Distribution line short circuit self-adaptive protection method and distribution line short circuit self-adaptive protection system

Abstract

The invention relates to a distribution line short-circuit self-adaptive protection method and system, and belongs to the technical field of power system protection. Comprising the steps that current signals at the two ends of a distribution line are collected, and the current signals comprise a first-end current signal and a second-end current signal; decomposing a first end transient component from the first end current signal, and decomposing a second end transient component from the second end current signal; calculating a transient energy function according to the first end transient component and the second end transient component; according to the transient energy function, real-time electric signal data and historical fault data of the distribution line are obtained, and the fault probability of the distribution line is predicted; obtaining a fault processing decision according to the transient energy function and the fault probability; outputting a control signal according to the fault processing decision; wherein before the control signal is output, time synchronization error correction is carried out on the two ends of the distribution line, and an encrypted communication tunnel is established. According to the invention, rapid detection, accurate isolation and intelligent recovery of the short-circuit fault of the distribution line are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system protection, and in particular to a short-circuit adaptive protection method and system for distribution lines. Background Art

[0002] With the continuous expansion of the scale of the power system and the rapid development of new energy, the short-circuit protection of distribution lines faces new challenges. Traditional distribution line short-circuit protection systems mainly rely on the monitoring of physical signals such as continuous current and voltage in the line, and achieve fault judgment by presetting fixed values.

[0003] However, in the context of the large-scale access of new energy and the wide application of distributed power sources, fault signals present new characteristics such as dynamics, non-linearity, and high noise interference. These changes make it difficult for traditional protection schemes to accurately capture fault transient information, resulting in a significant reduction in the accuracy and reliability of protection actions. Specifically, the output of new energy power generation is volatile and intermittent, which makes the characteristics of fault current such as amplitude and phase significantly different from traditional fault modes. Traditional protection schemes are difficult to accurately identify these new fault characteristics, resulting in delays or failures of protection actions. The access points of distributed power sources are scattered, the propagation path of fault signals is complex, and it is easy to be affected by noise interference. Traditional protection schemes are prone to misoperation or refusal to operate in this complex environment, further reducing the reliability of the protection system.

[0004] In addition, the communication methods between traditional protection devices also expose obvious deficiencies. These devices mostly use wired or wireless communication methods, and these communication means are prone to eavesdropping or malicious tampering, posing security risks and possibly affecting the overall stability of the protection system. Summary of the Invention

[0005] Therefore, the technical problem to be solved by the present invention is to overcome the difficulty in accurately capturing fault transient information in the prior art, resulting in a decrease in the accuracy and reliability of protection actions, and slow detection speed.

[0006] In a first aspect, to solve the above technical problem, the present invention provides a short-circuit adaptive protection method for distribution lines, including:

[0007] S1. Collect current signals at both ends of the distribution line, where the current signals include a first-end current signal and a second-end current signal;

[0008] S2. Decompose a first-end transient component from the first-end current signal and a second-end transient component from the second-end current signal;

[0009] S3. Calculate a transient energy function according to the first-end transient component and the second-end transient component;

[0010] S4. According to the transient energy function, obtain the real-time electrical signal data and historical fault data of the distribution line, and predict the fault probability of the distribution line;

[0011] S5. According to the transient energy function and the fault probability, obtain a fault handling decision; according to the fault handling decision, output a control signal;

[0012] Wherein, before outputting the control signal, it includes correcting the time synchronization error at both ends of the distribution line and establishing an encrypted communication tunnel.

[0013] In an embodiment of the present invention, while S4 is being executed, it includes calculating the statistical quantities of the transient energy function within a preset sliding window to obtain a mean function and a standard deviation function; setting an adaptive threshold according to the mean function and the standard deviation function; when the transient energy function is not less than the adaptive threshold, triggering a protection action.

[0014] In an embodiment of the present invention, for calculating the statistical quantities of the transient energy function within a preset sliding window to obtain a mean function and a standard deviation function, the expression of the mean function is:

[0015]

[0016] The expression of the standard deviation function is:

[0017]

[0018] Wherein, μ E (t) is the mean function, T w is the time length of the preset sliding window, ΔE(τ) is the transient energy function, σ E (t) is the standard deviation function, t is time, and τ is the integration time.

[0019] In an embodiment of the present invention, the steps of S2 for decomposing the first-end current signal into a first-end transient component and decomposing the second-end current signal into a second-end transient component are:

[0020] Select a wavelet basis suitable for short-time fault signal feature extraction and set the decomposition level;

[0021] Use statistical distribution to calculate the pre-collected signal samples during normal operation and faults to determine the threshold for each layer of wavelet decomposition;

[0022] According to the threshold for each layer of wavelet decomposition, perform wavelet decomposition on the first-end current signal and the second-end current signal respectively, and reconstruct each resolution component to obtain the first-end transient component and the second-end transient component.

[0023] In one embodiment of the present invention, in step S4, the steps of predicting the fault probability of the distribution line are as follows:

[0024] Preprocess the real-time electrical signal data, the historical fault data, and the transient energy function to obtain preprocessed data;

[0025] Input the preprocessed data into a fault prediction model, and output the fault probability;

[0026] Among them, the fault prediction model adopts a convolutional neural long short-term memory network. At least two convolutional or LSTM layers are set in the convolutional neural long short-term memory network, followed by a fully connected layer. The ReLU activation function and Dropout technology are used to prevent overfitting. During the training process of the convolutional neural long short-term memory network, cross-validation and dynamic learning rate adjustment strategies are used to reduce the prediction error, and the parameters of the fault prediction model are updated online to realize real-time fault trend prediction.

[0027] In one embodiment of the present invention, in step S5, the steps of obtaining a fault handling decision and generating an adaptive adjustment of protection parameters according to the transient energy function and the fault probability are as follows:

[0028] Establish the transient energy function and the fault probability into an input variable set;

[0029] Set a corresponding membership function for each item in the input variable set;

[0030] Construct a fuzzy rule base, use a fuzzy inference mechanism to synthesize the output of the fuzzy rule base, and after defuzzification, obtain a fault handling decision;

[0031] Introduce a self-learning closed-loop mechanism, use the feedback after actually executing the fault handling decision as training data, and adjust the parameters of the membership function and the rule weights through incremental learning.

