Data fusion method and device, electronic equipment, storage medium and product
By preprocessing the distribution network data and integrating the characteristic dimension probability allocation, the difficulty in determining fault locations caused by multi-source heterogeneous data in the distribution network is solved, and the accurate identification and rapid processing of fault locations are achieved, thereby improving power System troubleshooting efficiency.
Patent Information
- Application Number
- CN202410583690.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-11
- Publication Date
- 2024-11-22
AI Technical Summary
The multi-source and heterogeneity of data in the distribution network leads to the inability to effectively determine the fault location, affecting the efficiency of troubleshooting and processing.
By acquiring the distribution network data set for preprocessing, probability allocation and fusion are performed based on feature dimensions, feature fusion of multi-source heterogeneous data is achieved by using Gaussian distribution curves and evidence theory to determine the fault location.
It effectively solves the problem of determining fault locations caused by multi-source heterogeneous data, realizes accurate identification and rapid processing of fault locations, and improves the fault detection efficiency of the power system.
Smart Images

Figure CN119004355A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of electric power technology, and in particular to a data fusion method, device, electronic equipment, storage medium and product. Background Art
[0002] At present, with the continuous development of the power industry, it has become a trend to build an integrated data platform based on distribution network data and form a digital power grid. The large amount of distribution network data contained in the digital power grid can be acquired and shared instantly, enabling staff to perform distribution network-related operations based on distribution network data, such as fault diagnosis operations.
[0003] However, with the rapid advancement of the energy Internet strategy, the number of power sensors, distribution equipment, and status sensors in the distribution network has increased dramatically, resulting in a large amount of heterogeneous distributed data, and presenting problems of multiple types, multiple sources, and uncertainty. When a power system fails, it is necessary to determine the operation location based on the power data with abnormal conditions in the distribution network so that the staff can repair it in time. However, due to the multi-source and heterogeneous problems of data in the distribution network, it is often impossible to effectively determine the operation location, where the operation location is the measurement location of the distribution network data.
[0004] Therefore, a data fusion method is urgently needed to determine the operation location corresponding to the distribution network data. Summary of the invention
[0005] The present invention provides a data fusion method, device, electronic device, storage medium and product, which solves the problem that the fault location cannot be effectively determined due to the multi-source and heterogeneous data in the distribution network, and realizes the feature fusion of multi-source heterogeneous data.
[0006] In a first aspect, an embodiment of the present disclosure provides a data fusion method, including:
[0007] Acquire a distribution network data set, and preprocess the distribution network data set to obtain a data set to be fused;
[0008] Probability allocation is performed on the data set to be fused based on the feature dimensions corresponding to the data set to be fused, and a first probability allocation result corresponding to the number of the feature dimensions is determined;
[0009] The first probability distribution results are fused to obtain a second probability distribution result.
[0010] In a second aspect, an embodiment of the present disclosure provides a device for determining an associated object based on semantic extension, comprising:
[0011] A data set acquisition module is used to acquire a distribution network data set and pre-process the distribution network data set to obtain a data set to be fused;
[0012] A probability allocation module, configured to perform probability allocation on the data set to be fused based on the feature dimensions corresponding to the data set to be fused, and determine a number of first probability allocation results corresponding to the number of feature dimensions;
[0013] The feature data fusion module is used to fuse the first probability distribution results to obtain a second probability distribution result.
[0014] In a third aspect, an embodiment of the present disclosure provides an electronic device, including:
[0015] at least one processor; and
[0016] a memory communicatively connected to at least one processor; wherein,
[0017] The memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can execute a data fusion method provided by the above-mentioned first aspect embodiment.
[0018] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a processor to implement a data fusion method provided in the embodiment of the first aspect above when executed.
[0019] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements a data fusion method provided in the embodiment of the first aspect.
[0020] A data fusion method, device, equipment, medium and product of an embodiment of the present invention obtains a data set to be fused by acquiring a distribution network data set and preprocessing the distribution network data set; performs probability distribution on the data set to be fused based on the feature dimensions corresponding to the data set to be fused, and determines a first probability distribution result corresponding to the number of feature dimensions; and performs fusion processing on each of the first probability distribution results to obtain a second probability distribution result. The above technical solution performs probability distribution and fusion on the data set to be fused based on the feature dimensions, solves the problem that the fault location cannot be effectively determined due to the multi-source and heterogeneous problems of data in the distribution network, and realizes the feature fusion of multi-source heterogeneous data.
