A fish light complementary reactive power compensation control method and device

Abstract

The present application relates to the technical field of power system automation control, in particular to a fish-light complementary reactive power compensation control method and device; the method comprises the following steps: through real-time acquisition of photovoltaic output, fishery load operation intensity and water surface state and other multi-source information, after time sequence alignment and feature extraction, a unified fish-light complementary operation state index is constructed by fusion; short-term change trend analysis and evolution mode classification are carried out, potential reactive power demand in the future period is predicted, and a smooth and continuous short-term demand index is generated; according to the high and low levels of the demand index, a variety of resources are dynamically matched, a hierarchical priority scheduling and multi-resource parallel compensation strategy is formed; in the compensation execution process, output and system stability indexes are monitored in real time, and dynamic correction is carried out through synchronization deviation and stability deviation; at the same time, trend learning is carried out combined with long-term historical data, and the prediction model and resource allocation strategy are continuously optimized. The present application effectively improves the operation stability and energy efficiency of the fish-light complementary system.

Description

Technical Field

[0001] This invention relates to the field of power system automation control technology, specifically to a reactive power compensation control method and device for solar-fishery complementary systems. Background Technology

[0002] As a comprehensive utilization model that integrates photovoltaic power generation and aquaculture, solar-aquaculture complementarity has shown significant advantages in promoting the efficient use of land resources and the development of clean energy in recent years. In this system, the output of photovoltaic power generation is affected by the changes in sunshine, weather and seasons, exhibiting fluctuating characteristics, while the load of aquaculture production follows the rhythm of aquaculture operations and has periodic or intermittent characteristics. In addition, the surface conditions of the water body may have indirect effects on the operation of photovoltaic modules, such as reflection and shading, which makes the overall electrical operation characteristics of the system more complex and puts forward more refined management requirements for reactive power balance and voltage stability.

[0003] Chinese invention patent CN119298167B discloses a two-layer optimization method for the operation of a fishery-solar-storage complementary system. The method includes: collecting parameter data from photovoltaic and energy storage devices, selecting feeding equipment, aerators, and irrigation / drainage equipment as loads, and constructing a fishery-solar-storage complementary system model; constructing an upper-level objective function for the net load of the fishery-solar-storage complementary system considering the curtailment rate, and performing optimization to obtain the system's net load and energy storage charging / discharging power for each time period; performing a safety assessment of the grid connection risk of the photovoltaic system; if the assessment is not satisfactory, increasing the curtailment rate and performing optimization again; if the assessment is satisfactory, constructing a lower-level objective function for the low-carbon economic operation of the fishery-solar-storage complementary system considering the lowest carbon emissions and overall operating costs, and performing optimization to obtain the output power of each photovoltaic and energy storage device.

[0004] Currently, with the continuous improvement of the intelligence and digitalization level of power systems, optimizing the operation of complex energy systems through multi-source information fusion, dynamic prediction algorithms, and collaborative control strategies has become an important development direction. Under this trend, for the specific application scenario of fishery-solar complementary power generation, researching intelligent control methods that can comprehensively consider power generation, load, and environmental factors, and achieve proactive reactive power demand prediction and multi-resource collaborative regulation has positive practical significance and promotional value for improving system operation reliability, power quality, and overall energy efficiency. Summary of the Invention

[0005] The purpose of this invention is to address the problems existing in the background technology by proposing a reactive power compensation control method and device for fishery-solar complementary systems.

[0006] The technical solution of this invention: a reactive power compensation control method for solar-fishery complementary systems, comprising the following specific implementation steps: S1. Real-time acquisition of photovoltaic power output data, fishery load operation intensity data, and water surface state data in the fishery-solar hybrid system, and time-series alignment of the acquired multi-source data; construction of fishery-solar coupling response quantity based on the relationship between photovoltaic power output and fishery load operation intensity; calculation of operation disturbance correction quantity based on water surface state data; fusion of fishery-solar coupling response quantity and the operation disturbance correction quantity to construct a unified fishery-solar hybrid operation state index; S2. Conduct short-term trend analysis on the solar-fishery complementary operation status index to obtain short-term changes; classify the short-term changes into different evolution patterns based on their magnitude and periodicity; predict the potential reactive power demand increment in the future period based on the current evolution pattern and trend evolution function, and generate a smooth and continuous short-term potential reactive power demand index. S3. Based on the magnitude of the short-term potential reactive power demand index, divide it into different levels of demand; match the corresponding compensation resource type for each demand level and generate a resource priority sequence; based on the resource priority sequence and the real-time status of each compensation resource, determine the optimal compensation path and compensation amount for each resource under the current demand level. S4. Control each compensation resource to perform reactive power compensation according to the optimal compensation path, and monitor the actual compensation output of each resource and the system stability index in real time; calculate the synchronization deviation between the actual output and the predicted demand and the stability deviation of the system stability index; dynamically correct the compensation output of each resource according to the synchronization deviation and the stability deviation. S5. Collect and record historical data on photovoltaic output, fishery load, compensation output, and system operation indicators over a long period of time, and extract long-term trend characteristics; based on the long-term trend characteristics, use a predictive model to optimize the compensation strategy for future periods; evaluate the efficiency of each compensation resource based on the long-term implementation effect, dynamically adjust its weight and priority, and complete the adaptive optimization of the compensation strategy.

[0007] Preferably, in step S1, the specific method for constructing the fish-photonic coupling response is as follows: The changes in active power of the photovoltaic array and the changes in the operating intensity of the fishery load at adjacent sampling times were obtained. The photovoltaic power change and load change are weighted and summed to obtain the photovoltaic coupling response.