[0032] In one embodiment of the present invention, the steps of correcting the time synchronization error at both ends of the distribution line are as follows:

[0033] Determine the weight of each time synchronization error, and the expression of the weight is:

[0034]

[0035] According to the weight and the set initial estimated time, continuously iterate using the iteration formula until θ k+1 -θ k |<δ, and then obtain the final time offset; the expression of the iteration formula is:

[0036]

[0037] The control devices at both ends of the distribution line adjust their node clocks according to the time offset;

[0038] where w i is the weight, is the noise variance of the i-th time synchronization error, ∈ is a small positive value to prevent division by zero, δ is a preset value, θ k is the estimated value of the current calculation period, N is the total number of transmitted entangled photon pairs, Δt i is the time interval.

[0039] In an embodiment of the present invention, the step of establishing an encrypted communication tunnel is: generating a shared key, and using the shared key to establish an encrypted communication tunnel; using the encrypted communication tunnel to perform real-time encrypted transmission of the control signal.

[0040] In a second aspect, to solve the above technical problems, the present invention provides a short-circuit adaptive protection system for a distribution line, including:

[0041] A first-end acquisition module for acquiring the current signal at the first end of the distribution line; a second-end acquisition module for acquiring the current signal at the second end of the distribution line;

[0042] A first-end decomposition module for decomposing a first-end transient component from the first-end current signal; a second-end decomposition module for decomposing a second-end transient component from the second-end current signal;

[0043] A calculation module for calculating a transient energy function according to the first-end transient component and the second-end transient component;

[0044] A prediction module for predicting the fault probability of the distribution line according to the transient energy function and acquiring the real-time electrical signal data and historical fault data of the distribution line;

[0045] An output module for obtaining a fault handling decision according to the transient energy function and the fault probability; and outputting a control signal according to the fault handling decision;

[0046] Wherein, before outputting the control signal, it includes correcting the time synchronization error at both ends of the distribution line and establishing an encrypted communication tunnel.

[0047] In an embodiment of the present invention, it further includes a quantum communication device, and the quantum communication device includes a transmission module; the transmission module includes:

[0048] A bit sequence generation sub-module for randomly generating a bit sequence and selecting a polarization basis;

[0049] An original key sequence generation sub-module is used to exchange information about the polarization bases over a public channel, retain the bits that match the polarization bases from the information, and form an original key sequence;

[0050] An error correction sub-module is used to correct the original key sequence until the error rate is lower than a preset threshold to obtain a standard key sequence;

[0051] A key output sub-module performs privacy amplification on the standard key sequence and compresses it through a hash function to obtain a shared key.

[0052] In an embodiment of the present invention, it further includes a central quantum server for correcting the time synchronization error between protection devices at both ends of a distribution line; the central quantum server includes an entangled photon pair generation module and a time synchronization calculation module; both protection devices at both ends of the distribution line include time recording modules; wherein,

[0053] The entangled photon pair generation module is used to generate quantum entangled photon pairs and transmit the quantum entangled photon pairs;

[0054] The time recording module is used to receive the quantum entangled photon pairs and record the reception time;

[0055] The time synchronization calculation module is used to calculate the time synchronization error according to the reception time.

[0056] The above technical solution of the present invention has the following beneficial effects compared with the prior art:

[0057] (1) For a distribution line short-circuit adaptive protection method and system of the present invention, by extracting transient components from current signals, subtle changes occurring during a fault can be accurately captured. Through the analysis of transient components, fault characteristics can be identified earlier, enabling rapid detection and accurate isolation, and significantly improving the sensitivity of fault detection. Calculating the transient energy function based on transient components can quantify fault characteristics and intuitively reflect the severity and location information of the fault. Compared with traditional single current amplitude or frequency analysis, this method can more comprehensively evaluate the fault state and provide a more accurate basis for subsequent fault handling. By predicting the fault probability, early warning signals can be sent in advance to remind operation and maintenance personnel to take preventive maintenance measures, reduce the impact of sudden faults on the operation of the power grid, reduce the risk of equipment damage, and improve power supply reliability. In addition, the present invention corrects the time synchronization error at both ends of the distribution line, which can eliminate misjudgments caused by measurement errors or communication delays, thereby improving the accuracy of fault handling. At the same time, by establishing an encrypted communication tunnel, the security of data transmission is guaranteed, preventing data from being stolen or tampered with during transmission.

[0058] (2) The dynamic threshold algorithm of the present invention based on multi-resolution analysis and sliding window statistics can adjust the protection setting value according to real-time data, reduce the misoperation rate and shorten the response time, thereby improving the accuracy of adaptive protection.

[0059] (3) The present invention uses quantum entanglement states to achieve high-precision clock synchronization, adopts quantum key distribution to ensure data communication security, combines a fault detection method based on transient energy and an adaptive algorithm, and intelligent decision-making combining deep learning and fuzzy logic to achieve rapid detection, accurate isolation and intelligent recovery of short-circuit faults in distribution lines. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to make the content of the present invention easier to be clearly understood, the following further details the present invention according to specific embodiments of the present invention in conjunction with the drawings, wherein:

[0061] Figure 1 It is a flowchart of a method for adaptive protection against short circuits in a distribution line in a preferred embodiment of the present invention;

[0062] Figure 2 It is a flowchart of a method for obtaining transient components in a preferred embodiment of the present invention;

[0063] Figure 3 It is a flowchart of a method for obtaining a fault handling decision and generating an adaptive adjustment of protection parameters in a preferred embodiment of the present invention;

[0064] Figure 4 It is a structural diagram of a system for adaptive protection against short circuits in a distribution line in a preferred embodiment of the present invention.

[0065] Description of the reference numerals in the drawings: 1. First-end protection device; 2. Second-end protection device; 3. Central quantum server; 4. Circuit breaker; 5. Quantum entanglement photon pair distribution channel; 6. Optical fiber; 7. Distribution line; 8. Busbar. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0066] The following further describes the present invention in conjunction with the drawings and specific embodiments, so that those skilled in the art can better understand the present invention and implement it, but the embodiments cited do not limit the present invention.