[0021] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0023] Figure 1 is a flow chart of a data fusion method provided in Embodiment 1 of the present invention;
[0024] Figure 2 is a flow chart of a data fusion method provided in Embodiment 2 of the present invention;
[0025] Figure 3 This is an example display of a Gaussian distribution curve involved in a data fusion method provided in the second embodiment of the present invention;
[0026] Figure 4 This is an example diagram showing a correction process of an initial probability distribution result in a data fusion method provided in Embodiment 2 of the present invention;
[0027] Figure 5 is a structural schematic diagram of a data fusion device provided in Embodiment 3 of the present invention;
[0028] Figure 6 It is a structural schematic diagram of an electronic device provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION
[0029] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0030] It should be noted that the terms "first", "second" and "target" in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0031] Embodiment 1
[0032] Figure 1 It is a flow chart of a data fusion method provided in the first embodiment of the present invention. This embodiment is applicable to the situation where the probability distribution and fusion of the data sets to be fused are performed based on the feature dimensions. The method can be executed by a data fusion device, and the data fusion device can be implemented in the form of hardware and / or software.
[0033] like Figure 1 As shown, the method includes:
[0034] S101. Obtain a distribution network data set, and preprocess the distribution network data set to obtain a data set to be fused.
[0035] In this embodiment, the distribution network data set can be understood as a collection of distribution network data, and the distribution network data set in this embodiment is a collection of distribution network data from a measurement location. The data set to be fused can be understood as a standardized data set that can be used for subsequent data fusion after preprocessing the distribution network data set.
[0036] Specifically, the power data of the distribution network transmitted by power equipment such as power sensors, distribution equipment, and status sensors are obtained, and a distribution network data set is constructed based on each distribution network data, or a distribution network data set composed of distribution network data is directly obtained. The distribution network data set is initialized, and power transformation and data standardization are performed on the distribution network data set to obtain a standard data set to be fused.
[0037] S102 , performing probability distribution on the data set to be fused based on the feature dimensions corresponding to the data set to be fused, and determining a number of first probability distribution results corresponding to the number of feature dimensions.
[0038] In this embodiment, the feature dimension can be understood as the data source category corresponding to each distribution network data in the data set to be fused, such as voltage, current, active power, reactive power, etc. The first probability allocation result can be understood as the probability distribution result of the position to be determined corresponding to a single feature dimension. There can be several positions to be determined, and among the several positions to be determined, one position is the actual measurement position of the distribution network data. The probability allocation result is the basic probability allocation (Basic Probability Assignment, BPA).
[0039] Specifically, for a single feature dimension, Gaussian probability distribution calculation processing is performed on the distribution network data corresponding to each feature dimension in the data set to be fused, and the probability distribution information of each measurement position to be determined corresponding to the feature dimension is obtained, and the obtained probability distribution information is corrected to obtain an optimized first probability distribution result. It can be understood that each first probability distribution result corresponds to a feature dimension. For example, after the probability distribution processing is performed on the data set to be fused, the first probability distribution result of the relative voltage, the first probability distribution result of the relative current, the first probability distribution result of the relative active power, and the first probability distribution result of the relative reactive power are obtained.
[0040] S103: Fusing the first probability distribution results to obtain a second probability distribution result.
[0041] In this embodiment, the second probability distribution result can be understood as the final probability distribution result obtained by integrating the first probability distribution results of various feature dimensions, including the target probability value corresponding to each position to be determined.
[0042] Specifically, the weights corresponding to the first probability distribution results of each feature dimension are determined, and based on the weights, the first probability distribution results of each feature dimension are weighted fused, and the fused probability distribution results are fused again based on evidence theory to obtain the final second probability distribution result.
[0043] A data fusion method provided by an embodiment of the present invention includes obtaining a distribution network data set, and preprocessing the distribution network data set to obtain a data set to be fused; performing probability distribution on the data set to be fused based on the feature dimensions corresponding to the data set to be fused, and determining a first probability distribution result corresponding to the number of feature dimensions; and fusing each first probability distribution result to obtain a second probability distribution result. The above technical solution performs probability distribution and fusion on the data set to be fused based on the feature dimensions, solves the problem that the fault location cannot be effectively determined due to the multi-source and heterogeneous problems of data in the distribution network, and realizes the feature fusion of multi-source heterogeneous data.
[0044] As a first optional embodiment of this embodiment, the second probability allocation result includes a target probability value relative to each measurement position to be determined, and the method further includes:
[0045] The to-be-determined measurement position corresponding to the target probability value satisfying the position determination condition in the second probability allocation result is determined as the target measurement position.
[0046] In this embodiment, the second probability distribution result includes target probability values corresponding to the number of positions to be determined, and the target probability value can be understood as the probability value of the position to be determined. For example, the positions to be determined include position A, position B, and position C. The second probability distribution result includes a target probability value of 0.628 relative to position A, a target probability value of 0.230 relative to position B, and a target probability value of 0.142 relative to position C. The position determination condition can be understood as a condition for determining the target measurement position, for example, the target probability value can be the highest value in the second probability distribution result.
[0047] Specifically, the to-be-determined measurement position corresponding to the target probability value with the highest numerical value in the second probability distribution result is determined as the target measurement position, for example, position A is determined as the target measurement position.
[0048] Determining the target measurement location from among the measurement locations to be determined is beneficial for subsequent staff to troubleshoot and handle distribution network faults based on the target measurement location.
[0049] Embodiment 2
[0050] Figure 2 It is a flow chart of a data fusion method provided in the second embodiment of the present invention. This embodiment is a further optimization of any of the above embodiments, and can be applied to the situation where the data sets to be fused are probability assigned and fused based on feature dimensions. The method can be executed by a data fusion device, and the data fusion device can be implemented in the form of hardware and / or software.
[0051] like Figure 2 As shown, the method includes:
[0052] S201. Obtain an initialized distribution network data set.
[0053] S202 . Based on the data set format corresponding to the distribution network data set, perform power transformation and standardization processing on the distribution network data set to obtain the data set to be fused.
[0054] In this embodiment, the data set format can be understood as the composition form of the data set, such as a multidimensional array, a matrix, and a vector. Distribution network data set X = (X1, X2,, X p ) includes p distribution network data. Taking the matrix as an example, the distribution network data set Where n is the number of feature dimensions.
[0055] When the data set format corresponding to the distribution network data set is a multidimensional array (three-dimensional or higher array), the mean and standard deviation are solved along the multi-dimensionality of X, and then the data of X is normalized to return the converted high-dimensional array BC_Z = [X-mean(X)] / std(X), where mean(X) is the mean and std(X) is the standard deviation;
[0056] When the data set format corresponding to the power distribution network data set is vector, the vector conversion is performed directly, and the converted result vector is returned to BC_Z;
[0057] When the data set format corresponding to the distribution network data set is a matrix (two-dimensional array), data standardization is performed on the corresponding columns one by one based on the mean and standard deviation of the column vector of X, and the result matrix BC_Z is returned.
[0058] Specifically, for a single distribution network data x in the distribution network data set X, based on Perform generalized power transformation to obtain the transformed power grid data x * , where λ is the known power transformation parameter; based on the transformed power grid data x * Perform Z-score data standardization to obtain the transformed In the formula, i and j are the row data value and column data value respectively. is the mean value of variable Xj, is the standard deviation of variable Xj, For p distribution network data, the above power transformation and standardization processing are iterated in sequence to obtain the data set to be fused Each column Collect data conversion processing results.
[0059] S203: Classify the data to be fused in the data set to be fused according to the feature dimension to form at least one array to be fused.
[0060] In this embodiment, the data to be fused are the data of each distribution network after power transformation and standardization. The distribution network data has data characteristics, and the characteristic dimensions include voltage, current, active power, reactive power, etc. Correspondingly, the data to be fused also has data characteristics consistent with the distribution network data. The array to be fused is the array obtained by classifying the data to be fused based on the characteristic dimensions, and the data to be fused under the same characteristic dimension constitutes an array to be fused.
[0061] Specifically, for each feature dimension included in the data set to be fused, the data to be fused corresponding to the feature dimension are classified together to form an array to be fused, and finally the number of arrays to be fused corresponding to the number of feature dimensions is formed.
[0062] S204. For each array to be fused, the array to be fused is processed based on a pre-constructed Gaussian probability density function to obtain a Gaussian distribution curve relative to each measurement position to be determined, and an initial probability allocation result is determined based on each Gaussian distribution curve, wherein the measurement position to be determined is any measurement position in a pre-stored measurement position set.
[0063] In this embodiment, the measurement position to be determined can be understood as a measurement position that cannot be determined whether it is the target measurement position corresponding to the distribution network data set. It can be understood that among the measurement positions to be determined, there is a measurement position to be determined that is the real measurement position of the distribution network data set. The measurement position set is a set of measurement positions in the known distribution network system. The initial probability allocation result can be understood as the probability allocation result of the initially determined position relative to each position to be determined, and its effectiveness is relatively poor.
