Low-voltage transformer area wiring error electric leakage fault user detection method and detection device

Abstract

The invention provides a low-voltage transformer area wiring error electric leakage fault user detection method and detection device. The detection method comprises the following steps: (A) in an acquisition time period, acquiring residual current of a detected transformer area and load current of each user in the detected transformer area at a set sampling interval; calculating a complex weight matrix alpha = (AHA)-1AHb by using the following formula; and (B) if the real part of the calculated complex weight coefficient of the jth user is greater than or equal to a preset threshold value, judging that the jth user is a wiring error electric leakage user, otherwise, judging that the jth user is a normal user.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and particularly to the identification and location of users with wiring error and leakage faults in low-voltage distribution substations. Background Art

[0002] The low-voltage distribution network penetrates into the daily life and production areas of users. Its electrical equipment and power supply lines are spread all over the streets and alleys. While ensuring the power consumption of users, due to factors such as improper installation of equipment, harsh operating environment, and weak electrical safety awareness of personnel in some places and equipment, there are serious safety hazards. Poor contact of line contacts and terminal blocks, aging and insulation damage of lines increase the risk of electric shock between people and electrical fires day by day. In addition to insulation leakage faults, due to the differences in professional levels and work attitudes of electrical operators, subjective factors such as carelessness, in the distribution cabinet at the inlet position of the low-voltage power supply line, wiring error and leakage faults often occur where the neutral line and the ground line are reversely connected. Accurate and reliable evaluation of user leakage liability is the key to realizing the location of wiring error and leakage faults and protecting the leakage safety of the substation.

[0003] When this fault occurs, the ground line will become part of the user's power consumption loop. When the user uses electricity, the load current returns from the ground line to the neutral point of the transformer, which is converted into residual current, resulting in frequent misoperation of the residual current protection device in the substation. Moreover, the originally uncharged ground line passes through a large-amplitude load current, which is extremely easy to burn out the ground line terminal and damage the grounding protection function. According to the correlation between the abnormal residual current in the substation and the load current of the faulty user when the wiring error fault occurs, currently, a multiple linear regression equation can be constructed using the load current and the residual current amplitude to calculate the weight coefficient values of each user and judge the existence of abnormal users. However, only focusing on the current amplitude for multiple linear regression is a wrong description of the linear relationship between the residual current in the substation and the load current of users. In fact, it cannot effectively measure the contributions of the resistive and inductive components of the load current of each user to the residual current in the substation. There are relatively large regression errors in the real-number multiple regression in the research on the perception and location of wiring error and leakage faults, and it is impossible to accurately identify users with wiring error and leakage faults. When there are multiple abnormal users with wiring errors in the substation, the multiple linear regression method based on current amplitude also has significant false detections and missed detections. Summary of the Invention

[0004] The problem to be solved by the present invention is to provide a method and device for detecting users with wiring error and leakage faults in a low-voltage substation, aiming at the problem that when there are multiple abnormal users with wiring errors in the substation, the existing multiple linear regression method based on current amplitude also has significant false detections and missed detections.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is: A method for detecting users with wiring error and leakage faults in a low-voltage substation, characterized by comprising:

[0006] (A) During the acquisition time period, the residual current of the measured substation area and the load currents of each user in the measured substation area are acquired at a set sampling interval; the complex weight matrix α is calculated using the following formula:

[0007] α = (A H A) -1 A H b;

[0008] Among them, Among them, is the complex weight coefficient of the load current of the j-th user in the measured substation area with respect to the residual current of the measured substation area; is the i-th sampling value of the residual current of the measured substation area, is the i-th sampling value of the load current of the j-th user; i ∈ [1, m], m represents the number of sampling values of the residual current of the measured substation area; j ∈ [1, p], p represents the number of users in the measured substation area; m > p; (A H A) -1 represents the inverse matrix of A H A H represents the conjugate transpose matrix of A

[0009] (B) If the real part of the complex weight coefficient of the j-th user calculated is greater than or equal to the preset threshold, then the j-th user is determined to be a user with wiring error and leakage, otherwise, the j-th user is determined to be a normal user.