[0067] Referring to Figure 1 As shown, an embodiment of the present invention provides a method for adaptive protection against short circuits in a distribution line, including but not limited to the following steps:

[0068] S1. Collect current signals at both ends of the distribution line, where the current signals include a first-end current signal and a second-end current signal;

[0069] S2. Decompose the first - end current signal into a first - end transient component and decompose the second - end current signal into a second - end transient component;

[0070] S3. Calculate the transient energy function based on the first - end transient component and the second - end transient component;

[0071] S4. Predict the fault probability of the distribution line according to the transient energy function, and obtain the real - time electrical signal data and historical fault data of the distribution line;

[0072] S5. Obtain a fault handling decision according to the transient energy function and the fault probability; Output a control signal according to the fault handling decision;

[0073] Among them, before outputting the control signal, it includes time - synchronization error correction for both ends of the distribution line and establishing an encrypted communication tunnel.

[0074] An embodiment of the present invention provides a method for self - adaptive protection against short - circuits in a distribution line. By decomposing the transient component from the current signal, it is possible to capture the subtle changes more precisely when a fault occurs. The transient component usually contains high - frequency information at the moment of the fault, which is easily overlooked in conventional steady - state analysis. By analyzing the transient component, the fault characteristics can be detected earlier, thereby improving the sensitivity of fault detection to achieve rapid detection and accurate isolation. Calculate the transient energy function based on the transient component to further quantify the fault characteristics. The transient energy function can more intuitively reflect the severity and location information of the fault. Compared with single - current amplitude or frequency analysis, this method can more comprehensively evaluate the fault state. At the same time, when predicting the fault probability, not only the real - time transient energy function is considered, but also the historical fault data of the distribution line is combined. This comprehensive analysis method can fully consider the long - term operating state and potential hidden dangers of the line, thereby more accurately predicting the probability of a fault occurring. By predicting the fault probability, a warning signal can be sent in advance to remind the operation and maintenance personnel to take preventive maintenance measures. This helps to reduce the impact of sudden faults on the operation of the power grid, reduce the risk of equipment damage, and improve the power supply reliability. Making a fault - handling decision based on the transient energy function and the fault probability can more scientifically evaluate the urgency and handling priority of the fault. For example, for a severe fault with a high probability, isolation measures can be taken preferentially; while for a minor fault with a low probability, an observation or delayed - handling strategy can be adopted. Before outputting the control signal, correct the time - synchronization error at both ends of the distribution line to ensure the accuracy and reliability of the fault - handling decision. The time - synchronization error correction can eliminate misjudgments caused by measurement errors or communication delays, thereby improving the accuracy of fault handling. By establishing an encrypted communication tunnel, the security of data transmission is guaranteed, preventing data from being stolen or tampered with during transmission. This is crucial for protecting the privacy and integrity of power - grid operation data, especially when it comes to fault handling and control - signal transmission. The embodiment of the present invention reduces misjudgments and misoperations caused by single factors through comprehensive analysis of the transient energy function, fault probability, and time - synchronization error correction. This helps to improve the overall stability of the power system and avoid secondary faults caused by improper fault handling. This method not only improves the accuracy and prediction ability of fault detection, but also optimizes the fault - handling decision, enhancing the operation stability and intelligent level of the power system.

[0075] Specifically, for step S1, the first - end current signal i A (t) and the second - end current signal i B (t) at both ends of the distribution line are collected respectively. Among them, the distribution line refers to the line extending from the busbar, which is used to distribute electric power to different circuit breakers. The circuit breaker refers to a device used to control the flow of electric power and protect the circuit from overload or short - circuit.

[0076] Specifically, for step S2, discrete wavelet transform (DWT) and digital filter techniques are preferably used to process the current signal collected in step S1. Through these advanced signal processing methods, the current signal is effectively decomposed into two main components: the steady-state component and the transient component. Among them, the steady-state component represents the relatively stable and continuous part of the current signal, while the transient component captures the short-term and sudden parts of the signal. The mathematical relationships of the current signals, steady-state components, and transient components at both ends of the distribution line are as follows:

[0077] Δi A (t) = i A (t) - i Asteady ;

[0078] Δi B (t) = i B (t) - i Bsteady ;

[0079] Among them, Δi A (t) is the first-end transient component decomposed from the first-end current signal i A (t), and i Asteady is the first-end steady-state component decomposed from the first-end current signal i A (t); Δi B (t) is the second-end transient component decomposed from the second-end current signal i B (t), and i Bsteady is the second-end steady-state component decomposed from the second-end current signal i B (t).

[0080] Furthermore, referring to Figure 2 , the specific steps for decomposing the first-end transient component from the first-end current signal and the second-end transient component from the second-end current signal are as follows:

[0081] S210. To effectively extract the characteristics of short-time fault signals, suitable discrete wavelet bases (wavelet bases can also be called wavelet transform basis functions) are selected, such as Daubechies 4 or Symlet 8, which perform well in processing short-time signals. At the same time, an appropriate decomposition level L is set, such as L = 4, L = 5, L = 6, etc., to ensure that the high-frequency transient characteristics of the fault signal can be fully analyzed in the frequency domain, thereby improving the accuracy of fault detection.

[0082] S220. A specific threshold T j(j = 1, 2, ..., L). The determination of these thresholds is based on the signal samples collected in advance under normal operating conditions and fault conditions, and is set by analyzing the statistical distribution characteristics of these samples. Such a threshold setting strategy can ensure that when a fault occurs, abnormal signals can be accurately separated from the background noise, thereby improving the sensitivity and specificity of fault detection.

[0083] S230. According to the thresholds of each layer of wavelet decomposition, wavelet decomposition is performed on the first-terminal current signal and the second-terminal current signal respectively. After wavelet decomposition processing, each resolution component is reconstructed to obtain the first-terminal transient component and the second-terminal transient component, so as to calculate the transient energy function ΔE(t). This step is crucial for identifying and quantifying the energy change of the fault signal, and helps to further analyze the nature and severity of the fault.

