A feedback regulation system for reactive zero of new energy station grid-connected point

Abstract

This application relates to the field of power system automation control technology, specifically disclosing a feedback regulation system for reactive power zeroing at the grid connection point of renewable energy power plants, aiming to solve the problems of regulation lag and insufficient accuracy caused by renewable energy fluctuations. The system includes: a multi-dimensional high-frequency sensing unit for real-time extraction of reactive power components across the entire frequency domain; a power flow trend prediction unit for predicting reactive power demand trends and generating advance compensation commands; a collaborative allocation optimization unit for task allocation to heterogeneous equipment based on a multi-objective optimization model; a closed-loop feedback fine-tuning unit for real-time correction of residual deviations using adaptive regulation logic; and a health status monitoring unit for assessing equipment status and dynamically adjusting its participation weight. Through this system, this application achieves proactive prediction and precise compensation of reactive power, significantly improving the real-time tracking capability of reactive power zeroing at the grid connection point and the reliability of multi-equipment collaborative operation.

Description

Technical Field

[0001] This invention belongs to the field of power system automation control technology, specifically relating to a feedback regulation system for achieving zero reactive power at the grid connection point of new energy power plants. Background Technology

[0002] With the transformation of the global energy structure, new energy power generation technologies, represented by wind and solar power, have developed rapidly. New energy power plants are connected to the power system through grid connection points, and their operational stability has a significant impact on the power quality and voltage level of the grid. Reactive power management, as a core component of maintaining grid voltage balance, is directly related to the transmission efficiency and equipment operation safety of the power system. Implementing precise reactive power control at the grid connection point can effectively suppress voltage fluctuations and reduce line losses, making it a key component of modern smart grid construction.

[0003] Among these, reactive power zeroing regulation at the grid connection point of renewable energy power plants is an important technical direction in reactive power management. It aims to bring the reactive power exchange between the grid connection point and the power grid close to zero by dynamically adjusting the reactive power compensation equipment within the power plant. This technology is typically based on feedback control principles, monitoring the reactive power status of the grid connection point in real time and issuing adjustment commands to equipment such as inverters or static var generators within the power plant accordingly. Achieving real-time reactive power balance not only improves the operational efficiency of the power plant but also significantly enhances the power system's ability to accommodate fluctuating power sources.

[0004] Current technologies still face multiple technical challenges in achieving zero reactive power at grid connection points. Due to the high volatility and randomness of renewable energy output, traditional reactive power regulation methods often exhibit response lag, making it difficult to track instantaneous reactive power changes. This results in large reactive power offsets at the grid connection point and difficulty in guaranteeing regulation accuracy. Existing feedback regulation algorithms often lack sufficient robustness when handling complex operating conditions with multiple constraints, making them susceptible to interference from grid impedance changes and signal measurement noise, leading to oscillations or frequent limit violations in the control system. Furthermore, the coordinated scheduling logic among multiple reactive power sources within the power plant is not optimized enough, failing to fully utilize the response characteristics of each compensation device at different frequency bands, making it difficult to maintain a stable zero reactive power state during periods of severe load fluctuations. These problems limit the further improvement of renewable energy power plants' support capacity for the grid, necessitating a reactive power zeroing feedback regulation system capable of achieving high precision, rapid response, and adaptive capabilities. Summary of the Invention

[0005] The purpose of this invention is to provide a feedback regulation system for zeroing reactive power at the grid connection point of new energy power plants, so as to solve the problems of regulation lag, insufficient accuracy and low efficiency of multi-device coordination in the existing technology when the output of new energy is highly volatile, and to achieve high-precision dynamic balance and zeroing control of reactive power at the grid connection point.

[0006] The technical solution of this invention includes: a multi-dimensional high-frequency sensing unit, used to acquire electrical parameters of the grid connection point of the new energy power station and each branch within the station in real time; by instantaneously sampling voltage and current signals, it extracts the full-frequency reactive power components, including fundamental reactive power, harmonic reactive power, and impulsive reactive power, providing high-precision raw data input for subsequent regulation; a power flow trend prediction unit, used to predict the reactive power demand change trend in the very short time scale of the future based on historical sequence data acquired by the multi-dimensional high-frequency sensing unit, combined with the real-time active power output fluctuation slope within the station, using a dynamic spatiotemporal correlation analysis algorithm, and generating advance compensation commands; and a collaborative allocation optimization unit, used to receive the advance compensation commands generated by the power flow trend prediction unit. The system generates compensation commands and establishes a multi-objective optimization scheduling model based on the response bandwidth, remaining capacity, and loss characteristics of different types of reactive power compensation equipment within the power station. This model decomposes the total reactive power demand and allocates it to each specific compensation terminal. A closed-loop feedback fine-tuning unit monitors the residual reactive power deviation at the grid connection point in real time after the commands are executed. Through adaptive proportional-integral-derivative adjustment logic, it corrects the output of the collaborative allocation optimization unit in real time, ensuring that the reactive power exchange at the grid connection point continuously converges to near zero. A health status monitoring unit assesses the operating temperature rise, switching frequency, and electrical fatigue of each reactive power compensation device within the power station in real time and feeds the assessment results back to the collaborative allocation optimization unit to dynamically adjust the participation weight and adjustment priority of the devices.

[0007] In one embodiment of the present invention, the multi-dimensional high-frequency sensing unit is implemented by a synchronous sampling module configured at the grid connection point, with its sampling frequency set to above 10,000 Hz. By performing sliding window integration calculation on the collected discrete voltage and current sequences, the forced reactive power component caused by voltage fluctuations and the distorted reactive power component caused by nonlinear loads are separated. At the same time, the unit also independently decouples the positive sequence, negative sequence, and zero sequence components at the grid connection point to support subsequent precise adjustment for asymmetrical operating conditions.