[0008] Preferably, in step S2, a short-term trend analysis of the operating status index of the fishery-solar hybrid system is performed, specifically including: The difference in the operating status index at consecutive sampling times is calculated as a short-term change. Weighted moving averages are applied to short-term changes to suppress noise.

[0009] Preferably, in step S2, the evolution mode includes: Slow-evolution type, periodic fluctuation type, and sudden disturbance type; The classification is based on preset thresholds for small and large fluctuations, combined with autocorrelation analysis to determine periodicity.

[0010] Preferably, in step S2, the specific method for predicting the potential increase in reactive power demand in the future period is as follows: The corresponding model coefficients are selected according to the evolution pattern. The model coefficients, the short-term change amount and the trend evolution function are combined for calculation, and the calculation results are subjected to outlier detection and smoothing.

[0011] Preferably, in step S3, the division of the demand hierarchy is specifically as follows: Set thresholds for low-demand layers, upper limit thresholds for medium-demand layers, and safety reference thresholds for high-demand layers; The short-term potential reactive power demand index is compared with the threshold and divided into low demand layer, medium demand layer and high demand layer.

[0012] Preferably, in step S3, the matching of compensation resource types specifically involves: Photovoltaic inverters are used as the primary compensation resource for low-demand tiers. To match the combination of photovoltaic inverters and energy storage systems for mid-demand levels; To match the parallel regulation of photovoltaic inverters, energy storage systems and reactive power compensation capacitors for high-demand layers.

[0013] Preferably, in step S4, the specific method for dynamically correcting the compensation output of each resource is as follows: The comprehensive correction amount for each resource is calculated based on the synchronization deviation and the stability deviation. Based on the resource priority sequence and the comprehensive correction amount, the compensation output instructions for each resource are adjusted, and the adjustment instructions are low-pass filtered.

[0014] Preferably, in step S5, the specific method for extracting long-term trend features is as follows: By performing moving average and seasonal decomposition on historical reactive power demand index data, long-term trend components and seasonal components are obtained. Calculate the correlation coefficients between photovoltaic output, load fluctuations, reactive power demand, and voltage changes to form a long-term trend characteristic matrix.

[0015] The technical solution of this invention: a reactive power compensation control system for solar-fishery complementary systems, used to execute the aforementioned reactive power compensation control method for solar-fishery complementary systems, comprising: The joint sensing and state characterization module is used to synchronously collect photovoltaic power output, fishery load operation intensity and water surface state data and perform time-series alignment, construct fishery-solar coupling response quantity and operation disturbance correction quantity, and then fuse them to generate fishery-solar complementary operation state index. The potential demand identification and dynamic analysis module is used to calculate the short-term changes in the operating status index and classify the evolution patterns. Based on the pattern coefficients and trend evolution functions, it predicts the potential reactive power demand increment and generates a smooth short-term potential reactive power demand index. The short-term compensation strategy generation module is used to divide the short-term potential reactive power demand index into demand layers according to the threshold, match the compensation resource type for each layer and generate a priority sequence, and determine the optimal compensation path and the compensation amount of each resource based on the real-time status of the resources. The dynamic execution and real-time correction module is used to execute the optimal compensation path, monitor the actual output and system indicators, calculate the synchronization error and stability deviation, and dynamically correct the compensation output of each resource. The long-term trend optimization and intelligent enhancement module is used to store historical operation data for a long time, extract long-term trend features, optimize future compensation strategies using predictive models, and adaptively update resource weights and priorities based on long-term execution results.

[0016] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects: This invention designs a reactive power compensation control method and device for solar-fishery complementary systems. By integrating multi-dimensional information such as photovoltaic output, fishery load, and water conditions to construct a comprehensive operating state index, it achieves accurate characterization and advanced perception of the system's dynamic characteristics, significantly improving the predictability and accuracy of reactive power demand identification and overcoming the response lag problem caused by traditional methods relying on static parameters. Secondly, based on a potential demand prediction mechanism using short-term trend analysis and pattern classification, it can effectively identify different fluctuation types of the system and generate smooth and continuous demand guidance, providing a reliable basis for compensation decisions and enhancing the initiative and adaptability of control. Furthermore, by introducing a hierarchical demand matching and multi-resource collaborative scheduling mechanism, it achieves the separation of photovoltaic inverters, energy storage systems, and reactive power compensation devices. The optimized combination and priority allocation of distributed resources form a rapidly responsive and capacity-adaptive distributed control system, significantly improving resource utilization efficiency and economy while ensuring compensation effectiveness. Furthermore, the closed-loop execution and real-time correction mechanisms dynamically monitor compensation output and system stability indicators, ensuring key parameters such as voltage and power factor remain within safe ranges through synchronous adjustment and rapid correction, significantly enhancing the system's anti-disturbance capability and operational reliability. Finally, long-term trend learning and adaptive optimization functions enable the system to continuously accumulate operational experience, dynamically update prediction models and resource weights, achieving continuous self-improvement and intelligent enhancement of the control strategy, providing a stable, efficient, and long-term learning-capable reactive power compensation solution for the fishery-solar hybrid system. Attached Figure Description

[0017] Figure 1 This is a flowchart of a reactive power compensation control method for a solar-fishery complementary system proposed in this invention. Figure 2 This is a module architecture diagram of a reactive power compensation control device for a solar-fishery complementary system proposed in this invention. Detailed Implementation

