.

[0034] S104. Based on the mean and standard deviation of the dynamic disturbance within the preset time window, the mean is determined as the zero-impact disturbance threshold, and the sum of the mean and three times the standard deviation is determined as the maximum impact disturbance threshold.

[0035] During wind turbine operation, the equipment determines a preset time window for statistical analysis. The length of this window is set based on the wind turbine's operating characteristics and the rate of environmental change, typically ranging from 10 to 60 minutes. For relatively stable environments, a longer time window (e.g., 30-60 minutes) can be used to obtain more stable statistical results; for rapidly changing environments, a shorter time window (e.g., 10-20 minutes) can be used to adapt to environmental changes more quickly. The equipment then extracts all dynamic disturbance data within this time window from the historical database, typically containing hundreds to thousands of data points. Next, the equipment preprocesses this data, including outlier detection and handling (e.g., removing extreme values ​​outside a reasonable range, usually using the 3σ criterion or interquartile range method). After preprocessing, the equipment calculates the arithmetic mean of these data, which reflects the baseline level of dynamic disturbances under the current operating environment. Ideally, if the wind turbine operates exactly as calibrated, the mean should be close to zero; however, in real-world environments, due to various long-term factors (e.g., seasonal temperature changes, equipment aging), the mean usually has some deviation. The equipment defines this mean as the zero-impact disturbance threshold, indicating that disturbances at this level are considered the "new normal" and do not require special control adjustments. Simultaneously, the equipment calculates the standard deviation of these data, reflecting the volatility of dynamic disturbances; a larger standard deviation indicates an unstable operating environment or inconsistent turbine response. Then, the equipment adds the mean to three times the standard deviation to obtain the maximum impact disturbance threshold. This threshold is based on statistical principles; under the assumption of a normal distribution, approximately 99.7% of disturbance data will fall within the range of the mean ± three times the standard deviation. Therefore, disturbances exceeding this threshold are considered significantly abnormal and require the strongest adjustment measures. The equipment stores the calculated zero-impact disturbance threshold and maximum impact disturbance threshold in the system for subsequent safety margin factor calculations.

[0036] S105. By interpolating between the zero-impact disturbance threshold and the maximum-impact disturbance threshold, the safety margin factor corresponding to the dynamic disturbance at the current time is obtained and output.

[0037] The equipment acquires the current dynamic disturbance value within each control cycle. This value, calculated in step S103, reflects the degree to which the current operating state of the wind turbine deviates from the ideal stable state. Simultaneously, the equipment reads the zero-impact disturbance threshold and the maximum-impact disturbance threshold determined in step S104 from the system. Then, the equipment determines the positional relationship of the current dynamic disturbance relative to these two thresholds. If the dynamic disturbance is less than or equal to the zero-impact disturbance threshold, it indicates that the current disturbance has almost no impact on the wind turbine performance, and the equipment sets the safety margin factor to the maximum value of 1.0. If the dynamic disturbance is greater than or equal to the maximum-impact disturbance threshold, it indicates that the current disturbance has a significant impact on the wind turbine performance, and the equipment sets the safety margin factor to the minimum value of 0.0. If the dynamic disturbance is between the two thresholds, the equipment determines the safety margin factor through interpolation. Interpolation typically uses a linear interpolation method, that is, calculating the corresponding value of the safety margin factor between 0 and 1 according to the relative positional proportion of the dynamic disturbance between the two thresholds. The calculation formula is: Safety margin factor = 1.0 - (Current dynamic disturbance - Zero-impact disturbance threshold) / (Maximum-impact disturbance threshold - Zero-impact disturbance threshold). Furthermore, the equipment may employ nonlinear interpolation methods, such as quadratic or exponential interpolation, to more precisely reflect the impact of different disturbance levels on the safety margin. The calculated safety margin factor is temporarily stored in the system and may be displayed to operators through a human-machine interface, serving as an important basis for subsequent control strategy selection. The equipment also records the historical trend of the safety margin factor for long-term operational status analysis and predictive maintenance.

