CN113991652B - Data-driven multi-output calculation method for short-circuit current of distribution network containing IIDG - Google Patents
Data-driven multi-output calculation method for short-circuit current of distribution network containing IIDG Download PDFInfo
- Publication number
- CN113991652B CN113991652B CN202111255829.2A CN202111255829A CN113991652B CN 113991652 B CN113991652 B CN 113991652B CN 202111255829 A CN202111255829 A CN 202111255829A CN 113991652 B CN113991652 B CN 113991652B
- Authority
- CN
- China
- Prior art keywords
- distribution network
- short
- output
- iidg
- circuit current
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000004364 calculation method Methods 0.000 title claims abstract description 47
- 238000000034 method Methods 0.000 claims abstract description 47
- 238000012549 training Methods 0.000 claims description 23
- 238000004422 calculation algorithm Methods 0.000 claims description 16
- 238000011156 evaluation Methods 0.000 claims description 13
- 238000012360 testing method Methods 0.000 claims description 10
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 claims description 8
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 claims description 8
- 238000005259 measurement Methods 0.000 claims description 8
- 230000004927 fusion Effects 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 230000001360 synchronised effect Effects 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 abstract description 6
- 229910052799 carbon Inorganic materials 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 6
- 230000006978 adaptation Effects 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 238000013507 mapping Methods 0.000 description 4
- 230000009466 transformation Effects 0.000 description 4
- 238000011426 transformation method Methods 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000011217 control strategy Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/10—Power transmission or distribution systems management focussing at grid-level, e.g. load flow analysis, node profile computation, meshed network optimisation, active network management or spinning reserve management
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a data-driven multi-output calculation method for short-circuit current of a distribution network containing IIDGs. A large number of IIDGs are connected into the power distribution network under the double-carbon target, so that the contradiction between the calculation speed and the accuracy is increasingly outstanding in the calculation of the short-circuit current of the power distribution network based on mechanism modeling. The invention provides a data-driven power distribution network short-circuit current multi-output regression calculation model and a calculation method, which are based on XGBoost, adopt an MTRS method to carry out multi-output modeling, realize simultaneous calculation of short-circuit currents of all branches on a power distribution network, solve the problem of rapid and accurate calculation of the short-circuit currents of the whole network of the power distribution network containing IIDG, have better calculation performance than a single-output model, and avoid the problem of the number of models existing in the single-output model.
Description
Technical Field
The invention belongs to the field of power systems, in particular to a method for calculating short-circuit currents of all branches, which is provided by adopting a research mode of data science to research a machine learning multi-output algorithm, establishing a short-circuit current calculation multi-output model of a distribution network containing IIDGs.
Background
The national double-carbon strategic goal promotes the inverter type distributed power supply (INVERTER INTERFACED distributed generator, IIDG) to gradually present the development trend of high duty ratio in the power distribution network. Unlike traditional synchronous motor power supply, IIDG has strong nonlinear fault characteristic, and is affected by factors such as running characteristic, access position, control strategy and the like before fault. And a large amount of IIDGs are connected, so that the traditional short-circuit current calculation method is difficult to directly apply, and research on the calculation method of the short-circuit current of the power distribution network containing the IIDGs is urgently needed.
The current research on IIDG fault characteristics and short-circuit current calculation methods is mostly physical modeling research based on mechanisms. Because IIDG mechanism models are different under different control strategies, the short-circuit current calculation methods are large in difference under different fault types, and a large number of iterations are needed in the calculation process. When the IIDG access number or the network scale is increased, the physical modeling method is greatly increased due to time consumption of iterative calculation, and even the situation of non-convergence is likely to occur. In order to improve the calculation speed, the IIDG model is subjected to extremely simple processing, for example, the IIDG model is characterized by 1.2-2 times of rated current, and the calculation accuracy is necessarily sacrificed. Therefore, the existing physical modeling method is applied to a large-scale power distribution network with high IIDG ratio, and the contradiction between the computing rapidity and the computing accuracy is difficult to solve.
