Voltage sag domain online identification method and system based on dynamic fuzzy rule base

Abstract

The application discloses a voltage sag region online identification method and system based on a dynamic fuzzy rule base, and belongs to the field of power system power quality analysis and protection. The method comprises the following steps: calculating initial voltage sag critical points on each line in a power system to generate an initial voltage sag region AOV; acquiring voltage amplitude deviation, phase jump rate and fault point electric distance data of a load point, and performing fuzzy processing on the data by using a membership function to obtain a membership set, inputting the membership set into a TSK fuzzy reasoning model based on a dynamic updating rule base to output a critical point correction amount; and adaptively updating a region boundary according to the initial voltage sag critical points and the critical point correction amount to output an accurate voltage sag region. By introducing a dynamic fuzzy reasoning and online updating mechanism, the application effectively improves the accuracy and adaptability of voltage sag region identification, and solves the problems that existing methods cannot comprehensively consider multiple influence factors and lack online updating capability.

Description

Technical Field

[0001] This invention relates to the technical field of power quality analysis and protection in power systems, specifically to an online identification method and system for voltage sag domains based on a dynamic fuzzy rule base. Background Technology

[0002] The power distribution network is characterized by a high penetration rate of distributed generation (DG) and large-scale application of power electronic equipment. While this transformation has effectively increased the proportion of clean energy consumption, it has also made the interaction between system sources and loads increasingly complex, leading to increasingly prominent transient voltage quality issues such as voltage sags. Engineering practice shows that factors such as the high-frequency start-up and shutdown processes of power electronic equipment and the output fluctuations of distributed generation (such as the light sensitivity fluctuations of photovoltaic output and the gust effect of wind power) have caused the frequency of transient voltage disturbances in the power distribution network to increase by more than 35% year-on-year. These problems are particularly pronounced in industrial scenarios such as precision manufacturing and semiconductor production—when the voltage sag amplitude is less than 0.8 pu and the duration exceeds 10 ms, it is very easy to trigger protective shutdowns of sensitive loads such as programmable logic controller (PLC) control systems and servo motors. According to industry statistics, the average direct economic loss caused by a single shutdown can reach hundreds of thousands of yuan, and the indirect production interruption losses are amplified exponentially.

[0003] The current mainstream "passive governance" model (such as fixed installation of static reactive power generators, dynamic voltage restorers and other compensation devices) has significant engineering bottlenecks: on the one hand, its static configuration strategy cannot match the spatiotemporal characteristics of minute-level output fluctuations of distributed energy and dynamic load changes, resulting in the actual utilization rate of governance resources being less than 40% for a long time; on the other hand, traditional methods rely on offline simulation results under preset scenarios, which are difficult to cope with the changes in boundary conditions caused by dynamic reconfiguration of distribution network topology (such as ring main unit switching operation) and random load switching.

[0004] Voltage sag domain, as a key engineering indicator describing the risk of voltage sag on the system busbar, provides a quantitative basis for the location planning and protection scheme design of sensitive loads by quantifying the voltage sag coverage under different fault types and fault locations. However, existing research has three limitations in terms of engineering applicability: First, it relies too heavily on precise grid parameters (such as line impedance and transformer short-circuit impedance) to solve the fault voltage equation, but in actual distribution networks, there are many old line parameters with measured errors (usually reaching 15% to 20%), which can easily lead to numerical divergence in critical point calculations; second, it uses static topology models, which cannot respond to dynamic scenarios such as distribution network tie switch operation and distributed generation connection and disconnection; third, it evaluates single characteristics such as sag amplitude and duration in isolation, ignoring the coupling effect of multi-dimensional parameters (such as phase transition rate and electrical distance from the fault point)—engineering data shows that about 30% of sensitive load maloperations are triggered by the coupling of multiple features.

[0005] Therefore, there is an urgent need to design a voltage sag domain identification method to solve the above-mentioned technical problems. Summary of the Invention

[0006] The technical problem to be solved by this invention is how to address the shortcomings of existing voltage sag domain identification methods, such as their failure to comprehensively consider multiple influencing factors and lack of online update capabilities.

