阅读说明：本技术 基于边缘计算的电力系统稳态数据压缩方法 (Power system steady-state data compression method based on edge calculation ) 是由 �田�浩 叶小晖 赵二岗 王可庆 于 2021-07-30 设计创作，主要内容包括：本发明涉及一种基于边缘计算的电力系统稳态数据压缩方法,通过建立联合稀疏模型、建立稀疏冗余字典、明确测量矩阵、建立联合重构算法等阶段,结合压缩感知与分布式信源编码,完成边缘计算下电力系统稳态数据融合；利用小波变换算法,按分辨率将得到的数据融合结果分解至各个尺度水平上,得到高频系数与低频系数,经阈值处理高频系数后,采用无损编码技术输出压缩结果。本发明结合边缘计算方法,构建出稳态数据压缩方法,有效压缩电力系统数据,缩减储存空间与数据传输量。(The invention relates to a power system steady-state data compression method based on edge calculation, which combines compressed sensing and distributed source coding to complete power system steady-state data fusion under edge calculation by stages of establishing a joint sparse model, establishing a sparse redundant dictionary, determining a measurement matrix, establishing a joint reconstruction algorithm and the like; decomposing the obtained data fusion result to each scale level according to resolution by using a wavelet transform algorithm to obtain a high-frequency coefficient and a low-frequency coefficient, and outputting a compression result by adopting a lossless coding technology after the high-frequency coefficient is subjected to threshold processing. The invention combines the edge calculation method to construct a steady-state data compression method, effectively compresses the data of the power system, and reduces the storage space and the data transmission quantity.)

1. A power system steady-state data compression method based on edge calculation comprises the following steps:

s1, establishing a joint sparse model;

assume that the initial steady-state data signal is x_jThe common part and the innovation part are respectively z_c、z_jThen signal x_jSparse representation of (c) is as follows:

x_j＝z_c+z_j＝z_j＝ψθ_j

wherein the sparse matrix and the corresponding sparse coefficients are psi, theta_j；

S2, creating a sparse dictionary;

regarding the sparse decomposition stage, after the known initial sparse dictionary D is solved by the orthogonal matching pursuit algorithm to obtain sparse representation, the sparse representation coefficient shown in the following formula is obtained

The constraints are as follows:

wherein y represents a reconstructed steady-state data signal, a preset threshold is epsilon, N represents the number of atoms, and F represents a sparse coefficient matrix norm;

in the cyclic calculation during dictionary updating, the dictionary training algorithm only updates one atom at a time; when new atom d is fetched_kThen, the following equation holds:

in the above formula, the dilution coefficients with row numbers j and k in the sparse coefficient matrix X are respectivelyd_jThe expression ordinal number j is an atom;

if d is removed_kThe other atoms being offset byRewriting the above expression as the following expression:

suppose atom d_kIndex of the reconstructed signal of omega_k，N*ω_kIs omega_kIf divide by (ω)_k(i) I) the matrix elements other than the non-zero values are all zero values, the following expression is derived from the above expression,

index ω in which result divergence is avoided_kThe expression is shown as the following formula:

in the above formula, the k-th line is thinned outAfter zero value items in the row vector are removed, a row vector is obtainedNamely, it isAtom d within sparse coding phase_kThe deviation sequence isNamely, it is

Decomposing the deviation column by singular value decomposition strategyThe following decomposition expression was obtained:

wherein, the two orthogonal matrixes are U, V respectively, the diagonal matrix is delta, the first column of the two orthogonal matrixes U, V is obtained by decomposition, and the former is used for completing the atom d in the initial dictionary_kAfter multiplying the latter by the diagonal matrix Δ (1,1), the update and the replacement x are updated by the resulting product_jFurther acquiring a new sparse dictionary;

s3, establishing a measurement matrix;

a Gaussian measurement matrix is constructed, the dimension of a signal represented by the sparse matrix is reduced, and meanwhile, the accuracy of a reconstructed signal and the constraint isometric condition are ensured to be satisfied;