In order to solve the contradiction, the applicant provides a mechanism and data fusion driven calculation method for the short-circuit current of the distribution network containing IIDG, and the calculation method is shown in the volume P41-48 of the period 01 of the year 41 of 2021 of electric power automation equipment. The method is based on a machine learning idea, establishes a single-output model containing IIDG distribution network short-circuit current calculation, and compares the single-output model with a detailed physical modeling method, so that the calculation error is similar or smaller, and the calculation speed is high. Meanwhile, the power grid scale and the IIDG access quantity have small influence on the calculation performance of the method. However, this method uses a learning model of a single output, and when it is necessary to calculate a plurality of measurement point current values in the power distribution network, it is necessary to train a respective model for each measurement point, and thus a problem of model quantity disaster occurs.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, and provides a pure data-driven multi-output calculation method for the short-circuit current of the distribution network containing IIDG, which integrates machine learning and data-driven modeling into the calculation of the short-circuit current of the distribution network, obtains the mapping relation between input characteristics and a plurality of output quantities by training a multi-output model, can improve the calculation speed under the condition of accurately calculating the short-circuit current of the distribution network containing IIDG, outputs all branch short-circuit currents, and solves the problem of excessive number of single-output models.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
The invention comprises the following steps:
s1: establishing a short-circuit current sample set of the distribution network containing IIDGs, wherein the sample set comprises a training set and a testing set;
S2: the multiple output models are compared through a large number of analyses, and an MTRS-XGBoost multiple output model is selected and built;
S3: training an MTRS-XGBoost model by using a training set, and ending the training when the loss function is not reduced in the iteration process and the evaluation index is not reduced;
s4: and calculating the short-circuit current of the distribution network containing the IIDG by using the trained MTRS-XGBoost multi-output model, and outputting all branch currents under the corresponding running state.
Preferably, in step S1, the source of the short-circuit current sample set of the distribution network containing IIDG is, but not limited to, matlab/Simulink software simulation and actual distribution network operation data. The ratio of training set to test set in the sample set is 4:1.
Depending on the actual distribution network equipment situation, i.e. whether a micro synchronous phasor measurement unit (mu PMU) is installed or not, different input characteristics may be selected, the distribution network equipped with mu PMU may select the voltage amplitude, the phase and the active and reactive power values in the distribution network, and the distribution network not equipped with mu PMU may select the current value in the grid as the input characteristic.
Preferably, the MTRS-XGBoost multiple-output model in step S2 converts the multiple-output problem into a plurality of single-output problems using a multi-objective regression model fusion algorithm (multi-target regressor stacking, MTRS), wherein the base learner adopts a XGBoost method.
Preferably, the loss function in step S3 uses the mean absolute percentage error (Mean Absolute Percentage Error, MAPE), and the evaluation index uses the MAPE and the mean absolute error (Mean Absolute Error, MAE). The evaluation index is obtained by calculating the true value y j and the calculated valueThe smaller the evaluation index value, the higher the model accuracy.
The beneficial effects of the invention are as follows: the invention provides a data-driven multi-output model for calculating the short-circuit current containing IIDG, which can ensure the accuracy of calculating the short-circuit current and meet the requirement of quick calculation, can output the short-circuit current of all branches of a power distribution network at the same time, can meet the actual engineering requirement, and can be popularized and applied to the calculation of the short-circuit current of the power distribution network containing IIDG.
Drawings
Fig. 1 is a flow chart of the data-driven multi-output model establishment for calculating the short-circuit current containing IIDG according to the present invention.
Fig. 2 is a wiring diagram of an IEEE 34 node distribution network system with four distributed power sources connected.
Fig. 3 is a graph comparing performance of different multi-output methods at a high data loss ratio.