[0008] Furthermore, the power flow prediction unit integrates a variable step size differential analysis algorithm. This algorithm identifies the starting point and acceleration characteristics of reactive power fluctuations by calculating the power change rate between the current moment and the previous three sampling periods. By matching the real-time power curve with a pre-stored library of typical disturbance features, the unit can identify systemic reactive power fluctuations caused by drastic changes in wind speed or sudden changes in illumination. Based on this, it generates a compensation reference signal with a preset amplitude before the physical response lag occurs, thus realizing the transformation from passive following to active prediction.

[0009] In one embodiment of the present invention, the collaborative allocation optimization unit adopts a hierarchical allocation logic based on response speed priority. For high-frequency pulse reactive power demand with a frequency higher than 10 Hz, it is preferentially allocated to grid-connected inverter groups with fast adjustment capabilities, utilizing the extremely fast response characteristics of their control loops for offsetting. For steady-state or slowly changing reactive power demand with a frequency lower than 2 Hz, it is preferentially allocated to static var generators or switched capacitor banks with larger capacity. During the allocation process, the unit calculates the marginal adjustment losses of each device in real time, and selects the device combination scheme with the minimum total energy loss under the premise of meeting the zeroing target.

[0010] Furthermore, when handling multi-inverter parallel operation, the collaborative allocation optimization unit introduces a virtual impedance adjustment mechanism to artificially change the equivalent output impedance of each inverter at the grid connection point, thereby achieving automatic balancing of reactive load among multiple devices and preventing individual devices from being overloaded while others are idle. At the same time, the unit also performs remote voltage drop compensation on the issued commands based on the line impedance parameters of each branch, ensuring that the actual output after the command reaches the execution terminal can accurately correspond to the compensation requirements of the grid connection point.

[0011] In one embodiment of the present invention, the closed-loop feedback fine-tuning unit senses the impedance changes on the grid side in real time by constructing a control operator based on dynamic gain adjustment. When the system stability decreases due to a weakening of the grid strength, the unit automatically reduces the proportional coefficient of the feedback loop to enhance the damping ratio of the system and suppress possible low-frequency oscillations. When the reactive power offset at the grid connection point exceeds the set 1% dead zone threshold, the unit activates the fast correction mode and forces the reactive power to fall back to the zero value center region by superimposing an instantaneous pulse increment.

[0012] Furthermore, the closed-loop feedback fine-tuning unit has an error self-learning function, which can perform statistical analysis on the adjustment residuals over the past 24 hours, identify fixed offsets with periodic characteristics, and embed them as compensation benchmarks into the initial adjustment amount. In this way, the steady-state offset caused by sensor system errors or long-distance transmission delays can be offset.

[0013] In one embodiment of the present invention, the health status monitoring unit performs multi-dimensional correlation analysis on the casing temperature, bus voltage ripple, and current harmonic distortion rate of each power semiconductor device, and calculates the health score of each compensation device in real time. When the health score of a device is lower than 80, the collaborative allocation optimization unit will automatically reduce the reactive power output limit of the device. When the health score is lower than 60, the device will be cut off from the regulation sequence for forced cooling or maintenance, and its regulation share will be automatically taken over by other redundant devices, thereby ensuring the continuous and reliable operation of the entire system.

[0014] Furthermore, the system operates on a communication architecture based on microsecond-level deterministic industrial Ethernet, and the transmission delay of all sensor data and control commands is strictly controlled within 2 milliseconds; the internal time synchronization accuracy of the system reaches 1 microsecond level, ensuring that multiple sensing points and execution points deployed in different locations can coordinate their actions under a unified time reference.

[0015] As one embodiment of the present invention, the system also includes an offline simulation verification module, which synchronizes on-site operation data in real time and constructs a digital twin model of the station in a virtual space; by rehearsing adjustment strategies under extreme working conditions in the model, the module can continuously optimize the weight parameters in the collaborative allocation optimization unit and periodically update the optimized parameter set to the online operation system.

[0016] Furthermore, the system has the capability to provide reactive power support during low-voltage ride-through and high-voltage ride-through processes by real-time tracking of the voltage waveform at the grid connection point. During a grid fault, the system automatically switches from reactive power zeroing mode to emergency voltage support mode, rapidly injecting inductive or capacitive reactive power according to the proportional coefficient required by the grid dispatching specifications. After the fault is eliminated and the voltage returns to the normal range, the system smoothly switches back to reactive power zeroing feedback regulation mode.

[0017] As one embodiment of the present invention, the multi-dimensional high-frequency sensing unit uses an improved phase-shift-free filtering algorithm based on an integral chain when extracting reactive power components. This algorithm can effectively filter out the fundamental frequency deviation noise caused by grid frequency fluctuations, ensuring that the phase error of the extracted reactive power is less than 0.1 degrees within the frequency variation range of 45 Hz to 55 Hz. This high-precision phase capture capability is the physical basis for achieving precise zeroing of the grid connection point.

[0018] Furthermore, when performing reactive power allocation, the collaborative allocation optimization unit also considers the tap position status of the transformers in the station; through coordinated control with the on-load tap-changing transformer, the system can optimize the voltage distribution in a larger spatial range, reduce the problem of limited reactive power output of equipment due to excessively high or low voltage, and increase the availability of reactive power regulation capacity to over 95%.

[0019] As one embodiment of the present invention, the system also has an environmental adaptive adjustment logic, which can automatically adjust the overload multiple limit of power equipment such as inverters based on the ambient temperature and humidity data obtained by meteorological sensors; under low ambient temperature conditions, the system is allowed to extract reactive current exceeding 1.2 times the rated value for a short time to cope with extreme instantaneous fluctuations and improve the robustness of the system under complex climatic conditions.

[0020] Furthermore, the integral element of the closed-loop feedback fine-tuning unit adopts anti-saturation processing technology. When the adjustment command reaches the physical output limit of the device, the accumulation of integral increments is automatically stopped, avoiding large-scale adjustment overshoot when the system exits the saturation state, thus ensuring the smoothness and stability of the adjustment process.