[0010] The applicant's research found that the synthesis of substation area current is not a simple algebraic sum of amplitudes, but a phasor superposition of its amplitude and phase. When there are multiple users with wiring error abnormalities in the substation area, the multivariate linear regression method based on current amplitude also has significant false detections and missed detections. In this application, using the complex multivariate linear regression method, through a set of phasor sampling values of the residual current of the substation area and the load currents of each user, a set of complex weight coefficients that minimize the sum of the squared amplitudes of the residual current regression error is determined in the complex domain, as an evaluation index for the responsibility of each user for the abnormal residual current in the substation area, so that normal users and abnormal users can still be accurately distinguished when there are multiple users with wiring error abnormalities in the substation area.

[0011] In the above technical solution, the preset threshold is 0.9.

[0012] In the above technical solution, the value range of the set sampling interval is [5 min, 30 min].

[0013] In the above technical solution, m ≥ 200.

[0014] Based on the same inventive concept, the present invention also provides a detection device for users with wiring error and leakage faults in a low-voltage power distribution area, including a computer device; the computer device is configured or programmed to execute the steps of the above-mentioned detection method for users with wiring error and leakage faults in a low-voltage power distribution area.

[0015] The advantages and positive effects of the present invention are as follows: The present invention uses the multiple linear regression method in the complex domain to calculate the complex weight coefficients of the load currents of each user with respect to the abnormal residual current in the power distribution area, and can accurately distinguish normal users and abnormal users under various fault scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 It is a schematic diagram of the steps of the detection method for users with wiring error and leakage faults in a low-voltage power distribution area according to an embodiment of the present invention;

[0018] Figure 2-1 It is a schematic diagram of the superposition of the amplitudes of the residual currents in the power distribution area according to an embodiment of the present invention;

[0019] Figure 2-2 It is a schematic diagram of the phasor superposition of the residual currents in the power distribution area according to an embodiment of the present invention;

[0020] Figure 3 It is a curve of the change of the resistive components of the currents in the power distribution area when a single user has a wiring error according to an embodiment of the present invention.

[0021] Figure 4 It is a curve of the change of the inductive components of the currents in the power distribution area when a single user has a wiring error according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0023] In the present invention, considering the phasor characteristics of multi-source electrical quantities in the distribution area, from the perspective of the inductive component and resistive component of the current, the contribution of the load current of each user at a certain moment to the residual current in the distribution area is quantitatively evaluated. Based on the sampled values of the current phasor, a complex multiple linear regression model is constructed, and the complex weight coefficients of the load current of each user are calculated according to the least squares theory. In complex fault scenarios, accurate guiding information can also be provided for the detection and troubleshooting of leakage faults. To achieve the above object, the present invention proposes a method for identifying users with wiring error leakage faults based on complex multiple linear regression.

[0024] The present invention discloses a method for identifying users with neutral line and ground line wiring error leakage faults based on complex multiple linear regression. Based on the phasor model of the residual current in the distribution area and the linear relationship between the load current of users during wiring error leakage faults, considering the phasor characteristics of the electrical quantities in the distribution area, a complex multiple linear regression equation is constructed. In the complex domain, the optimal complex weight coefficients of the multiple linear regression equation are solved according to the least squares theory, and these are used as evaluation indicators to measure the contribution degree of each user under the distribution area to the abnormal residual current in the distribution area. The present invention more detailedly and appropriately describes the influence of users' wiring error leakage faults on the residual current in the distribution area from the perspective of phasors, effectively improves the identification accuracy of users with wiring error faults, provides prior guidance for leakage fault troubleshooting, improves the operation rate of leakage protection in the distribution area, and ensures the electricity use safety of users.

[0025] As Figure 1 shown, the present invention provides a method for detecting users with wiring error leakage faults in a low-voltage distribution area, including:

[0026] (A) During the acquisition time period, the residual current of the measured distribution area and the load current of each user in the measured distribution area are acquired at a set sampling interval; the complex weight matrix α is calculated using the following formula:

[0027] α = (A H A) -1 A H b;

[0028] Among them, α is a matrix with p rows and 1 column, and b is a matrix with m rows and 1 column;

[0029] Among them, is the complex weight coefficient of the load current of the j-th user in the measured distribution area with respect to the residual current of the measured distribution area; is the i-th sampled value of the residual current of the measured distribution area, is the i-th sampled value of the load current of the j-th user; i ∈ [1, m], m represents the number of sampled values of the residual current of the measured distribution area; j ∈ [1, p], p represents the number of users in the measured distribution area; m > p; (A H A) -1 represents the inverse matrix of A H AH Denotes the conjugate transpose matrix of A. Sampling can be carried out at a fixed period, and the i-th sampling value of the residual current and the i-th sampling value of the user's load current are collected at the same moment (i.e., the sampling point of the i-th sampling period).