[0038] S106. Select the target wind turbine adjustment strategy based on the safety margin factor and adjust the weight parameters of the PID algorithm based on the wind turbine adjustment strategy.

[0039] Within each control cycle, the equipment acquires the safety margin factor calculated in step S105, which reflects the safety level of the current wind turbine operating state. Then, based on the value range of the safety margin factor, the equipment selects the appropriate control mode from a preset modal strategy library. The modal strategy library typically contains at least three basic control modes: when the safety margin factor is close to 1.0 (e.g., greater than 0.8), a conservative control mode is selected, emphasizing system stability and avoiding over-adjustment; when the safety margin factor is in the intermediate range (e.g., between 0.3 and 0.8), a balanced control mode is selected, achieving a balance between stability and response speed; when the safety margin factor is close to 0 (e.g., less than 0.3), an aggressive control mode is selected, emphasizing rapid response to quickly handle severe disturbances. Next, the equipment extracts the corresponding PID parameter configuration set from the selected control mode, including the baseline proportional coefficient (Kp), integral coefficient (Ki), and derivative coefficient (Kd). To achieve finer control, the equipment further adjusts the baseline PID parameters based on the specific position of the safety margin factor within the current modal range. Adjustments typically employ linear or quadratic interpolation algorithms to smoothly transition between parameter configurations of adjacent modes, avoiding abrupt changes in the control strategy. Ultimately, the equipment obtains PID algorithm weight parameters optimized for the current operating state; these parameters will be used in subsequent turbine adjustment parameter calculations. The equipment also records historical parameter adjustment data for long-term optimization of control strategies and parameter configurations.

[0040] S107. Calculate the fan adjustment parameters that ensure the fan speed matches the target speed using dynamic disturbance and PID algorithm.

[0041] In each control cycle, the equipment acquires the actual speed of the current fan and the target speed set by the system, and calculates the speed deviation between the two. Simultaneously, the equipment acquires the dynamic disturbance calculated in step S103, which reflects the degree of influence of external factors on the fan operation. Then, the equipment inputs the speed deviation into the PID control algorithm, which uses the weight parameters (Kp, Ki, Kd) adjusted according to the safety margin factor in step S106. The PID algorithm calculation includes three parts: a proportional term (current deviation multiplied by the proportional coefficient Kp), an integral term (cumulative sum of deviations multiplied by the integral coefficient Ki), and a derivative term (rate of change of deviation multiplied by the derivative coefficient Kd). The sum of these three parts constitutes the basic control quantity. Next, the equipment combines the dynamic disturbance as a feedforward compensation term with the basic control quantity. The purpose of feedforward compensation is to proactively address detected disturbances, rather than relying solely on feedback control to respond to deviations. The combination method typically involves multiplying the dynamic disturbance by a feedforward coefficient and then adding it to the PID basic control quantity. The feedforward coefficient can be a fixed value or dynamically adjusted according to the disturbance characteristics. Finally, the equipment converts the combined control quantities into specific fan adjustment parameters, such as the frequency setpoint of the inverter, the voltage or current command of the motor, etc. The specific conversion method depends on the type of fan drive system. The equipment sends the calculated fan adjustment parameters to the actuator (such as the inverter, motor driver, etc.) to achieve precise control of the fan speed.

[0042] In this embodiment, by employing virtual back EMF constant calibration and dynamic threshold determination techniques, the system can adaptively adjust the disturbance judgment criteria, avoiding misjudgments caused by fixed thresholds, solving the problem of insufficient adaptability of fixed parameter PID in complex environments, and improving the accuracy and stability of fan speed control.

[0043] While the above embodiment can achieve single-unit adaptive control, in a multi-fan parallel system, when local airflow paths are blocked (e.g., a server rack overheating leading to increased local air resistance, or severe dust accumulation on filters in a certain area of ​​the airflow wall), the obstructed fan will blindly increase its speed under traditional methods, resulting in a surge in energy consumption with minimal actual airflow contribution, and may even enter the surge zone. Meanwhile, fans in unobstructed areas may reduce their output due to changes in the main duct pressure. This operating mode may lead to a decrease in the overall system efficiency.