[0021] Compared with the prior art, the advantages and positive effects of the present invention are as follows: This invention introduces a power flow trend prediction unit, changing the passive mode of traditional feedback regulation that can only respond to deviations that have already occurred. Through differential analysis and feature matching, the system can predict the reactive power change trend in a very short time and implement advance compensation, effectively overcoming the control lag caused by communication delay, calculation time and the response inertia of power electronic devices. This reduces the dynamic deviation range of reactive power at the grid connection point by more than 60%, and significantly improves the real-time tracking capability for reactive power zeroing.

[0022] This invention achieves efficient coordination of heterogeneous reactive power sources within a power station through a collaborative allocation optimization unit. The system breaks away from the previous isolated adjustment of each device and performs frequency domain decoupling allocation based on the physical characteristics of the devices. This leverages both the high-speed adjustment advantage of the inverter and the capacity advantage of the large reactive power compensation equipment. This hierarchical allocation architecture not only improves adjustment accuracy but also reduces the overall operating losses of the power station and extends the service life of key components by optimizing the participation weight of the devices.

[0023] This invention constructs a comprehensive sensing system based on multi-dimensional high-frequency sensing and health status monitoring, enabling the system to maintain extremely high robustness in complex power grid environments. The closed-loop feedback fine-tuning unit can automatically optimize the control gain according to changes in power grid strength through adaptive parameter adjustment, effectively solving the control oscillation problem that traditional systems are prone to cause under weak power grid conditions. At the same time, the introduction of a health evaluation mechanism ensures the safe and smooth transfer of regulation tasks among various devices, greatly enhancing the continuity and reliability of new energy power plants' support for the power grid.

[0024] This invention does not rely on expensive external dispatch communication links at all. By constructing a high-precision sensing and feedback closed loop inside the power station, it realizes autonomous management and zero-level control of reactive power at the grid connection point. This highly integrated intelligent regulation scheme provides a solid technical guarantee for the large-scale grid connection of new energy sources, significantly improves the grid's ability to accept fluctuating power sources, and has important technical driving value for optimizing the energy structure and ensuring the safe and stable operation of the power system.

[0025] This invention achieves comprehensive coverage of complex operating conditions such as nonlinearity and asymmetry by combining fully digital control logic with high-frequency sampling technology. The system can not only eliminate fundamental reactive power, but also effectively suppress false reactive power caused by harmonics, ensuring that the grid connection point can maintain an extremely high power factor and near-zero reactive power exchange under various complex operating conditions, thus meeting the stringent requirements of modern smart grids for high controllability of the power supply side. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the overall technical solution architecture proposed in this invention; Figure 2 This is a schematic diagram of the core principle framework of reactive power zeroing regulation based on prediction-allocation-feedback closed-loop drive in this invention. Figure 3 This is a flowchart illustrating the logic of multi-dimensional high-frequency sensing and full-frequency reactive component extraction in this invention. Figure 4 This is a schematic diagram of the multi-level interaction relationship and data flow of the collaborative allocation and multi-objective optimization scheduling of heterogeneous reactive power compensation equipment in this invention; Figure 5 This is a logical framework diagram of the dynamic adjustment and closed-loop fine-tuning of the weights of devices based on health status evaluation in this invention. Detailed Implementation

[0027] Example 1 Please refer to the appendix. Figure 1 With appendix Figure 2 This embodiment discloses a feedback regulation system for achieving reactive power zeroing at the grid connection point of a renewable energy power plant. The system is deployed at the physical connection point between the renewable energy power plant and the public power grid, i.e., the grid connection point. Through deep sensing, predictive analysis, and command coordination of the entire power plant's electrical status, it achieves real-time zeroing of reactive power at the grid connection point. Structurally, the system consists of multiple highly integrated hardware modules and software algorithm logic clusters, with its underlying architecture relying on a high-bandwidth, low-latency industrial communication network.

[0028] Please refer to the attached document. Figure 3 The multi-dimensional high-frequency sensing unit, serving as the data source for the entire feedback regulation system, is installed on the bus side of the grid connection point of the renewable energy power plant and at key nodes in each internal generation branch. The multi-dimensional high-frequency sensing unit integrates high-precision current transformers, voltage transformers, and matching high-speed analog-to-digital conversion circuits. To capture extremely short-term non-stationary random fluctuations in the power system, the sampling frequency of this unit is set to above 10,000 Hz, meaning that within each 50 Hz power frequency cycle, the multi-dimensional high-frequency sensing unit can acquire at least 200 discrete data sampling points. These sampling points constitute a high-dimensional raw sequence describing the electrical characteristics of the grid connection point.

[0029] After acquiring discrete voltage and current sequences, the multi-dimensional high-frequency sensing unit does not directly perform conventional root-mean-square (RMS) calculations. Instead, it performs in-depth analysis using a sliding window integration operator configured within its processor. The length of the sliding window is dynamically adjusted according to the system frequency, typically set to one power frequency cycle, while the sliding step size is set to 100 microseconds, thus ensuring that the refresh rate of the sensing data reaches the microsecond level. During data processing, the multi-dimensional high-frequency sensing unit applies an improved phase-shift-free filtering algorithm based on an integral chain. This algorithm, through the cascading of multiple integration stages, filters out high-frequency switching noise while strictly controlling the phase shift during signal processing.

[0030] The core function of the multi-dimensional high-frequency sensing unit is to extract reactive power components across the entire frequency domain. It extracts the fundamental reactive power component by orthogonally decomposing the instantaneous product of voltage and current waveforms. Addressing the harmonic issues generated by the extensive integration of power electronic equipment in renewable energy power plants, this unit utilizes discrete Fourier transform or wavelet transform logic to calculate the phase relationship between each harmonic current and the voltage at the same frequency, thereby extracting the harmonic reactive power components. Furthermore, for the impulsive reactive power generated by the switching of large transformers or line faults within the power plant, the multi-dimensional high-frequency sensing unit calculates the instantaneous power spectral density, capturing abrupt changes in the energy spectrum to achieve quantitative analysis of the impulsive reactive power. To support precise regulation under asymmetrical operating conditions, the unit also performs real-time decoupling operations on positive-sequence, negative-sequence, and zero-sequence components, calculating the reactive power contribution corresponding to each sequence component, providing a physical basis for subsequent asymmetrical compensation.

