An active power distribution network electromechanical equivalent modeling method, system, medium and computer device

Abstract

This invention discloses a method, system, medium, and computer device for electromechanical equivalent modeling of active distribution networks, relating to the field of power system technology. It aims to construct an equivalent model of an active distribution network that maintains both accuracy and computational efficiency. The method includes grouping similar components in the model according to characteristic similarity using a K-means clustering algorithm to form equivalent units; constructing the distribution network equivalent model, which includes a virtual bus, equivalent impedance, and equivalent units. The equivalent units are connected to the virtual bus (POC) and then to the point of common coupling (PCC) via the distribution network equivalent impedance; and employing trajectory sensitivity analysis to screen key parameters. By calculating the impact of parameter changes on the active and reactive power of the PCC, the parameters with the greatest impact on the model output are identified. This invention comprehensively reflects the load characteristics of distribution networks widely connected to photovoltaic power generation systems and provides technical support for accurate and rapid simulation calculations on the grid-connected side of power systems.

Description

Technical Field

[0001] This invention relates to a method, system, medium, and computer device for electromechanical equivalent modeling of active power distribution networks, belonging to the field of power system technology. Background Technology

[0002] Against the backdrop of accelerated construction of new power systems, the process of electricity substitution is accelerating, and the proportion of electricity in final energy consumption is steadily increasing, placing higher demands on the safe and stable operation of the power grid. Power system simulation, as a key method for understanding system characteristics and ensuring safety and stability, is becoming increasingly important. In recent years, the installed capacity of new energy sources, represented by wind power and photovoltaics, has expanded rapidly, with a significant increase in the proportion of distributed photovoltaics. The large-scale integration of distributed photovoltaics into the distribution network is driving the transformation of traditional passive distribution networks into more complex active distribution networks, profoundly changing load-side characteristics and posing new challenges to the safe and stable operation of the power system. To accurately analyze the dynamic behavior and impact of a high proportion of distributed photovoltaics, a high-precision active distribution network model is essential. However, real-world large power grids often contain hundreds or thousands of active distribution networks. Simulating all of them using detailed models would result in low computational efficiency, convergence difficulties, or even unsolvable problems. Therefore, constructing an equivalent model of an active distribution network that maintains both accuracy and good computational efficiency has become a practical requirement for conducting large-scale power grid safety and stability analysis. Summary of the Invention

[0003] The purpose of this invention is to propose an electromechanical equivalent modeling method, system, medium, and computer device for active distribution networks, which comprehensively reflects the load characteristics of distribution networks widely connected to photovoltaic power generation systems, and provides technical support for accurate and rapid simulation calculations on the grid-connected side of the power system.

[0004] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0005] In a first aspect, this invention proposes a method for electromechanical equivalent modeling of active power distribution networks, comprising:

[0006] Based on the K-means clustering algorithm, similar components in the target active distribution network are grouped according to their characteristic similarity to form equivalent units; the equivalent units include distributed photovoltaic equivalent units, capacitor equivalent units, static load equivalent units, and dynamic load equivalent units;

[0007] Based on the topology and power flow data of the target active distribution network and the equivalent unit, an equivalent model of the distribution network including the virtual bus POC, the equivalent impedance and the equivalent unit is constructed. The equivalent unit is connected to the virtual bus POC and connected to the common coupling point PCC through the equivalent impedance.

[0008] The parameters to be determined in the equivalent model of the distribution network are obtained. Based on the trajectory sensitivity analysis method, the influence of the parameters to be determined on the external power characteristics of the model is determined. Based on the influence, all parameters to be determined are divided into key parameters and minor parameters. For the key parameters, the simulated annealing algorithm is used to optimize and identify them to obtain their optimal values, and the optimal values ​​are used as the parameters of the equivalent model of the distribution network. For the minor parameters, the statistical synthesis method is used to take typical values ​​and fix them.

[0009] Furthermore, the similar components include distributed photovoltaic (PV) systems. These distributed PV systems are clustered, and several distributed PV systems with similar operating characteristics are considered equivalent to typical distributed PV power plants, simplifying the modeling scale. This includes:

[0010] Distributed photovoltaic (PV) systems are categorized based on the PV capacity connected to the distribution network at various voltage levels. This includes current-controlled mode (Category A) distributed PV, power-controlled mode (Category A) distributed PV, and Category B distributed PV capacities. The proportion of current-controlled mode (Category A) distributed PV capacity connected is [missing information]. The proportion of Class A distributed photovoltaic grid connection capacity under power control mode is: The proportion of Class B distributed photovoltaic grid connection capacity is ;