[0007] This invention solves the above-mentioned technical problems through the following technical means: an online identification method for voltage sag domains based on a dynamic fuzzy rule base, comprising:

[0008] S1. Calculate the initial voltage sag boundary point on each line in the power system to generate the initial area of ​​vulnerability (AOV).

[0009] S2. Obtain the voltage amplitude deviation, phase jump rate, and electrical distance data of the load point, set corresponding fuzzy subsets for the voltage amplitude deviation, phase jump rate, and electrical distance data of the fault point, and use a preset membership function to fuzzify the collected data into a membership set.

[0010] S3. Input the membership set into the TSK (Takagi-Sugeno-Kang) fuzzy inference model. The TSK fuzzy inference model performs inference based on a dynamically updated fuzzy rule base and outputs the critical point correction amount.

[0011] S4. Based on the initial voltage sag boundary obtained in step S1 and the critical point correction obtained in step S3, the boundary of the initial voltage sag domain AOV is adaptively updated to output the final accurate voltage sag domain.

[0012] This invention calculates the initial voltage sag boundary point on each line in a power system, providing a reliable foundation for the Initial Voltage Sag Domain (AOV). It collects real-time data on voltage amplitude deviation, phase jump rate, and electrical distance to fault points, and then fuzzifies this data to align with actual engineering requirements. Based on a dynamically updated fuzzy rule base, a TSK fuzzy inference model accurately fits the complex nonlinear relationships of the power grid, outputting reliable critical point corrections. Finally, by adaptively updating the AOV boundary, the AOV accurately matches the real-time operating conditions of the power grid, providing strong support for power grid risk identification and safe, stable operation.

[0013] Furthermore, the setting of the corresponding fuzzy subset in step S2 includes:

[0014] A fuzzy subset used to describe voltage amplitude deviation, the fuzzy subset including severe voltage drop, slight voltage drop, normal, slight overvoltage, and severe overvoltage;

[0015] A fuzzy subset used to describe the phase transition rate, the fuzzy subset including fast transitions, slow transitions, and hold;

[0016] A fuzzy subset used to describe the electrical distance to the fault point, the fuzzy subset including nearby, medium distance and far distance.

[0017] Furthermore, the formula for calculating the voltage amplitude deviation in step S2 is as follows:

[0018]

[0019] In the formula, This is the measured effective value of the voltage. This is the nominal voltage.

[0020] Furthermore, the formula for calculating the phase transition rate in step S2 is as follows:

[0021]

[0022] In the formula, The initial phase angle, The current phase angle, For time intervals.

[0023] Furthermore, the electrical distance to the fault point mentioned in step S2 is estimated using the line impedance method, and its calculation formula is as follows:

[0024]

[0025]

[0026] In the formula, This is the voltage before the fault. This is the current before the fault. For conjugate, The square of the modulus, The virtual part, The impedance per unit length of the line, including positive sequence. and zero order Quantity, This represents the total length of the line.

[0027] Furthermore, the formula for calculating the critical point correction amount in step S3 is as follows:

[0028]

[0029]

[0030] In the formula, This is the critical point correction amount. For the first r Each rule activates the weight. For a linear function of the regular consequent, R For the total number of rules, These are the membership functions corresponding to voltage amplitude deviation, phase jump rate, and electrical distance from the fault point, respectively, and their expressions are as follows:

[0031]

[0032]

[0033]

[0034] In the formula, For input variables, and These are the center value and width of the Gaussian membership function, respectively. , , , These represent the left boundary, left core point, right core point, and right boundary of the trapezoidal membership function, respectively. , , These represent the left boundary, core point, and right boundary of the membership function of the triangle, respectively.

[0035] Furthermore, the fuzzy rule form of the TSK fuzzy inference model described in step S3 is as follows:

[0036]

[0037] In the formula, These are the fuzzy subsets of each input variable under the corresponding fuzzy rule. u, v, w These are the indices of the fuzzy subsets under the corresponding fuzzy rules. uThe value can be 1, 2, 3, 4, or 5. v The value can be 1, 2, or 3. w The value can be 1, 2, or 3. AND is the fuzzy connection operator. These are the preset coefficients for the corresponding fuzzy rules. It is a constant.