[0018] Example 1, as Figure 1 As shown, the specific implementation steps of the reactive power compensation control method for solar-fishery complementary systems proposed in this invention are as follows: S1. By jointly sensing and extracting multi-source operational information such as photovoltaic output, fishery load operating rhythm, and water condition in the fishery-solar hybrid system, the focus is on characterizing the "changing trend" of the operational status rather than static electrical parameters. The impact of fishery production behavior and the water environment on photovoltaic stability is mapped into a unified operational status index in an engineering manner. The specific implementation process is as follows: S11. Synchronously collect and time-unify the photovoltaic array power, fishery production load intensity, and water surface condition data. Focus on extracting the rate of change of power and load to characterize the operating rhythm, avoiding the omission of potentially sensitive system states due to asynchronous data or focusing only on instantaneous values. Specifically: During the operation of the solar-fishery complementary system, the output power of the photovoltaic side, the operating status of the water surface, and the operating characteristics of the fishery production load are collected respectively. The information from different sources is aligned with a unified time reference to form a comparable basic operating sequence. Let the basic runtime collected at time t be: ; in, This represents the active power output of the photovoltaic array at time t, which is obtained in real time by the photovoltaic inverter or photovoltaic monitoring system. This represents the change in photovoltaic array power at adjacent sampling times, calculated by differencing historical sampling values. Indicators representing the state of the water surface can be obtained through monitoring water surface disturbances (such as water fluctuation amplitude, reflectivity changes, shading distribution, etc.). This indicator represents the intensity of fishery production load, and is derived from real-time power or operating status data collected from fishery electrical equipment such as aerators and feeding equipment. This represents the set of multi-source basic operational information of the fishery-solar hybrid system collected at time t; S12. Based on the changing relationship between photovoltaic output and the operating rhythm of fishery load, a fishery-photovoltaic coupled response is constructed. By normalizing and superimposing the rate of change, the impact of synchronous or staggered changes between the generation side and the load side on system stability is reflected, enabling reactive power control to detect demand evolution trends in advance. Specifically: After obtaining the basic operating data in a unified time series, we further analyzed the correlation between changes in fishery production activities and fluctuations in photovoltaic output, and extracted key change features that can reflect the turning point in the system's operating state. Constructing runtime coupled response variables: ; in, This represents the coupled response of the solar-fishery operation, used to comprehensively characterize the impact of photovoltaic power output fluctuations and fishery load changes on the system state. The weighting coefficient for the rate of change of photovoltaic power is used to measure the degree of impact of photovoltaic fluctuations on the system's operating status. It can be determined based on historical operating experience or statistical analysis. The weighting coefficient for the rate of change of fishery load is used to measure the degree of impact of load-side changes on the system's operating status. It is obtained through on-site operational surveys or empirical estimations. This indicates the change in the intensity of fishery load operation at adjacent sampling times; S13. The influence of water surface conditions on photovoltaic power output stability is transformed into an operational disturbance correction quantity, which is then amplified and mapped to the sensitivity of photovoltaic power changes. Specifically: Considering that changes in the water surface state can indirectly affect the operational stability of the photovoltaic array through reflection conditions and local shading, a mapping factor of water surface state to electrical fluctuations is introduced to correct the coupling response results. Construct water surface impact correction terms: ; in, This represents the correction amount for operational disturbances caused by the water surface condition, used to correct the coupled response amount to reflect the indirect impact of the water surface on electrical fluctuations; The water surface impact coefficient represents the amplification or buffering effect of water surface disturbance on photovoltaic power output fluctuations. It is obtained through field experience or data fitting based on different fishery-solar hybrid layout conditions and water characteristics. S14. By integrating the fishery-solar coupling response and the water disturbance correction, a unified fishery-solar complementary operating state index is constructed to characterize the system's potential sensitivity to reactive power regulation. This index serves as the core input for subsequent reactive power demand evolution analysis and control strategy selection. Specifically: Based on the aforementioned coupled response and water surface correction quantities, a comprehensive operational status characterization of the fishery-solar complementary system is constructed to indicate the current system's potential sensitivity to reactive power support. Define the comprehensive operating status index: ; in, The index represents the operational status of the solar-fishery complementary system, which comprehensively reflects the impact of photovoltaic power output fluctuations, load changes, and water surface conditions on the system's operational stability.