Detailed Description
The invention is further described below with reference to the accompanying drawings, which comprise the following steps:
According to fig. 1, a flow chart is established for a data-driven short-circuit current multi-output model of a distribution network containing IIDG, wherein the multi-output model has the capability of simultaneously outputting all branch short-circuit currents, and the flow chart comprises the following steps:
s1: establishing a short-circuit current sample set of the distribution network containing IIDGs, wherein the sample set comprises a training set and a testing set;
In the step S1, the source of the IIDG-containing power distribution network short-circuit current sample set is, but not limited to, matlab/Simulink software simulation and actual power distribution network operation data. The ratio of training set to test set in the sample set is typically set to 4:1.
Depending on the actual distribution network installation, i.e. whether a miniature synchronous phasor measurement unit (mu PMU) is installed, different sample characteristics can be selected. For a power distribution network with K nodes and E branches, a system power supply access node is generally provided, and if R IIDG access nodes are provided, the input sample characteristics of the short circuit current calculation model are as follows: x 1 corresponds to the distribution network with the mu PMU and X 2 corresponds to the distribution network without the mu PMU:
In the formula, |V|, theta and P, Q are node voltage amplitude values, phases, branch active power and reactive power which can be directly obtained by the mu PMU; the subscript w represents a node number, and the number 1 is a system power supply access node; t represents a branch line number; f line denotes a faulty line number; f loc represents the position percentage of the fault position from the line head end; f type represents a fault type, and the invention is provided with a three-phase short circuit, and other types of faults can be also arranged; DG h is the h IIDG access node; Is the h IIDG capacity; i is the obtained current effective value of each branch without the mu PMU and with the current transformer; v is a voltage effective value obtained by installing a voltage transformer at the connection position of a system power supply and the IIDG; the V, theta and P, Q, I, V all contain A/B/C three-phase electrical values, so that the method can be applied to an asymmetric distribution network.
Besides taking the actual measurement conditions of the power distribution network into consideration, the two types of electric quantity are adopted as the characteristic information of the operation mode of the power distribution network, and fault information of fault positions and fault types is also needed to form the input characteristics of the sample. The difference of each characteristic value in the sample set is large, so that normalization processing is needed, and the convergence capacity in the training process is enhanced. The label of the sample is obtained through modeling simulation, and specifically, the steady-state value of the short-circuit current of each line corresponding to the input characteristic is obtained.
S2: establishing an MTRS-XGBoost multi-output model;
the subtask association method of multiple output regression can be roughly divided into two types, namely problem transformation and algorithm adaptation.
The core idea of the problem transformation is to transform the corresponding mapping relation into a plurality of single-output forms, and then splice the plurality of single-output models into a multi-output model so as to achieve the purpose of simultaneously outputting a plurality of predicted values. Algorithms for problem transformation include single-target method (ST), multiple-target regression model fusion algorithm (MTRS, multi-target regressor stacking), regression chain algorithm (RC), and the like. Aiming at different practical problems, different problem conversion methods are selected according to the correlation among output tasks, so that the optimal multi-output performance is achieved.
Because the phase angles of the short-circuit currents provided by the IIDG power supply and the system power supply are often different, the amplitude values of steady-state currents of all branches do not explicitly meet kirchhoff current law, but the current vectors still meet the kirchhoff current law, so that certain correlation exists between outputs, and an MTRS or RC method needs to be selected. In addition, the single-output model spliced by the problem transformation method is called a base learner, and under the same splicing method, the stronger the performance of the base learner, the better the splicing output effect. Currently, XGBoost, lightGBM and the like are base learners with better performance, and the base learners are more representative in the brand-new corner of each large competition in recent years.
The algorithm adaptation method is to modify a single-output model according to actual problems, and simultaneously output a plurality of predicted values by using one model. The output mapping of the algorithm adaptation method not only needs to consider input variables, but also needs to consider the relation among different output variables, so that the accuracy of a plurality of predicted values is improved. The algorithm adaptation method comprises a statistical method, a multi-output support vector machine, a multi-output regression tree, a kernel function method, a neural network and other methods. The algorithm is suitable for a single-output model before modification, and parameters such as input, model structure, output label, loss function and the like in the algorithm need to be modified according to different practical problems. The different algorithm adaptation methods are not exactly the same.