[0044] Please see Figure 2 This is another flowchart illustrating a fan speed adaptive control method in an embodiment of this application.

[0045] S201. When the target wind turbine is running stably, based on the stable operating parameters monitored under multiple preset power levels, the virtual back EMF constant between the back EMF and the rotational speed is determined through calibration using a preset correlation model.

[0046] S202. Input the real-time monitored operating parameters into the preset correlation model to obtain the real-time virtual back electromotive force.

[0047] S203. The difference between the real-time virtual back EMF and the virtual back EMF constant is used as the dynamic disturbance quantity.

[0048] Steps S201 to S203 are similar to steps S101 to S103, and will not be described in detail here.

[0049] S204. Within a preset number of control cycles, traverse the multiple virtual back electromotive force constants of all wind turbines in the target wind turbine array to which the target wind turbine belongs, and obtain the reference disturbance value of the target wind turbine array based on statistical analysis.

[0050] Among them, the target wind turbine array refers to the wind turbine group where the target wind turbine is located, which is usually composed of multiple wind turbines with similar functions and close locations.

[0051] During the wind turbine array coordinated control phase, the equipment determines the time range for statistical analysis, i.e., a preset number of control cycles. This value is typically set based on the size of the wind turbine array and the characteristics of the operating environment. For example, 50-100 control cycles might be chosen for a small array (5-10 turbines), while 200-500 control cycles might be chosen for a large array (dozens of turbines) to ensure sufficient data collection. Then, the equipment acquires the virtual back-EMF constant data of all turbines in the target wind turbine array during this time period via network communication or a central control system. This data may come from the calculation results of each turbine itself or be calculated uniformly by the central system. Next, the equipment preprocesses all collected virtual back-EMF constant data, including outlier detection and processing (such as removing significantly deviating extreme values) and data standardization. After preprocessing, the equipment performs statistical analysis, calculating the statistical characteristics of the virtual back-EMF constant of the entire wind turbine array, such as the mean, median, and standard deviation. In most cases, the equipment chooses the median as the benchmark disturbance value because the median is not sensitive to extreme values ​​and can more robustly reflect the overall level. In certain special cases, the equipment may also choose a weighted average or other statistical measure as the baseline disturbance value, with the weights possibly related to the location, capacity, or importance of the wind turbine. The equipment stores the calculated baseline disturbance value in the system for subsequent abnormal wind turbine identification and coordinated control. In addition, the equipment also records the historical trend of the baseline disturbance value for analyzing the long-term operating status of the entire wind turbine array.

[0052] In some embodiments, the baseline perturbation value can be determined in several ways: Optionally, the device can use a hierarchical clustering method to determine the baseline perturbation value. Specifically, the device first performs hierarchical clustering analysis on the virtual back EMF constant data of all collected wind turbines, dividing the data into several groups. Then, the device identifies the largest data cluster, which typically represents the group of wind turbines operating normally. Next, the device calculates the center value (such as the mean or median) of this main data cluster. Finally, the device determines this center value as the baseline perturbation value. This method can automatically eliminate the influence of abnormal wind turbines and obtain a more accurate baseline. Optionally, the device can also use a spatiotemporal correlation analysis method to determine the baseline perturbation value. Specifically, the device first establishes a spatial distribution model of the wind turbine array, considering the physical location relationship of each wind turbine. Then, the device analyzes the spatial distribution pattern of the virtual back EMF constant, identifying possible regional differences (such as local airflow changes due to terrain or obstacles). Next, the device calculates the baseline value separately for each region, forming a spatially distributed baseline perturbation value field. Finally, the device extracts the corresponding local baseline value from the baseline perturbation value field according to the location of the target wind turbine. It is understandable that other methods can be used to determine the baseline disturbance value, such as pattern matching based on historical data or adaptive filtering methods, which are not limited here. In addition, it should be noted that when determining the baseline disturbance value, the equipment will also consider time factors, such as working hours (day shift / night shift), seasonal changes, etc., and different processing strategies may be adopted for data under different time conditions.