[0019] S2. Based on the integrated operation status index of fishery-solar hybrid power generation output in step S1, by analyzing its short-term change trend, classifying its evolution mode, predicting trends, and continuously smoothing updates, the potential reactive power demand of the system can be identified in advance. This provides a quantitative and sustainable decision-making basis for subsequent compensation strategy selection, forming an active, state-driven reactive power demand evolution identification method. The specific implementation process is as follows: S21. Short-term trend analysis of the comprehensive operational status index: Calculate the differential change of the status index between consecutive samples, and smooth the fluctuations using a weighted moving average to capture the dynamic fluctuation characteristics of the fishery-solar complementary system on a short timescale, providing a basic quantitative signal for pattern recognition. Specifically: Set short-term sampling interval The value is determined based on the maximum power change rate of the photovoltaic array and the rhythm of the fishery equipment; in this embodiment, the value is taken as 1 to 5 minutes. Calculate the differential change in the continuously acquired state index: ; To suppress noise, in Perform a weighted moving average on the above: ; in, This represents the short-term change in the operating status index, characterizing the performance of the solar-fishery hybrid system. The magnitude of changes in operating status over a period of time; It represents the smoothed short-term change, used to suppress the effects of sampling noise and instantaneous fluctuations; The moving average weighting coefficient is used to adjust the contribution of different historical sample values ​​to the average value; N represents the moving window length, which is the number of historical sample points used to calculate the average. S22. Classification and Feature Extraction of Short-Term Change Patterns: Smoothed short-term changes are classified into three types based on amplitude and periodicity: slowly evolving, periodically fluctuating, and suddenly disturbing. Periodic characteristics are identified through autocorrelation or short-time Fourier transform, providing a basis for different compensation strategies. Specifically: Based on historical operational data statistical analysis of typical short-term fluctuations in solar-aquaculture hybrid systems, threshold parameters are set as follows: (Small fluctuation threshold, used to determine if the system is in a slow evolutionary state) (The threshold for large fluctuations is used to determine if the system is in a state of sudden disturbance). By analyzing the smoothed short-term changes The statistical distribution is analyzed to ensure that the threshold division reflects the true operating characteristics of the system and avoids misjudgment due to transient noise. For each sampling time t, based on the threshold and the change pattern, the short-term changes are discretized into three evolutionary modes: ; Autocorrelation function to determine periodicity: ; like With a certain delay The presence of a significant peak indicates the existence of periodic fluctuations, which can be identified as "periodic fluctuation type". in, This represents the autocorrelation function value, reflecting the short-term change series. In delay The degree of repetition or periodicity in the time series is used to determine whether there are regular fluctuations in the time series, i.e., the identification of periodic fluctuation patterns. The larger the value, the more regular the fluctuation pattern in the time series. The higher the similarity, the more obvious the periodicity; This indicates the type of short-term change pattern, including "slow evolution type", "cyclical fluctuation type", and "sudden disturbance type". S23. Trend Forecasting and Quantification of Potential Reactive Power Demand: Based on the current change pattern, select the pattern coefficient, combine short-term changes with the trend evolution function, and predict the incremental potential reactive power demand within the future time window. This achieves the transformation from change patterns to quantifiable reactive power demand, providing input for compensation resource scheduling. Specifically: Change pattern of output in step S22 Set the corresponding mode-related incremental coefficients : Slow-evolving type: Small coefficients, emphasizing smooth compensation; Cyclical fluctuation type: The coefficient is moderate, taking into account short-term oscillations; Sudden disturbance type: With a large coefficient, it is used for rapid response to potential reactive power demands; Choose the trend evolution function This reflects the changing characteristics of future demand under different models: Slow-evolution type: linear extrapolation or small exponential increase; Cyclical fluctuation type: based on prediction using periodic models (such as sine or autocorrelation fitting); For sudden disturbances: extrapolate exponential decay or rapid trends to ensure timely response; Based on the pattern coefficient and trend function, calculate the potential increase in reactive power demand in the near future: ; For the calculated Perform outlier detection: like Then correct ; like Then correct ; Introducing weighted smoothing: ; in, Indicates the incremental coefficients related to the pattern; This represents a trend evolution function that describes a future time window. A function of latent internal demand changing over time; This indicates the potential increase in reactive power demand, predicted within a time window. The amount of reactive power compensation that the internal system may need represents the intensity of the system’s short-term demand for reactive power support. and These represent the upper and lower limits of potential reactive power demand, and the maximum and minimum allowable range of the predicted value, respectively. This represents the smoothed forecast of potential reactive power demand. This represents the smoothing coefficient, a coefficient that controls the weighting ratio between new and old forecasts; S24. Continuous discrimination and state update of evolutionary models: By weighted fusion of the previous period's predicted value and the current period's potential demand, a smooth and continuous update of the short-term reactive power demand index is achieved, ensuring the stability and reliability of the discrimination results and providing a continuous and sustainable reference for compensation strategy selection. Specifically: The previous prediction value Compared with the current forecast Perform weighted fusion: ; in, This represents the smoothed short-term potential reactive power demand index, used as a reference for continuous control. This represents the smoothing coefficient, which controls the fusion weight of the old and new predictions.