According to the invention, through a large number of analysis and comparison, the MTRS-XGBoost multi-output model with optimal comprehensive performance is finally selected. The multi-output model is to convert the multi-output problem into a plurality of single-output problems by using a multi-objective regression model fusion algorithm MTRS, wherein a base learner adopts a XGBoost method.
The MTRS method is a problem transformation method requiring two-stage processing. The sample of the first group is provided with N groups of samples, X l is the input of the first group, Y l is the output corresponding to the first group, namely the input and the output of the samples of the first group are as follows:
In the method, in the process of the invention, For the ith input feature in the first group, i e {1, …, M }, M being the number of input features; /(I)For the j-th output value in the first group, j ε {1, …, D }, D is the number of output labels. In the first order of MTRS, mapping between all outputs and inputs is established respectively, D single-output models are required to be established for D outputs, and each single-output model is responsible for a single output predicted value, that is, training set D MTRS-1 of the j-th output is:
the MTRS second order is to output the first order And original input quantity X l are combined into new input X *l:
The second order uses the new input X *l and the original output Y l as training sets of D output models, i.e., training set D MTRS-2 of the j-th output is:
the second order takes the first order output predicted value as the input of the training set, so that potential relations among a plurality of outputs are considered to a certain extent, the deviation of the problem transformation predicted value is corrected to a certain extent, and the prediction accuracy can be improved.
S3: training an MTRS-XGBoost model by using a training set, and ending the training when the loss function is not reduced in the iteration process and the evaluation index is not reduced;
The loss function in step S3 uses the mean absolute percentage error (Mean Absolute Percentage Error, MAPE), and the evaluation index uses the MAPE and the mean absolute error (Mean Absolute Error, MAE). The evaluation index is obtained by calculating the true value y j and the calculated value The smaller the evaluation index value, the higher the model accuracy. The evaluation indexes MAPE and MAE are defined as follows:
S4: the trained MTRS-XGBoost multi-output model is used as a multi-output model for calculating the short-circuit current of the distribution network containing IIDG, and when the input characteristics of the distribution network in a certain fault state are input, all branch short-circuit currents in the corresponding running state can be rapidly output.
Application example
In order to verify the effectiveness of the short-circuit current multi-output calculation method of the distribution network containing IIDGs based on data driving, the IEEE 34 node distribution network system shown in fig. 2 is taken as an example for verification. The node of the number 800 is a system power supply access point, the system is an unbalanced system, single-phase and three-phase feeder lines exist, and model parameter settings are the same as those in the literature [ Zheng Xiang, wang Huifang, jiang Kuan, and the like ]. Mechanism and data fusion driving method [ J ]. Electric power automation equipment, 2021, 41 (01): 41-48 ]. The running mode of the distribution network is determined by randomly generating equivalent power supply impedance, IIDG capacity and load of the system within [0.8,1.2] times of the original value.
The fault is set in the simulation to the fault at f 1 in fig. 2, i.e., a three-phase short circuit fault occurs on lines 834-860. In order to accurately explain the capability of simultaneously outputting short-circuit current in the multi-output model, the short-circuit current of three measuring points marked by ①、②、③ in the dashed line frame of fig. 2 is selected for calculation, and the data in the short-circuit current calculation method of the distribution network with IIDG driven by mechanism and data fusion of the reference P41-48 of the 01-41-year-2021 of the electric power automation equipment are adopted as a comparison group. Meanwhile, in order to verify the advantage of XGBoost of the base learner, lightGBM is selected as a contrast base learner in the same method, and the effectiveness of MTRS-XGBoost is demonstrated through comparison. Table 1 shows the predicted short-circuit current results and performance for various methods.