[0053] S205. Based on the distribution of multiple virtual back EMF constants relative to the reference disturbance value, determine that the wind turbines exceeding the preset dynamic discrimination threshold are high-load abnormal wind turbines.

[0054] After calculating the baseline disturbance value, the equipment acquires the virtual back EMF constant data of all wind turbines in the target wind turbine array, as well as the baseline disturbance value calculated in step S204. Then, the equipment calculates the deviation between the virtual back EMF constant of each wind turbine and the baseline disturbance value, which can be an absolute deviation (direct subtraction) or a relative deviation (deviation divided by the baseline value to obtain a percentage). Next, the equipment analyzes the distribution of deviations across all wind turbines, calculating statistical characteristics such as mean, standard deviation, and quartiles, and constructs a probability density function or cumulative distribution function for the deviation distribution. Based on these distribution characteristics, the equipment determines a preset dynamic discrimination threshold. This threshold is not fixed but dynamically calculated based on the current distribution, typically set as the mean plus a specific multiple (e.g., 2 or 2.5 times) of the standard deviation, or a specific quantile (e.g., the 95th percentile). The equipment compares the deviation value of each wind turbine with the dynamic discrimination threshold, identifying wind turbines with deviations exceeding the threshold as high-load abnormal wind turbines. The virtual back electromotive force constants of these wind turbines are significantly higher than the overall array level, indicating that they may face greater loads or resistance and require special attention and handling.

[0055] S206. For the cooperating fan that is closest to the physical location of the high-load abnormal fan and whose virtual back EMF constant is not greater than the reference disturbance value, the target speed of the cooperating fan shall be increased proportionally.

[0056] After identifying a high-load abnormal fan, the device acquires information on all high-load abnormal fans identified in step S205, including their ID, physical location coordinates, and virtual back EMF constant value. Then, the device extracts the physical location information of all fans from the fan array layout database, typically represented in three-dimensional coordinates (x, y, z) or two-dimensional coordinates (x, y) plus a height value. Next, for each high-load abnormal fan, the device calculates its physical distance to all other fans in the array, using the Euclidean distance formula or a weighted distance formula considering the actual airflow path. The device sorts all fans from closest to furthest from the high-load abnormal fan and filters out fans whose virtual back EMF constant is not greater than the baseline disturbance value; these fans are considered potential cooperating fan candidates. From these candidates, the device typically selects the 2-5 closest fans as actual cooperating fans; the specific number may be dynamically determined based on the degree of abnormality of the high-load abnormal fan and the number of available candidates. For each selected cooperating fan, the device calculates the target speed increase. The calculation typically considers several factors: the degree of abnormality of the high-load abnormal fan (the magnitude of the deviation of the virtual back EMF constant from the baseline value), the distance between the cooperating fan and the abnormal fan (the closer the distance, the greater the increase), and the load status of the cooperating fan itself (the lower the virtual back EMF constant, the greater the increase). The increase is usually expressed as a percentage increase in the original target speed, such as 5%, 8%, or 10%. The equipment sends the calculated new target speed value to the control system of each cooperating fan to achieve speed adjustment. Simultaneously, the equipment continuously monitors the response of the cooperating fans and the status changes of the high-load abnormal fan to evaluate the effectiveness of the coordinated control.

[0057] In some embodiments, when the device detects multiple high-load abnormal fans simultaneously, it first constructs a collaboration graph, where nodes represent all fans (including abnormal fans and potential collaborative fans), and edges represent collaboration relationships (from a collaborative fan to a supported abnormal fan). Then, the device identifies those fans that have received multiple collaboration requests (represented in the graph as nodes with multiple outgoing edges). For these fans, instead of simply summing the speed increase values ​​of all requests, the device employs a more complex allocation strategy: it can allocate based on the degree of abnormality, prioritizing requests from fans with more severe abnormalities; it can allocate based on an inverse proportion of distance, providing greater support to closer abnormal fans; or it can use a global optimization algorithm to find the collaboration configuration that maximizes the improvement of overall system performance, which will not be elaborated here.

