Harmonic emission level estimation method, device, equipment, medium and program product

Abstract

The invention provides a harmonic emission level estimation method, device and equipment, a medium and a program product, and relates to the technical field of harmonic detection, and the method comprises the steps: estimating the harmonic emission level of a PCC point accessed by a harmonic source based on a trained BP neural network optimized by a cuckoo search algorithm; wherein the trained BP neural network optimized by the cuckoo search algorithm is obtained by training the BP neural network optimized by the cuckoo search algorithm on the basis of a data set constructed by active power discrete data, reactive power discrete data and THDU tag values. The method improves the utilization rate of mass monitoring data of the harmonic source, gets rid of the dependence of an impedance estimation method for estimating the harmonic emission level on network topology and multi-point monitoring, and improves the accuracy of harmonic emission level estimation.

Description

Technical Field

[0001] The present invention relates to the field of harmonic detection technology, and in particular to a harmonic emission level estimation method, device, equipment, medium and program product. Background Art

[0002] Harmonic pollution is a common power quality problem in distribution networks and microgrids. In recent years, the number of distributed power sources and nonlinear loads has continued to increase, making the harmonic problem of public power grids increasingly prominent. Accurately estimating the harmonic emission level plays a vital role in the division of harmonic pollution responsibilities at the Point of Common Coupling (PCC).

[0003] Most of the existing harmonic emission level assessment methods are based on the impedance estimation of the system, which can be basically divided into two categories. One is "interventional", which is to estimate the harmonic impedance by injecting harmonic current into the system or disconnecting a branch. Although the intervention method can obtain effective instantaneous data during the experiment, its method will have an adverse effect on the normal operation of the power system, thus limiting its scope of use. The other is "non-interventional", which is to use existing harmonic sources or measurable harmonic parameters to calculate the harmonic impedance. It is easy to implement and will not affect the normal operation of the system. Therefore, the non-intervention method has been widely adopted.

[0004] However, non-intrusive impedance estimation methods require a full understanding of the system and electrical equipment topology, and the network parameters of some nodes in the system are unmeasurable, which reduces the accuracy of impedance estimation and further reduces the accuracy of harmonic emission level estimation. Summary of the invention

[0005] The present invention provides a harmonic emission level estimation method, device, equipment, medium and program product, which are used to solve the defect of reducing the accuracy of harmonic emission level estimation in the prior art and improve the accuracy of harmonic emission level estimation.

[0006] The present invention provides a harmonic emission level estimation method, comprising:

[0007] Based on the trained BP neural network optimized by cuckoo search algorithm, the harmonic emission level of the PCC point connected to the harmonic source is estimated;

[0008] The trained BP neural network optimized by the cuckoo search algorithm is based on the active power discrete data, reactive power discrete data and THD U The data set constructed by the label value is obtained by training the BP neural network optimized by the cuckoo search algorithm.

[0009] In some embodiments, before estimating the harmonic emission level of the PCC point connected to the harmonic source based on the trained BP neural network optimized by the cuckoo search algorithm, the method further includes:

[0010] Based on the cuckoo search algorithm, the initial weight value and the bias value of the BP neural network are optimized to obtain the optimal solution of the initial weight value and the optimal solution of the bias value;

[0011] The optimal solution of the initial weight value and the optimal solution of the bias value are assigned to the BP neural network to obtain the BP neural network optimized by the cuckoo search algorithm.

[0012] In some embodiments, before estimating the harmonic emission level of the PCC point connected to the harmonic source based on the trained BP neural network optimized by the cuckoo search algorithm, the method further includes:

[0013] Perform power measurement and harmonic measurement on the PCC point where the harmonic source is connected to obtain continuous data of active power, continuous data of reactive power and THD U Continuous data;

[0014] Discretizing the active power continuous data and the reactive power continuous data respectively to obtain active power discrete data and reactive power discrete data corresponding to each segment;

[0015] Determine the THD corresponding to each segment according to the active power discrete data and the reactive power discrete data corresponding to each segment U Continuous data;

[0016] Based on the THD corresponding to each segment U Continuous data, determine the THD corresponding to each segment U Tag value;

[0017] According to the active power discrete data and reactive power discrete data corresponding to all segments, as well as the THD corresponding to all segments U Label values, construct the dataset.

[0018] In some embodiments, the step of discretizing the active power continuous data and the reactive power continuous data to obtain active power discrete data and reactive power discrete data corresponding to each segment includes:

[0019] Splitting the active power continuous data into the active power continuous data corresponding to a plurality of segments respectively;

[0020] Taking the minimum Euclidean distance between the active power continuous data and the active power discrete data corresponding to each moment as the discretization target, a particle swarm algorithm is used for solving to obtain the active power discrete data corresponding to each segment;

[0021] Splitting the reactive power continuous data into the reactive power continuous data corresponding to a plurality of segments respectively;

[0022] The minimization of the Euclidean distance between the reactive power continuous data and the reactive power discrete data corresponding to each moment is taken as the discretization target, and a particle swarm algorithm is used to solve the problem, so as to obtain the reactive power discrete data corresponding to each segment.

[0023] In some embodiments, the THD corresponding to each segment U Continuous data, determine the THD corresponding to each segment U Tag values ​​include

[0024] The THD corresponding to each segment U For continuous data, the average value, maximum value and minimum value are calculated to obtain the THD corresponding to each segment. U Average value, THD U Maximum value and THD U Minimum value;

[0025] The THD corresponding to each segment U Average value, THD U Maximum value and THD U The minimum value is rounded down to get three rounded-down results corresponding to each segment;

[0026] Based on the three round-down results corresponding to each segment, determine the THD corresponding to each segment U Tag value.

[0027] In some embodiments, the THD corresponding to each segment is determined based on the three round-down results corresponding to each segment. U Tag values, including:

[0028] For the current segment, if the three rounding-down results corresponding to the current segment are all the same integer value, the same integer value is used as the THD corresponding to the current segment. U Tag value;

[0029] When the three rounding-down results corresponding to the current segment are not the same integer value, determine the Euclidean distance between all the active power discrete data and the corresponding active power continuous data, or the Euclidean distance between all the reactive power discrete data and the corresponding reactive power continuous data in the current segment;

[0030] The Euclidean distances between all the active power discrete data and the corresponding active power continuous data, or the Euclidean distances between all the reactive power discrete data and the corresponding reactive power continuous data, are sorted in ascending order, and the THD corresponding to the preset number of Euclidean distances ranked first is selected. U Continuous data;

[0031] According to the THD screened U The average value, maximum value and minimum value corresponding to the continuous data determine the THD corresponding to the current segment U Tag value;

[0032] Traverse all segments and determine the THD corresponding to each segment U Tag value.

[0033] The present invention also provides a harmonic emission level estimation device, comprising:

[0034] An estimation module, used for estimating the harmonic emission level of the PCC point connected to the harmonic source based on the trained BP neural network optimized by the cuckoo search algorithm;

[0035] The trained BP neural network optimized by the cuckoo search algorithm is based on the active power discrete data, reactive power discrete data and THD U The data set constructed by the label value is obtained by training the BP neural network optimized by the cuckoo search algorithm.

[0036] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the above-mentioned methods for estimating harmonic emission levels is implemented.

[0037] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for estimating the harmonic emission level as described above is implemented.

[0038] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for estimating the harmonic emission level as described above is implemented.

[0039] The harmonic emission level estimation method, device, equipment, medium and program product provided by the present invention first optimize the BP neural network through the cuckoo search algorithm to obtain the BP neural network optimized by the cuckoo search algorithm, so as to effectively avoid the problem that the BP neural network is easy to fall into the local optimum due to the initial weight value and bias value, and at the same time improve the convergence speed of the BP neural network; then, the active power discrete data, the reactive power discrete data and the THD are used to estimate the harmonic emission level. U The dataset constructed by label values ​​is used to train the BP neural network optimized by the cuckoo search algorithm to obtain the trained BP neural network optimized by the cuckoo search algorithm, which improves the utilization rate of massive monitoring data of harmonic sources and gets rid of the dependence of the impedance estimation method on network topology and multi-point monitoring when estimating the harmonic emission level. Finally, the harmonic emission level of the PCC point connected to the harmonic source is estimated by the trained BP neural network optimized by the cuckoo search algorithm, which improves the accuracy of the harmonic emission level estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 It is one of the flow charts of the harmonic emission level estimation method provided by the present invention;

[0042] Figure 2 This is the second flow chart of the harmonic emission level estimation method provided by the present invention;

[0043] Figure 3 It is a structural schematic diagram of the harmonic emission level estimation device provided by the present invention;

[0044] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other in the absence of conflict. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0046] In order to facilitate a clearer understanding of the embodiments of the present invention, some relevant technical knowledge is first introduced as follows.

[0047] (1) BP Neural Network

[0048] Back propagation neural network (BPNN) is a concept proposed by scientists led by Rumelhart and McClelland in 1986. It is a multi-layer feedforward neural network trained according to the error back propagation algorithm and is the most widely used neural network.

[0049] The BP neural network adds several layers (one or more layers) of neurons between the input layer and the output layer. These neurons are called hidden units. They have no direct connection with the outside world, but changes in their states can affect the relationship between the input and output. Each layer can have several nodes.

[0050] (2) Cuckoo search algorithm

[0051] The cuckoo search (CS) algorithm is a new heuristic algorithm proposed by Professor Xinshe Yang and S. Deb of Cambridge University in 2009. According to long-term observations and studies by entomologists, some cuckoos raise their young in a parasitic way. Instead of building nests, they lay their eggs in the nests of other birds, which then hatch and raise their chicks. However, if these foreign bird eggs are discovered by the host, the host will abandon them or build a new nest.

[0052] Yang et al. assumed three ideal states for cuckoo egg-laying behavior:

[0053] 1) Cuckoos lay only one egg at a time and randomly choose a nest location to hatch it;

[0054] 2) Among a group of randomly selected bird nests, the best one will be kept for the next generation;

[0055] 3) The number of parasitic nests that can be selected is fixed, and the probability that the owner of the parasitic nest finds foreign bird eggs is P a , where 0≤P a ≤1.

[0056] Based on these three ideal states, Yang et al. used Levy flight to predict the location of the next generation of bird nests. To update:

[0057]

[0058] Where: represents the position of the i-th bird nest in the t+1th generation, i∈[1,2,…,n], and n represents the total number of bird nests in each generation; represents the position of the i-th bird's nest in the t-th generation; Represents point-to-point multiplication; α represents the step size control amount, which is used to control the step size. Usually, α=1; Levy(λ) represents the search path of Levy flight; λ is the input parameter.