s4, establishing a joint reconstruction algorithm;

establishing a joint reconstruction algorithm by fusing a synchronous orthogonal matching tracking algorithm and a learning training algorithm; firstly, reconstructing the acquired steady-state data by using the former algorithm, and then updating the sparse dictionary by using the latter algorithm; the algorithm operation flow is described in detail as follows:

s4-1, initializing and processing relevant parameters of the joint reconstruction algorithm; for the initial residual r₀Its residual r corresponding to the p-th node_pAre in equal relation; index value omega_k0; index set Λ₀Is an empty set;

s4-2, and converting the initial signal matrix X_n*sInitial dictionary psi_n*nMeasurement matrix phi_m*nMinimum reconstructed signal-to-noise ratio SNR_defAs an input item, where s is the number of nodes, and p ═ 1,2,. multidata, s }, the data length is n, and the number of measurements is m;

s4-3, establishing a sensing matrix to obtain the following expression:

A_m*n＝ψ_m*n*Φ_n*n

s4-4, solving each row residual error r by adopting the following formula_pWith each row of sensing matrix A_qTwo norm sum between:

according to the obtained maximum value of the two norm sum, the corresponding sensing matrix row index is reserved, and after the sensing matrix row index is fused with the index set, a new index set is obtained, as follows:

Λ_τ＝[Λ_τ-1 ξ_p]

s4-5, after solving the relevant parameters by the least square algorithm, updating the residual error by using the following expression:

s4-6, respectively solving the reconstructed intermediate signal and the relative root mean square error and the reconstructed signal-to-noise ratio thereof by adopting the following calculation formulas:

finally, the lowest reconstructed SNR is compared_defWhen the SNR is smaller, updating dictionary atoms by using a dictionary training algorithm, and returning to S4-4; otherwise, an output result, namely a reconstruction result x is obtained_j' with dictionary atoms used;

s5, compressing the steady-state data of the power system under the wavelet transformation;

known even sequence e_j+1Odd sequence o_j+1Then the wavelet decomposition process is described using the following expression:

split(x′_j)＝(e_j+1,o_j+1)

in the above formula, the even number sequence e_j+1＝a_j+1-U(b_j+1) Odd sequence o_j+1＝b_j+1+P(a_j+1) (ii) a Wherein, a_j+1And b_j+1Respectively representing low and high frequency coefficients in the sequence, Y (b)_j+1) And P (a)_j+1) Respectively representing the updating result of the high-frequency coefficient and the prediction result of the low-frequency coefficient;

from this, a compressed reconstructed signal representation is derived, as follows:

x″_j＝merge(e_j+1,o_j+1)

wherein merge represents a merge sort algorithm.

2. The method for compressing steady-state data of the power system based on the edge calculation as claimed in claim 1, wherein: in S3, there is a constant δ in the range of 0 to 1_kFor all sparse coefficient matrices X, the following inequality holds for the measurement matrix Φ:

3. a compression effect evaluation method of a power system steady-state data compression method based on edge calculation is characterized by comprising the following steps: connecting an electric energy acquisition device to the electric interface to acquire steady-state data of a research object; carrying out quantitative evaluation on three indexes of the space ratio of data compression, the normalized mean square error and the data compression ratio; wherein the content of the first and second substances,

Technical Field

The invention relates to a power system steady-state data compression method based on edge calculation. Belongs to the technical field of power grids.

Background

The national economy is continuously rising, the scale of the power grid is increasingly large, and the power system is gradually developed towards diversification and complication. The power management and control requirements such as data analysis, fault monitoring and wide-area measurement can be met by effectively recording mass power data. Therefore, if a large-scale data generated during the operation of the power system is stored and transmitted, the operation rate and the storage space burden are greatly increased, and even the intelligent development of the power grid in a crossing manner is hindered.