[0064] Specifically, there may be multiple measurement locations in the distribution network system to collect distribution network data, and these measurement locations constitute a measurement location set. Several or all measurement locations in the measurement location set are determined as locations to be determined. For each array to be fused corresponding to each feature dimension, the array to be fused is input into a pre-constructed Gaussian probability density function corresponding to the feature dimension. In the formula, X is the array to be fused, μ is the mean of X, and σ is the variance of X. The Gaussian distribution curve of the Gaussian function output relative to the measurement position to be determined is obtained, and the initial probability distribution result is determined based on the intersection position between the Gaussian distribution curves corresponding to each measurement position to be determined.
[0065] For example, Figure 3 This is an example display of a Gaussian distribution curve in a data fusion method provided in the second embodiment of the present invention; Figure 3 As shown, the positions A, B, and C to be determined correspond to a Gaussian distribution curve respectively. In the figure, the horizontal axis represents the characteristic value s j ′, the ordinate is the probability value, and the intersection points h1, h2, and h3 are the eigenvalues s j ′ and the measurement positions A, B, C to be determined on the feature s j The intersection of the Gaussian function under s j The probability value of the relative measurement position to be determined under the coordinates is determined based on the intersection positions and the intersection relationship between the Gaussian distribution curves. The part without the intersection relationship is taken to determine the initial probability distribution result. The initial probability distribution result can be, for example, Figure 3The above formula includes m(C), m(A,C), and m(A,B,C), where M(C) = h3-h2, M(A,C) = h2-h1, and M(A,B,C) = h1. For example, m(B), m(B,C), and m(B,C,A) may also be included, as shown in the following formula. The initial probability distribution result of the relative voltage U feature is S U :m(B)=0.039,m(B,C)=0.184,m(B,C,A)=0.777, the initial probability distribution result of the relative current I characteristic is S I :m(B)=0.023,m(B,C)=0.051,m(B,C,A)=0.925, the initial probability distribution result of the relative active power P feature is S P :m(C)=0.120,m(B,C)=0.859,m(B,C,A)=0.00, the initial probability distribution result of the relative reactive power Q feature is S Q :m(C)=0.177, m(B,C)=0.801, m(B,C,A)=0.001.
[0066] It can be understood that before processing the fused array based on the pre-built Gaussian probability density function, based on the training data set and the maximum likelihood method, get and Will The value of is determined as the mean μ, The value of is determined as the variance σ.
[0067] Exemplarily, the distribution network dataset contains 150 samples, of which 50 samples are from three locations, which are represented by A, B, and C for convenience. Each location corresponds to 4 features - voltage U, current I, active power P, and reactive power Q. The sample selection method is as follows: if the original training data volume is N*t, where N is the total number and 0<t≤1 is the training ratio. For a feature, let the mean of an event be μ, the filling ratio be p, where 0<p≤1, and then fill (N*t)*p samples with a value of μ, and the size of the filled data set is (N*t)*(1+p).
[0068] The proportion of data extracted from the training data set used to construct the Gaussian probability density function is called the training proportion. As a preparation, all the data are first shuffled, and then data with an 80% training proportion are randomly selected instead of using the proportion data in each data set to ensure the authenticity during training and then the accuracy during application. These data are then used to generate a Gaussian probability distribution, and the function is determined based on the maximum likelihood estimate. The filling ratio is set to 40% of the data, with the value being the average of the extracted data to reduce overfitting. The mean and variance values of the Gaussian distribution obtained for the four features at three positions are shown in Table 1 below:
[0069] Table 1 μ, σ of different types of features obtained through training sets
[0070]
[0071] Random distribution network data samples with four characteristics are selected in the data set and their locations are shown in Table 2:
[0072] Table 2. Characteristics and true locations of selected samples
[0073]
[0074] S205. For each initial probability distribution result, modify the initial probability distribution result based on the confidence value corresponding to the initial probability distribution result to determine the corresponding first probability distribution result.
[0075] In this embodiment, the confidence value can be understood as the value of the uncertainty corresponding to the initial probability distribution result. The lower the confidence value, the higher the uncertainty.
[0076] Specifically, for each initial probability allocation result, the initial probability allocation result is reconstructed by assigning the probability values of the parts corresponding to the high uncertainty in the initial probability allocation result to the low uncertainty values associated with them, and the reconstructed initial probability allocation result is transformed in a reverse manner to obtain the final first probability allocation result.