Table 1 Single output model and Multi-output model results
As can be seen from table 1, XGBoost single output models can only predict the short circuit current of a single measurement point, and the percentage error APE is within an acceptable range; the short-circuit current predicted by the multi-output model has lower percentage error than that of single output because of considering the relativity among the outputs. In addition, short-circuit currents at different measurement points have slightly different short-circuit current error conditions by using different base learner models. In conclusion, the multi-output model is adopted to calculate the short-circuit current of the distribution network containing the IIDG, so that the method is effective and higher in accuracy.
Corresponding discussions are also made regarding multiple-output model performance impact calculations among different multiple-output methods, base learners, and input features. And meanwhile, the influence of two input features of the formulas (1) and (2) on the calculation performance is compared. The multi-output method uses two problem transformation methods of MTRS and RC, and the base learner adopts two algorithms of XGBoost and LightGBM. The results are shown in Table 2, where training time is the time for model training to be completed using 8000 samples, and test time is the total consumption of completing 2000 sample tests, with bold values in the table indicating the performance optima for each multiple output model.
As can be seen from Table 2, the 4 models and the two input features described above achieved good performance, with maximum and average MAPE and MAE values within acceptable ranges. The calculation speed is more dependent on the adopted problem transformation method, the MTRS method is obviously slower than the RC method, the maximum test time is 11.75062s, namely, the maximum test time of each group is 5.875ms on average and is faster than that of a single-output model, and therefore, the scene with high calculation speed requirement can be met. In addition, the two types of features have little effect on the output accuracy of the multi-output model, and the feature of formula (2) can be used when the configuration of mu PMUs in a power distribution network has not been popularized. The XGBoost base learner has a slightly smaller error than the LightGBM base learner.
TABLE 2 comparison of Performance of different multiple output models, different base learners, different input features
In practical application, partial characteristic data loss may occur in the data acquisition process. The method with better comprehensive performance in the table 2 is selected, namely, the XGBoost algorithm of MTRS and RC splicing is respectively used, the characteristic of the formula (1) is selected, and the data loss is represented in a random zero setting mode. The performance of the two models at different data loss ratios is shown in table 3 and fig. 3, with the bolded values in table 3 representing smaller error data.
The data in table 3 shows that the error is small for both models when 1% of the data is lost. The multi-output model can still keep better performance under the condition of losing 5% and 10% of data volume, wherein the error of the MTRS model is slightly better than that of the RC model.
TABLE 3 comparison of the Performance of multiple output models at different data loss ratios
FIG. 3 shows MAPE-mean values for different multiple output models at data loss ratios above 10%. When the data loss is less than 30%, the error values of the MTRS and RC methods are smaller, and after the data loss proportion exceeds 30%, the average value of the errors of the MTRS and RC is rapidly increased. Therefore, under the condition of small loss of data, the MTRS and RC methods have better anti-interference capability, and when the data loss ratio is high, the errors of the two methods are increased rapidly along with the increase of the data loss ratio.
According to the application example, the data-driven multi-output calculation method for the short-circuit current of the distribution network containing IIDGs can simultaneously meet the two requirements of calculation accuracy and calculation rapidness, can simultaneously calculate the short-circuit current of all branches in one operation, can be applied to calculation of the short-circuit current of the distribution network system containing IIDGs, and is popularized to all stages of novel power systems with a large amount of new energy access, so that the rapid and accurate calculation of the short-circuit current is realized.
