Wind and light storage and charging preview method and system based on digital twinning and storage medium

Abstract

The invention provides a digital twinborn-based wind and light storage and charging rehearsal method and system and a storage medium. The method comprises the following steps: acquiring a system digital twinborn model, a topological connection relationship and real-time operation data, determining an initial fault probability in combination with an equipment health state index and historical alarm data, sampling to generate an initial fault event, and constructing an event vector; according to the system frequency deviation rate, the node voltage offset and the energy storage SOC balance degree, performing weighted calculation on a transient pressure index of a key node; updating a state transition matrix based on the index, calculating secondary fault probability distribution in combination with the system state at the previous moment, and determining a new fault component; synchronously calculating multi-dimensional risk vectors including new energy removal power, load power loss and power grid topology connectivity decline; and when a termination condition is satisfied, terminating rehearsal, and outputting a fault propagation path, a system state snapshot and a multi-dimensional risk vector.

Description

Technical Field

[0001] This application belongs to the field of simulation, and in particular relates to a method, system and storage medium for simulation of wind, solar and energy storage based on digital twins. Background Technology

[0002] Integrated wind-solar-storage-charging systems incorporate power electronic equipment. Even minor initial disturbances, under specific operating conditions, can propagate through various pathways, including electrical and informational ones, threatening the safe and stable operation of the power system. Therefore, it is necessary to simulate the propagation process of cascading faults and assess potential risks. Current power grid risk assessment methods, based on offline component failure rate data, struggle to reflect the real-time impact of operating conditions on fault evolution paths and cannot represent the relationships between different links in the fault chain, leading to discrepancies between simulation results and actual conditions.

[0003] In the field of cascading failure simulation and analysis, existing technologies employ models based on the "Nk" criterion or deterministic stability criteria to simulate fault propagation. These models typically trigger the disconnection of a line or transformer component after detecting that its physical quantity exceeds a preset threshold, thus initiating the next round of power flow or stability calculations. However, they neglect the impact of transient pressure differences at each node and the component's own health status on fault propagation. In fact, voltage deviations, frequency fluctuations, or the adjustment status of energy storage resources around a node can all affect the vulnerability of components in that area to disturbances, thereby altering the direction and speed of fault propagation. Furthermore, existing risk assessment indicators focus on load loss or the final state of system disconnection, failing to provide a comprehensive and multi-dimensional risk representation of the entire cascading failure process from multiple dimensions, including renewable energy losses, the degree of damage to the grid topology, and load loss. Therefore, a cascading failure simulation method is needed that can integrate real-time pressure and component health status, and provide multi-dimensional risk representation. Summary of the Invention

[0004] This invention proposes a wind-solar-storage-charging pre-simulation method based on digital twins to address the problems of existing technologies that cannot integrate real-time pressure and component health status, and cannot perform multi-dimensional risk representation for cascading failure pre-simulation. The method includes:

[0005] Acquire the digital twin model, topological connection relationship and real-time operation data of the wind-solar-storage-charging system; determine the initial failure probability of each component based on the health status index and historical alarm data of each equipment component, and sample and generate initial failure events from them; combine the real-time operation data to determine the failure location, failure type and initial disturbance energy of the event, and form an initial failure event vector.

[0006] Within each time step of the rehearsal, the transient pressure index of the key nodes in the system is calculated.

[0007] Based on the transient stress index of each key node, update a state transition matrix representing the propagation of faults among components; based on the updated state transition matrix and the system state at the previous time step, calculate the probability distribution of secondary faults occurring within the current step size, and identify new faulty components.

[0008] Simultaneously calculate a multidimensional risk vector consisting of renewable energy power cut-off, load loss, and grid topology connectivity reduction; terminate the chain reaction simulation when the termination condition is met; output the chain reaction propagation path, system state snapshots at each stage, and the multidimensional risk vector.