[0038] Furthermore, the fuzzy rule base has a dynamic update mechanism for rule coefficients. Updates are triggered when the system topology changes or when the PMU (Phasor Measurement Unit) detects a sag event, adjusting the rule coefficients online according to the following formula:

[0039]

[0040] In the formula, To preset the historical coefficient weights, For learning rate, These represent the deviations between the measured and predicted values ​​for voltage amplitude deviation, phase jump rate, and electrical distance from the fault point, respectively.

[0041] This invention also provides an online voltage sag domain identification system based on a dynamic fuzzy rule base, comprising:

[0042] Critical point calculation module: used to calculate the initial voltage sag critical point on each line in the power system and generate the initial voltage sag domain (AOV).

[0043] Data acquisition and fuzzification module: used to acquire voltage amplitude deviation, phase jump rate and electrical distance data of load point, set corresponding fuzzy subsets for voltage amplitude deviation, phase jump rate and electrical distance data of fault point respectively, and use preset membership function to fuzzify the acquired data into membership set;

[0044] Fuzzy inference module: used to input the membership degree set into the TSK fuzzy inference model, which performs inference based on a dynamically updated fuzzy rule base and outputs the critical point correction amount;

[0045] Voltage Sag Domain Update Module: Based on the initial voltage sag boundary obtained by the critical point calculation module and the critical point correction amount obtained by the fuzzy inference module, the module adaptively updates the boundary of the initial voltage sag domain AOV and outputs the final accurate voltage sag domain.

[0046] The present invention also provides a processing device, including at least one processor and at least one memory communicatively connected to the processor, wherein: the memory stores program instructions executable by the processor, and the processor can execute the above-described method steps by calling the program instructions.

[0047] The advantages of this invention are:

[0048] This invention collects voltage amplitude deviation, phase jump rate, and electrical distance data of fault points in real time. After fuzzing processing to resolve data uncertainty, compared with traditional methods that rely on static parameters, it achieves accurate perception of the real-time status of the power grid and provides core data support for dynamic updates of AOV.

[0049] This invention uses the TSK fuzzy inference model based on a dynamically updated rule base to accurately fit the complex nonlinear and time-varying relationships between power grid data, outputting reliable critical point correction values, avoiding the failure problem of fixed rule models when operating conditions change, and ensuring correction accuracy.

[0050] This invention dynamically adjusts the boundary based on the initial voltage sag threshold and critical point correction, effectively compensating for errors caused by ignoring real-time operating condition fluctuations and simplifying the model in the initial calculation. This ensures that the final AOV accurately matches the actual state of the power grid, significantly improving the accuracy of voltage sag risk characterization and providing a reliable basis for risk identification. Attached Figure Description

[0051] Figure 1 This is a flowchart of the online voltage sag domain identification method based on a dynamic fuzzy rule base according to Embodiment 1 of the present invention;

[0052] Figure 2 This is a schematic diagram of the numerical-data dual-drive architecture of Embodiment 1 of the present invention;

[0053] Figure 3 This is a membership function graph of the voltage amplitude deviation in Embodiment 1 of the present invention;

[0054] Figure 4 This is a membership function diagram of the phase jump rate in Embodiment 1 of the present invention;

[0055] Figure 5 This is a membership function diagram of the electrical distance to the fault point in Embodiment 1 of the present invention. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Example 1

[0058] like Figure 1As shown, the online identification method for voltage sag domains based on a dynamic fuzzy rule base includes:

[0059] S1. Calculate the initial voltage sag boundary point on each line in the power system and generate the initial voltage sag domain AOV.

[0060] Specifically, based on power flow calculations, the pre-fault voltages of all nodes in the system are obtained, positive-sequence, negative-sequence, and zero-sequence impedance matrices are constructed, and then... LU Decompose to generate sequence impedance parameters.

