A quantum key application method and device suitable for distribution network protection business

Abstract

The present invention discloses a quantum key application method and device suitable for distribution network protection services, the method comprising obtaining distribution network protection service operation data samples from multiple dimensions, using artificial intelligence technology to predict the distribution network protection service operation status, and generating a quantum key application strategy according to the distribution network protection service operation status prediction result. The present invention predicts the distribution network protection line service action status based on artificial intelligence technology, updates the quantum key before the control instructions with high real-time, security and reliability requirements such as switch actions are triggered, and uses "one-time one-key" encryption control instructions to issue, which reduces the delay caused by key update to the control service and improves the security of service transmission.

Description

Technical Field

[0001] The present invention belongs to the technical field of quantum secure communication, and specifically relates to a quantum key application method and device suitable for distribution network protection services. Background Art

[0002] Distribution network protection services, represented by distribution network differential protection services and distribution network regional protection, have high requirements for communication bandwidth, real-time performance, reliability and security. At present, the distribution network protection service is mainly carried by optical fiber private network or 5G power virtual private network. The optical fiber private network has the characteristics of large bandwidth, low latency and high reliability, which can well meet the needs of distribution network protection services. However, the distribution network protection service data does not use any encryption method, which poses the risk of being eavesdropped. 5G communication has the characteristics of large bandwidth (eMBB), low latency (uRLLC), and wide connection (mMTC), which can be used to carry distribution network protection services. However, at present, 5G-related standards and their industrialization are still in the stage of gradual development, and most commercial 5G networks only support the R15 standard. Therefore, in order to ensure the low latency characteristics, the power system supports horizontal communication between two 5G communication terminals through self-built UPF (User Plane Function) without passing through the core network, through UPF forwarding, which greatly reduces the transmission delay. However, the 5G network is built on the operator's public network, and the security and isolation of the slices need to be further demonstrated. At the same time, the encryption scheme based on computational complexity is currently widely used in 5G networks, and the randomness and real-time performance of the keys need to be improved, which poses a security risk to the increase in computing power.

[0003] Quantum secure communication technology, based on the unconditional and secure distribution of quantum keys in information theory and "one-time one-key", can achieve end-to-end secure distribution of business data, providing a feasible way to solve the above problems. However, due to factors such as the accuracy of optical components, the amount of quantum keys generated is limited, and differentiated quantum key application methods need to be adopted for different business types. At the same time, the distribution and update of quantum keys will introduce delays, so for distribution network protection services with high real-time requirements, it is necessary to provide differentiated quantum key supply services for different business types. Summary of the invention

[0004] The purpose of the present invention is to provide a quantum key application method and device suitable for distribution network protection services. The method predicts the action status of distribution network protection line services based on artificial intelligence technology, updates the quantum key before the control instructions with high real-time, security and reliability requirements such as switch actions are triggered, and uses "one-time one-key" encryption to issue control instructions, thereby reducing the delay caused by key update to control services and improving the security of service transmission.

[0005] In order to achieve the above technical objectives, the present invention is implemented by the following technical solutions:

[0006] The present invention provides a quantum key application method applicable to distribution network protection services, comprising:

[0007] Obtain multi-dimensional data samples of distribution network protection business operation;

[0008] Based on the obtained multi-dimensional data samples of the distribution network protection service operation, a model for predicting the distribution network protection service operation status is constructed; the distribution network protection service operation status includes a normal state and a line fault state;

[0009] Based on the real-time acquired distribution network protection service data, the model for predicting the distribution network protection service operation status is used to obtain the distribution network protection service data under the current normal state and the distribution network protection service data under the line fault state; the real-time acquired distribution network protection service data includes distribution network protection equipment data and distribution network protection line data;

[0010] Predict the future operation status of distribution network protection services based on the distribution network protection service data under the current line fault status;

[0011] Different quantum key application methods are selected based on the predicted future distribution network protection service operation status.

[0012] Furthermore, the obtaining of the multi-dimensional data sample of the distribution network protection service operation includes:

[0013] Real-time and historical data of distribution network protection equipment data, distribution network protection line data, and data affected by external factors;

[0014] The distribution network protection equipment data includes equipment inventory data, equipment operation data, equipment inspection data and equipment maintenance data;

[0015] The distribution network protection line data includes line ledger data, line operation data, line fault data, line inspection data, line maintenance data and line load change data;

[0016] The external factors influencing data include meteorological data, energy load change data and power grid dispatching data.