[0030] (B) If the real part of the complex weight coefficient of the j-th user calculated is greater than or equal to the preset threshold, then it is determined that the j-th user is a user with wiring error and leakage, otherwise, it is determined that the j-th user is a normal user.

[0031] The preset threshold is 0.9. The value range of the set sampling interval is [5 min, 30 min]. m≥200.

[0032] The present invention also provides a device for detecting users with wiring error and leakage faults in a low-voltage power distribution area, including a computer device; the computer device is configured or programmed to execute the steps of the above-mentioned method for detecting users with wiring error and leakage faults in a low-voltage power distribution area.

[0033] The method for detecting users with wiring error and leakage by complex multiple linear regression of the present invention first collects the residual current on the distribution transformer side of the power distribution area and the load current data of each subordinate user within a certain period of time. According to the superposition theorem and the phasor synthesis characteristics, a complex multiple linear regression equation set is constructed. According to the least square theory, the optimal complex weight coefficients of the multiple linear regression equation set are solved. According to the magnitudes of the complex weight coefficients of each user obtained, the users with wiring error and leakage faults are located. The establishment and solution steps of the complex multiple linear regression equation set are as follows:

[0034] A1. Within an analysis period of length m, collect the residual current on the distribution transformer side of a certain power supply area with p (where p < m) users and the phasor data of the load current of each user, including the phase magnitude and the amplitude magnitude.

[0035] B1. At the t-th measurement point, the residual current of the power distribution area and the load current of each user are respectively denoted as and The leakage current of the power supply line of the power distribution area at this moment is denoted as Construct the following complex multiple linear regression equation set:

[0036]

[0037] C1. Represent the complex multiple linear regression equation set in matrix form:

[0038] b = Aα + ε

[0039] Where b is an m×1 complex vector, including the sampled values of the residual current in the transformer area at each moment; A is an m×p complex matrix, including the sampled values of the load current of each user at each moment; α is a p×1 complex vector, including the complex weight coefficients of the load current of different users with respect to the residual current; ε is an m×1 complex vector, including the leakage current values of the power supply lines in the transformer area at each moment.

[0040] D1. Calculate the complex weight coefficient that minimizes the sum of the squared norms Q of the residual current in the transformer area. The objective function is:

[0041]

[0042] E1. Considering that the objective function is a real-valued complex variable function, which does not satisfy the Cauchy-Riemann conditions and is difficult to directly differentiate to find the least squares solution according to the extreme value theory. Then, use Wirtinger Calculation to calculate the derivative of the objective function Q with respect to α 1 α 2 ... α n respectively and make its derivative value equal to 0, solve the stationary point of the objective function, and finally the least squares solution in the complex domain can be obtained as α = (A H A) -1 A H b.

[0043] F1. Set the threshold for distinguishing users with wiring error leakage faults to 0.99. If the real part of the complex weight coefficient phasor of a certain user is greater than this threshold, then this user is regarded as an abnormal user.

[0044] Considering the requirements of detection timeliness and the need to solve the regression equation, the sampling duration in step A1 is usually not shorter than 6 hours and not longer than 24 hours.

[0045] In step B1, represents the magnitude of the leakage current of the j-th user in the transformer area at time t, is used to measure the user leakage current and the user load current The relationship between them represents the complex weight coefficient of the user load current with respect to the residual current in the transformer area in the complex multiple linear regression equation.

[0046] During the sampling period, the complex weight coefficients of the load currents of each user with respect to the residual current in the transformer area are regarded as constants.

[0047] In step E1, the Wirtinger R-derivative of the objective function Q with respect to α j is:

[0048]

[0049] Before solving the complex weight coefficients of each user's load current with respect to the abnormal residual current in the transformer substation area, first construct an auxiliary multiple regression model of each user's load current with respect to other load currents, calculate the variance inflation factor of each user, and verify the linear independence of each user's load current during this period.

[0050] When a wiring error leakage fault occurs, the user's load current is converted into the residual current in the transformer substation area, and there is a significant correlation between it and the abnormal residual current in the transformer substation area. Some studies have shown that the weight coefficients of each load current with respect to the residual current in the transformer substation area can be solved by using multiple linear regression of the residual current in the transformer substation area and the amplitudes of each user's load current to determine whether there is a wiring error leakage fault for the user. The multiple linear regression model is shown as follows:

[0051] I r =β 1 I L.1 +β 2 I L.2 +...+β p I L.p +I r0 (1)

[0052] In the formula, I r is the amplitude of the residual current in the transformer substation area, I L.j , j ∈ [1, P] represents the load current magnitudes of P users in the transformer substation area, β j represents the conversion ratio of the j-th user with respect to the residual current in the transformer substation area, and β j is a real number. However, the current synthesis is not a simple algebraic sum of its amplitudes, but a phasor superposition of its amplitude and phase. As shown in Figure 2-1 and Figure 2-2 , due to the differences in the user phase and current phase, the phasor synthesis result will be enhanced or weakened. If, as shown in Figure 2-1 , the amplitude superposition method of the existing technology is adopted, that is, only relying on the current amplitude to construct a multiple linear regression model in the real number domain to calculate the influence of each user's load current in the transformer substation area on the abnormal residual current in the transformer substation area, it is a wrong description of the linear relationship between the residual current in the transformer substation area and the user's load current. In fact, it cannot effectively measure the contributions of the resistive component and inductive component of each user's load current to the residual current in the transformer substation area. There is a large regression error fundamentally in the real number multiple regression in the research on the perception and location of wiring error leakage faults. Figure 2-2 This is the method proposed by the present invention (that is, the method of phasor superposition), that is, according to the phasor superposition in the complex number domain is the current synthesis result that conforms to the actual situation, that is, the principle explanation of the method proposed by the present invention.

[0053] Based on the above problems and combined with the phasor characteristics of the substation area current, the present invention proposes a method for detecting users with wiring error and leakage faults based on complex multiple linear regression. In the complex domain, the correlation between the phasors of the load currents of each user and the residual current phasor of the substation area is calculated.

[0054] Under normal operating conditions, there is only a small amount of equipment and line leakage current in the user's indoor area. The magnitude and ratio of its resistive component and capacitive component will change with the switching of equipment and the user's electricity consumption behavior. Therefore, it can be considered that the user's leakage current and the user's load current have a certain relationship. At the same moment, the leakage current part contained in the user's load current can be calculated by the following formula:

[0055]

[0056] where That is, the conversion ratio of the resistive component and capacitive component in the user's load current with respect to the leakage current is measured by the complex weight coefficient . At this time, the natural residual current of the substation area is composed of the leakage currents of each user and the leakage current of the substation area power supply line, and all the currents in the substation area are in phasor form. Therefore, according to the phasor superposition theorem of current, there is a certain linear relationship among the three, and a complex multiple linear regression model can be constructed as shown below:

[0057]

[0058] Due to the intermittency of the user's electricity consumption behavior, it can be assumed that within a period of time, the composition of the electrical appliances participating in electricity consumption in each user's indoor area and the user's electricity consumption behavior approximately do not change, that is, the conversion ratio of the user's load current with respect to the leakage current remains unchanged, and the corresponding complex weight coefficient is constant. Then when a leakage fault occurs in the neutral line or ground wire of a certain user in the substation area, the abnormal user load current returns to the transformer neutral point through the ground wire. At this time, the conversion ratio of the abnormal user load current with respect to the residual current of the substation area is completely converted, and there is a strong correlation between the abnormal user and the residual current of the substation area. The complex weight coefficient of this user in the formula will change. Since the amplitude of the abnormal user's leakage current far exceeds the natural residual current of the substation area, as Figure 3 , Figure 4 shown, the change curves of the resistive component and inductive component of the abnormal user load current will be highly approximated to the change curves of the resistive component and inductive component of the residual current after phasor synthesis in the substation area. According to Figure 3 , Figure 4It can be seen that after a wiring error occurs, the fluctuations of the resistive and inductive parts of the residual current in the substation area and the load current of abnormal users will be highly similar. This is because the load current of abnormal users will all be converted into residual current, dominating the fluctuations of the residual current in the substation area.

[0059] The sampling interval of the present invention is generally 15 minutes, and the number of sample points should be at least greater than the number of variables (i.e., the number of users). The number of sample points is generally at least 200.

[0060] Based on the method of the present application, within the analyzed acquisition time period (for example, an analysis period of length m, m > p), it is approximately considered that the electricity consumption behaviors and electrical appliance compositions of each user in the substation area do not change. The residual current in the substation area and the load current of each user at the t-th measurement point are respectively denoted as and The leakage current of the power supply line in the substation area at this moment is denoted as Thus, a complex multivariate linear equation system of the user load current with respect to the residual current in the substation area during this period can be listed:

[0061]

[0062] Let:

[0063]

[0064]

[0065]

[0066]

[0067] Then the complex multivariate linear equation system can be expressed in matrix form:

[0068] b = Aα + ε (9)

[0069] Where b is an m×1 complex vector, including the sampling values of the residual current in the substation area at each moment; A is an m×p complex matrix, including the sampling values of the load current of each user at each moment; α is a p×1 complex vector, including the complex weight coefficients of the load current of different users with respect to the residual current; ε is an m×1 complex vector, including the leakage current values of the power supply line in the substation area at each moment.

[0070] In the formula for calculating the real part of the complex weight coefficient and subsequent calculations, the expression of is not used. This formula is an error term, and in subsequent calculations such as residuals and regression, this term has been ignored.

[0071] According to the least squares theory, the optimal complex weight coefficient can be obtained by calculating the minimum value of the sum of the squared moduli of the residuals of the residual current in the substation area. The target residual function Q is:

[0072]

[0073] Furthermore, from the formula for the sum of squared residuals, it can be seen that this objective function represents a real-valued complex variable function that maps from the complex domain to the real domain. It usually does not satisfy the Cauchy-Riemann conditions and is not a holomorphic function. Therefore, its complex derivative cannot be obtained using the derivative rules and chain rules in the real domain. For such functions, in practical applications, more attention is paid to determining the stationary points of the function. Thus, Wiringer Calculation can be introduced to calculate the partial derivative of the residual function Q with respect to the complex weight coefficients, and then the least squares solution of the complex weight coefficients can be solved.

[0074] 3. Wirtinger defined the R-derivative and conjugate R-derivative of a complex variable function:

[0075]

[0076]

[0077] When calculating the R-derivative or conjugate R-derivative, z or z will be regarded as two independent variables, and they can be differentiated separately. When is satisfied, the stationary point of the real-valued complex variable function f(z) can be obtained.

[0078] Furthermore, in the multiple linear regression model of the load current of users in the substation area with respect to the residual current, for the complex weight coefficient vector α = [α 1 , α 2 , …, α p T ∈ C P , the Wirtinger partial derivative of the residual function Q with respect to α is denoted as:

[0079]

[0080] It is called the complex gradient of the function Q with respect to α, denoted as When and only when is satisfied, the stationary point of Q can be calculated. Then, let Q be differentiated with respect to α 1 α 2 ... α n to find the Wirtinger R-derivative respectively, and make the derivative value equal to 0. Its expression is:

[0081]

[0082] where

[0083] ​

[0084] It can be seen that

[0085]

[0086] Finally, the matrix form obtained is as follows:

[0087]

[0088] The solution is:

[0089] α = (A H A) -1 A H b (18)

[0090] Based on the real part size of the optimal complex weight coefficient of the residual current in the transformer substation for the load currents of each user obtained, the wiring error leakage user discrimination threshold is set to 0.99, which can effectively judge the abnormal users in the transformer substation. Among them, represents the conjugate matrix of A. A H represents the conjugate transpose matrix of A. The formula α = (A H A) -1 A H b gives a complex number, that is, the analytical solution of the complex multiple regression equation obtained through the above derivation.

[0091] According to the least squares solution of the complex weight coefficient of the complex multiple linear regression equation obtained by Wirtinger Calculation, it is only the solution of the stationary point, but not necessarily the unique solution. Therefore, it is necessary to further consider the multicollinearity among the load currents of users. Before constructing the multiple linear regression model, the collinearity of the explanatory variables is evaluated through the Variance Inflation Factor (VIF) to ensure that the columns of the complex matrix A are full rank. VIF is used to measure the collinearity degree of a single explanatory variable in the regression model with respect to other explanatory variables, and its calculation formula is as follows:

[0092]

[0093] Among them is the coefficient of determination of the auxiliary regression model constructed by the explanatory variable I L.j with respect to other explanatory variables. The larger the VIF value, the higher the collinearity degree of the corresponding explanatory variable. Usually, 10 is used as the boundary to measure whether there is multicollinearity among the explanatory variables. If VIF ≥ 10, there is collinearity, indicating that the result is not available. It is necessary to expand the sample points to reduce the linear relationship. Measuring the multicollinearity among independent variables by the variance inflation factor is a commonly used method in statistical analysis.

[0094] Table 1 Details of the phase affiliations of each user

[0095] Phase difference User number Phase A 1，4，7，10，13，16 Phase B 2，5，8，11，14，17 Phase C 3，6，9，12，15，18

[0096] Collect the amplitude and phase of the residual current in the substation area and the load current of each user for 18 users of a certain supply tape (i.e., the number of users is 18). The sampling interval is 1 minute, the total number of samples is 1440 points, the time span is 1 day, and a sample set is constructed. The phase types of each user in the substation area are shown in Table 1. Further, a simulation experiment is carried out using the SDP solver in the CVX package on the MATLAB platform. Multivariate linear regression equations are established and the optimal complex weight coefficient vectors are calculated for the scenarios of single-user wiring error leakage fault, two users with wiring error leakage fault in the same phase, two users with wiring error leakage fault in different phases, and three users with wiring error leakage fault in different phases, respectively, so as to identify and locate abnormal users.

[0097] First, construct an auxiliary multivariate linear regression model of the load current of each user among the 18 users with respect to the load current of other users, and find the optimal solution to obtain the coefficient of determination of the auxiliary multivariate linear regression model. Calculate the variance inflation factor of each user as shown in Table 2. There is no significant multicollinearity among the phasor data of the load current of each user.

[0098] Table 2 Calculation results of variance inflation factor VIF for 18 users

[0099] User VIF User VIF 1 1.1621 10 4.9792 2 1.1837 11 5.4241 3 1.3571 12 1.4405 4 1.2832 13 1.3081 5 1.8375 14 1.3705 6 1.0172 15 1.0585 7 2.3820 16 2.4595 8 1.7288 17 2.4846 9 1.9259 18 1.4123

[0100] In this embodiment, in the scenario of single-user wiring error (single-user anomaly), the 2nd user is preset as the abnormal user. In the scenario of two users with wiring error in the same phase (two users with anomaly in the same phase), the 2nd and 5th users are preset as the abnormal users. In the scenario of two users with wiring error in different phases (two users with anomaly in different phases), the 2nd and 4th users are preset as the abnormal users. In the scenario of three users with wiring error in different phases (three users with anomaly in different phases), the 2nd, 6th, and 10th users are preset as the abnormal users. The complex weight coefficients of the load current of each user with respect to the abnormal residual current in the substation area are calculated using the multivariate regression method in the complex domain. The calculation results are shown in Table 3.

[0101] As shown in Table 3, for the preset abnormal users, the real parts of the complex weight coefficients calculated according to the detection method of the present invention are all greater than 0.9. Therefore, according to the detection method proposed by the present invention, by judging whether the real part of the complex weight coefficient of the user is greater than the threshold value of 0.9, the users with wiring error leakage faults can be effectively identified in various fault scenarios.

[0102] Table 3 Complex weight coefficients of the load current of each user under various fault scenarios obtained according to the detection method of the present invention

[0103]

[0104] In the embodiments of the present invention, for the unbalanced three - user leakage fault scenario, users No. 2, No. 6, and No. 10 are preset as abnormal users. The amplitude regression method of the existing technology and the detection method of the present invention (i.e., the complex regression method) are respectively used to calculate the optimal weight coefficient of each user's load current with respect to the residual current in the transformer substation area. Taking the RMSE value of the residual current time - series data in the transformer substation area as the evaluation index, the advantages of the proposed method are analyzed. The RMSE calculation formula is as follows:

[0105]

[0106] In the formula, is the estimated value of the residual current, b is the measured value of the residual current, and m is the sample length.