[0061] The calculation of fault voltage includes sequence impedance, transmission impedance, and pre-fault voltage, using variables. p Expressed as:

[0062]

[0063]

[0064]

[0065] In the formula, The input impedance at fault location K is... and The input impedances on buses F and T are respectively. The line impedance between buses F and T is... The sequence impedance between buses F and T is given. The sequence transfer impedance between bus s and fault location K. The sequence transmission impedance between bus s and F. The sequence transfer impedance between bus s and T. This represents the pre-fault voltage at fault location K. The voltage before the fault on bus F. This represents the voltage before the fault on the T-bus. The ratio of the fault location length to the total length of the line FT (0 <p<1)。

[0066] Based on the grid impedance parameters, the initial critical point is calculated using the fault voltage equation. The calculation formula is as follows:

[0067]

[0068] In the formula, This is the fault voltage of phase A. The voltage before the fault at bus S is... , , These are the zero-sequence, positive-sequence, and negative-sequence transfer impedances between bus s and fault location K, respectively. , , These are the zero-sequence, positive-sequence, and negative-sequence impedances input at the fault location K, respectively.

[0069] By calculating and storing the voltage sag values ​​caused by faults on each bus, the voltage sag value vector of bus s is obtained. V mag The calculation formula is as follows:

[0070]

[0071] In the formula, For nodes i ( i =1,2,…, n The voltage dip caused by a fault at bus S. n This indicates the number of system nodes.

[0072] Define a node vulnerability index (BVI), which is determined by analyzing data from... V mag Subtract the given voltage threshold from the middle V th The calculation formula is as follows:

[0073]

[0074]

[0075] In the formula, for V mag and V th The difference, For nodes i Vulnerability indicators Represents a node i When a fault occurs, the bus voltage is below the threshold. V th It belongs to AOV.

[0076] Define the Vulnerability Index of Line LVI End node F j and T j Between j lines LVI By using the bus BVI The values ​​are added together to determine the result, and the calculation formula is as follows:

[0077]

[0078] In the formula, End node F j and T j Between j ( j =1,2,…, m The vulnerability index of the line, where m is the number of lines in the system. For nodes F j Vulnerability indicators For nodes T j Vulnerability indicators Indicates the line j Completely outside of AOV This indicates that one end of the line is within the AOV. This indicates that both ends of the line are within the AOV.

[0079] Based on the above, the calculation of the initial voltage transient threshold and the AOV boundary only applies to... The lines need to be processed:

[0080] (1) Case 1:

[0081] Calculate the midpoint of the route ( p =0.5) transient decrease value | f (0.5)|, using the amplitudes at both endpoints and the midpoint, a second interpolation is performed to construct the equation. Where a, b, and c are the coefficients of the quadratic interpolation equation. Given the line position variable to be solved, solve for the roots of the interpolation equation. p ic As the initial value for the secant method, the precise initial voltage transient threshold is calculated iteratively using the secant method. .

[0082] (2) Case 2:

[0083] The maximum descent point is found using the golden section search method. p max and amplitude | f ( p max )|, determine whether AOV needs to be segmented, if V th >| f ( p max If |, then the entire line is within the AOV; otherwise, use the endpoints and p max Perform quadratic interpolation on the points to find the roots. pic1 , p ic2 Using the secant method as initial values, the AOV segment is determined, and the initial voltage transient threshold is obtained. .

[0084] S2. Obtain the voltage amplitude deviation, phase jump rate, and electrical distance data of the load point. Set corresponding fuzzy subsets for the voltage amplitude deviation, phase jump rate, and electrical distance data of the fault point, and use the preset membership function to fuzzify the collected data into a membership set.

