Wind turbine generator ups residual power prediction method based on multi-parameter correction and related device

Abstract

This invention discloses a method and related device for predicting the remaining power of a wind turbine UPS based on multi-parameter correction, belonging to the field of UPS power prediction technology. The method involves real-time acquisition of UPS system operating parameters, including at least load parameters, battery parameters, and system status parameters. Feature extraction and preprocessing are performed on the acquired operating parameters to obtain key feature parameters. Based on these key feature parameters, dynamic battery capacity calibration is performed using a preset battery capacity dynamic calibration model. Multi-dimensional parameter dynamic correction is then performed based on the key feature parameters, calculating multiple correction factors and predicting the load power. Finally, based on the dynamically calibrated battery capacity and the multiple correction factors, combined with the predicted load power, the remaining power supply prediction time is calculated. This invention significantly improves the accuracy and reliability of wind turbine UPS remaining power time prediction, offering higher safety and adaptability.

Description

Technical Field

[0001] This invention belongs to the field of UPS power prediction technology, and relates to a method and related device for predicting the remaining power of a wind turbine UPS based on multi-parameter correction. Background Technology

[0002] Uninterruptible power supplies (UPS) are core backup equipment for critical control systems in wind farms (such as main control, pitch, and yaw systems). Their function is to provide continuous and stable backup power to wind turbines when grid power is interrupted, ensuring a safe and orderly shutdown process and preventing equipment damage. Accurately predicting the remaining power supply time of the UPS batteries (i.e., remaining battery life) is crucial, directly providing wind farm maintenance personnel with decision-making information to determine whether to initiate emergency procedures or perform remote intervention, thereby ensuring the safety of both the wind turbines and the grid. Currently, mainstream methods for predicting UPS remaining battery life rely primarily on two traditional techniques: the battery voltage lookup table method and the current integration method.

[0003] The voltage lookup table method is based on the correspondence between battery terminal voltage and remaining capacity. The system estimates the remaining power by measuring the battery pack's terminal voltage in real time and consulting a preset battery discharge curve (voltage-capacity curve), thus calculating the remaining power supply time. However, the prediction accuracy of this method is greatly affected by external factors and has inherent defects. Battery voltage is a dynamic parameter, not only related to the remaining capacity but also significantly affected by instantaneous load current, battery internal resistance, ambient temperature, and battery aging. In scenarios like wind farms with dynamically changing loads, the power consumption of the wind turbine control system fluctuates drastically with operating conditions. When the load suddenly increases, the voltage drop across the battery's internal resistance causes the terminal voltage to drop instantaneously. In this case, the lookup table method will severely underestimate the remaining time, causing false alarms. Conversely, when the load is light, the terminal voltage is artificially high, and the lookup table method will overestimate the remaining time, posing a serious risk of power outages to the equipment. Therefore, this method has a large prediction error and very low reliability under conditions of frequent load fluctuations.

[0004] The current integration method calculates the discharged capacity (Ah) by monitoring the battery's discharge current in real time and integrating it over time. The remaining capacity is then calculated by subtracting the discharged capacity from the battery's nominal capacity, and this remaining time is estimated by combining this with the current load power. Compared to the voltage lookup table method, the current integration method offers improved accuracy under stable loads. However, its fundamental flaw lies in its assumption that the battery's "usable capacity" is a constant nominal value. In reality, the battery's actual usable capacity is influenced by a combination of factors, including discharge rate (C-rate), ambient temperature, and state of health (SOH). According to the PGY effect, during high-current discharge, the battery's effective usable capacity is significantly less than its nominal capacity. If the calculation is still based on the nominal capacity, the predicted remaining time will be much longer than the actual time, potentially leading to unexpected power outages with disastrous consequences. Furthermore, battery capacity naturally decays (ages) with increasing cycle count, and a fixed nominal capacity cannot reflect this decay, resulting in completely inaccurate predictions in the later stages of equipment use.

[0005] In summary, existing methods are simple and fixed, failing to systematically incorporate key variables such as load power transients, battery temperature, state of health (SOH), internal resistance changes, and discharge history into real-time prediction models. Especially in dynamic load scenarios like wind turbines, the prediction errors of existing methods can be as high as 30%-50%, which cannot meet the needs of refined and highly reliable operation and maintenance of wind farms. Summary of the Invention

[0006] The purpose of this invention is to provide a method and related device for predicting the remaining power of a wind turbine UPS based on multi-parameter correction, so as to solve the technical problem that the existing technology is difficult to adapt to factors such as dynamic load, battery aging and environmental changes, resulting in poor accuracy and reliability of the remaining power prediction results.

