A method for predicting the load level of branch line distribution transformers based on user-side smart meter data

Abstract

The present invention relates to a method for predicting the load level of a branch line distribution transformer based on user-side smart meter data. The method calculates branch reactive power and line loss through the power flow relationship of the distribution network, and designs an algorithm for estimating reactive load based on the smart meter data. The distribution characteristics of the three-phase source and load of the distribution network are judged according to the data collected by the meter, and then the branch power flow is iteratively calculated at a fixed point in combination with the new energy output data and the network topology. When the voltage difference obtained by the iteration is within a preset range, accurate three-phase reactive load data is obtained. For a distribution network with asymmetrical distribution of three-phase source and load, the method further considers the line charging effect and combines the reactive load prediction data and active load collection data of the distribution network to evaluate the load level of the branch line distribution transformer. The method solves the problem that the load level of the branch line distribution transformer is difficult to accurately evaluate due to the inability to directly measure the reactive load on the user side of the distribution network, thereby significantly improving the automation level of the distribution network and the safe and stable operation of the distribution network.

Description

Technical Field

[0001] The present invention belongs to the technical field of electricity consumption analysis, and in particular relates to a method for predicting the load level of a branch line distribution transformer based on user-side smart meter data. Background Art

[0002] With the gradual increase in the scale and proportion of distributed new energy and electric vehicle access, the safe and stable operation of distribution networks has been greatly challenged. The load level of distribution transformers is closely related to the operating status of the distribution network. Based on the load level of distribution transformers, it is possible to determine whether they are overloaded. This is an important basis for adjusting distribution network operation strategies, planning distribution network expansion, and assessing distribution network operation risks. Accurate assessment of the load level of distribution transformers is particularly important for distribution networks with uneven source load distribution and large source load fluctuations. The load level of distribution transformers depends on the active and reactive loads of the users connected to them. The widespread use of smart meters on the user side provides sufficient data support for the evaluation of distribution network transformer loads. Smart meters can obtain real-time voltage, current, and active load data on the user side, but there is still a lack of direct sources of reactive load data on the user side.

[0003] At present, reactive load is mainly obtained indirectly based on flow analysis or data-driven methods. The flow analysis-based method is to analyze the distribution network flow through voltage and active power monitoring data, and set the user-side power factor constant to estimate the reactive load. However, this method depends on whether the power factor setting is reasonable. The uncertainty of the user-side source load power such as household photovoltaics and electric vehicles makes it difficult for the actual operating power factor to be at a stable value. Therefore, the reactive load estimated based on flow analysis often has limited accuracy; the data-driven method is based on the historical data of distribution network operation, such as network loss, voltage and different types of active load data, and uses artificial intelligence methods such as neural networks and deep learning to train an accurate flow regression model, and use this model to predict the reactive load distribution of the distribution network online. However, this method requires massive data to establish an accurate regression model, and the model needs to be adjusted as the source load level of the distribution network develops, and its applicable scenarios are limited.

[0004] Therefore, designing a branch line distribution transformer load level prediction method based on user-side smart meter data that can adapt to different distribution network conditions and improve prediction accuracy has become a technical problem that needs to be solved urgently. Summary of the Invention

[0005] In order to solve the above problems, the present invention provides an algorithm for distinguishing abnormal electricity consumption in enterprises based on electricity data analysis.

[0006] The technical solution of the present invention is a method for predicting the load level of a branch line distribution transformer based on user-side smart meter data. The prediction method includes the following steps:

[0007] 1) Taking the distribution network branch as the object, collect the renewable energy output data and the active load, node voltage, branch impedance and current data of the user-side smart meter, and initialize the number of iterations k = 1;

[0008] 2) Determine the asymmetry of the three-phase source-load distribution of the distribution network based on node voltage data. For distribution networks with asymmetric distribution, consider the line charging effect. Obtain the branch flow relationship of current and voltage from the collected data and integrate it into a flow relationship matrix. This flow relationship matrix includes the voltage drop vector, current vector, impedance matrix, ground susceptance, and current flow direction matrix.

[0009] 3) Calculate the active power and reactive power of the branch using the current data and voltage drop vector;

[0010] 4) Representing active power and reactive power as active power loss, user-side active load, renewable energy active output, and reactive power loss, user-side reactive load, renewable energy reactive output, establish a power relationship expression, and introduce the impedance matrix and current flow matrix into the power relationship expression to obtain a power matrix expression;

[0011] 5) Based on the linear relationship between active power and reactive power and distribution line parameters, the power matrix expression is derived to obtain the reactive loss expression. The active power, user-side active load, and renewable energy active output are then substituted into the expression to calculate the reactive loss. The reactive loss is then substituted into the power matrix expression to calculate the reactive load.

[0012] 6) Use active load and reactive load to represent the current vector in the power flow relationship matrix, and calculate the voltage drop vector value. Compare it with the voltage drop vector of the node voltage data to determine whether the calculated voltage drop vector value is correct. Then determine whether the calculated voltage drop vector value converges. If the judgment result is incorrect or does not converge, iterate the node voltage, and the number of iterations k increases by 1;

[0013] 7) If the judgment result of step 6) is both correct and converged, the apparent power of the transformer is calculated by the active load and reactive load, and the load level of the branch line distribution transformer corresponding to the apparent power is used to obtain the prediction result.

[0014] As a further improvement of the present invention, the determination of the asymmetry of the three-phase source-load distribution of the distribution network includes calculating an evaluation index of the three-phase imbalance of the distribution network, and the calculation formula is formula (1): Among them U m is the value of the m-phase node voltage collected by the meter, U m,avg is the average value of the m-phase node voltage within a judgment cycle.

[0015] As a further improvement of the present invention, the branch power flow relationship between the current and voltage is expressed as formula (2) Formula (3) in, is the current of the m-phase (a-1, a) branch, is the current flowing from node a of phase m to the user, is the voltage of the N node of phase m, is the impedance of the m-phase (a-1, a) branch; under a distribution network branch line with N nodes, node 1 is the branch line distribution transformer node, a=(2,3,…,N).

[0016] As a further improvement of the present invention, the expression of the tidal relationship matrix is as follows: in, is the voltage drop vector of phase m of the distribution network, any element represents the voltage drop between node 1 and node a, is the current vector of phase m of the distribution network, any element represents the current flowing from node a to the user, is the impedance matrix of phase m of the distribution network, any element represents the impedance between adjacent nodes a-1 and a, is the user-side current flow matrix of phase m of the distribution network.

[0017] As a further improvement of the present invention, the calculation expressions of the active power and reactive power are as follows: in as well as They correspond to the complex power, active power, and reactive power between adjacent nodes a-1 and a, respectively.

[0018] As a further improvement of the present invention, the power relationship expression is formula (6)

[0019] ,

[0020] in, and They correspond to the active loss and reactive loss of the m-phase (a-1, a) branch, and They correspond to the new energy active output and new energy reactive output of node a in phase m respectively. and They correspond to the active load and reactive load of node a of phase m respectively; the power matrix expression is formula (7) in, as well as They correspond to active power vector, active loss vector, active load vector and new energy active output vector respectively. as well as They correspond to reactive power vector, reactive loss vector, reactive load vector and new energy reactive output vector respectively.

[0021] As a further improvement of the present invention, the reactive power loss expression is as follows: in as well as For the matrix The matrix formed by the imaginary and real parts of the elements in is The elements correspond to the line reactance and resistance of the distribution network; Represents the division of the corresponding elements in the matrix; the calculation formula of the reactive load is formula (9) The active load and reactive load are brought into the power flow relationship matrix to obtain the voltage drop vector calculation formula (10): in, is the node voltage vector of phase m of the distribution network, any element Represents the voltage at node a.

[0022] As a further improvement of the present invention, the above-mentioned formula corresponding to the distribution network with asymmetric three-phase source-load distribution is obtained by correcting the branch current flow relationship and the voltage drop vector calculation formula by taking into account the line charging effect, which are respectively formula (11)

[0023]

[0024] And formula (12) in It represents the earth susceptance of node a on phase m of the distribution network.

[0025] As a further improvement of the present invention, the iterative processing of the node voltage includes the following processing formula, formula (13) And formula (14)

[0026] Asymmetric; the judgment of whether the calculated value of the pressure drop vector converges includes the convergence discriminant formula (15)

[0027] Among them, ||||2 is the 2-norm representation and ε is the convergence parameter.

[0028] As a further improvement of the present invention, the calculation formula of the apparent power is formula (16)

[0029] Among them, ||||1 is the 1-norm representation,

[0030] S TR is the apparent power.

[0031] After adopting the above method, the reactive load of the distribution network is accurately calculated through the active load data and node voltage data collected by the smart meter on the user side, which is used to predict the load level of the branch line distribution transformer. The charging effect of the distribution network line under the asymmetric source-load distribution is considered through the branch flow model of the distribution network branch line, and the distribution network branch status is comprehensively analyzed. The application scenarios are comprehensive, providing technical support for the life prediction of distribution transformers, the adjustment of distribution network operation strategies and the perception of distribution network operation status, which has a significant improvement on the safe and stable operation of the distribution network.

[0032] By using branch flow fixed-point iteration to calculate reactive load, there is no need for a fixed power factor assumption on the user side. It can be compared with the voltage data collected by the smart meter, and then the node voltage can be updated according to the three-phase source-load distribution status of the distribution network, eliminating the numerical error generated in the reactive load calculation process. The calculation process is simple and has application advantages in multi-node distribution network branch lines containing a high proportion of distributed renewable energy. It is conducive to improving the accuracy of reactive load calculation of unbalanced distribution networks and load level prediction of distribution transformers, and plays an important guiding role in reactive load management and voltage adjustment of distribution network branch lines. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 Shown is a schematic diagram of user-side data collection of the present invention.

[0034] Figure 2 Shown is a schematic diagram of the transformer branch line of the present invention.

[0035] Figure 3 Shown is a schematic diagram of the load level prediction process of the present invention. DETAILED DESCRIPTION

[0036] like Figure 1-3 A method for predicting the load level of a branch line distribution transformer based on user-side smart meter data is shown. The prediction method includes the following steps:

[0037] 1) Taking the distribution network branch as the object, collect the renewable energy output data and the active load, node voltage, branch impedance and current data of the user-side smart meter, and initialize the number of iterations k = 1;

[0038] 2) Determine the asymmetry of the three-phase source-load distribution of the distribution network based on node voltage data. For distribution networks with asymmetric distribution, consider the line charging effect. Obtain the branch flow relationship of current and voltage from the collected data and integrate it into a flow relationship matrix. This flow relationship matrix includes the voltage drop vector, current vector, impedance matrix, ground susceptance, and current flow direction matrix.

[0039] 3) Calculate the active power and reactive power of the branch using the current data and voltage drop vector;

[0040] 4) Representing active power and reactive power as active power loss, user-side active load, renewable energy active output, and reactive power loss, user-side reactive load, renewable energy reactive output, establish a power relationship expression, and introduce the impedance matrix and current flow matrix into the power relationship expression to obtain a power matrix expression;

[0041] 5) Based on the linear relationship between active power and reactive power and distribution line parameters, the power matrix expression is derived to obtain the reactive loss expression. The active power, user-side active load, and renewable energy active output are then substituted into the expression to calculate the reactive loss. The reactive loss is then substituted into the power matrix expression to calculate the reactive load.

[0042] 6) Use active load and reactive load to represent the current vector in the power flow relationship matrix, and calculate the voltage drop vector value. Compare it with the voltage drop vector of the node voltage data to determine whether the calculated voltage drop vector value is correct. Then determine whether the calculated voltage drop vector value converges. If the judgment result is incorrect or does not converge, iterate the node voltage, and the number of iterations k increases by 1;

[0043] 7) If the judgment result of step 6) is both correct and converged, the apparent power of the transformer is calculated by the active load and reactive load, and the load level of the branch line distribution transformer corresponding to the apparent power is used to obtain the prediction result.

[0044] The determination of the asymmetry of the three-phase source-load distribution of the distribution network includes calculating the evaluation index of the three-phase imbalance of the distribution network, and the calculation formula is formula (1): Among them U m is the value of the m-phase node voltage collected by the meter, U m,avg is the average value of the m-phase node voltage within a judgment cycle.

[0045] The branch power flow relationship between the current and voltage is expressed as formula (2) Formula (3) in, is the current of the m-phase (a-1, a) branch, is the current flowing from node a of phase m to the user, is the voltage of the N node of phase m, is the impedance of the m-phase (a-1, a) branch; under a distribution network branch line with N nodes, node 1 is the branch line distribution transformer node, a=(2,3,…,N).

[0046] The expression of the power flow relationship matrix is formula (4) in, is the voltage drop vector of phase m of the distribution network, any element represents the voltage drop between node 1 and node a, is the current vector of phase m of the distribution network, any element represents the current flowing from node a to the user, is the impedance matrix of phase m of the distribution network, any element represents the impedance between adjacent nodes a-1 and a, is the user-side current flow matrix of phase m of the distribution network.

[0047] The calculation expressions of active power and reactive power are as follows: in as well as They correspond to the complex power, active power, and reactive power between adjacent nodes a-1 and a, respectively.

[0048] The power relationship expression is formula (6)

[0049]

[0050] ,in, and They correspond to the active loss and reactive loss of the m-phase (a-1, a) branch, and They correspond to the new energy active output and new energy reactive output of node a in phase m respectively. and They correspond to the active load and reactive load of node a of phase m respectively; the power matrix expression is formula (7) in, as well as They correspond to active power vector, active loss vector, active load vector and new energy active output vector respectively. as well as They correspond to the reactive power vector, reactive loss vector, reactive load vector and new energy reactive output vector respectively.

[0051] The reactive power loss expression is: in as well as For the matrix The matrix formed by the imaginary and real parts of the elements in is The elements correspond to the line reactance and resistance of the distribution network; Represents the division of the corresponding elements in the matrix; the calculation formula of the reactive load is formula (9) The active load and reactive load are brought into the power flow relationship matrix to obtain the voltage drop vector calculation formula (10): in, is the node voltage vector of phase m of the distribution network, any element Represents the voltage at node a.

[0052] The above-mentioned formula corresponding to the distribution network with asymmetric three-phase source-load distribution is obtained by correcting the branch current flow relationship and the voltage drop vector calculation formula by considering the line charging effect, which are respectively:

[0053]

[0054] And formula (12) in It represents the earth susceptance of node a on phase m of the distribution network.

[0055] The iterative processing of the node voltage includes the following processing formula, formula (13)

[0056] And formula (14)

[0057] Asymmetric; the judgment of whether the calculated value of the pressure drop vector converges includes the convergence discriminant formula (15)

[0058] Among them, ||||2 is the 2-norm representation and ε is the convergence parameter.

[0059] As a further improvement of the present invention, the calculation formula of the apparent power is formula (16)

[0060] Among them, ||||1 is the 1-norm representation,

[0061] S TR is the apparent power.

[0062] The active load data and node voltage data collected by the user-side smart meters are used to accurately calculate the reactive load of the distribution network, which is used to predict the load level of the branch line distribution transformer. The charging effect of the distribution network line under the asymmetric source-load distribution is considered through the branch line flow model of the distribution network, and the distribution network branch status is comprehensively analyzed. The application scenarios are comprehensive, providing technical support for the life prediction of distribution transformers, the adjustment of distribution network operation strategies and the perception of distribution network operation status, which significantly improves the safe and stable operation of the distribution network.

[0063] By using branch flow fixed-point iteration to calculate reactive load, there is no need for a fixed power factor assumption on the user side. It can be compared with the voltage data collected by the smart meter, and then the node voltage can be updated according to the three-phase source-load distribution status of the distribution network, eliminating the numerical error generated in the reactive load calculation process. The calculation process is simple and has application advantages in multi-node distribution network branch lines containing a high proportion of distributed renewable energy. It is conducive to improving the accuracy of reactive load calculation of unbalanced distribution networks and load level prediction of distribution transformers, and plays an important guiding role in reactive load management and voltage adjustment of distribution network branch lines.

Claims

1. A method for predicting the load level of a branch line distribution transformer based on user-side smart meter data, characterized by: The prediction method The following steps are included: 1) Taking the distribution network branch as the object, collect the renewable energy output data and the active load, node voltage, branch impedance and current data of the user-side smart meter, and initialize the number of iterations k = 1; 2) Determine the asymmetry of the three-phase source-load distribution of the distribution network based on node voltage data. For distribution networks with asymmetric distribution, consider the line charging effect. Obtain the branch flow relationship of current and voltage from the collected data and integrate it into a flow relationship matrix. This flow relationship matrix includes the voltage drop vector, current vector, impedance matrix, ground susceptance, and current flow direction matrix. 3) Calculate the active power and reactive power of the branch using the current data and voltage drop vector; 4) Representing active power and reactive power as active power loss, user-side active load, renewable energy active output, and reactive power loss, user-side reactive load, renewable energy reactive output, establish a power relationship expression, and introduce the impedance matrix and current flow matrix into the power relationship expression to obtain a power matrix expression; 5) Based on the linear relationship between active power and reactive power and distribution line parameters, the power matrix expression is derived to obtain the reactive loss expression. The active power, user-side active load, and renewable energy active output are then substituted into the expression to calculate the reactive loss. The reactive loss is then substituted into the power matrix expression to calculate the reactive load. 6) Use active load and reactive load to represent the current vector in the power flow relationship matrix, and calculate the voltage drop vector value. Compare it with the voltage drop vector of the node voltage data to determine whether the calculated voltage drop vector value is correct. Then determine whether the calculated voltage drop vector value converges. If the judgment result is incorrect or does not converge, iterate the node voltage, and the number of iterations k increases by 1; 7) If the judgment result of step 6) is both correct and converged, the apparent power of the transformer is calculated by the active load and reactive load, and the load level of the branch line distribution transformer corresponding to the apparent power is used to obtain the prediction result.

2. The method for predicting the load level of a branch line distribution transformer based on user-side smart meter data according to claim 1, characterized in that: The determination of the asymmetry of the three-phase source-load distribution of the distribution network includes calculating the evaluation index of the three-phase imbalance of the distribution network, and the calculation formula is formula (1): Among them U m is the value of the m-phase node voltage collected by the meter, U m,avg is the average value of the m-phase node voltage within a judgment cycle.

3. The method for predicting the load level of a branch line distribution transformer based on user-side smart meter data according to claim 1, characterized in that: The branch power flow relationship between the current and voltage is expressed as formula (2) in, is the current of the m-phase (a-1, a) branch, is the current flowing from node a of phase m to the user, is the voltage of the N node of phase m, is the impedance of the m-phase (a-1, a) branch; under a distribution network branch line with N nodes, node 1 is the branch line distribution transformer node, a=(2,3,…,N).

4. The method for predicting the load level of a branch line distribution transformer based on user-side smart meter data according to claim 3, characterized in that: The expression of the power flow relationship matrix is formula (4) in, is the voltage drop vector of phase m of the distribution network, any element represents the voltage drop between node 1 and node a, is the current vector of phase m of the distribution network, any element represents the current flowing from node a to the user, is the impedance matrix of phase m of the distribution network, any element represents the impedance between adjacent nodes a-1 and a, is the user-side current flow matrix of phase m of the distribution network.

5. The method for predicting the load level of a branch line distribution transformer based on user-side smart meter data according to claim 4, characterized in that: The calculation expressions of active power and reactive power are as follows: in as well as They correspond to the complex power, active power, and reactive power between adjacent nodes a-1 and a, respectively.

6. The method for predicting the load level of a branch line distribution transformer based on user-side smart meter data according to claim 5, characterized in that: The power relationship expression is formula (6) , in, and They correspond to the active loss and reactive loss of the m-phase (a-1, a) branch, and They correspond to the new energy active output and new energy reactive output of node a in phase m respectively. and They correspond to the active load and reactive load of node a of phase m respectively; the power matrix expression is formula (7) in, as well as They correspond to active power vector, active loss vector, active load vector and new energy active output vector respectively. as well as They correspond to the reactive power vector, reactive loss vector, reactive load vector and new energy reactive output vector respectively.

7. The method for predicting the load level of a branch line distribution transformer based on user-side smart meter data according to claim 6, characterized in that: The reactive power loss expression is: in as well as For the matrix The matrix formed by the imaginary and real parts of the elements in is The elements correspond to the line reactance and resistance of the distribution network; Represents the division of the corresponding elements in the matrix; the calculation formula of the reactive load is formula (9) The active load and reactive load are brought into the power flow relationship matrix to obtain the voltage drop vector calculation formula (10): in, is the node voltage vector of phase m of the distribution network, any element Represents the voltage at node a.

8. The method for predicting the load level of a branch line distribution transformer based on user-side smart meter data according to claim 7, characterized in that: The above-mentioned formula corresponding to the distribution network with asymmetric three-phase source-load distribution is obtained by correcting the branch current flow relationship and the voltage drop vector calculation formula by considering the line charging effect, which are respectively: And formula (12) in It represents the earth susceptance of node a on phase m of the distribution network.

9. The method for predicting the load level of a branch line distribution transformer based on user-side smart meter data according to claim 8, characterized in that: The iterative processing of the node voltage includes the following processing formula, formula (13) And formula (14) Asymmetric; the judgment of whether the calculated value of the pressure drop vector converges includes the convergence discriminant formula (15) Among them, ||||2 is the 2-norm representation and ε is the convergence parameter.

10. The method for predicting the load level of a branch line distribution transformer based on user-side smart meter data according to claim 7, characterized in that: The calculation formula of the apparent power is formula (16) Among them, ||||1 is the 1-norm representation, S TR is the apparent power.

Images