[0085] Specifically, the TSK (Takagi-Sugeno-Kang) fuzzy inference model is an intelligent inference model that integrates fuzzy logic and numerical computation. Proposed by Takagi, Sugeno, and Kang in the 1980s, it is widely used in complex nonlinear system modeling and multi-dimensional feature analysis. Its core structure comprises three parts: first, a fuzzification layer that maps input variables to a predefined fuzzy subset and quantifies feature membership through a membership function; second, a rule base consisting of several "IF-THEN" rules, where the rule conclusions use linear expressions of the input variables, rather than the fuzzy sets of traditional fuzzy models; and third, a defuzzification layer that transforms the fuzzy inference results into precise numerical outputs through weighted summation. The TSK fuzzy inference model's multi-dimensional feature collaborative processing capability adapts to engineering complexity, combining the robustness of fuzzy logic with the accuracy of numerical computation. Its efficient inference characteristics meet real-time requirements, and the interpretability of the rule base facilitates engineering debugging and iteration. It is the synergistic effect of these characteristics that makes the TSK fuzzy model applicable in engineering fields such as power system transient analysis, industrial process control, and intelligent equipment fault diagnosis.

[0086] This invention considers the impact of three factors on sag domain identification: voltage amplitude deviation, phase transition rate, and electrical distance to the fault point. For each of these factors, corresponding fuzzy subsets are defined. Based on actual conditions and understanding, voltage amplitude deviation is defined as five fuzzy subsets: "Severe Voltage Drop NB", "Minor Voltage Drop NS", "Normal ZO", "Minor Overvoltage PS", and "Severe Overvoltage PB"; phase transition rate is defined as three fuzzy subsets: "FastDrop", "SlowDrop", and "Hold"; and electrical distance to the fault point is defined as three fuzzy subsets: "Near", "Medium", and "Far". For the input parameters of each set of influencing factors, the membership function of each fuzzy subset is used to determine the corresponding membership value.

[0087] Figure 2This diagram illustrates a numerical-data dual-drive architecture. By inputting grid parameters and an impedance matrix, the initial AOV boundary can be obtained using the golden section algorithm and the secant method. Real-time monitoring data from the PMU is used to dynamically update the fuzzy rule base. Finally, the critical point correction is obtained through the TSK fuzzy inference model, and the AOV boundary is adaptively updated.

[0088] Based on the linear function of the rule consequent, fuzzy subsets corresponding to the three influencing factors are defined, as shown in Table 1.

[0089] Table 1 Fuzzy Subset Settings

[0090]

[0091] The formula for calculating voltage amplitude deviation is:

[0092]

[0093] In the formula, This is the measured effective value of the voltage. This is the nominal voltage.

[0094] The formula for calculating the phase transition rate is:

[0095]

[0096] In the formula, The initial phase angle, The current phase angle, For time intervals.

[0097] The electrical distance to the fault point is estimated using the horizontal impedance method, and the calculation formula is as follows:

[0098]

[0099]

[0100] In the formula, This is the voltage before the fault. This is the current before the fault. For conjugate, The square of the modulus, The virtual part, The impedance per unit length of the line, including positive sequence. and zero order Quantity, This represents the total length of the line.

[0101] S3. Input the membership set into the TSK fuzzy inference model. The TSK fuzzy inference model performs inference based on the dynamically updated fuzzy rule base and outputs the critical point correction amount.

[0102] The expressions for the critical point correction and weights are as follows:

[0103]

[0104]

[0105] In the formula, This is the critical point correction amount. For the first r Each rule activates the weight. For a linear function of the regular consequent, R For the total number of rules, These are the membership functions corresponding to voltage amplitude deviation, phase jump rate, and electrical distance from the fault point, respectively, and their expressions are as follows:

[0106]

[0107]

[0108]

[0109] The membership function parameters of the TSK fuzzy inference model are defined as shown in Tables 2-4, and the specific membership function graph is shown below. Figures 3-5 As shown.

[0110] Table 2 Membership function parameters for voltage amplitude deviation

[0111]

[0112] Table 3 Membership function parameters for phase transition rate

[0113]

[0114] Table 4 Membership function parameters of electrical distance from fault point

[0115]

[0116] The form of fuzzy rules is as follows:

[0117]

[0118] In the formula, These represent the fuzzy subsets of each input variable under the corresponding fuzzy rule, and AND is the fuzzy connection operator. These are the preset coefficients for the corresponding fuzzy rules. It is a constant. u, v, w This represents the index of the fuzzy subset corresponding to the fuzzy rule. u The value can be 1, 2, 3, 4, or 5. v The value can be 1, 2, or 3. wIt can take the value 1, 2 or 3.