[0009] Furthermore, this invention also relates to a wind-solar-storage-charging pre-simulation system based on digital twins, comprising the following modules:

[0010] The first determining module is used to acquire the digital twin model, topological connection relationship and real-time operation data of the wind-solar-storage-charging system; based on the health status index and historical alarm data of each equipment component, determine the initial failure probability of each component, and sample and generate initial failure events from them; combined with the real-time operation data, determine the failure location, failure type and initial disturbance energy of the event, and form an initial failure event vector.

[0011] The calculation module is used to calculate the transient pressure index of the key nodes in the system at each time step of the pre-simulation.

[0012] The second determination module is used to update a state transition matrix representing the propagation of faults among components based on the transient pressure index of each key node; and to calculate the probability distribution of secondary faults occurring within the current step size based on the updated state transition matrix and the system state at the previous time step, and to determine the new faulty component.

[0013] The output module is used to synchronously calculate a multi-dimensional risk vector consisting of renewable energy power cut-off, load loss, and grid topology connectivity reduction; when the termination condition is met, the current chain reaction simulation is terminated; the propagation path of the chain reaction, system state snapshots at each stage, and the multi-dimensional risk vector are output.

[0014] This invention improves the realism of fault simulation by generating initial fault events through the fusion of health status indices of equipment components and historical alarm data. It utilizes a transient stress index composed of system frequency deviation, node voltage offset, and energy storage balance, using this index as a key factor influencing fault propagation, enabling the evolution path of cascading reactions to reflect local system vulnerability differences. Furthermore, by constructing a multi-dimensional risk vector including renewable energy power cut-off, load loss, and grid topology connectivity, the consequences of cascading faults can be better assessed. Attached Figure Description

[0015] Figure 1 A flowchart of the first embodiment;

[0016] Figure 2 This diagram illustrates the initial failure probability determination and event generation.

[0017] Figure 3 This is a schematic diagram illustrating the transient pressure index.

[0018] Figure 4 This is a schematic diagram for determining the termination condition of the pre-rehearsal. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] The term "multiple" in this application refers to two or more. Furthermore, it should be understood that the terms "first," "second," etc., used in the description of this application are used only for descriptive purposes and should not be construed as indicating or implying relative importance, nor as indicating or implying order.

[0021] In the first embodiment, the present invention proposes a wind-solar-storage-charging pre-simulation method based on digital twins, such as... Figure 1 ,include:

[0022] S1. Obtain the digital twin model, topological connection relationship and real-time operation data of the wind-solar-storage-charging system; determine the initial failure probability of each component based on the health status index and historical alarm data of each equipment component, and sample and generate initial failure events from them; combine the real-time operation data to determine the failure location, failure type and initial disturbance energy of the event, and form an initial failure event vector.

[0023] Using professional simulation software such as MATLAB / Simulink or DIgSILENT Power Factory, a digital twin model of the wind-solar-storage-charging system is constructed, including wind turbines, photovoltaic arrays, energy storage units, charging loads, conventional generators, transformers, transmission lines, and corresponding control and protection systems. The current topology connection relationship is exported by reading the database of the power grid energy management system (EMS) or the dispatching system (SCADA), and stored in the form of a node-branch table or adjacency matrix. Through phasor measurement units (PMUs) deployed in substations, sensors installed on various devices, and the battery management system (BMS), real-time operating data of voltage, frequency, active power, reactive power, and energy storage state of charge (SOC) of each node are collected and uploaded to the pre-simulation calculation platform via a high-speed communication network.

[0024] In an optional embodiment, determining the initial failure probability of each component and sampling from it to generate initial failure events includes:

[0025] Using a logistic regression model, the health status index of the equipment components and the historical alarm frequency per unit time are used as inputs to calculate the initial failure probability of each component.

[0026] The Monte Carlo sampling method is used to randomly select a component as the source of the initial failure event based on the calculated probability distribution.

