Data recovery method, system and equipment of power distribution network and medium

Abstract

The invention discloses a data recovery method, system and device for a power distribution network and a medium, and the method is characterized in that the method comprises the steps: obtaining multi-dimensional perception data of the power distribution network; dynamically determining a channel allocation scheme corresponding to each data level based on each channel state scoring result and the data level corresponding to each type of sensing data in the multi-dimensional sensing data, and performing data transmission according to the channel allocation scheme by using a multi-channel parallel transmission mechanism to form a data transmission result corresponding to each channel; and performing redundant coding on the data transmission result to generate redundant coded data, and inputting the redundant coded data into a sparse signal recovery model to solve the sparse signal recovery model to obtain recovery data of the power distribution network. The data recovery accuracy and efficiency of the power distribution network can be improved so as to ensure the stability of the data.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power distribution network intelligent sensing, and in particular to a data recovery method, system, device and medium for a power distribution network. BACKGROUND

[0002] With the rapid development of smart power distribution networks, a large number of optical fiber sensing, low-voltage carrier communication and wireless sensing terminals are deployed in power grid operation sensing systems to collect multi-source operation information such as current, voltage, temperature, humidity, stress and air pressure. Due to the wide distribution of these sensing devices and the complex working environment, the data is extremely susceptible to electromagnetic interference, channel fading and communication congestion during the collection, transmission and fusion process, resulting in packet loss, damage or delay of part of the sensing data. If the lost data cannot be recovered in time, it will directly affect the complete monitoring of the power grid operation state and the accuracy of fault identification, and thus endanger the safe and stable operation of the power distribution system. Therefore, in order to ensure the continuity and integrity of the sensing information, it is a key technical requirement to realize high-precision recovery of the lost data in the smart power distribution network to ensure the reliability of monitoring.

[0003] However, the existing technology still has obvious deficiencies in the recovery of multi-source sensing data. Traditional methods rely on single-channel transmission and fixed redundancy mechanisms, which are difficult to cope with multi-dimensional data loss in complex communication environments. Although some systems use simple error correction coding, the recovery capability is limited, and only a small amount of continuous errors can be repaired, and effective recovery of randomly lost or cross-channel missing data cannot be achieved. In addition, existing sensing systems generally lack adaptive recovery mechanisms based on data characteristics and channel states, and do not fully utilize the sparsity characteristics of sensing data for efficient reconstruction, resulting in insufficient recovery accuracy and serious error accumulation. Especially in the multi-source heterogeneous data scenario, the correlation between different types of data is not fully utilized, and the recovery model cannot balance accuracy and efficiency, making it difficult to meet the demand of the power distribution network for real-time, stable and highly reliable data recovery. Therefore, there is an urgent need for a multi-dimensional sensing data recovery mechanism that combines redundancy coding and sparse reconstruction algorithms, with multi-channel collaboration and adaptive optimization capabilities, to achieve high-precision restoration and complete recovery of lost data. SUMMARY

[0004] The present application provides a data recovery method, system, device and medium for a power distribution network, which can improve the accuracy and efficiency of data recovery for the power distribution network, thereby ensuring the stability of the data.

[0005] In a first aspect, an embodiment of the present application provides a data recovery method for a power distribution network, comprising:

[0006] acquiring multi-dimensional sensing data of the power distribution network, wherein the multi-dimensional sensing data includes operation data and environmental data;

[0007] determine a channel allocation scheme corresponding to each data level based on the channel state score results and data levels corresponding to each type of perception data in the multi-dimensional perception data, perform data transmission according to the channel allocation scheme by using a multi-channel parallel transmission mechanism, and form data transmission results corresponding to each channel, wherein each channel state score result is calculated based on a state parameter of each channel;

[0008] perform redundant encoding on the data transmission results to generate redundant encoding data, input the redundant encoding data into a sparse signal recovery model to solve the sparse signal recovery model, and obtain recovery data of the power distribution network, wherein the sparse signal recovery model is calculated based on a linear mapping relationship between historical recovery data and historical data transmission results.

[0009] The embodiment of the present application realizes comprehensive perception of the power grid operation state by obtaining multi-dimensional perception data of the power distribution network, enables the data recovery process to model the influence of environmental changes on data fluctuations, thereby improving the identification and repair accuracy of the recovery model for abnormal or missing data; based on the channel state score results and the data levels corresponding to each type of perception data, a channel allocation scheme is dynamically determined, and a multi-channel parallel transmission mechanism is used for data transmission, so that high-priority data is transmitted through high-quality channels first, reducing the risk of loss of critical data and reducing the burden of data recovery from the source and improving the integrity of the recovery data; the data transmission results are redundantly encoded, and the redundant encoding data is input into the sparse signal recovery model, which is solved by a sparse recovery algorithm, which can accurately complete the data in the case of partial data loss by using the principle of sparse reconstruction, thereby enhancing the accuracy and stability of data recovery; in summary, through the synergistic effect of multi-dimensional perception fusion, dynamic channel allocation and sparse reconstruction mechanism, the data recovery quality and reliability of the power distribution network in a complex communication environment are effectively improved, and the high-precision recovery of key perception data and the protection of data integrity are realized.

[0010] Further, the multi-dimensional perception data of the power distribution network is obtained, including:

[0011] obtain initial multi-dimensional perception data of the power distribution network, wherein the initial multi-dimensional perception data includes initial operation data and initial environment data;

[0012] normalize the initial operation data and the initial environment data respectively to obtain operation normalization results and environment normalization results;

[0013] perform sparse sampling on the operation normalization results based on a preset compressive sensing sampling matrix to generate first compressed data, and perform compression on the environment normalization results by using a differential encoding method to generate second compressed data;

[0014] determine multi-dimensional perception data of a power distribution network based on the first compressed data and the second compressed data.

[0015] The embodiment of the application realizes unified processing and efficient compression of data by normalizing, sparsely sampling and differentially compressing multi-source perception data, retains key power grid parameter information, improves data transmission efficiency and reconstruction accuracy, enhances real-time performance and reliability of multi-dimensional state perception of the power distribution network, thereby providing high-quality input for subsequent sparse reconstruction, and improves recovery accuracy and stability of multi-dimensional perception data of the power distribution network.

[0016] Further, the channel allocation scheme corresponding to each data level is dynamically determined based on the channel state score results and the data levels corresponding to each type of perception data in the multi-dimensional perception data, and the channel allocation scheme comprises:

[0017] comprehensive evaluation of data characteristics corresponding to each type of perception data in the multi-dimensional perception data is performed to determine data levels corresponding to each type of perception data;

[0018] weighting and normalization calculation of the state parameters of each channel are performed to generate state score results corresponding to each channel;

[0019] based on the data levels, the channel state score results and a preset weight coefficient, a preset data allocation model is called to allocate each type of perception data to a corresponding channel to obtain a channel allocation scheme, wherein the weight coefficient is calculated from the adaptability of each type of perception data to each channel.