[0107] The calculation results of each regression method are shown in Table 4. When there are multiple users with wiring - error leakage faults in the transformer substation area, the amplitude regression method cannot effectively distinguish normal users from abnormal users, and the regression accuracy is significantly worse than that of the complex multiple linear regression method. The example shows that the proposed method for identifying users with wiring - error leakage faults by complex multiple linear regression (i.e., the detection method of the present invention) can accurately distinguish the existence of abnormal users, and also has extremely high accuracy and discrimination ability in the case of multiple - user anomalies. As shown in Table 4, using the amplitude regression method, the real part of the complex weight coefficient of user No. 2 and user No. 6 calculated is less than 0.9, and only the real part of the complex weight coefficient of user No. 10 is greater than 0.9 in the calculation result. That is, using the existing amplitude regression method, only the fault of user No. 10 can be identified, but the faults of user No. 2 and user No. 6 cannot be identified. According to the detection method of the present invention, the real parts of the complex weight coefficients of user No. 2, user No. 6, and user No. 10 calculated are all greater than 0.9. Therefore, using the detection method of the present invention, when multiple users have faults, each faulty user can be accurately identified. As shown in Table 4, the calculated RMSE value obtained according to the existing technology is significantly higher than the method proposed by the present invention.

[0108] Table 4 Calculation results of each regression method in the case of unbalanced three - user faults

[0109] User number Amplitude regression method in the prior art Detection method of the present invention 1 0.0417 0.0000+0.0000i 2 <![CDATA 0.4215 > <![CDATA 0.9992-0.0001i > 3 -0.1941 0.0012+0.0003i 4 -0.1505 -0.0006-0.0001i 5 -0.0591 0.0021+0.0013i 6 <![CDATA 0.4179 > <![CDATA 0.9955-0.0013i > 7 -0.0041 -0.0002+0.0005i 8 0.0110 -0.0003-0.0001i 9 -0.0097 -0.0013+0.0002i 10 <![CDATA 0.9789 > <![CDATA 0.9995-0.0001i > 11 0.0250 -0.0003+0.0001i 12 -0.2080 -0.0003+0.0017i 13 -0.0278 -0.0007-0.0007i 14 -0.0274 -0.0014-0.0005i 15 -0.0060 0.0004+0.0005i 16 -0.0102 -0.0002-0.0006i 17 0.0101 -0.0018-0.0007i 18 0.0230 0.0003+0.0000i RMSE 0.3334 0.0045

[0110] It should be noted that the embodiments in this specification are all described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other. The above has described the embodiments of the present invention in detail, but the content described is only the preferred embodiment of the present invention and cannot be considered as used to limit the scope of the embodiments of the present invention. All equivalent changes and improvements made according to the scope of the present invention should still fall within the scope covered by this application. After reading the present invention, various equivalent forms of modification by those skilled in the art to the present invention all fall within the scope defined by the appended claims of this application. Without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

Claims

1. A method for detecting users with wiring error and leakage fault in low-voltage power distribution areas, characterized in that: It includes: (A) During the acquisition time period, the residual current of the measured power distribution area and the load currents of each user in the measured power distribution area are acquired at a set sampling interval; the complex weight matrix α is calculated using the following formula: α = (A H A) -1 A H b; Among them, Among them, is the complex weight coefficient of the load current of the j-th user in the measured power distribution area with respect to the residual current of the measured power distribution area; is the i-th sampling value of the residual current of the measured power distribution area, is the i-th sampling value of the load current of the j-th user; i ∈ [1, m], m represents the number of sampling values of the residual current of the measured power distribution area; j ∈ [1, p], p represents the number of users in the measured power distribution area; m > p; (A H A) -1 represents the inverse matrix of A H A H represents the conjugate transpose matrix of A (B) If the real part of the complex weight coefficient of the j-th user obtained by calculation is greater than or equal to the preset threshold, it is determined that the j-th user is a user with wiring error and leakage. Otherwise, it is determined that the j-th user is a normal user.

2. The method for detecting users with wiring error and leakage fault in low-voltage power distribution areas according to claim 1, characterized in that: The preset threshold is 0.

9.

3. The method for detecting users with wiring error and leakage fault in low-voltage power distribution areas according to claim 1, characterized in that: The value range of the set sampling interval is [5 min, 30 min].

4. The method for detecting users with wiring error and leakage fault in low-voltage power distribution areas according to claim 1, characterized in that: m≥200。 5. A device for detecting users with wiring error and leakage fault in low-voltage power distribution areas, characterized in that, It includes a computer device; The computer device is configured or programmed to execute the steps of the method for detecting users with wiring error and leakage fault in low-voltage power distribution areas according to any one of claims 1-4.

Images