[0084] Through the above steps S210 to S230, it is helpful to more accurately analyze and identify abnormal conditions in the distribution line, thereby improving the detection and response capabilities of the protection device to faults.

[0085] Specifically, for step S3, according to the first-terminal transient component Δi A (t) and the second-terminal transient component Δi B (t) decomposed in step S2, and the fault trigger start time t0, calculate the transient energy function ΔE(t), and its specific mathematical expression is:

[0086]

[0087] where t is time and τ is the integration time.

[0088] Specifically, for step S4, preferably, a deep neural network comprehensively analyzes the real-time electrical signal data and the change of the transient energy function ΔE(t), and combines historical fault data to predict the probability of a fault occurring in the power system and the possible fault area. Among them, the real-time electrical signal data includes the real-time current signal i(t) and the voltage signal v(t). Since the deep neural network can learn from complex data patterns and extract features, it can achieve high-precision identification of the precursors of distribution line faults. Therefore, in this embodiment, a fault prediction model is constructed based on the deep neural network, and then the fault prediction model is used to predict the fault probability P fault and the fault area. The specific steps for constructing the fault prediction model and realizing the fault probability prediction are as follows:

[0089] S410. Adopt a hybrid model that combines a convolutional neural network (CNN) and a long short-term memory network (LSTM). The network formed by the fusion of the convolutional neural network and the long short-term memory network can be simply referred to as a convolutional neural long short-term memory network. The input layer of this model is used to receive preprocessed data, where the preprocessed data is time series data of the preprocessed current signal i(t), transient energy function ΔE(t), voltage signal v(t), and historical fault record data. Such an input design aims to provide a comprehensive data perspective for the model, thereby improving the accuracy and reliability of fault detection.

[0090] S420. In the network structure, at least two convolutional layers or LSTM layers are set to fully extract the features of the time series data. These layers are then connected to a fully connected layer, using the ReLU activation function to enhance the nonlinear expression ability of the model, and the Dropout technique is used to prevent overfitting. Finally, the model outputs the fault probability P fault and the fault area location index, providing direct reference information for fault diagnosis.

[0091] S430. In the training stage, a cross-validation method is used to evaluate the generalization ability of the model, and the learning rate is dynamically adjusted to optimize the training process. These strategies ensure that the model can quickly converge to a lower prediction error, and the model parameters are updated online to achieve real-time prediction of the fault trend. This dynamic learning and updating mechanism enables the model to adapt to changes in the operating conditions of the power system, improving the real-time and accuracy of prediction. In addition, the form of the prediction function can be expressed as P fault = f(i(t), v(t), ΔE(t), H), where H represents the historical fault data vector, and f(·) is the neural network function obtained through training.

[0092] The fault prediction model designed in the embodiment of the present invention can not only significantly improve the accuracy of fault detection, but also narrow the range of possible fault areas, thereby providing strong decision-making support for the maintenance and management of the power system. In addition, through continuous learning and optimization, the model can adapt to changes in the operating conditions of the power system, further enhancing the reliability and practicality of prediction.

[0093] Specifically, while step S4 is being executed, it also includes calculating the statistics of the transient energy function obtained in step S3 within a preset sliding window to obtain the mean function and the standard deviation function, and setting an adaptive threshold according to the mean function and the standard deviation function; when the transient energy function is not less than the adaptive threshold, a protection action is triggered. The specific steps are as follows:

[0094] First, calculate the statistics of the transient energy function obtained in step S3 within a preset sliding window to obtain the corresponding mean function μ E (t) and the standard deviation function σ E(t). Mean function μ E (t) and standard deviation function σ E (t) are expressed as follows:

[0095]

[0096] where T w is the preset sliding window time length. The preset sliding window time length can be determined according to the characteristics of the data and the analysis objective, such as 50 ms.

[0097] Secondly, according to the mean function μ E (t) and the standard deviation function σ E (t), calculate the adaptive threshold E th (t). The setting of the adaptive threshold is to dynamically adjust the detection standard, enabling it to automatically optimize according to the real-time changes of the data, thereby improving the accuracy and robustness of the detection. This method is applicable to application scenarios where the data characteristics change over time, and can ensure that even when the data distribution shifts, high-efficiency performance can still be maintained. In this embodiment, the mathematical expression of the adaptive threshold E th (t) is:

[0098] E th (t) = μ E (t) + α · σ E (t);

[0099] where α is an empirical coefficient determined according to the fault historical data.

[0100] Finally, compare the adaptive threshold E th (t) with the transient energy function ΔE(t) in step S3. When ΔE(t) ≥ E th (t), the protection mechanism will be automatically triggered. Specifically, according to the identified fault types, such as single-phase ground fault, two-phase short circuit fault or three-phase short circuit fault, the corresponding breaker hierarchical control instructions will be intelligently generated and issued. This process ensures the rapid response and precise control of the power system, thereby effectively limiting the scope of the fault impact and guaranteeing the stability and safety of the system.

[0101] Through this differential processing strategy based on the fault type, the impact on the non-fault part of the power system can be minimized, the resource allocation can be optimized, and the operation efficiency and reliability of the overall power system can be improved.

[0102] Specifically, for step S5, according to the transient energy function ΔE(t) obtained in step S3 and the fault probability P obtained in step S4 fault, and based on the regional load status, perform intelligent fault isolation and reclosing operations. This process accurately identifies the fault location and responds quickly to minimize the impact of the fault on the power system. In addition, to improve the flexibility and adaptability of the protection system, a feedback closed-loop mechanism is constructed. This mechanism can adaptively adjust the protection parameters according to the results of actual operations and the actual operating status of the system. Refer to Figure 3 , the specific steps to obtain the fault handling decision and generate the adaptive adjustment of protection parameters are as follows:

[0103] S510. Establish an input variable set that includes the transient energy function ΔE(t) and the fault probability P fault .