[0020] The embodiment of the application determines data levels by feature evaluation of multi-dimensional perception data, and realizes dynamic allocation in combination with channel state scores, so that important data is preferentially transmitted through high-quality channels, thereby reducing the risk of loss of key data; the adaptability weight and the data allocation model are introduced to optimize the channel allocation scheme, thereby improving multi-channel utilization efficiency and transmission stability, effectively reducing transmission error and packet loss rate, ensuring the integrity of the transmitted data, and further improving the accuracy and reliability of subsequent data recovery.

[0021] Further, the data transmission is performed according to the channel allocation scheme by using a multi-channel parallel transmission mechanism to form data transmission results corresponding to each channel, and the data transmission comprises:

[0022] each type of perception data in the channel allocation scheme is transmitted in parallel by using a multi-channel parallel transmission mechanism to generate an initial data transmission result;

[0023] Real-time acquisition of feedback information of each channel in the data transmission process, construction of a channel feedback optimization model based on the feedback information, and calculation of the state adjustment amount of each channel based on the channel feedback optimization model, dynamic adjustment of the use priority and resource allocation ratio of each channel according to the state adjustment amount;

[0024] The use priority and the resource allocation ratio are applied to the initial data transmission result to form a data transmission result corresponding to each channel.

[0025] The embodiment of the application realizes synchronous transmission of data through a multi-channel parallel transmission mechanism, improves the data transmission efficiency; by constructing a channel feedback optimization model, dynamically adjusting the channel priority and resource allocation ratio according to real-time feedback information such as bandwidth, delay and error rate, the transmission process can adapt to the change of channel state, thereby maintaining efficient and stable data transmission performance, and further providing more complete and higher quality input data support for subsequent data recovery.

[0026] Further, the redundant encoding of the data transmission result to generate redundant encoding data comprises:

[0027] The data transmission result is redundantly encoded to obtain a generator matrix and a redundancy matrix, and a redundancy symbol is determined based on the generator matrix and the redundancy matrix;

[0028] The redundancy symbol is combined with the corresponding original data symbol in the data transmission result to generate redundant encoding data.

[0029] The embodiment of the application can effectively correct and recover lost or damaged data symbols in the data transmission process by redundantly encoding the data transmission result, improving the data error correction and anti-interference capability; by generating a generator matrix and a redundancy matrix to determine a redundancy symbol, and combining it with an original data symbol to form redundant encoding data, the original information can still be accurately recovered in the case of partial data loss, ensuring the integrity of the transmitted data in a complex channel environment, thereby improving the accuracy and stability of subsequent data recovery.

[0030] Further, the construction process of the sparse signal recovery model comprises:

[0031] The historical redundant encoding data is linearly mapped by a preset measurement matrix to obtain observation data, and the observation data and the measurement matrix are input into the sparse signal recovery model to iteratively optimize the sparse signal using a preset sparse recovery algorithm to obtain a reconstructed signal, wherein the sparse signal is obtained by sparse transformation of the historical redundant encoding data;

[0032] The trained sparse signal recovery model is obtained by minimizing the error between the reconstructed signal and the observation data when a threshold condition of a preset non-zero element quantity limit is satisfied.

[0033] The embodiment of the present application can still obtain observation information that can be used for reconstruction in the case of partial data loss by inputting redundant coded data into the sparse signal recovery model and performing linear mapping using the measurement matrix, thereby effectively preserving the characteristics of the original signal; high-precision signal reconstruction can be realized under the condition of compressed sampling by iteratively optimizing the sparse signal through the sparse recovery algorithm, thereby improving the stability and robustness of data recovery; the recovery process takes into account both accuracy and sparsity by minimizing the error between the reconstructed signal and the observation data and combining the constraint condition of the non-zero element quantity limit, so that the optimal recovered data is obtained, thereby ensuring the integrity and reliability of the power distribution network sensing data and improving the accuracy and quality of data recovery.

[0034] Further, after obtaining the recovered data of the power distribution network, the method further comprises:

[0035] The receiving end decodes the recovered data through a polynomial interpolation-based error correction decoding algorithm after receiving the recovered data, obtains first recovered data, and reconstructs the first recovered data through a L1 regularization-based sparse recovery algorithm, thereby obtaining second recovered data;

[0036] The equal probability of the second recovered data and the expected data set is calculated, and when the equal probability is greater than or equal to a preset consistency threshold, the target recovered data of the power distribution network is determined.

[0037] The embodiment of the present application can evaluate the accuracy and reliability of the recovered data in real time by verifying the consistency of the recovered data and the expected data set; by calculating the equal probability and comparing it with the preset threshold, abnormalities or errors in the recovered data can be found in time, thereby ensuring that the finally output data is highly matched with the actual power grid state, effectively improving the reliability of the data recovery result, and providing an accurate and reliable data basis for the monitoring, analysis and decision-making of the power distribution network.

[0038] In a second aspect, the embodiment of the present application provides a data recovery system for a power distribution network, which comprises an acquisition module, a transmission module and a recovery module.

[0039] The acquisition module is configured to acquire multi-dimensional sensing data of the power distribution network, wherein the multi-dimensional sensing data comprises operating data and environmental data.

[0040] The transmission module is configured to dynamically determine a channel allocation scheme corresponding to each data level based on each channel state score result and a data level corresponding to each type of perception data in the multi-dimensional perception data, and perform data transmission according to the channel allocation scheme by using a multi-channel parallel transmission mechanism to form a data transmission result corresponding to each channel, wherein each channel state score result is calculated based on a state parameter of each channel.

[0041] The recovery module is configured to perform redundancy encoding on the data transmission result to generate redundancy encoding data, input the redundancy encoding data into a sparse signal recovery model, solve the sparse signal recovery model to obtain recovery data of the power distribution network, wherein the sparse signal recovery model is calculated based on a linear mapping relationship between historical recovery data and historical data transmission results.

[0042] The embodiments of the present application realize efficient collection, reliable transmission and accurate reconstruction of multi-dimensional perception data of the power distribution network by constructing a data recovery system including an acquisition module, a transmission module and a recovery module; the acquisition module can comprehensively collect operation data and environmental data to provide complete information basis for subsequent recovery; the transmission module improves the transmission reliability of key data, reduces the risk of loss and error code, and provides high-quality input for the recovery process by using dynamic channel allocation and a multi-channel parallel transmission mechanism; the recovery module realizes high-precision reconstruction of lost or damaged data by combining redundancy encoding and a sparse signal recovery model, thereby obtaining recovery data of the power distribution network; and the system effectively improves the accuracy, integrity and stability of data recovery of the power distribution network, and provides reliable data guarantee for power grid monitoring and operation analysis.

[0043] In a third aspect, the embodiments of the present application provide a terminal device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface complete communication with each other through the communication bus.

[0044] The memory is configured to store at least one executable instruction, and the executable instruction causes the processor to perform the operations of the power distribution network data recovery method described in the present application.

[0045] In a fourth aspect, the embodiments of the present application provide a computer readable storage medium, which comprises a stored computer program, wherein when the computer program runs, the computer readable storage medium controls a device or system where the computer readable storage medium is located to perform the power distribution network data recovery method described in the present application.