[0058] S207. Based on the mean and standard deviation of the dynamic disturbance within the preset time window, the mean is determined as the zero-impact disturbance threshold, and the sum of the mean and three times the standard deviation is determined as the maximum impact disturbance threshold.

[0059] S208. By interpolating between the zero-impact disturbance threshold and the maximum-impact disturbance threshold, the safety margin factor corresponding to the dynamic disturbance at the current time is obtained and output.

[0060] Steps S207 to S208 are similar to steps S104 to S105, and will not be described in detail here.

[0061] S209. Perform a discrete Fourier transform on the time series of dynamic disturbances within the preset time window to obtain the spectrum data.

[0062] During wind turbine operation, the equipment determines the length of a preset time window for spectrum analysis. This length is typically set based on the characteristics of the disturbance to be analyzed. For example, a longer window (e.g., 1-5 minutes) might be chosen to analyze low-frequency disturbances (such as slow environmental changes), while a shorter window (e.g., 1-10 seconds) might be chosen to analyze high-frequency disturbances (such as mechanical vibrations). The equipment then extracts sample data of all dynamic disturbance quantities within the preset time window from the system cache. This data is usually recorded at a fixed sampling frequency (e.g., one sample every 10-100 milliseconds), forming a disturbance time series. Next, the equipment preprocesses the extracted time series data, including removing linear trends, applying window functions (such as Hanning or Hamming windows) to reduce spectral leakage, and zero-padding to improve frequency resolution. After preprocessing, the equipment applies a Fast Fourier Transform (FFT, an efficient implementation of the Discrete Fourier Transform) to the processed time series, converting the time-domain signal into a frequency-domain representation. The raw result of the FFT calculation is a series of complex values. The equipment calculates the modulus of these complex numbers to obtain the energy amplitude at each frequency point, forming the spectrum data. Spectral data is typically represented as a series of frequency points and their corresponding energy amplitudes, ranging from 0 Hz (DC component) to half the sampling frequency (Nyquist frequency). The device stores the calculated spectral data in the system for subsequent frequency characteristic analysis and disturbance classification. Additionally, the device may calculate the power spectral density (PSD) or other spectral characteristics to more comprehensively describe the frequency characteristics of the disturbance.

[0063] S210. Calculate the sum of all energy amplitudes in the spectrum data whose frequencies are lower than the preset gradual change cutoff frequency to obtain the total value of the gradual change disturbance energy.

[0064] Among them, the energy amplitude refers to the energy level corresponding to each frequency point in the spectrum data, which is usually the square of the modulus of the complex FFT result; the total energy of the slowly varying disturbance represents the total energy of the low-frequency disturbance components, reflecting the degree of influence of slowly changing factors in the system (such as changes in ambient temperature, gradual aging of equipment, etc.).

[0065] After completing the spectrum data calculation, the device obtains a preset slow-change cutoff frequency value from the system configuration. This value is typically preset based on the physical characteristics of the wind turbine system and the actual application scenario; for example, it might be set to 0.1Hz for a large industrial wind turbine, indicating that changes with a period greater than 10 seconds are considered slow-change disturbances. Then, the device extracts all data points with frequencies lower than the preset slow-change cutoff frequency from the spectrum data calculated in step S209. These data points include the frequency value and the corresponding energy amplitude. Next, the device sums the energy amplitudes of these low-frequency data points to obtain the total energy value of the slow-change disturbance. During the calculation, the device may apply a frequency-weighted method, such as giving higher weight to ultra-low frequencies closer to the DC component (0Hz), as these components typically represent slower and more persistent system changes. Furthermore, the device may normalize the energy amplitude, such as dividing it by the total spectrum energy, to obtain the proportion of the slow-change disturbance in the total disturbance, which helps in subsequent proportional relationship analysis. The device stores the calculated total energy value of the slow-change disturbance in the system for subsequent disturbance characteristic analysis and control strategy adjustment. In addition, the equipment may also record the changing trends of historical slowly varying disturbance energy values ​​for long-term system behavior analysis and predictive maintenance.