[0007] To achieve the above objectives, the present invention employs the following technical solution: In a first aspect, the present invention provides a method for predicting the remaining power capacity of a wind turbine UPS based on multi-parameter correction, comprising the following steps: Real-time acquisition of UPS system operating parameters, including at least load parameters, battery parameters, and system status parameters; The collected operating parameters are subjected to feature extraction and preprocessing to obtain key feature parameters; Based on the aforementioned key feature parameters, dynamic calibration of battery capacity is performed using a preset dynamic calibration model for battery capacity. Based on the key feature parameters, multi-dimensional parameter dynamic correction is performed, multiple correction factors are calculated, and load power is predicted. Based on the dynamically calibrated battery capacity and the multiple correction factors, combined with the predicted load power, the remaining power supply prediction time is calculated.

[0008] Furthermore, the load parameters include real-time total output power, nonlinear load percentage, and load power change rate; the battery parameters include total battery pack voltage, voltage of each individual battery cell / module, real-time discharge current, and battery temperature; and the system status parameters include input voltage status, charging module operating status, and historical discharge records.

[0009] Furthermore, the step of extracting and preprocessing the collected operating parameters to obtain key feature parameters specifically includes: The collected operating parameters are cleaned, denoised, and transformed to identify key features that have a significant impact on power prediction. These key features include load features, current features, and battery features. For load characteristics, calculate real-time load rate and load fluctuation coefficient; identify load transient events; For the current characteristics, calculate the average discharge current and the current variation trend; For battery characteristics, the real-time internal resistance of the battery pack is estimated by the correspondence between voltage transients and current changes, abnormal voltage / current data is eliminated, and missing individual cell voltage data is supplemented.

[0010] Furthermore, the step of dynamically calibrating the battery capacity based on the key feature parameters and using a preset battery capacity dynamic calibration model specifically includes: The most recent complete discharge record is used as the core sample, combined with recent discharge fragment data and periodic self-check discharge data; A battery capacity prediction model based on LSTM is constructed. The input parameters include battery internal resistance, single cell voltage equalization, ambient temperature, number of cycles, and historical discharge capacity decay trend. The model dynamically outputs the actual maximum usable total capacity C_1(t) and real-time charging capacity C_c(t) of the battery pack at time t.

[0011] Furthermore, the step of dynamically correcting multi-dimensional parameters based on the key feature parameters and calculating multiple correction factors specifically includes: Load transient correction: If a rapid increase in load power is detected, a short-term reduction factor λ_L=1-k×R(t) / U0 is introduced based on the rising slope k and the current internal resistance R(t). At the same time, linear fitting is used in combination with historical load transient data of the same type to predict the load power W(t) at time t. Temperature correction: Based on the real-time battery temperature, the system automatically queries the temperature-capacity curve in the battery specification sheet. At the same time, it combines the real-time battery temperature T(t) and the ambient temperature T_env, and corrects the baseline curve through a second fitting to obtain the dynamic temperature correction coefficient λ_t(t) = a×T(t)² + b×T(t) + c. Internal resistance correction: Calculate λ_R = R0 / R(t) based on the ratio of the real-time estimated internal resistance value R(t) to the internal resistance R0 of the new battery. Voltage trend prediction: By analyzing the rate of decrease of battery voltage k_U through linear regression, when the voltage approaches the preset shutdown voltage threshold U_th and k_U (V / s) exceeds the preset threshold, a prediction time correction coefficient λ_U=1-|k_U| / k_U0 is actively added.

[0012] Furthermore, the formula for calculating the remaining power supply prediction time is as follows: T_y(t)=[(C_1(t)+ C_c(t))*λ_L*λ_R×V(t)]* λ_U / (W(t)*λ_t(t)×φ) In the formula, C_1 is the total available battery capacity Ah of the UPS at time t, output by the dynamic calibration step; C_c(t) is the real-time rechargeable capacity Ah at time t, which is included if the charging module is working, otherwise it is 0; V(t) is the predicted battery voltage V(t) at time t, obtained by voltage trend fitting; W(t) is the predicted load power W(t), output by the load transient correction step; λ_t(t) is the dynamic temperature correction coefficient; and φ is the UPS output power factor.

[0013] Furthermore, the method also includes: Complete parameter data for each discharge process are recorded. The predicted remaining time curve is compared with the actual discharge time curve, and the prediction error E(t) at each time point is calculated. Based on the error E(t), the correction factor is fine-tuned using the gradient descent method.

