Power on-line fault detection method based on source network load storage integration and electronic equipment

Abstract

The invention relates to the technical field of electric power, in particular to an electric power online fault detection method based on source-network-load-storage integration and electronic equipment, and aims to solve the problems that in the prior art, accurate system operation state evaluation cannot be provided, subsequent accurate fault detection and positioning are not facilitated, and the fault detection efficiency is low. And follow-up dynamic adjustment of an energy storage strategy to optimize the system stability is not facilitated. By processing multi-dimensional data, providing accurate system operation state evaluation and utilizing an iterative algorithm to obtain high-precision voltage amplitude and phase angle estimation, the method is suitable for traditional components and energy storage and new energy equipment, faults can be accurately detected and positioned, an energy storage strategy can be dynamically adjusted to optimize the system stability, and the reliability of the system is improved. And the anti-noise capability is enhanced through weighted residual errors, the energy storage electric quantity is updated by selecting a proper time step length, the complex calculation is simplified, the method is particularly suitable for a source network load storage integrated system, and the management efficiency and the operation reliability are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of power technology, and more specifically, to a power online fault detection method and an electronic device based on source-network-load-storage integration. Background Art

[0002] With the continuous development of the power system, source-network-load-storage integration has become an important development direction of the power system. Source-network-load-storage integration can achieve efficient utilization of energy and stable operation of the power system. However, in practical applications, the power system often faces various faults, such as power supply faults, grid faults, load faults, and energy storage faults, etc. If these faults are not detected and processed in time, it may lead to the paralysis of the power system and energy waste.

[0003] The patent application with the publication number CN113820564A discloses a fault detection method suitable for a complex power grid of source-network-load-storage. The principle of this method is simple and reliable, and it is not affected by the transition resistance and the distributed capacitance of the line at the same time, providing technical support for the safe and reliable operation of the distribution network. To implement the above invention, the present invention adopts the following technical solutions: First, the measurement element performs data acquisition and calculates the current differential. It is determined whether the protection starting element is satisfied according to the current variance. When it is greater than the set threshold, it enters the next step, otherwise the protection returns. Calculate the intrinsic mode singular value entropy Gini coefficient and the current integral, and respectively input them into the fault identification and fault pole selection programs. The fault identification program uses the intrinsic mode singular value entropy Gini coefficient to identify the fault type. If the intrinsic mode singular value entropy Gini coefficient is greater than the threshold, it is determined as an in-zone fault, and the protection enters the next step; otherwise, it is determined as an out-of-zone fault, and the protection returns. When the fault identification program and the fault pole selection both meet the conditions, an action signal is sent to the fault pole. However, the above reference patent expresses the fault characteristics through the intrinsic mode singular value Gini coefficient, enhances the fault resistance ability to 700Ω and improves the anti-noise performance, improving the accuracy and reliability of DC distribution network fault detection. However, it cannot provide an accurate assessment of the system operation state, which is not conducive to subsequent accurate detection and location of faults, not conducive to subsequent dynamic adjustment of the energy storage strategy to optimize the system stability, cannot accurately and efficiently identify faults, cannot quickly judge system anomalies and respond in time, and at the same time cannot achieve fast and accurate fault location, reducing the fault response speed and unable to ensure the safe and stable operation of the power system.

[0004] Therefore, we propose a power online fault detection method and an electronic device based on source-network-load-storage integration for the above problems. Summary of the Invention

[0005] The object of the present invention is to provide a power online fault detection method and an electronic device based on source-network-load-storage integration, which solves the problems that the prior art cannot provide an accurate evaluation of the system operation state, is not conducive to subsequent accurate detection and location of faults, is not conducive to subsequent dynamic adjustment of the energy storage strategy to optimize the system stability, cannot accurately and efficiently identify faults, cannot quickly judge system anomalies and respond in a timely manner, and at the same time cannot achieve fast and accurate fault location, reduce the fault response speed, and cannot guarantee the safe and stable operation of the power system.