[0066] In some embodiments, the calculation of slowly varying disturbance energy can be achieved in several ways: Optionally, the device can use a piecewise weighted integral method to calculate the slowly varying disturbance energy. Specifically, the device first divides the low-frequency region (0Hz to a preset slowly varying cutoff frequency) into multiple sub-bands, such as the ultra-low frequency band (0-0.01Hz), the extremely low frequency band (0.01-0.05Hz), and the low frequency band (0.05-preset slowly varying cutoff frequency). Then, the device assigns different weighting coefficients to each sub-band, typically with the ultra-low frequency band having the highest weight, which gradually decreases as the frequency increases. Next, the device calculates the sum of the energy amplitudes within each sub-band and applies the corresponding weighting coefficients. Finally, the device sums the weighted energy values ​​of each sub-band to obtain a more accurate total value of slowly varying disturbance energy that reflects the slowly varying characteristics. Optionally, the device can also use a spectral morphology analysis method to calculate the slowly varying disturbance energy. Specifically, the device first analyzes the morphological characteristics of the spectrum in the low-frequency region, such as the smoothness of the energy distribution, peak positions, and attenuation rates. Then, based on these morphological characteristics, the device identifies low-frequency components with significant physical meaning, such as frequencies related to diurnal environmental variations (approximately 1 / 86400Hz) or frequencies related to the device's thermal cycling. Next, the device focuses on calculating the energy contribution of these physically significant low-frequency components. Finally, the device separates the energy of these key low-frequency components from other low-frequency background energies to obtain more targeted characteristic values ​​of slowly varying disturbance energy. It is understood that other methods can also be used to calculate slowly varying disturbance energy, such as multi-resolution energy analysis methods based on wavelet decomposition or adaptive filtering energy extraction methods; this is not limited here.

[0067] S211. Calculate the sum of all energy amplitudes in the spectrum data whose frequencies are higher than the preset mutation start frequency to obtain the total value of mutation disturbance energy.

[0068] The device first obtains a preset mutation start frequency value from the system configuration. This value is typically preset based on the dynamic response characteristics of the fan system and the actual application scenario; for example, it might be set to 1Hz for a medium-sized industrial fan, indicating that changes with a period of less than 1 second are considered mutation disturbances. Then, the device extracts all data points with frequencies higher than the preset mutation start frequency from the spectrum data calculated in step S209. These data points include the frequency value and the corresponding energy amplitude. Next, the device sums the energy amplitudes of these high-frequency data points to obtain the total energy value of the mutation disturbance. During the calculation, the device may apply a frequency-weighted method, such as giving higher weights to specific frequency ranges (e.g., regions close to the system's inherent frequency), because disturbances at these frequencies may have a more significant impact on system stability. Furthermore, the device may normalize the energy amplitude, such as dividing it by the total spectrum energy, to obtain the proportion of the mutation disturbance in the total disturbance, which helps in subsequent proportional analysis.

[0069] S212. Based on the ratio between the total energy of slowly varying disturbances and the total energy of sudden disturbances, a disturbance characteristic correction coefficient is generated, and the weight parameters in the target PID parameter configuration set are proportionally adjusted using the disturbance characteristic correction coefficient to obtain the adjusted weight parameters.

[0070] The target PID parameter configuration set refers to the pre-set combination of PID control algorithm parameters, including the proportional coefficient (Kp), integral coefficient (Ki), and derivative coefficient (Kd); the weight parameters are the three coefficients in the PID algorithm that determine the response characteristics of the control system to deviations.