[0020] S3. By addressing the evolutionary characteristics of potential reactive power demand, optimal path selection and adaptive scheduling for multi-resource compensation are achieved. Based on the continuously smoothed potential reactive power demand index output in step S2, demand is layered, resources are matched, dynamic switching is performed, and closed-loop optimization is conducted to ensure efficient, stable, and sustainable compensation, while also considering response speed and cost. This achieves proactive prediction-driven reactive power control. The specific implementation process is as follows: S31. Analyze the magnitude and rate of change of the potential reactive power demand index, divide the demand into three levels: low, medium, and high, and generate corresponding resource priority sequences for each level to realize the basis of hierarchical scheduling. Specifically: Set stratification threshold , , Based on historical fluctuations in photovoltaic power output, patterns in fishery loads, and dynamic determination of system capacity, three demand levels are established: Potential reactive power demand index Corresponding to hierarchical classification: ; Calculate the rate of change of the demand index: ; According to the demand layer Based on the rate of change in the demand index, allocate available compensation resource types and priorities: Low-demand tier: Primarily uses photovoltaic inverters with adjustable reactive power, fast response speed, and low cost; Mid-demand layer: Photovoltaic inverter + energy storage system combination compensation to ensure a balance between capacity and response; High-demand layer: Photovoltaic inverter + energy storage system + main transformer or compensation capacitor are adjusted in parallel to achieve rapid large-capacity compensation; And generate a priority sequence for each level: ; in, This represents the threshold for the low-demand layer, used to determine the lower limit of potential reactive power demand within the "low-demand layer". This indicates the upper limit threshold of the demand layer, meaning that rapid reactive power adjustment alone is insufficient to fully cover the demand, and an additional compensation capacity needs to be introduced to reach the critical boundary. This represents the safety reference threshold for the high-demand layer, used to define the warning range where there may be risks of voltage exceeding limits or power factor deterioration. It is determined by comprehensively considering grid connection specifications, transformer capacity limits, and compensation device limits, and is used to trigger strong intervention strategies. This indicates the rate of change in the demand index; Represents the resource number, sorted by a combination of response speed, capacity, and cost; Indicates a sequence of resource priorities; This represents the potential demand layer; S32. Match available compensation resources according to each demand level, construct a resource matrix, select the optimal compensation path that meets capacity, response speed, and cost constraints, and realize multi-resource joint scheduling, specifically as follows: Obtain the real-time available reactive power capacity of each resource. Response rate and cost indicators Establish a resource matrix: ; Perform optimal path search for each level of requirement L: Define constraints: ; Define the optimization objective: minimize It balances response speed and energy consumption; A multi-resource parallel compensation strategy is adopted for high-demand layers, while photovoltaic inverters are used for fine-tuning in low-demand layers, forming a "layered matching + parallel adjustment" mechanism. in, This indicates that at time t, the maximum reactive power regulation capacity that the i-th compensation resource can provide is determined by the rated capacity of the equipment, the current operating conditions, and the margin limit. This represents the response rate of the i-th resource, which is the response speed required for the compensation resource to complete the reactive power output change from receiving the control command. It is determined by the equipment type and control characteristics, such as the inverter responding quickly and the mechanical switching device responding slowly. This represents the cost index of the i-th resource, which is the comprehensive cost of using the resource for reactive power compensation. It is set by comprehensively considering factors such as energy loss, equipment lifespan, and maintenance frequency. This represents the resource matching matrix under the demand level, describing the matching relationship between compensation demand and the capabilities of various resources under a certain reactive power demand level L. This indicates resource efficiency, which is determined based on equipment characteristics (such as inverter efficiency, capacitor losses, etc.). Indicates the amount of compensation for each resource; This indicates the reactive power demand at the current demand level. This represents the response speed weighting coefficient, used to comprehensively consider response speed and cost optimization; N represents the total number of resources. S33. Real-time monitoring of the error between actual compensation and predicted demand. When the error exceeds a threshold, dynamic path switching is triggered, prioritizing the adjustment of resources with fast response and low cost to ensure that the compensation strategy adapts to demand fluctuations in real time. Specifically: Real-time monitoring of total compensation amount With potential demand Error: ; like Triggering the path switching mechanism: Prioritize resources with fast response times and low costs; If the demand layer is high and the error is large, then the auxiliary reactive power device will be activated. It should be noted that the previous path state is recorded during the switch to avoid system oscillation caused by frequent switching; in, This indicates the total amount of compensation currently being implemented; This indicates the error between demand and actual compensation. This indicates the error tolerance threshold, which is set according to system stability requirements and the allowable range of voltage fluctuations. S34. The execution path is corrected through closed-loop feedback, the compensation amount of each resource is adjusted according to the correction coefficient, and the resource priority is updated to achieve adaptive optimization and long-term stable operation, ensuring that compensation is synchronized with demand. Specifically: Calculate the actual compensation correction factor: ; Adjust the compensation amount for each resource: ; Update the resource matrix and priority weights to achieve adaptive optimization: Resources under prolonged high load will automatically reduce their usage frequency. Resources with fast response times should be prioritized during periods of frequent fluctuations. The revised resource allocation is fed back to the path selection in the next cycle, forming a closed-loop optimization cycle; in, This represents the compensation correction coefficient, used to adjust the actual allocation to match the predicted demand; Indicates a small quantity to prevent division by zero; This indicates the compensation amount for each resource in the next cycle, and the current allocation is adjusted according to the correction coefficient.