[0027] Specifically, a logistic regression model is constructed and trained. The model's output is a probability value between 0 and 1, mathematically expressed as follows: Where P is the failure probability. Represents a health status index. This represents the frequency of historical warnings, while These are the weight coefficients obtained by training the model using historical data. For example, for a transformer with a health status index of 0.6 and 3 alarms in the past month, assuming the trained model weights are... The initial failure probability is then calculated as follows: This process will be repeated for all critical components in the power grid to obtain an initial list of failure probabilities for each component.

[0028] The Monte Carlo method is used to arrange the failure probabilities of all components into a probability space. For example, suppose there are three components A, B, and C, with calculated failure probabilities of 0.214, 0.05, and 0.1, respectively. These three probability values ​​are normalized to obtain their relative probabilities. A random number between 0 and 1 is generated. If the random number falls within the probability interval representing component A, then component A is selected as the initial failure event. Figure 2The process ensures that the component with the highest theoretical failure probability is more likely to be selected as the trigger point for the simulation, thus making the simulation closer to actual risks.

[0029] In an optional embodiment, dissolved gas analysis data, winding hot spot temperature, and partial discharge quantity are collected in the transformer oil. A health status index HI between 0 and 1 is output using a fuzzy comprehensive evaluation method or a trained neural network model. Simultaneously, the number of overload and overvoltage alarms in the past year is counted and recorded as follows: The initial failure probability of the transformer can then be determined by a logistic regression model, for example... ,in and These are weighting coefficients; calculated for all components in the system. Then, a roulette wheel selection method is used for sampling to select a component as the initial fault component; for example, if line L5 is selected, the fault location is determined to be L5; based on the historical fault statistics of the line, the occurrence probability of three-phase short circuit and single-phase ground fault types is set, and the current fault is determined to be a three-phase short circuit by sampling again; based on the real-time operating data at the time of the fault occurrence, the short-circuit current on L5 is calculated, and the initial value of the generated Joule heat integral or transient energy function is used as the initial disturbance energy; an initial fault event vector is formed, such as [line L5, three-phase short circuit, 30 km from the head end, 5 megajoules].

[0030] S2, at each time step of the pre-simulation, calculate the transient pressure index of the key nodes in the system;

[0031] The renewable energy collection point, major load centers, and tie line outlet busbar were selected as key nodes. After applying an initial fault event in the digital twin model, transient simulations were performed with a step size of 10ms. At the end of each step, the system frequency f and the voltage of key node i were read from the simulation. and the SOC value of the energy storage unit that is electrically closest to the node. ; Calculate the system frequency deviation rate Δf = |f - 50| / 50; Calculate the voltage offset at node i. ,in Given the rated voltage of the node; calculate the SOC balance of the energy storage unit. ,in The average state of charge of all energy storage units; the transient stress exponent of critical node i. ,in These are preset weighting coefficients, for example, 0.4, 0.4, and 0.2 respectively.

[0032] In an optional embodiment, calculating the node transient pressure index for key nodes in the system at each time step of the pre-simulation includes:

[0033] The transient pressure exponent at node i at time t is calculated using the following formula. :

[0034]

[0035] in, The system frequency deviation rate, Let be the voltage offset at node i. For the SOC balance of energy storage units, and These are the system's rated frequency and rated voltage, respectively.

[0036] This formula integrates three key power grid transient stability indicators, with weighting coefficients of 0.4, 0.4, and 0.2 reflecting the importance of frequency stability, voltage stability, and energy storage system coordination to overall power grid security, respectively. Frequency deviation and voltage offset, which represent power angle and voltage stability, are given higher weights; while the SOC balance of energy storage units reflects the coordination and sustainability of backup support capabilities, with a relatively lower weight but still indispensable.

[0037] To calculate the transient stress index of a specific node, it is necessary to first obtain the power grid operating data at that moment through transient simulation. Assume the rated frequency... The rated voltage of a node i is 50Hz. The voltage is 220 kV. At time t, the simulation results show that the system frequency drops to 49.7 Hz, the voltage at node i drops to 210 kV, and the SOC balance index of each energy storage unit decreases. The calculated value is 0.8. The sub-indicators are calculated as follows: frequency deviation rate is 0.006; voltage offset is approximately 0.045; energy storage imbalance is 0.2. Substituting these values ​​into the formula, the transient pressure index TPI(i,t) of this node is obtained as 0.0604. Figure 3 The higher the dimensionless value, the greater the pressure the node bears, and the more prone the connected components are to failure.