[0031] Based on the continuous output of high-precision electrical parameters by the multi-dimensional high-frequency sensing unit, the power flow prediction unit executes the logic for predicting future states. The power flow prediction unit integrates a variable step-size differential analysis algorithm, the mathematical expression model of which is as follows: In the above formula, This represents the reactive power demand value at the predicted time. This represents the actual reactive power value measured by the multi-dimensional high-frequency sensing unit at the current moment. and These represent the first-order rate of change weighting coefficient and the second-order acceleration weighting coefficient, respectively. These coefficients are dynamically updated online based on historical adjustment errors.

[0032] The power flow prediction unit identifies the initial trend of reactive power fluctuations by calculating the power change rate between the current moment and the previous three sampling periods. For example, when a wind farm experiences a drastic change in wind speed or a photovoltaic power station is rapidly blocked by clouds, the active power output of the station will change drastically, accompanied by a corresponding change in the reactive power output of the inverter group. By matching the real-time power fluctuation curve with an internally stored library of typical disturbance characteristics, the power flow prediction unit can identify the type of disturbance and its evolution pattern. Once the matching degree exceeds a set threshold of 90%, the unit generates a compensation command with anticipatory attributes before physical response lag occurs. This shift from passive feedback to proactive prediction effectively offsets the inherent time constant required for data transmission and equipment operation, enabling the system to respond to reactive power demands from the grid side in advance.

[0033] Please refer to the attached document. Figure 4 The collaborative allocation optimization unit receives advance compensation instructions from the power flow trend prediction unit and is responsible for scientifically distributing the total reactive power compensation task to various compensation terminals within the power station. New energy power stations typically contain multiple types of reactive power sources, such as grid-connected inverter groups, static var generators (SVA), and switched capacitor banks. These devices have different physical characteristics, response bandwidths, and operating costs. The collaborative allocation optimization unit employs a hierarchical allocation logic based on response speed priority.

[0034] Specifically, for high-frequency, pulsed reactive power demand above 10 Hz, the collaborative allocation optimization unit defines it as a transient compensation task and prioritizes its allocation to grid-connected inverter groups with microsecond-level adjustment capabilities. By adjusting the inverter's pulse width modulation signal, the high bandwidth characteristics of its internal current loop are utilized for rapid reactive power offsetting. For fluctuating reactive power between 2 Hz and 10 Hz, the static var generator (SVM) primarily handles the load. For steady-state or slowly varying reactive power demand below 2 Hz, basic support is provided by controlling stepped or continuously switched capacitor banks. During the allocation process, the collaborative allocation optimization unit establishes a multi-objective optimization scheduling model in real time. This model uses minimizing the reactive power residual at the grid connection point, minimizing the total equipment loss across the entire station, and balancing equipment health as optimization objectives to solve for the optimal allocation ratio coefficient.

[0035] When dealing with the complex operating conditions of large-scale inverter parallel operation, the collaborative load distribution optimization unit introduces a virtual impedance regulation mechanism. This mechanism simulates the physical impedance characteristics in the control algorithm, changing the equivalent output impedance of each inverter at the grid connection point. When excessive reactive current is detected in a branch inverter, the collaborative load distribution optimization unit automatically increases the virtual impedance parameter of that branch, thereby guiding the reactive current to other inverter branches with remaining capacity margin, achieving automatic spatial balancing of reactive load. Simultaneously, the unit also incorporates the impedance parameters of the internal power collection lines of the substation to perform remote voltage drop compensation on the issued control commands. This means that the commands issued by the system are not simply power values, but include corrections for voltage losses due to line transmission, ensuring that the commands, after reaching the execution terminal and being converted into actual output, can accurately act on the grid connection point.

[0036] The closed-loop feedback fine-tuning unit is a key component in ensuring the accuracy of reactive power zeroing. This unit monitors the residual reactive power deviation at the grid connection point in real time after the collaborative allocation optimization unit has executed its functions. Despite the advance compensation from the power flow prediction unit, slight reactive power exchange may still exist at the grid connection point due to model errors and environmental noise. The closed-loop feedback fine-tuning unit uses adaptive proportional-integral-derivative (PID) adjustment logic to finely adjust the output command.

[0037] The closed-loop feedback fine-tuning unit internally constructs a control operator based on dynamic gain adjustment, and its control logic follows the formula below: In this formula, For the first Fine-tuning correction amount per adjustment cycle This represents the real-time reactive power deviation at the grid connection point. The core feature of the system lies in its proportional gain. Integral coefficient and differential coefficients It is not a fixed value, but is adjusted in real time according to the changes in the impedance characteristics of the grid side. When the closed-loop feedback fine-tuning unit senses that the voltage waveform at the grid connection point is distorted or the frequency fluctuation is increased, and predicts that the grid strength is weakening, the unit will automatically reduce the proportional coefficient to increase the phase margin of the control system, suppress possible low-frequency oscillations, and prevent the system from falling into an unstable state.

[0038] The closed-loop feedback fine-tuning unit has a 1% dead zone threshold. When the reactive power offset at the grid connection point is within this threshold range, the system maintains its current output to avoid fatigue of power electronic devices caused by frequent small-amplitude movements. Once the offset exceeds the 1% dead zone range, the unit immediately activates a fast correction mode, forcing the reactive power to quickly return to the zero-value center region by superimposing an instantaneous pulse increment on the output. In addition, the unit also has an error self-learning function, which can perform statistical analysis on the adjustment data over the past 24 hours. If a fixed offset component is found to exist in the system under specific time periods or operating conditions, the closed-loop feedback fine-tuning unit will embed this component as a static compensation benchmark into the initial adjustment, thereby offsetting the steady-state error caused by systematic sensor drift or long-distance signal transmission delay.