[0059] Levy(λ) is the Levy distribution:

[0060] Levy:μ=t -λ ,1≤λ≤3

[0061] In the formula, μ represents the output result of the Levy() function, t represents the tth generation, λ represents the input parameter of the Levy() function, and the value range of λ is [1,3].

[0062] The specific process of the cuckoo search algorithm is:

[0063] Step 1: Set the number of bird nests to n, the dimension of the search space to d, and initialize the location of the bird nests And find the best nest location and the optimal solution f min .

[0064] Step 2: Loop body:

[0065] (1) Keep the best nest location of the previous generation t is an integer;

[0066] (2) Using Levy flight to obtain a new set of bird nest locations

[0067] (3) The location of the bird's nest corresponding to the previous generation For example, and For comparison, and Compare and replace the bird's nest positions with better fitness values ​​with the bird's nest positions with worse fitness values, so as to obtain a set of better bird's nest positions

[0068] (4) After the location is updated, the probability P of the host discovering foreign bird eggs is calculated by using the random number R∈[0,1] a Compare, if R>P a , then for g t The nest position in the random change, retaining g tThe position of the bird's egg with the lowest probability is found in the random change of the position of the bird's nest with the highest probability, and a new set of bird's nest positions is obtained, which is consistent with g t Compare the bird's nest positions in the test, replace the bird's nest positions with poor test values ​​with the bird's nest positions with better test values, and get a set of better bird's nest positions

[0069] Step 3, from the final obtained P t Determine the optimal nest location and the optimal solution f min If the iteration condition (the specified number of iterations or the required accuracy) is met, the global optimal value f is output min and the global optimal nest location Otherwise, return to step 2 to continue iterative updating.

[0070] The cuckoo search algorithm has a balance between global exploration and local exploitation performance as well as population diversity.

[0071] Figure 1 is one of the flow charts of the harmonic emission level estimation method provided by the present invention, such as Figure 1 As shown, the present invention provides a method for estimating harmonic emission levels, the method comprising the following steps:

[0072] Step 110, estimating the harmonic emission level of the PCC point to which the harmonic source is connected based on the trained BP neural network optimized by the cuckoo search algorithm;

[0073] Among them, the trained BP neural network optimized by the cuckoo search algorithm is based on the active power discrete data, reactive power discrete data and THD U The data set constructed by label values ​​is obtained by training the BP neural network optimized by the cuckoo search algorithm.

[0074] Specifically, the cuckoo search algorithm has the characteristics of improving local and global optimization. In order to improve the dependence of the BP neural network on the initial weight value and bias value, the cuckoo search algorithm is applied to the BP neural network to obtain the BP neural network optimized by the cuckoo search algorithm.

[0075] After obtaining the BP neural network optimized by the cuckoo search algorithm, the active power discrete data, reactive power discrete data and voltage total harmonic distortion (THD) are used to calculate the voltage total harmonic distortion (THD) U ) label values ​​are divided into a training data set and a test data set, and the BP neural network optimized by the cuckoo search algorithm is trained and the accuracy is tested to obtain a trained BP neural network optimized by the cuckoo search algorithm.

[0076] After obtaining the trained BP neural network optimized by the cuckoo search algorithm, the trained BP neural network optimized by the cuckoo search algorithm is applied to the estimation of the harmonic emission level of the PCC point connected to the harmonic source. The specific process can be:

[0077] Perform real-time power monitoring on the PCC point connected to the harmonic source to obtain continuous active power data and continuous reactive power data; input the continuous active power data and continuous reactive power data into the trained BP neural network optimized by the cuckoo search algorithm to obtain the THD output of the trained BP neural network optimized by the cuckoo search algorithm. U Tag estimate (i.e., harmonic emission level estimate).

[0078] The harmonic emission level estimation method provided by the embodiment of the present invention first optimizes the BP neural network through the cuckoo search algorithm to obtain the BP neural network optimized by the cuckoo search algorithm, so as to effectively avoid the problem that the BP neural network is easy to fall into the local optimum due to the initial weight value and the bias value, and at the same time improve the convergence speed of the BP neural network; then, the active power discrete data, the reactive power discrete data and the THD are used to obtain the harmonic emission level estimation method; U The dataset constructed by label values ​​is used to train the BP neural network optimized by the cuckoo search algorithm to obtain the trained BP neural network optimized by the cuckoo search algorithm, which improves the utilization rate of massive monitoring data of harmonic sources and gets rid of the dependence of the impedance estimation method on network topology and multi-point monitoring when estimating the harmonic emission level. Finally, the harmonic emission level of the PCC point connected to the harmonic source is estimated by the trained BP neural network optimized by the cuckoo search algorithm, which improves the accuracy of the harmonic emission level estimation.

[0079] In some embodiments, before estimating the harmonic emission level of the PCC point connected to the harmonic source based on the trained BP neural network optimized by the cuckoo search algorithm, the method further includes:

[0080] Based on the cuckoo search algorithm, the initial weight value and bias value of the BP neural network are optimized to obtain the optimal solution of the initial weight value and the optimal solution of the bias value;

[0081] The optimal solution of the initial weight value and the optimal solution of the bias value are assigned to the BP neural network to obtain the BP neural network optimized by the cuckoo search algorithm.