[0046] The above description is merely an overview of the technical solutions of the embodiments of the present invention. In order to better understand the technical means of the embodiments of the present invention and to implement them in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the embodiments of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0047] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0048] Figure 1 This is a flowchart illustrating one embodiment of a data recovery method for a power distribution network provided in this application;

[0049] Figure 2 This is a flowchart illustrating steps S201 to S203 provided in this application;

[0050] Figure 3 This is a flowchart illustrating steps S301 to S302 provided in this application;

[0051] Figure 4 This is a flowchart illustrating steps S401 to S402 provided in this application;

[0052] Figure 5 This is a schematic diagram of an embodiment of a data recovery system for a power distribution network provided in this application. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0055] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0056] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0057] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0058] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0059] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0060] With the rapid development of smart distribution networks, a large number of fiber optic sensors, low-voltage carrier communication devices, and wireless sensing terminals have been deployed in power grid operation sensing systems to collect multi-source operational information such as current, voltage, temperature, humidity, stress, and air pressure. However, due to the wide distribution of sensing devices and the complex working environment, data is susceptible to electromagnetic interference, channel fading, and communication congestion during acquisition, transmission, and fusion. This can lead to packet loss, corruption, or delays in some sensing data. If lost data cannot be recovered in a timely manner, it will directly affect the accuracy of complete monitoring of the power grid's operational status and fault identification, thereby jeopardizing the safe and stable operation of the distribution system. Existing technologies still have significant shortcomings in the recovery of multi-source sensing data. Traditional methods mostly rely on single-channel transmission and fixed redundancy mechanisms, which are difficult to cope with the problem of multi-dimensional data loss in complex communication environments. Although some systems have introduced error correction coding, their recovery capabilities are limited, and they can only repair a small number of consecutive errors, failing to effectively handle randomly lost or missing data across channels. At the same time, existing sensing systems generally lack adaptive recovery mechanisms based on data characteristics and channel states, failing to fully utilize the sparsity characteristics and multi-source correlation of sensing data for efficient reconstruction. This results in insufficient recovery accuracy and serious error accumulation, making it difficult to meet the requirements of smart distribution networks for real-time, stable, and highly reliable data recovery.

[0061] See Figure 1 To improve the accuracy and efficiency of data recovery in power distribution networks and ensure data stability, an embodiment of the present invention provides a data recovery method for power distribution networks, including steps S101 to S103.

[0062] Step S101: Obtain multi-dimensional sensing data of the power distribution network, wherein the multi-dimensional sensing data includes operational data and environmental data;

[0063] In some embodiments, a data source identification and type classification process is performed to determine the source type of each data and its corresponding physical attributes: Fiber optic sensor data identification: Fiber optic sensors are mainly used to collect operating parameters such as temperature, current, stress, and displacement. Their output signals typically manifest as continuously changing light intensity reflection values ​​or interference signals. By analyzing the reflection intensity distribution, spectral characteristics, and wavelength drift information of the fiber optic signal, this data type can be identified as fiber optic sensor data. Low-voltage carrier communication data identification: Low-voltage carrier communication terminals are used to collect electrical energy parameters such as voltage, current, and power factor. Their signals exhibit obvious periodic fluctuations over time, and their frequency is closely related to changes in the power grid load. By detecting carrier frequency characteristics, amplitude fluctuation range, and periodic characteristic values, this data source can be identified as low-voltage carrier communication data. Wireless sensor data identification: Wireless sensor nodes are mainly used to collect external environmental parameters such as ambient temperature, humidity, and air pressure. Their sampling results are mostly discrete data and are easily affected by environmental noise and communication fading. By parsing the device ID, channel frequency, and communication protocol type of the wireless terminal, this data type can be identified as wireless sensor data. The fiber optic sensing data and the low-voltage carrier communication data are classified as operational data, and the wireless sensing data is classified as environmental data.

[0064] Through the above steps, accurate identification and classification of multi-source sensing data are achieved, ensuring that different types of sensing data can be modeled in a targeted manner based on their physical properties and signal characteristics during subsequent fusion and processing, thereby improving the accuracy and reliability of multi-dimensional data processing in the power distribution network.

[0065] Step S102: Based on the channel state score results and the data level corresponding to each type of sensing data in the multidimensional sensing data, dynamically determine the channel allocation scheme corresponding to each data level, and use the multi-channel parallel transmission mechanism to transmit data according to the channel allocation scheme to form the data transmission result corresponding to each channel. The channel state score results are calculated from the state parameters of each channel.

[0066] In some embodiments, acquiring multidimensional sensing data of the distribution network includes: acquiring initial multidimensional sensing data of the distribution network, wherein the initial multidimensional sensing data includes initial operating data and initial environmental data; normalizing the initial operating data and the initial environmental data respectively to obtain operating normalization results and environmental normalization results; sparsely sampling the operating normalization results based on a preset compressed sensing sampling matrix to generate first compressed data; compressing the environmental normalization results using a differential coding method to generate second compressed data; and determining the multidimensional sensing data of the distribution network based on the first compressed data and the second compressed data.

[0067] In some embodiments, initial multidimensional sensing data of the distribution network is acquired, including initial operational data and initial environmental data. Specifically, this involves real-time acquisition of operational status and environmental parameters through multiple types of sensor nodes in the distribution network. Fiber optic sensors are used to collect operational parameters such as temperature, current, stress, and displacement; low-voltage carrier communication terminals are used to collect electrical energy parameters such as voltage, current, and power factor; and wireless sensor nodes are used to collect environmental information such as temperature, humidity, and air pressure. By identifying the device type, communication protocol, and signal characteristics of the acquisition terminals, automatic identification and classification of data source types are achieved: when the rate of change of light intensity and the interference spectrum characteristics of the detected fiber optic reflected signal meet a preset threshold range, it is determined to be fiber optic sensing operational data; when the detected signal has obvious periodic fluctuations and the carrier frequency is in the power frequency harmonic range, it is determined to be low-voltage carrier operational data; and when discrete data with a specific device ID and channel frequency is received, it is determined to be wireless sensing environmental data. The fiber optic sensing operational data and the low-voltage carrier operational data are classified as initial operational data, and the wireless sensing environmental data is classified as initial environmental data.

[0068] In some embodiments, the initial operating data and the initial environmental data are normalized to obtain operating normalization results and environmental normalization results, respectively. Specifically, the fiber optic sensing data is first normalized. Fiber optic sensing data mainly reflects continuous operating parameters of power distribution equipment such as temperature, current, stress, and displacement. Its original signal is related to the amplitude of light intensity reflection and is easily affected by environmental interference, resulting in significant amplitude differences. To eliminate dimensional differences and improve data comparability, the mean μ of the fiber optic sensing data is calculated. fiber With standard deviation σ fiber And according to the standard deviation normalization formula:

[0069]

[0070] Where, x fiber For raw data from fiber optic sensors, μ fiber and σ fiber x′ represents the mean and standard deviation of the fiber optic data, respectively. fiber The first step is to normalize the fiber optic data. Next, linear normalization is performed on the low-voltage carrier communication data. Low-voltage carrier communication data mainly includes periodic operating parameters such as voltage, current, and power, whose numerical ranges are fixed but fluctuate frequently. To suppress the impact of abnormal peak values ​​on the model calculation, the maximum value of the carrier communication data, max(x), is calculated. carrier ) and minimum value min(x) carrier And use the linear proportional normalization formula:

[0071]

[0072] Among them, x carrier is the carrier communication data, min(x carrier ) and max(x carrier ) are respectively the minimum and maximum values of the carrier communication data, and x′ carrier is the normalized carrier communication data. The normalized optical fiber data x′ fiber and the normalized carrier communication data x′ carrier are the operation normalization results, enabling the power parameters of different sampling nodes to be mapped to the interval [0, 1], improving the fusion consistency of cross-regional data. Again, standardize the environmental data collected by the wireless sensors. The wireless sensors mainly collect environmental parameters such as temperature, humidity, and air pressure. The data characteristics are easily affected by climate change and fluctuate greatly. To reduce the interference caused by environmental drift, calculate the mean μ wireless and standard deviation σ wireless of the environmental data, and according to the standardization formula:

[0073]

[0074] Among them, x wireless is the wireless sensor data, μ wireless and σ wireless are respectively the mean and standard deviation of the wireless sensor data, and x ′ wireless is the normalized wireless sensor data. Perform normalization on the wireless sensor data to obtain the environmental normalization result, enabling environmental features to participate in subsequent analysis and transmission under a unified scale.

[0075] In some embodiments, perform sparse sampling on the operation normalization result based on a preset compressive sensing sampling matrix to generate the first compressed data, and use the differential coding method to compress the environmental normalization result to generate the second compressed data. Specifically: Set the time series signal length of the optical fiber and carrier data to N, set the sampling length M according to the compression ratio requirement (where M < N), and construct a sparse matrix with dimensions M×N. The construction method of the sparse matrix can adopt the pseudo-random sparse generation algorithm. Its core idea is to make only some elements in the matrix non-zero under the premise of ensuring signal recoverability to achieve low-redundancy compression. Specifically, the element generation rule of matrix Φ is as follows:

[0076]

[0077] Among them, rand(i, j) represents a random number between 0 and 1, p is the sparse probability of non-zero elements (usually taking 0.1 - 0.3), ω i is the sampling weight coefficient, used to highlight the importance of key feature signals. Further, the weight coefficient ω iDynamic adjustments are made based on the characteristics of the distribution network operation data. For example, when a sudden change in current or stress is detected in the fiber optic sensing signal, the corresponding ω value in that interval is adjusted. i Increase ω to 1.5–2.0 to enhance the contribution of this part of the signal in compressed sampling; while for carrier communication data segments with relatively stable fluctuations, ω will be increased. i The value is set to 0.8–1.0 to balance the overall sampling energy. This "sparse + weighted" design reduces invalid sampling points while preserving key mutation information in the running data, ensuring high-precision recovery in the subsequent sparse reconstruction stage. After completing the sparse matrix construction, compressed sensing sampling is performed, the mathematical expression of which is:

[0078] y = Φx;

[0079] Where Φ is the sparse matrix after weighted design, x is the normalized signal vector formed by splicing fiber optic sensing data and carrier communication data, and y is the generated compressed sensing observation vector. Through this sampling process, the original high-dimensional signal can be reduced to a low-dimensional sparse representation, generating the first compressed data. This compression result significantly reduces the data volume and alleviates communication bandwidth consumption while maintaining the main feature information. Secondly, for the environmental data (i.e., the second compressed data), due to its slow change and long sampling period, a differential coding (Δ coding) method is used for lightweight compression. Specifically, the continuous environmental sampling data {x} is compressed... t ,x t-1 ,x t-2 ,…} are encoded as follows:

[0080] Δx t =x t -x t-1 ;

[0081] Where, Δx t For, x t For, x t-1 To reduce redundant storage, only the amount of data change is recorded, generating a second compressed data set. For example, fiber optic sensor data is compressed using a weighted sparse matrix of M=256 and N=1024, achieving a compression ratio of 1:4. Carrier communication data uses the same sparsity but with lower weights to achieve balanced compression. Environmental data is reduced by approximately 70% after differential coding.

[0082] In some embodiments, multi-dimensional sensing data of the distribution network is determined based on the first compressed data and the second compressed data. Specifically, the first compressed operating data is subjected to sparse reconstruction processing, and the corresponding current, voltage, power and node state information are recovered by matching pursuit (OMP) or least squares reconstruction algorithm; the second compressed environmental data is subjected to differential decoding and error correction recovery to extract environmental parameters such as temperature, humidity and external stress; then the recovered operating features and environmental features are time-synchronized and spatially mapped, and they are paired and fused according to the sensor acquisition timestamp and geographical location index to generate a multi-dimensional sensing data matrix, so as to realize the joint representation of the operating state and environmental state of the distribution network.

[0083] In some embodiments, dynamically determining the channel allocation scheme corresponding to each data level based on the channel state score results and the data levels corresponding to various types of sensing data in the multidimensional sensing data includes: comprehensively evaluating the data features corresponding to various types of sensing data in the multidimensional sensing data to determine the data levels corresponding to various types of sensing data; performing weighted normalization calculation on the state parameters of each channel to generate state score results corresponding to each channel; and, based on the data levels, the channel state score results, and preset weight coefficients, calling a preset data allocation model to allocate various types of sensing data to corresponding channels to obtain a channel allocation scheme, wherein the weight coefficients are calculated based on the degree of adaptation between various types of sensing data and each channel.

[0084] In some embodiments, the data characteristics corresponding to various types of sensing data in the multidimensional sensing data are comprehensively evaluated to determine the data level corresponding to each type of sensing data. Specifically, the following steps are taken: First, for multidimensional sensing data from fiber optic sensors, low-voltage carrier communication devices, and wireless sensor networks, timeliness indicators, volatility indicators, and historical anomaly frequency indicators are extracted respectively. The timeliness indicator reflects the data update frequency and latency sensitivity, the volatility indicator characterizes the dynamic change amplitude of the data, and the historical anomaly frequency is used to assess the number of times anomalies occurred in this type of data over a past period. Second, a comprehensive priority score P for each type of data is calculated based on the above indicators. data The calculation formula is as follows:

[0085] P data =α·T efficiency +β·V volatility +γ·F frequency ;

[0086] Among them, T efficiency V represents the timeliness score. volatility F represents the volatility score. frequencyThis represents the historical anomaly frequency score, with α, β, and γ as weighting coefficients to reflect the importance of different feature dimensions. For example, to emphasize the priority of real-time power data transmission, the weighting coefficients are set to α = 0.5, β = 0.3, and γ = 0.2. After calculation, fiber optic sensor data, due to its high timeliness and high accuracy, has a higher P... data Values ​​are usually higher than 0.8; low-voltage carrier communication data has a priority between 0.5 and 0.7 because the fluctuations are relatively gentle; wireless environmental monitoring data has a lower priority, generally below 0.4.