[0039] Please refer to the attached document. Figure 5 The health status monitoring unit provides a safety guarantee for the long-term stable operation of the entire system. It uses sensors deployed at key locations in each compensation device to acquire real-time data on the casing temperature of power semiconductor devices, the ripple coefficient of the DC-side bus voltage, and the harmonic distortion rate of the AC-side current. Through multi-dimensional correlation analysis, the health status monitoring unit calculates a real-time health score for each online device.

[0040] If the health score of a device drops below 80 points due to excessive temperature rise from continuous high-load operation, the health status monitoring unit will immediately send a weight adjustment signal to the collaborative allocation optimization unit, automatically lowering the device's participation priority in the allocation strategy. If the health score further drops below 60 points, the unit will trigger forced isolation logic, removing the damaged device from the adjustment sequence and initiating a redundancy replacement scheme to smoothly transfer its previously borne reactive power share to other devices in good health within the station. This mechanism ensures the continuity of the adjustment process and avoids reactive power outages at the grid connection point caused by the failure of a single device.

[0041] The system in this embodiment operates on a communication architecture based on microsecond-level deterministic industrial Ethernet. This network employs a time-sensitive networking protocol to ensure that the transmission latency of all electrical signal data uploaded by sensors and the control commands issued is strictly locked within 2 milliseconds. Internally, the system uses a high-precision clock source provided by the Global Positioning System (GPS) or the BeiDou Navigation Satellite System for time synchronization, achieving a time synchronization accuracy of 1 microsecond. This extreme time synchronization capability ensures that data acquired by multiple sensing points distributed across different areas of the site are perfectly aligned on the time axis, providing a unified time reference for multi-point coordinated adjustment.

[0042] In addition, this system includes an offline simulation verification module. This module runs on a background computing server, synchronizing all operational data, equipment parameters, and grid status in real time, constructing a digital twin model in virtual digital space that is completely equivalent to the physical site. The offline simulation verification module continuously trains and corrects the model using actual operational data. In the event of extreme weather or early warnings of large grid disturbances, this module can pre-simulate various combinations of regulation strategies in the virtual environment, find the optimal set of control parameters under the current conditions, and periodically push these verified parameters to the online regulation system to achieve closed-loop iterative optimization of regulation performance.

[0043] Under special operating conditions of abnormal fluctuations in the power grid, this system exhibits strong adaptive capabilities. When the voltage at the grid connection point falls below or rises above the safe operating range, triggering the low-voltage ride-through or high-voltage ride-through mechanism, the system automatically and rapidly switches from reactive power zeroing mode to emergency voltage support mode. At this time, the closed-loop feedback fine-tuning unit temporarily suppresses the zeroing logic and instead injects large-current inductive or capacitive reactive power into the grid according to the voltage-reactive power support curve preset in the power grid dispatching specifications, assisting in the restoration of grid voltage stability. After the grid voltage returns to the normal range of 0.9 to 1.1 times the rated voltage, the system gradually reduces the support amount and reactivates the zeroing regulation link through a smooth transition algorithm, ensuring that no secondary disturbances occur during the switching process.

[0044] The multi-dimensional high-frequency sensing unit also utilizes an improved instantaneous power algorithm when extracting reactive power components. By performing coordinate transformation in a three-phase stationary coordinate system, the three-phase voltage and current signals are converted into alpha and beta axis components, thereby eliminating the influence of frequency fluctuations on reactive power calculation. This algorithm can maintain a reactive power phase capture error of less than 0.1 degrees over a wide frequency range of 45 Hz to 55 Hz. It is this extremely high sensing accuracy at the physical level that forms the solid foundation for the entire system to achieve reactive power zeroing.

[0045] When performing reactive power distribution, the collaborative allocation optimization unit also monitors the on-load tap changer status of the main transformers in the substation in real time. When it detects that the reactive power regulation capacity is close to saturation due to voltage level limitations, the system automatically sends an adjustment request to the transformer tap changer controller. By collaboratively adjusting the transformer turns ratio, it optimizes the voltage distribution throughout the substation and releases the suppressed reactive power regulation capacity. This cross-level collaborative control method increases the availability of the substation's reactive power regulation capacity to over 95%, greatly improving the system's capacity margin under extreme operating conditions.

[0046] The system also features environmental adaptive adjustment logic. By accessing meteorological sensor data within the site, such as ambient temperature, humidity, and irradiance, the health status monitoring unit dynamically adjusts the physical limit thresholds of the power equipment. In low-temperature conditions such as winter, due to improved heat dissipation, the system automatically relaxes the inverter's overload limits, allowing for the extraction of reactive current exceeding 1.2 times the rated value for a short period to cope with sudden power system faults. Conversely, in high-temperature environments, the system automatically tightens the limits to protect critical power electronic components from overheating damage.

[0047] The integral term of the closed-loop feedback fine-tuning unit employs advanced anti-saturation processing technology. When the reactive power demand at the grid connection point exceeds the physical output limit of all station equipment, the integral term automatically stops accumulating values ​​and locks its internal state at the maximum output value. This approach avoids severe overshoot caused by the slow regression of the integral term after the reactive power demand returns to the normal range, ensuring that the reactive power exchange at the grid connection point always smoothly converges to near zero.

[0048] Example 2 Based on the feedback regulation system for achieving zero reactive power at the grid connection point of new energy power plants described in Embodiment 1 above, this embodiment focuses on describing the specific operating mode and enhanced regulation logic of the system under weak grid conditions. When a new energy power plant is connected to a distant, high-impedance grid terminal, i.e., under the so-called weak grid conditions, the system faces serious challenges in stable operation.