[0119] This invention also defines a dynamic update mechanism for rule coefficients. When the system topology changes (such as circuit breaker operation / DG switching), the PMU detects a sag event or a large data change, triggering the update condition, according to GB / T

[0120] The definition of voltage sag, the requirements for detection time resolution and phase transition rate on the impact of voltage sag in 30137-2013 "Power Quality Voltage Sags and Short Interruptions" are as follows:

[0121] 1. The continuous time when the voltage is below 0.9 pu is ≥ 10ms.

[0122] 2. The effective value change rate of adjacent cycles voltage is >5% / ms.

[0123] 3. Accompanying phase transition rate | When |>2° / ms, the voltage change rate threshold is relaxed to 8%.

[0124] Adjust the rule coefficients online using the following formula:

[0125]

[0126] In the formula, To preset the historical coefficient weights, The learning rate is initially 0.3, decreasing as the error decreases. This represents the deviation between the measured and predicted values ​​for voltage amplitude deviation, phase jump rate, and electrical distance from the fault point.

[0127] S4. Based on the initial voltage sag boundary obtained in step S1 and the critical point correction obtained in step S3, the boundary of the initial voltage sag domain AOV is adaptively updated to output the final accurate voltage sag domain. The update formula is as follows:

[0128]

[0129] In the formula, The initial voltage transient threshold is obtained in step S1. The set, This is the critical point adjustment amount for the first update. This is the critical point for the first update. , The first y , y The critical point of +1 update, For the first y +1 update of critical point correction amount.

[0130] Example 2

[0131] Based on Embodiment 1, Embodiment 2 of the present invention also provides an online voltage sag domain identification system based on a dynamic fuzzy rule base, including:

[0132] Critical point calculation module: used to calculate the initial voltage sag critical point on each line in the power system and generate the initial voltage sag domain (AOV).

[0133] Data acquisition and fuzzification module: used to acquire voltage amplitude deviation, phase jump rate and electrical distance data of load points, set corresponding fuzzy subsets for voltage amplitude deviation, phase jump rate and electrical distance data of fault points respectively, and use preset membership function to fuzzify the acquired data into membership set;

[0134] Specifically, in the data acquisition and fuzzification module, corresponding fuzzy subsets are set for the voltage amplitude deviation, phase transition rate, and electrical distance data to the fault point, including:

[0135] A fuzzy subset used to describe voltage amplitude deviation, the fuzzy subset including severe voltage drop, slight voltage drop, normal, slight overvoltage, and severe overvoltage.

[0136] A fuzzy subset used to describe the phase transition rate, the fuzzy subset including fast transitions, slow transitions, and hold.

[0137] A fuzzy subset used to describe the electrical distance to the fault point, the fuzzy subset including nearby, medium distance and far distance.

[0138] The formula for calculating voltage amplitude deviation in the data acquisition and fuzzification module is as follows:

[0139]

[0140] In the formula, This is the measured effective value of the voltage. This is the nominal voltage.

[0141] The formula for calculating the phase transition rate in the data acquisition and fuzzification module is as follows:

[0142]

[0143] In the formula, The initial phase angle, The current phase angle; For time intervals.

[0144] In the data acquisition and fuzzification module, the electrical distance to the fault point is estimated using the row impedance method, and the calculation formula is as follows:

[0145]

[0146]

[0147] In the formula, This is the voltage before the fault. This is the current before the fault. For conjugate, The square of the modulus, The virtual part, The impedance per unit length of the line, including positive sequence. and zero order Quantity, This represents the total length of the line.

[0148] Fuzzy Inference Module: This module is used to input the membership degree set into the TSK fuzzy inference model. The TSK fuzzy inference model performs inference based on a dynamically updated fuzzy rule base and outputs the critical point correction amount.