[0006] The object of the present invention is achieved through the following technical solutions: A power online fault detection method based on source-network-load-storage integration includes the following steps: Step 1: Collect the real-time operation data of the source-network-load-storage integrated power system, and perform preprocessing operations on the collected operation data; Step 2: Based on the collected operation data, perform state estimation on the power system to obtain an accurate estimated value of the system operation state; Step 3: Based on the results of the state estimation, detect whether a fault occurs in the power system; Step 4: Use the state estimation results and the fault detection results for fault location to determine the location where the fault occurs; Step 5: According to the fault detection results and the fault location results, dynamically adjust the operation of the energy storage system.

[0007] As a preferred embodiment of the present invention, the specific process of performing state estimation on the power system in Step 2 is as follows: Obtain the preprocessed operation data, where the operation data includes operation voltage, operation current, operation power, operation frequency, energy storage state, and new energy power generation; Simplify the source-network-load-storage integrated power system into a power network including n nodes, and each node can be connected to a generator, a load, an energy storage device, and a new energy power generation device; Node: Use i to represent the node number; Branch: Use (i,j) to represent the branch connecting node i and node j; Generator: The active power of generator k is , and the reactive power is ; Load: The active power of load k is , and the reactive power is ; Energy storage device: The active power of energy storage device k is ; New energy power generation device: The active power of new energy power generation device k is ; The following measurement data are obtained from the pre - processed operating data: The voltage magnitude measured at certain nodes i and the voltage phase angle measured at certain nodes i and the active power measured in certain branches (i, j) and the reactive power measured in certain branches (i, j) and the measured active power of generator k and the measured reactive power of generator k and the power measurement of energy storage device k and the power measurement of new - energy power generation device k .

[0008] As a preferred embodiment of the present invention, the state variables are represented as a vector x, the state variables include the voltage magnitudes and phase angles of all nodes, the measured values are represented as a vector z, and the measurement equation is z = h(x)+e, where h(x) is a non - linear function of the measured values and e is the measurement error; The objective function of state estimation is: where R is the covariance matrix of the measurement error; An iterative algorithm is used to minimize the objective function, and the iterative formula is: where H is the Jacobian matrix, and its elements are ; The calculation steps for the iterative algorithm to minimize the objective function are as follows: T1: Assign an initial value to the state variable x; T2: According to the current state variable , use the power flow calculation method of the power system to calculate the predicted measured values ; T3: Calculate the Jacobian matrix H; T4: Update the state vector using the iterative formula; T5: If the change in the objective function value or the state variable is less than the set threshold, stop the iteration, otherwise, repeat T2 - T4; The finally obtained x contains the estimated values of the voltage magnitudes and phase angles of all nodes.

[0009] As a preferred embodiment of the present invention, according to the state estimation result, the voltage of each node is represented as a complex number: where is the voltage magnitude, is the voltage phase angle; For the branch connecting node i and node j, its power flow is calculated by the following formula : , where is the impedance of the branch; Calculate the active power on the branch and reactive power : ; , where is the phase angle of; The finally calculated result is the power flow estimation value of the branch; State estimation provides the power output of each energy storage device at the current moment , and a time series of state estimation results is required. The time series of state estimation results is achieved through the following steps: S1: Select a suitable time step ∆t; S2: For each time step, update the energy storage power according to the power output of the energy storage device: If > 0: ; If < 0: ; S3: Repeat S2 until the required moment is calculated; The finally calculated result is the charge and discharge state estimation value of the energy storage device; If the state estimation includes the power measurement value of new energy generation, then this measurement value is the output estimation value of new energy generation.