[0049] In weak grid environments, the voltage at the grid connection point is extremely sensitive to reactive power changes; even minor reactive power adjustments can trigger significant voltage fluctuations, leading to system oscillations. To address this technical challenge, the multi-dimensional high-frequency sensing unit in this embodiment adds a real-time assessment function for the short-circuit ratio at the grid connection point. This unit estimates the equivalent internal impedance of the grid online by observing the correlation slope between voltage and current fluctuations at the grid connection point and utilizing the Thevenin equivalent principle.

[0050] When the estimated short-circuit ratio is below 3.0, the power flow prediction unit automatically switches to conservative prediction mode. In this mode, the step size of the prediction command is reduced, and a constraint factor based on voltage sensitivity is added. The predicted advance compensation no longer merely aims for instantaneous zeroing of reactive power, but prioritizes ensuring that the voltage fluctuation rate at the grid connection point is within a safe threshold.

[0051] In weak grid conditions, the collaborative allocation optimization unit significantly enhances the role of the virtual impedance regulation mechanism. By introducing a virtual reactance that matches the equivalent impedance of the grid, the power transmission impedance characteristics between the renewable energy plant and the grid are artificially altered, thereby increasing the system's damping ratio. In this case, when allocating reactive power, the collaborative allocation optimization unit adds a phase compensation component to offset the voltage phase drift caused by impedance in weak grid conditions.

[0052] In weak grid environments, the closed-loop feedback fine-tuning unit employs multi-model adaptive control logic. This unit pre-stores control parameter packages for different grid strengths. When the health monitoring unit indicates a decrease in grid strength, the closed-loop feedback fine-tuning unit switches in real-time to a parameter set with a higher damping coefficient. Simultaneously, the unit's error self-learning function shortens the learning cycle from 24 hours to 1 hour, enabling more agile detection of rapid changes in grid impedance.

[0053] In this enhanced mode, the system's communication frequency is further increased. Control messages in the industrial Ethernet are prioritized to the highest level, ensuring that any minute adjustment commands under weak power grid conditions reach the actuator in the shortest possible time. Through this targeted logic enhancement, even in extremely weak power grid environments with very low short-circuit ratios, this system can still maintain high-precision zeroing of reactive power at the grid connection point without inducing any resonance or instability.

[0054] Example 3 This embodiment discloses the fault-tolerant regulation mode of the feedback regulation system for reactive power zeroing at the grid connection point of a new energy power plant when dealing with large-scale equipment failures within the power plant. When a large area of ​​power generation equipment or compensation equipment in a new energy power plant is disconnected from the grid due to lightning strikes, wildfires, or severe line short circuits, the system will automatically activate the reconfiguration strategy described in this embodiment.

[0055] Under these extreme conditions, the health status monitoring unit will immediately capture offline signals from a large area of ​​equipment via industrial Ethernet and instantly calculate the total remaining available reactive power capacity. If the remaining capacity is insufficient to cover the total reactive power demand of the grid connection point, the collaborative allocation optimization unit will initiate a priority-based stripping logic. This logic prioritizes reactive power support for high-power branches based on the active power contribution of each branch to the grid connection point, while restricting low-power or less efficient branches.

[0056] During this process, the multi-dimensional high-frequency sensing unit enhances the monitoring of transient energy flow to prevent damage to remaining online equipment caused by instantaneous overvoltage due to large-area equipment tripping. The closed-loop feedback fine-tuning unit immediately enters the nonlinear large-signal adjustment mode, skipping the dead zone limit and directly performing maximum slope tracking compensation based on the magnitude of the residual reactive power deviation, striving to restore the reactive power balance of the grid connection point in the shortest possible time.

[0057] Meanwhile, the offline simulation verification module initiates an emergency evolution simulation for the current faulty topology within milliseconds. It evaluates the optimal load allocation scheme using a digital twin model and injects the calculation results into the collaborative allocation optimization unit via a fast channel. In this fault-tolerant mode, the system can minimize the backfeeding or excessive absorption of reactive power from the grid connection point to the grid by maximizing the utilization of remaining limited resources, until the substation returns to normal operating topology.

[0058] This embodiment also relates to system operation protection under extreme high humidity conditions. When the health status monitoring unit detects that the ambient humidity exceeds 95% through the humidity sensor, it automatically triggers the anti-condensation operation logic of the internal power devices. The collaborative allocation optimization unit periodically allocates a small amount of circulating reactive current to the inverter in standby mode, using the heat generated by the device operation to maintain a dry environment inside the cabinet. This cross-dimensional environmental adaptability enables this reactive power zeroing feedback regulation system to not only achieve the ultimate in electrical performance, but also possess extremely high engineering robustness in terms of physical environmental adaptability.

[0059] In summary, this invention constructs a complete closed-loop system with predictive capabilities, efficient collaborative capabilities, self-healing capabilities, and high-precision control capabilities through the deep collaboration of multi-dimensional high-frequency sensing units, power flow prediction units, collaborative allocation optimization units, closed-loop feedback fine-tuning units, and health status monitoring units. Data flow between units is achieved through a microsecond-level synchronous industrial network, ensuring real-time zero reactive power at the grid connection point while significantly improving the adaptability of renewable energy power plants to complex grid environments. The complex algorithmic logic and precise hardware sensing architecture within the system work together to effectively solve the pain points of existing technologies in terms of adjustment lag, limited accuracy, and multi-device conflicts, providing crucial technical support for the construction of smart grids. The design concept of this system fully considers various extreme and abnormal operating conditions in actual engineering, possessing high practical value and broad application prospects. Through in-depth expansion and logical refinement of each functional unit, this implementation clearly and completely demonstrates the technical path to achieve the invention's objectives. All technical descriptions are based on objective physical quantities, clear logical flows, and rigorous control strategies, guiding engineers in the construction and deployment of this system in actual production. The parameter values ​​and ranges involved in this invention are all optimal references derived from numerous engineering experiments, aiming to ensure that the system maintains optimal performance balance under various operating boundaries. Through this comprehensive, multi-level feedback regulation system, the reactive power management level of new energy power plants will achieve a qualitative leap, making a significant contribution to the safe, stable, and efficient operation of the power grid.

