New energy facility intelligent operation and maintenance and adaptive control method

Abstract

The invention discloses an intelligent operation and maintenance and adaptive control method for new energy facilities, and the method achieves the collaborative optimization scheduling of multi-source equipment, reduces the start-stop frequency of a conventional generator set, and improves the charging and discharging efficiency of an energy storage system. An equipment health monitoring mechanism reduces the risk of sudden failures, and a self-adaptive control strategy enhances the stability of the system under external disturbance. Wind and light power generation prediction and real-time scheduling are combined, the utilization rate of renewable energy sources is improved, meanwhile, the overall operation cost is reduced through data-driven fine control, fine control over the charging and discharging process of the energy storage equipment is achieved, the system power adjustment requirement and the battery health protection requirement are effectively balanced, and the system reliability is improved. Energy conversion in a low-efficiency period is avoided in charge and discharge opportunity selection, efficiency loss caused by a battery polarization effect is reduced in charge and discharge rate control, and meanwhile, the risks of over-charge, over-discharge and overheating of the battery are reduced through a real-time monitoring and dynamic adjustment mechanism.

Description

Technical Field

[0001] This application relates to the field of operation and maintenance technology, and in particular to a method for intelligent operation and maintenance and adaptive control of new energy facilities. Background Technology

[0002] With the accelerated pace of global energy transition, the new energy industry has become a crucial direction for energy sector development, playing a key role in addressing climate change and achieving carbon neutrality. Within the new energy industry, the new energy power system, as the core infrastructure supporting the large-scale grid connection of renewable energy sources such as wind and solar power, directly impacts the success or failure of the energy transition in terms of stability and economic viability. New energy power systems typically consist of various heterogeneous devices, including wind turbines, photovoltaic arrays, energy storage systems, and traditional generator sets. These devices exhibit significant differences in operating characteristics, response speeds, and control methods. Achieving coordinated and optimized scheduling of multiple types of equipment while ensuring healthy equipment operation and system supply-demand balance has become a core challenge in the operation and maintenance management of new energy facilities.

[0003] Current practices in the operation and maintenance (O&M) and control of new energy facilities have several shortcomings. First, in the coordinated scheduling of multi-source heterogeneous equipment, traditional methods often employ rule-based scheduling strategies, which struggle to cope with the strong volatility and uncertainty of wind and solar power generation. This results in low renewable energy utilization rates, frequent start-ups and shutdowns of traditional generator sets, and increased operating costs and equipment wear. Second, in the charging and discharging management of energy storage systems, there is a lack of refined control over the timing, rate, and depth of charging and discharging, easily leading to overcharging and discharging. This not only reduces the conversion efficiency of the energy storage system but also accelerates battery degradation and shortens equipment lifespan. Third, existing O&M models are mostly reactive fault responses, lacking proactive early warning mechanisms based on equipment health status. This makes it difficult to predict and intervene before faults occur, leading to a high risk of sudden outages. Furthermore, the adaptive capability of system control strategies is insufficient. When faced with external disturbances such as sudden load changes and drastic weather variations, the response speed is slow and the adjustment accuracy is low, easily causing supply-demand imbalances and grid fluctuations, affecting the safe and stable operation of the system. Summary of the Invention

[0004] The main objective of this application is to provide an intelligent operation and maintenance and adaptive control method for new energy facilities to solve the problems mentioned in the background.

[0005] To achieve the above objectives, this application provides the following technical solution: A method for intelligent operation and maintenance and adaptive control of new energy facilities, comprising the following specific steps: S1. Data Acquisition and Monitoring: Deploy sensors and monitoring equipment in new energy facilities to acquire data on grid load, wind speed, solar radiation, energy storage system charging status, and equipment operating health status, and transmit the data to the central control system through Internet of Things (IoT) technology. S2. Data cleaning and preprocessing: Denoising, outlier removal, and missing value imputation are performed on the raw data to ensure data accuracy, consistency, and integrity. Data from different sources are also standardized and fused. S3. Demand and power generation forecasting: Based on historical data and real-time information, forecasting models are used to predict future electricity demand and the power generation of wind and solar power, taking into account external factors such as weather changes and seasonal fluctuations. S4. System-level scheduling optimization: Based on real-time data and power generation demand forecasts, optimization algorithms are used to schedule the operation of wind power, photovoltaic, energy storage systems and traditional generator sets, giving priority to the use of renewable energy, reducing the start-up and shutdown frequency of traditional generator sets, and rationally scheduling energy storage devices. S5. Energy storage device charging and discharging control optimization: Based on system-level scheduling optimization, the timing, rate and depth of charging and discharging of energy storage devices are adjusted to ensure efficient charging and discharging of energy storage devices and reduce efficiency loss and equipment damage caused by overcharging and discharging. S6. Equipment health monitoring: Deploys intelligent sensors and uses machine learning algorithms to monitor the operating status of energy storage systems and traditional generator sets, assesses equipment health status, and promptly detects potential faults and issues early warnings. S7. Adaptive control adjusts the system control strategy based on real-time operating data and feedback information to ensure stable system operation and respond quickly to external disturbances, maintaining supply and demand balance and system stability.

[0006] Preferably, the data acquisition and monitoring in step S1 are carried out in the following manner: S1.1 Deploy wind speed sensors, solar radiation sensors, voltage sensors, current sensors, energy storage system charging status sensors, and equipment health monitoring sensors in new energy facilities; S1.2 Collect sensor data through Internet of Things (IoT) technology, including grid load, wind speed, solar radiation, energy storage system charging status and equipment health status, and transmit the data to the central control system.

[0007] Preferably, the data cleaning and preprocessing in step S2 are carried out in the following manner: S2.1. The original data is denoised using a preset digital filtering algorithm. Anomalies are identified and removed from the denoised data based on a preset statistical anomaly detection method. Data is filled in using a preset missing value imputation method. S2.2. Use a preset standardization or normalization method to unify the dimensions of data of different physical quantities and map them to a unified numerical range. Perform time alignment processing on data with different sampling frequencies and use a resampling method to make the time resolution of each data source consistent. Use a preset multi-source data fusion technology to fuse the standardized and time-aligned data and construct a comprehensive dataset.

[0008] Preferably, the demand and power generation forecasting in step S3 is carried out in the following manner: S3.1 Extract time characteristic parameters from historical load data from the comprehensive dataset. The time characteristic parameters include hourly load fluctuation characteristics, daily load change characteristics, weekly load cycle characteristics, and monthly load trend characteristics. Obtain real-time meteorological data and date attribute information as input variables for external influencing factors. Real-time meteorological data includes ambient temperature data, relative humidity data, and wind speed data. Date attribute information includes weekday type identifier, holiday identifier, and season identifier. Use a preset prediction algorithm to input the time characteristic parameters and external influencing factor input variables into the power demand prediction model and generate power demand prediction curves for future prediction periods. S3.2 Extract historical wind speed data, historical wind direction data, and historical power output data of wind turbine generators; obtain real-time and forecasted wind speed parameters, wind direction parameters, and atmospheric pressure parameters; establish a conversion relationship model between wind speed and power and generate wind power output curves for future periods; extract historical solar radiation intensity data, historical ambient temperature data, and historical power output data of photovoltaic arrays; obtain real-time and forecasted solar radiation intensity parameters, cloud cover parameters, and ambient temperature parameters; establish a conversion relationship model between solar radiation and power and generate photovoltaic power output curves for future periods; superimpose the wind power output curves and photovoltaic power output curves to generate predicted data for total renewable energy power generation.

[0009] Preferably, the system-level scheduling optimization in step S4 is implemented as follows: S4.1 Determine the scheduling optimization objective function, which includes maximizing renewable energy utilization, minimizing system operating costs, and minimizing carbon emissions. Set the system operation constraint set, which includes power supply and demand balance constraints, upper and lower limits of wind turbine output, upper and lower limits of photovoltaic array output, upper and lower limits of energy storage system state of charge, upper and lower limits of energy storage system charging and discharging power, upper and lower limits of traditional generator output, traditional generator ramp rate constraints, and grid frequency and voltage safe operation constraints. Establish renewable energy priority scheduling rules, construct energy storage system auxiliary scheduling strategies, calculate power deficit or power surplus based on the deviation between renewable energy output curve and power demand forecast curve, and determine the charging period, discharging period, and charging and discharging power allocation scheme of energy storage system. S4.2. Solve the multi-source collaborative scheduling optimization model using a preset optimization algorithm and generate a system scheduling scheme, including: numerically solving the multi-objective optimization function through the optimization algorithm and obtaining the optimal scheduling solution under the premise of satisfying the system operation constraint set, generating a system scheduling scheme. The system scheduling scheme includes power scheduling instructions for wind turbine generators, power scheduling instructions for photovoltaic arrays, charging and discharging scheduling instructions for energy storage systems, and start-up and shutdown scheduling instructions and output scheduling instructions for traditional generators. The energy storage system smooths out short-term power fluctuations and reduces the number of starts and stops of traditional generators during the scheduling cycle. The system scheduling scheme is then distributed to the wind turbine generator control unit, photovoltaic array control unit, energy storage system control unit, and traditional generator control unit through the communication network.