[0104] S520. For each input variable, set detailed membership functions. These membership functions are used to map the values of the input variables into fuzzy sets, and their membership degree parameters are determined based on a large amount of historical data statistical analysis. Such parameter setting helps to more accurately describe the membership degree of the input variables in different states, thereby improving the accuracy of fuzzy inference.

[0105] S530. Construct a fuzzy rule base, which contains some typical rules as follows:

[0106] 1) "If ΔE(t) is high and P fault is high, then the isolation area is in an emergency state";

[0107] 2) "If ΔE(t) is medium and P fault is medium, then attempt to restore power supply with a time-delay reclosing".

[0108] These rules are summarized based on expert knowledge and historical data and are used to guide the fault handling decision.

[0109] In the fuzzy inference process, adopt a fuzzy inference mechanism, such as algorithms like the Mamdani inference method, to synthesize the outputs of each rule. Then, use defuzzification algorithms such as the centroid method to convert the synthesized fuzzy result into a specific decision output, that is, obtain the fault handling decision. This process realizes the conversion from fuzzy input to accurate decision-making and provides clear guidance for fault handling.

[0110] S540. To improve the self-adaptability and learning ability, introduce a self-learning closed-loop mechanism. This mechanism uses the feedback information after actual protection actions (such as the effect of fault elimination, the success rate of reclosing, etc.) as training data, and dynamically adjusts the membership function parameters and rule weights through an incremental learning algorithm. This self-learning mechanism can continuously learn from experience and optimize its own decision-making ability, thereby improving the accuracy and reliability of fault prediction and handling.

[0111] Through steps S510 to S540, an input variable set, membership functions, and a fuzzy rule base are established, and by using fuzzy inference and a self-learning closed-loop mechanism, intelligent prediction and handling of distribution line faults can be achieved. This method based on fuzzy logic and machine learning provides strong technical support for improving the stability and reliability of the power system.

[0112] Further, before the control signal is sent out with the fault handling decision in step S5, time synchronization error correction is performed on the control devices (such as protection devices) at both ends of the distribution line. The preferred correction method is the iterative weighted least squares correction model, and the specific correction steps are as follows:

[0113] Step 1: Determine the weight w of each time synchronization error i , and the weight is mainly determined by its signal-to-noise ratio (SNR). The weight w i is specifically defined as:

[0114]

[0115] where is the noise variance of the i-th time synchronization error, and ∈ is a small positive value to prevent division by zero. This weight allocation method ensures that time synchronization errors with better signal quality have a greater influence in subsequent correction processes.

[0116] Step 2: The correction process starts from the set initial estimated time θ0 and is gradually optimized using the iterative formula according to the weight w i . The iterative process continues until a specific convergence condition is met, that is, |θ k+1 -θ k | < δ, indicating that the iteration stops when the difference in time offsets between two consecutive iteration results is less than the preset value δ (the value of δ can be 0.1 nanosecond or lower). Among them, the expression of the iterative formula is:

[0117]

[0118] where θ k is the estimated value of the current calculation period, N is the total number of transmitted entangled photon pairs, and Δt i is the time interval. This iterative method helps to gradually improve the accuracy of time synchronization and ensure the reliability of the final result.

[0119] Step 3: The obtained time offset θ = θ k+1 will be used to adjust the clocks of each node. This adjustment step ensures that the time synchronization error ΔT of the control devices at both ends of the distribution line is ≤ 1 ns, thus meeting the predetermined requirements and further achieving high-precision time synchronization of the entire power system.

[0120] Through the above steps, the time synchronization performance at both ends of the distribution line can be significantly improved, providing a solid foundation for various time-sensitive applications.

[0121] Further, before the control signal is sent according to the fault handling decision obtained in step S5, an encrypted communication tunnel is established. The steps for establishing this communication tunnel are as follows: First, a shared key K is generated by applying the Quantum Key Distribution (QKD) protocol; Second, an encrypted communication tunnel is established using the shared key K to perform real-time encrypted transmission of the transmitted data. Among them, the transmitted data includes control signals and electrical quantities.

[0122] Further, the key steps for generating the shared key K are as follows: First, randomly select the polarization basis of the quantum state to encode information; Second, perform information comparison on the public channel to ensure that the quantum state measurement results of both parties are consistent; Then, eliminate those bits found to be mismatched during the comparison process to exclude potential errors; Then, through the implementation of the quantum key distribution error correction protocol and privacy amplification technology, error correction and privacy protection processing are performed on the remaining bits; Finally, a secure key that meets the predetermined quantum bit error rate standard is generated, and this secure key is the shared key K.

[0123] The above security measures are crucial for preventing unauthorized access and protecting the key from being leaked. Through these strict security measures, it can be ensured that the generated key is both secure and reliable, providing strong support for the protection of the distribution line. Thus, ensuring the security of communication and the reliability of the key.

[0124] Embodiment 2

[0125] Based on the same inventive concept, this embodiment provides a short-circuit adaptive protection system for a distribution line. The principle of solving the problem is similar to that of a short-circuit adaptive protection method provided in Embodiment 1, and the repeated parts will not be elaborated.

[0126] This embodiment provides a short-circuit adaptive protection system for a distribution line, including:

[0127] A first-end acquisition module for acquiring the current signal at the first end of the distribution line; A second-end acquisition module for acquiring the current signal at the second end of the distribution line;

[0128] A first-end decomposition module for decomposing the first-end transient component from the first-end current signal; A second-end decomposition module for decomposing the second-end transient component from the second-end current signal;

[0129] A calculation module for calculating the transient energy function according to the first-end transient component and the second-end transient component;

[0130] A prediction module, configured to predict the fault probability of a distribution line according to a transient energy function and obtain real-time electrical signal data and historical fault data of the distribution line;

[0131] An output module, configured to obtain a fault handling decision according to the transient energy function and the fault probability; and output a control signal according to the fault handling decision;

[0132] Before outputting the control signal, time synchronization error correction is performed on both ends of the distribution line and an encrypted communication tunnel is established.