[0149] Specifically, the formula for calculating the critical point correction amount in the fuzzy inference module is as follows:

[0150]

[0151]

[0152] In the formula, This is the critical point correction amount. For the first r Each rule activates the weight. For a linear function of the regular consequent, R For the total number of rules, These are the membership functions corresponding to voltage amplitude deviation, phase jump rate, and electrical distance from the fault point, respectively, and their expressions are as follows:

[0153]

[0154]

[0155]

[0156] In the formula, For input variables, and These are the center value and width of the Gaussian membership function, respectively. , , , These represent the left boundary, left core point, right core point, and right boundary of the trapezoidal membership function, respectively. , , These represent the left boundary, core point, and right boundary of the membership function of the triangle, respectively.

[0157] The fuzzy rule form of the TSK fuzzy inference model in the fuzzy inference module is as follows:

[0158]

[0159] In the formula, These are the fuzzy subsets of each input variable under the corresponding fuzzy rule. u , v , w These are the indices of the fuzzy subsets under the corresponding fuzzy rules. u The value can be 1, 2, 3, 4, or 5. v The value can be 1, 2, or 3. w The value can be 1, 2, or 3. AND is the fuzzy connection operator. These are the preset coefficients for the corresponding fuzzy rules. It is a constant.

[0160] The fuzzy inference module also features a dynamic rule coefficient update mechanism. When the system topology changes or the PMU detects a temporary descent event, the update condition is triggered, and the rule coefficients are adjusted online according to the following formula:

[0161]

[0162] In the formula, To preset the historical coefficient weights, For learning rate, These represent the deviations between the measured and predicted values ​​for voltage amplitude deviation, phase jump rate, and electrical distance from the fault point, respectively.

[0163] Voltage Sag Domain Update Module: Based on the initial voltage sag boundary obtained by the critical point calculation module and the critical point correction amount obtained by the fuzzy inference module, the module adaptively updates the boundary of the initial voltage sag domain AOV and outputs the final accurate voltage sag domain.

[0164] Example 3

[0165] Based on Embodiment 1, Embodiment 3 of the present invention also provides a processing device, including at least one processor and at least one memory communicatively connected to the processor, wherein: the memory stores program instructions executable by the processor, and the processor can execute the method steps of Embodiment 1 by calling the program instructions.

[0166] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for online identification of voltage sag domains based on a dynamic fuzzy rule base, characterized in that, include: S1. Calculate the initial voltage sag boundary point on each line in the power system and generate the initial voltage sag domain AOV. S2. Obtain the voltage amplitude deviation, phase jump rate, and electrical distance data of the load point, set corresponding fuzzy subsets for the voltage amplitude deviation, phase jump rate, and electrical distance data of the fault point, and use a preset membership function to fuzzify the collected data into a membership set. S3. Input the membership set into the TSK fuzzy inference model. The TSK fuzzy inference model performs inference based on a dynamically updated fuzzy rule base and outputs the critical point correction amount. The formula for calculating the critical point correction amount in step S3 is as follows: In the formula, This is the critical point correction amount. For the first r Each rule activates the weight. For a regular consequent linear function, R For the total number of rules, These are the membership functions corresponding to voltage amplitude deviation, phase jump rate, and electrical distance from the fault point, respectively, and their expressions are as follows: In the formula, For input variables, and These are the center value and width of the Gaussian membership function, respectively. , , , These represent the left boundary, left core point, right core point, and right boundary of the trapezoidal membership function, respectively. , , These represent the left boundary, core point, and right boundary of the triangle membership function, respectively. The fuzzy rule base has a dynamic update mechanism for rule coefficients. The update is triggered when the system topology changes or the PMU detects a temporary descent event, and the rule coefficients are adjusted online according to the following formula: In the formula, To preset the historical coefficient weights, For learning rate, These are the deviations between the measured and predicted values ​​for voltage amplitude deviation, phase jump rate, and electrical distance from the fault point, respectively. S4. Based on the initial voltage sag boundary obtained in step S1 and the critical point correction obtained in step S3, the boundary of the initial voltage sag domain AOV is adaptively updated to output the final accurate voltage sag domain.