[0071] After calculating the total energy of slowly varying disturbances and the total energy of abrupt disturbances, the equipment calculates the ratio between the two, typically expressed as the percentage of slowly varying disturbance energy to the total disturbance energy (slow plus abrupt), or the ratio of slowly varying disturbance energy to abrupt disturbance energy. This ratio directly reflects the main characteristics of the current system disturbance: a higher ratio indicates that the system is mainly affected by slowly changing factors, while a lower ratio indicates that the system is mainly affected by rapidly changing factors. Then, based on this ratio, the equipment generates disturbance characteristic correction coefficients through a preset mapping function. The mapping function is usually a nonlinear function that maps the ratio to a coefficient range suitable for adjusting control parameters (e.g., 0.5-2.0). For example, when slowly varying disturbances dominate (high ratio), the correction coefficients may favor enhancing integral action and weakening derivative action; when abrupt disturbances dominate (low ratio), the correction coefficients may favor enhancing derivative action and weakening integral action. Next, the device retrieves the target PID parameter configuration set for the current operating mode from the control system configuration. This configuration set includes baseline proportional (Kp), integral (Ki), and derivative (Kd) values. Then, the device adjusts these baseline parameters using disturbance characteristic correction factors. The adjustments are typically differentiated: for the proportional (Kp) coefficient, the square root of the correction factor may be applied; for the integral (Ki) coefficient, a correction factor positively correlated with the proportion of slowly varying disturbances may be applied; and for the derivative (Kd) coefficient, a correction factor positively correlated with the proportion of abrupt disturbances may be applied. Finally, the device uses the adjusted PID parameters as the final weighting parameters for subsequent control algorithm calculations. The device also records historical data on parameter adjustments for long-term optimization of control strategies and parameter configurations.

[0072] S213. Select the target wind turbine adjustment strategy based on the safety margin factor and adjust the weight parameters of the PID algorithm based on the wind turbine adjustment strategy.

[0073] S214. Calculate the fan adjustment parameters that ensure the fan speed matches the target speed using dynamic disturbance and PID algorithm.

[0074] Steps S213 and S214 are similar to steps S106 to S107, and will not be described in detail here.

[0075] In a further embodiment of this application, high-load abnormal fans are identified by calculating the benchmark disturbance value. Combined with a collaborative control mechanism, the system can automatically adjust resource allocation in the event of partial blockage, avoid ineffective energy consumption of blocked fans, improve overall system efficiency, and provide maintenance personnel with accurate fault location information.

[0076] The exemplary device 300 provided in the embodiments of this application is described below. Figure 3This is an exemplary hardware structure diagram of the device 300 provided in the embodiments of this application.

[0077] In some embodiments, the device 300 is a computer device or includes a computer device. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores data. The network interface of the computer device is used to communicate with other external terminals or servers via a network connection. In some embodiments, the network interface can be a wired network interface; in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, it implements the methods in the embodiments of this application.

[0078] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements. The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. 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 scope of the technical solutions of the embodiments of this application.

[0079] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0080] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0081] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for adaptive speed control of a fan, characterized in that, The method includes: When the target wind turbine is running stably, based on the stable operating parameters monitored under multiple preset power levels, a preset correlation model is used for calibration to determine the virtual back EMF constant between the back EMF and the rotational speed when the target wind turbine is running stably. The real-time monitored operating parameters are input into a preset correlation model to obtain the real-time virtual back electromotive force; The difference between the real-time virtual back electromotive force and the virtual back electromotive force constant is used as the dynamic disturbance quantity. Based on the mean and standard deviation of the dynamic disturbance within a preset time window, the mean is determined as the zero-impact disturbance threshold, and the sum of the mean and 3 times the standard deviation is determined as the maximum impact disturbance threshold. By interpolating between the zero-impact disturbance threshold and the maximum-impact disturbance threshold, the safety margin factor corresponding to the dynamic disturbance amount at the current time is obtained and output. Based on the safety margin factor, a wind turbine adjustment strategy for the target wind turbine is selected, and the weight parameters of the PID algorithm are adjusted based on the wind turbine adjustment strategy. The dynamic disturbance and PID algorithm are used to calculate the fan adjustment parameters that ensure the fan speed matches the target speed.

2. The method according to claim 1, characterized in that, The steps of selecting the wind turbine adjustment strategy for the target wind turbine based on the safety margin factor and adjusting the weight parameters of the PID algorithm based on the wind turbine adjustment strategy specifically include: Based on the preset value range of the safety margin factor, modal mapping is performed in the modal strategy library to determine the target control mode. The modal strategy library contains at least three preset control modes, and the three preset control modes correspond to different PID parameter configuration sets. After extracting the corresponding target PID parameter configuration set according to the target control mode, the target PID parameter configuration set is corrected based on the relative position of the safety margin factor within the preset numerical range to obtain the weight parameters of the PID algorithm. The correction is calculated between the parameter configurations of adjacent modes through linear interpolation or quadratic interpolation algorithms.