[0133] Specifically, referring to Figure 4 , in this embodiment, protection devices are provided at both ends of the distribution line 7, namely a first-end protection device 1 and a second-end protection device 2, and the two-end protection devices use quantum communication optical fibers for information transmission. In Figure 4 , the 10 kV bus 8 is a key part of the power system and is responsible for distributing 10 kV of electricity. The first-end protection device 1 includes a first-end acquisition module and a first-end decomposition module, and the second-end protection device 2 includes a second-end acquisition module and a second-end decomposition module; both the first-end protection device 1 and the second-end protection device 2 include a calculation module, a prediction module, an output module, a fuzzy logic decision module, and a quantum communication device. Among them, the fuzzy logic decision module is configured to perform intelligent fault isolation and reclosing operations after inputting the transient energy function, the fault probability, and the regional load status, and form a feedback closed loop to adaptively adjust the protection parameters.

[0134] To ensure the time synchronization of the first-end protection device 1 and the second-end protection device 2 when sending control signals to the circuit breakers at each end, it is necessary to perform time synchronization calibration on the first-end protection device 1 and the second-end protection device 2.

[0135] Specifically, the distribution line short-circuit adaptive protection system provided in this embodiment further includes a central quantum server 3, configured to correct the time synchronization error between the protection devices at both ends of the distribution line 7 (i.e., the first-end protection device 1 and the second-end protection device 2). The central quantum server 3 includes an entangled photon pair generation module, which uses a laser pump and a nonlinear crystal (e.g., a β-BBO crystal) to generate quantum entangled photon pairs, and each pair of photons is transmitted to the protection devices at both ends of the distribution line 7 through a quantum entangled photon pair distribution channel 5. The mathematical expression of the entangled state |ψ> of the quantum entangled photons is:

[0136]

[0137] where |H> and |V> respectively represent the horizontal and vertical polarization states of the photon pair, and the subscripts A and B represent the quantum communication devices in the protection devices located at both ends of the distribution line 7.

[0138] Further, the central quantum server 3 transmits the entangled photon pairs to the quantum communication devices in the protection devices at both ends of the power distribution line 7 through the low-loss optical fiber 6. Each end's quantum communication device respectively receives the entangled photon pairs and uses a high-precision clock to record the reception time t A,i and t B,i , where i represents the i-th transmission of the entangled photon pairs by the central quantum server 3. Each end's quantum communication device includes a time recording module, which is used to receive the quantum entangled photon pairs and preferably uses a single-photon detector to record the reception time. Subsequently, the time recording module sends the received time to the time synchronization calculation module of the central quantum server 3 again.

[0139] Further, the time synchronization calculation module calculates the time synchronization error Δt according to the reception time i . The specific calculation formula is:

[0140] Δt i =|t A,i -t B,i |.

[0141] Further, the central quantum server 3 also includes a correction module, which corrects the clock deviation according to the time synchronization error and uses the iterative weighted least squares method based on maximum likelihood estimation. Among them, the clock deviation refers to the time difference existing in the clocks used for time recording in the protection devices at both ends of the power distribution line 7. The working principle of the correction module can refer to the relevant descriptions of correction steps 1 to 3 in Embodiment 1.

[0142] Specifically, each end's quantum communication device also includes a transmission module, which is used to generate a shared key K by applying the quantum key distribution (QKD) protocol, establish an encrypted communication tunnel using the shared key K, and perform real-time encrypted transmission on the transmission data. Among them, the transmission data includes control signals and electrical quantities.

[0143] Further, the process of generating the shared key includes several key steps: First, randomly select the polarization basis of the quantum state to encode the information; Second, perform information comparison on the public channel to ensure that the quantum state measurement results of both parties are consistent; Then, eliminate those bits found to be mismatched during the comparison process to exclude potential errors; Finally, through the implementation of the quantum key distribution error correction protocol and privacy amplification technology, perform error correction and privacy protection processing on the remaining bits. This series of steps aims to generate a secure key that meets the predetermined quantum bit error rate standard, thereby ensuring the security of communication and the reliability of the key. To implement the above quantum key distribution process, the transmission module includes a bit sequence generation sub-module, a raw key sequence generation sub-module, an error correction sub-module, and a key output sub-module. The specific functions of each module are as follows:

[0144] The bit sequence generation sub-module is used to randomly generate a bit sequence and select a polarization basis. Specifically, it randomly generates a bit sequence of a predetermined length L Line and independently selects polarization bases (such as linear polarization basis and diagonal polarization basis) to encode and measure these bit sequences. This design ensures that the quantum states generated by the bit sequence generation sub-modules at both ends are random and unpredictable.

[0145] The raw key sequence generation sub-module is used to exchange the information of the polarization basis over a public channel, retain the bits with matching polarization bases from the information, and form a raw key sequence K raw . Specifically, after the bit sequence generation sub-module completes encoding and measurement, the quantum communication devices at both ends will exchange the information of the polarization bases they used over a public channel. Through this exchange process, both parties will identify and only retain those bits with matching polarization bases, thereby forming the raw key sequence. This design helps to ensure that only the matching quantum states are used to generate the key, thus improving the security of the key.

[0146] The error correction sub-module is used to correct the raw key sequence K raw until the error rate is lower than a preset threshold, and obtain a standard key sequence. Specifically, in this module, a quantum key distribution (QKD) error correction interactive protocol is adopted to correct the generated raw key sequence. During the error correction process, through segmented comparison and multiple rounds of iteration, the error rate in the key is gradually reduced until it drops below the preset threshold ∈1. This error correction mechanism is crucial for improving the accuracy and reliability of the key.

[0147] The key output sub-module performs privacy amplification on the standard key sequence and compresses it through a hash function to obtain a shared key K. Specifically, after the error correction sub-module completes error correction, privacy amplification processing will be performed on the key sequence. By selecting a suitable hash function, the key sequence is compressed to generate the final key. This design ensures that even if there are potential eavesdroppers, the information entropy that can be obtained is much lower than the preset security threshold ∈2, thus ensuring the security of communication.

[0148] It should be noted that during the entire public channel information exchange of the raw key sequence generation sub-module and the error correction process of the error correction sub-module, the risk of information leakage must be strictly controlled. All parameters and intermediate information used should only be used for the key negotiation process and not directly expose the content of the final key. This security measure is crucial for preventing unauthorized access and protecting the key from being leaked. Through these strict security measures, it can be ensured that the key generated by the quantum communication device is both secure and reliable, providing strong support for the protection of the distribution line 7.