[0010] As a preferred embodiment of the present invention, the specific process of step three for detecting whether a fault occurs in the power system is as follows: Obtain the power system state estimation value, including: the voltage amplitude and phase angle estimation values of each node , the power flow estimation values of each branch , the charge and discharge state estimation values of the energy storage device ; Obtain the real-time measurement value, including: the voltage amplitude and phase angle measurement values of each node , the power flow measurement values of each branch , the charge and discharge state measurement values of the energy storage device ; Calculate the residuals between the state estimation value and the real-time measurement value. For each type of measurement value, calculate its corresponding residual: , ; , ; 。

[0011] As a preferred embodiment of the present invention, due to the different precisions of different measurement values, it is necessary to perform weighted processing on the residuals. The elements of the weight matrix W are inversely proportional to the variances of the measurement values: , where is the variance of the k-th measurement; The weighted residual is: , where r is the residual vector containing all types of residuals; Calculate the sum of squares of the weighted residuals: , if exceeds the preset threshold , it indicates that a fault has occurred in the power system, where is the significance level.

[0012] As a preferred embodiment of the present invention, the specific process of determining the location where the fault occurs in step four is as follows: Obtain the historical state estimation results and historical fault detection results of the power system, and obtain the state estimation results before the fault therefrom: including all node voltages , branch power flows , and energy storage states , and obtain the state estimation results after the fault therefrom: including the estimated values of node voltages at the same moment after the fault occurs , estimated values of branch power flows , and estimated values of energy storage states ; Calculate the differences between the state estimation values before and after the fault for all nodes: .

[0013] As a preferred embodiment of the present invention, the differences in node voltage changes , differences in branch power flow changes and as well as differences in energy storage state changes for each node are combined to construct a fault risk prediction matrix GFJ. The fault risk prediction matrices GFJ of all nodes are used as the input of the machine learning model, and the risk probability values of faults corresponding to the fault risk prediction matrices GFJ of each node are used as the output of the machine learning model. With the risk probability value of the fault as the prediction target and minimizing the sum of prediction errors of all training data as the training target, the machine learning model is trained until the sum of prediction errors reaches convergence and then the training is stopped to obtain a node fault risk prediction model. The expression formula of the node fault risk prediction model is: ; where λ1, λ2, λ3, and λ4 are all regression coefficients, and η represents the error term; When a fault occurs in the power system is detected, the real-time state estimation result of the power system is immediately obtained. After processing the real-time state estimation result of the power system, it is input into the node fault risk prediction model to obtain the real-time risk probability values of all nodes having faults. The node corresponding to the maximum risk probability value is the location where the fault occurs.

[0014] As a preferred embodiment of the present invention, the specific process of dynamically adjusting the operation of the energy storage system in step five is as follows: Obtain the fault detection result and fault location result of the power system. The fault detection result is whether a fault occurs, and the fault location result is the node location where the fault occurs. Combining the fault detection and location results, a comprehensive strategy of a multi-level protection mechanism and adaptive control is used to dynamically adjust the operation of the energy storage system.

[0015] As a preferred embodiment of the present invention, an electronic device includes a memory, a processor, and a computer program stored on the memory. The processor executes the computer program to implement the steps of the power online fault detection method based on source-network-load-storage integration.

[0016] Compared with the prior art, the advantages of the present invention are as follows: (1) In the present invention, by processing multi-dimensional data, accurate system operation state evaluation is provided. High-precision voltage amplitude and phase angle estimation are obtained by using an iterative algorithm. It is applicable to traditional components, energy storage, and new energy devices. It can accurately detect and locate faults, dynamically adjust the energy storage strategy to optimize system stability, enhance the anti-noise ability through weighted residuals, select an appropriate time step to update the energy storage power, simplify complex calculations, and is particularly suitable for source-network-load-storage integrated systems, significantly improving management efficiency and operation reliability; (2) In the present invention, faults are identified by comparing the state estimation value with the real-time measurement value, which has the characteristics of high accuracy and strong adaptability. Weighted residual analysis is performed on various measurement values to effectively process measurement errors and model uncertainties. The chi-square distribution is used to set thresholds, which can quickly judge system anomalies and respond in a timely manner, ensuring the safe and stable operation of the power system and providing strong support for the reliable operation of the system; (3) In the present invention, a fault risk prediction matrix GFJ is constructed by calculating the changes in node voltage, branch power flow, and energy storage state before and after a fault to accurately locate the fault. The model is trained using historical data and the real-time state estimation result is used to update the model when a fault is detected, achieving fast and accurate fault location. The effectiveness of the model is ensured by minimizing the prediction error and convergence criteria, improving the fault response speed, helping to prioritize the processing of high-risk areas, and ensuring the safe and stable operation of the power system. Description of the Drawings