[0082] Specifically, the specific process of optimizing the initial weight value and bias value of the BP neural network based on the cuckoo search algorithm is as follows:

[0083] (1) Each BP neural network is regarded as a bird's nest position, and the initial weight value and bias value of each BP neural network are initialized. The weight value and bias value of the BP neural network are the parameters to be optimized.

[0084] (2) Perform forward propagation on each BP neural network, calculate the output of each BP neural network, and calculate the fitness value corresponding to each BP neural network. The fitness value can be evaluated based on the performance indicators of the BP neural network (such as mean square error).

[0085] (3) Compare with the previous generation of BP neural network, and replace the BP neural network with poor fitness value with the BP neural network with better fitness value, so as to obtain a set of better BP neural network.

[0086] (4) Select the weight value and bias value of the BP neural network with the highest fitness value from the better BP neural networks as the optimal solution of the current initial weight value and the optimal solution of the bias value.

[0087] (5) Perturb the optimal solution of the current initial weight value and the optimal solution of the bias value to generate a new BP neural network and calculate its fitness value. If the fitness value of the newly generated BP neural network is better, replace the BP neural network with the worst fitness value in the better BP neural network; otherwise, discard the newly generated BP neural network.

[0088] (6) Repeat steps 2 to 5 until the stopping condition is met (e.g., the maximum number of iterations is reached or a satisfactory solution is found).

[0089] (7) Return the BP neural network with the highest fitness value, and use the weight value and bias value of the BP neural network with the highest fitness value as the final optimal solution of the initial weight value and the optimal solution of the bias value.

[0090] The optimal solution of the initial weight value and the optimal solution of the bias value are assigned to the BP neural network to obtain the BP neural network optimized by the cuckoo search algorithm.

[0091] The harmonic emission level estimation method provided in the embodiment of the present invention optimizes the initial weight values ​​and bias values ​​of the BP neural network through the cuckoo search algorithm, further effectively avoiding the problem that the BP neural network is easily trapped in the local optimum due to the initial weight values ​​and bias values, and at the same time can improve the convergence speed of the BP neural network.

[0092] In some embodiments, Figure 2 FIG. 2 is a flow chart of the harmonic emission level estimation method provided by the present invention. Figure 2 As shown, before estimating the harmonic emission level of the PCC point connected to the harmonic source based on the trained BP neural network optimized by the cuckoo search algorithm, it also includes:

[0093] Step 210: perform power measurement and harmonic measurement on the PCC point connected to the harmonic source to obtain continuous data of active power, continuous data of reactive power and THD. U Continuous data;

[0094] Step 220, discretizing the active power continuous data and the reactive power continuous data respectively to obtain active power discrete data and reactive power discrete data corresponding to each segment;

[0095] Step 230: Determine the THD corresponding to each segment based on the active power discrete data and reactive power discrete data corresponding to each segment. U Continuous data;

[0096] Step 240, based on the THD corresponding to each segment U Continuous data, determine the THD corresponding to each segment U Tag value;

[0097] Step 250: according to the active power discrete data and reactive power discrete data corresponding to all the segments, and the THD corresponding to all the segments U Integer value, constructs the dataset.

[0098] Specifically, a portable power meter for measuring harmonics is used to measure power and harmonics at the PCC point where the harmonic source is connected, and obtain continuous data on active power, continuous data on reactive power and THD. U Continuous data.

[0099] In order to improve the practicability of parameters and suppress the random fluctuation of continuous parameters, the active power continuous data and reactive power continuous data are discretized respectively to obtain the active power discrete data and reactive power discrete data corresponding to each segment.

[0100] The obtained discrete data (active power discrete data and reactive power discrete data) are compared with the measured THD U Continuous data is segmented to establish discrete data (active power discrete data and reactive power discrete data) and THD U The correspondence between continuous data determines the THD corresponding to each segment U Continuous data. Different discrete data (active power discrete data and reactive power discrete data) correspond to THD that varies within a certain range. U Continuous data.

[0101] Based on the THD corresponding to each segment U Continuous data, determine the THD corresponding to each segment U Label value. Optionally, the THD corresponding to each segment UThe continuous data is averaged and the average value is used as the THD corresponding to each segment. U Label value. Optionally, the THD corresponding to each segment U The standard deviation of continuous data is calculated and the obtained standard deviation is used as the THD corresponding to each segment. U Tag value.

[0102] According to the active power discrete data and reactive power discrete data corresponding to all segments, as well as the THD corresponding to all segments U The label value is used to construct a data set. The data set can be regarded as a discrete data of active power, discrete data of reactive power and THD. U A matrix with label values ​​as columns.

[0103] The harmonic emission level estimation method provided by the embodiment of the present invention obtains continuous data of active power, continuous data of reactive power and THD U Continuous data; Discretize the active power continuous data and reactive power continuous data respectively to obtain the active power discrete data and reactive power discrete data corresponding to each segment, aggregate the data characteristics and reduce the amount of calculation, while suppressing the random fluctuation of continuous data caused by load fluctuation and measurement noise; Determine the THD corresponding to each segment based on the active power discrete data and reactive power discrete data corresponding to each segment U Continuous data; based on the THD corresponding to each segment U Continuous data, determine the THD corresponding to each segment U Label value; based on the active power discrete data and reactive power discrete data corresponding to all segments, as well as the THD corresponding to all segments U Integer values ​​are used to construct a data set, so that the data set can be used to train the BP neural network optimized by the cuckoo search algorithm, which is further conducive to the subsequent use of the trained BP neural network optimized by the cuckoo search algorithm to estimate the harmonic emission level of the PCC point connected to the harmonic source.