[0017] Furthermore, the model for predicting the operation status of the distribution network protection service is constructed based on the obtained multi-dimensional data sample of the distribution network protection service operation, including:

[0018] A support vector machine-based method is used to build a model for predicting the operating status of distribution network protection services.

[0019] Furthermore, before building a model for predicting the operation status of distribution network protection services, the following steps are also included:

[0020] Clean and normalize the multi-dimensional data samples of distribution network protection business operation;

[0021] Quantitative data are directly normalized; qualitative data are weighted together with quantitative data and unified into two types: normal operation of distribution network protection service and line fault.

[0022] Furthermore, based on the distribution network protection service data under the current line fault state, the future distribution network protection service operation state is predicted, including:

[0023] The LBP algorithm is used to learn the distribution network protection service data under the current line fault state, and the data of distribution network protection actions caused by external factors and distribution network protection actions caused by line faults are extracted;

[0024] The extracted data and the weighted factors of external environmental influencing factors are weighted using the optimal weighted combination and input into the long short-term memory network to predict the future operation status of distribution network protection services.

[0025] Furthermore, the different quantum key application methods are selected based on the predicted future distribution network protection service operation status, including:

[0026] For control services, the control instructions are encrypted and issued using the quantum key "one-time one-key" updated at time t-1 before the distribution network protection device is activated;

[0027] For collection services, the historical quantum key in the quantum key pool is used to encrypt and distribute the service data; the historical quantum key refers to the quantum key farthest from the current moment.

[0028] Furthermore, for control services, if the quantum key acquisition fails at time t-1, the control instruction is encrypted and issued using the latest quantum key stored in the quantum key pool.

[0029] Furthermore, all used quantum keys in the quantum key pool are deleted after use;

[0030] The quantum key is generated and stored in real time based on an optical fiber quantum key distribution device, a satellite quantum key distribution device, or a quantum random number generator device; and quantum key distribution is performed based on an optical fiber private network or a wireless channel.

[0031] Furthermore, it also includes:

[0032] After the control command is encrypted and issued, the operating status of the distribution network protection line is judged to determine whether a false operation occurs;

[0033] If an erroneous action occurs, it is fed back to the training phase of the long short-term memory network.

[0034] The present invention also provides a quantum key application device applicable to a distribution network protection service, the device is used to implement the aforementioned quantum key application method applicable to a distribution network protection service, the device comprising:

[0035] Initialize the data acquisition module to obtain multi-dimensional data samples of distribution network protection business operation;

[0036] A prediction model building module, used to build a model for predicting the distribution network protection service operation status based on the obtained multi-dimensional data samples of the distribution network protection service operation; the distribution network protection service operation status includes a normal state and a line fault state;

[0037] A data classification module, which is used to obtain the distribution network protection service data under the current normal state and the distribution network protection service data under the line fault state based on the distribution network protection service data obtained in real time by using the model for predicting the distribution network protection service operation state; the distribution network protection service data obtained in real time includes distribution network protection equipment data and distribution network protection line data;

[0038] A state prediction module is used to predict the future operation state of the distribution network protection service based on the distribution network protection service data under the current line fault state;

[0039] The configuration module is used to select different quantum key application methods based on the predicted future distribution network protection service operation status.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] (1) The present invention predicts the future operation status of distribution network protection line services. For high-security control services, quantum key updates are performed before control instructions are triggered, which greatly reduces the delay caused by key updates to control services.

[0042] (2) The present invention provides differentiated key application strategies for distribution network protection collection services and control services, which can effectively solve the problem of insufficient quantum key generation. For control services, the present invention uses real-time updated, "one-time one-key" quantum keys to encrypt service data, which greatly improves the security of service transmission. For collection services, historically updated, expanded and amplified quantum keys are used to encrypt service data, which has better effects than existing methods in terms of key randomness and update frequency. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 A flow chart of a quantum key application method applicable to distribution network protection services provided in this embodiment;

[0044] Figure 2 A flow chart of data sample feature extraction provided for this embodiment;

[0045] Figure 3 A flow chart of the future distribution network protection service operation status prediction model training provided in this embodiment;

[0046] Figure 4 Different quantum key application strategies are provided for this embodiment. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the scope of protection of the present invention.

[0048] Example 1

[0049] This embodiment provides a quantum key application method applicable to distribution network protection services, and the specific implementation process is as follows:

[0050] S1. Obtain multi-dimensional data samples of distribution network protection business operation

[0051] The operation status of the distribution network protection business will be affected by the interaction of internal factors such as distribution network protection equipment and distribution network protection lines, as well as external factors such as environment, energy load, and power grid dispatch. Among them, the impact of some data is periodic and regular, such as: the gradual change of current information before and after the distribution network protection equipment and line failure in internal factors, and environmental data and energy load data in external factors. This type of data can be quantitatively modeled for data mining and analysis. The impact of some data is sudden and irregular, such as: power grid dispatch in external factors, etc. This type of data is difficult to quantify and model, and can be qualitatively analyzed and superimposed on the aforementioned influencing factors.

[0052] In summary, the sources of distribution network protection business operation data samples include: distribution network protection equipment data, distribution network protection line data, real-time data and historical data in dimensions such as external factors, as shown in Table 1 below.

[0053] Table 1 Distribution network protection service operation data sample source

[0054]

[0055]

[0056] S2. Based on the obtained multi-dimensional data samples of distribution network protection business operation, a model for predicting the operation status of distribution network protection business is constructed;

[0057] In this embodiment, the distribution network protection service operation status is divided into two categories: normal operation status and line fault status.

[0058] Under normal operating conditions, the distribution network protection equipment only collects status data such as the current at both ends of the line (referred to as collection services) and performs real-time comparison;

[0059] When a line fails, the distribution network protection equipment needs to perform a switch action (referred to as control service) to isolate the faulty section.

[0060] In this embodiment, based on the distribution network protection service operation data samples, a model for predicting the distribution network protection service operation status is constructed, and data samples of normal operation and line failure are extracted. Figure 2 As shown, the specific implementation process is:

[0061] S2-1. Data cleaning and normalization

[0062] Data cleaning is used to remove data that is obviously problematic;

[0063] Data normalization is to quantify the quantifiable data, conduct qualitative analysis on the difficult-to-quantify data, and weight the quantifiable data to unify them into two states: normal operation of distribution network protection business and fault of distribution network protection line, forming key feature quantities;

[0064] S2-2, using data mining algorithms such as support vector machine (SVM) to analyze the temporal and spatial characteristics of samples, extracting data samples during normal operation and line faults, as follows:

[0065] Most (80%) samples containing n key features (such as the influence of distribution network protection equipment, distribution network protection lines and external environment when the switch is in action) are used as training data and as input for SVM training to construct a classifier and establish an SVM model for predicting the operating status of distribution network protection services.

[0066] The remaining samples (20%) containing n key feature quantities are used as test data for testing and compared with the original evaluation results (historical data) to obtain the prediction accuracy and determine whether they are normal operation samples or line fault samples.

[0067] (1) The basic idea of ​​support vector machine is: given l sample data (x 1 ,y 1 ),(x 2 ,y 2 ),…,(x l ,y l )∈R n ×R, where x is the sample input and y is the sample output. First, a nonlinear mapping is used to transform the input vector from the original space R nMapped to a high-dimensional feature space (Hibert space); then, in this high-dimensional feature space, the principle of structural risk minimization is used to construct the optimal decision function, and the kernel function of the original space is used to replace the dot product operation of the high-dimensional feature space to avoid complex operations, thereby converting the nonlinear function estimation problem into a linear function problem in the high-dimensional feature space.

[0068] Assume that the constructed optimal decision function has the following form:

[0069]

[0070] The purpose of solving the above equation is to use the principle of structural risk minimization to find the parameter ω T and b.

[0071] Find ω T and b are equivalent to solving the following optimization problem:

[0072]

[0073] Through the dual form of the above formula, we can find its optimal solution.

[0074] The dual form of the above formula can establish the Lagarangian function according to the objective function and constraints, and calculate the partial derivative of the established Lagarangian function. According to its partial derivative equation, the optimization problem to be solved can be transformed into solving a linear equation, and the kernel function K(x,x,) = (x)(x,) is defined to replace the nonlinear mapping. Finally, the estimation function (i.e., the system model) of the least squares support vector machine is obtained as follows:

[0075]

[0076] Where K(x,x i ) is any symmetric function satisfying Mercer's condition.

[0077] (2) Feature extraction method of distribution network line protection service data samples

[0078] Any one of the model parameters trained by the support vector machine is selected as the result variable. For the remaining parameters, existing samples are used for training. When a new set of data is received, the corresponding result variable is inferred by analyzing the new data using the regression method of the support vector machine. The predicted value is compared with the true value. If the true recorded value is within the error range, no fault alarm occurs. If the difference is large, it indicates that the parameter is abnormal and the alarm may indicate a fault, thereby achieving the purpose of line fault prediction.

[0079] The definition of training samples is similar to that in (1). When a certain factor is selected as the result set, the remaining factors are used as input quantities, and the output set Y=R here, that is, Y is a continuous real number within the value range of the factor. Use a nonlinear mapping to map the data to a high-dimensional feature space, and then perform linear regression in the high-dimensional feature space. The optimal decision function is the same as in (1), which is:

[0080]

[0081] The ε-insensitive loss function is introduced in support vector machine regression to ignore the error within a certain range of the true value, which is defined as follows:

[0082]

[0083] So we get the following optimized form:

[0084]

[0085] The constraints are:

[0086]

[0087] Optimize the dual form of the above equation and introduce nonlinear mapping to map the data to a high-dimensional feature space, that is, replace the inner product with a kernel function. Then the problem becomes to find the maximum value of the following function:

[0088]

[0089] Solving for α i and The value of f(x) can be obtained as:

[0090]

[0091] In this way, the characteristic value of a certain factor of the distribution network protection line status data to be analyzed can be obtained. By comparing it with the recorded value, it can be judged whether the current status of the distribution network protection line is normal or there may be a fault, which corresponds to the data samples when the distribution network line protection service is operating normally and the line is faulty.

[0092] S3. Based on the real-time acquired distribution network protection service data, the above-constructed SVM model for predicting the operation status of the distribution network protection service is used to obtain the distribution network protection service data under the current normal state and the distribution network protection service data under the line fault state; wherein, the real-time acquired distribution network protection service data includes distribution network protection equipment data and distribution network protection line data.

[0093] S4. Predicting the future operation status of the distribution network protection service based on the distribution network protection service data under the current line fault status;

[0094] For the specific implementation process, see Figure 3 ,include:

[0095] S4-a, perform deep learning on the data samples when the distribution network line fails, and use the LBP algorithm to extract the corresponding data sample features during the distribution network protection action process of "external influence-distribution network protection line failure-distribution network protection device action" (the line action is due to external factors) or "distribution network protection line failure-distribution network protection device action" (the line action is the line itself Fault).

[0096] S4-b, based on the extracted data sample characteristics and the weighted factors of external environmental influencing factors, the long short-term memory network (LSTM) algorithm is used to predict whether the future operation status of the distribution network protection service is a control service.

[0097] First, the extracted data sample characteristics and the weighted factors of the external environment are weighted to obtain the pre-processed data weights, and the optimal weighted combination is used for weighted processing. First, the deviation matrix E is obtained, that is:

[0098]

[0099] Where N is the total number of load samples, e 1t and e 2t are the errors between the predicted value and the true value of the extracted data sample features and the weighted factors of the external environment at time t.

[0100] The optimal weight can be obtained by the Lagrange multiplier method, as shown in the following formula:

[0101]

[0102] In the formula, w 1 and w 2 are the weight coefficients of the extracted data sample features and the weighted factors of the external environment, and the sum of the coefficients is 1, R = [1,1] T .

[0103] In summary, the final load forecast result can be obtained as follows:

[0104]

[0105] In the formula, is the load forecast result of the weighted result at time t, and They are the load forecast results at time t based on the extracted data sample characteristics and the weighted factors of the external environment influence.

[0106] Then the preprocessed data is divided into two parts: training set and test set according to 9:1;

[0107] Build an LSTM network to train and verify the data, and obtain a prediction model for the future operation status of the distribution network protection service (whether it is a control service).

[0108] The specific steps of the LSTM algorithm are as follows:

[0109] The LSTM network changes the state by controlling the gate information, where the forget gate is set to f t , the input gate is set to o t The cell state C at time t t The state at this time is expressed as:

[0110]

[0111] f t The calculation formula is:

[0112] f t =σ(W f ·[h t-1 ,x t ]+b f );

[0113] Sigmoid is selected as the activation function, and tanh is used to update the unit state.

[0114] set up Update the value of the cell state, i t is the input gate vector, and the input gate and unit state update values ​​of LSTM are calculated as follows:

[0115]

[0116] The previous hidden state and the current input are passed to the sigmoid function, and then the new state is passed to the tanh function. The output of tanh is multiplied by the output of sigmoid to determine the information that the hidden state should carry. The calculation formula for the output gate is:

[0117]

[0118] By adding b o The mean of is initialized to 1, which can make LSTM achieve the same effect as GRU.

[0119] On this basis, the purpose of predicting the operating status of distribution network protection and control services can be achieved.

[0120] S5. Generate different quantum key application strategies based on the predicted future distribution network protection service operation status

[0121] In this embodiment, based on the prediction results of the future distribution network protection service operation status, different quantum key application strategies are adopted for control services and collection services. The data sample features are refined into the t-1 moment before the distribution network protection device acts, the t moment during the action, and the t+1 moment after the action.

[0122] See also Figure 4 For control services,

[0123] The control instructions are encrypted and issued using the quantum key "one-time one-key" updated at time t-1 before the distribution network protection equipment is activated to complete the switch action and cut off the faulty section of the distribution network line.

[0124] If the quantum key acquisition fails at time t-1, the most recently stored quantum key in the quantum key pool, that is, the quantum key closest to time t (such as t-2), is used, and the "one-time one-key" encryption control instruction is issued to complete the switching action and cut off the faulty section of the distribution network line.

[0125] See also Figure 4 For collection services,

[0126] Due to the large flow of business data and high real-time nature, the business data is encrypted using the historical key in the quantum key pool (the one farthest from time t. For example, if there are keys corresponding to N times before time t, then the key at time tN, that is, the oldest key, is used first).

[0127] It should be noted that the keys in the quantum key pool are all stored with time tags. In this embodiment, the historical keys can be used starting from the ones with the earliest time tags, that is, the ones farthest from time t. If there are keys corresponding to N times before time t, then the key at time tN is used first, that is, the oldest key is used.

[0128] It should be noted that the key can be amplified and the key update cycle can be determined according to the amount of power business data to be encrypted and the amount of keys stored locally, which can generally be set to different levels such as hours and days.

[0129] It should be noted that all used quantum keys in the quantum key pool will be deleted after use and will not be reused. Quantum keys can be generated and stored in real time based on optical fiber quantum key distribution equipment, satellite quantum key distribution equipment, or quantum random number generator equipment. At the same time, quantum key distribution can be based on optical fiber private networks or wireless channels (including operator public networks and 5G power virtual private networks).

[0130] In this embodiment, it also includes: analyzing the operating status of the distribution network protection line after quantum encryption transmission,

[0131] Whether a false operation occurs is determined by the operating status of the distribution network protection line (for example, after the switch is cut off, there will be no current, etc.). If a false operation occurs, it will be fed back to the training stage of the LSTM prediction model as historical experience data to prevent calculation errors next time.

[0132] Example 2

[0133] This embodiment provides a quantum key application device applicable to a distribution network protection service. The device is used to implement the quantum key application method applicable to a distribution network protection service of the aforementioned embodiment 1. The device includes:

[0134] Initialize the data acquisition module to obtain multi-dimensional data samples of distribution network protection business operation;

[0135] A prediction model building module, used to build a model for predicting the distribution network protection service operation status based on the obtained multi-dimensional data samples of the distribution network protection service operation; the distribution network protection service operation status includes a normal state and a line fault state;

[0136] A data classification module, which is used to obtain the distribution network protection service data under the current normal state and the distribution network protection service data under the line fault state based on the distribution network protection service data obtained in real time by using the model for predicting the distribution network protection service operation state; the distribution network protection service data obtained in real time includes distribution network protection equipment data and distribution network protection line data;

[0137] A state prediction module is used to predict the future operation state of the distribution network protection service based on the distribution network protection service data under the current line fault state;

[0138] The configuration module is used to select different quantum key application methods based on the predicted future distribution network protection service operation status.

[0139] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0140] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0141] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0142] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0143] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all fall within the protection of the present invention.

Claims

1. A quantum key application method suitable for distribution network protection services, It is characterized in that include: Obtain multi-dimensional data samples of distribution network protection business operation, including: distribution network protection equipment data, distribution network protection line data, real-time data and historical data of external factors affecting data; the distribution network protection equipment data includes equipment inventory data, equipment operation data, equipment inspection data and equipment maintenance data; the distribution network protection line data includes line inventory data, line operation data, line fault data, line inspection data, line maintenance data and line load change data; the external factors affecting data include meteorological data, energy load change data and power grid dispatching data; Based on the obtained multi-dimensional data samples of the distribution network protection service operation, a model for predicting the distribution network protection service operation status is constructed; the distribution network protection service operation status includes a normal state and a line fault state; Based on the real-time acquired distribution network protection service data, the model for predicting the distribution network protection service operation status is used to obtain the distribution network protection service data under the current normal state and the distribution network protection service data under the line fault state; the real-time acquired distribution network protection service data includes distribution network protection equipment data and distribution network protection line data; Based on the distribution network protection service data under the current line fault state, the future distribution network protection service operation state is predicted, including: using the LBP algorithm to learn the distribution network protection service data under the current line fault state, extracting the data of the distribution network protection action caused by external factors and the distribution network protection action caused by the line fault itself; using the optimal weighted combination to weight the extracted data and the weighted factors of the external environmental influencing factors, and inputting them into the long short-term memory network to predict the future distribution network protection service operation state; Different quantum key application methods are selected based on the predicted future distribution network protection service operation status, including: for control services, the control instructions are encrypted and issued using the quantum key updated at time t-1 before the distribution network protection equipment is operated; for collection services, the service data is encrypted and issued using the historical quantum key in the quantum key pool; the historical quantum key refers to the quantum key farthest from the current moment.

2. According to claim 1, a quantum key application method suitable for distribution network protection services, It is characterized in that The constructing a model for predicting the operation status of the distribution network protection service based on the obtained multi-dimensional data sample of the distribution network protection service operation includes: A support vector machine-based method is used to build a model for predicting the operating status of distribution network protection services.

3. A quantum key application method suitable for distribution network protection services according to claim 2, It is characterized in that Before building a model for predicting the operation status of distribution network protection services, the following are also required: Perform data cleaning and normalization on multi-dimensional data samples of distribution network protection business operation; Quantitative data are directly normalized; qualitative data are weighted together with quantitative data and unified into two types: normal operation of distribution network protection service and line fault.

4. According to claim 1, a quantum key application method suitable for distribution network protection services, It is characterized in that For control services, if the quantum key acquisition fails at time t-1, the control instructions are encrypted and issued using the latest quantum key stored in the quantum key pool.

5. According to claim 1, a quantum key application method suitable for distribution network protection services, It is characterized in that All used quantum keys in the quantum key pool are deleted after use; The quantum key is generated and stored in real time based on an optical fiber quantum key distribution device, a satellite quantum key distribution device, or a quantum random number generator device; and quantum key distribution is performed based on an optical fiber private network or a wireless channel.

6. A quantum key application method suitable for distribution network protection services according to claim 1, It is characterized in that Also includes, After the control command is encrypted and issued, the operating status of the distribution network protection line is judged to determine whether a false operation occurs; If an erroneous action occurs, it is fed back to the training phase of the long short-term memory network.

7. A quantum key application device suitable for distribution network protection services, It is characterized in that The device is used to implement the quantum key application method applicable to the distribution network protection service according to any one of claims 1 to 6, and the device includes: Initialize the data acquisition module to obtain multi-dimensional data samples of distribution network protection business operation; A prediction model building module, used to build a model for predicting the distribution network protection service operation status based on the obtained multi-dimensional data samples of the distribution network protection service operation; the distribution network protection service operation status includes a normal state and a line fault state; A data classification module, which is used to obtain the distribution network protection service data under the current normal state and the distribution network protection service data under the line fault state based on the distribution network protection service data obtained in real time by using the model for predicting the distribution network protection service operation state; the distribution network protection service data obtained in real time includes distribution network protection equipment data and distribution network protection line data; A state prediction module is used to predict the future operation state of the distribution network protection service based on the distribution network protection service data under the current line fault state; The configuration module is used to select different quantum key application methods based on the predicted future distribution network protection service operation status.

Images