[0038] S3, based on the transient pressure index of each key node, update a state transition matrix representing the propagation of faults among components; based on the updated state transition matrix and the system state at the previous time step, calculate the probability distribution of secondary faults occurring within the current step size, and determine the new faulty component.

[0039] Let the currently faulty component be s, the connection node be m, and any healthy target component be t, with connection node n; the impedance matrix is ​​obtained by inverting the admittance matrix. Electrical distance between source and target components That is, the mutual impedance between nodes m and n. The modulus; the transition probability of target component t Proportional to the transient pressure exponent of node n Inversely proportional to electrical distance ,Right now , where K is the normalization coefficient; calculate the transition probability for all healthy components, forming one row of the state transition matrix; the system state at the previous time step is the set of components that have failed; take all failed components as sources, calculate the sum of their transition probabilities to the same healthy target component t, and obtain the total probability of secondary failure occurring in t within this step; perform Monte Carlo sampling on all healthy components, and if the random number between 0 and 1 generated for component t is less than the total failure probability, then component t is determined to be a new failed component.

[0040] To establish a mathematical model representing the failure propagation path and probability, in an optional embodiment, updating a state transition matrix representing the propagation of failure between components based on the transient stress index of each key node includes:

[0041] Calculate the transition probability from source component j to target component i. The calculation formula is:

[0042]

[0043] in, The transient stress index is the number of nodes connected to the target component i. The electrical distance between component j and component i. Let j be the set of potential target components that are electrically connected to source component j, and let j be the normalization factor for all potential target components in the denominator.

[0044] Specifically, the model posits that the probability of a fault propagating from a failed component j to a healthy component i is proportional to the transient stress at the node containing the target component i and inversely proportional to the electrical distance between the two components. The transient stress TPI(i,t) represents the vulnerability of the target component; higher stress increases the likelihood of failure. The electrical distance... This represents the resistance to fault propagation; the greater the distance, the weaker the electrical connection, and the more difficult it is for the fault to propagate. (Numerator) The combination of these two factors constitutes the original propagation tendency from j to i.

[0045] Assume a faulty transformer j is directly connected to three lines i, k, and m. At time t, the calculated transient pressure indices at each line connection node are TPI(i,t) = 0.06, TPI(k,t) = 0.04, and TPI(m,t) = 0.03. The electrical distances between transformer j and the three lines are as follows: , , Calculate the initial propagation probabilities for each line: 1.2 for line i; 0.4 for line k; and 0.375 for line m. Calculate the normalized denominator, i.e., the sum of all probabilities, which is 1.975. Calculate the transition probabilities from transformer j to each line: the probability of transitioning to line i... The probability of reaching line k The probability of reaching line m The probability values ​​constitute part of the state transition matrix.

[0046] In an optional embodiment, the step of calculating the probability distribution of secondary faults occurring within the current step size based on the updated state transition matrix and the system state at the previous time step, and determining new faulty components, includes:

[0047] Multiply the system fault state vector of the previous time t-1 with the updated state transition matrix to obtain a cumulative effect vector representing the impact of fault propagation on each non-faulty component;

[0048] The cumulative effect vector is normalized to obtain the probability distribution of secondary faults occurring within the current step size;

[0049] The component with the highest probability value in the probability distribution is selected as the new fault component within the current step size.

[0050] All components in the system constitute a state vector. For example, in a 5-component system, if components 2 and 4 have failed, the fault state vector at time t-1 can be represented as [0,1,0,1,0]. Simultaneously, there is a state transition matrix P updated at time t, where the elements... This represents the probability that a fault propagates from j to i.