[0104] In some embodiments, the active power continuous data and the reactive power continuous data are discretized respectively to obtain the active power discrete data and the reactive power discrete data corresponding to each segment, including:

[0105] Splitting the active power continuous data into active power continuous data corresponding to a plurality of segments;

[0106] Taking the minimum Euclidean distance between the active power continuous data and the active power discrete data corresponding to each moment as the discretization target, the particle swarm algorithm is used to solve the problem and obtain the active power discrete data corresponding to each segment.

[0107] Splitting the reactive power continuous data into reactive power continuous data corresponding to a plurality of segments;

[0108] Taking the minimum Euclidean distance between the continuous reactive power data and the discrete reactive power data corresponding to each moment as the discretization target, the particle swarm algorithm is used to solve the problem and the discrete reactive power data corresponding to each segment is obtained.

[0109] Specifically, the discrete process of active power continuous data is taken as an example to illustrate that the active power continuous data is divided into M segments. Optionally, for the active power continuous data within 1 day, M does not exceed 24, because, in actual steady-state operation, the load characteristics and composition change little within an hour. M segments correspond to M-1 segmentation points, which are recorded as:

[0110] m=[m1,m2,…,m M-1 ] T

[0111] In the formula, m represents the column vector composed of M-1 segmentation points, m M-1 Indicates the M-1th segmentation point.

[0112] Assuming that the active power is recorded every 3 minutes for 1 day, the active power continuous data can be represented as a 480×1 column vector Similarly, the active power discrete data can be represented as a 480×1 column vector It is expressed as:

[0113]

[0114] In the formula, Indicates active power discrete data, M indicates the number of segments, Represents the active power discrete data of the Mth segment.

[0115] The goal of discretization is to minimize the difference between the continuous data of active power and the discrete data of active power, that is, to minimize the Euclidean distance between the continuous data of active power and the discrete data of active power corresponding to each time t. Therefore, the discrete objective function is set as:

[0116]

[0117] In the formula, Represents active power discrete data, represents the active power continuous data, and m represents the column vector consisting of M-1 segmentation points.

[0118] Taking the maximum active power value recorded during the monitoring time as the reference value, the monitored active power continuous data is converted into a per-unit value, so the values ​​of the active power continuous data and the active power discrete data are both within [0,1]. The constraints of the above discrete objective function are:

[0119]

[0120] In the formula, represents the active power discrete data, M represents the number of segments, m1 represents the order of the active power discrete data corresponding to the first segment point, and m M-1 Indicates the order of the active power discrete data corresponding to the M-1th segment point.

[0121] The solution of the above discrete objective function belongs to a mixed integer programming problem. The particle swarm algorithm is used to solve the discrete objective function to determine the discrete parameters and scattered points, that is, to determine the active power discrete data corresponding to each segment.

[0122] The discrete process of reactive power continuous data is similar to that of active power continuous data, and will not be described again here to avoid repetition.

[0123] The harmonic emission level estimation method provided by the embodiment of the present invention splits the active power continuous data into multiple segments, takes the minimum Euclidean distance between the active power continuous data corresponding to each moment and the active power discrete data as the discretization target, adopts the particle swarm algorithm to solve, obtains the active power discrete data corresponding to each segment, realizes the discretization of the active power continuous data, and is beneficial to the construction of subsequent data sets; splits the reactive power continuous data into multiple segments, takes the minimum Euclidean distance between the reactive power continuous data corresponding to each moment and the reactive power discrete data as the discretization target, adopts the particle swarm algorithm to solve, obtains the reactive power discrete data corresponding to each segment, realizes the discretization of the reactive power continuous data, and is beneficial to the construction of subsequent data sets.

[0124] In some embodiments, based on the THD corresponding to each segment U Continuous data, determine the THD corresponding to each segment U Tag values, including:

[0125] THD corresponding to each segment U For continuous data, the average value, maximum value and minimum value are calculated to obtain the THD corresponding to each segment. U Average value, THD U Maximum value and THD U Minimum value;

[0126] THD corresponding to each segment U Average value, THDU Maximum value and THD U The minimum value is rounded down to get three rounded-down results corresponding to each segment;

[0127] Based on the 3 floor results for each segment, determine the THD for each segment U Tag value.

[0128] Specifically, THD U The calculation formula is as follows:

[0129]

[0130] Where U h It represents the hth harmonic amplitude of voltage, U1 represents the fundamental amplitude of voltage, and H represents the total number of harmonics.

[0131] The THD corresponding to each segment U The average, maximum and minimum values ​​of continuous data are calculated to obtain the THD corresponding to each segment. U Average value C1, THD U Maximum C2 and THD U Minimum value C3.

[0132] THD U The calculation formula of the average value C1 is as follows:

[0133]

[0134] THD U The maximum value C2 is calculated as follows:

[0135] C2=max{THD U,1 ,THD U,2 ,…,THD U,γ}

[0136] THD U The calculation formula for the minimum value C3 is as follows:

[0137] C3=min{THD U,1 ,THD U,2 ,…,THD U,Υ}

[0138] Where γ represents the THD in the current segment U Total amount of continuous data, THD U,Υ Indicates the Yth THD in the current segment U Continuous data.