[0077] By reconstructing the initial probability distribution results, the elements in the probability distribution results are dispersed so that more effective data can be obtained in the subsequent fusion.
[0078] Optionally, the initial probability distribution result is corrected based on the confidence value corresponding to the initial probability distribution result to determine the corresponding first probability distribution result, including:
[0079] S2051. Reconstruct the initial probability distribution result based on the confidence value corresponding to the initial probability distribution result, and determine the intermediate probability distribution result.
[0080] In this embodiment, the intermediate probability distribution result can be understood as the probability distribution result after reconstructing the initial probability distribution result.
[0081] Specifically, for BPA, the more elements pointed to, the greater the uncertainty of BPA, and the more ambiguous the information contained. For each initial probability allocation result, based on the reconstruction equation Processing, where m(A j ) is the BPA focal element probability value before reconstruction, m r (A i ) is the probability value of the BPA focal element after reconstruction, m(Θ) is the identification frame before reconstruction, m r (Θ) is the reconstructed recognition frame. A i , A j is the focal element of the identification frame Θ, |A j | is A j The modulo operation is also equal to A j The number of elements contained in Indicates A j The number of possible outcomes in , which is a measure of uncertainty, and n is the number of focal elements contained in BPAΘ. Through this operation, the data of each BPA comes not only from itself, but also from its upper set, measuring the degree of association between individual BPAs. When the focal element of a BPA is BPAΘ, its only source of data is itself. According to The reconstructed BPA is normalized to form an intermediate probability distribution result to meet the construction criteria of BPA and facilitate subsequent operations.
[0082] In the above example, It can be seen that the generated BPA value tends to be the BPA with more focal elements, for example, (B, C, A) under the current characteristic reaches 0.925. At this time, the BPAs are constructed by reconstructing the equation to reduce the uncertainty of the BPA and obtain the intermediate probability distribution result. Taking the voltage characteristic as an example, The intermediate probability distribution results of the seven relative voltage dimensions are obtained. Obviously, the initial probability distribution results containing only three probabilities are reconstructed to obtain the intermediate probability distribution results containing seven probabilities. The reconstructed intermediate probability distribution results relative to each feature dimension are shown in Table 3. It can be seen that the BPA with the highest uncertainty, such as m(B, C, A), is reduced, while the BPA with low uncertainty, such as m(B) and m(C), is increased.
[0083] Table 3. Reconstructed BPA
[0084]
[0085]
[0086] S2052: Perform an inverse transformation on the intermediate probability distribution result to determine a first probability distribution result.
[0087] In this embodiment, the BPA is inverted, and the reconstructed BPA (intermediate probability allocation result) is transformed into nrBPA (first probability allocation result), where nrBPA includes several probability values mnr.
[0088] Specifically, let m Ai is the BPA on the identification frame Θ, is m Ai The complement of Negative BPA is defined as That is, through Take the inverse transformation of the intermediate probability distribution result to obtain the first probability distribution result nrBPA, where N = |Θ| is the number of all focal elements in the recognition frame Θ, 2 N -2 is the sum of the reciprocals of all BPAs of the recognition frame Θ.
[0089] For example, the equation To calculate the inverse of the reconstructed BPA, the result is nrBPA mnr. When the number of focal elements contained in the BPA is 1, mnr is simply converted to a difference relative to 1. When the number of focal elements is greater than 1, the higher the uncertainty, the smaller the value obtained by dividing by a number greater than 1, and the smaller the value obtained in the fusion result, which corresponds to the increased uncertainty of the evidence. The process of taking the inverse of mr to obtain the nrBPA of the voltage feature U is as follows:
[0090]
[0091]
[0092]
[0093]
[0094] The final first probability distribution results relative to each feature dimension are shown in Table 4 below:
[0095] Table 4. nrBPA
[0096]
[0097]
[0098] S206. For each first probability allocation result, determine the information entropy of the first probability allocation result, and determine the feature weight of the feature dimension corresponding to the first probability allocation result according to the information entropy.
[0099] In this embodiment, the information entropy is the confidence entropy, which is used to calculate the weight. The feature weight is the weight corresponding to each feature dimension, and the sum of the weights of each feature dimension is 1.
[0100] Specifically, after determining each first probability allocation result, by In the formula, E Md is the information entropy, m(A) is the first probability allocation result of position A, and the uncertainty of the first probability allocation result is measured based on the improved information entropy. Calculate the weight of each feature, where E i is the information entropy E of each feature i Md .