[0149] This embodiment uses quantum entanglement states to achieve high-precision clock synchronization, adopts quantum key distribution (QKD) to ensure data communication security, and combines a fault detection method based on transient energy and an adaptive algorithm, as well as intelligent decision-making combining deep learning and fuzzy logic, to realize a protection system for rapid detection, accurate isolation, and intelligent recovery of short-circuit faults in distribution line 7. To more clearly elaborate the working process of this system, the following specific steps are exemplarily given:

[0150] Step 1: Quantum entanglement state synchronization. The central quantum server 3 uses laser pumping and a nonlinear crystal (such as a β-BBO crystal) to generate entangled photon pairs. Each pair of photons is respectively sent to the protection devices at both ends of the distribution line 7 and transmitted through the optical fiber 6 channel. The quantum communication devices at each end use single-photon detectors to record the reception time and return the data to the central quantum server 3. The central quantum server 3 calculates the time synchronization error through the maximum likelihood estimation and iterative weighted least squares correction methods to ensure that the clock synchronization error is lower than 1 ns.

[0151] Step 2: Quantum key distribution. After completing the time synchronization, the protection devices at each end use the BB84 protocol for quantum key distribution. Each end generates a random bit sequence and uses random polarization encoding of the straight basis and diagonal basis. After the two sides exchange the polarization basis information, only the bits of the matching basis are retained for key generation, and the bit error rate is reduced to less than 0.1% through the CASCADE error correction protocol. Finally, privacy amplification technology is used to ensure the security of the key.

[0152] Step 3: Data acquisition and preprocessing. The protection devices at each end collect the current signals of the distribution line 7 in real time (sampling rate 100 kHz) and use discrete wavelet transform (DWT) to extract the transient components. The wavelet basis is selected as Daubechies 4 (db4), and the decomposition level L is set to 4 to ensure the ability to extract the high-frequency characteristics of the fault signal. At the same time, an FIR digital filter is used to remove the power frequency steady-state components and only retain the short-time transient signals.

[0153] Step 4: Transient energy function calculation. After detecting a short-circuit fault, the protection device calculates the mean and standard deviation of the transient energy function through a sliding window (the length of the sliding window is 50 ms). The empirical coefficient α for setting the adaptive threshold ranges from 1.5 to 2.0, such as 1.5, 1.58, 1.6, etc., to balance the sensitivity and false alarm rate of fault detection.

[0154] Step 5: Fault identification and action trigger. If the transient energy exceeds the threshold value, the protection device determines that a short circuit has occurred and further distinguishes the fault type in combination with the voltage sag situation:

[0155] Single-phase grounding fault: The zero-sequence current I0 increases significantly, and the voltage sag is less than 30%.

[0156] Phase - to - phase short - circuit fault: The sudden change in current of two or three phases increases significantly, and the voltage drop is greater than 40%.

[0157] After a fault is detected, the protection device issues a tripping command for circuit breaker 4 within 2 ms, and its operation time is significantly better than that of traditional protection devices (5 - 10 milliseconds).

[0158] Step 6: Deep - learning fault prediction. This step is carried out simultaneously with Step 5. The system uses a convolutional neural network (CNN) + long short - term memory network (LSTM) to analyze historical fault data and combines the current transient signals to predict the fault area. The neural network uses 3 convolutional layers (convolution kernel size 3×3) and 2 LSTM layers (number of hidden units 64). The training data set contains fault records for the past 6 months. Such a design has an accuracy rate of over 95%.

[0159] Step 7: Fuzzy - logic decision - making. Based on transient energy, fault probability, and system load status, the system uses fuzzy - logic decision - making for intelligent fault isolation:

[0160] If the load in the fault area is light, the faulty section is directly isolated.

[0161] If the load is heavy, the system attempts a delayed reclosing (200 ms).

[0162] If the fault recovery fails, the protection setting value is readjusted and secondary isolation is performed.

[0163] Step 8: Feedback optimization and adaptive adjustment. The protection system records the fault clearing situation after each fault, updates the fault mode database, and optimizes the neural network parameters through incremental learning to improve the accuracy of future fault identification and isolation.

[0164] The distribution line short - circuit adaptive protection system provided in this embodiment ensures that the clock error between protection devices is less than 1 ns through quantum entanglement state synchronization technology combined with iterative weighted least - squares correction, greatly improving the time - synchronization accuracy. The communication link constructed by using the quantum key distribution protocol BB84 to generate shared keys, combined with the interactive protocol CASCADE error correction and privacy amplification technology in quantum key distribution, provides physical - layer security protection against man - in - the - middle attacks and significantly enhances communication security. The dynamic threshold algorithm based on multi - resolution analysis and sliding - window statistics can adjust the protection setting value according to real - time data, reduce the misoperation rate and shorten the response time, thereby improving the accuracy of adaptive protection. In addition, the intelligent decision - making module combining deep learning and fuzzy logic realizes the adaptive optimization of fault - area location and isolation decision - making, reduces the unnecessary power - outage range, ensures the stable operation of the system, and makes the intelligent decision - making more accurate. The introduction of quantum technologies (such as quantum entanglement state and quantum key distribution) can achieve sub - nanosecond - level time synchronization and physical - layer encrypted transmission, bringing new adaptive and intelligent technical support to the protection system.

[0165] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0166] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0167] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0168] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0169] Obviously, the above embodiments are merely examples given for clear illustration and are not limitations on the implementation manners. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to enumerate all the implementation manners here. And the obvious changes or modifications derived therefrom still fall within the protection scope of the present invention.