[0014] Secondly, the present invention provides a wind turbine UPS remaining power prediction system based on multi-parameter correction, comprising: The data acquisition module is used to collect the operating parameters of the UPS system in real time. The operating parameters include at least load parameters, battery parameters and system status parameters. The key feature extraction module is used to extract and preprocess the collected operating parameters to obtain key feature parameters; The battery capacity calibration module is used to perform dynamic calibration of battery capacity based on the key feature parameters and a preset dynamic calibration model for battery capacity. The multi-parameter correction module is used to perform multi-dimensional parameter dynamic correction based on the key feature parameters, calculate multiple correction factors, and predict the load power. The calculation module is used to calculate the remaining power supply prediction time based on the dynamically calibrated battery capacity and the multiple correction factors, combined with the predicted load power.

[0015] Thirdly, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method for predicting the remaining power of a wind turbine UPS based on multi-parameter correction as described above.

[0016] Fourthly, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method for predicting the remaining power of a wind turbine UPS based on multi-parameter correction.

[0017] Compared with the prior art, the present invention has the following beneficial effects: This invention discloses a method and related device for predicting the remaining power of a wind turbine UPS based on multi-parameter correction. By collecting multiple operating parameters such as load, battery, and system status in real time, it comprehensively understands the operating status of the UPS system, providing a rich data foundation for accurate prediction. Key feature parameters are obtained through feature extraction and preprocessing of the collected parameters. Calibration is performed using a preset dynamic battery capacity calibration model, which can promptly correct deviations in battery capacity caused by usage, environment, and other factors, ensuring the accuracy and reliability of capacity data. Then, a correction factor matrix is ​​constructed and calculated to accurately capture the coupled effects of multiple factors. Finally, the dynamically calibrated capacity parameters and multi-dimensional correction factors are fused, combined with the predicted load power, to accurately calculate and output the remaining power supply time. This invention improves prediction accuracy significantly compared to traditional methods by integrating multiple parameters such as load, temperature, internal resistance, charging, and voltage fluctuations, and taking into account load changes, temperature changes, and the influence of internal resistance. It can provide a conservative (shorter) prediction time under heavy load impacts, avoiding misjudgments by administrators, triggering emergency procedures in advance, and ensuring the safety of critical business operations. The model can automatically adjust as the battery ages, maintaining prediction reliability throughout the entire battery lifespan, and exhibiting strong adaptability. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a schematic diagram of the system of the present invention; Figure 3 This is a flowchart illustrating the UPS remaining power prediction process according to an embodiment of the present invention. Detailed Implementation

[0020] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0021] The following detailed description is exemplary and intended to provide further detailed explanation of the invention. Unless otherwise specified, all technical terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this invention is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention.

[0022] See Figure 1 and Figure 3 The embodiments of the present invention disclose A method for predicting the remaining power capacity of a wind turbine UPS based on multi-parameter correction includes the following steps: S1, Real-time acquisition of UPS system operating parameters, including at least load parameters, battery parameters and system status parameters; Real-time acquisition of multi-dimensional parameters during UPS system operation provides comprehensive data support for subsequent modeling. The acquisition frequency is no less than 1Hz to ensure parameter real-time performance. These parameters include, but are not limited to: Load parameters: real-time output total power, non-linear load percentage, load power change rate (power fluctuation value per unit time). Battery parameters: total battery pack voltage, voltage of each individual battery cell / module, real-time discharge current, and battery temperature (multiple temperature measurement points).

[0023] System status parameters: input voltage status (normal / abnormal), charging module operating status, and historical discharge records.

[0024] S2, perform feature extraction and preprocessing on the collected operating parameters to obtain key feature parameters; The collected raw data was cleaned, denoised, and feature-transformed to identify key features that significantly impact power consumption prediction. Details are as follows: Load characteristics: Calculate real-time load rate (current load power / rated power), load fluctuation coefficient (load power standard deviation / average load power), and load transient identification (when the load power change rate exceeds a preset threshold, it is determined to be a load transient).

[0025] Current characteristics: Calculate the average discharge current (mean of sliding window, window size can be configured from 10s to 60s) and the current change trend (fit the slope of the current change over time through linear regression).

[0026] Battery characteristics: The real-time internal resistance of the battery pack is estimated by the correspondence between voltage transients and current changes (ΔV / ΔI), abnormal voltage / current data is eliminated (using the 3σ criterion), and missing cell voltage data is supplemented (based on the mean interpolation method of adjacent cell voltages).

[0027] S3, Based on the key feature parameters, perform dynamic calibration of battery capacity using a preset dynamic calibration model for battery capacity; By combining historical data with real-time parameters, an AI model dynamically corrects the actual usable battery capacity, solving the error problem caused by traditional methods that calculate based on nominal capacity. Data basis: The core sample is the most recent complete discharge record (from full charge to shutdown voltage), combined with discharge segment data from the past 3 months and regular self-check discharge data (once a month, discharge depth 30%). Calibration Model: Construct a battery capacity prediction model based on LSTM (Long Short-Term Memory Network). The input parameters include battery internal resistance, single cell voltage equalization, ambient temperature, number of cycles, and historical discharge capacity decay trend. The dynamic output is the actual maximum usable total capacity C_1(t) of the battery pack at time t and the real-time charging capacity C_c(t). Calibration frequency: Real-time calibration (capacity parameters are updated every 5 minutes) + forced calibration after full discharge (to ensure that the capacity data is consistent with the actual state).

[0028] S4. Based on the key feature parameters, perform multi-dimensional parameter dynamic correction, calculate multiple correction factors, and predict the load power. Dynamic correction: Establish a correction factor matrix / model, and dynamically correct the prediction results through multi-dimensional parameters. The core is to capture the coupled effects of factors such as load transients, temperature, internal resistance, and voltage trends. Load transient correction: If a rapid increase in load power is detected (the rate of change exceeds a preset threshold), a short-term reduction factor λ_L is introduced based on the rise slope k (kW / s) and the current internal resistance R(t). The formula is λ_L=1-k×R(t) / U0 (U0 is the nominal battery voltage) to characterize the impact of the rapid increase in load power. At the same time, linear fitting is used in combination with historical similar load transient data to predict the load power W(t) at time t, so as to achieve a rapid response to the sudden increase in load.

[0029] Temperature Correction: Based on the real-time battery temperature, the system automatically queries the temperature-capacity curve in the battery specification sheet. At the same time, it combines the real-time battery temperature T(t) and the ambient temperature T_env, and corrects the baseline curve through quadratic fitting to obtain the dynamic temperature correction coefficient λ_t(t). The formula is λ_t(t)=a×T(t)²+b×T(t)+c (a, b, and c are calibration coefficients based on battery type), covering the full operating temperature range of -20℃ to 60℃.

[0030] Internal resistance correction: Based on the ratio of the real-time estimated internal resistance value R(t) to the internal resistance R0 of the new battery, calculate λ_R=R0 / R(t). When the internal resistance increases, λ_R decreases. Adjust the usable capacity a second time to correct the problem of shortened discharge time caused by increased internal resistance.

[0031] Voltage trend prediction: The rate of decrease of battery voltage k_U (V / s) is analyzed by linear regression. When the voltage approaches the preset shutdown voltage threshold U_th and k_U (V / s) exceeds the preset threshold, a prediction time correction coefficient λ_U is actively added, λ_U=1-|k_U| / k_U0 (k_U0 is the normal voltage decrease slope threshold), to actively shorten the prediction time and avoid prediction inaccuracies caused by sudden voltage drops.

[0032] S5. Based on the dynamically calibrated battery capacity and the multiple correction factors, combined with the predicted load power, calculate the remaining power supply prediction time.

[0033] Based on dynamically calibrated capacity parameters and multi-dimensional correction factors, combined with load power prediction values, the final remaining power supply prediction time is calculated using the following formula: T_y(t)=[(C_1(t)+ C_c (t))*λ_L*λ_R×V(t)]* λ_U / (W(t)*λ_t(t)×φ) in: C_1: Total available UPS battery capacity (Ah) at time t, output by the dynamic calibration process; C_c(t): Real-time rechargeable capacity Ah at time t. If the charging module is working, it is included; otherwise, it is 0. V(t): The predicted battery voltage V(t) at time t, obtained by fitting the voltage trend; W(t): Predicted load power W(t), output from the load transient correction step; λ_t(t): Dynamic temperature correction coefficient φ: UPS output power factor. In one feasible embodiment of the present invention, it further includes: The S6 features a built-in self-learning module that continuously optimizes the correction factor model by comparing actual discharge data with predicted data, thereby enhancing its personalized adaptation capabilities. S601, Data Logging: Completely records all parameter data for each discharge process (from the start of discharge to the UPS shutting down due to low voltage), including real-time total output power, voltage of each individual battery, discharge current, battery temperature, ambient temperature, actual discharge time, etc. S602, Error Analysis: Compare the predicted remaining time curve with the actual discharge time curve, and calculate the prediction error at each time point: E(t) = |Ty(t) Tactual(t)∣ / Tactual(t) S603, Model Fine-tuning: Based on error data, the gradient descent method is used to fine-tune the coefficients in the correction factor matrix (such as the calculated coefficients of λ_L and λ_t(t)) to adapt the model to the battery aging characteristics and load fluctuation patterns of the UPS system, and realize the closed-loop iteration of "prediction-verification-optimization".