[0101] Taking EMd(U) of feature U as an example, the calculation is as follows:
[0102]
[0103] When the number of focal elements increases, the improved confidence entropy (information entropy) also takes into account the BPA data of the subset of BPA. The information entropy of all features can be obtained by the same method. The information entropy of the first probability distribution result corresponding to each feature dimension is shown in Table 5 below:
[0104] Table 5 Information entropy of all features
[0105] EU EI EP EQ 1.852 1.846 1.868 1.870
[0106] According to the equation Determine the weight WU of feature U as Similarly, the weights of all features are calculated, as shown in Table 6 below:
[0107] Table 6 Weights of all features
[0108] WU WI WP WQ 0.251 0.252 0.249 0.249
[0109] S207 . Based on the feature weights of the feature dimensions corresponding to the first probability distribution results, the first probability distribution results are fused to obtain a fused probability distribution result.
[0110] In this embodiment, the fused probability allocation result may be understood as a single probability allocation result obtained by fusing the first probability allocation results corresponding to each feature dimension.
[0111] Specifically, the entropy weight method is used to weight each first probability distribution result, and the final BPA of each focal element is obtained by multiplying the obtained BPA with their corresponding weight values obtained by the improved entropy weight method, and then multiplying the BPAs of different BPAs but the same focal element to obtain the final BPA of each focal element.i For example, the final m′(A i ) is calculated as m′(A i )=w1·m1(A i )+w2·m2(A i )...w M ·m M (A i ), where M is the total number of features, w is the weight, m1(A i ) is the first probability distribution result, m′(A i ) is the fusion probability distribution result
[0112] The importance of different features to fusion is distinguished by the entropy weights of different features. For example, the probability value of mw(A) corresponding to position A is calculated as: m(A)=0.148*0.251+0.145*0.252+0.167*0.249+0.167*0.249=0.157, where mw(A) is the probability value of relative position A in the fusion probability distribution result. Similarly, other fusion probability values are calculated. The fusion probability distribution result is shown in Table 7 below:
[0113] Table 7 Weighted BPA mw (fusion probability distribution result)
[0114] mw(A) mw(B) mw(C) mw(A,B) mw(A,C) mw(B,C) mw(A,B,C) 0.157 0.128 0.116 0.157 0.157 0.130 0.157
[0115] S208: Fusing the reference probability values in the fused probability distribution result relative to each measurement position to be determined, to obtain a second probability distribution result relative to each measurement position to be determined.
[0116] In this embodiment, the reference probability value can be understood as a probability value associated with the measurement position to be determined. For example, the reference probability value of position A includes mw(A), mw(A,B), mw(A,C) and mw(A,B,C); the reference probability value of position B includes mw(B), mw(A,B), mw(B,C) and mw(A,B,C); the reference probability value of position C includes mw(C), mw(A,C), mw(B,C) and mw(A,B,C).
[0117] Specifically, based on the evidence theory combination algorithm, the reference probability values corresponding to each measurement position to be determined are The fusion is performed to obtain a unique probability value relative to each measurement position to be determined, that is, a target probability value, and each target probability value is combined to form a second probability distribution result. The second probability distribution result is shown in Table 8 below:
[0118] Table 8: Second probability distribution results
[0119] A B C 0.628 0.230 0.142
[0120] Figure 4 is an example diagram showing a correction process of an initial probability distribution result in a data fusion method provided in Embodiment 2 of the present invention. Figure 4 As shown, the initial probability allocation result is reconstructed and inverted to obtain several first probability allocation results for each dimension. After the first probability allocation results are fused based on the feature dimension, they are further fused according to each measurement position to be determined to obtain m(A), m(B) and m(C).