[0017] Figure 1 This is the flowchart of the power online fault detection method based on source-network-load-storage integration in the present invention; Figure 2 This is the flowchart of the calculation steps for minimizing the objective function of the iterative algorithm in the present invention; Figure 3 This is the flowchart of the steps for calculating the charge-discharge state estimation value of the energy storage device in the present invention; Figure 4 This is the schematic structural diagram of an electronic device provided in Embodiment III of the present invention. Detailed implementation manners

[0018] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0019] Embodiment I: As Figure 1 , Figure 2 and Figure 3 shown, the power online fault detection method based on source-network-load-storage integration proposed by the present invention includes the following steps: Step 1: Collect the real-time operation data of the source-network-load-storage integrated power system, and perform preprocessing operations on the collected operation data. The preprocessing operations include but are not limited to data cleaning, filtering, and normalization; Through data cleaning, filtering, and normalization, high-quality and reliable preprocessed data can be obtained, which will provide a solid foundation for subsequent analysis, modeling, and control, and ultimately improve the operation efficiency, stability, and reliability of the source-network-load-storage integrated power system.

[0020] Step 2: Based on the collected operation data, perform state estimation on the power system to obtain an accurate estimation value of the system operation state; The specific process of performing state estimation on the power system in Step 2 is as follows: Obtain the preprocessed operation data. The operation data includes operation voltage, operation current, operation power, operation frequency, energy storage state, and new energy generation power; Simplify the source-network-load-storage integrated power system into a power network including n nodes, and each node can be connected to a generator, a load, an energy storage device, and a new energy generation device; Node: Use i to represent the node number; Branch: Use (i, j) to represent the branch connecting node i and node j; Generator: The active power of generator k is , and the reactive power is ; Load: The active power of load k is , and the reactive power is ; Energy storage device: The active power of energy storage device k is ; New energy power generation device: The active power of new energy power generation device k is ; The following measurement data are obtained from the preprocessed operation data: The voltage amplitude measured at certain nodes i , the voltage phase angle measured at certain nodes i , the active power measured in certain branches (i, j) , the reactive power measured in certain branches (i, j) , the measured value of the active power of generator k , the measured value of the reactive power of generator k , the measured value of the power of energy storage device k , the measured value of the power of new energy power generation device k ; The state variables are represented as vector x, where the state variables include the voltage amplitudes and phase angles of all nodes. The measured values are represented as vector z, and the measurement equation is z = h(x) + e, where h(x) is a non-linear function of the measured values and e is the measurement error; The objective function of state estimation is: , where R is the covariance matrix of the measurement error; An iterative algorithm is used to minimize the objective function, and the iterative formula is: , where H is the Jacobian matrix, and its elements are ; The calculation steps for the iterative algorithm to minimize the objective function are as follows: T1: Assign an initial value to the state variable x; T2: According to the current state variable , use the power flow calculation method of the power system to calculate the predicted measured value ; T3: Calculate the Jacobian matrix H; T4: Update the state vector using the iterative formula; T5: If the change in the objective function value or the state variable is less than the set threshold, stop the iteration; otherwise, repeat T2 - T4; The finally obtained x contains the estimated values of the voltage amplitudes and phase angles of all nodes; According to the state estimation results, the voltage of each node is represented as a complex number: , where is the voltage amplitude, is the voltage phase angle; For the branch connecting node i and node j, its power flow is calculated by the following formula : , where is the impedance of the branch; Calculate the active power and reactive power on the branch: ; , where is phase angle; The finally calculated result is the power flow estimation value of the branch; State estimation provides the power output of each energy storage device at the current moment (a positive value indicates discharging, and a negative value indicates charging). A time series of state estimation results is required. The time series of state estimation results is achieved through the following steps: S1: Select a suitable time step ∆t (e.g., 1 second or shorter); S2: For each time step, update the energy storage power according to the power output of the energy storage device: If > 0: ; If < 0: ; S3: Repeat S2 until the required moment is calculated; The finally calculated result is the charge and discharge state estimation value of the energy storage device; If the power measurement value of new energy generation is included in the state estimation, then this measurement value is the output estimation value of new energy generation; By processing multi-dimensional data, it provides an accurate assessment of the system operating state, obtains high-precision voltage amplitude and phase angle estimates using an iterative algorithm, is applicable to traditional components and energy storage and new energy devices, can accurately detect and locate faults, dynamically adjust the energy storage strategy to optimize system stability, enhance the anti-noise ability through weighted residuals, select a suitable time step to update the energy storage power, simplify complex calculations, is particularly suitable for source-grid-load-storage integrated systems, and significantly improves management efficiency and operation reliability.