[0139] THD corresponding to each segment U Average value, THDU Maximum value and THD U The minimum value is rounded down to get the 3 rounded down results corresponding to each segment. Based on the 3 rounded down results corresponding to each segment, determine the THD corresponding to each segment U Optionally, you can average the three rounding-down results corresponding to each segment, round down the average value, and use the rounding-down result corresponding to the average value as the THD corresponding to each segment. U Tag value.

[0140] The harmonic emission level estimation method provided by the embodiment of the present invention is to calculate the THD corresponding to each segment. U For continuous data, the average value, maximum value and minimum value are calculated to obtain the THD corresponding to each segment. U Average value, THD U Maximum value and THD U Minimum value; THD corresponding to each segment U Average value, THD U Maximum value and THD U The minimum value is rounded down to obtain three rounding results corresponding to each segment; based on the three rounding results corresponding to each segment, the THD corresponding to each segment is determined U The label value is conducive to the construction of subsequent data sets.

[0141] In some embodiments, the THD corresponding to each segment is determined based on the three floor results corresponding to each segment. U Tag values, including:

[0142] For the current segment, if the three rounding-down results corresponding to the current segment are all the same integer value, the same integer value is used as the THD corresponding to the current segment. U Tag value;

[0143] When the three rounding-down results corresponding to the current segment are not the same integer value, determine the Euclidean distance between all active power discrete data and the corresponding active power continuous data, or the Euclidean distance between all reactive power discrete data and the corresponding reactive power continuous data in the current segment;

[0144] Sort the Euclidean distances between all active power discrete data and the corresponding active power continuous data, or the Euclidean distances between all reactive power discrete data and the corresponding reactive power continuous data, in ascending order, and select the THDs corresponding to the preset number of Euclidean distances that are ranked first. U Continuous data;

[0145] According to the THDU The average value, maximum value and minimum value corresponding to the continuous data determine the THD corresponding to the current segment U Tag value;

[0146] Traverse all segments and determine the THD corresponding to each segment U Tag value.

[0147] Specifically, for the current segment, if the three rounding-down results corresponding to the current segment are all the same integer value, the same integer value is used as the THD corresponding to the current segment. U Tag value.

[0148] When the three rounding-down results corresponding to the current segment are not the same integer value, determine the Euclidean distance between all active power discrete data and the corresponding active power continuous data, or the Euclidean distance between all reactive power discrete data and the corresponding reactive power continuous data in the current segment.

[0149] The calculation formula of Euclidean distance is as follows:

[0150]

[0151] Where, the subscript X represents active power or reactive power; d X Indicates the Euclidean distance between active power discrete data and active power continuous data, or the Euclidean distance between reactive power discrete data and reactive power continuous data; Indicates active power discrete data or reactive power discrete data, Indicates active power continuous data or reactive power continuous data.

[0152] Sort all Euclidean distances corresponding to the current segment (the Euclidean distances between all active power discrete data and the corresponding active power continuous data, or the Euclidean distances between all reactive power discrete data and the corresponding reactive power continuous data) in ascending order, and select the THD corresponding to the Euclidean distances with a preset number (e.g., the first 60%) of the top order. U Continuous data.

[0153] Since active power affects THD U The influence of is greater, so it is preferred to filter out THD according to the Euclidean distance between the current segmented active power discrete data and the active power continuous data. U Continuous data.

[0154] The THD of the screened U Calculate the average, maximum and minimum values ​​of continuous data. UThe average value, maximum value, and minimum value corresponding to the continuous data are rounded down respectively. If the three rounding results are all the same integer value, then the integer value is the THD corresponding to the current segment U Label value; if the three rounded-down results are still not the same integer value, the filtered THD U The continuous data are screened again until the three rounding-down results are the same integer value.

[0155] According to the above steps, traverse all segments and determine the THD corresponding to each segment U Tag value.

[0156] The harmonic emission level estimation method provided by the embodiment of the present invention, if the three rounding-down results corresponding to the current segment are all the same integer value, the same integer value is used as the THD corresponding to the current segment U Label value; if the three rounding-down results corresponding to the current segment are not the same integer value, based on the Euclidean distance between all active power discrete data and the corresponding active power continuous data in the current segment, or the Euclidean distance between all reactive power discrete data and the corresponding reactive power continuous data, THD U Continuous data is filtered and the THD is filtered out U The average value, maximum value and minimum value corresponding to the continuous data determine the THD corresponding to the current segment U Tag value, realizing THD U The accuracy of the tag value further helps to improve the accuracy of the harmonic emission level estimation.

[0157] The harmonic emission level estimation device provided by the present invention is described below. The harmonic emission level estimation device described below and the harmonic emission level estimation method described above can be referred to each other.

[0158] Figure 3 is a schematic diagram of the structure of the harmonic emission level estimation device provided by the present invention, such as Figure 3 As shown, the present invention provides a harmonic emission level estimation device, the device comprising: an estimation module 310;

[0159] An estimation module 310 is used to estimate the harmonic emission level of the PCC point connected to the harmonic source based on the trained BP neural network optimized by the cuckoo search algorithm;

[0160] The trained BP neural network optimized by the cuckoo search algorithm is based on the active power discrete data, reactive power discrete data and THD U The data set constructed by the label value is obtained by training the BP neural network optimized by the cuckoo search algorithm.