[0051] Multiplying the state vector by the transition matrix allows us to calculate the total fault stress experienced by each healthy component from all failed components. For example, for healthy component 1, the total fault stress is equal to the sum of the stress from component 2 and the stress from component 4, i.e. This calculation is performed on all healthy components, resulting in a cumulative effect vector. Assume that after calculation, the cumulative effect values ​​for healthy components 1, 3, and 5 are 0.3, 0.8, and 0.1, respectively. Normalizing this vector yields the probability distribution of secondary failures for each component: the failure probability for component 1 is 0.25, for component 3 it is approximately 0.67, and for component 5 it is approximately 0.08. Based on the maximum probability criterion, component 3, with the highest probability value, is selected as the new failure component within the current step t.

[0052] S4 synchronously calculates a multidimensional risk vector consisting of renewable energy power cut-off, load loss, and grid topology connectivity reduction; when the termination condition is met, terminate this chain reaction simulation; output the chain reaction propagation path, system state snapshots at each stage, and the multidimensional risk vector.

[0053] At each step, the total power of wind and solar power units affected by protection actions or abnormal shutdowns is statistically analyzed using the simulation model, serving as the renewable energy power cut-off; the total load power lost due to line disconnection or low-voltage load shedding is statistically analyzed, serving as the load loss; the power grid topology is treated as a graph, and the ratio of the square of the current number of nodes to the total number of edges is calculated to represent the topological connectivity, and the decrease in topological connectivity compared to the initial state is calculated; these three factors constitute a multidimensional risk vector. Calculate the Euclidean modulus of the vector. The growth rate is Simultaneously, the sum of the maximum power that all energy storage units can currently generate is calculated. The rate of change The first threshold is set to 0.01 and the second threshold is set to 0.005. When the risk vector magnitude growth rate is less than 0.01 for two consecutive steps and the total adjustable power change rate of energy storage is less than 0.005, it is determined that a new quasi-steady state has been reached and the pre-simulation is terminated.

[0054] During the rehearsal, whenever a new component fails, the name of the component and the time step of the failure are recorded. These records are then chained together in chronological order to form a propagation path, for example: 0.1 seconds: three-phase short circuit on line L5 -> 0.5 seconds: generator G2 disconnects due to loss of synchronization -> 0.9 seconds: transformer T1 shuts down due to overload. Simultaneously, after each fault event, all major electrical quantities in the digital twin model, such as the voltage amplitude and phase angle of all buses, the power flow of all branches, and the output of all generators, are saved as a data file as a snapshot of the state at that stage. After the rehearsal ends, the specific values ​​of the multidimensional risk vector for the last step are output, such as 300 MW of renewable energy power disconnection, 150 MW of load power loss, and a 25% decrease in topology connectivity.

[0055] To assess the losses caused by each cascading failure from three dimensions and form a risk vector R(t), in an optional embodiment, the synchronous calculation of a multi-dimensional risk vector consisting of renewable energy power cut-off, load loss, and grid topology connectivity degradation includes:

[0056] The difference between the output of photovoltaic and wind turbines before and after the simulation will be used as the power cut-off of new energy sources.

[0057] The total power loss on the user side caused by the fault is taken as the load power loss.

[0058] The decrease in grid topology connectivity is represented by calculating the change in the maximum eigenvalue of the grid adjacency matrix before and after the simulation.

[0059] Specifically, the first dimension, renewable energy power cut-off, focuses on the impact on clean energy power generation. For example, before a new component failure occurs at time t, the total output of photovoltaic and wind turbines is 800 MW. After the failure, some renewable energy plants are cut off, and the total output drops to 650 MW. In this case, the risk value for this dimension is 150 MW.

[0060] The second dimension, load loss, represents the impact on electricity users. Before the fault, the total load was 5000 MW. The fault caused the grid to disconnect in a certain area, resulting in a 300 MW load loss in that area. Therefore, the risk value for this dimension is 300 MW. The third dimension, the decrease in grid topology connectivity, assesses the system's integrity from a network structure perspective. Assume the maximum eigenvalue of the grid adjacency matrix before the fault was 15.4. The fault caused a critical tie line to break, changing the network structure, and the maximum eigenvalue of the new adjacency matrix decreased to 12.1. Therefore, the risk value for this dimension is 3.3. In summary, the risk vector at time t is [150, 300, 3.3].

[0061] In an optional embodiment, terminating the current chain reaction simulation when the termination condition is met includes:

[0062] Normalize each component of the multidimensional risk vector R(t) to obtain a dimensionless normalized risk vector. ;

[0063] Calculate the Euclidean norm of the normalized risk vector ;

[0064] When the growth rate The simulation is terminated when the power is less than the first threshold for two consecutive steps and the total adjustable power change rate of all energy storage units in the system is less than the second threshold.

[0065] The multidimensional risk vector containing different physical units is normalized, transforming it into a dimensionless vector with a uniform scale. For example, based on the system's total installed capacity and total load baseline, the risk vector [150 MW, 300 MW, 3.3] is normalized to [0.1, 0.06, 0.2]. The Euclidean norm of the normalized vector is calculated, which is approximately 0.231. This norm value comprehensively reflects the overall risk level at the current moment.

[0066] The simulation terminates based on the simultaneous fulfillment of two conditions. The first condition is that risk growth slows down. This involves continuously tracking changes in the risk range value. Assume that at three consecutive times t-2, t-1, and t, the calculated risk range values ​​are 0.220, 0.228, and 0.231, respectively. The growth rate from t-2 to t-1 is approximately 0.036, less than the first threshold of 0.05. The growth rate from t-1 to t is approximately 0.013, also less than 0.05. At this point, the first condition has been met for two consecutive time steps. The second condition is that the adjustment capability stabilizes. For example, at time t-1, the total adjustable power of the energy storage unit is 50 MW, while at time t it becomes 49.6 MW, a change rate of 0.8%, lower than the second threshold of 1%. Since both conditions are met, the chain reaction is considered complete, and the simulation terminates. Figure 4 .

[0067] In the second embodiment, the present invention also proposes a wind-solar-storage-charging pre-simulation system based on digital twins, comprising the following modules:

[0068] The first determining module is used to acquire the digital twin model, topological connection relationship and real-time operation data of the wind-solar-storage-charging system; based on the health status index and historical alarm data of each equipment component, determine the initial failure probability of each component, and sample and generate initial failure events from them; combined with the real-time operation data, determine the failure location, failure type and initial disturbance energy of the event, and form an initial failure event vector.

[0069] The calculation module is used to calculate the transient pressure index of the key nodes in the system at each time step of the pre-simulation.

[0070] The second determination module is used to update a state transition matrix representing the propagation of faults among components based on the transient pressure index of each key node; and to calculate the probability distribution of secondary faults occurring within the current step size based on the updated state transition matrix and the system state at the previous time step, and to determine the new faulty component.

[0071] The output module is used to synchronously calculate a multi-dimensional risk vector consisting of renewable energy power cut-off, load loss, and grid topology connectivity reduction; when the termination condition is met, the current chain reaction simulation is terminated; the propagation path of the chain reaction, system state snapshots at each stage, and the multi-dimensional risk vector are output.

[0072] In an optional embodiment, determining the initial failure probability of each component and sampling from it to generate initial failure events includes:

[0073] Using a logistic regression model, the health status index of the equipment components and the historical alarm frequency per unit time are used as inputs to calculate the initial failure probability of each component.

[0074] The Monte Carlo sampling method is used to randomly select a component as the source of the initial failure event based on the calculated probability distribution.

[0075] In an optional embodiment, calculating the node transient pressure index for key nodes in the system at each time step of the pre-simulation includes:

[0076] The transient pressure exponent at node i at time t is calculated using the following formula. :

[0077]

[0078] in, The system frequency deviation rate, Let be the voltage offset at node i. For the SOC balance of energy storage units, and These are the system's rated frequency and rated voltage, respectively.

[0079] In an optional embodiment, updating a state transition matrix representing the propagation of faults among components based on the transient stress index of each critical node includes:

[0080] Calculate the transition probability from source component j to target component i. The calculation formula is:

[0081]

[0082] in, The transient stress index is the number of nodes connected to the target component i. The electrical distance between component j and component i. Let j be the set of potential target components that are electrically connected to source component j, and let j be the normalization factor for all potential target components in the denominator.

[0083] In an optional embodiment, the step of calculating the probability distribution of secondary faults occurring within the current step size based on the updated state transition matrix and the system state at the previous time step, and determining new faulty components, includes:

[0084] Multiply the system fault state vector of the previous time t-1 with the updated state transition matrix to obtain a cumulative effect vector representing the impact of fault propagation on each non-faulty component;

[0085] The cumulative effect vector is normalized to obtain the probability distribution of secondary faults occurring within the current step size;

[0086] The component with the highest probability value in the probability distribution is selected as the new fault component within the current step size.

[0087] In an optional embodiment, the synchronous calculation of a multi-dimensional risk vector consisting of renewable energy power cut-off, load loss, and grid topology connectivity degradation includes:

[0088] The difference between the output of photovoltaic and wind turbines before and after the simulation will be used as the power cut-off of new energy sources.

[0089] The total power loss on the user side caused by the fault is taken as the load power loss.

[0090] The decrease in grid topology connectivity is represented by calculating the change in the maximum eigenvalue of the grid adjacency matrix before and after the simulation.

[0091] In an optional embodiment, terminating the current chain reaction simulation when the termination condition is met includes:

[0092] Normalize each component of the multidimensional risk vector R(t) to obtain a dimensionless normalized risk vector. ;

[0093] Calculate the Euclidean norm of the normalized risk vector ;

[0094] When the growth rate The simulation is terminated when the power is less than the first threshold for two consecutive steps and the total adjustable power change rate of all energy storage units in the system is less than the second threshold.

[0095] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments in this specification are not limited to the described order of actions, because according to the embodiments in this specification, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments in this specification.

[0096] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0097] The preferred embodiments disclosed above are merely illustrative of this specification. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the embodiments described herein. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the embodiments, thereby enabling those skilled in the art to better understand and utilize this specification. This specification is limited only by the claims and their full scope and equivalents.

Claims

1. A wind-solar-storage-charging pre-simulation method based on digital twins, characterized in that, Includes the following steps: Acquire the digital twin model, topological connection relationship and real-time operation data of the wind-solar-storage-charging system; determine the initial failure probability of each component based on the health status index and historical alarm data of each equipment component, and sample and generate initial failure events from them; combine the real-time operation data to determine the failure location, failure type and initial disturbance energy of the event, and form an initial failure event vector. Within each time step of the rehearsal, the transient stress index of the key nodes in the system is calculated. Based on the transient stress index of each key node, update a state transition matrix representing the propagation of faults among components; Based on the updated state transition matrix and the system state at the previous time step, calculate the probability distribution of secondary faults occurring within the current step size and identify new faulty components. Simultaneously calculate a multi-dimensional risk vector consisting of renewable energy power cut-off, load loss, and grid topology connectivity reduction. When the termination condition is met, the chain reaction simulation is terminated; the propagation path of the chain reaction, system state snapshots at each stage, and multidimensional risk vectors are output.

2. The method according to claim 1, characterized in that, The process of determining the initial failure probability of each component and sampling it to generate initial failure events includes: Using a logistic regression model, the health status index of the equipment components and the historical alarm frequency per unit time are used as inputs to calculate the initial failure probability of each component. The Monte Carlo sampling method is used to randomly select a component as the source of the initial failure event based on the calculated probability distribution.