[0021] Step 3: Based on the results of state estimation, detect whether a fault has occurred in the power system; The specific process of step three for detecting whether a fault occurs in the power system is as follows: Obtain the power system state estimation values, including: the voltage magnitude and phase angle estimation values of each node , and the power flow estimation values of each branch , and the charge and discharge state estimation value of the energy storage device ; Obtain the real-time measurement values, including: the voltage magnitude and phase angle measurement values of each node , and the power flow measurement values of each branch , and the charge and discharge state measurement value of the energy storage device ; Calculate the residuals between the state estimation values and the real-time measurement values. For each type of measurement value, calculate its corresponding residual: (voltage magnitude residual), (voltage phase angle residual); (active power residual), (reactive power residual); (energy storage state residual); Since the accuracies of different measurement values are different, it is necessary to perform weighted processing on the residuals. The elements of the weight matrix W are inversely proportional to the variances of the measurement values: , where is the variance of the kth measurement; The weighted residual is: , where r is the residual vector containing all types of residuals; Calculate the sum of the squares of the weighted residuals: , if exceeds the preset threshold , it indicates that a fault has occurred in the power system, where is the significance level, and the threshold can be determined by looking up the chi-square distribution table; By comparing the state estimation values with the real-time measurement values to identify faults, it has the characteristics of high accuracy and strong adaptability. By performing weighted residual analysis on multiple measurement values, it can effectively handle measurement errors and model uncertainties. By setting the threshold using the chi-square distribution, it can quickly judge system anomalies and respond in a timely manner, ensuring the safe and stable operation of the power system and providing strong support for the reliable operation of the system.

[0022] Step four: Use the state estimation results and fault detection results for fault location to determine the location where the fault occurs; The specific process of step four for determining the location where the fault occurs is as follows: Obtain the historical state estimation results and historical fault detection results of the power system, and obtain the state estimation results before the fault from them: including all node voltages Branch power flow and energy storage state to obtain the post-fault state estimation result: including the node voltage estimation value at the same moment after the fault occurs branch power flow estimation value and energy storage state estimation value ; Calculate the difference between the pre-fault and post-fault state estimation values of all nodes. These differences reflect the system state changes caused by the fault: ; Combine the node voltage change difference branch power flow change difference and as well as the energy storage state change difference of each node to construct the fault risk prediction matrix GFJ. Use the fault risk prediction matrix GFJ of all nodes as the input of the machine learning model, and use the risk probability value of the fault occurrence corresponding to the fault risk prediction matrix GFJ of each node as the output of the machine learning model. Take the risk probability value of the fault occurrence as the prediction target, and take minimizing the sum of the prediction errors of all training data as the training target to train the machine learning model until the sum of the prediction errors reaches convergence and then stop training to obtain the node fault risk prediction model. The expression formula of the node fault risk prediction model is: ; where λ1, λ2, λ3, and λ4 are all regression coefficients, and η represents the error term; When a fault occurs in the power system is detected, immediately obtain the real-time state estimation result of the power system, process the real-time state estimation result of the power system and then input it into the node fault risk prediction model to obtain the real-time risk probability values of all nodes for the fault occurrence. The node corresponding to the maximum risk probability value is the location where the fault occurs; Construct the fault risk prediction matrix GFJ by calculating the changes in node voltage, branch power flow, and energy storage state before and after the fault, accurately locate the fault, train the model using historical data, and update the model using the real-time state estimation result when a fault is detected, realizing fast and accurate fault location. Use the minimization of prediction error and convergence criterion to ensure the effectiveness of the model, improve the fault response speed, help prioritize the processing of high-risk areas, and ensure the safe and stable operation of the power system.