Claims

1. A short-circuit self-adaptive protection method for a distribution line, characterized in that, Including: S1. Collect current signals at both ends of the distribution line, where the current signals include the first - end current signal and the second - end current signal; S2. Decompose the first - end transient component from the first - end current signal and decompose the second - end transient component from the second - end current signal; S3. Calculate the transient energy function according to the first - end transient component and the second - end transient component; S4. According to the transient energy function, and obtain the real - time electrical signal data and historical fault data of the distribution line, predict the fault probability of the distribution line; S5. Obtain a fault handling decision according to the transient energy function and the fault probability; Output a control signal according to the fault handling decision; Among them, before outputting the control signal, it includes time - synchronization error correction for both ends of the distribution line and establishing an encrypted communication tunnel.

2. The adaptive short-circuit protection method for a distribution line according to claim 1, characterized in that While S4 is being executed, it includes calculating the statistics of the transient energy function within a preset sliding window to obtain a mean function and a standard - deviation function; setting an adaptive threshold according to the mean function and the standard - deviation function; When the transient energy function is not less than the adaptive threshold, trigger a protection action; where the expression of the mean function is: The expression of the standard - deviation function is: Among them, μ E (t) is the mean function, T w is the preset sliding window time length, ΔE(τ) is the transient energy function, σ E (t) is the standard deviation function, t is time, and τ is the integration time.

3. The adaptive short-circuit protection method for a power distribution line according to claim 1, wherein, The step of S2, decomposing the first - end transient component from the first - end current signal and decomposing the second - end transient component from the second - end current signal is: Select a wavelet basis suitable for short - time fault signal feature extraction and set the decomposition level; Calculate using the statistical distribution for the pre - collected signal samples during normal operation and faults to determine the threshold for each layer of wavelet decomposition; According to the threshold for each layer of wavelet decomposition, perform wavelet decomposition on the first - end current signal and the second - end current signal respectively, and reconstruct each resolution component to obtain the first - end transient component and the second - end transient component.

4. A short-circuit adaptive protection method for a distribution line according to claim 1, characterized in that, The step of S4, predicting the fault probability of the distribution line is: Pre - process the real - time electrical signal data, the historical fault data and the transient energy function to obtain pre - processed data; Input the pre - processed data into a fault prediction model and output the fault probability; Among them, the fault prediction model uses a convolutional neural long - short - term memory network. At least two convolutional or LSTM layers are set in the convolutional neural long - short - term memory network, followed by a fully - connected layer. The ReLU activation function and Dropout technology are used to prevent over - fitting; during the training process of the convolutional neural long - short - term memory network, cross - validation and dynamic learning rate adjustment strategies are used to reduce the prediction error, and the parameters of the fault prediction model are updated online to achieve real - time fault trend prediction.

5. The adaptive short-circuit protection method for a distribution line according to claim 1, wherein, The step of S5, obtaining a fault handling decision according to the transient energy function and the fault probability and generating an adaptive adjustment of protection parameters is: Establish the transient energy function and the fault probability into an input variable set; Set a corresponding membership function for each item in the input variable set; Construct a fuzzy rule base, use a fuzzy inference mechanism to synthesize the output of the fuzzy rule base, and after defuzzification, obtain a fault handling decision; Introduce a self-learning closed-loop mechanism, using the feedback after actually executing the fault handling decision as training data, and adjusting the parameters of the membership function and the rule weights through incremental learning.

6. The adaptive short-circuit protection method for a distribution line according to claim 1, wherein The steps for correcting the time synchronization error at both ends of the distribution line are as follows: Determine the weight of each time synchronization error, and the expression of the weight is: According to the weight and the set initial estimation time, continuous iteration is performed using the iterative formula until |θ k+1 -θ k | < δ, and then the final time offset is obtained; the expression of the iterative formula is: The control devices at both ends of the distribution line adjust their respective node clocks according to the time offset. where, w i is the weight, is the noise variance of the i-th time synchronization error, ∈ is a small positive value to prevent division by zero, δ is a preset value, θ k is the estimated value of the current calculation period, N is the total number of times of transmitting entangled photon pairs, Δt i is the time interval.

7. A short-circuit adaptive protection method for a distribution line according to claim 1, characterized in that The steps for establishing an encrypted communication tunnel are: generating a shared key, and using the shared key to establish an encrypted communication tunnel; using the encrypted communication tunnel to perform real-time encrypted transmission of the control signal.

8. An adaptive short-circuit protection system for a distribution line, characterized in that, Include: A first-end acquisition module for acquiring the current signal at the first end of the distribution line. A second-end acquisition module for acquiring the current signal at the second end of the distribution line. A first-end decomposition module for decomposing the first-end transient component from the first-end current signal. A second-end decomposition module for decomposing the second-end transient component from the second-end current signal. A calculation module for calculating the transient energy function according to the first-end transient component and the second-end transient component. A prediction module for predicting the fault probability of the distribution line according to the transient energy function, and acquiring the real-time electrical signal data and historical fault data of the distribution line. An output module for obtaining a fault handling decision according to the transient energy function and the fault probability. Output a control signal according to the fault handling decision. Among them, before outputting the control signal, it includes correcting the time synchronization error at both ends of the distribution line and establishing an encrypted communication tunnel.

9. The adaptive short-circuit protection system for a distribution line according to claim 8, characterized in that, It further includes a quantum communication device, and the quantum communication device includes a transmission module; the transmission module includes: A bit sequence generation sub-module for randomly generating a bit sequence and selecting a polarization basis. A raw key sequence generation sub-module for exchanging the information of the polarization basis on a public channel, and retaining the bits that match the polarization basis from the information to form a raw key sequence. An error correction sub-module for correcting the raw key sequence until the error rate is lower than a preset threshold to obtain a standard key sequence. A key output sub-module for performing privacy amplification on the standard key sequence and compressing it through a hash function to obtain a shared key.

10. A distribution line short-circuit adaptive protection system according to claim 8, characterized in that, It further includes a central quantum server for correcting the time synchronization error between the protection devices at both ends of the distribution line; the central quantum server includes an entangled photon pair generation module and a time synchronization calculation module; both protection devices at both ends of the distribution line include a time recording module; among them, The entangled photon pair generation module is used to generate quantum entangled photon pairs and transmit the quantum entangled photon pairs. The time recording module is used to receive the quantum entangled photon pairs and record the reception time. The time synchronization calculation module is used to calculate the time synchronization error according to the reception time.

Images