[0021] S4. Through four stages—real-time monitoring, dynamic stability constraints, rapid correction, and feedback optimization—a closed loop of prediction, execution, and optimization is formed to ensure that the compensation output matches the potential demand, while also taking into account system security and resource utilization efficiency. The specific implementation process is as follows: S41. Real-time monitoring of the output and potential demand index of each compensation resource, calculation of synchronization deviation, and weight correction of resources with lagging response to ensure that multi-resource compensation actions are synchronized with demand evolution, avoiding voltage fluctuations and over-compensation without reactive power, specifically: The current compensation resource combination and output quantity are collected in step S3. Including but not limited to photovoltaic inverters, energy storage systems, and reactive power regulators for main transformers; Calculate the actual output and predicted demand of each resource. Deviation: ; Introduce advance / delay adjustment strategies for resources with response lag: ; Establish a synchronous monitoring and control cycle relationship, with the control cycle typically being slightly shorter than the sampling cycle, to ensure that the compensation output can quickly follow changes in demand. in, This represents the real-time reactive power output of the i-th compensation resource at time t, which is obtained through real-time monitoring data of each compensation resource, such as the output of photovoltaic inverters, energy storage systems, and reactive power regulators of main transformers. This represents the synchronization deviation of the i-th resource, i.e., the difference between the actual output and the predicted demand under this resource allocation ratio; This represents the weight of the i-th resource in the total compensation demand; This represents the adjustment output for the i-th resource at the next time step; This represents the synchronization deviation adjustment coefficient, used to amplify or reduce the adjustment amount, and is set according to the resource response characteristics (speed, capacity limit); S42. Establish multi-index dynamic stability constraints, including voltage, current, power factor, and total compensation, set upper and lower limits, and trigger safety constraints for exceeding limits or potential unstable states to proactively address rapid disturbances and ensure the system operates within a safe range. Specifically: Key metrics to collect: ; Set dynamic upper and lower limits for each indicator: Voltage Power factor The line current does not exceed 80%~90% of the rated capacity, and short-time voltage fluctuations... ; Stability deviation for exceeding the limit index: ; If the deviation exceeds the set threshold, the following constraints are triggered: reduce the output of some compensation resources; activate auxiliary devices (such as capacitor bank quick-connection); adjust the resource combination path to ensure system stability as a priority. in, Represents a set of dynamic stability indices; This represents the effective value of the system voltage, i.e., the real-time voltage sampling. Indicates the actual current in the line; Indicates the system power factor; This represents the current total compensation amount of the system, which is the sum of the outputs of each compensation resource. This represents the short-time voltage fluctuation amplitude, calculated by sampling and differential calculation using a voltage sensor. It represents a certain stability index of the system, such as voltage, current or power factor; Indicates the upper and lower limits or safety thresholds of the corresponding indicator; This indicates the degree to which a certain indicator exceeds its limit; S43. Calculate the compensation correction amount for each resource based on the synchronization deviation and stability deviation, adjust the output according to priority, and combine with low-pass filtering to prevent frequent switching or oscillation, thereby realizing dynamic correction and closed-loop response of compensation execution, specifically as follows: Calculate the overall correction amount: ; Adjust the compensation output of each resource according to priority to ensure that the total compensation amount is close to the expected value. At the same time, it satisfies stability constraints; Low-pass filtering is applied to correct for short-term fluctuations: ; in, This represents the total number of stability indicators, such as voltage deviation, current deviation, power factor deviation, transient fluctuation deviation, etc. This represents the compensation adjustment amount for the i-th resource; Indicates the synchronization deviation adjustment weight; This represents the stability deviation adjustment weight, which controls the degree of impact of stability deviation on resource adjustment. This indicates the correction amount after low-pass filtering; This indicates the length of the low-pass filter window, which is set according to the dynamic characteristics of the system. S44. Evaluate the effectiveness of compensation execution by comparing the actual output with the predicted potential demand to calculate the error, and feed the execution results back to the path selection and resource matching matrix to update priorities and adjustment coefficients, thereby achieving long-term adaptive optimization and system robustness improvement. Specifically: Real-time collection of system metrics after compensation execution , , ; Calculate the execution error: ; Update resource allocation weights, priorities, and correction coefficients based on execution error and stability metrics; Statistical analysis of long-term trends can identify potential resource overload or shortage, allowing for proactive adjustment of the resource matching matrix and adaptive optimization. in, This indicates the deviation between the actual total compensation amount and the predicted demand.

[0022] S5. Through long-term data acquisition, trend analysis, predictive optimization, adaptive weight adjustment, and closed-loop intelligent enhancement, the system achieves comprehensive control from short-term dynamic response to long-term adaptive optimization of the solar-fishery complementary reactive power compensation system. This improves the system's operating efficiency, stability, and intelligence level, while ensuring sustainable strategy iteration and forward-looking adjustment capabilities. The specific implementation process is as follows: S51. By collecting long-term data on photovoltaic output, fishery load, compensation output, power grid indicators, and environmental data, a multi-dimensional trend feature matrix is ​​constructed to extract periodic, seasonal, and long-term variation patterns, providing a data foundation for subsequent forecasting and compensation strategy optimization. Specifically: The following metrics are recorded in real time, including but not limited to: photovoltaic output power. Fishery load power Total reactive power compensation ,Voltage Power factor PF(t), environmental variables (solar radiation intensity) Temperature ); Historical data Perform moving average and seasonal decomposition: ; Using the cosine similarity algorithm, the correlation coefficients between key indicators are calculated to identify the long-term coupling relationship between photovoltaic power output, load fluctuation, reactive power demand and voltage change. Forming a long-term trend characteristic matrix Record the trend, periodic amplitude, and fluctuation range of each indicator to provide a reference for prediction and optimization; in, This represents the actual active power output of the photovoltaic module at time t; This represents the actual active power of the fishery load at time t; This represents the total reactive power compensation of the system at time t; PF(t) represents the effective value of the system voltage; PF(t) represents the system power factor. Indicates light intensity; Indicates ambient temperature; This represents the historical potential reactive power demand index; Indicates the long-term trend component; The seasonal component is derived from the periodic decomposition of historical data; This indicates the length of the moving average window used for trend decomposition, i.e., how many consecutive moments of historical data are used to calculate the average trend. It represents a long-term trend characteristic matrix, including multi-dimensional characteristics such as the trend of each indicator, periodic amplitude, and fluctuation range; S52. Utilize the long-term trend feature matrix to predict potential reactive power demand, and combine it with the Long Short-Term Memory (LSTM) network model to proactively optimize the allocation, priority, and adjustment coefficients of compensation resources, and pre-schedule resources during peak periods to reduce response latency. Specifically: Long-term trend feature matrix Input an enhanced LSTM model to predict potential demand in future periods. ; Based on the predicted potential demand, optimize the output, priority, and adjustment coefficient of each compensation resource: ; Pre-allocate the output of the energy storage system and the reactive capacity of the main transformer during periods of potential high demand to reduce compensation delays and fluctuations. Output optimization strategy matrix, forming Record the forward-looking allocation strategy of each resource, including output, priority and adjustment coefficient, to provide a reference for short-term closed-loop control; in, This represents the index indicating the projected future potential reactive power demand. This represents the optimized compensation strategy matrix, which includes the output, priority, and adjustment coefficient of each compensation resource. This indicates that the i-th compensation resource is at the predicted time. The optimized compensation amount is achieved by optimizing the mapping function. Generated based on predicted demand and strategies; This represents an optimized mapping function that maps future demand, resource capabilities, and weight matrices to specific outputs. This represents the adaptive weight matrix, calculated using historical execution efficiency and stability metrics. It should be noted that optimizing the mapping function It is an intelligent decision-making mechanism used to map the predicted future reactive power demand, the adaptive weights of each compensation resource, and the current strategy state into specific compensation outputs. This function dynamically generates the output and adjustment coefficient of each compensation device by comprehensively considering future load change trends, photovoltaic power output fluctuations, resource response capabilities, and historical execution efficiency, so as to achieve reactive power compensation that combines foresight, continuity, and stability. The function adopts multi-dimensional data fusion and adaptive adjustment strategies to enable resource priorities to be automatically optimized with long-term operating performance, while ensuring the smoothness of the compensation process and system stability, so as to achieve intelligent compensation control that can learn itself and optimize in the long term. S53. Evaluate the efficiency of each compensation resource based on long-term execution error and stability indicators, dynamically adjust resource weights, priorities, and adjustment coefficients to form an adaptive weight matrix, thereby strengthening and optimizing the compensation strategy in the long term, enabling efficient resource utilization and enhancing system robustness. Specifically: Combined with historical execution errors and stability indicators Quantify the long-term efficiency of each resource: ; Resources that are inefficient, frequently overloaded, or slow to respond are de-prioritized, while high-performing resources are given higher weights, forming an adaptive weight matrix. ; Combining forward-looking forecast results Dynamically adjust the adjustment coefficients of each resource and the path priority to enhance and optimize the compensation strategy; Forming a long-term adaptive optimization strategy matrix This is used to guide the implementation and dynamic adjustment of short-term compensation; in, This represents the total number of stability indicators; Indicates the length of the long-term statistical window; This represents the actual reactive power output of the i-th compensation resource at historical time k. This represents the long-term efficiency index of the i-th compensation resource, which measures the long-term contribution and performance of the resource and is used for adaptive weight adjustment. This represents the long-term optimization strategy matrix; S54. By combining long-term optimization strategies and adaptive weights with short-term compensation closed loops, the prediction model, resource allocation, and adjustment strategies are iteratively updated in real time to form a long-term adaptive intelligent closed loop. This enables the system to achieve continuous optimization, stable and efficient operation, and intelligent enhancement capabilities. Specifically: Will and Combine input short-term path selection (step S3) and compensation execution correction (step S4). Based on actual execution results, long-term trend deviations, and short-term errors, the prediction model, weight matrix, and resource priorities are iteratively updated to achieve system self-adjustment. The strategy matrix and correction coefficients are periodically updated based on seasonal load, photovoltaic changes, and long-term trends. As runtime increases, the system automatically optimizes prediction accuracy, resource allocation, and adjustment strategies to achieve long-term intelligent enhancement and stable, efficient operation.