[0023] Embodiment 2: The technical solution of this embodiment of the present invention is different from that of Embodiment 1 in that: As Figure 1As shown in the figure, Step 5: Dynamically adjust the operation of the energy storage system according to the fault detection result and the fault location result; The specific process of dynamically adjusting the operation of the energy storage system in Step 5 is as follows: Obtain the fault detection result and the fault location result of the power system. The fault detection result is whether a fault occurs, and the fault location result is the node position where the fault occurs. Combining the fault detection and location results, adopt a comprehensive strategy of a multi-level protection mechanism and adaptive control to dynamically adjust the operation of the energy storage system. The specific content of the comprehensive strategy is: The first layer: Emergency protection Objective: Immediately suppress the fault expansion and prevent the system from crashing; Measures: When any abnormal situation (such as abnormal battery pack voltage, inverter overload, overhigh temperature, etc.) is detected, immediately stop the charge and discharge operations of all energy storage units; If the system has a hardware isolation function, immediately isolate the faulty unit or module; Send an emergency shutdown signal to the upper control system and other parts of the power system, such as the microgrid control system or the power dispatching center; The second layer: Fault isolation and local control Objective: Maximize the operating capacity of healthy units while ensuring system safety; Measures: Accurately identify the faulty unit or module according to the fault location result; Isolate the faulty unit and monitor the surrounding units to prevent the fault from spreading; According to the power of the faulty unit, redistribute the corresponding power to healthy energy storage units to maintain the system output power; Adjust the voltage or frequency control parameters according to the fault location and type to stabilize the local power grid; The third layer: Adaptive control and optimization strategy Objective: Optimize the operation strategy of the energy storage system after the fault recovery to improve the system efficiency and reliability; Measures: Conduct a comprehensive state assessment of the energy storage system and the power system, including the health status of the battery pack, environmental temperature, power system load, etc.; According to the state assessment result, adjust the charge and discharge strategy of the energy storage system, such as adjusting the charge and discharge power curve, optimizing the power distribution algorithm, adjusting the voltage / frequency control parameters, etc.; According to the fault history data and the state assessment result, predict potential faults and formulate a preventive maintenance plan to avoid future faults.

[0024] Adopt machine learning or artificial intelligence algorithms to analyze the operation data of the energy storage system, continuously optimize the control strategy, and improve the system efficiency and reliability; Dynamically adjust the energy storage system through multi-level protection and adaptive control strategies. The emergency protection layer quickly isolates faults to prevent their spread; the fault isolation layer precisely processes faulty units and maintains system stability; the adaptive control layer optimizes strategies, predicts potential faults and formulates prevention plans, and continuously optimizes using machine learning to enhance system flexibility and response speed, ensuring the reliable operation of the power system.