[0060] The embodiments described above are merely exemplary descriptions of the system architecture and working principle of this invention. In practical engineering applications, adjustments can be made to the internal parameter settings, communication protocol details, and specific optimization algorithms of each unit, without departing from the core concept of this invention, based on the specific scale of the new energy power station, equipment type, and the special requirements of the power grid. Such subtle changes based on actual engineering constraints should all be covered within the scope of protection of this invention. The entire feedback regulation system, as an organic whole, achieves the high-precision dynamic benchmark correction target of reactive power at the new energy grid connection point through strict spatiotemporal alignment and logical coupling between modules. Every detail of the system design, from high-frequency sampling to trend prediction, from collaborative allocation to feedback fine-tuning, and then to health monitoring and fault-tolerant operation, is designed to meet the high controllability requirements of modern power systems for clean energy access. This fully digital and highly automated regulation scheme not only improves the operating efficiency of individual power stations but also lays a solid technical foundation for the collaborative control of future large-scale new energy base clusters. In future technological evolution, this system can further integrate deep reinforcement learning algorithms in the field of artificial intelligence, and further improve the recognition accuracy and response speed of complex power flow fluctuation patterns through autonomous mining of massive amounts of operating data. Regardless of how technology evolves, the "prediction-allocation-feedback" three-in-one core architecture described in this invention remains the physical foundation for ensuring the reactive power zeroing control performance of the grid connection point.

[0061] The implementation of this invention will significantly reduce the grid assessment risks faced by new energy power plants due to excessive reactive power deviation. Simultaneously, by optimizing the voltage distribution across the entire station, it reduces line losses, indirectly improving the power generation economic efficiency of the plant. The system's temperature rise management and fatigue control of hardware equipment effectively extend the service life of expensive power electronic compensation equipment. This design approach, which balances technological leadership with engineering economy, is a significant advantage of this invention compared to traditional solutions. In actual deployment, the system's high integration also greatly simplifies on-site commissioning and maintenance, shortening the construction cycle. Through comprehensive digital sensing and intelligent decision-making, this invention truly realizes the transformation of new energy power plants from "passive grid connection" to "active support," providing key technical guarantees and practical examples for building a new power system with new energy as the main body. The application of each sub-technology, such as 10000 Hz high-frequency sampling, microsecond-level time synchronization, and load balancing based on virtual impedance, aims to find the most precise balance point in this high-dimensional, nonlinear dynamic adjustment process, ensuring that the reactive power exchange at the grid connection point continuously and stably converges to zero. This represents the ultimate pursuit in the field of power system control, and it is also the core value embodiment of the technical solution of this invention.

Claims

1. A feedback regulation system for achieving zero reactive power at the grid connection point of a new energy power plant, characterized in that, include: It includes a multi-dimensional high-frequency sensing unit, a trend prediction unit, a collaborative allocation optimization unit, a closed-loop feedback fine-tuning unit, and a health status monitoring unit; The multi-dimensional high-frequency sensing unit is used to synchronously collect voltage and current signals at the grid connection point of the new energy power station and each branch inside the station, extract the full-frequency reactive component of the grid connection point, and estimate the equivalent internal impedance of the power grid and the short-circuit ratio of the grid connection point online by observing the correlation slope between the voltage fluctuation and current fluctuation at the grid connection point. The power flow prediction unit is used to generate advance compensation instructions based on the full-frequency reactive component and the power output fluctuation information of the power station, and to switch to conservative prediction mode when the short-circuit ratio at the grid connection point is lower than the preset short-circuit ratio threshold. In the conservative prediction mode, the prediction step size of the advance compensation instructions is reduced, and a voltage sensitivity constraint factor is introduced so that the generated advance compensation instructions prioritize limiting the voltage fluctuation rate at the grid connection point. The collaborative allocation optimization unit is used to receive the advance compensation command, allocate the reactive power compensation task to the grid-connected inverter group, static var generator or switched capacitor bank according to the frequency characteristics of reactive power demand, and introduce a virtual reactance that matches the equivalent internal impedance of the grid and a phase compensation component to offset voltage phase drift when the short-circuit ratio at the grid connection point is lower than the preset short-circuit ratio threshold. At the same time, it combines the line impedance parameters of each branch in the station to perform remote voltage drop compensation for the control commands sent to each compensation terminal. The closed-loop feedback fine-tuning unit is used to monitor the residual reactive power deviation of the grid connection point after each compensation terminal executes the control command in real time, and to switch the control parameter package corresponding to different grid strengths according to the equivalent internal impedance of the grid or the short-circuit ratio of the grid connection point, so as to dynamically correct the control command output by the collaborative allocation optimization unit, so that the reactive power exchange of the grid connection point converges to near zero. The health status monitoring unit is used to collect the operating status parameters of each compensation device, generate a health score for each compensation device, and adjust the participation priority, reactive power output limit, or redundant takeover status of each compensation device in the collaborative allocation optimization unit according to the health score.

2. The feedback regulation system for achieving zero reactive power at the grid connection point of a new energy power station according to claim 1, characterized in that, The multi-dimensional high-frequency sensing unit includes a synchronous sampling subunit, a full-frequency domain reactive power decoupling subunit, and a sequence component independent decoupling subunit. The synchronous sampling subunit is used to collect voltage and current signals from grid connection points and key nodes inside the station at a sampling frequency of not less than 10,000 Hz. The full-frequency domain reactive decoupling subunit is used to separate the fundamental reactive component, harmonic reactive component and impulsive reactive component from the acquired signal using sliding window integration processing. The sequence component independent decoupling subunit is used to perform independent decoupling operations on the positive sequence component, negative sequence component, and zero sequence component of the grid connection point voltage signal and current signal.