The power system is a key component of the power grid enterprise. The society and science and technology are rapidly developed, the intelligent degree and the informatization degree are increasingly deepened, along with the rapid improvement of a power grid communication technology, the data transmission scale is gradually strengthened, and the data compression is promoted to gradually develop into one of the hot research subjects in the electric power field.

Disclosure of Invention

The invention aims to overcome the defects and provides a power system steady-state data compression method based on edge calculation.

The purpose of the invention is realized as follows:

a steady-state data compression method of a power system based on edge calculation is characterized by comprising the following steps: the method comprises the following steps:

s1, establishing a joint sparse model;

assume that the initial steady-state data signal is x_jThe common part and the innovation part are respectively z_c、z_jThen signal x_jSparse representation of (c) is as follows:

x_j＝z_c+z_j＝z_j＝ψθ_j

wherein the sparse matrix and the corresponding sparse coefficients are psi, theta_j；

S2, creating a sparse dictionary;

regarding the sparse decomposition stage, after the known initial sparse dictionary D is solved by the orthogonal matching pursuit algorithm to obtain sparse representation, the sparse representation coefficient shown in the following formula is obtained

The constraints are as follows:

wherein y represents a reconstructed steady-state data signal, a preset threshold is epsilon, N represents the number of atoms, and F represents a sparse coefficient matrix norm;

in the cyclic calculation during dictionary updating, the dictionary training algorithm only updates one atom at a time; when new atom d is fetched_kThen, the following equation holds:

in the above formula, the dilution coefficients with row numbers j and k in the sparse coefficient matrix X are respectivelyd_jThe expression ordinal number j is an atom;

if d is removed_kThe other atoms being offset byRewriting the above expression as the following expression:

suppose atom d_kIndex of the reconstructed signal of omega_k，N*ω_kIs omega_kIf divide by (ω)_k(i) I) the matrix elements other than the non-zero values are all zero values, the following expression is derived from the above expression,

index ω in which result divergence is avoided_kThe expression is shown as the following formula:

in the above formula, the k-th line is thinned outAfter zero value items in the row vector are removed, a row vector is obtainedNamely, it isAtom d within sparse coding phase_kThe deviation sequence isNamely, it is

Decomposing the deviation column by singular value decomposition strategyThe following decomposition expression was obtained:

wherein, the two orthogonal matrixes are U, V respectively, the diagonal matrix is delta, the first column of the two orthogonal matrixes U, V is obtained by decomposition, and the former is used for completing the atom d in the initial dictionary_kAfter multiplying the latter by the diagonal matrix Δ (1,1), the update and the replacement x are updated by the resulting product_jFurther acquiring a new sparse dictionary;

s3, establishing a measurement matrix;

a Gaussian measurement matrix is constructed, the dimension of a signal represented by the sparse matrix is reduced, and meanwhile, the accuracy of a reconstructed signal and the constraint isometric condition are ensured to be satisfied;

s4, establishing a joint reconstruction algorithm;

establishing a joint reconstruction algorithm by fusing a synchronous orthogonal matching tracking algorithm and a learning training algorithm; firstly, reconstructing the acquired steady-state data by using the former algorithm, and then updating the sparse dictionary by using the latter algorithm; the algorithm operation flow is described in detail as follows:

s4-1, initializing and processing relevant parameters of the joint reconstruction algorithm; for the initial residual r₀Its residual r corresponding to the p-th node_pAre in equal relation; index value omega_k0; index set Λ₀Is an empty set;

s4-2, and converting the initial signal matrix X_n*sInitial dictionary psi_n*nMeasurement matrix phi_m*nMinimum reconstructed signal-to-noise ratio SNR_defAs an input item, where s is the number of nodes, and p ═ 1,2,. multidata, s }, the data length is n, and the number of measurements is m;

s4-3, establishing a sensing matrix to obtain the following expression:

A_m*n＝ψ_m*n*Φ_n*n

s4-4, solving each row residual error r by adopting the following formula_pWith each row of sensing matrix A_qTwo norm sum between:

according to the obtained maximum value of the two norm sum, the corresponding sensing matrix row index is reserved, and after the sensing matrix row index is fused with the index set, a new index set is obtained, as follows:

Λ_τ＝[Λ_τ-1ξ_p]

s4-5, after solving the relevant parameters by the least square algorithm, updating the residual error by using the following expression:

s4-6, respectively solving the reconstructed intermediate signal and the relative root mean square error and the reconstructed signal-to-noise ratio thereof by adopting the following calculation formulas:

finally, the lowest reconstructed SNR is compared_defWhen the SNR is smaller, updating dictionary atoms by using a dictionary training algorithm, and returning to S4-4; otherwise, an output result, namely a reconstruction result x is obtained_j' with dictionary atoms used;

s5, compressing the steady-state data of the power system under the wavelet transformation;

known even sequence e_j+1Odd sequence o_j+1Then the wavelet decomposition process is described using the following expression:

split(x′_j)＝(e_j+1,o_j+1)

in the above formula, the even number sequence e_j+1＝a_j+1-U(b_j+1) Odd sequence o_j+1＝b_j+1+P(a_j+1) (ii) a Wherein, a_j+1And b_j+1Respectively representing low and high frequency coefficients in the sequence, Y (b)_j+1) And P (a)_j+1) Respectively representing the updating result of the high-frequency coefficient and the prediction result of the low-frequency coefficient;

from this, a compressed reconstructed signal representation is derived, as follows:

x″_j＝merge(e_j+1,o_j+1)

wherein merge represents a merge sort algorithm.

Further, in S3, there is a constant δ in the range of 0 to 1_kFor all sparse coefficient matrices X, the following inequality holds for the measurement matrix Φ:

a compression effect evaluation method of a power system steady-state data compression method based on edge calculation is characterized by comprising the following steps: connecting an electric energy acquisition device to the electric interface to acquire steady-state data of a research object; carrying out quantitative evaluation on three indexes of the space ratio of data compression, the normalized mean square error and the data compression ratio; wherein the content of the first and second substances,

compared with the prior art, the invention has the beneficial effects that:

(1) the invention relates to a power system steady-state data compression method based on edge calculation, which combines compressed sensing and distributed information source coding to complete power system steady-state data fusion under edge calculation through stages of establishing a joint sparse model, establishing a sparse redundant dictionary, determining a measurement matrix, establishing a joint reconstruction algorithm and the like; decomposing the obtained data fusion result to each scale level according to resolution by using a wavelet transform algorithm to obtain a high-frequency coefficient and a low-frequency coefficient, and outputting a compression result by adopting a lossless coding technology after the high-frequency coefficient is subjected to threshold processing.

(2) According to the method for compressing the steady-state data of the power system based on the edge calculation, the quantitative evaluation results of three indexes of the space occupation ratio of data compression, the normalized mean square error and the data compression ratio are verified, the method has obvious compression advantages, and the waveform fitting degree of a compressed signal and an actual sampling signal is high.

Drawings

Fig. 1 is a schematic diagram of the principle of reconstruction compression.

Fig. 2 is a flow chart of fusion and compression of steady-state data signals of the power system.

Fig. 3 is a schematic diagram of a steady-state data sampling signal.

Fig. 4 is a waveform diagram of a reconstructed signal fused based on multiple compression methods.

Fig. 5 is a waveform diagram of an error signal based on fusion of multiple compression modes.

Fig. 6 is a waveform diagram of a reconstructed signal based on tensor Tucker decomposition.

Fig. 7 is a waveform diagram of an error signal based on tensor Tucker decomposition.

Fig. 8 is a waveform diagram of a reconstructed signal according to the present invention.

FIG. 9 is a waveform diagram of an error signal according to the present invention.

FIG. 10 is a diagram illustrating the evaluation index results of normalized mean square error and data compression ratio.

FIG. 11 is a diagram illustrating steady-state data compression space ratio indicators for different methods.

Detailed Description

The following description of the embodiments of the present invention will be made with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.

Based on the distributed compressed sensing technology, the method completes the steady-state data fusion of the power system under the edge calculation through four steps of establishing a joint sparse model, establishing a sparse redundant dictionary, determining a measurement matrix and establishing a joint reconstruction algorithm. Decomposing the obtained data fusion result to each scale level according to resolution by using a wavelet transform algorithm to obtain a high-frequency coefficient and a low-frequency coefficient, and outputting a compression result by adopting a lossless coding technology after the high-frequency coefficient is subjected to threshold processing.

The invention comprises the following steps:

s1, establishing a joint sparse model;

the steady-state data of the power system in the time domain has no sparsity, and only sparse basis sparse decomposition is utilizedAfter the steady state data, the distributed compressed sensing technology can be used for collection and fusion. Because the signal frequency of the steady-state data contains subharmonic waves and fundamental waves, the initial sparse basis is represented by a Fourier forward transform matrix, and a joint sparse model is established to acquire the steady-state data. There is no common part in the signal, so sparse representation of the innovation part (i.e. the difference between the coefficient vector and the common part) can be done by one sparse basis. Assume that the initial steady-state data signal is x_jThe common part and the innovation part are respectively z_c、z_jThen signal x_jSparse representation of (c) is as follows:

x_j＝z_c+z_j＝z_j＝ψθ_j (1)

wherein the sparse matrix and the corresponding sparse coefficients are psi, theta_j。

S2, creating a sparse dictionary;

the sparsity degree is inversely related to the number of atoms and the amount of uploaded data. In order to ensure that the atoms and the initial signals are subjected to self-adaptive matching, a learning type sparse dictionary in a dictionary training algorithm is introduced, and through multiple dictionary updating, the deviation between the initial signals and the reconstructed signals is reduced, so that the signal-to-noise ratio of the reconstructed signals meets a preset threshold value.

Regarding the sparse decomposition stage, after the known initial sparse dictionary D is solved by the orthogonal matching pursuit algorithm to obtain sparse representation, the sparse representation coefficient shown in the following formula is obtained

The constraints are as follows:

wherein y represents a reconstructed steady-state data signal, the preset threshold is epsilon, N represents the number of atoms, and F represents a sparse coefficient matrix norm.

In the cyclic calculation during dictionary updating, the dictionary training algorithm only performs updating processing on one atom at a time. When new atom d is fetched_kThen, the following equation holds:

in the above formula, the dilution coefficients with row numbers j and k in the sparse coefficient matrix X are respectivelyd_jThe expression ordinal number j is an atom.

If d is removed_kThe other atoms being offset byRewriting the above expression as the following expression:

suppose atom d_kIndex of the reconstructed signal of omega_k，N*ω_kIs omega_kIf divide by (ω)_k(i) I) the matrix elements other than the non-zero values are all zero values, the following expression is derived from the above expression, wherein the index ω to avoid divergence of the result is_kThe expression is shown in formula (7):

in the above formula, the k-th line is thinned outAfter zero value items in the row vector are removed, a row vector is obtainedNamely, it isAtom d within sparse coding phase_kThe deviation sequence isNamely, it is

Decomposing the deviation column by singular value decomposition strategyThe following decomposition expression was obtained:

wherein, the two orthogonal matrixes are U, V respectively, the diagonal matrix is delta, the first column of the two orthogonal matrixes U, V is obtained by decomposition, and the former is used for completing the atom d in the initial dictionary_kAfter multiplying the latter by the diagonal matrix Δ (1,1), the update and the replacement x are updated by the resulting product_jAnd then a new sparse dictionary is obtained.

S3, establishing a measurement matrix;

and a Gaussian measurement matrix is constructed, the dimensionality of the signals represented by the sparse matrix is reduced, and the accuracy of the reconstructed signals and the constraint equidistant condition are ensured to be satisfied. That is, there is a constant δ in the range of values from 0 to 1_kFor all sparse coefficient matrices X, the following inequality holds for the measurement matrix Φ:

s4, establishing a joint reconstruction algorithm;

and establishing a joint reconstruction algorithm by fusing a synchronous orthogonal matching tracking algorithm and a learning training algorithm. The former algorithm is used for reconstructing the acquired steady-state data, and the latter algorithm is used for updating the sparse dictionary. The algorithm operation flow is described in detail as follows:

and S4-1, initializing and processing relevant parameters of the joint reconstruction algorithm. For the initial residual r₀Its residual r corresponding to the p-th node_pAre in equal relation; index value omega_k0; index set Λ₀Is an empty set;

s4-2, and converting the initial signal matrix X_n*sInitial dictionary psi_n*nMeasurement matrix phi_m*nMinimum reconstructed signal-to-noise ratio SNR_defAs an input item, where s is the number of nodes, and p ═ 1,2,. multidata, s }, the data length is n, and the number of measurements is m;

s4-3, establishing a sensing matrix to obtain the following expression:

A_m*n＝ψ_m*n*Φ_n*n (10)

s4-4, solving each row residual error r by adopting the following formula_pWith each row of sensing matrix A_qTwo norm sum between:

according to the obtained maximum value of the two norm sum, the corresponding sensing matrix row index is reserved, and after the sensing matrix row index is fused with the index set, a new index set is obtained, as follows:

Λ_τ＝[Λ_τ-1ξ_p] (12)

s4-5, after solving the relevant parameters by the least square algorithm, updating the residual error by using the following expression:

s4-6, respectively solving the reconstructed intermediate signal and the relative root mean square error and the reconstructed signal-to-noise ratio thereof by adopting the following calculation formulas:

finally, the lowest reconstructed SNR is compared_defWhen the SNR is smaller, updating dictionary atoms by using a dictionary training algorithm, and returning to S4-4; otherwise, an output result, namely a reconstruction result x is obtained_j' with the dictionary atoms used.

S5, compressing the steady-state data of the power system under the wavelet transformation;

based on the steady-state data fusion result obtained by edge calculation, decomposing the steady-state data fusion result to each scale level according to resolution by using a wavelet transform algorithm to obtain a high-frequency coefficient and a low-frequency coefficient, zeroing the relatively small high-frequency coefficient by threshold processing, and only keeping the low-frequency coefficient and the high-frequency coefficient with signal feature presentation capability. The conversion from integer to integer is fundamentally realized, the floating point calculation steps are reduced, and the method is more suitable for the practical application of the power system. The reconstruction compression based on the wavelet transform algorithm is divided into several stages of splitting, predicting, updating and the like, and the principle of the reconstruction compression is shown in fig. 1.

Known even sequence e_j+1Odd sequence o_j+1Then the wavelet decomposition process is described using the following expression:

split(x′_j)＝(e_j+1,o_j+1) (17)

in the above formula, the even number sequence e_j+1＝a_j+1-U(b_j+1) Odd sequence o_j+1＝b_j+1+P(a_j+1). Wherein, a_j+1And b_j+1Respectively representing low and high frequency coefficients in the sequence, Y (b)_j+1) And P (a)_j+1) Respectively representing the update result of the high-frequency coefficient and the prediction result of the low-frequency coefficient.

From this, a compressed reconstructed signal representation is derived, as follows:

x″_j＝merge(e_j+1,o_j+1) (18)

wherein merge represents a merge sort algorithm.

The implementation flow of the method for fusing and compressing the steady-state data signals of the power system is shown in fig. 2. The method comprises the steps of firstly fusing a steady-state data signal by using an edge algorithm, then performing multi-scale transformation processing by using a wavelet algorithm, processing a high-frequency coefficient by using a threshold value, and improving a compression ratio by using a lossless coding technology.