2. The online identification method for voltage sag domain based on a dynamic fuzzy rule base according to claim 1, characterized in that, The setting of the corresponding fuzzy subset in step S2 includes: A fuzzy subset used to describe voltage amplitude deviation, the fuzzy subset including severe voltage drop, slight voltage drop, normal, slight overvoltage, and severe overvoltage; A fuzzy subset used to describe the phase transition rate, the fuzzy subset including fast transitions, slow transitions, and hold; A fuzzy subset used to describe the electrical distance to the fault point, the fuzzy subset including nearby, medium distance and far distance.

3. The online identification method for voltage sag domain based on a dynamic fuzzy rule base according to claim 1, characterized in that, The formula for calculating the voltage amplitude deviation in step S2 is as follows: In the formula, This is the measured effective value of the voltage. This is the nominal voltage.

4. The online identification method for voltage sag domain based on a dynamic fuzzy rule base according to claim 1, characterized in that, The formula for calculating the phase transition rate in step S2 is as follows: In the formula, The initial phase angle, The current phase angle, For time intervals.

5. The online identification method for voltage sag domain based on a dynamic fuzzy rule base according to claim 1, characterized in that, The electrical distance to the fault point mentioned in step S2 is estimated using the line impedance method, and its calculation formula is as follows: In the formula, This is the voltage before the fault. This is the current before the fault. For conjugate, The square of the modulus, The virtual part, The impedance per unit length of the line, including positive sequence. and zero order Quantity, This represents the total length of the line.

6. The online identification method for voltage sag domain based on a dynamic fuzzy rule base according to claim 1, characterized in that, The fuzzy rule form of the TSK fuzzy inference model described in step S3 is as follows: In the formula, These are the fuzzy subsets of each input variable under the corresponding fuzzy rule. u , v , w These are the indices of the fuzzy subsets under the corresponding fuzzy rules. u The value can be 1, 2, 3, 4, or 5. v The value can be 1, 2, or 3. w The value can be 1, 2, or 3. AND is the fuzzy connection operator. These are the preset coefficients for the corresponding fuzzy rules. It is a constant.

7. An online voltage sag domain identification system based on a dynamic fuzzy rule base, characterized in that, include: Critical point calculation module: used to calculate the initial voltage sag critical point on each line in the power system and generate the initial voltage sag domain (AOV). Data acquisition and fuzzification module: used to acquire voltage amplitude deviation, phase jump rate and electrical distance data of load point, set corresponding fuzzy subsets for voltage amplitude deviation, phase jump rate and electrical distance data of fault point respectively, and use preset membership function to fuzzify the acquired data into membership set; Fuzzy inference module: used to input the membership degree set into the TSK fuzzy inference model, which performs inference based on a dynamically updated fuzzy rule base and outputs the critical point correction amount; The formula for calculating the critical point correction is: In the formula, This is the critical point correction amount. For the first r Each rule activates the weight. For a regular consequent linear function, R For the total number of rules, These are the membership functions corresponding to voltage amplitude deviation, phase jump rate, and electrical distance from the fault point, respectively, and their expressions are as follows: In the formula, For input variables, and These are the center value and width of the Gaussian membership function, respectively. , , , These represent the left boundary, left core point, right core point, and right boundary of the trapezoidal membership function, respectively. , , These represent the left boundary, core point, and right boundary of the triangle membership function, respectively. The fuzzy rule base has a dynamic update mechanism for rule coefficients. The update is triggered when the system topology changes or the PMU detects a temporary descent event, and the rule coefficients are adjusted online according to the following formula: In the formula, To preset the historical coefficient weights, For learning rate, These are the deviations between the measured and predicted values ​​for voltage amplitude deviation, phase jump rate, and electrical distance from the fault point, respectively. Voltage Sag Domain Update Module: Based on the initial voltage sag boundary obtained by the critical point calculation module and the critical point correction amount obtained by the fuzzy inference module, the module adaptively updates the boundary of the initial voltage sag domain AOV and outputs the final accurate voltage sag domain.

8. A processing device, characterized in that, The method includes at least one processor and at least one memory communicatively connected to the processor, wherein the memory stores program instructions executable by the processor, and the processor invokes the program instructions to perform the method as described in any one of claims 1 to 6.

Images