[0121] A data fusion method provided by an embodiment of the present invention includes obtaining an initialized distribution network data set; based on a data set format corresponding to the distribution network data set, performing power transformation processing and data standardization processing on the distribution network data set to obtain a data set to be fused; classifying the data to be fused in the data set to be fused according to feature dimensions to form at least one array to be fused; for each array to be fused, processing the array to be fused based on a pre-constructed Gaussian probability density function to obtain a Gaussian distribution curve relative to each measurement position to be determined, and determining an initial probability allocation result based on each Gaussian distribution curve, wherein the measurement position to be determined is any measurement position in a pre-stored measurement position set. The invention relates to a method for realizing a method for realizing a first probability distribution result in which a plurality of measurement positions to be determined are as follows: a first probability distribution result is a first probability distribution result in which a plurality of measurement positions to be determined are as follows: a first probability distribution result is a first probability distribution result in which a plurality of measurement positions to be determined are as follows: a first probability distribution result is a first probability distribution result in which a plurality of measurement positions to be determined are as follows: a first probability distribution result is a first probability distribution result in which a plurality of measurement positions to be determined are as follows: a first probability distribution result is a first probability distribution result in which a plurality of measurement positions to be determined are as follows: a first probability distribution result is a first probability distribution result in which a plurality of measurement positions to be determined are as follows: a second probability distribution result is a second probability distribution result in which a plurality of measurement positions to be determined are as follows: The above technical scheme, according to the multi-source characteristics of distribution network data, utilizes a unified processing method for transforming the dimensions and orders of magnitude of each distribution network data source based on the generalized power transform Z-score to standardize the multi-source data, and obtain a unified data set to be fused for performing data fusion, laying the foundation for the subsequent data fusion process; by reconstructing the initial probability distribution result, the elements in the probability distribution result are dispersed so that more effective data can be obtained in the subsequent fusion; through information entropy weighting and evidence theory algorithm, BPA fusion of multiple feature dimensions is realized to obtain the final probability distribution result, realizing effective, accurate and highly reliable data fusion.
[0122] Embodiment 3
[0123] Figure 5 Schematic diagram of the structure of a data fusion device provided by Embodiment 3 of the present invention. Figure 5As shown, the device comprises:
[0124] The data set acquisition module 31 is used to acquire the distribution network data set and pre-process the distribution network data set to obtain the data set to be fused;
[0125] A probability allocation module 32 is used to perform probability allocation on the data set to be fused based on the feature dimensions corresponding to the data set to be fused, and determine a number of first probability allocation results corresponding to the number of feature dimensions;
[0126] The feature data fusion module 33 is used to fuse the first probability distribution results to obtain a second probability distribution result.
[0127] The data fusion device adopted in this technical solution solves the problem that the fault location cannot be effectively determined due to the multi-source and heterogeneous data in the distribution network, and realizes the feature fusion of multi-source heterogeneous data.
[0128] Optionally, the probability allocation module 32 includes:
[0129] A feature classification unit, used for classifying the data to be fused in the data set to be fused according to the feature dimension to form at least one array to be fused;
[0130] an initial probability allocation unit, configured to process each array to be fused based on a pre-constructed Gaussian probability density function to obtain a Gaussian distribution curve relative to each measurement position to be determined, and determine an initial probability allocation result based on each Gaussian distribution curve, wherein the measurement position to be determined is any measurement position in a pre-stored measurement position set;
[0131] The initial probability correction unit is used to correct each of the initial probability distribution results based on the confidence value corresponding to the initial probability distribution result to determine the corresponding first probability distribution result.
[0132] Optionally, the initial probability correction unit is specifically used to:
[0133] Reconstructing the initial probability distribution result based on the confidence value corresponding to the initial probability distribution result, and determining an intermediate probability distribution result;
[0134] The intermediate probability distribution result is subjected to an inverse transformation process to determine a first probability distribution result.
[0135] Optionally, the feature data fusion module 33 is specifically used for:
[0136] For each first probability allocation result, determine the information entropy of the first probability allocation result, and determine the feature weight of the feature dimension corresponding to the first probability allocation result according to the information entropy;
[0137] Based on the feature weights of the feature dimensions corresponding to the first probability allocation results, the first probability allocation results are fused to obtain a fused probability allocation result;
[0138] The reference probability values in the fused probability distribution results relative to each measurement position to be determined are fused to obtain a second probability distribution result relative to each measurement position to be determined.
[0139] Optionally, the data set acquisition module 31 is specifically used for:
[0140] Get the initialized distribution network dataset;
[0141] Based on the data set format corresponding to the distribution network data set, power transformation processing and standardization processing are performed on the distribution network data set to obtain the data set to be fused.
[0142] Optionally, the second probability allocation result includes a target probability value relative to each measurement position to be determined, and the device also includes a measurement position determination module, which is used to determine the measurement position to be determined corresponding to the target probability value that meets the position determination condition in the second probability allocation result as the target measurement position.
[0143] The associated object determination device based on semantic extension provided in the embodiment of the present invention can execute the data fusion method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
[0144] Embodiment 4
[0145] Figure 6 : is a structural schematic diagram of an electronic device provided in Embodiment 4 of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.
[0146] like Figure 6As shown, the electronic device 40 includes at least one processor 41, and a memory connected to the at least one processor 41, such as a read-only memory (ROM) 42, a random access memory (RAM) 43, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 41 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 42 or the computer program loaded from the storage unit 48 to the random access memory (RAM) 43. In the RAM 43, various programs and data required for the operation of the electronic device 40 can also be stored. The processor 41, the ROM 42, and the RAM 43 are connected to each other through a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.