3. The feedback regulation system for achieving zero reactive power at the grid connection point of a new energy power station according to claim 1, characterized in that, The power trend prediction unit includes a power change rate identification subunit, a disturbance pattern matching subunit, and a lead command generation subunit. The power change rate identification subunit is used to identify the starting point and acceleration characteristics of reactive power fluctuations based on the power data of the current sampling period and at least three previous sampling periods. The disturbance pattern matching subunit is used to match the real-time power fluctuation curve with a pre-stored typical disturbance feature library to identify the type of systemic reactive power disturbance caused by changes in wind speed or light intensity. The advance instruction generation subunit is used to generate a reactive power compensation reference signal with advance attributes before the physical response lag occurs when the matching degree exceeds a preset matching threshold.

4. The feedback regulation system for achieving zero reactive power at the grid connection point of a new energy power station according to claim 1, characterized in that, The collaborative allocation optimization unit includes a frequency domain response hierarchical subunit and a multi-objective optimization scheduling subunit; The frequency domain response hierarchical subunit is used to allocate high-frequency pulse reactive power demand with a frequency higher than 10 Hz to the grid-connected inverter group, allocate fluctuating reactive power demand with a frequency between 2 Hz and 10 Hz to the static var generator, and allocate steady-state or slowly changing reactive power demand with a frequency lower than 2 Hz to the switching capacitor bank. The multi-objective optimization scheduling subunit is used to solve the reactive power allocation ratio of each compensation terminal with the optimization objectives of minimizing the reactive power residual at the grid connection point, minimizing the total equipment loss of the entire station, and balancing the equipment health.

5. The feedback regulation system for achieving zero reactive power at the grid connection point of a new energy power station according to claim 4, characterized in that, The collaborative allocation optimization unit also includes a virtual impedance balancing subunit and a line voltage drop compensation subunit; The virtual impedance balancing subunit is used to adjust the virtual impedance parameter of the corresponding branch according to the magnitude of the reactive current borne by each branch of the grid-connected inverter when multiple grid-connected inverters are running in parallel. When the reactive current of a branch exceeds the preset branch limit, the virtual impedance parameter of the branch is increased to guide the reactive current to the inverter branch that still has capacity margin. The line voltage drop compensation subunit is used to add line transmission voltage drop correction to the control commands sent to each compensation terminal based on the line impedance parameters of the internal collection lines of the substation, so that the actual output of each compensation terminal corresponds to the compensation requirements of the grid connection point.

6. The feedback regulation system for achieving zero reactive power at the grid connection point of a new energy power station according to claim 1, characterized in that, The preset short-circuit ratio threshold is 3.0; when the short-circuit ratio at the grid connection point is lower than 3.0, the power flow trend prediction unit reduces the prediction step size of the advance compensation command, the collaborative allocation optimization unit enables a virtual reactance that matches the equivalent internal impedance of the power grid and superimposes a phase compensation component, and the closed-loop feedback fine-tuning unit switches to a control parameter package with a higher damping coefficient.

7. The feedback regulation system for achieving zero reactive power at the grid connection point of a new energy power station according to claim 1, characterized in that, The closed-loop feedback fine-tuning unit includes a dynamic gain adjustment subunit, a fast correction activation subunit, an error self-learning subunit, and an integral anti-saturation processing subunit. The dynamic gain adjustment subunit is used to adjust the proportional coefficient, integral coefficient and derivative coefficient according to the change of the equivalent internal impedance of the power grid, and to reduce the proportional coefficient to enhance the system damping ratio when the power grid strength weakens. The fast correction activation subunit is used to superimpose an instantaneous pulse increment in the control command when the reactive power offset at the grid connection point exceeds the preset dead zone threshold, so that the reactive power at the grid connection point falls back to the zero value center region. The error self-learning subunit is used to statistically analyze historical adjustment residuals, identify periodic fixed offsets, and embed the periodic fixed offsets as static compensation benchmarks into the initial adjustment amount. The integral anti-saturation processing subunit is used to stop the accumulation of integral terms when the control command reaches the physical output limit of the compensation device.

8. The feedback regulation system for achieving zero reactive power at the grid connection point of a new energy power station according to claim 1, characterized in that, The health status monitoring unit includes a multi-dimensional status perception subunit, a health score calculation subunit, and a dynamic weight adjustment subunit. The multi-dimensional state sensing subunit is used to collect the power device housing temperature, bus voltage ripple, and current harmonic distortion rate of each compensation device. The health score calculation subunit is used to generate a health score for each compensation device based on the power device housing temperature, bus voltage ripple, and current harmonic distortion rate. The dynamic weight adjustment subunit is used to reduce the upper limit of reactive power output of the compensation device when the health score of the compensation device is lower than the first health threshold, and to cut the compensation device out of the adjustment sequence when the health score of the compensation device is lower than the second health threshold, so that the reactive power share it undertakes is redundantly taken over by other compensation devices.

9. The feedback regulation system for achieving zero reactive power at the grid connection point of a new energy power station according to claim 1, characterized in that, The system operates on a microsecond-level deterministic industrial Ethernet communication architecture, which ensures that the transmission delay of sensor data and control commands does not exceed 2 milliseconds, and enables each sensing point and execution point within the system to achieve a time synchronization accuracy of no more than 1 microsecond through a high-precision time synchronization protocol.

10. The feedback regulation system for achieving zero reactive power at the grid connection point of a new energy power station according to claim 1, characterized in that, It also includes an offline simulation verification module and a fault ride-through support module; The offline simulation verification module is used to construct a digital twin model of the new energy power station based on the field operation data, equipment parameters and power grid status, and to pre-simulate the adjustment strategy under extreme operating conditions in the digital twin model, and update the control parameter set verified by the pre-simulation to the collaborative allocation optimization unit or the closed-loop feedback fine-tuning unit. The fault ride-through support module is used to switch the system from reactive power zeroing mode to voltage emergency support mode when the grid connection point voltage is lower or higher than the safe operating range, and inject inductive or capacitive reactive power according to the preset voltage-reactive power support curve. After the grid connection point voltage recovers to the normal range, it smoothly switches back to reactive power zeroing mode.

Images