[0087] In some embodiments, the acquired state parameters of each channel are weighted and normalized to generate a state score result corresponding to each channel. Specifically, this involves: performing state detection on multiple available transmission channels in the communication system (including fiber optic communication channels, low-voltage carrier communication channels, and wireless communication channels), and collecting the bandwidth parameter C of each channel. i Delay parameter T i To comprehensively reflect the channel transmission capability, the channel status score calculation formula is set as follows:

[0088]

[0089] Where S is the channel state score, and C i Let T be the bandwidth of the i-th channel. i Let n be the transmission delay and n be the number of communication channels. A higher channel score indicates better channel performance and suitability for carrying high-priority data. After calculation, the status scores of all channels are linearly normalized to distribute their values ​​within the [0,1] interval for consistent comparison with data priority. For example, when the bandwidth of an optical fiber channel is 100Mbps, the delay is 2ms, and the bit error rate is 0.001, its status score is approximately 0.95; the status score of a low-voltage carrier communication channel is approximately 0.72; and the status score of a wireless channel, due to its high delay and low bandwidth, is approximately 0.41.

[0090] In some embodiments, based on the data level, the channel state scoring results, and preset weight coefficients, a preset data allocation model is invoked to allocate various types of sensing data to corresponding channels, resulting in a channel allocation scheme. The weight coefficients are calculated based on the degree of adaptation between various types of sensing data and each channel, specifically by constructing a data-channel adaptation matrix W. ij , used to characterize the matching degree between the i-th type of sensing data and the j-th channel. Matrix element W ij The value depends on the physical characteristics of the communication medium and the real-time requirements of the data type. For example, for fiber optic communication channels, due to their high bandwidth and low latency, they are suitable for high-priority, high-precision data, such as current, voltage, stress, and information on sudden changes in equipment status; the corresponding weight is set to W. fiber=0.9; For low-voltage carrier channels, suitable for power operation data with moderate real-time and reliability requirements, such as power factor and voltage fluctuation information, the corresponding weight is set to W. carrier =0.6; For wireless channels, adapting to environmental monitoring data with low real-time and low bandwidth requirements, such as temperature, humidity, and air pressure, the corresponding weight is set to W. wireless =0.4. Secondly, based on priority scoring P... data With channel state score S j The data allocation model is constructed as follows:

[0091]

[0092] Among them, D i W represents the allocation amount for the i-th type of data. ij The data type and channel matching degree, m is the total number of channels, P data For data priority, S j The state score of channel j is determined. During the initial channel allocation phase, various data types are allocated to the channels with the highest scores in descending order of priority. For example, when the fiber optic channel has the highest state score, high-priority data output from fiber optic sensors is prioritized; when the carrier channel state is moderate, medium-priority data such as current and voltage are allocated; and when the wireless channel state is poor, it is used for low-priority environmental data transmission. After obtaining the initial allocation scheme, the adaptation matrix W is used... ij The allocation results for each data category are adjusted. When uneven channel load or fluctuations in the state of a channel are detected, the algorithm dynamically corrects the allocation ratio. For example, when high interference occurs in the wireless channel, causing the state score S to drop... wireless When the priority is reduced to 0.3, the allocation of low-priority data will be automatically reduced, and some medium-priority data will be temporarily switched to carrier channels to maintain overall transmission stability. Finally, a target channel allocation scheme is generated, ensuring that high-priority data is always allocated to high-quality channels, and medium- and low-priority data are allocated to the remaining channels, thereby achieving optimal resource utilization and real-time performance in bandwidth-constrained multi-media communication environments.

[0093] Please refer to Figure 2 In some embodiments, the process of using a multi-channel parallel transmission mechanism to transmit data according to the channel allocation scheme and forming data transmission results corresponding to each channel includes steps S201 to S203.

[0094] Step S201: Use a multi-channel parallel transmission mechanism to transmit various types of sensing data in the channel allocation scheme in parallel to generate initial data transmission results;

[0095] In some embodiments, according to a channel allocation scheme, different types of sensing data are divided into several parallel data streams, including fiber optic channel transmission streams, low-voltage carrier channel transmission streams, and wireless channel transmission streams. Each data stream corresponds to different types of data characteristics and priority levels. For example, the fiber optic channel carries high-priority physical quantity data (such as current, voltage, stress, etc.), requiring high bandwidth and low latency; the low-voltage carrier channel transmits power operation parameters (such as power, phase, power quality, etc.), with moderate real-time requirements; and the wireless channel carries environmental monitoring data (such as temperature, humidity, air pressure, etc.), allowing for a certain delay and error range. To improve the robustness of data transmission, during the data distribution phase, each data packet is assigned a redundancy number based on a channel state score and transmitted in parallel according to a redundancy strategy. The core of the redundancy strategy is based on the real-time packet loss rate of each channel. i Dynamically allocating data load enables reconstruction and recovery based on redundant data from other channels when packet loss occurs on any channel. In practice, parallel transmission and recovery are performed according to the following model:

[0096]

[0097] Among them, D recovered For the recovered data, loss i Let D be the packet loss rate of the i-th channel. i This refers to data transmitted through the i-th channel. During transmission, the packet loss rate and delay changes of each channel are periodically monitored. When the packet loss rate of a channel exceeds a preset threshold (e.g., 5%), the redundancy coefficient of the corresponding data on that channel is automatically increased, and the load ratio of other channels is simultaneously reduced to balance the total bandwidth usage and ensure recovery accuracy. During this process, a multi-channel error correction and verification mechanism is also used to compare and verify all transmitted data. When a partial data block is detected as missing, data copies from channels with low packet loss rates are preferentially used for differential recovery. If the recovery accuracy is still insufficient, a cross-channel joint recovery algorithm is triggered to reconstruct the lost data fragments by comparing the timestamps, coding features, and sequence numbers of the data received from multiple channels, forming the initial data transmission result.

[0098] Step S202: Real-time acquisition of feedback information from each channel during data transmission; construction of a channel feedback optimization model based on the feedback information; calculation of the state adjustment amount of each channel based on the channel feedback optimization model; and dynamic adjustment of the usage priority and resource allocation ratio of each channel according to the state adjustment amount.

[0099] In some embodiments, after each round of data transmission, real-time feedback parameters from each channel are collected, including bandwidth feedback. j): Represents the available bandwidth of the current channel, reflecting the channel's instantaneous transmission capability; delay feedback (feedback_delay) j ): Represents the average delay from data packet transmission to reception, used to evaluate the real-time performance of the channel; error rate feedback (error) j ): This represents the bit error rate or packet error rate in the transmitted data, used to assess channel stability. Based on the above feedback information, a channel feedback optimization model is constructed:

[0100] ΔS j =α·feedback j +β·error j ;

[0101] Where, ΔS j Represents the state adjustment amount of the j-th channel, feedback j For channel integration feedback information (including bandwidth and delay), error j The error rate is represented by α and β, which are weighting coefficients used to balance the importance of channel performance and stability. In practical applications: for fiber optic channels, due to their high bandwidth and low latency, bandwidth and transmission efficiency are more important, so α = 0.7 and β = 0.3 are set; for wireless channels, due to greater susceptibility to interference, the error rate weight β is increased accordingly (e.g., α = 0.5, β = 0.5) to enhance the focus on channel reliability. After the channel feedback optimization model is calculated, based on ΔS... j The adjustment results are used to update channel usage priorities and bandwidth allocation in real time. For example, when the bandwidth of a channel decreases or the error rate increases, its status score S is updated. j The bandwidth weight will be dynamically reduced, and the corresponding data traffic will be transferred to channels with higher status scores. If a channel recovers stability, its bandwidth weight will be gradually increased to achieve adaptive optimization. In addition, during the channel optimization process, the system also introduces a periodic balancing mechanism, that is, every preset time window (e.g., 10 seconds), the average feedback value of all channels is re-aggregated, and the weights of α and β are re-adjusted to adapt to the dynamically changing communication environment and prevent channel resource allocation imbalance caused by short-term fluctuations.