[0010] Preferably, the optimization of the energy storage device's charging and discharging control in step S5 is as follows: S5.1. Obtain the current state of charge (SOC) value, battery temperature value, internal resistance value, and historical cycle count data of the energy storage system. Calculate the charging and discharging timing parameters based on the power magnitude and duration of the power supply and demand deficit or surplus. The charging and discharging timing parameters include the charging start time, charging end time, discharging start time, and discharging end time. Calculate the charging and discharging rate parameters based on the preset battery characteristic curve and the current SOC value. Set the upper and lower threshold values ​​of the SOC as charging and discharging depth limit parameters. S5.2 Generate energy storage charging and discharging control commands that include charging and discharging timing parameters, charging and discharging power setpoints, and target values ​​for the state of charge. Send the energy storage charging and discharging control commands to the energy storage system control unit through the communication network. During the charging and discharging process, collect actual charging and discharging power data, state of charge change rate data, battery temperature change data, and voltage and current response data. When the battery temperature exceeds the preset temperature safety threshold, the state of charge approaches the boundary value, or the charging and discharging efficiency is lower than the preset efficiency lower limit, adjust the charging and discharging power value or terminate the charging and discharging process.

[0011] Preferably, the equipment health monitoring in step S6 is carried out in the following manner: S6.1 Collect battery cell voltage data, battery temperature data, charge / discharge current data, internal resistance change rate data, cycle count data, and capacity decay rate data of the energy storage system; collect bearing vibration signal data, lubricating oil temperature data, cooling water temperature data, exhaust temperature data, speed fluctuation data, and power output data of the traditional generator set; perform signal processing on the collected operating status data and extract feature parameters, including time-domain feature parameters and frequency-domain feature parameters. Time-domain feature parameters include mean, variance, peak value, and crest factor; frequency-domain feature parameters include spectral energy distribution, characteristic frequency amplitude, and harmonic components. Combine the operating status data and feature parameters to form an equipment health feature vector. S6.2. Train the equipment health assessment model using historical normal operation data and historical fault data. Input the real-time collected equipment health feature vector into the equipment health assessment model to obtain the equipment health index. The equipment health index is a value from 0 to 100. Set health warning threshold and fault warning threshold. When the equipment health index is lower than the health warning threshold, trigger the equipment degradation warning signal. When the equipment health index is lower than the fault warning threshold, trigger the equipment fault warning signal. Transmit the warning signal to the operation and maintenance terminal through the central control system and output the warning report. The warning report includes equipment identification information, equipment health index value, abnormal characteristic parameters and maintenance suggestions.

[0012] Preferably, the adaptive control in step S7 is implemented in the following manner: S7.1 Acquire real-time system operating data, including actual output data of wind turbine generators, actual output data of photovoltaic arrays, actual output data of traditional generators, actual charging and discharging power data of energy storage systems, real-time grid load data, system frequency data, and system voltage data. Compare the real-time system operating data with the target operating parameters in the system scheduling scheme to calculate and obtain the deviation, including power deviation, frequency deviation, and voltage deviation. Identify external disturbance events by monitoring the rate of change of data. External disturbance events include load change events, renewable energy output change events, equipment failure events, or grid failure events. Set allowable deviation thresholds, including allowable power deviation thresholds, allowable frequency deviation thresholds, and allowable voltage deviation thresholds. When the deviation exceeds the allowable deviation threshold or a major disturbance event of a preset level is detected, a control strategy adjustment trigger signal is generated. S7.2. Based on the magnitude of the deviation and the trend of its time series change, a preset adaptive control algorithm is used to calculate the control strategy adjustment scheme. The control strategy adjustment scheme includes the output adjustment of the wind turbine generator set, the output adjustment of the photovoltaic array, the output adjustment of the traditional generator set, the charging and discharging power adjustment of the energy storage system, and the load-side response adjustment. Rapid response measures are formulated for different types of external disturbance events. Adjusted equipment control commands are generated and sent to the wind turbine generator set control unit, the photovoltaic array control unit, the energy storage system control unit, and the traditional generator set control unit through the communication network. The adjusted system operating status is continuously monitored. When the deviation of the system operating status is lower than the allowable deviation threshold, the current control strategy is maintained. When the deviation of the system operating status still exceeds the allowable deviation threshold, the control strategy adjustment calculation process is repeated and a new control strategy adjustment scheme is generated.

[0013] Compared with the prior art, the beneficial effects of the present invention are: 1. This application achieves coordinated and optimized scheduling of multi-source equipment, reducing the start-stop frequency of traditional generator sets and improving the charging and discharging efficiency of energy storage systems. Equipment health monitoring mechanisms reduce the risk of sudden failures, and adaptive control strategies enhance system stability under external disturbances. The combination of wind and solar power generation forecasting and real-time scheduling improves the utilization rate of renewable energy, while data-driven refined control reduces overall operating costs.

[0014] 2. This application can comprehensively acquire diverse data required for the operation of new energy facilities, such as simultaneously collecting environmental wind speed and equipment vibration data, providing data support for power generation prediction and equipment health assessment. It achieves efficient data transmission through IoT technology, for example, enabling real-time data transmission from thousands of sensors in large photovoltaic power plants, avoiding control command delays caused by data transmission latency. This solution constructs a complete data acquisition system, laying a reliable data foundation for subsequent data processing, predictive analysis, and optimized control. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the steps of the method described in this application. Detailed Implementation

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0017] The terms "first," "second," and "third" in this application are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first," "second," or "third" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. All directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of this application are only used to explain the relative positional relationships and movements between components in a specific orientation (as shown in the figures). If the specific orientation changes, the directional indications also change accordingly. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0018] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0019] Example 1: Please refer to Figure 1 A method for intelligent operation and maintenance and adaptive control of new energy facilities, the specific steps of which are as follows: S1. Data Acquisition and Monitoring: Deploy sensors and monitoring equipment in new energy facilities to acquire data on grid load, wind speed, solar radiation, energy storage system charging status, and equipment operating health status, and transmit the data to the central control system through Internet of Things (IoT) technology. S2. Data cleaning and preprocessing: Denoising, outlier removal, and missing value imputation are performed on the raw data to ensure data accuracy, consistency, and integrity. Data from different sources are also standardized and fused. S3. Demand and power generation forecasting: Based on historical data and real-time information, forecasting models are used to predict future electricity demand and the power generation of wind and solar power, taking into account external factors such as weather changes and seasonal fluctuations. S4. System-level scheduling optimization: Based on real-time data and power generation demand forecasts, optimization algorithms are used to schedule the operation of wind power, photovoltaic, energy storage systems and traditional generator sets, giving priority to the use of renewable energy, reducing the start-up and shutdown frequency of traditional generator sets, and rationally scheduling energy storage devices. S5. Energy storage device charging and discharging control optimization: Based on system-level scheduling optimization, the timing, rate and depth of charging and discharging of energy storage devices are adjusted to ensure efficient charging and discharging of energy storage devices and reduce efficiency loss and equipment damage caused by overcharging and discharging. S6. Equipment health monitoring: Deploys intelligent sensors and uses machine learning algorithms to monitor the operating status of energy storage systems and traditional generator sets, assesses equipment health status, and promptly detects potential faults and issues early warnings. S7. Adaptive control adjusts the system control strategy based on real-time operating data and feedback information to ensure stable system operation and respond quickly to external disturbances, maintaining supply and demand balance and system stability.

[0020] In this embodiment: In the prior art, new energy power systems consist of various heterogeneous devices with significantly different operating characteristics. Traditional methods, employing experience-based rule-based scheduling strategies, struggle to cope with the volatility and uncertainty of wind and solar power generation, resulting in low renewable energy utilization rates, frequent start-ups and shutdowns of traditional generator units, and increased operating costs. Energy storage system charging and discharging management lacks refined control, easily leading to overcharging and discharging, reducing conversion efficiency and accelerating equipment wear. Existing operation and maintenance models are mostly reactive fault responses, lacking equipment health status early warning mechanisms, resulting in a high risk of sudden outages. System control strategies lack adaptive capabilities, responding slowly to sudden load changes or drastic weather variations, easily causing supply-demand imbalances and grid fluctuations.

[0021] To address the aforementioned issues, achieving coordinated scheduling of multiple types of equipment becomes crucial. Considering the volatility of wind and solar power generation, predictive models are needed to anticipate power generation and demand. Regarding energy storage efficiency, charging and discharging parameters need to be optimized to balance power fluctuations. For equipment health monitoring, potential faults can be identified through real-time data acquisition and machine learning algorithms. In terms of system stability, adjusting control strategies based on real-time feedback can improve response speed. Thus, a technical solution combining data acquisition, predictive analysis, optimized scheduling, and adaptive control is gradually being developed.