[0147] A number of components in the electronic device 40 are connected to the I / O interface 45, including: an input unit 46, such as a keyboard, a mouse, etc.; an output unit 47, such as various types of displays, speakers, etc.; a storage unit 48, such as a disk, an optical disk, etc.; and a communication unit 49, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 49 allows the electronic device 40 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.
[0148] The processor 41 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 41 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 41 performs the various methods and processes described above, such as a data fusion method.
[0149] In some embodiments, the data fusion method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 48. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 40 via the ROM 42 and / or the communication unit 49. When the computer program is loaded into the RAM 43 and executed by the processor 41, one or more steps of the data fusion method described above may be performed. Alternatively, in other embodiments, the processor 41 may be configured to perform the data fusion method in any other suitable manner (e.g., by means of firmware).
[0150] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
[0151] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.
[0152] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0153] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).
[0154] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.
[0155] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.
[0156] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.
[0157] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A data fusion method, characterized in that: include: Acquire a distribution network data set, and preprocess the distribution network data set to obtain a data set to be fused; Probability allocation is performed on the data set to be fused based on the feature dimensions corresponding to the data set to be fused, and a first probability allocation result corresponding to the number of the feature dimensions is determined; The first probability distribution results are fused to obtain a second probability distribution result.
2. The method according to claim 1, characterized in that The performing probability distribution on the data set to be fused based on the feature dimensions corresponding to the data set to be fused, and determining a first probability distribution result corresponding to the number of the feature dimensions, includes: Classifying the data to be fused in the data set to be fused according to the feature dimension to form at least one array to be fused; For each array to be fused, the array to be fused is processed based on a pre-constructed Gaussian probability density function to obtain a Gaussian distribution curve relative to each measurement position to be determined, and an initial probability allocation result is determined based on each Gaussian distribution curve, wherein the measurement position to be determined is any measurement position in a pre-stored measurement position set; For each of the initial probability distribution results, the initial probability distribution result is corrected based on the confidence value corresponding to the initial probability distribution result to determine the corresponding first probability distribution result.
3. The method according to claim 2, characterized in that The correcting the initial probability distribution result based on the confidence value corresponding to the initial probability distribution result to determine the corresponding first probability distribution result includes: Reconstructing the initial probability distribution result based on the confidence value corresponding to the initial probability distribution result, and determining an intermediate probability distribution result; The intermediate probability distribution result is subjected to an inverse transformation process to determine a first probability distribution result.
4. The method according to claim 1, characterized in that: The fusing the first probability distribution results to obtain a second probability distribution result includes: For each first probability allocation result, determine the information entropy of the first probability allocation result, and determine the feature weight of the feature dimension corresponding to the first probability allocation result according to the information entropy; Based on the feature weights of the feature dimensions corresponding to the first probability allocation results, the first probability allocation results are fused to obtain a fused probability allocation result; The reference probability values in the fused probability distribution results relative to each measurement position to be determined are fused to obtain a second probability distribution result relative to each measurement position to be determined.
5. The method according to claim 1, characterized in that The step of obtaining a distribution network data set and preprocessing the distribution network data set to obtain a data set to be fused includes: Get the initialized distribution network dataset; Based on the data set format corresponding to the distribution network data set, power transformation processing and standardization processing are performed on the distribution network data set to obtain the data set to be fused.
6. The method according to claim 1, characterized in that The second probability allocation result includes a target probability value relative to each measurement position to be determined, and the method further includes: The to-be-determined measurement position corresponding to the target probability value satisfying the position determination condition in the second probability allocation result is determined as the target measurement position.
7. A data fusion device, characterized in that: include: A data set acquisition module is used to acquire a distribution network data set and pre-process the distribution network data set to obtain a data set to be fused; A probability allocation module, configured to perform probability allocation on the data set to be fused based on the feature dimensions corresponding to the data set to be fused, and determine a number of first probability allocation results corresponding to the number of feature dimensions; The feature data fusion module is used to fuse the first probability distribution results to obtain a second probability distribution result.
8. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform a data fusion method according to any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement a data fusion method according to any one of claims 1 to 6 when executed.
10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the computer program implements a data fusion method according to any one of claims 1-6.
Citation Information
Cited By
Multi-source reactive compensation control method, system, program, medium and equipment
CN119231559A