3. The method according to claim 1, characterized in that, Within each time step of the pre-simulation, the transient pressure index of the key nodes in the system is calculated, including: The transient pressure index is weighted by the system frequency deviation rate, the voltage offset of the node, and the SOC balance of the energy storage unit. The transient pressure exponent at node i at time t is calculated using the following formula. : in, The system frequency deviation rate, Let be the voltage offset at node i. For the SOC balance of energy storage units, and These are the system's rated frequency and rated voltage, respectively.

4. The method according to claim 3, characterized in that, The transient stress index of each key node is used to update a state transition matrix representing the propagation of faults among components. include: Among them, the probability of a fault transfer from the source component to the target component is inversely proportional to the electrical distance between them and directly proportional to the transient stress exponent of the node to which the target component is connected; Calculate the transition probability from source component j to target component i. The calculation formula is: in, The transient stress index is the value of the node connected to the target component i. The electrical distance between component j and component i. Let j be the set of potential target components that are electrically connected to source component j, and let j be the normalization factor for all potential target components in the denominator.

5. The method according to claim 1, characterized in that, Based on the updated state transition matrix and the system state at the previous time step, the probability distribution of secondary faults occurring within the current step size is calculated, and new faulty components are identified, including: Multiply the system fault state vector of the previous time t-1 with the updated state transition matrix to obtain a cumulative effect vector representing the impact of fault propagation on each non-faulty component; The cumulative effect vector is normalized to obtain the probability distribution of secondary faults occurring within the current step size; The component with the highest probability value in the probability distribution is selected as the new fault component within the current step size.

6. The method according to claim 1, characterized in that, The synchronous calculation of a multi-dimensional risk vector, consisting of renewable energy power cut-off, load loss, and grid topology connectivity degradation, includes: The difference between the output of photovoltaic and wind turbines before and after the simulation will be used as the power cut-off of new energy sources. The total power loss on the user side caused by the fault is taken as the load power loss. The decrease in grid topology connectivity is represented by calculating the change in the maximum eigenvalue of the grid adjacency matrix before and after the simulation.

7. The method according to any one of claims 1-6, characterized in that, The termination of the current chain reaction simulation when the termination condition is met includes: Normalize each component of the multidimensional risk vector R(t) to obtain a dimensionless normalized risk vector. ; Calculate the Euclidean norm of the normalized risk vector ; When the growth rate The simulation is terminated when the power is less than the first threshold for two consecutive steps and the total adjustable power change rate of all energy storage units in the system is less than the second threshold.

8. A wind-solar-storage-charging pre-simulation system based on digital twins, characterized in that, Includes the following modules: The first determining module is used to acquire the digital twin model, topological connection relationship and real-time operation data of the wind-solar-storage-charging system; based on the health status index and historical alarm data of each equipment component, determine the initial failure probability of each component, and sample and generate initial failure events from them; combined with the real-time operation data, determine the failure location, failure type and initial disturbance energy of the event, and form an initial failure event vector. The calculation module is used to calculate the transient pressure index of the key nodes in the system at each time step of the pre-simulation. The second determination module is used to update a state transition matrix representing the propagation of faults among components based on the transient stress index of each key node. Based on the updated state transition matrix and the system state at the previous time step, calculate the probability distribution of secondary faults occurring within the current step size and identify new faulty components. The output module is used to synchronously calculate a multi-dimensional risk vector consisting of renewable energy power cut-off, load loss, and grid topology connectivity reduction. When the termination condition is met, the chain reaction simulation is terminated; the propagation path of the chain reaction, system state snapshots at each stage, and multidimensional risk vectors are output.

9. The system according to claim 8, characterized in that, The process of determining the initial failure probability of each component and sampling it to generate initial failure events includes: Using a logistic regression model, the health status index of the equipment components and the historical alarm frequency per unit time are used as inputs to calculate the initial failure probability of each component. The Monte Carlo sampling method is used to randomly select a component as the source of the initial failure event based on the calculated probability distribution.

10. A computer-readable storage medium storing a computer program thereon, characterized in that, The computer program, when executed by a processor, implements the method as described in any one of claims 1-7.

Images