Claims (3)
1. The data-driven multi-output calculation method for the short-circuit current of the distribution network containing the IIDG is characterized by comprising the following steps of:
s1: establishing a short-circuit current sample set of the distribution network containing IIDGs, wherein the sample set comprises a training set and a testing set;
S2: selecting and establishing an MTRS-XGBoost multi-output model through a large number of analysis and comparison multi-output models, wherein the multi-output model has the capacity of outputting all branch short-circuit currents simultaneously;
S3: training an MTRS-XGBoost model by using a training set, and ending the training when the loss function is not reduced in the iteration process and the evaluation index is not reduced;
S4: calculating the short-circuit current of the distribution network containing IIDG by using a trained MTRS-XGBoost multi-output model, and simultaneously outputting all branch currents under the corresponding running state;
In step S1, the source of the sample set of the short-circuit current of the distribution network containing IIDG is, but not limited to, matlab/Simulink software simulation and actual distribution network operation data, and the ratio of the training set to the testing set in the sample set is 4:1, a step of;
According to the actual power distribution network equipment condition, namely whether a miniature synchronous phasor measurement unit mu PMU is installed or not, different input characteristics can be selected; the distribution network with the mu PMU can select voltage amplitude, phase and active and reactive power values in the distribution network, and the distribution network without the mu PMU can select current values in the power network as input characteristics;
For a power distribution network with K nodes and E branches, a system power supply access node is assumed to have R IIDG access nodes, and the input sample characteristics of the short circuit current calculation model are as follows: x 1 corresponds to the distribution network with the mu PMU and X 2 corresponds to the distribution network without the mu PMU:
In the formula, |V|, Q, P and Q are node voltage amplitude values, phases, branch active power and reactive power which can be directly obtained by the mu PMU; the subscript w represents a node number, and the number 1 is a system power supply access node; t represents a branch line number; f line denotes a faulty line number; f loc represents the position percentage of the fault position from the line head end; f type represents a fault type; DG h is the h IIDG access node; s DGh is the h IIDG capacity; i is the obtained current effective value of each branch without the mu PMU and with the current transformer; v is a voltage effective value obtained by installing a voltage transformer at the connection position of a system power supply and the IIDG; the I V I q, P, Q, I, V contain A/B/C three-phase electric quantity values, and can be applied to an asymmetric power distribution network.
2. The data-driven multi-output calculation method for the short-circuit current of the distribution network containing IIDG according to claim 1, wherein the method is characterized by comprising the following steps: the MTRS-XGBoost multi-output model in the step S2 is to convert the multi-output problem into a plurality of single-output problems by using a multi-objective regression model fusion algorithm MTRS, wherein a base learner adopts a XGBoost method.
3. The data-driven multi-output calculation method for the short-circuit current of the distribution network containing IIDG according to claim 1, wherein the method is characterized by comprising the following steps: the loss function in the step S3 adopts average absolute percentage error MAPE, and the evaluation index adopts average absolute percentage error MAPE and average absolute error MAE; the evaluation index is obtained by calculating the true value y j and the calculated valueThe smaller the evaluation index value, the higher the model accuracy.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111255829.2A CN113991652B (en) | 2021-10-27 | 2021-10-27 | Data-driven multi-output calculation method for short-circuit current of distribution network containing IIDG |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111255829.2A CN113991652B (en) | 2021-10-27 | 2021-10-27 | Data-driven multi-output calculation method for short-circuit current of distribution network containing IIDG |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113991652A CN113991652A (en) | 2022-01-28 |
| CN113991652B true CN113991652B (en) | 2024-05-17 |
Family
ID=79742587