[0022] Therefore, this application proposes methods for data acquisition and monitoring, data cleaning and preprocessing, demand and power generation forecasting, system-level scheduling optimization, energy storage device charging and discharging control optimization, equipment health monitoring, and adaptive control. Specifically, this includes: deploying sensors in renewable energy facilities to acquire grid load, wind speed, solar radiation, energy storage status, and equipment health data, and transmitting this data to a central system via the Internet of Things; denoising, outlier processing, and missing value imputation of the data; standardizing and fusing multi-source data; forecasting power demand and power generation based on historical data and real-time information; employing optimization algorithms to schedule the operation of wind power, photovoltaic, energy storage, and traditional generating units; adjusting energy storage charging and discharging parameters to optimize efficiency; monitoring equipment health status through sensors and machine learning; and adjusting control strategies based on real-time data and feedback to maintain system stability.

[0023] Data cleaning and preprocessing refers to denoising, outlier removal, and missing value imputation of raw data. This can be achieved using digital filtering algorithms and statistical anomaly detection methods to ensure data accuracy and consistency. Demand and power generation forecasting involves building predictive models based on historical data and real-time meteorological information. This can be implemented using time series analysis or machine learning algorithms to guide system scheduling. System-level scheduling optimization involves coordinating the operation of multi-source equipment through multi-objective optimization algorithms. This can be implemented using linear programming or genetic algorithms to prioritize the use of renewable energy. Energy storage device charge and discharge control optimization involves adjusting charge and discharge parameters according to battery status and power demand. This can be achieved using state-of-charge thresholds and charge / discharge rate limits to reduce equipment wear. Equipment health monitoring involves collecting operating parameters and assessing equipment status through sensors. This can be achieved using vibration signal analysis and temperature monitoring to provide early warnings of potential faults. Adaptive control involves adjusting control commands based on real-time deviations. This can be achieved using proportional-integral-derivative algorithms or model predictive control to quickly respond to external disturbances.

[0024] The sensor network collects real-time data on grid load, environmental parameters, and equipment operation, transmitting it to the central system via the Internet of Things (IoT). The data cleaning module removes noise through filtering and anomaly detection, fills in missing values, and generates a standardized dataset. The prediction module combines historical load characteristics and weather forecast data to generate future electricity demand curves and wind and solar power output forecasts. The optimization scheduling module calculates the output plan for each device with the goal of maximizing renewable energy utilization and smooths short-term fluctuations through energy storage systems. The energy storage control module dynamically adjusts charging and discharging power based on battery status to avoid overcharging and discharging. The health monitoring module extracts characteristic parameters such as equipment vibration and temperature, assesses health status using machine learning models, and triggers early warnings. The adaptive control module monitors system frequency and voltage deviations in real time, quickly adjusting equipment output when disturbances are detected to maintain supply and demand balance.

[0025] Compared to existing technologies, traditional methods rely on fixed-rule scheduling, which is difficult to adapt to the randomness of wind and solar power generation. This solution, however, dynamically adjusts the scheduling strategy through predictive models and optimization algorithms, improving the renewable energy absorption capacity. Existing energy storage control often uses fixed charge and discharge thresholds; this solution optimizes charge and discharge parameters based on battery status and system requirements, extending equipment lifespan. Passive operation and maintenance cannot predict equipment failures; this solution achieves proactive early warning through multi-dimensional feature extraction and health assessment. Conventional control strategies are often lagging in adjustment; this solution improves the system's resilience to disturbances based on real-time data feedback and adaptive algorithms.

[0026] Through the above technical solutions, this application achieves coordinated and optimized scheduling of multi-source equipment, reduces the start-stop frequency of traditional generator sets, and improves the charging and discharging efficiency of energy storage systems. Equipment health monitoring mechanisms reduce the risk of sudden failures, and adaptive control strategies enhance system stability under external disturbances. The combination of wind and solar power generation forecasting and real-time scheduling improves the utilization rate of renewable energy, while data-driven refined control reduces overall operating costs.

[0027] Example 2: Please refer to Figure 1 The data collection and monitoring in step S1 are carried out in the following ways: S1.1 Deploy wind speed sensors, solar radiation sensors, voltage sensors, current sensors, energy storage system charging status sensors, and equipment health monitoring sensors in new energy facilities; S1.2 Collect sensor data through Internet of Things (IoT) technology, including grid load, wind speed, solar radiation, energy storage system charging status and equipment health status, and transmit the data to the central control system.

[0028] In this embodiment: This application further proposes to deploy wind speed sensors, solar radiation sensors, voltage sensors, current sensors, energy storage system charging status sensors, and equipment health monitoring sensors in new energy facilities, collect sensor data through Internet of Things technology, including grid load, wind speed, solar radiation, energy storage system charging status, and equipment health status, and transmit the data to the central control system.

[0029] Wind speed sensors are devices used to measure the wind speed in the environment where wind turbine generators are located. They can be implemented using ultrasonic wind speed sensors or mechanical anemometers, and their function is to provide basic data for wind power output prediction. Solar radiation sensors are devices used to monitor the intensity of solar radiation received by photovoltaic arrays. They can be implemented using thermopile-type radiation sensors or photoelectric radiation sensors, and their function is to provide input parameters for photovoltaic power output prediction. Voltage and current sensors are devices used to measure voltage and current signals at power grid nodes. They can be implemented using Hall effect sensors or resistive voltage divider sensors, and their function is to provide real-time electrical measurement data for power grid load monitoring and power supply and demand balance analysis. Energy storage system charging status sensors are devices used to detect the state of charge of energy storage batteries. They can be implemented using coulomb metering chips or open-circuit voltage-based estimation modules, and their function is to provide status feedback for energy storage system charging and discharging control. Equipment health monitoring sensors are a combination of sensors used to collect equipment operating status parameters. They can be implemented using vibration sensors, temperature sensors, and acoustic emission sensors, and their function is to provide multi-dimensional data input for equipment health assessment. Internet of Things (IoT) technology refers to a data transmission architecture based on wireless communication protocols, which can be implemented using LoRa, NB-IoT, or 5G communication modules. Its function is to enable real-time acquisition and remote transmission of sensor data.

[0030] The aforementioned sensor arrays are installed at key locations in new energy facilities. For example, wind speed sensors are deployed on the top of wind turbine towers, solar radiation sensors are installed on the surface of photovoltaic arrays, voltage and current sensors are configured at grid connection points, charging status sensors are integrated inside energy storage battery packs, and vibration sensors are placed at the generator bearings. These sensors continuously collect data at preset sampling frequencies; for example, wind speed sensors collect data every 5 seconds, and solar radiation sensors collect data every minute. Multi-source heterogeneous data is uploaded to the central control system via an IoT communication module. For example, the data is initially encapsulated using an edge computing gateway and then transmitted to a cloud server via a 5G network. After receiving the data, the central control system stores it in a unified format, such as storing wind speed data as a structured data table associated with timestamps and equipment numbers, providing input for subsequent data cleaning, prediction models, and scheduling algorithms.

[0031] Compared to existing technologies, traditional methods typically deploy only a single type of sensor, such as collecting only voltage and current signals while ignoring device health parameters, resulting in insufficient data dimensions. Existing technologies often use wired data transmission methods, such as RS-485 bus connections, which suffer from complex wiring and poor scalability. This solution, through the collaborative deployment of multiple types of sensors, achieves the synchronous acquisition of power grid operating parameters, environmental parameters, and device status parameters. Simultaneously, it employs wireless IoT technology to construct a flexible data transmission network, solving the applicability problem of traditional wired transmission methods in the decentralized deployment scenarios of new energy facilities.

[0032] Through the above technical solution, this application can comprehensively acquire diverse data required for the operation of new energy facilities, such as simultaneously collecting environmental wind speed and equipment vibration data, providing data support for power generation prediction and equipment health assessment. Efficient data transmission is achieved through Internet of Things (IoT) technology, for example, enabling real-time data transmission from thousands of sensors in large photovoltaic power plants, avoiding control command delays caused by data transmission latency. This solution constructs a complete data acquisition system, laying a reliable data foundation for subsequent data processing, predictive analysis, and optimized control.

[0033] Example 3: Please refer to Figure 1 The data cleaning and preprocessing in step S2 are carried out in the following ways: S2.1. The original data is denoised using a preset digital filtering algorithm. Anomalies are identified and removed from the denoised data based on a preset statistical anomaly detection method. Data is filled in using a preset missing value imputation method. S2.2. Use a preset standardization or normalization method to unify the dimensions of data of different physical quantities and map them to a unified numerical range. Perform time alignment processing on data with different sampling frequencies and use a resampling method to make the time resolution of each data source consistent. Use a preset multi-source data fusion technology to fuse the standardized and time-aligned data and construct a comprehensive dataset.

[0034] In this embodiment: This application further proposes specific methods for data cleaning and preprocessing, including using digital filtering algorithms to denoise the original data, using statistical anomaly detection methods to identify and remove outliers from the denoised data, using missing value imputation methods to complete the data for missing data, using standardization or normalization methods to unify the dimensions of data of different physical quantities and map them to a unified numerical range, using time alignment processing for data of different sampling frequencies and using resampling methods to make the time resolution of each data source consistent, and using multi-source data fusion technology to fuse the standardized and time-aligned data and construct a comprehensive dataset.

[0035] Digital filtering algorithms remove high-frequency noise components from signals using preset filter parameters. This can be achieved using low-pass filters or moving average filters, and is used to eliminate random interference signals in sensor-acquired data. Statistical anomaly detection methods identify data points deviating from the normal range based on probability distribution models. This can be achieved using the three-sigma rule or box plot analysis, and is used to exclude abnormal data caused by equipment failure or transmission errors. Missing value imputation methods fill in missing values ​​caused by data acquisition interruptions using interpolation. This can be achieved using linear interpolation or time series forecasting, and is used to ensure the continuity of the data sequence. Standardization or normalization methods convert physical quantities with different dimensions into uniform dimensionless values. This can be achieved using Z-score standardization or Min-Max normalization, and is used to eliminate dimensional differences between multi-source data. Time alignment matching involves matching timestamps of data with different sampling frequencies. This can be achieved using linear interpolation resampling or nearest-neighbor interpolation, and is used to unify the time reference of multi-source data. Multi-source data fusion technology refers to the association and integration of heterogeneous data that has been standardized and time-aligned. Specifically, it can be achieved by using data cube construction technology or feature-level fusion algorithms to form a comprehensive dataset with spatiotemporal consistency.

[0036] The data cleaning stage first filters out high-frequency noise from sensor signals using digital filters, such as a Butterworth low-pass filter with a cutoff frequency of 5Hz to eliminate electromagnetic interference noise. Then, the three-sigma rule is used to detect anomalies in each data channel, marking data points exceeding three standard deviations of the mean as outliers and removing them. For time periods with missing data, time-series-based linear interpolation is used to fill in the missing values. For example, if a sensor loses data between 10:00 and 10:05, linear fitting is performed based on data points from the preceding and following five minutes to generate complete data. In the preprocessing stage, parameters with different dimensions, such as temperature, humidity, and voltage, are normalized using Min-Max, mapping all parameters to the 0-1 range. For load data collected every minute and meteorological data collected every 5 minutes, linear interpolation is used to resample the meteorological data to a minute-level time resolution. Finally, feature-level fusion technology is used to align the standardized multi-source data by timestamp, forming a comprehensive dataset containing multi-dimensional features such as grid load, meteorological parameters, and equipment status.

[0037] Compared to existing technologies, traditional methods typically employ manual screening to handle outlier data, lacking automated processing workflows. For example, they rely on maintenance personnel to visually inspect data curves to identify anomalies, leading to low processing efficiency and subjective errors. Existing technologies often process data from different devices independently, such as analyzing wind speed and voltage data separately, resulting in the ineffective utilization of the spatiotemporal correlations between multi-source data. This solution, by constructing a standardized automated data processing workflow, achieves the systematic integration of multi-source heterogeneous data. For instance, it unifies previously scattered meteorological data, equipment status data, and power grid load data from independent systems into a comprehensive dataset with spatiotemporal correlations, providing high-quality input for subsequent predictive models.

[0038] Through the above technical solutions, this application solves the problem of decreased analysis accuracy caused by acquisition errors and format differences in multi-source heterogeneous data. For example, it eliminates the impact of wind speed sensor noise on power generation prediction and avoids scheduling command delays caused by data time asynchrony. Automated anomaly detection and data completion mechanisms reduce the need for manual intervention, such as automatically identifying and repairing missing energy storage system status data caused by communication interruptions. Standardization processing makes data from different devices comparable; for example, it unifies the kilowatt-level power data of photovoltaic arrays and the percentage state of charge data of energy storage into dimensionless values, facilitating collaborative calculations by optimization algorithms. Time alignment processing ensures the temporal consistency of multi-source data; for example, it allows for correlation analysis of load data updated every minute and solar radiation data collected every 5 minutes under the same time base, improving the timeliness of scheduling commands.

[0039] Example 4: Please refer to Figure 1 In step S3, the demand and power generation forecasting are conducted in the following ways: S3.1 Extract time characteristic parameters from historical load data from the comprehensive dataset. The time characteristic parameters include hourly load fluctuation characteristics, daily load change characteristics, weekly load cycle characteristics, and monthly load trend characteristics. Obtain real-time meteorological data and date attribute information as input variables for external influencing factors. Real-time meteorological data includes ambient temperature data, relative humidity data, and wind speed data. Date attribute information includes weekday type identifier, holiday identifier, and season identifier. Use a preset prediction algorithm to input the time characteristic parameters and external influencing factor input variables into the power demand prediction model and generate power demand prediction curves for future prediction periods. S3.2 Extract historical wind speed data, historical wind direction data, and historical power output data of wind turbine generators; obtain real-time and forecasted wind speed parameters, wind direction parameters, and atmospheric pressure parameters; establish a conversion relationship model between wind speed and power and generate wind power output curves for future periods; extract historical solar radiation intensity data, historical ambient temperature data, and historical power output data of photovoltaic arrays; obtain real-time and forecasted solar radiation intensity parameters, cloud cover parameters, and ambient temperature parameters; establish a conversion relationship model between solar radiation and power and generate photovoltaic power output curves for future periods; superimpose the wind power output curves and photovoltaic power output curves to generate predicted data for total renewable energy power generation.

[0040] In this embodiment: This application further proposes a specific implementation method for the demand and power generation forecasting steps in the intelligent operation and maintenance and adaptive control method for new energy facilities: Time characteristic parameters of historical load data are extracted from the comprehensive dataset. These time characteristic parameters include hourly load fluctuation characteristics, daily load change characteristics, weekly load cycle characteristics, and monthly load trend characteristics. Real-time meteorological data and date attribute information are obtained as input variables for external influencing factors. The real-time meteorological data includes ambient temperature data, relative humidity data, and wind speed data. The date attribute information includes weekday type identifiers, holiday identifiers, and season identifiers. A forecasting algorithm is used to input the time characteristic parameters and external influencing factor input variables into the power demand forecasting model and generate... The system generates a forecast curve for electricity demand during the future forecast period. It extracts historical wind speed, wind direction, and wind turbine output data to obtain real-time and forecasted wind speed, wind direction, and atmospheric pressure parameters. A model for the conversion between wind speed and power is established, generating a wind power output curve for the future period. Similarly, it extracts historical solar radiation intensity, historical ambient temperature, and photovoltaic array output data to obtain real-time and forecasted solar radiation intensity, cloud cover, and ambient temperature parameters. A model for the conversion between solar radiation and power is established, generating a photovoltaic power output curve for the future period. Finally, the wind power output curve and the photovoltaic power output curve are superimposed to generate a forecast of total renewable energy power generation.

[0041] Time characteristic parameters refer to the load variation patterns at different time scales extracted from historical load data. Specifically, time series analysis methods can be used to decompose these into hourly fluctuations, daily variations, weekly cycles, and monthly trends to capture short-term load fluctuations and long-term evolution patterns. External influencing factors, or input variables, refer to the dynamic impact of real-time weather conditions and date attributes on electricity demand. For example, ambient temperature data is obtained through temperature sensors, and holiday markers are identified through calendar systems. These are used to correct the forecasting model's response to weather sensitivity and social activity patterns. Forecasting algorithms are machine learning models used to establish the mapping relationship between input variables and electricity demand. Examples include long short-term memory neural networks or gradient boosting decision trees, which generate high-precision demand forecast curves by training historical and real-time data. The wind speed-power conversion model is a mathematical model based on the aerodynamic characteristics of wind turbine generators. For example, wind speed parameters can be mapped to theoretical power generation through power curve fitting methods, and actual output can be corrected by combining wind direction and atmospheric pressure parameters. The solar radiation-power conversion model refers to a mathematical model based on the photoelectric conversion efficiency of photovoltaic arrays. For example, the theoretical output can be calculated using radiation intensity and temperature compensation coefficient, and the actual power generation can be corrected by combining cloud cover parameters.

[0042] In the demand forecasting phase, historical load data is decomposed into temporal characteristics, such as hourly load fluctuations during weekday morning peak hours, daily load variations on weekends, weekly load cycles from industrial production periods, and monthly load trends for seasonality. These are combined with real-time meteorological data (temperature, humidity, wind speed, etc.) and holiday markers from date attributes, and input into a trained forecasting model to generate a 24-hour electricity demand forecast curve. In the power generation forecasting phase, for wind power, the correlation between historical wind speed and turbine output is used, for example, by fitting power curves to obtain theoretical output at different wind speeds. This is combined with real-time wind speed forecasts and atmospheric pressure data to generate a wind power output curve for future periods. For photovoltaic (PV) power, based on the relationship between solar radiation intensity and PV module efficiency, such as using a radiation-temperature-output model and incorporating cloud cover forecasts to correct for radiation attenuation, a PV power output curve is generated. Finally, the predicted output curves for wind and PV are superimposed to obtain total renewable energy power generation forecast data, providing input for subsequent dispatch optimization.

[0043] Compared to existing technologies, traditional methods for predicting electricity demand and renewable energy generation typically rely solely on historical data at a single timescale or fail to adequately integrate external influencing factors, such as ignoring the abrupt changes in load caused by holidays or neglecting the dynamic attenuation of photovoltaic output due to cloud cover. This proposed solution significantly improves the adaptability and accuracy of the prediction model through multi-timescale feature decomposition and external factor fusion, such as combining hourly fluctuations with monthly trends and incorporating real-time meteorological and date attributes. Furthermore, traditional wind speed and power conversion models often neglect the influence of wind direction and atmospheric pressure, leading to significant deviations in power output predictions. This solution, by establishing multi-dimensional parameter transformation relationships, such as incorporating corrections for wind turbine yaw losses, enhances the accuracy of wind power generation prediction.

[0044] Through the above technical solutions, this application can more accurately predict electricity demand and renewable energy generation. For example, it reduces the timing error of load forecasting through time feature decomposition and reduces the impact of sudden weather changes on power generation forecasting through meteorological data fusion, thereby providing a reliable basis for system-level dispatch. This helps optimize the coordinated operation of energy storage devices and traditional generator sets, such as reducing power deficits or surpluses caused by forecasting errors, improving the renewable energy absorption capacity, and reducing the frequent adjustment losses of traditional units.

[0045] Example 5: Please refer to Figure 1 The system-level scheduling optimization in step S4 is implemented as follows: S4.1 Determine the scheduling optimization objective function, which includes maximizing renewable energy utilization, minimizing system operating costs, and minimizing carbon emissions. Set the system operation constraint set, which includes power supply and demand balance constraints, upper and lower limits of wind turbine output, upper and lower limits of photovoltaic array output, upper and lower limits of energy storage system state of charge, upper and lower limits of energy storage system charging and discharging power, upper and lower limits of traditional generator output, traditional generator ramp rate constraints, and grid frequency and voltage safe operation constraints. Establish renewable energy priority scheduling rules, construct energy storage system auxiliary scheduling strategies, calculate power deficit or power surplus based on the deviation between renewable energy output curve and power demand forecast curve, and determine the charging period, discharging period, and charging and discharging power allocation scheme of energy storage system.

[0046] A represents the total available power, B represents the predicted wind power, C represents the predicted photovoltaic power, D represents the equivalent power that energy storage can release / absorb in the current cycle, Amax represents the physical upper limit, and a, b, and c are weight values, and a, b, and c are all non-zero, and a+b+c=1. S4.2. Solve the multi-source collaborative scheduling optimization model using a preset optimization algorithm and generate a system scheduling scheme, including: numerically solving the multi-objective optimization function through the optimization algorithm and obtaining the optimal scheduling solution under the premise of satisfying the system operation constraint set, generating a system scheduling scheme. The system scheduling scheme includes power scheduling instructions for wind turbine generators, power scheduling instructions for photovoltaic arrays, charging and discharging scheduling instructions for energy storage systems, and start-up and shutdown scheduling instructions and output scheduling instructions for traditional generators. The energy storage system smooths out short-term power fluctuations and reduces the number of starts and stops of traditional generators during the scheduling cycle. The system scheduling scheme is then distributed to the wind turbine generator control unit, photovoltaic array control unit, energy storage system control unit, and traditional generator control unit through the communication network.

[0047]

[0048] For the output of conventional units in this cycle, E is the real-time load, and A is the total available power. For the actual output of traditional units in the previous cycle, To suppress over-adjustment caused by sudden deviations, Let κ∈(0,1] be the robust threshold, representing the climbing factor, selected to make... ; In this embodiment: This application further proposes, in system-level scheduling optimization, determining the scheduling optimization objective function, which includes maximizing renewable energy utilization, minimizing system operating costs, and minimizing carbon emissions; setting a set of system operation constraints, which includes power supply and demand balance constraints, upper and lower limits of wind turbine output, upper and lower limits of photovoltaic array output, upper and lower limits of energy storage system state of charge, upper and lower limits of energy storage system charging and discharging power, upper and lower limits of traditional generator output, traditional generator ramp rate constraints, and grid frequency and voltage safety operation constraints; establishing renewable energy priority scheduling rules; constructing an energy storage system auxiliary scheduling strategy; and calculating power deficit or power surplus based on the deviation between the renewable energy output curve and the power demand forecast curve, and determining the energy storage... The system includes charging and discharging periods and power allocation schemes for the energy system. A pre-defined optimization algorithm is used to solve the multi-source collaborative scheduling optimization model and generate a system scheduling scheme. This includes numerically solving a multi-objective optimization function using the optimization algorithm and obtaining the optimal scheduling solution while satisfying the system operation constraints. The system scheduling scheme includes power scheduling commands for wind turbine generators, photovoltaic arrays, energy storage systems, and traditional generator start-up and shutdown commands and output scheduling commands. The energy storage system helps to mitigate short-term power fluctuations and reduce the number of starts and stops of traditional generators during the scheduling cycle. The system scheduling scheme is then distributed to the wind turbine generator control unit, photovoltaic array control unit, energy storage system control unit, and traditional generator control unit control unit via a communication network.

[0049] The scheduling optimization objective function refers to transforming renewable energy utilization, operating costs, and carbon emissions into quantifiable optimization indicators through mathematical modeling. Specifically, it can be implemented using multi-objective weighted methods or Pareto optimal solution set selection methods to balance conflicting relationships between different optimization objectives. The system operation constraint set refers to the aggregated expression of equipment physical operation limitations and grid safety requirements. Specifically, it can be implemented using linear inequality constraints or nonlinear constraint equations to ensure that the scheduling scheme conforms to actual operating conditions. The renewable energy priority scheduling rule refers to the scheduling logic that uses wind and solar power output as the basic power supply source. Specifically, it can be implemented using priority ranking algorithms or dynamic adjustment coefficient methods to maximize the absorption of renewable energy power generation. The energy storage system auxiliary scheduling strategy refers to the strategy of dynamically adjusting the charging and discharging state of energy storage based on power deviations. Specifically, it can be implemented using model predictive control or rolling optimization methods to compensate for the volatility of renewable energy power generation. The multi-source collaborative scheduling optimization model refers to the joint scheduling model integrating wind power, solar power, energy storage, and traditional generator units. Specifically, it can be implemented using mixed integer programming or stochastic optimization algorithms to generate scheduling instructions for the coordinated operation of multiple devices.

[0050] During system scheduling, a multi-objective function incorporating renewable energy utilization, operating costs, and carbon emissions is first established, and constraints are used to limit the output range of each device and the grid safety boundary. Based on the deviation analysis between renewable energy output forecast and load demand forecast, the charging and discharging periods of the energy storage system are dynamically divided. For example, during peak photovoltaic output periods, the energy storage is set to charging mode, and during peak nighttime load periods, it is switched to discharging mode. An optimization algorithm is used to obtain a scheduling scheme that balances economic efficiency and environmental friendliness, generating full-load commands for wind and photovoltaic power and peak-shaving commands for traditional generator sets. Furthermore, the energy storage system is utilized to absorb short-term power fluctuations; for example, during sudden wind speed changes, the charging and discharging levels of the energy storage system suppress power fluctuations, thereby reducing the frequent start-stop operations of traditional generator sets.

[0051] Compared to existing technologies, traditional scheduling methods often employ fixed-priority scheduling strategies, which cannot dynamically respond to fluctuations in renewable energy output, resulting in energy storage systems being used only as backup power. This solution, however, establishes a multi-objective optimization model and dynamic scheduling rules to achieve synergistic interaction between the energy storage system and renewable energy sources. For example, when wind and solar resources are abundant, excess electricity is prioritized for storage; when output is insufficient, the stored electricity is quickly released to replace the peak-shaving function of traditional generator sets. This dynamic scheduling mechanism reduces the number of start-ups and shutdowns of traditional units compared to fixed strategies, while simultaneously improving the utilization rate of energy storage devices.

[0052] Through the above technical solutions, this application achieves coordinated and optimized scheduling of multiple types of power sources, prioritizing the consumption of renewable energy generation while ensuring grid security. The flexible charging and discharging control of the energy storage system effectively smooths out the intermittent fluctuations of wind and solar power generation, reducing the frequency regulation pressure on traditional generator sets. The establishment of a multi-objective optimization model ensures that the scheduling scheme simultaneously meets the requirements of economy, environmental protection, and equipment operation safety, solving the problem of low overall system efficiency caused by single-objective optimization in traditional scheduling methods.

[0053] Example 6: Please refer to Figure 1 The optimization of energy storage device charging and discharging control in step S5 is as follows: S5.1. Obtain the current state of charge (SOC) value, battery temperature value, internal resistance value, and historical cycle count data of the energy storage system. Calculate the charging and discharging timing parameters based on the power magnitude and duration of the power supply and demand deficit or surplus. The charging and discharging timing parameters include the charging start time, charging end time, discharging start time, and discharging end time. Calculate the charging and discharging rate parameters based on the preset battery characteristic curve and the current SOC value. Set the upper and lower threshold values ​​of the SOC as charging and discharging depth limit parameters. S5.2 Generate energy storage charging and discharging control commands that include charging and discharging timing parameters, charging and discharging power setpoints, and target values ​​for the state of charge. Send the energy storage charging and discharging control commands to the energy storage system control unit through the communication network. During the charging and discharging process, collect actual charging and discharging power data, state of charge change rate data, battery temperature change data, and voltage and current response data. When the battery temperature exceeds the preset temperature safety threshold, the state of charge approaches the boundary value, or the charging and discharging efficiency is lower than the preset efficiency lower limit, adjust the charging and discharging power value or terminate the charging and discharging process.

[0054] In this embodiment, this application further proposes specific steps for optimizing the charging and discharging control of energy storage devices, including acquiring the current state of charge (SOC) value, battery temperature value, internal resistance value, and historical cycle count data of the energy storage system; calculating charging and discharging timing parameters based on the magnitude and duration of power supply and demand deficits or surpluses, including charging start time, charging end time, discharging start time, and discharging end time; calculating charging and discharging rate parameters based on preset battery characteristic curves and the current SOC value; setting upper and lower thresholds for the SOC as charging and discharging depth limitation parameters; generating energy storage charging and discharging control commands containing charging and discharging timing parameters, charging and discharging power settings, and target SOC values; sending the energy storage charging and discharging control commands to the energy storage system control unit via a communication network; collecting actual charging and discharging power data, SOC change rate data, battery temperature change data, and voltage and current response data during the charging and discharging process; adjusting the charging and discharging power value or terminating the charging and discharging process when the battery temperature exceeds a preset temperature safety threshold, the SOC approaches a boundary value, or the charging and discharging efficiency is lower than a preset efficiency lower limit value.

[0055] The state of charge (SOC) value refers to the percentage of the current remaining charge of an energy storage battery relative to its rated capacity. It can be calculated by real-time acquisition of battery terminal voltage and charge / discharge current data using voltage and current sensors, combined with a battery capacity model, and is used to assess the available energy reserves of the energy storage system. Battery temperature refers to the measured temperature of the battery's interior or surface, which can be obtained using thermocouples or infrared temperature sensors. It is used to monitor whether the battery's operating status is within a safe range. Internal resistance refers to the battery's internal equivalent resistance, which can be measured using AC impedance spectroscopy or pulse discharge methods, and is used to assess the degree of battery aging. Charge / discharge timing parameters refer to the charging / discharging time window of the energy storage device dynamically determined based on the power supply and demand balance requirements. It can be calculated using optimization algorithms combined with load forecast curves and power generation output curves, and is used to match system power regulation needs. Charge / discharge rate parameters refer to the ratio of the energy storage device's charging / discharging power to its capacity. It can be dynamically adjusted based on the charge / discharge rate characteristic curve provided by the battery manufacturer and the current SOC value, and is used to balance charging / discharging efficiency and battery life. The upper and lower thresholds of the state of charge refer to the safe range of the battery's state of charge set to prevent overcharging or over-discharging. For example, the upper threshold can be set to 95% of the state of charge, and the lower threshold can be set to 20% of the state of charge, in order to protect the battery's health.

[0056] The energy storage system control unit collects real-time data on battery cell voltage, temperature, internal resistance, and historical cycle count. Combined with power demand based on supply and demand imbalances or surpluses, it dynamically calculates charging and discharging time windows and power allocation schemes using optimization algorithms. For example, when renewable energy output is excessive, charging is prioritized during periods when the state of charge (SBC) is below the upper limit threshold, and the charging rate is adjusted based on battery temperature trends. During peak load periods, discharging is initiated based on the SBC lower limit threshold and battery health status, while limiting the discharge power according to the rate of change in internal resistance. During charging and discharging, battery temperature, voltage and current response, and efficiency changes are continuously monitored. When the temperature exceeds a safety threshold or the efficiency falls below a preset value, the charging and discharging power is automatically reduced or the operation is terminated to prevent battery overheating or excessive wear.

[0057] Compared to existing technologies, traditional energy storage control methods typically rely solely on state-of-charge thresholds for charge / discharge start / stop control, lacking comprehensive consideration of battery temperature, internal resistance, and cycle count. This makes them susceptible to control strategy failure due to ambient temperature fluctuations or battery aging. In contrast, this solution employs a multi-parameter fusion dynamic adjustment mechanism to adapt to real-time changes in battery state. For example, it automatically reduces the charge / discharge rate in high-temperature environments to prevent thermal runaway, or adjusts power distribution based on changes in internal resistance during battery aging, thereby improving control accuracy and safety.

[0058] Through the above technical solutions, this application achieves refined control of the charging and discharging process of energy storage devices, effectively balancing the system power regulation requirements and battery health protection requirements. For example, it avoids energy conversion during inefficient periods in selecting charging and discharging times, reduces efficiency loss caused by battery polarization effects in charging and discharging rate control, and reduces the risks of battery overcharging, over-discharging, and overheating through real-time monitoring and dynamic adjustment mechanisms, thereby extending the service life of energy storage devices.

[0059] Example 7: Please refer to Figure 1 The equipment health monitoring in step S6 is carried out in the following manner: S6.1 Collect battery cell voltage data, battery temperature data, charge / discharge current data, internal resistance change rate data, cycle count data, and capacity decay rate data of the energy storage system; collect bearing vibration signal data, lubricating oil temperature data, cooling water temperature data, exhaust temperature data, speed fluctuation data, and power output data of the traditional generator set; perform signal processing on the collected operating status data and extract feature parameters, including time-domain feature parameters and frequency-domain feature parameters. Time-domain feature parameters include mean, variance, peak value, and crest factor; frequency-domain feature parameters include spectral energy distribution, characteristic frequency amplitude, and harmonic components. Combine the operating status data and feature parameters to form an equipment health feature vector. S6.2. Train the equipment health assessment model using historical normal operation data and historical fault data. Input the real-time collected equipment health feature vector into the equipment health assessment model to obtain the equipment health index. The equipment health index is a value from 0 to 100. Set health warning threshold and fault warning threshold. When the equipment health index is lower than the health warning threshold, trigger the equipment degradation warning signal. When the equipment health index is lower than the fault warning threshold, trigger the equipment fault warning signal. Transmit the warning signal to the operation and maintenance terminal through the central control system and output the warning report. The warning report includes equipment identification information, equipment health index value, abnormal characteristic parameters and maintenance suggestions.

[0060] In this embodiment, the application further proposes the following specific method for equipment health monitoring: Collecting battery cell voltage data, battery temperature data, charge / discharge current data, internal resistance change rate data, cycle count data, and capacity decay rate data from the energy storage system; collecting bearing vibration signal data, lubricating oil temperature data, cooling water temperature data, exhaust temperature data, speed fluctuation data, and power output data from a traditional generator set; performing signal processing on the collected operating status data and extracting feature parameters, including time-domain and frequency-domain feature parameters. Time-domain feature parameters include mean, variance, peak value, and crest factor; frequency-domain feature parameters include spectral energy distribution, characteristic frequency amplitude, and harmonic components. The operating status data is then analyzed. Data and feature parameters are combined to form an equipment health feature vector. The equipment health assessment model is trained using historical normal operation data and historical fault data. The real-time collected equipment health feature vector is input into the equipment health assessment model to obtain the equipment health index, which is a value from 0 to 100. Health warning thresholds and fault warning thresholds are set. When the equipment health index is lower than the health warning threshold, an equipment degradation warning signal is triggered. When the equipment health index is lower than the fault warning threshold, an equipment fault warning signal is triggered. The warning signal is transmitted to the operation and maintenance terminal through the central control system and a warning report is output. The warning report includes equipment identification information, equipment health index value, abnormal feature parameters, and maintenance suggestions.

[0061] Battery cell voltage data refers to the measured voltage of a single battery cell in an energy storage system. This data can be collected using voltage sensors and is used to monitor the charge / discharge state and voltage balance of the battery cell. Battery temperature data refers to the temperature parameters of the battery during operation. This data can be collected using temperature sensors and is used to assess the battery's thermal management status. Time-domain feature parameters refer to the statistical features of the signal extracted from the time dimension. These can be calculated using a sliding window to determine parameters such as mean and variance, and are used to characterize the steady-state characteristics of the equipment. Frequency-domain feature parameters refer to the signal spectrum features extracted through Fourier transform. This can be calculated using a fast Fourier transform algorithm to determine the spectral energy distribution and is used to identify abnormal vibrations or electromagnetic interference in the equipment. The equipment health assessment model is a machine learning-based equipment status classification model. This model can be implemented using support vector machines or neural network algorithms and is used to quantitatively assess the health status of the equipment.

[0062] Energy storage systems collect real-time operating data from individual battery cells using voltage and temperature sensors, while traditional generator sets collect mechanical operating data using vibration and temperature sensors. After preprocessing, the raw signals are used to extract time-domain statistical features and frequency-domain spectral features, forming multi-dimensional feature vectors. Feature vectors from historical data covering normal and fault conditions are used to train a classification model, enabling the model to identify feature patterns corresponding to different health states. During online monitoring, the real-time feature vectors are input into the model, outputting a health score of 0-100. A score below a preset threshold triggers a tiered warning. For example, when the rate of change of internal resistance of the energy storage battery abnormally increases and the health score is below 60, the system automatically generates a maintenance work order and pushes it to the maintenance personnel's terminal.

[0063] Compared to existing technologies, traditional equipment health monitoring methods typically rely on a single parameter threshold for judgment, such as monitoring only whether the temperature exceeds the limit. This solution, by fusing multi-dimensional sensor data and extracting composite time-frequency features, can more comprehensively characterize the equipment degradation process. Existing manual periodic inspection methods suffer from response lag, while this solution uses a machine learning model to achieve real-time quantitative assessment of health status, significantly improving the timeliness of fault warnings.

[0064] Through the above technical solution, this application realizes online health status assessment of key equipment in new energy facilities, enabling the identification of abnormal characteristics and triggering early warnings in the early stages of equipment performance degradation, thus avoiding system downtime caused by sudden failures. By differentiating the degree of equipment degradation through a tiered early warning mechanism, it provides differentiated maintenance decision support for operation and maintenance personnel, effectively reducing the number of unplanned repairs and equipment replacement costs.

[0065] Example 8: Please refer to Figure 1 The adaptive control in step S7 is implemented in the following way: S7.1 Acquire real-time system operating data, including actual output data of wind turbine generators, actual output data of photovoltaic arrays, actual output data of traditional generators, actual charging and discharging power data of energy storage systems, real-time grid load data, system frequency data, and system voltage data. Compare the real-time system operating data with the target operating parameters in the system scheduling scheme to calculate and obtain the deviation, including power deviation, frequency deviation, and voltage deviation. Identify external disturbance events by monitoring the rate of change of data. External disturbance events include load change events, renewable energy output change events, equipment failure events, or grid failure events. Set allowable deviation thresholds, including allowable power deviation thresholds, allowable frequency deviation thresholds, and allowable voltage deviation thresholds. When the deviation exceeds the allowable deviation threshold or a major disturbance event of a preset level is detected, a control strategy adjustment trigger signal is generated. S7.2. Based on the magnitude of the deviation and the trend of its time series change, a preset adaptive control algorithm is used to calculate the control strategy adjustment scheme. The control strategy adjustment scheme includes the output adjustment of the wind turbine generator set, the output adjustment of the photovoltaic array, the output adjustment of the traditional generator set, the charging and discharging power adjustment of the energy storage system, and the load-side response adjustment. Rapid response measures are formulated for different types of external disturbance events. Adjusted equipment control commands are generated and sent to the wind turbine generator set control unit, the photovoltaic array control unit, the energy storage system control unit, and the traditional generator set control unit through the communication network. The adjusted system operating status is continuously monitored. When the deviation of the system operating status is lower than the allowable deviation threshold, the current control strategy is maintained. When the deviation of the system operating status still exceeds the allowable deviation threshold, the control strategy adjustment calculation process is repeated and a new control strategy adjustment scheme is generated.

[0066] In this embodiment, the application further proposes that the adaptive control process includes the following steps: collecting real-time system operating data, including actual output data of wind turbine generators, actual output data of photovoltaic arrays, actual output data of traditional generators, actual charging and discharging power data of energy storage systems, real-time grid load data, system frequency data, and system voltage data; comparing and calculating the deviation between the real-time system operating data and the target operating parameters in the system scheduling scheme, including power deviation, frequency deviation, and voltage deviation; identifying external disturbance events by monitoring the rate of change of data, including load change events, renewable energy output change events, equipment failure events, or grid failure events; and setting allowable deviation thresholds, including allowable power deviation thresholds, allowable frequency deviation thresholds, and allowable voltage deviation thresholds. When the deviation exceeds the allowable deviation threshold or a major disturbance event of a preset level is detected, a control strategy adjustment trigger signal is generated. Based on the magnitude of the deviation and its time series trend, an adaptive control algorithm is used to calculate the control strategy adjustment scheme, which includes adjustments to the output of wind turbine generators, photovoltaic arrays, traditional generators, energy storage system charging and discharging power, and load-side response. Rapid response measures are formulated for different types of external disturbance events, and adjusted equipment control commands are generated and sent to the control unit through the communication network. The adjusted system operating status is continuously monitored. When the deviation of the system operating status is lower than the allowable deviation threshold, the current control strategy is maintained. When the deviation of the system operating status still exceeds the allowable deviation threshold, the control strategy adjustment calculation process is repeated and a new control strategy adjustment scheme is generated.

[0067] Real-time system operation data refers to the set of physical quantities that reflect the current operating status of new energy facilities, obtained through sensors and monitoring devices. Specifically, it can be implemented using distributed data acquisition modules and communication protocol parsing technology to construct a complete view of the system's dynamic operation.

[0068] Deviation refers to the quantitative indicator of the difference between the actual operating parameters and the scheduling target. It can be implemented by difference calculation or percentage error algorithm and is used to evaluate the degree of deviation of the system's operating status.

[0069] External disturbance events refer to unplanned operating conditions that cause sudden changes in the system's operating state. Specifically, they can be implemented using sliding window variance analysis or mutation point detection algorithms to identify sudden situations such as power grid anomalies or equipment failures.

[0070] The allowable deviation threshold refers to the range of parameter fluctuations that the system can withstand to maintain stable operation. It can be set according to the power grid safety operation standards or the equipment tolerance capabilities, and is used to determine the critical conditions that trigger the adjustment of the control strategy.

[0071] Adaptive control algorithms are mathematical methods that dynamically adjust control parameters based on real-time feedback. Specifically, they can be implemented using model predictive control or fuzzy PID control algorithms to generate multi-device coordinated adjustment commands.

[0072] During system operation, the multi-source data acquisition module continuously acquires the actual output data of each power generation unit and the grid operating parameters, and compares them in real time with the target values ​​in the scheduling scheme. When a power deviation exceeding a preset threshold is detected or a disturbance event such as a sudden load change is identified, the control strategy reconfiguration mechanism is triggered. Based on the time-series characteristics of the deviation, the output commands of each power generation device are redistributed through an optimization algorithm, and the charging and discharging power of the energy storage system is adjusted first to quickly compensate for the power gap. In response to grid frequency or voltage anomalies, the output of traditional generator sets and the load-side response strategies are adjusted simultaneously to form multi-dimensional collaborative control. After the control command is executed, the closed-loop monitoring module continuously tracks the system status. If the deviation is not eliminated, a secondary optimization calculation is initiated until the system returns to stable operation.

[0073] Compared to existing technologies, traditional methods typically use fixed thresholds to trigger power adjustments for single devices, lacking a design for collaborative response mechanisms across multiple devices, and relying on manual intervention during secondary adjustments. This solution, by constructing a closed-loop feedback control system, achieves dynamic collaborative optimization across multiple types of devices. It can automatically generate composite adjustment strategies based on real-time deviation characteristics, significantly improving the system's ability to suppress external disturbances.

[0074] Through the above technical solution, this application can quickly generate multi-device collaborative control commands when there are sudden changes in grid load or fluctuations in renewable energy output, effectively smoothing out power and frequency deviations. By employing a closed-loop monitoring and iterative optimization mechanism, it avoids the continuous instability problem caused by insufficient single adjustments in traditional methods, thereby improving the system's anti-interference capability and operational reliability under complex operating conditions.

[0075] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units. The above are merely embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

[0076] The specific embodiments of the invention have been described in detail above, but they are only examples, and this application is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions to the invention are also within the scope of this application. Therefore, all equivalent changes, modifications, and improvements made without departing from the spirit and principles of this application should be covered within the scope of this application.

Claims

1. A method for intelligent operation and maintenance and adaptive control of new energy facilities, characterized in that: The specific steps are as follows: S1. Data Acquisition and Monitoring: Deploy sensors and monitoring equipment in new energy facilities to acquire data on grid load, wind speed, solar radiation, energy storage system charging status, and equipment operating health status, and transmit the data to the central control system through Internet of Things (IoT) technology. S2. Data cleaning and preprocessing: Denoising, outlier removal, and missing value imputation are performed on the raw data to ensure data accuracy, consistency, and integrity. Data from different sources are also standardized and fused. S3. Demand and power generation forecasting: Based on historical data and real-time information, forecasting models are used to predict future electricity demand and the power generation of wind and solar power, taking into account external factors such as weather changes and seasonal fluctuations. S4. System-level scheduling optimization: Based on real-time data and power generation demand forecasts, optimization algorithms are used to schedule the operation of wind power, photovoltaic, energy storage systems and traditional generator sets, giving priority to the use of renewable energy, reducing the start-up and shutdown frequency of traditional generator sets, and rationally scheduling energy storage devices. S5. Energy storage device charging and discharging control optimization: Based on system-level scheduling optimization, the timing, rate and depth of charging and discharging of energy storage devices are adjusted to ensure efficient charging and discharging of energy storage devices and reduce efficiency loss and equipment damage caused by overcharging and discharging. S6. Equipment health monitoring: Deploys intelligent sensors and uses machine learning algorithms to monitor the operating status of energy storage systems and traditional generator sets, assesses equipment health status, and promptly detects potential faults and issues early warnings. S7. Adaptive control adjusts the system control strategy based on real-time operating data and feedback information to ensure stable system operation and respond quickly to external disturbances, maintaining supply and demand balance and system stability.

2. The intelligent operation and maintenance and adaptive control method for new energy facilities according to claim 1, characterized in that, The data collection and monitoring in step S1 are carried out in the following ways: S1.1 Deploy wind speed sensors, solar radiation sensors, voltage sensors, current sensors, energy storage system charging status sensors, and equipment health monitoring sensors in new energy facilities; S1.2 Collect sensor data through Internet of Things (IoT) technology, including grid load, wind speed, solar radiation, energy storage system charging status and equipment health status, and transmit the data to the central control system.

3. The intelligent operation and maintenance and adaptive control method for new energy facilities according to claim 2, characterized in that, The data cleaning and preprocessing in step S2 are carried out in the following ways: S2.

1. The original data is denoised using a preset digital filtering algorithm. Anomalies are identified and removed from the denoised data based on a preset statistical anomaly detection method. Data is filled in using a preset missing value imputation method. S2.

2. Use a preset standardization or normalization method to unify the dimensions of data of different physical quantities and map them to a unified numerical range. Perform time alignment processing on data with different sampling frequencies and use a resampling method to make the time resolution of each data source consistent. Use a preset multi-source data fusion technology to fuse the standardized and time-aligned data and construct a comprehensive dataset.

4. The intelligent operation and maintenance and adaptive control method for new energy facilities according to claim 3, characterized in that, The demand and power generation forecasting in step S3 is described in the following manner: S3.1 Extract time characteristic parameters from historical load data from the comprehensive dataset. The time characteristic parameters include hourly load fluctuation characteristics, daily load change characteristics, weekly load cycle characteristics, and monthly load trend characteristics. Obtain real-time meteorological data and date attribute information as input variables for external influencing factors. Real-time meteorological data includes ambient temperature data, relative humidity data, and wind speed data. Date attribute information includes weekday type identifier, holiday identifier, and season identifier. Use a preset prediction algorithm to input the time characteristic parameters and external influencing factor input variables into the power demand prediction model and generate power demand prediction curves for future prediction periods. S3.2 Extract historical wind speed data, historical wind direction data, and historical power output data of wind turbine generators; obtain real-time and forecasted wind speed parameters, wind direction parameters, and atmospheric pressure parameters; establish a conversion relationship model between wind speed and power and generate wind power output curves for future periods; extract historical solar radiation intensity data, historical ambient temperature data, and historical power output data of photovoltaic arrays; obtain real-time and forecasted solar radiation intensity parameters, cloud cover parameters, and ambient temperature parameters; establish a conversion relationship model between solar radiation and power and generate photovoltaic power output curves for future periods; superimpose the wind power output curves and photovoltaic power output curves to generate predicted data for total renewable energy power generation.

5. The intelligent operation and maintenance and adaptive control method for new energy facilities according to claim 4, characterized in that, The system-level scheduling optimization in step S4 is implemented as follows: S4.1 Determine the scheduling optimization objective function, which includes maximizing renewable energy utilization, minimizing system operating costs, and minimizing carbon emissions. Set the system operation constraint set, which includes power supply and demand balance constraints, upper and lower limits of wind turbine output, upper and lower limits of photovoltaic array output, upper and lower limits of energy storage system state of charge, upper and lower limits of energy storage system charging and discharging power, upper and lower limits of traditional generator output, traditional generator ramp rate constraints, and grid frequency and voltage safe operation constraints. Establish renewable energy priority scheduling rules, construct energy storage system auxiliary scheduling strategies, calculate power deficit or power surplus based on the deviation between renewable energy output curve and power demand forecast curve, and determine the charging period, discharging period, and charging and discharging power allocation scheme of energy storage system. S4.

2. Solve the multi-source collaborative scheduling optimization model using a preset optimization algorithm and generate a system scheduling scheme, including: numerically solving the multi-objective optimization function through the optimization algorithm and obtaining the optimal scheduling solution under the premise of satisfying the system operation constraint set, generating a system scheduling scheme. The system scheduling scheme includes power scheduling instructions for wind turbine generators, power scheduling instructions for photovoltaic arrays, charging and discharging scheduling instructions for energy storage systems, and start-up and shutdown scheduling instructions and output scheduling instructions for traditional generators. The energy storage system smooths out short-term power fluctuations and reduces the number of starts and stops of traditional generators during the scheduling cycle. The system scheduling scheme is then distributed to the wind turbine generator control unit, photovoltaic array control unit, energy storage system control unit, and traditional generator control unit through the communication network.

6. The intelligent operation and maintenance and adaptive control method for new energy facilities according to claim 5, characterized in that, The optimization of energy storage device charging and discharging control in step S5 is as follows: S5.

1. Obtain the current state of charge (SOC) value, battery temperature value, internal resistance value, and historical cycle count data of the energy storage system. Calculate the charging and discharging timing parameters based on the power magnitude and duration of the power supply and demand deficit or surplus. The charging and discharging timing parameters include the charging start time, charging end time, discharging start time, and discharging end time. Calculate the charging and discharging rate parameters based on the preset battery characteristic curve and the current SOC value. Set the upper and lower threshold values ​​of the SOC as charging and discharging depth limit parameters. S5.2 Generate energy storage charging and discharging control commands that include charging and discharging timing parameters, charging and discharging power setpoints, and target values ​​for the state of charge. Send the energy storage charging and discharging control commands to the energy storage system control unit through the communication network. During the charging and discharging process, collect actual charging and discharging power data, state of charge change rate data, battery temperature change data, and voltage and current response data. When the battery temperature exceeds the preset temperature safety threshold, the state of charge approaches the boundary value, or the charging and discharging efficiency is lower than the preset efficiency lower limit, adjust the charging and discharging power value or terminate the charging and discharging process.

7. The intelligent operation and maintenance and adaptive control method for new energy facilities according to claim 6, characterized in that, The equipment health monitoring in step S6 is carried out in the following manner: S6.1 Collect battery cell voltage data, battery temperature data, charge / discharge current data, internal resistance change rate data, cycle count data, and capacity decay rate data of the energy storage system; collect bearing vibration signal data, lubricating oil temperature data, cooling water temperature data, exhaust temperature data, speed fluctuation data, and power output data of the traditional generator set; perform signal processing on the collected operating status data and extract feature parameters, including time-domain feature parameters and frequency-domain feature parameters. Time-domain feature parameters include mean, variance, peak value, and crest factor; frequency-domain feature parameters include spectral energy distribution, characteristic frequency amplitude, and harmonic components. Combine the operating status data and feature parameters to form an equipment health feature vector. S6.

2. Train the equipment health assessment model using historical normal operation data and historical fault data. Input the real-time collected equipment health feature vector into the equipment health assessment model to obtain the equipment health index. The equipment health index is a value from 0 to 100. Set health warning threshold and fault warning threshold. When the equipment health index is lower than the health warning threshold, trigger the equipment degradation warning signal. When the equipment health index is lower than the fault warning threshold, trigger the equipment fault warning signal. Transmit the warning signal to the operation and maintenance terminal through the central control system and output the warning report. The warning report includes equipment identification information, equipment health index value, abnormal characteristic parameters and maintenance suggestions.

8. The intelligent operation and maintenance and adaptive control method for new energy facilities according to claim 7, characterized in that, The adaptive control in step S7 is implemented as follows: S7.1 Acquire real-time system operating data, including actual output data of wind turbine generators, actual output data of photovoltaic arrays, actual output data of traditional generators, actual charging and discharging power data of energy storage systems, real-time grid load data, system frequency data, and system voltage data. Compare the real-time system operating data with the target operating parameters in the system scheduling scheme to calculate and obtain the deviation, including power deviation, frequency deviation, and voltage deviation. Identify external disturbance events by monitoring the rate of change of data. External disturbance events include load change events, renewable energy output change events, equipment failure events, or grid failure events. Set allowable deviation thresholds, including allowable power deviation thresholds, allowable frequency deviation thresholds, and allowable voltage deviation thresholds. When the deviation exceeds the allowable deviation threshold or a major disturbance event of a preset level is detected, a control strategy adjustment trigger signal is generated. S7.

2. Based on the magnitude of the deviation and the trend of its time series change, a preset adaptive control algorithm is used to calculate the control strategy adjustment scheme. The control strategy adjustment scheme includes the output adjustment of the wind turbine generator set, the output adjustment of the photovoltaic array, the output adjustment of the traditional generator set, the charging and discharging power adjustment of the energy storage system, and the load-side response adjustment. Rapid response measures are formulated for different types of external disturbance events. Adjusted equipment control commands are generated and sent to the wind turbine generator set control unit, the photovoltaic array control unit, the energy storage system control unit, and the traditional generator set control unit through the communication network. The adjusted system operating status is continuously monitored. When the deviation of the system operating status is lower than the allowable deviation threshold, the current control strategy is maintained. When the deviation of the system operating status still exceeds the allowable deviation threshold, the control strategy adjustment calculation process is repeated and a new control strategy adjustment scheme is generated.

Images