[0025] Example 3: Please refer to Figure 4 As shown, this example publicly provides an electronic device, including a memory, a processor, and a computer program stored on the memory. The processor executes the computer program to implement the steps of the power online fault detection method based on source-grid-load-storage integration. Since the electronic device introduced in this example is the electronic device used to implement the power online fault detection method based on source-grid-load-storage integration in Example 1 of the present application, based on the power online fault detection method based on source-grid-load-storage integration introduced in the embodiments of the present application, those skilled in the art can understand the specific implementation manners and various variations of the electronic device in this example. Therefore, the specific implementation of how this electronic device implements the method in the embodiments of the present application will not be described in detail here; as long as those skilled in the art implement the electronic device used in the power online fault detection method based on source-grid-load-storage integration in the embodiments of the present application, it falls within the scope of protection of the present application.

[0026] The above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its improved conceptions, makes equivalent substitutions or changes, and all should be covered by the protection scope of the present invention.

Claims

1. An online power fault detection method based on source-network-load-storage integration, characterized in that It includes the following steps: Step 1: Collect the real-time operation data of the source-network-load-storage integrated power system, and perform preprocessing operations on the collected operation data; Step 2: Based on the collected operation data, perform state estimation on the power system to obtain accurate estimated values of the system operation state; Step 3: Based on the results of state estimation, detect whether a fault occurs in the power system; Step 4: Use the state estimation results and fault detection results to locate the fault and determine the location where the fault occurs; Step 5: According to the fault detection results and fault location results, dynamically adjust the operation of the energy storage system.

2. The power online fault detection method based on source-network-load-storage integration according to claim 1, wherein The specific process of performing state estimation on the power system in Step 2 is as follows: Obtain the preprocessed operation data, where the operation data includes operation voltage, operation current, operation power, operation frequency, energy storage state, and new energy generation power; Simplify the source-network-load-storage integrated power system into a power network with n nodes, and each node can be connected to generators, loads, energy storage devices, and new energy generation devices; Node: Use i to represent the node number; Branch: Use (i,j) to represent the branch connecting node i and node j; Generator: The active power of generator k is , and the reactive power is ; Load: The active power of load k is , and the reactive power is ; Energy storage device: The active power of the energy storage device k is ; New energy power generation equipment: The active power of new energy power generation equipment k is ; Obtain the following measurement data from the preprocessed operation data: The voltage magnitude measured at certain nodes i , the voltage phase angle measured at certain nodes i , the active power measured in certain branches (i, j) , the reactive power measured in certain branches (i, j) , the measured active power of generator k , the measured reactive power of generator k , the power measurement of energy storage device k , the power measurement of new energy generation device k .

3. The power online fault detection method based on source-network-load-storage integration according to claim 2, wherein, Represent the state variables as a vector x, where the state variables include the voltage amplitudes and phase angles of all nodes, represent the measured values as a vector z, and the measurement equation is z = h(x) + e, where h(x) is a non-linear function of the measured values and e is the measurement error; The objective function of state estimation is: , where R is the covariance matrix of the measurement error; Use an iterative algorithm to minimize the objective function, and the iterative formula is: , where H is the Jacobian matrix, the elements of which are ; The calculation steps for the iterative algorithm to minimize the objective function are as follows: T1: Assign an initial value to the state variable x; T2: Calculate the predicted measured values using the power flow calculation method of the power system according to the current state variables ; ​ T3: Calculate the Jacobian matrix H; T4: Update the state vector using the iterative formula; T5: If the change in the objective function value or the state variable is less than the set threshold, stop the iteration, otherwise, repeat T2 - T4; The finally obtained x contains the estimated values of the voltage amplitudes and phase angles of all nodes.

4. The power online fault detection method based on source-network-load-storage integration according to claim 3, wherein According to the state estimation results, the voltage of each node is represented as a complex number: , where is the voltage amplitude, is the voltage phase angle; For the branch connecting node i and node j, its power flow is calculated by the following formula :[[]]END]] , where is the impedance of the branch; Calculate the active power and reactive power : ； , where is phase angle of The finally calculated result is the estimated value of the branch power flow; State estimation provides the power output of each energy storage device at the current moment , and a time series of state estimation results is required. The time series of state estimation results is achieved through the following steps: S1: Select a suitable time step ∆t; S2: For each time step, update the stored energy in the energy storage device according to the power output of the energy storage device: If > 0: ; If <0: ; S3: Repeat S2 until the required time is calculated; The finally calculated result is the estimated value of the charge and discharge state of the energy storage device; If the state estimation includes the power measurement value of new energy generation, then this measurement value is the estimated output of new energy generation.