Embodiment I, power system steady state data compression experiment analysis

S1, a preparation phase;

the method is used for carrying out static data compression test on a trial-run power system of a certain power grid company and verifying feasibility and applicability of the method. An EAC5000D type electric energy acquisition device is connected to the electric interface to acquire steady-state data of a research object, and a sampling signal of the steady-state data is shown in figure 3. In the acquisition process, the sampling rate is 50kHz, and the number of sampling points of the initial voltage signal is 36000.

In order to measure the data compression effect, firstly, a method of fusing a plurality of compression modes and tensor Tucker decomposition and a method are integrated, the acquired static data are compressed one by one, and the waveforms of a reconstruction signal and an error signal are observed; and then quantitative evaluation is carried out by using three indexes of the space ratio of data compression, the normalized mean square error and the data compression ratio. The calculation formulas of the indexes are respectively as follows:

in the three indexes, except that the data compression space ratio index value and the compression effect are in positive correlation, the other two compression evaluation indexes are in negative correlation with the compression effect, and the smaller the index value is, the more ideal the compression effect is.

S2, analyzing the compression effect of the steady-state data based on the sampling signal;

the static data compression reconstructed signal and the error signal of the different methods are shown in fig. 4-6, respectively. As is apparent from the signal waveform diagram, the method fuses the acquired steady-state data by using the edge algorithm and decomposes the fused signal to each scale level by using the wavelet transform algorithm, so that the final compressed signal and the actual steady-state data sampling signal waveform (see fig. 3) have higher fitting degree.

Combining signal waveforms and their normalized mean square errors of different methods with the data compression ratio evaluation index results (as shown in fig. 7), it can be seen that the invention realizes the conversion from integer to integer fundamentally, reduces the floating point calculation step greatly, and improves the compression ratio by the lossless coding technique because of introducing the wavelet transform algorithm, therefore, the index values of data compression ratio and normalized mean square error are far smaller than the index values of other two methods. This shows that the method can remove more redundant data, and the original signal characteristics are better preserved, and the compression advantage is significant.

S3, analyzing the steady-state data compression effect based on the data specification;

in order to detect the influence of the data size on the compression effect, the steady-state data compression effect of different methods is evaluated by adopting the data compression space ratio aiming at the steady-state data with the specifications of 64kB, 128kB, 256kB, 512kB and 1MB respectively. The results of the index data are shown in fig. 8.

Therefore, the method only retains the low-frequency coefficient and the high-frequency coefficient with signal characteristic presenting capability by zeroing the relatively small high-frequency coefficient through threshold processing, so that the compression space occupation ratio is larger and the compression effect is more ideal compared with the other two methods. From the trend of the space ratio curve of the method, the index value is reduced with the increase of the data specification, which shows that a certain correlation exists between the compression effect and the data size of the method, and the index value can be taken as the research focus of the next stage to deal with the massive data scale of the power system in the information age.

The method completes the steady-state data fusion of the power system under the edge calculation by establishing a joint sparse model, establishing a sparse redundant dictionary, determining a measurement matrix, establishing a joint reconstruction algorithm and the like and combining compressed sensing and distributed source coding. Decomposing the obtained data fusion result to each scale level according to resolution by using a wavelet transform algorithm to obtain a high-frequency coefficient and a low-frequency coefficient, and outputting a compression result by adopting a lossless coding technology after the high-frequency coefficient is subjected to threshold processing. In the test, a compression test is expanded for static data of a test run power system of a certain power grid company, and the method has obvious compression advantages and high fitting degree of a compressed signal and an actual sampling signal waveform according to quantitative evaluation results of three indexes of a data compression space ratio, a normalized mean square error and a data compression ratio.

In the above embodiments, the present invention is described only by way of example, but those skilled in the art, after reading the present patent application, may make various modifications to the present invention without departing from the spirit and scope of the present invention.