[0161] In some embodiments, the apparatus further comprises:

[0162] An optimization module, used for optimizing the initial weight value and the bias value of the BP neural network based on the cuckoo search algorithm, to obtain an optimal solution for the initial weight value and the optimal solution for the bias value;

[0163] The first obtaining module is used to assign the optimal solution of the initial weight value and the optimal solution of the bias value to the BP neural network to obtain the BP neural network optimized by the cuckoo search algorithm.

[0164] In some embodiments, the apparatus further comprises:

[0165] Acquisition module, used to measure the power and harmonics of the PCC point connected to the harmonic source, and obtain continuous data of active power, continuous data of reactive power and THD U Continuous data;

[0166] The second obtaining module is used to discretize the active power continuous data and the reactive power continuous data respectively to obtain active power discrete data and reactive power discrete data corresponding to each segment;

[0167] The first determination module is used to determine the THD corresponding to each segment according to the active power discrete data and the reactive power discrete data corresponding to each segment. U Continuous data;

[0168] The second determination module is used to determine the THD corresponding to each segment based on the U Continuous data, determine the THD corresponding to each segment U Tag value;

[0169] A building module is used to calculate the active power discrete data and reactive power discrete data corresponding to all segments, and the THD corresponding to all segments. U Label values, construct the dataset.

[0170] In some embodiments, the second obtaining module is specifically used for:

[0171] Splitting the active power continuous data into the active power continuous data corresponding to a plurality of segments respectively;

[0172] Taking the minimum Euclidean distance between the active power continuous data and the active power discrete data corresponding to each moment as the discretization target, a particle swarm algorithm is used for solving to obtain the active power discrete data corresponding to each segment;

[0173] Splitting the reactive power continuous data into the reactive power continuous data corresponding to a plurality of segments respectively;

[0174] The minimization of the Euclidean distance between the reactive power continuous data and the reactive power discrete data corresponding to each moment is taken as the discretization target, and a particle swarm algorithm is used to solve the problem, so as to obtain the reactive power discrete data corresponding to each segment.

[0175] In some embodiments, the second determining module includes:

[0176] A submodule for obtaining the THD corresponding to each segment U For continuous data, the average value, maximum value and minimum value are calculated to obtain the THD corresponding to each segment. U Average value, THD U Maximum value and THD U Minimum value;

[0177] The rounding submodule is used to calculate the THD corresponding to each segment. U Average value, THD U Maximum value and THD U The minimum value is rounded down to get three rounded-down results corresponding to each segment;

[0178] A determination submodule is used to determine the THD corresponding to each segment based on the three rounding-down results corresponding to each segment. U Tag value.

[0179] In some embodiments, the determining submodule is specifically used to:

[0180] For the current segment, if the three rounding-down results corresponding to the current segment are all the same integer value, the same integer value is used as the THD corresponding to the current segment. U Tag value;

[0181] When the three rounding-down results corresponding to the current segment are not the same integer value, determine the Euclidean distance between all the active power discrete data and the corresponding active power continuous data, or the Euclidean distance between all the reactive power discrete data and the corresponding reactive power continuous data in the current segment;

[0182] The Euclidean distances between all the active power discrete data and the corresponding active power continuous data, or the Euclidean distances between all the reactive power discrete data and the corresponding reactive power continuous data, are sorted in ascending order, and the THD corresponding to the preset number of Euclidean distances ranked first is selected. U Continuous data;

[0183] According to the THD screened U The average value, maximum value and minimum value corresponding to the continuous data determine the THD corresponding to the current segment U Tag value;

[0184] Traverse all segments and determine the THD corresponding to each segment U Tag value.

[0185] It should be noted here that the above-mentioned harmonic emission level estimation device provided by the present invention can implement all the method steps implemented by the above-mentioned method embodiment, and can achieve the same technical effect. The parts and beneficial effects that are the same as the method embodiment in this embodiment will not be described in detail here.

[0186] Figure 4 is a schematic diagram of the structure of the electronic device provided by the present invention, such as Figure 4 As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the harmonic emission level estimation method, which includes:

[0187] Based on the trained BP neural network optimized by cuckoo search algorithm, the harmonic emission level of the PCC point connected to the harmonic source is estimated;

[0188] The trained BP neural network optimized by the cuckoo search algorithm is based on the active power discrete data, reactive power discrete data and THD U The data set constructed by the label value is obtained by training the BP neural network optimized by the cuckoo search algorithm.

[0189] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0190] On the other hand, the present invention further provides a computer program product, the computer program product comprising a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the harmonic emission level estimation method provided by the above methods, the method comprising:

[0191] Based on the trained BP neural network optimized by cuckoo search algorithm, the harmonic emission level of the PCC point connected to the harmonic source is estimated;

[0192] The trained BP neural network optimized by the cuckoo search algorithm is based on the active power discrete data, reactive power discrete data and THD U The data set constructed by the label value is obtained by training the BP neural network optimized by the cuckoo search algorithm.

[0193] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to perform the harmonic emission level estimation method provided by the above methods, the method comprising:

[0194] Based on the trained BP neural network optimized by cuckoo search algorithm, the harmonic emission level of the PCC point connected to the harmonic source is estimated;

[0195] The trained BP neural network optimized by the cuckoo search algorithm is based on the active power discrete data, reactive power discrete data and THD U The data set constructed by the label value is obtained by training the BP neural network optimized by the cuckoo search algorithm.

[0196] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0197] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0198] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0199] It should be further noted that, in this article, the terms "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0200] In the present invention, "at least one" means one or more, and "more than one" means two or more than two. The terms "first", "second", "third", "fourth", etc. (if any) in the present invention are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0201] In the embodiments of the present invention, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0202] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for estimating harmonic emission levels, characterized in that: include: Based on the trained BP neural network optimized by cuckoo search algorithm, the harmonic emission level of the PCC point connected to the harmonic source is estimated; The trained BP neural network optimized by the cuckoo search algorithm is based on the active power discrete data, reactive power discrete data and THD U The data set constructed by the label value is obtained by training the BP neural network optimized by the cuckoo search algorithm.

2. The harmonic emission level estimation method according to claim 1, characterized in that: Before estimating the harmonic emission level of the PCC point connected to the harmonic source based on the trained BP neural network optimized by the cuckoo search algorithm, the method further includes: Based on the cuckoo search algorithm, the initial weight value and the bias value of the BP neural network are optimized to obtain the optimal solution of the initial weight value and the optimal solution of the bias value; The optimal solution of the initial weight value and the optimal solution of the bias value are assigned to the BP neural network to obtain the BP neural network optimized by the cuckoo search algorithm.

3. The harmonic emission level estimation method according to claim 1, characterized in that: Before estimating the harmonic emission level of the PCC point connected to the harmonic source based on the trained BP neural network optimized by the cuckoo search algorithm, the method further includes: Perform power measurement and harmonic measurement on the PCC point where the harmonic source is connected to obtain continuous data of active power, continuous data of reactive power and THD U Continuous data; Discretizing the active power continuous data and the reactive power continuous data respectively to obtain active power discrete data and reactive power discrete data corresponding to each segment; Determine the THD corresponding to each segment according to the active power discrete data and the reactive power discrete data corresponding to each segment U Continuous data; Based on the THD corresponding to each segment U Continuous data, determine the THD corresponding to each segment U Tag value; According to the active power discrete data and reactive power discrete data corresponding to all segments, as well as the THD corresponding to all segments U Label values, construct the dataset.

4. The harmonic emission level estimation method according to claim 3, characterized in that: The step of discretizing the active power continuous data and the reactive power continuous data to obtain active power discrete data and reactive power discrete data corresponding to each segment includes: Splitting the active power continuous data into the active power continuous data corresponding to a plurality of segments respectively; Taking the minimum Euclidean distance between the active power continuous data and the active power discrete data corresponding to each moment as the discretization target, a particle swarm algorithm is used for solving to obtain the active power discrete data corresponding to each segment; Splitting the reactive power continuous data into the reactive power continuous data corresponding to a plurality of segments respectively; The minimization of the Euclidean distance between the reactive power continuous data and the reactive power discrete data corresponding to each moment is taken as the discretization target, and a particle swarm algorithm is used to solve the problem, so as to obtain the reactive power discrete data corresponding to each segment.

5. The harmonic emission level estimation method according to claim 3, characterized in that: The THD corresponding to each segment U Continuous data, determine the THD corresponding to each segment U Tag values ​​include The THD corresponding to each segment U For continuous data, the average value, maximum value and minimum value are calculated to obtain the THD corresponding to each segment. U Average value, THD U Maximum value and THD U Minimum value; The THD corresponding to each segment U Average value, THD U Maximum value and THD U The minimum value is rounded down to get three rounded-down results corresponding to each segment; Based on the three round-down results corresponding to each segment, determine the THD corresponding to each segment U Tag value.

6. The harmonic emission level estimation method according to claim 5, characterized in that: The THD corresponding to each segment is determined based on the three rounding-down results corresponding to each segment. U Tag values, including: For the current segment, if the three rounding-down results corresponding to the current segment are all the same integer value, the same integer value is used as the THD corresponding to the current segment. U Tag value; When the three rounding-down results corresponding to the current segment are not the same integer value, determine the Euclidean distance between all the active power discrete data and the corresponding active power continuous data, or the Euclidean distance between all the reactive power discrete data and the corresponding reactive power continuous data in the current segment; The Euclidean distances between all the active power discrete data and the corresponding active power continuous data, or the Euclidean distances between all the reactive power discrete data and the corresponding reactive power continuous data, are sorted in ascending order, and the THD corresponding to the preset number of Euclidean distances ranked first is selected. U Continuous data; According to the THD screened U The average value, maximum value and minimum value corresponding to the continuous data determine the THD corresponding to the current segment U Tag value; Traverse all segments and determine the THD corresponding to each segment U Tag value.

7. A harmonic emission level estimation device, characterized in that: include: An estimation module, used for estimating the harmonic emission level of the PCC point connected to the harmonic source based on the trained BP neural network optimized by the cuckoo search algorithm; The trained BP neural network optimized by the cuckoo search algorithm is based on the active power discrete data, reactive power discrete data and THD U The data set constructed by the label value is obtained by training the BP neural network optimized by the cuckoo search algorithm.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the harmonic emission level estimation method according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the harmonic emission level estimation method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the harmonic emission level estimation method according to any one of claims 1 to 6 is implemented.

Images