[0023] Example 2, as Figure 2 As shown, the present invention proposes a reactive power compensation control device for fishery-solar complementary systems, which is used to execute a reactive power compensation control method for fishery-solar complementary systems proposed in Embodiment 1. The device includes: a joint sensing and state characterization module, a potential demand identification and dynamic analysis module, a short-term compensation strategy generation module, a dynamic execution and real-time correction module, and a long-term trend optimization and intelligent enhancement module.

[0024] The joint sensing and state characterization module is responsible for collecting multi-source data from the solar-fishery hybrid system in real time, including but not limited to photovoltaic output, fishery load, voltage, current, power factor, and environmental parameters such as light intensity and temperature. The module integrates a data processing unit to denoise, normalize, and extract features from the collected data, thereby achieving a joint characterization of the system's operating state and generating multi-dimensional feature data. The potential demand identification and dynamic analysis module, based on multi-dimensional feature data, makes high-precision predictions of load fluctuations and photovoltaic output in the short term, and performs dynamic response calculations by combining system voltage stability and power factor changes. The module adopts an adaptive prediction algorithm and a real-time evaluation mechanism to generate early warnings and adjustment strategies for possible load peaks or photovoltaic fluctuations, providing short-term optimization targets for reactive power compensation. The short-term compensation strategy generation module generates specific compensation strategies based on short-term forecast results and historical execution data, combined with the priority, capacity, and response characteristics of each compensation resource. These strategies include the output amount, adjustment path, and adjustment order of each resource. The dynamic execution and real-time correction module is responsible for executing the compensation strategy output by the strategy generation module among various resources and monitoring the execution effect in real time. By comparing the target with the actual output, it calculates the execution error and system stability index, and dynamically corrects the strategy. The module integrates a feedback analysis unit and an adaptive adjustment unit, which adjust the output priority and adjustment coefficient of each compensation resource according to the real-time feedback to achieve short-term closed-loop control. The long-term trend optimization and intelligent enhancement module uses historical operating data and trend characteristics to analyze the long-term evolution trend of the system, and combines predicted output and execution feedback to optimize and intelligently enhance the compensation strategy; it adjusts the weights of each resource and the strategy matrix according to long-term performance to achieve forward-looking optimization and self-reinforcing control.