5. The power online fault detection method based on source-network-load-storage integration according to claim 1, wherein The specific process of detecting whether a fault occurs in the power system in Step 3 is as follows: Obtain the power system state estimation values, including: the voltage magnitude and phase angle estimation values of each node , the power flow estimation values of each branch , the charge and discharge state estimation values of energy storage devices ; Obtain real-time measurement values, including: voltage amplitude and phase angle measurement values of each node , power flow measurement values of each branch , charge and discharge state measurement values of energy storage devices ; Calculate the residuals between the state estimation values and the real-time measurement values, and for each type of measurement value, calculate its corresponding residual: ， ； ， ； 。 6. The power online fault detection method based on source-network-load-storage integration according to claim 5, wherein, Due to the different precisions of different measurement values, it is necessary to perform weighted processing on the residuals. The elements of the weight matrix W are inversely proportional to the variances of the measurement values: , where is the variance of the k-th measurement; The weighted residual is: , where r is the residual vector, containing all types of residuals; Calculate the sum of squares of weighted residuals: , if exceeds a preset threshold , it indicates that a fault has occurred in the power system, where is the significance level.

7. The power online fault detection method based on source-network-load-storage integration according to claim 1, wherein The specific process of determining the location where the fault occurs in Step 4 is as follows: Obtain the historical state estimation results and historical fault detection results of the power system, and obtain the pre-fault state estimation results therefrom: including all node voltages , branch power flows , and energy storage states , and obtain the post-fault state estimation results therefrom: including the estimated values of node voltages at the same moment after the fault occurs , estimated values of branch power flows , and estimated values of energy storage states ; Calculate the difference in the state estimation values of all nodes before and after the fault: 。 8. The power online fault detection method based on source-network-load-storage integration according to claim 7, characterized in that The difference in the change of the node voltage of each node , the difference in the change of the branch power flow and as well as the difference in the change of the energy storage state are combined to construct a fault risk prediction matrix GFJ. The fault risk prediction matrix GFJ of all nodes is used as the input of the machine learning model, and the risk probability value of the occurrence of a fault corresponding to the fault risk prediction matrix GFJ of each node is used as the output of the machine learning model. Taking the risk probability value of the occurrence of a fault as the prediction target and minimizing the sum of the prediction errors of all training data as the training target, the machine learning model is trained until the sum of the prediction errors reaches convergence and then the training is stopped to obtain a node fault risk prediction model. The expression formula of the node fault risk prediction model is: ； where λ1, λ2, λ3, and λ4 are all regression coefficients, and η represents the error term; When it is detected that a fault occurs in the power system, immediately obtain the real-time state estimation results of the power system, process the real-time state estimation results of the power system and input them into the node fault risk prediction model to obtain the real-time risk probability values of all nodes having a fault, and the node corresponding to the maximum risk probability value is the location where the fault occurs.

9. The power online fault detection method based on source-network-load-storage integration according to claim 1, wherein, The specific process of dynamically adjusting the operation of the energy storage system in Step Five is as follows: Obtain the fault detection result and fault location result of the power system. The fault detection result indicates whether a fault has occurred, and the fault location result indicates the node location where the fault occurred. Combining the fault detection and location results, adopt a comprehensive strategy of a multi-level protection mechanism and adaptive control to dynamically adjust the operation of the energy storage system.

10. An electronic device, comprising a memory, a processor, and a computer program stored on the memory, characterized in that, The processor executes the computer program to implement the steps of the power online fault detection method based on source-network-load-storage integration according to any one of claims 1-9.

Images