[0034] See Figure 2 This invention discloses a wind turbine UPS remaining power prediction system based on multi-parameter correction, including a data acquisition module, a key feature extraction module, a battery capacity calibration module, a multi-parameter correction module, and a calculation module. As the data source foundation for the entire prediction process, it collects three core parameters: load, battery, and system status. The acquisition frequency is no less than 1Hz to ensure real-time performance, providing comprehensive and accurate data support for subsequent feature processing and model building.

[0035] The data acquisition module is used to collect the operating parameters of the UPS system in real time. The operating parameters include at least load parameters, battery parameters and system status parameters. The key feature extraction module is used to extract and preprocess the collected operating parameters to obtain key feature parameters. Specifically, it cleans, denoises, and transforms the collected raw data to extract key and effective features such as load rate, current change trend, and real-time internal resistance, removes abnormal data and fills in missing data to improve data quality and provide reliable input for capacity calibration and fusion prediction.

[0036] The battery capacity calibration module is used to perform dynamic calibration of battery capacity based on the key feature parameters and through a preset battery capacity dynamic calibration model. Specifically, based on the LSTM model, combined with historical discharge data and real-time parameters, it dynamically outputs the actual maximum usable total battery capacity C_1(t) and real-time charging capacity C_c(t) at time t, thus solving the error problem caused by the traditional method of calculating based on nominal capacity.

[0037] The multi-parameter correction module is used to perform multi-dimensional parameter dynamic correction based on the key feature parameters, calculate multiple correction factors, and predict the load power. Specifically, based on the preprocessed feature data, a correction factor matrix composed of load transient correction factor λ_L, temperature correction factor λ_t, internal resistance correction factor λ_R, and voltage trend correction factor λ_U is constructed and calculated to accurately capture the coupled effects of multiple factors.

[0038] The calculation module is used to calculate the remaining power supply prediction time based on the dynamically calibrated battery capacity and the multiple correction factors, combined with the predicted load power. Preferably, a model self-correction module is added as the core link of closed-loop optimization. This module records the actual discharge full parameter data, compares the predicted values ​​with the actual values ​​and calculates the error, and uses the gradient descent method to fine-tune the correction factor coefficients. These adjustments are then fed back to the multi-parameter fusion prediction and battery capacity dynamic calibration stages to achieve continuous iterative optimization of the model and ensure long-term prediction reliability.

[0039] In one embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used in the operation of a wind turbine UPS remaining power prediction method based on multi-parameter correction.

[0040] This invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the method for predicting the remaining power of a wind turbine UPS based on multi-parameter correction in the above embodiments.

[0041] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0042] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0043] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0044] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for predicting the remaining power capacity of a wind turbine UPS based on multi-parameter correction, characterized in that, Includes the following steps: Real-time acquisition of UPS system operating parameters, including at least load parameters, battery parameters, and system status parameters; The collected operating parameters are subjected to feature extraction and preprocessing to obtain key feature parameters; Based on the aforementioned key feature parameters, dynamic calibration of battery capacity is performed using a preset dynamic calibration model for battery capacity. Based on the key feature parameters, multi-dimensional parameter dynamic correction is performed, multiple correction factors are calculated, and load power is predicted. Based on the dynamically calibrated battery capacity and the multiple correction factors, combined with the predicted load power, the remaining power supply prediction time is calculated.

2. The method for predicting the remaining power of a wind turbine UPS based on multi-parameter correction according to claim 1, characterized in that, The load parameters include real-time total output power, nonlinear load percentage, and load power change rate; the battery parameters include total battery pack voltage, voltage of each individual battery cell / module, real-time discharge current, and battery temperature; the system status parameters include input voltage status, charging module operating status, and historical discharge records.

3. The method for predicting the remaining power of a wind turbine UPS based on multi-parameter correction according to claim 1, characterized in that, The step of extracting and preprocessing the collected operating parameters to obtain key feature parameters specifically includes: The collected operating parameters are cleaned, denoised, and transformed to identify key features that have a significant impact on power prediction. These key features include load features, current features, and battery features. For load characteristics, calculate real-time load rate and load fluctuation coefficient; identify load transient events; For the current characteristics, calculate the average discharge current and the current variation trend; For battery characteristics, the real-time internal resistance of the battery pack is estimated by the correspondence between voltage transients and current changes, abnormal voltage / current data is eliminated, and missing individual cell voltage data is supplemented.

4. The method for predicting the remaining power of a wind turbine UPS based on multi-parameter correction according to claim 1, characterized in that, The step of dynamically calibrating the battery capacity based on the key feature parameters and using a preset battery capacity dynamic calibration model specifically includes: The most recent complete discharge record is used as the core sample, combined with recent discharge fragment data and periodic self-check discharge data; A battery capacity prediction model based on LSTM is constructed. The input parameters include battery internal resistance, single cell voltage equalization, ambient temperature, number of cycles, and historical discharge capacity decay trend. The model dynamically outputs the actual maximum usable total capacity C_1(t) and real-time charging capacity C_c(t) of the battery pack at time t.

5. The method for predicting the remaining power of a wind turbine UPS based on multi-parameter correction according to claim 1, characterized in that, The step of dynamically correcting multi-dimensional parameters based on the key feature parameters and calculating multiple correction factors specifically includes: Load transient correction: If a rapid increase in load power is detected, a short-term reduction factor λ_L=1-k×R(t) / U0 is introduced based on the rising slope k and the current internal resistance R(t). At the same time, linear fitting is used in combination with historical load transient data of the same type to predict the load power W(t) at time t. Temperature correction: Based on the real-time battery temperature, the system automatically queries the temperature-capacity curve in the battery specification sheet. At the same time, it combines the real-time battery temperature T(t) and the ambient temperature T_env, and corrects the baseline curve through a second fitting to obtain the dynamic temperature correction coefficient λ_t(t) = a×T(t)² + b×T(t) + c. Internal resistance correction: Calculate λ_R = R0 / R(t) based on the ratio of the real-time estimated internal resistance value R(t) to the internal resistance R0 of the new battery. Voltage trend prediction: By analyzing the rate of decrease of battery voltage k_U through linear regression, when the voltage approaches the preset shutdown voltage threshold U_th and k_U (V / s) exceeds the preset threshold, a prediction time correction coefficient λ_U=1-|k_U| / k_U0 is actively added.

6. The method for predicting the remaining power of a wind turbine UPS based on multi-parameter correction according to claim 1, characterized in that, The formula for calculating the remaining power supply prediction time is as follows: T_y(t)=[(C_1(t)+ C_c(t))*λ_L*λ_R×V(t)]* λ_U / (W(t)*λ_t(t)×φ) In the formula, C_1 is the total available battery capacity Ah of the UPS at time t, output by the dynamic calibration step; C_c(t) is the real-time rechargeable capacity Ah at time t, which is included if the charging module is working, otherwise it is 0; V(t) is the predicted battery voltage V(t) at time t, obtained by voltage trend fitting; W(t) is the predicted load power W(t), output by the load transient correction step; λ_t(t) is the dynamic temperature correction coefficient; and φ is the UPS output power factor.

7. The method for predicting the remaining power of a wind turbine UPS based on multi-parameter correction according to claim 1, characterized in that, Also includes: Complete parameter data for each discharge process are recorded. The predicted remaining time curve is compared with the actual discharge time curve, and the prediction error E(t) at each time point is calculated. Based on the error E(t), the correction factor is fine-tuned using the gradient descent method.

8. A wind turbine UPS remaining power prediction system based on multi-parameter correction, characterized in that, include: The data acquisition module is used to collect the operating parameters of the UPS system in real time. The operating parameters include at least load parameters, battery parameters and system status parameters. The key feature extraction module is used to extract and preprocess the collected operating parameters to obtain key feature parameters; The battery capacity calibration module is used to perform dynamic calibration of battery capacity based on the key feature parameters and a preset dynamic calibration model for battery capacity. The multi-parameter correction module is used to perform multi-dimensional parameter dynamic correction based on the key feature parameters, calculate multiple correction factors, and predict the load power. The calculation module is used to calculate the remaining power supply prediction time based on the dynamically calibrated battery capacity and the multiple correction factors, combined with the predicted load power.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for predicting the remaining power of a wind turbine UPS based on multi-parameter correction as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for predicting the remaining power of a wind turbine UPS based on multi-parameter correction as described in any one of claims 1-7.

Images