[0102] Step S203: Apply the usage priority and the resource allocation ratio to the initial data transmission result to form the data transmission result corresponding to each channel.

[0103] In some embodiments, the application of priority and resource allocation ratio includes: remapping the initial transmitted data to each channel according to the state adjustment amount and dynamic priority of each channel, and adjusting the number of data fragments and the redundancy allocation strategy. For example, high-priority electrical parameter data is sent in full on the main channel, while partially redundant copies are sent on low-priority channels; environmental monitoring data is sent in full or partially redundant on low-bandwidth channels according to the allocation ratio. Feedback information is continuously monitored during transmission, and the data scheduling strategy is updated in real time to ensure that the data transmission results of each channel meet both timeliness requirements and loss tolerance and system robustness. Finally, the data transmission results corresponding to each channel can be used for subsequent data fusion, compression reconstruction, and multi-dimensional state analysis of the distribution network.

[0104] Through the above steps, the priority levels of multi-dimensional sensing data and the scoring results of each channel status are comprehensively analyzed to dynamically generate the optimal channel allocation scheme. The multi-channel parallel transmission and redundancy recovery mechanism is adopted to achieve efficient and reliable data transmission, thereby achieving optimal resource utilization and real-time guarantee in a bandwidth-constrained multi-media communication environment.

[0105] Step S103: Redundant encoding is performed on the data transmission result to generate redundant encoded data. The redundant encoded data is then input into the sparse signal recovery model to solve the sparse signal recovery model and obtain the recovery data of the distribution network. The sparse signal recovery model is calculated based on the linear mapping relationship between historical recovery data and historical data transmission results.

[0106] Please refer to Figure 3 In some embodiments, the step of performing redundant encoding on the data transmission result to generate redundant encoded data includes steps S301 to S302.

[0107] Step S301: Redundancy encoding is performed on the data transmission result to obtain a generator matrix and a redundancy matrix, and redundancy symbols are determined based on the generator matrix and the redundancy matrix.

[0108] In some embodiments, the RS code parameters (n,k) are selected based on the required error correction capability and transmission overhead, where n is the code length (total number of symbols), k is the number of information symbols, and the number of redundant symbols is r = nk. The RS code operates in the finite field GF(2^n). m In this context, m is determined by the bit width of a single symbol (e.g., 1 byte per symbol when m = 8). The error correction capability is... Select the finite field GF(2) m And specify the primitive polynomial p(x) (for example, in this embodiment, GF(2) is selected). 3 If x is taken, then the primitive polynomial takes x 3+x+1). Construct the field elements and the rules of addition and multiplication according to the primitive polynomial, and determine the primitive element α of the field. Select n pairwise distinct evaluation points {α}. 0 ,α 1 ,…,α n-1 As evaluation points for RS codes, construct a generator matrix G of Vandermonde form or systematize it into a systematized generator matrix G. sys =[I k |P]: First construct a k×n Vandermonde matrix (rows correspond to information symbol counts from 0 to k-1, columns correspond to evaluation points):

[0109] G V (i,j)=(α j-1 ) i-1 ,i=1…k,j=1…n;

[0110] For G V Perform column transformations or Gaussian elimination to obtain the systematic form, which consists of a first k-column matrix A and a last r-column matrix B. Calculate P = A. -1 B, yielding the systematic generation matrix:

[0111] G sys =[I k |P];

[0112] G sys The P matrix in the matrix is ​​transposed or formalized as needed to generate a redundant matrix E (in some implementations, E = P). T The encoding relationship is satisfied:

[0113] C=G sys ·D=[I k |P]·D;

[0114] Where, D∈GF(2 m ) k×1 Given an information column vector, the generated redundant symbol column vector p = P·D, and the encoded codeword C = [D] T p T ] T .

[0115] Step S302: Combine the redundant symbols with the corresponding original data symbols in the data transmission result to generate redundant coded data.

[0116] In some embodiments, the obtained information symbol D and redundant symbol p are combined in a systematic manner to form a complete codeword C:

[0117] C = [c1, c2, ..., c k ,c k+1 ,…,c n=[D1,…,D k ,p1,…,p r ;

[0118] Among them, the first k symbols are original data symbols, and the last r symbols are redundant check symbols. This combination is the encoded redundant encoded data and serves as the encoded block to be transmitted. To adapt to the underlying communication protocol, the codeword C is sliced according to the selected transmission unit (for example, in terms of the number of symbols or bytes). Each slice is appended with metadata fields, including: slice sequence number, block number (blockid), the length n of the codeword to which it belongs, the information length k, the identification of the finite field type GF(2 m ), timestamp, and checksum (CRC). This metadata facilitates recombination and verification at the receiving end. According to the previous channel allocation scheme and channel status, different symbols or redundant copies of the codeword are allocated to different channels for parallel transmission: for high-priority codewords, a complete copy of the codeword can be sent to the primary channel (such as an optical fiber) if the bandwidth permits, and sliced copies of the redundant symbols are sent to the backup channel (such as a carrier or wireless); for medium- and low-priority codewords, only a single copy of the codeword can be sent and some redundant symbols are transmitted through a low-bandwidth channel. At the sending end, integrity verification (CRC) is performed on the constructed redundant encoded data, and the sending record and the corresponding codeword and metadata are stored in the retransmission buffer, facilitating subsequent triggering of retransmission or feedback-based recovery strategies. After receiving the slices, the receiving end reconstructs the codeword based on the metadata. If some symbols are missing, the RS decoder (such as the Berlekamp-Massey algorithm or the Euclidean algorithm) is called to recover the missing symbols based on polynomial root finding and field operations; if an interleaving strategy is applied, the received sequence is first deinterleaved and then decoded.

[0119] Please refer to Figure 4 , in some embodiments, the construction process of the sparse signal recovery model includes steps S401 to step S402;

[0120] Step S401, linearly map the historical redundant encoded data through a preset measurement matrix to obtain observed data, and input the observed data and the measurement matrix into the sparse signal recovery model to iteratively optimize the sparse signal using a preset sparse recovery algorithm to obtain a reconstructed signal, where the sparse signal is obtained by sparsely transforming the historical redundant encoded data;

[0121] In some embodiments, the measurement matrix is an M×N order matrix, denoted as Φ, where M < N, and is used to implement compressive sensing measurement mapping. Specifically, first, for the redundant encoded data D c =[d1,d2,…,d N T ​Sparse transformation is performed by using wavelet transform or discrete cosine transform to map the original signal to the sparse domain, resulting in a sparse signal vector x = [x1, x2, ..., xn]. N ] T In the sparse domain, the signal has only a few coefficients that are significantly non-zero, thus exhibiting sparsity characteristics. Subsequently, the sparse signal is linearly mapped using the measurement matrix Φ to obtain compressed observation data:

[0122] y = Φx;

[0123] Where y represents the observed data, indicating the low-dimensional measurement results obtained under compressed sampling conditions. The elements of the measurement matrix Φ can be randomly generated and satisfy the linear independence between column vectors to ensure the invertibility and stability of the reconstruction. For example, in a specific example, the original signal dimension is N=4, and the selected measurement matrix is:

[0124]

[0125] When redundant coded data undergoes a sparse transformation, the result is x = [1, 0, 3, 0]. T The observation data are then obtained through the above measurement matrix:

[0126] y = Φx = [4, 3] T ;

[0127] Therefore, an effective compressed representation of the original four-dimensional sparse signal can be achieved using only two observation data points. During the sparse signal recovery process, the sparse signal recovery model employs a joint optimization strategy based on least squares and L1 regularization, reconstructing the signal by iteratively solving the following objective function:

[0128]

[0129] Where λ is the regularization coefficient, used to balance reconstruction error and sparsity constraints; λ represents the sum of squared observation errors; ||x||1 represents the L1 norm penalty term, used to maintain the sparsity of the signal. In a specific embodiment, λ is taken between 0.01 and 0.05, and iterative updates are performed using gradient descent or coordinate descent methods to gradually reduce the observation errors. When the rate of change of error is less than 10... -5 When convergence is considered complete, the optimal output solution is the reconstructed signal.

[0130] Step S402: By minimizing the error between the reconstructed signal and the observed data, a trained sparse signal recovery model is obtained when the threshold condition of the preset non-zero element quantity limit is met.

[0131] In some embodiments, a threshold value of k for the number of non-zero elements is set, meaning that a maximum of k non-zero components are allowed in the sparse signal. The reconstruction optimization satisfies the following constraints:

[0132]

[0133] Where ||x||0 represents the number of non-zero components in the signal. During the iteration process, the number of non-zero components of the current sparse signal is counted in real time. When ||x||0≤k and the observation error is... (where ε is the preset error tolerance, such as 10) -4 When the reconstruction process converges, the reconstructed signal is considered complete. After inverse sparse transformation (such as inverse wavelet transform or inverse discrete cosine transform) to restore it to the original domain, the recovered data D of the distribution network is obtained. r This data represents the high-fidelity observation results recovered under conditions of packet loss, interference, or channel interruption. It can be directly input into subsequent monitoring and analysis modules to achieve complete reconstruction of current, voltage, stress, or environmental quantities. In one specific embodiment, to address the different sparsity characteristics of different types of signals (such as electrical quantities, fiber strain, and environmental parameters), the transform domain can be automatically selected according to the signal type: time-domain sparse modeling is used for abrupt signals such as current and voltage; spatial-domain sparse reconstruction is used for fiber strain signals; and wavelet-domain sparse representation is used for slowly varying signals such as temperature and humidity. Through adaptive sparse domain selection and hierarchical reconstruction mechanisms, the recovery accuracy and system fault tolerance of multimodal data in the distribution network can be further improved.

[0134] In some embodiments, after obtaining the recovery data of the distribution network, the method further includes: after receiving the recovery data, the receiving end decodes the recovery data using a polynomial interpolation-based error correction decoding algorithm to obtain first recovery data, and reconstructs the first recovery data using an L1 regularization-based sparse recovery algorithm to obtain second recovery data; calculates the probability of equality between the second recovery data and the expected dataset, and determines the target recovery data of the distribution network when the probability of equality is greater than or equal to a preset consistency threshold.

[0135] In some embodiments, after receiving the recovered data, the receiving end decodes the recovered data using a polynomial interpolation-based error correction decoding algorithm to obtain first recovered data, and then reconstructs the first recovered data using an L1 regularized sparse recovery algorithm to obtain second recovered data. Specifically, the receiving end first analyzes the received compressed and redundant encoded data to determine the original data symbols and redundant symbol sets contained therein; when a missing or corrupted data packet is detected, the receiving end calls the error correction decoding module based on the Reed-Solomon coding mechanism. The error correction decoding module is implemented through the following steps: Error detection: Calculate the check polynomial for the recovered data to determine whether there are erroneous or missing symbols; Error location: If an error is detected, construct a set of check equations for the received symbols using a polynomial interpolation method, and solve the error location polynomial based on the Berlekamp-Massey algorithm or the Euclidean algorithm to determine the location of the erroneous symbol; Error correction: Calculate the correction value of the erroneous symbol based on the location result and the redundant symbol set, and replace the data symbol at the corresponding position to obtain the first recovered data after error correction. Through the above steps, the receiver can effectively detect and correct data based on redundant symbols, recovering a data sequence close to the original state, even when some data is lost or corrupted. The receiver further reconstructs the signal from the first recovered data using a sparse recovery algorithm based on L1 regularization constraints to obtain the second recovered data. Specifically, the receiver constructs a sparse reconstruction model based on a preset measurement matrix Φ and observation data y, solving the following optimization problem:

[0136]

[0137] Where x represents the target sparse signal, y represents the observation vector extracted from the first recovered data, Φ represents the measurement matrix, and λ represents the regularization coefficient, used to balance reconstruction accuracy and sparsity constraints. The receiver iteratively minimizes the above objective function, employing either an Iterative Shrinkage-Thresholding Algorithm (ISTA) or a Basis Pursuit (BP) method to progressively optimize the sparse signal until the reconstruction error converges below a preset threshold. Ultimately, the obtained second recovered data can accurately reconstruct the main features of the original signal even with a high data loss rate or incomplete observations, achieving high-fidelity data reconstruction.

[0138] In some embodiments, the probability of equality between the second restored data and the expected dataset is calculated. When the probability of equality is greater than or equal to a preset consistency threshold, the target restored data for the distribution network is determined. Specifically, the receiving end calculates the probability P(D) of equality between the second restored data and the expected dataset. reconstructed=D expected ), and compare it with the preset consistency threshold θ; if the condition P(D) is satisfied, reconstructed =D expected If the threshold value is greater than or equal to θ, the second recovered data is deemed valid and identified as the target recovered data for the distribution network. If the threshold condition is not met, the sparse reconstruction step is re-executed or the redundant data compensation mechanism is triggered until the consistency of the recovered data meets the requirements. The consistency threshold θ can be adjusted according to the application scenario, data type, and system fault tolerance: when the data belongs to key operating parameters such as current and voltage, θ can be set to 0.95 or higher to ensure high-precision recovery; when the data is environmental monitoring information (such as temperature, humidity, stress, etc.), θ can be set between 0.85 and 0.9 to achieve a balance between accuracy and recovery efficiency.

[0139] Through the above steps, combined with redundant error correction coding and sparse signal reconstruction, highly reliable transmission and high-precision recovery of multi-source sensing data in complex channel environments are achieved, effectively improving the anti-interference, fault tolerance and reconstruction accuracy of distribution network data, and ensuring the accuracy of subsequent monitoring and analysis.

[0140] like Figure 5 As shown, based on the above method embodiments, corresponding apparatus embodiments are provided;

[0141] An embodiment of the present invention provides a schematic diagram of the structure of a data recovery system for a power distribution network, including: an acquisition module 100, a transmission module 200, and a recovery module 300;

[0142] The acquisition module 100 is used to acquire multi-dimensional sensing data of the power distribution network, wherein the multi-dimensional sensing data includes operational data and environmental data;

[0143] The transmission module 200 is used to dynamically determine the channel allocation scheme corresponding to each data level based on the channel state score results and the data level corresponding to each type of sensing data in the multidimensional sensing data, and to perform data transmission according to the channel allocation scheme using a multi-channel parallel transmission mechanism to form data transmission results corresponding to each channel. The channel state score results are calculated from the state parameters of each channel.

[0144] The recovery module 300 is used to perform redundant encoding on the data transmission results to generate redundant encoded data, and input the redundant encoded data into the sparse signal recovery model to solve the sparse signal recovery model to obtain the recovery data of the distribution network. The sparse signal recovery model is calculated based on the linear mapping relationship between historical recovery data and historical data transmission results.

[0145] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention, and can implement the data recovery method for a power distribution network provided by any of the above-described method embodiments of the present invention. More detailed workflows and principles of this system can be found, but are not limited to, the relevant descriptions of the above methods.

[0146] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can specifically be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0147] Based on the above-described embodiment of a data recovery method for a power distribution network, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements a data recovery method for a power distribution network according to any embodiment of the present invention.

[0148] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0149] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0150] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0151] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute a data recovery method for a power distribution network as described in any of the above-described method embodiments of the present invention.

[0152] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0153] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A data recovery method for a power distribution network, characterized in that, include: Acquire multi-dimensional sensing data of the power distribution network, wherein the multi-dimensional sensing data includes operational data and environmental data; Based on the channel state score results and the data level corresponding to each type of sensing data in the multidimensional sensing data, the channel allocation scheme corresponding to each data level is dynamically determined, and the data is transmitted according to the channel allocation scheme using a multi-channel parallel transmission mechanism to form the data transmission result corresponding to each channel. The channel state score results are calculated from the state parameters of each channel. The data transmission results are redundantly encoded to generate redundant encoded data. The redundant encoded data is then input into a sparse signal recovery model to solve the sparse signal recovery model and obtain the recovery data of the distribution network. The sparse signal recovery model is calculated based on a linear mapping relationship between historical recovery data and historical data transmission results.

2. The data recovery method for a power distribution network as described in claim 1, characterized in that, The acquisition of multi-dimensional sensing data of the distribution network includes: Acquire initial multidimensional sensing data of the power distribution network, wherein the initial multidimensional sensing data includes initial operating data and initial environmental data; The initial running data and the initial environment data are normalized respectively to obtain running normalization results and environment normalization results; Based on a preset compressed sensing sampling matrix, the normalized running result is sparsely sampled to generate first compressed data. The environment normalized result is then compressed using a differential coding method to generate second compressed data. Multidimensional sensing data of the distribution network is determined based on the first compressed data and the second compressed data.

3. The data recovery method for a power distribution network as described in claim 1, characterized in that, The dynamic determination of the channel allocation scheme corresponding to each data level based on the channel state scoring results and the data levels corresponding to various types of sensing data in the multidimensional sensing data includes: A comprehensive evaluation is performed on the data features corresponding to various types of sensing data in the multidimensional sensing data to determine the data level corresponding to each type of sensing data. The state parameters of each channel are obtained and weighted and normalized to generate the state score results corresponding to each channel. Based on the data level, the channel state score results, and the preset weight coefficient, a preset data allocation model is invoked to allocate various types of sensing data to the corresponding channels to obtain a channel allocation scheme. The weight coefficient is calculated based on the degree of compatibility between various types of sensing data and each channel.

4. The data recovery method for a power distribution network as described in claim 1, characterized in that, The process of using a multi-channel parallel transmission mechanism to transmit data according to the channel allocation scheme, forming data transmission results corresponding to each channel, includes: The various types of sensing data in the channel allocation scheme are transmitted in parallel using a multi-channel parallel transmission mechanism to generate initial data transmission results. The system acquires feedback information from each channel during data transmission in real time, constructs a channel feedback optimization model based on the feedback information, calculates the state adjustment amount of each channel based on the channel feedback optimization model, and dynamically adjusts the usage priority and resource allocation ratio of each channel according to the state adjustment amount. The usage priority and the resource allocation ratio are applied to the initial data transmission results to form the data transmission results corresponding to each channel.

5. The data recovery method for a power distribution network as described in claim 1, characterized in that, The step of performing redundant encoding on the data transmission result to generate redundant encoded data includes: The data transmission results are redundantly encoded to obtain a generator matrix and a redundancy matrix, and redundancy symbols are determined based on the generator matrix and the redundancy matrix. The redundant symbols are combined with the corresponding original data symbols in the data transmission result to generate redundant coded data.

6. The data recovery method for a power distribution network as described in claim 1, characterized in that, The construction process of the sparse signal recovery model includes: Historical redundant coded data is linearly mapped using a preset measurement matrix to obtain observation data. The observation data and the measurement matrix are then input into the sparse signal recovery model to iteratively optimize the sparse signal using a preset sparse recovery algorithm to obtain a reconstructed signal. The sparse signal is obtained by sparse transformation of the historical redundant coded data. By minimizing the error between the reconstructed signal and the observed data, a well-trained sparse signal recovery model is obtained when a threshold condition limiting the number of non-zero elements is met.

7. A data recovery method for a power distribution network as described in any one of claims 1-6, characterized in that, After obtaining the power distribution network recovery data, the following is also included: After receiving the recovered data, the receiving end decodes the recovered data using a polynomial interpolation-based error correction decoding algorithm to obtain the first recovered data, and then reconstructs the first recovered data using an L1 regularization-based sparse recovery algorithm to obtain the second recovered data. Calculate the probability of equality between the second restored data and the expected dataset. When the probability of equality is greater than or equal to a preset consistency threshold, determine the target restored data for the distribution network.

8. A data recovery system for a power distribution network, characterized in that, The system includes: an acquisition module, a transmission module, and a recovery module; The acquisition module is used to acquire multi-dimensional sensing data of the power distribution network, wherein the multi-dimensional sensing data includes operational data and environmental data; The transmission module is used to dynamically determine the channel allocation scheme corresponding to each data level based on the channel state score results and the data level corresponding to each type of sensing data in the multidimensional sensing data, and to perform data transmission according to the channel allocation scheme using a multi-channel parallel transmission mechanism to form the data transmission results corresponding to each channel. The channel state score results are calculated from the state parameters of each channel. The recovery module is used to perform redundant encoding on the data transmission results to generate redundant encoded data, and input the redundant encoded data into the sparse signal recovery model to solve the sparse signal recovery model to obtain the recovery data of the distribution network. The sparse signal recovery model is calculated based on the linear mapping relationship between historical recovery data and historical data transmission results.

9. A terminal device, characterized in that, The device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements a data recovery method for a power distribution network as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, include: A stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform a data recovery method for a power distribution network as described in any one of claims 1-7.

Images