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111255829.2A Active CN113991652B (en) | 2021-10-27 | 2021-10-27 | Data-driven multi-output calculation method for short-circuit current of distribution network containing IIDG |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113991652B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114971388B (en) * | 2022-06-20 | 2023-06-13 | 山东安能信息技术有限公司 | Power distribution network line loss fine management system based on big data |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111064182A (en) * | 2019-11-25 | 2020-04-24 | 国网浙江省电力有限公司湖州供电公司 | Short-circuit current calculation method based on power grid characteristics |
| CN111159841A (en) * | 2019-11-25 | 2020-05-15 | 国网浙江省电力有限公司湖州供电公司 | Power distribution network short-circuit current calculation method based on data fusion |
| CN111262238A (en) * | 2020-02-12 | 2020-06-09 | 浙江大学 | Machine learning-based method for predicting short-circuit current of power distribution network containing IIDG |
-
2021
- 2021-10-27 CN CN202111255829.2A patent/CN113991652B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111064182A (en) * | 2019-11-25 | 2020-04-24 | 国网浙江省电力有限公司湖州供电公司 | Short-circuit current calculation method based on power grid characteristics |
| CN111159841A (en) * | 2019-11-25 | 2020-05-15 | 国网浙江省电力有限公司湖州供电公司 | Power distribution network short-circuit current calculation method based on data fusion |
| CN111262238A (en) * | 2020-02-12 | 2020-06-09 | 浙江大学 | Machine learning-based method for predicting short-circuit current of power distribution network containing IIDG |
Non-Patent Citations (3)
| Title |
|---|
| A Novel Machine Learning-Based Short-Circuit Current Prediction Method for Active Distribution Networks;Xiang Zheng等;《energies》;1-15 * |
| DSTARS: A Multi-Target Deep Structure for Tracking Asynchronous Regressor Stack;Saulo Martiello Mastelini等;《2017 Brazilian Conference on Intelligent Systems》;19-24 * |
| 基于数据驱动的含IIDG配电网短路电流计算模型研究;郑翔;《中国优秀硕士学位论文全文数据库(工程科技Ⅱ辑)》;C042-204 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113991652A (en) | 2022-01-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Azmy et al. | Artificial neural network-based dynamic equivalents for distribution systems containing active sources | |
| CN105938578A (en) | Large-scale photovoltaic power station equivalent modeling method based on clustering analysis | |
| CN110311398B (en) | Connection topology, control system and method of novel energy storage battery system | |
| CN110797874A (en) | State estimation method for alternating current-direct current hybrid power distribution network containing power electronic transformer | |
| CN105762777A (en) | Pilot protection method containing multi-T-connection inverter interfaced distributed generation power distribution network | |
| Chen et al. | A novel fusion model based online state of power estimation method for lithium-ion capacitor | |
| CN111262238B (en) | Machine learning-based method for predicting short-circuit current of power distribution network containing IIDG | |
| CN113991652B (en) | Data-driven multi-output calculation method for short-circuit current of distribution network containing IIDG | |
| CN113937764A (en) | Low-voltage distribution network high-frequency measurement data processing and topology identification method | |
| CN115640748A (en) | Dynamic frequency response prediction method for generators after disturbance of power system | |
| CN115034493A (en) | Wind power plant black start path optimization method considering unit operation state | |
| CN116187082A (en) | Single-machine equivalent modeling method for wind power plant | |
| An et al. | IGBT open circuit fault diagnosis method for a modular multilevel converter based on PNN-MD | |
| CN112670966B (en) | Self-adaptive current protection method for photovoltaic power distribution network | |
| CN115236457A (en) | Method, system, equipment and storage medium for positioning short-circuit fault section of oil field distribution network | |
| CN113241793A (en) | Prevention control method for power system with IPFC (intelligent power flow controller) considering wind power scene | |
| CN112564158B (en) | Direct current commutation failure prediction method | |
| Liu et al. | A data-driven harmonic modeling method for electric vehicle charging stations | |
| CN112769124A (en) | Power system rapid operation risk assessment method based on power flow transfer and tracking | |
| CN110957723A (en) | Data-driven method for rapidly evaluating transient voltage safety of power grid on line | |
| Netto et al. | On the use of smart meter data to estimate the voltage magnitude on the primary side of distribution service transformers | |
| Wen et al. | Equivalent Modeling Based on Long Short-term Memory Neural Network for Virtual Synchronous Generator | |
| de Toledo Silva et al. | Faster-than-Real-Time Simulation of a Large Brazilian AC/DC Grid to Analyze Electromagnetic & Electromechanical Transients as Well as Commutation Failures | |
| Doosthosseini et al. | Hardware-in-the-Loop Testbed Development for Validating Novel Photovoltaic Battery Energy Storage System Concepts | |
| Zhang et al. | Accuracy Evaluation Method for Electromechanical Transient Model of Dynamic Reactive Power Compensation Device Applied to Renewable Power Station |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |