New energy base field group cooperative control and protection integration system

Abstract

The application discloses a new energy base field group cooperative control and protection integrated system, and relates to the technical field of power supply systems.The system comprises a panoramic state sensing unit, a cooperative sensing reasoning engine, an integrated decision hub, a dynamic coordination execution unit, a hierarchical response device, a cooperative power regulation device and a feedback closed loop unit.The application constructs a unified state snapshot of the whole field group and evaluates the safety level and disturbance trend;the control target and protection constraint are included in a unified optimization framework to jointly solve and generate a strategy;the execution timing of the control and protection is dynamically adjusted according to the real-time state, and the rate or strategy recalculation is automatically adjusted or triggered when a risk or topology mutation is caused.The application breaks the information barrier and timing conflict of the traditional separate structure, effectively avoids the protection misoperation and chain reaction caused by the control regulation, and significantly improves the transient stability and power supply reliability of the new energy base.

Description

Technical Field

[0001] This invention relates to the field of power supply system technology, specifically to an integrated system for coordinated control and protection of new energy base clusters. Background Technology

[0002] With the deepening of the global energy transition, new energy power generation technologies, represented by wind power and photovoltaics, are developing rapidly. The construction and grid connection of large-scale new energy bases have become an important trend in power system development. New energy bases typically consist of multiple power plant clusters working together, transmitting clean energy to remote load centers via ultra-high-voltage direct current transmission or AC grid interconnection. Against this backdrop, how to achieve the safe, stable, and efficient operation of new energy base clusters has become a core research topic urgently needing study in the power system field. Among these, the collaborative control and protection technology of new energy base clusters is a key link in ensuring the safe operation of the system. Collaborative control aims to achieve power balance and stable operation among multiple power plants through optimized scheduling strategies, while protection devices are responsible for quickly isolating faulty areas when a fault occurs to prevent the accident from escalating. Both work together to maintain the transient stability and power supply reliability of the system. Currently, existing new energy base clusters generally adopt a separate control and protection architecture, with the control master station focusing on steady-state optimization and regulation, and protection devices operating independently only based on local electrical quantities. The lack of information sharing and timing coordination between the two means that power fluctuations caused by control and regulation during grid disturbances may trigger maloperation of protection systems, leading to disorderly disconnection of generating units from the grid. Furthermore, topological changes after protection tripping can easily cause control strategy mismatch, triggering a chain reaction and seriously threatening the transient stability and power supply reliability of high-proportion renewable energy sending-end systems. Summary of the Invention

[0003] The purpose of this invention is to provide an integrated system for collaborative control and protection of new energy base clusters, so as to solve the problems of information non-sharing, timing mismatch, and easy triggering of protection malfunctions and chain reactions caused by the separate control and protection architecture in the prior art.

[0004] To solve the above problems, the technical solution adopted by the present invention is as follows:

[0005] A new energy base cluster collaborative control and protection integrated system includes:

[0006] The panoramic status perception unit is used to collect electrical quantity data, non-electrical quantity data and network topology data of each station in the new energy base cluster in real time, and build a unified status snapshot of the entire cluster.

[0007] The collaborative perception inference engine is used to collaboratively analyze the state snapshots collected by the panoramic state perception unit, identify the security level of the current operating state, and detect potential disturbance trends.

[0008] An integrated decision-making center is used to generate unified control and protection strategies based on the security level and disturbance trend output by the collaborative perception and inference engine. The integrated decision-making center has a built-in collaborative decision-making algorithm module, which incorporates the control objective function and protection action constraints into a unified optimization framework for joint solution.

[0009] The dynamic coordination execution unit is used to receive the unified control strategy and protection strategy output by the integrated decision center, and dynamically adjust the execution sequence of control commands and protection actions according to the real-time operating status; when the risk of protection maloperation caused by control adjustment is detected, the dynamic coordination execution unit automatically adjusts the control adjustment rate to avoid the dead zone of protection action; when the network topology change is caused by protection tripping is detected, the dynamic coordination execution unit immediately triggers the integrated decision center to perform adaptive recalculation of control strategy to suppress chain reaction;

[0010] A graded response device is used to execute the protection strategy output by the dynamic coordination execution unit;

[0011] A coordinated power regulation device is used to execute the control strategy output by the dynamic coordination execution unit;

[0012] The feedback closed-loop unit is used to monitor the execution effects of the hierarchical response device and the collaborative power adjustment device in real time, and to feed back the execution results to the collaborative perception inference engine to update the state snapshot and adjust the subsequent decision-making strategy.

[0013] Furthermore, the electrical quantity data includes the AC side current, voltage, active power, and reactive power of each inverter; the non-electrical quantity data includes the photovoltaic module temperature, wind turbine blade angle, and ambient wind speed; and the network topology data includes bus connection relationships, switch on / off status, and line parameters.

[0014] Furthermore, the safety level is divided into four levels: normal, early warning, emergency, and severe; the disturbance trend includes power fluctuation trend, voltage change trend, and frequency offset trend; the collaborative perception inference engine performs the collaborative analysis based on preset state transition rules and field group coupling relationship model.

[0015] Furthermore, the collaborative perception inference engine also includes a disturbance source localization submodule; the disturbance source localization submodule achieves accurate localization of the disturbance source based on the temporal correlation analysis of the electrical quantity data; the temporal correlation analysis uses a mutual information algorithm to calculate the correlation between the electrical quantity data of each station and the disturbance occurrence time; the mutual information algorithm obtains the mutual information value by calculating the sum of the logarithms of the products of the joint probability distribution of the electrical quantity data sequence of the station and the disturbance occurrence time sequence and their respective marginal probability distributions, the larger the mutual information value, the higher the correlation between the electrical quantity data of the station and the disturbance; the localization result of the disturbance source localization submodule is simultaneously input into the integrated decision center to generate targeted control strategies and protection strategies.

[0016] This technical solution addresses the complex operating conditions during disturbances by employing time-series correlation analysis and mutual information algorithms. Relying solely on the probability distribution correlation between electrical quantity data sequences and the disturbance occurrence time, it can accurately identify and locate the disturbance source. This method offers fast computation speed and strong anti-interference capabilities, saving valuable time for subsequent development of targeted control and protection strategies.