3. The method according to claim 1, characterized in that, After the step of using the difference between the real-time virtual back electromotive force and the virtual back electromotive force constant as a dynamic disturbance, the method further includes: Within a preset number of control cycles, the multiple virtual back electromotive force constants of all wind turbines in the target wind turbine array to which the target wind turbine belongs are traversed, and the reference disturbance value of the target wind turbine array is obtained based on statistical analysis. Based on the distribution of the multiple virtual back EMF constants relative to the reference disturbance value, wind turbines exceeding a preset dynamic discrimination threshold are identified as high-load abnormal wind turbines. The preset dynamic discrimination threshold is determined based on the discrete values ​​of the distribution. For the cooperating fan that is closest to the physical location of the high-load abnormal fan and whose virtual back EMF constant is not greater than the reference disturbance value, the target speed of the cooperating fan is increased proportionally.

4. The method according to claim 1, characterized in that, Before the step of correcting the target PID parameter configuration set based on the relative position of the safety margin factor within the preset numerical range, the method further includes: The time series of the dynamic disturbance within the preset time window is subjected to discrete Fourier transform to obtain spectral data, which consists of a series of frequency points and the energy amplitude corresponding to the frequency points. The sum of all energy amplitudes with frequencies lower than a preset gradual cutoff frequency in the spectrum data is calculated to obtain the total value of the gradual disturbance energy. The gradual cutoff frequency is determined by disturbance events caused by gradual load changes or slow environmental drift. The sum of all energy amplitudes with frequencies higher than the preset mutation start frequency in the spectrum data is calculated to obtain the total value of mutation disturbance energy. The mutation start frequency is determined by a disturbance event caused by a transient impact or a sudden change in load. Based on the ratio between the total energy of the slowly varying disturbance and the total energy of the abrupt disturbance, a disturbance characteristic correction coefficient is generated, and the weight parameters in the target PID parameter configuration set are proportionally adjusted using the disturbance characteristic correction coefficient to obtain the adjusted weight parameters.

5. The method according to claim 4, characterized in that, After the step of generating a perturbation characteristic correction coefficient based on the ratio between the total energy of the slowly varying perturbation and the total energy of the abrupt perturbation, the method further includes: If the proportion of the total slowly varying disturbance energy in the total disturbance energy of the wind turbine exceeds the first preset diagnostic threshold, then first diagnostic information is generated; If the proportion of the total energy of the sudden disturbance in the total energy of the wind turbine disturbance exceeds the second preset diagnostic threshold, then second diagnostic information is generated, which is different from the first diagnostic information.

6. The method according to claim 4, characterized in that, The method further includes: If the proportion of the total slowly varying disturbance energy in the total disturbance energy of the wind turbine exceeds the first preset diagnostic threshold, then linear regression is performed on the disturbance time series to calculate the predicted disturbance amount for the next control cycle. According to the preset weighting coefficient, the predicted disturbance amount and the dynamic disturbance amount of the current control cycle are weighted and averaged to obtain the comprehensive disturbance amount, and the comprehensive disturbance amount is used to replace the dynamic disturbance amount.

7. The method according to claim 5, characterized in that, The method further includes: When the preset safety margin factor is lower than the preset minimum safety threshold, an alarm message is generated and output. The alarm message is one of the first diagnostic message and the second diagnostic message.

8. A device, characterized in that, The device includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory being used to store computer program code, the computer program code including computer instructions, the one or more processors invoking the computer instructions to cause the device to perform the method as described in any one of claims 1-7.

9. A computer program product containing instructions, characterized in that, When the computer program product is run on the device, the device causes the device to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instructions are executed on the device, the device causes the device to perform the method as described in any one of claims 1-7.

Images