[0025] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A fish light complementary reactive power compensation control method, characterized in that, The specific implementation steps include the following: S1. Real-time acquisition of power data of photovoltaic output, operation intensity data of fishery load and water surface status data of fishery-solar complementary system, and time-series alignment of the acquired multi-source data; A fishery-solar coupled response quantity is constructed based on the relationship between the changes in photovoltaic output and the intensity of fishery load operation. The operational disturbance correction amount is calculated based on the water surface state data; the fishing-solar coupling response amount and the operational disturbance correction amount are integrated to construct a unified fishing-solar complementary operational state index. S2. Conduct short-term trend analysis on the solar-fishery complementary operation status index to obtain short-term changes; classify the short-term changes into different evolution patterns based on their magnitude and periodicity; predict the potential reactive power demand increment in the future period based on the current evolution pattern and trend evolution function, and generate a smooth and continuous short-term potential reactive power demand index. S3. Based on the magnitude of the short-term potential reactive power demand index, divide it into different levels of demand; match the corresponding compensation resource type for each demand level and generate a resource priority sequence; based on the resource priority sequence and the real-time status of each compensation resource, determine the optimal compensation path and compensation amount for each resource under the current demand level. S4. Control each compensation resource to perform reactive power compensation according to the optimal compensation path, and monitor the actual compensation output of each resource and the system stability index in real time; calculate the synchronization deviation between the actual output and the predicted demand and the stability deviation of the system stability index; dynamically correct the compensation output of each resource according to the synchronization deviation and the stability deviation. S5. Collect and record historical data on photovoltaic output, fishery load, compensation output, and system operation indicators over a long period of time, and extract long-term trend characteristics; based on the long-term trend characteristics, use a predictive model to optimize the compensation strategy for future periods; evaluate the efficiency of each compensation resource based on the long-term implementation effect, dynamically adjust its weight and priority, and complete the adaptive optimization of the compensation strategy.

2. The reactive power compensation control method for solar-fishery complementary systems according to claim 1, characterized in that, In step S1, the specific method for constructing the fish-optical coupling response is as follows: The changes in active power of the photovoltaic array and the changes in the operating intensity of the fishery load at adjacent sampling times were obtained. The photovoltaic power change and load change are weighted and summed to obtain the photovoltaic coupling response.

3. The reactive power compensation control method for solar-fishery complementary systems according to claim 2, characterized in that, In step S2, a short-term trend analysis of the operating status index of fishery-solar hybridization is performed, specifically including: The difference in the operating status index at consecutive sampling times is calculated as a short-term change. Weighted moving averages are applied to short-term changes to suppress noise.

4. The reactive power compensation control method for solar-fishery complementary systems according to claim 3, characterized in that, In step S2, the evolutionary modes include: Slow-evolution type, periodic fluctuation type, and sudden disturbance type; The classification is based on preset thresholds for small and large fluctuations, combined with autocorrelation analysis to determine periodicity.

5. The reactive power compensation control method for solar-fishery complementary systems according to claim 4, characterized in that, In step S2, the specific method for predicting the potential increase in reactive power demand in the future period is as follows: The corresponding model coefficients are selected according to the evolution pattern. The model coefficients, the short-term change amount and the trend evolution function are combined for calculation, and the calculation results are subjected to outlier detection and smoothing.

6. The reactive power compensation control method for solar-fishery complementary systems according to claim 5, characterized in that, In step S3, the division of the requirement hierarchy is as follows: Set thresholds for low-demand layers, upper limit thresholds for medium-demand layers, and safety reference thresholds for high-demand layers; The short-term potential reactive power demand index is compared with the threshold and divided into low demand layer, medium demand layer and high demand layer.

7. The reactive power compensation control method for solar-fishery complementary systems according to claim 6, characterized in that, In step S3, the matching of compensation resource types specifically involves: Photovoltaic inverters are used as the primary compensation resource for low-demand tiers. To match the combination of photovoltaic inverters and energy storage systems for mid-demand levels; To match the parallel regulation of photovoltaic inverters, energy storage systems and reactive power compensation capacitors for high-demand layers.

8. The reactive power compensation control method for solar-fishery complementary systems according to claim 7, characterized in that, In step S4, the specific method for dynamically correcting the compensation output of each resource is as follows: The comprehensive correction amount for each resource is calculated based on the synchronization deviation and the stability deviation. Based on the resource priority sequence and the comprehensive correction amount, the compensation output instructions for each resource are adjusted, and the adjustment instructions are low-pass filtered.

9. The reactive power compensation control method for solar-fishery complementary systems according to claim 8, characterized in that, In step S5, the specific method for extracting long-term trend features is as follows: By performing moving average and seasonal decomposition on historical reactive power demand index data, long-term trend components and seasonal components are obtained. Calculate the correlation coefficients between photovoltaic output, load fluctuation, reactive power demand and voltage changes to form a long-term trend characteristic matrix.

10. A reactive power compensation control system for a fishery-solar hybrid system, used to execute the reactive power compensation control method for a fishery-solar hybrid system as described in any one of claims 1 to 9, characterized in that, include: The joint sensing and state characterization module is used to synchronously collect photovoltaic power output, fishery load operation intensity and water surface state data and perform time-series alignment, construct fishery-solar coupling response quantity and operation disturbance correction quantity, and then fuse them to generate fishery-solar complementary operation state index. The potential demand identification and dynamic analysis module is used to calculate the short-term changes in the operating status index and classify the evolution patterns. Based on the pattern coefficients and trend evolution functions, it predicts the potential reactive power demand increment and generates a smooth short-term potential reactive power demand index. The short-term compensation strategy generation module is used to divide the short-term potential reactive power demand index into demand layers according to the threshold, match the compensation resource type for each layer and generate a priority sequence, and determine the optimal compensation path and the compensation amount of each resource based on the real-time status of the resources. The dynamic execution and real-time correction module is used to execute the optimal compensation path, monitor the actual output and system indicators, calculate the synchronization error and stability deviation, and dynamically correct the compensation output of each resource. The long-term trend optimization and intelligent enhancement module is used to store historical operation data for a long time, extract long-term trend features, optimize future compensation strategies using predictive models, and adaptively update resource weights and priorities based on long-term execution results.

Images