[0017] Furthermore, the control objective function includes a power balance objective, a voltage stability objective, and a frequency regulation objective; the protection action constraints include protection selectivity constraints and action time constraints; the power balance objective requires minimizing the deviation between the total power output of the power plant group and the scheduling plan value; the voltage stability objective requires that the voltage of each bus be maintained within the allowable range; the frequency regulation objective requires that the system frequency be maintained near the rated value; the protection selectivity constraint requires that the protection action be executed according to the set selective coordination relationship; and the action time constraint requires that the protection action be completed within a specified time window.

[0018] Furthermore, the collaborative decision-making algorithm module adopts a model predictive control framework; within each control cycle, the model predictive control framework predicts the system behavior for multiple future time steps based on the current state snapshot, and selects the control sequence that minimizes the comprehensive objective function as the optimal strategy; the comprehensive objective function is a weighted fusion of indicators such as power regulation error, voltage deviation, frequency deviation, and the number of protection actions.

[0019] Furthermore, the dynamic coordination execution unit also includes a timing matching submodule, which is used to calculate the time window matching degree between the control adjustment action and the protection action. During the calculation, the timing matching submodule determines whether there is a time overlap between the expected effective time period of the control adjustment and the preset action time window of the protection device. If there is no time overlap and the control adjustment is expected to be completed before the protection action, the time window matching degree is calculated using the difference between the expected effective end time of the control adjustment and the start time of the preset action time window. When the time window matching degree is lower than a preset threshold, the dynamic coordination execution unit automatically extends the transition time of the control adjustment or prioritizes the execution of the protection action.

[0020] This technical solution quantifies the risk of action interference by calculating the time window matching degree between control and protection actions. When the matching degree is too low, the transition time of control is automatically extended or the protection action is executed first. Through extremely precise timing staggered management, the dead zone of the protection device is perfectly avoided, eliminating the risk of misjudgment caused by control.

[0021] Furthermore, the graded response device is divided into three levels: station-level protection, regional-level protection, and system-level protection. The station-level protection is responsible for quickly isolating internal faults within the station, the regional-level protection is responsible for isolating fault propagation between stations, and the system-level protection is responsible for disconnecting or reconnecting the new energy base to the main grid during a fault. The action time limit of the graded response device adopts an adaptive adjustment mechanism, dynamically correcting the action time limit based on the safety level output by the collaborative perception inference engine. When the safety level is emergency, the action time limit is shortened to 50% of the normal time limit; when the safety level is severe, the action time limit is shortened to 25% of the normal time limit.

[0022] The coordinated power regulation device adopts a hierarchical coordination architecture, including an active power regulation module and a reactive power regulation module. The upper-level coordinator is responsible for allocating the overall active power regulation target and reactive power regulation target across the entire power plant group, while the lower-level actuator is responsible for distributing the regulation target allocated by the upper-level coordinator to specific inverter units within each power plant. The active power regulation module adjusts the active power output allocation of each power plant according to the instructions of the integrated decision-making center, and the reactive power regulation module dynamically allocates the reactive power compensation capacity of each power plant according to the voltage stability target.

[0023] The feedback closed-loop unit compares the monitoring data with the expected target and calculates the execution deviation using the root mean square error method. When the execution deviation exceeds a preset threshold, the feedback closed-loop unit generates a deviation alarm and sends it to the collaborative perception inference engine, triggering the collaborative perception inference engine to reassess the operating status and security level. The feedback closed-loop unit also includes a model self-learning function, which uses historical execution data and employs recursive least squares or Kalman filtering algorithms to perform online optimization of the collaborative decision-making algorithm module and the disturbance trend prediction algorithm.

[0024] In this technical solution, the action time limits of system-level and regional-level protection are no longer fixed and rigid parameters, but are adaptively shortened significantly according to the current safety level, achieving a balance between "fast action" under extreme conditions and "error prevention" under normal conditions. At the same time, the feedback closed-loop unit uses historical execution data to perform online iteration and self-learning optimization of the collaborative decision-making algorithm module, giving the system an intelligent collaborative operation mechanism that continuously evolves with changes in operating conditions.

[0025] Compared with the prior art, the beneficial effects of the present invention are:

[0026] (1) This invention constructs a global architecture integrating perception, reasoning, decision-making, execution, and feedback; through a panoramic state perception unit and a collaborative perception reasoning engine, it achieves unified aggregation and state assessment of information across the entire field, generates a unified strategy through an integrated decision-making center, and performs time-series control by a dynamic coordination execution unit. This mechanism effectively solves the problems of information sharing and timing mismatch under the traditional separate control and protection architecture. It actively avoids protection malfunctions when control regulation may cause them, and immediately recalculates the strategy when protection tripping causes topology changes, thereby greatly improving the transient stability and power supply reliability of high-proportion renewable energy sending-end systems.

[0027] (2) The integrated decision center incorporates the control objective function (power, voltage, frequency) and protection action constraints into a unified optimization framework for joint solution. In particular, it combines the model predictive control framework, enabling the system to predict the system behavior of future time steps in the control cycle and find the optimal strategy, thus completely avoiding the technical defects of mutual restraint and conflict between control regulation and protection action in traditional discrete systems.

[0028] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, embodiments of the present invention are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0029] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 This is a block diagram of the integrated system for collaborative control and protection of new energy base clusters as described in this invention;

[0031] Figure 2 This is a core principle block diagram of the integrated decision-making center in this invention;

[0032] Figure 3 This is a data processing flowchart of the panoramic state perception unit and the collaborative perception inference engine in this invention. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0034] This invention provides specific implementation methods for an integrated system of coordinated control and protection for new energy base clusters. Please refer to the appendix for details. Figure 1 Appendix Figure 2 and attached Figure 3 This needs to be understood. It should be noted that the various unit modules described in this embodiment of the invention, such as the panoramic state perception unit, the collaborative perception inference engine, and the integrated decision-making center, may physically include a processor and a memory. The memory is used to store computer program instructions, and the processor is used to read and execute the instructions to implement the specific logical functions of the aforementioned units. The processor can be an industrial control chip such as a central processing unit (CPU), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), or a field-programmable gate array (FPGA).

[0035] I. System Overall Architecture Description

[0036] The integrated control and protection system for new energy base clusters consists of seven core functional units, which interact and transmit commands via a high-speed communication network. The system's data flow follows a complete chain: "panoramic state perception → collaborative perception and reasoning → integrated decision-making → dynamic coordinated execution → hierarchical response and power regulation → feedback closed loop." These functional units can be physically deployed at the same control center or distributed across station-level and regional control nodes. A schematic diagram of the overall system architecture is attached. Figure 1As shown.

[0037] The core design concept of the system is to transform the traditional separate architecture of control and protection functions into an integrated architecture. Through a unified perception, reasoning, decision-making, and execution mechanism, deep synergy between control and protection is achieved. The panoramic state perception unit is responsible for constructing a unified state snapshot of the entire field group. The collaborative perception and reasoning engine performs collaborative analysis on the state data and identifies the safety level and disturbance trend. The integrated decision center generates unified control and protection strategies based on the analysis results. The dynamic coordination and execution unit is responsible for the timing coordination of control commands and protection actions. The hierarchical response device and the collaborative power adjustment device execute protection and control actions respectively. The feedback closed-loop unit feeds back the execution results to the reasoning engine to achieve continuous learning and adaptive adjustment.

[0038] II. Specific Implementation of the Panoramic Status Perception Unit

[0039] The panoramic status perception unit is the data foundation of the entire integrated system, responsible for real-time collection of electrical and non-electrical data, as well as network topology data, from each station within the new energy base cluster. This unit is typically deployed in the monitoring room of each station, completing data acquisition tasks through sensors and measuring devices deployed at the electrical equipment sites.

[0040] The electrical quantity data acquisition scope includes AC side current, voltage, active power, and reactive power of each inverter. Current and voltage are acquired using electronic current transformers and voltage transformers, achieving a measurement accuracy of 0.2%, which meets the accuracy requirements for relay protection and control decision-making. Active and reactive power calculations are performed internally within the data acquisition unit using a calculation method based on instantaneous power theory, reflecting the inverter's power output status in real time. The electrical quantity data acquisition frequency is set to 128 points per cycle, i.e., a sampling interval of 156.25 μs under 50Hz power frequency conditions. This sampling frequency meets the requirements for synchronous phasor measurement, providing sufficient data resolution for subsequent dynamic performance analysis.

[0041] The non-electrical data collection scope includes photovoltaic module temperature, wind turbine blade angle, and ambient wind speed. Photovoltaic module temperature is collected using a platinum resistance temperature sensor deployed on the backsheet of the photovoltaic module, with a measurement range of -40℃ to 125℃ and a measurement accuracy of ±0.5℃. Wind turbine blade angle is collected using a blade angle sensor, directly reading the angle feedback signal from the pitch system, with a measurement range of 0 to 90° and a resolution of 0.1°. Ambient wind speed is collected using an ultrasonic anemometer deployed within the site's meteorological station, with a measurement range of 0 to 60 meters per second and a measurement accuracy of ±0.1 meters per second. The collection frequency of non-electrical data is relatively low, typically set to 1 to 10 times per second, with the specific frequency determined based on the site type and operational requirements.

[0042] The network topology data collection scope includes bus connection relationships, switch open / closed status, and line parameters. Bus connection relationships are collected by reading the substation automation system's topology configuration file, which is updated in real time to ensure the system can grasp the latest electrical wiring methods. Switch open / closed status is collected by reading the merging unit and process layer GOOSE messages of the smart substation, with a timestamp accuracy of 1ms for status changes. Line parameters include line impedance, admittance, and capacitance to ground. These parameters are pre-stored in the system database and updated based on measured data after line maintenance.

[0043] The panoramic state perception unit also includes a data preprocessing module. This module filters, denoises, and performs synchronization phasor calculations on the acquired raw data. Filtering uses a digital notch filter to remove interference signals at the power frequency and its integer multiples. Denoising employs an adaptive Kalman filter algorithm to reduce measurement noise. Synchronization phasor calculation uses the Park transform method based on instantaneous sampled values ​​to calculate the amplitude, phase, and frequency information of voltage and current. The preprocessed data is encapsulated into real-time data frames and sent to the collaborative perception inference engine at a rate of 60 frames per second. The data frame format includes fields such as site number, measurement type, data value, timestamp, and quality flag.

[0044] III. Specific Implementation of the Collaborative Perception Reasoning Engine

[0045] The collaborative perception inference engine is the core of the system's intelligent analysis. It is used to collaboratively analyze the overall field group status data collected by the panoramic status perception unit, identify the safety level of the current operating status, and detect potential disturbance trends. Deployed in the regional control center and implemented using a high-performance computing server, this engine can process real-time data streams from dozens of stations and thousands of measurement points.

[0046] The safety level is classified into four levels: Normal, Warning, Emergency, and Critical. The Normal level indicates that the system is operating within its rated parameter range, all indicators are within the safe range, and no abnormal trends have been detected. The Warning level indicates that the system operating parameters have deviated, or a slight disturbance trend has been detected, but it has not yet exceeded the safety boundary. The Emergency level indicates that the system operating parameters are close to or slightly exceed the safety boundary, or a significant disturbance trend has been detected, which may trigger protection actions in a short time. The Critical level indicates that the system operating parameters have seriously exceeded the safety boundary, or a severe disturbance trend has been detected, the fault is developing and evolving, and immediate emergency control or protection measures are required.

[0047] The safety level is determined based on a comprehensive assessment of multiple indicators, including electrical quantity deviation indicators, non-electrical quantity anomaly indicators, and network topology stability indicators. Electrical quantity deviation indicators include voltage deviation, frequency deviation, power deviation, and phase angle difference deviation. Each deviation value is compared to a preset threshold, and the degree to which it exceeds the threshold serves as the basis for level determination. Non-electrical quantity anomaly indicators include abnormal component temperature, abnormal blade angle, and abnormal ambient wind speed. These indicators are compared to normal operating ranges to determine the degree of anomaly. Network topology stability indicators are based on switch state changes and fault signal detection. When a topology change is detected, this indicator directly indicates a higher safety level.

[0048] The aforementioned field-group coupling relationship model is a network mathematical model established using graph theory, specifically represented as a node impedance matrix or admittance matrix, used to characterize the electrical coupling strength between each new energy power station and between the power station and the main grid nodes; the state transition rules include state machine logic pre-set based on historical operating data, when the system operating parameters exceed the limits or the topology changes abruptly, the system is triggered to determine the state transition between four safety levels: normal, early warning, emergency, and severe, based on the severity of the limit exceedance and the influence range of the coupled nodes.

[0049] The detection of disturbance trends employs a combination of temporal analysis and pattern recognition. The collaborative perception inference engine pre-defines multiple disturbance patterns, including power fluctuation, voltage change, and frequency shift patterns. Each pattern corresponds to specific temporal characteristics. For example, a typical characteristic of power fluctuation is a large variance in the power curve over a short period; a typical characteristic of voltage change is a monotonically increasing or decreasing trend in voltage amplitude; and a typical characteristic of frequency shift is a deviation from the rated value and a continuous drift in a certain direction. The engine's internal pattern matching algorithm continuously monitors the real-time data stream. When the detected data features match a certain disturbance pattern with a pre-defined threshold, it is determined that a disturbance trend of that type exists.

[0050] The detection results of disturbance trends are output in quantitative form, including disturbance type, disturbance intensity, disturbance velocity, and estimated development time. Disturbance intensity is represented by a normalized value from 0 to 100, with higher values ​​indicating more severe disturbances. Disturbance velocity represents the rate of change of the disturbance parameter, with units determined according to the disturbance type; for example, the velocity of power fluctuations is measured in megawatts per second, and the velocity of voltage changes in kilovolts per second. The estimated development time represents the estimated time from the current moment until the disturbance reaches the protection action threshold or triggers a fault, measured in seconds.

[0051] The collaborative sensing inference engine also includes a disturbance source localization submodule. This submodule achieves precise location of the disturbance source based on temporal correlation analysis of the electrical quantity data of the entire field group. The temporal correlation analysis uses a mutual information algorithm to calculate the correlation between the electrical quantity data of each station and the time of disturbance occurrence. The calculation formula for the mutual information algorithm is as follows:

[0052]

[0053] Where X represents the electrical quantity data sequence of a certain substation, and Y represents the time sequence of disturbance occurrence. For specific value samples in sequence X, For specific value samples in sequence Y, Describes the joint probability distribution. and These represent the marginal probability distributions. A higher mutual information value indicates a stronger correlation between the data from that site and the disturbance, making the site more likely to be a source of the disturbance or close to it. The location results from the disturbance source localization submodule are simultaneously input into the integrated decision-making center to generate targeted control and protection strategies.

[0054] The working principle framework of the collaborative perception inference engine is as follows: Figure 3 As shown, the engine receives real-time data frames from the panoramic state perception unit, and after processing steps such as data preprocessing, feature extraction, pattern matching and trend prediction, it outputs the safety level assessment results and disturbance trend analysis results.

[0055] IV. Specific Implementation of the Integrated Decision-Making Center

[0056] The integrated decision-making center is the core decision-making unit of the system, used to generate unified control and protection strategies based on the security level and disturbance trends output by the collaborative perception inference engine. This center is deployed using a high-reliability industrial control server with redundant hardware to ensure system availability.

[0057] The integrated decision-making center incorporates a collaborative decision-making algorithm module. This module integrates the control objective function and protection action constraints into a unified optimization framework for joint solution. This design breaks the limitation of separate decision-making for control and protection in traditional architectures, enabling the system to simultaneously consider control effects and protection requirements within the same optimization framework, thus avoiding conflicts between control adjustments and protection actions.

[0058] The control objective function comprises three sub-objectives: power balance, voltage stability, and frequency regulation. The power balance objective requires minimizing the deviation between the total power output of the power plant cluster and the planned dispatch value; the objective function is the sum of squares of the differences between the actual and target active power values ​​of each power plant. The voltage stability objective requires maintaining the voltage of each bus within the allowable range; the objective function is the weighted sum of squares of voltage deviations. The frequency regulation objective requires maintaining the system frequency near its rated value; the objective function is the sum of squares of frequency deviations.

[0059] Specifically, the mathematical expressions for the control objective function and each sub-objective are as follows:

[0060] Integrated control objective function for:

[0061]

[0062] in:

[0063] (1) Power balance sub-objective :

[0064]

[0065] (2) Voltage stability sub-target :

[0066]

[0067] (3) Frequency adjustment sub-target :

[0068]

[0069] In the above formula, The total number of stations, To monitor the total number of busbars; and The first The actual and planned active power values ​​of each station; and The first The actual voltage value of the busbar is different from the rated reference value. For the first Voltage weighting coefficient for each busbar; and These are the actual and rated values ​​of the system frequency, respectively. , , These are the normalized weight coefficients for the three sub-objectives mentioned above.

[0070] When solving the above objective function, the protection action constraints must be satisfied. The protection action constraints include two constraints: protection selectivity constraints and action time constraints.

[0071] (1) The protection selectivity constraint requires the protection action to be performed according to the set selective coordination relationship, that is, when a fault occurs, only the protection device closest to the fault point is activated to avoid the upper protection from overstepping the level.

[0072] The protection selectivity constraint is: ,in For the first The operational status of each protective device Selective action allowable set based on topology (i.e., only protection actions closest to the fault point are allowed).

[0073] (2) The action time limit constraint requires the protection action to be completed within the specified time window. Different types and locations of protection correspond to different action time limits.

[0074] The action time limit constraint is: ,in For the first The actual operating delay of each protection device The maximum allowed duration of its actions.

[0075] The collaborative decision-making algorithm module employs a Model Predictive Control Framework (MMC) as the solution algorithm. Within each control cycle, the MMC predicts the system behavior for multiple future time steps based on a snapshot of the current state and selects the control sequence that minimizes the comprehensive objective function as the optimal strategy. The prediction time domain of the MMC is typically set to 1 to 10 seconds, and the control time domain to 0.1 to 1 second. Optimization is achieved using quadratic programming or mixed-integer programming algorithms. The comprehensive objective function incorporates multiple indicators, including power regulation error, voltage deviation, frequency deviation, and the number of protection actions. The weighting coefficients for each indicator are set according to system operation and safety requirements.

[0076] The model predictive control framework predicts future system behavior based on a discrete-time state-space prediction model. The basic mathematical expression of its prediction model is as follows:

[0077]

[0078] in, For the system in The state vector at any given moment (such as node voltage, system frequency, active / reactive output of equipment, etc.); for The control strategy sequence vector at any given time (such as the set active / reactive power adjustment command). A represents a measurable disturbance sequence (such as the bus voltage drop caused by a short-circuit fault); A is the system state matrix, B is the control input matrix, and D is the disturbance matrix. These matrix parameters are updated in real time using the field-group coupling model described above. The algorithm iteratively solves this prediction model in the prediction time domain to find the optimal control sequence that minimizes the comprehensive objective function. .

[0079] The specific workflow of the integrated decision-making center is as follows: First, it receives the security level and disturbance trend information output by the collaborative perception inference engine; then, it selects the corresponding decision mode according to the security level: routine control optimization is performed at the normal level, preventive control optimization is performed at the early warning level, emergency control optimization is performed at the emergency level, and emergency control and protection actions are prepared at the severe level; next, according to the selected decision mode, the collaborative decision-making algorithm module is run to generate control and protection strategies; finally, the generated strategy instructions are sent to the dynamic coordination execution unit.

[0080] The core principle framework of the integrated decision-making center is as follows: Figure 2 As shown in the diagram, the framework diagram illustrates the complete decision-making chain from input information to policy output, as well as the switching logic between various decision modes.

[0081] V. Specific Implementation of Dynamic Coordination Execution Unit

[0082] The dynamic coordination execution unit is the core of the system's timing coordination. It receives unified control and protection strategies output from the integrated decision-making center and dynamically adjusts the execution timing of control commands and protection actions based on real-time operating status. This unit is deployed in the regional control center and connected to the integrated decision-making center via an internal communication bus.

[0083] The core function of the dynamic coordination execution unit is to resolve the timing coordination problem between control regulation and protection actions. In traditional discrete architectures, control regulation and protection actions are executed by independent systems, lacking an effective coordination mechanism. This can easily lead to problems such as control regulation causing protection malfunctions or protection actions causing control failures. In the system of this invention, the dynamic coordination execution unit ensures that the timing coordination of control and protection is performed in an optimized state through real-time monitoring and dynamic adjustment.

[0084] When a control adjustment is detected as potentially causing a protection malfunction, the dynamic coordination execution unit automatically adjusts the control adjustment rate to avoid the protection's dead zone. The protection dead zone refers to the inactive area of ​​the protection device, typically set between 95% and 105% of the protection setting. When control adjustments cause electrical quantity changes to enter the protection dead zone, the protection device may make incorrect judgments and malfunction. The dynamic coordination execution unit predicts the trajectory of electrical quantity changes after control adjustments. When it predicts that the electrical quantity will enter the protection dead zone, it automatically reduces the control adjustment rate or performs adjustments in stages, allowing the electrical quantity changes to smoothly pass through the dead zone and preventing protection malfunctions.

[0085] When a protection trip causes a topology shift, the dynamic coordination execution unit immediately triggers an adaptive recalculation of the control strategy to suppress cascading effects. Protection tripping leads to changes in switch states, which in turn cause network topology changes. These topology changes result in power flow redistribution, potentially triggering overloads or voltage exceedances in other stations. Upon detecting a topology change, the dynamic coordination execution unit immediately sends a recalculation request to the integrated decision-making center. The integrated decision-making center recalculates the control strategy based on the new topology, and the dynamic coordination execution unit immediately implements the new strategy, suppressing cascading effects by rapidly adjusting control parameters.

[0086] The dynamic coordination execution unit also includes a timing matching submodule, which calculates the time window matching degree between control and protection actions. The time window matching degree is calculated based on statistical analysis of the action time series. The matching degree is low when the control action is expected to have a significant impact within the protection action's time window; and high when there is a sufficient time interval between the control and protection actions or when they have no mutual influence. When the time window matching degree is lower than a preset threshold, the dynamic coordination execution unit automatically extends the transition time of the control adjustment or prioritizes the execution of the protection action to ensure system safety. Specifically, the expected effective time period of the control adjustment is defined as... The preset operating time window of the protection device is ,in, To control the expected start time of the adjustment, To control the expected effective and end time of the adjustment, The preset operating time window start time for the protection device. This is the end time of the preset action time window for the protection device. It defines the time window matching degree. This refers to the time margin indicators that do not overlap between the two.

[0087] First, determine if there is a time overlap between the two: when the condition is met... and If time overlap is detected, the matching degree is directly adjusted. The penalty is a minimum value (e.g., 0);

[0088] When there is no time overlap, and the control adjustment is expected to be completed before the protection action (i.e.) When matching degree is ), The calculation formula is:

[0089]

[0090] When the calculated matching degree When the time interval is below a preset threshold (e.g., 20%, indicating that the time interval is too short and there is a risk of mutual interference), the dynamic coordination execution unit automatically extends the transition time of control adjustment or prioritizes the execution of protective actions to ensure system safety.

[0091] VI. Specific Implementation of the Graded Response Device

[0092] The graded response device is the system's protection execution unit, used to execute the protection strategy output by the dynamic coordination execution unit. This device adopts a three-tiered protection architecture: site-level protection, area-level protection, and system-level protection. These three levels coordinate their operations according to a pre-defined cooperation relationship.

[0093] Station-level protection is responsible for quickly isolating faults within the station. Typical actions of station-level protection include inverter DC-side short-circuit protection, inverter AC-side overcurrent protection, transformer differential protection, station bus protection, and station line protection. The operating time of station-level protection is set in milliseconds, typically 20~100ms, to ensure isolation is completed before the fault current reaches its peak value. Station-level protection devices operate independently of the upper-level control system and have the highest priority, ensuring reliable operation even in the event of communication failures.

[0094] Regional-level protection is responsible for isolating fault propagation between stations. Its main functions include regional tie-line protection, regional busbar protection, and regional grid disconnection protection. The operating time limit for regional-level protection is set to 1.5 to 2 times that of station-level protection, i.e., 100 to 200 ms, to ensure that station-level protection operates with priority. Regional-level protection needs to receive status information from multiple stations for comprehensive judgment, thus requiring data interaction with the dynamic coordination execution unit and the collaborative sensing inference engine.

[0095] System-level protection is responsible for disconnecting or reconnecting the renewable energy base to the main grid in the event of a severe fault. Triggering conditions for system-level protection include grid-side faults, severe internal faults within the renewable energy base leading to unstable operation, and severe over-limiting of power channels between the renewable energy base and the main grid. The action time limit for system-level protection is set to 2 to 3 times that of regional-level protection, i.e., 200 to 500 ms, but can be automatically shortened in emergency situations. The action of system-level protection will result in the physical connection between the renewable energy base and the main grid being disconnected or reconnected; therefore, the action requires strict safety verification.

[0096] The tiered response device employs an adaptive adjustment mechanism for its action time limit. The action time limit is dynamically adjusted based on the safety level output by the collaborative perception inference engine. When the safety level is emergency, the protection action time limit is shortened to 50% of the normal time limit to accelerate fault handling. When the safety level is severe, the protection action time limit is shortened to 25% of the normal time limit, and preset emergency control measures are activated simultaneously. This adaptive adjustment mechanism allows the system to flexibly adjust the protection response speed according to the operating status, maximizing system operating efficiency while ensuring safety.

[0097] VII. Specific Implementation of the Coordinated Power Regulation Device

[0098] The coordinated power regulation device is the system's control execution unit, used to execute the control strategy output by the dynamic coordination execution unit. This device adopts a hierarchical coordination architecture, including an active power regulation module and a reactive power regulation module.

[0099] The active power regulation module adjusts the active power output distribution of each power station according to the instructions of the integrated decision-making center. Active power regulation is implemented by adjusting the DC-side power commands of the inverters, adjusting the tracking angle of photovoltaic modules, or adjusting the blade angle of wind turbines. The issuance of regulation commands follows a hierarchical coordination principle: the upper-level coordinator is responsible for allocating the total active power regulation target across the entire power station cluster, that is, distributing the total regulation amount to each power station; the lower-level actuators are responsible for allocating the allocated regulation target to specific inverter units within each power station. The upper-level coordinator and the lower-level actuators exchange commands via a high-speed communication network, with communication latency controlled within 10ms to ensure real-time regulation response.

[0100] The reactive power regulation module dynamically allocates reactive power compensation capacity to each substation based on voltage stability targets. Reactive power regulation is implemented through methods such as adjusting inverter power factor commands, switching reactive power compensation capacitor banks, and adjusting static var compensators (SVCs) or static synchronous compensators (SNCs). Reactive power regulation also employs a hierarchical coordination architecture. The upper-level coordinator calculates the reactive power compensation needs of each substation based on the overall grid voltage distribution, while the lower-level actuators allocate the reactive power compensation targets to each reactive power compensation device.

[0101] The hierarchical coordination architecture of the collaborative power regulation unit ensures that regulation commands can be quickly and accurately transmitted to each execution unit. The upper-level coordinator, deployed in the regional control center, possesses a view and information processing capability across the entire field group and can perform global optimization calculations. The lower-level actuators, deployed in the control rooms of each site, have direct control capabilities over local equipment and can translate regulation targets into specific equipment commands. The two layers interact through a standardized communication protocol, the format of which includes fields such as regulation target value, regulation priority, regulation time limit, and feedback requirements.

[0102] VIII. Specific Implementation of the Feedback Closed-Loop Unit

[0103] The feedback closed-loop unit is the core of the system's learning and optimization. It is used to monitor the performance of the hierarchical response device and the collaborative power regulation device in real time, and to feed the performance results back to the collaborative perception inference engine to update the overall group state data and adjust subsequent decision-making strategies.

[0104] The monitoring content of the feedback closed-loop unit includes the execution status and results of protection actions, as well as the execution status and effects of power regulation. Monitoring of protection actions is achieved by reading the action reports of the protection devices, which include protection type, action time, action current, and action voltage. Monitoring of power regulation is achieved by reading the operating data of the inverter and compensation device, which includes actual power values, power change rate, and regulation response time.

[0105] The feedback loop unit compares the monitored data with the expected target and calculates the execution deviation. The execution deviation is calculated using the root mean square error (RMSE) method, reflecting the overall degree of deviation between the execution result and the target value. When the execution deviation exceeds a preset threshold, the feedback loop unit generates a deviation alarm and sends it to the collaborative perception inference engine, triggering the inference engine to reassess the operating status and safety level. When the execution deviation is within the allowable range, the feedback loop unit archives and stores the execution data for subsequent performance analysis and model optimization.

[0106] The feedback loop unit also includes a model self-learning function, which utilizes historical execution data to optimize the collaborative decision-making algorithm module and the disturbance trend prediction algorithm online. The model self-learning employs recursive least squares or Kalman filtering algorithms. Based on new observation data, it continuously updates the system state matrix elements in the predictive control model, the index weight coefficients in the integrated objective function, and the threshold parameters in the disturbance pattern matching algorithm using recursive least squares or Kalman filtering algorithms. Through this model self-learning function, the system can continuously adapt to changes in operating conditions, constructing an intelligent collaborative operation mechanism with self-optimization capabilities.

[0107] IX. System Integration and Communication Architecture

[0108] The functional units of the integrated control and protection system for new energy base clusters are connected via a high-speed redundant communication network. The communication network adopts a dual-network redundancy architecture, with the primary network and backup network operating simultaneously. In the event of a failure in either network, it automatically switches to the other network to ensure communication reliability.

[0109] The communication network adopts the IEC 61850 standard as its communication protocol foundation, which defines the communication architecture and data model of the substation automation system. In this system, the panoramic status perception unit sends real-time data frames to the collaborative perception inference engine via the MMS protocol. The collaborative perception inference engine and the integrated decision-making center transmit inference results and control requests via the GOOSE protocol. The integrated decision-making center and the dynamic coordination execution unit transmit policy instructions via the MMS protocol. The dynamic coordination execution unit and the hierarchical response device and the collaborative power regulation device transmit execution instructions via the GOOSE protocol. The feedback closed-loop unit collects execution result data via the MMS protocol.

[0110] The latency requirements of the communication network are determined based on different data streams: the transmission latency of real-time electrical quantity data should not exceed 10ms, the transmission latency of protection action commands should not exceed 5ms, the transmission latency of control strategy commands should not exceed 20ms, and the transmission latency of feedback data should not exceed 50ms. The bandwidth requirements of the communication network are determined based on the data volume. The panoramic state perception unit sends the largest amount of data to the collaborative perception inference engine, approximately 10 megabytes per second, while the bandwidth requirements of other data streams are relatively smaller.

[0111] The system uses the IEEE 1588 precision time protocol for time synchronization. The internal clocks of each functional unit are synchronized with the time server in the regional control center, achieving a time synchronization accuracy of 1μs. Time synchronization is fundamental to ensuring the coordinated operation of the system, especially in the timing coordination of protection actions and the rate adjustment of control regulation, where time accuracy is crucial.

[0112] 10. System Workflow Examples

[0113] To illustrate the collaborative relationships between the various units of the system, a specific operational scenario is used as an example. Assume the new energy base cluster is operating normally, and a short-circuit fault suddenly occurs near one of the stations. The system handles this according to the following procedure:

[0114] The first step involves the panoramic state perception unit immediately acquiring electrical quantity data, non-electrical quantity data, and network topology data at the moment the fault occurred. Electrical quantity data shows a sudden drop in bus voltage and a rapid increase in short-circuit current near the fault point; non-electrical quantity data shows no significant change; network topology data indicates the fault occurred between the station's incoming switch and the bus. The data preprocessing module filters the raw data and performs synchronous phasor calculations, generating real-time data frames and sending them to the collaborative perception inference engine.

[0115] The second step involves the collaborative perception inference engine receiving real-time data frames and performing a safety level assessment and disturbance trend analysis. Based on the abnormal characteristics of the electrical quantity data, the engine determines the current safety level to be severe, the disturbance type to be a short-circuit fault, the disturbance intensity to be 100, and the estimated development time to be 0 seconds, indicating that the fault has already occurred. Simultaneously, the engine activates the disturbance source localization submodule, which quickly locates the fault point at the site's incoming line where the short-circuit fault occurred based on temporal correlation analysis. The analysis results are then sent to the integrated decision-making center.

[0116] Third, upon receiving information about the severe safety level and short-circuit fault, the integrated decision-making center immediately enters emergency decision-making mode. The collaborative decision-making algorithm module within the center rapidly generates protection and emergency control strategies: the protection strategy requires station-level protection to immediately trip the fault point, area-level protection to prepare backups, and system-level protection to monitor fault development; the emergency control strategy requires other stations to increase output to compensate for the output loss at the faulty station, while simultaneously adjusting reactive power compensation to maintain voltage stability. These strategy instructions are then sent to the dynamic coordination and execution unit.

[0117] Fourthly, after receiving the policy instructions, the dynamic coordination execution unit first checks the timing coordination between control adjustments and protection actions. The detection results show that the protection actions will cause network topology changes, potentially triggering power fluctuations in other stations. The dynamic coordination execution unit immediately sends a policy recalculation request to the integrated decision center and simultaneously activates the timing matching submodule to calculate the optimal execution timing. The integrated decision center quickly updates the control policy based on the new topology prediction, and the updated policy is sent to the dynamic coordination execution unit.

[0118] Fifth, the dynamic coordination execution unit issues execution commands according to the optimal timing. Upon receiving the protection strategy, the tiered response device immediately activates the station-level protection device, tripping the incoming switch at the fault point and isolating the fault. Upon receiving the control strategy, the coordinated power regulation device rapidly increases the output of the active power regulation modules at each station, while the reactive power regulation modules dynamically adjust their compensation capacity. Real-time monitoring data during the execution process is sent to the feedback closed-loop unit.

[0119] Step 6: Feedback closed-loop unit monitoring execution effect: protection actions are completed correctly, fault points are isolated; output from other stations increases, compensating for fault losses, and system power remains basically balanced; voltage is maintained within allowable range. Execution deviations are within allowable ranges, and feedback data is archived and stored. The collaborative perception inference engine updates the status snapshot based on the latest operational data, the safety level is adjusted from severe to warning, and the system enters a new steady-state operating state.

[0120] The entire processing flow is completed within hundreds of milliseconds, demonstrating the system's high-speed response and collaborative processing capabilities. Through a complete closed loop of perception, reasoning, decision-making, execution, and feedback, the system achieves integrated coordination of control and protection, effectively avoiding problems such as information sharing issues, timing mismatches, and chain reactions found in traditional discrete architectures.

[0121] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An integrated system for coordinated control and protection of new energy base clusters, characterized in that, include: The panoramic status perception unit is used to collect electrical quantity data, non-electrical quantity data and network topology data of each station in the new energy base cluster in real time, and build a unified status snapshot of the entire cluster. A collaborative perception inference engine is used to collaboratively analyze the state snapshots collected by the panoramic state perception unit, identify the safety level of the current operating state, and detect potential disturbance trends. The collaborative perception inference engine includes a disturbance source localization submodule. The disturbance source localization submodule achieves accurate localization of the disturbance source based on the temporal correlation analysis of the electrical quantity data. The temporal correlation analysis uses a mutual information algorithm to calculate the correlation between the electrical quantity data of each station and the disturbance occurrence time. The mutual information algorithm obtains the mutual information value by calculating the sum of the logarithms of the products of the joint probability distribution of the electrical quantity data sequence of the station and the disturbance occurrence time sequence and their respective marginal probability distributions. The larger the mutual information value, the higher the correlation between the electrical quantity data of the station and the disturbance. The location results of the disturbance source location submodule are simultaneously input into the integrated decision center to generate targeted control and protection strategies. An integrated decision-making center is used to generate unified control and protection strategies based on the security level and disturbance trend output by the collaborative perception and inference engine. The integrated decision-making center has a built-in collaborative decision-making algorithm module, which incorporates the control objective function and protection action constraints into a unified optimization framework for joint solution. The dynamic coordination execution unit is used to receive the unified control strategy and protection strategy output by the integrated decision center, and dynamically adjust the execution sequence of control commands and protection actions according to the real-time operating status; when the risk of protection maloperation caused by control adjustment is detected, the dynamic coordination execution unit automatically adjusts the control adjustment rate to avoid the dead zone of protection action; when the network topology change is caused by protection tripping is detected, the dynamic coordination execution unit immediately triggers the integrated decision center to perform adaptive recalculation of control strategy to suppress chain reaction; A graded response device is used to execute the protection strategy output by the dynamic coordination execution unit; A coordinated power regulation device is used to execute the control strategy output by the dynamic coordination execution unit; The feedback closed-loop unit is used to monitor the execution effects of the hierarchical response device and the collaborative power adjustment device in real time, and to feed back the execution results to the collaborative perception inference engine to update the state snapshot and adjust the subsequent decision-making strategy.

2. The integrated control and protection system for new energy base clusters according to claim 1, characterized in that, The electrical quantity data includes the AC side current, voltage, active power, and reactive power of each inverter; the non-electrical quantity data includes the photovoltaic module temperature, wind turbine blade angle, and ambient wind speed; the network topology data includes bus connection relationships, switch on / off status, and line parameters.

3. The integrated system for coordinated control and protection of new energy base clusters according to claim 1, characterized in that, The safety level is divided into four levels: normal, early warning, emergency, and severe; the disturbance trend includes power fluctuation trend, voltage change trend, and frequency offset trend. The collaborative perception reasoning engine performs the collaborative analysis based on preset state transition rules and field group coupling relationship models.

4. The integrated system for coordinated control and protection of new energy base clusters according to claim 1, characterized in that, The control objective function includes a power balance objective, a voltage stability objective, and a frequency regulation objective; the protection action constraints include protection selectivity constraints and action time constraints; the power balance objective requires minimizing the deviation between the total power output of the power plant group and the scheduling plan value; the voltage stability objective requires that the voltage of each bus be maintained within the allowable range; the frequency regulation objective requires that the system frequency be maintained near the rated value; the protection selectivity constraint requires that the protection action be executed according to the set selective coordination relationship; and the action time constraint requires that the protection action be completed within a specified time window.

5. The integrated system for coordinated control and protection of new energy base clusters according to claim 1, characterized in that, The collaborative decision-making algorithm module adopts a model predictive control framework. In each control cycle, the model predictive control framework predicts the system behavior for multiple future time steps based on the current state snapshot, and selects the control sequence that minimizes the comprehensive objective function as the optimal strategy. The comprehensive objective function is a weighted fusion of indicators such as power regulation error, voltage deviation, frequency deviation, and the number of protection actions.

6. The integrated system for coordinated control and protection of new energy base clusters according to claim 1, characterized in that, The dynamic coordination execution unit further includes a timing matching submodule, which is used to calculate the time window matching degree between control adjustment actions and protection actions. During the calculation, the timing matching submodule determines whether there is a time overlap between the expected effective time period of the control adjustment and the preset action time window of the protection device. If there is no time overlap and the control adjustment is expected to be completed before the protection action, the time window matching degree is calculated using the difference between the expected effective end time of the control adjustment and the start time of the preset action time window. When the time window matching degree is lower than a preset threshold, the dynamic coordination execution unit automatically extends the transition time of the control adjustment or prioritizes the execution of the protection action.

7. The integrated system for coordinated control and protection of new energy base clusters according to claim 1, characterized in that, The tiered response device is divided into three levels: station-level protection, regional-level protection, and system-level protection. Station-level protection is responsible for quickly isolating internal station faults; regional-level protection is responsible for isolating fault propagation between stations; and system-level protection is responsible for disconnecting or reconnecting the new energy base to the main grid during a fault. The action time limit of the tiered response device adopts an adaptive adjustment mechanism, dynamically correcting the action time limit based on the safety level output by the collaborative perception inference engine. When the safety level is emergency, the action time limit is shortened to 50% of the normal time limit. When the security level is severe, the action time limit is shortened to 25% of the normal time limit.

8. The integrated system for coordinated control and protection of new energy base clusters according to claim 1, characterized in that, The coordinated power regulation device adopts a hierarchical coordination architecture, including an active power regulation module and a reactive power regulation module. The upper-level coordinator is responsible for allocating the overall active power regulation target and reactive power regulation target across the entire power plant group, while the lower-level actuator is responsible for distributing the regulation target allocated by the upper-level coordinator to specific inverter units within each power plant. The active power regulation module adjusts the active power output allocation of each power plant according to the instructions of the integrated decision-making center, and the reactive power regulation module dynamically allocates the reactive power compensation capacity of each power plant according to the voltage stability target.

9. The integrated system for coordinated control and protection of new energy base clusters according to claim 1, characterized in that, The feedback closed-loop unit compares the monitoring data with the expected target and calculates the execution deviation using the root mean square error method. When the execution deviation exceeds a preset threshold, the feedback closed-loop unit generates a deviation alarm and sends it to the collaborative perception inference engine, triggering the collaborative perception inference engine to reassess the operating status and security level. The feedback closed-loop unit also includes a model self-learning function, which uses historical execution data and employs recursive least squares or Kalman filtering algorithms to perform online optimization of the collaborative decision-making algorithm module and the disturbance trend prediction algorithm.

Images