[0011] Based on the distributed photovoltaic (PV) cluster feature vector, the K-means clustering algorithm is used to divide PV power stations with similar characteristics in both the current-controlled (Type A) and power-controlled (Type A) distributed PV systems into the same sub-cluster, making the dynamic characteristics of the PV power stations within the sub-cluster similar. The distributed PV cluster feature vector is obtained by selecting the inverter's outer loop control parameters. , The actual dynamic characteristics of distributed photovoltaic power generation are represented as follows:

[0012] (1)

[0013] In the formula, The feature vector of the distributed photovoltaic cluster; , These are the integral parameters and proportional coefficients of the outer loop PI control of the photovoltaic inverter, respectively.

[0014] After K-means clustering, the photovoltaic arrays of various types are evaluated as single-unit equivalents to determine the output power, photovoltaic array parameters, inverter normal operation control parameters, and ride-through control parameters of each type of single-unit equivalent photovoltaic array.

[0015] The further proposed solution employs the K-means clustering algorithm, using the inverter's outer-loop control parameters as cluster feature vectors to group photovoltaic power plants with similar characteristics into typical units. This method addresses the modeling challenges posed by the large number and dispersed nature of distributed photovoltaic systems. It determines the optimal number of clusters using the elbow method and performs single-unit equivalence based on the power conservation principle, significantly improving the model's computational efficiency.

[0016] Furthermore, the step of using the K-means clustering algorithm to divide distributed photovoltaic power stations with similar characteristics in the current control mode A-type distributed photovoltaic system into the same sub-cluster includes:

[0017] S1: Obtain a sample dataset of Class A distributed photovoltaic systems with current control mode within the target area. ,in The number of Class A distributed photovoltaic systems under current control mode. For the i-th current control mode, type A distributed photovoltaic system, i=1,2,… ;

[0018] S2: The number of clusters of the current control mode A distributed photovoltaic system is determined using the elbow method. This includes: letting the cluster number k be in [2, ... Traverse the interval, calculate the sum of squared errors corresponding to different k values, and use the k corresponding to the inflection point of the sum of squared errors curve as the number of clusters of the current control mode A-type distributed photovoltaic system. Among them, the sum of squared errors The calculation formula is:

[0019] (2)

[0020] In the formula, The initial cluster centers of the i-th class are randomly selected; For the i-th distributed photovoltaic sub-cluster; This represents the sum of squared errors when the number of clusters is k.

[0021] S3: Random selection One sample is used as the initial cluster center for the current control mode A-type distributed photovoltaic;

[0022] S4: Calculate the distance between each sample index and the randomly selected initial cluster centers. The expression is:

[0023] (3)

[0024] In the formula, This represents the v-th attribute of the i-th distributed photovoltaic system. The v-th attribute represents the initial cluster center of the i-th class selected randomly. The dimension of the feature vector, i.e., the number of clustering indicators;

[0025] S5: Based on the distance Each sample is assigned to the nearest cluster to form a sub-cluster, and the mean of the samples in each sub-cluster is used as the new cluster center.

[0026] S6: Repeat S4 and S5 until the cluster centers no longer change or the maximum number of iterations is reached, and output the results of the clustering of the distributed photovoltaic cluster.

[0027] Perform the same clustering process as steps S1-S6 on the power control mode A-type distributed photovoltaic system to complete the division of the power control mode A-type distributed photovoltaic subgroup.

[0028] Furthermore, the step of performing single-unit equivalence on various photovoltaic systems after K-means clustering includes keeping the grid-connected output power unchanged before and after equivalence, so that the equivalent single-unit output power is equal to the sum of the output power of all distributed photovoltaic power stations in the sub-cluster; keeping the output power and output voltage of the photovoltaic array unchanged before and after equivalence, so that the output current of the equivalent array is equal to the sum of the output current of all photovoltaic power stations in the cluster, and adjusting the number of parallel photovoltaic panels in the equivalent array accordingly to complete the calculation of the number of series and parallel connections.

[0029] Furthermore, the distributed photovoltaic equivalent unit adopts a photovoltaic power generation system model, including a type A photovoltaic model and a type B photovoltaic model; wherein the type A photovoltaic model represents a distributed photovoltaic system with low / high voltage ride-through capability, and its low voltage state transition relationship is as follows:

[0030] When the distributed photovoltaic system is in normal condition, if the voltage at the new energy generator terminal is... If the voltage drop is below the low voltage ride-through threshold for new energy sources, the system transitions from the normal state to the ride-through state.

[0031] When the distributed photovoltaic system is in a pass-through state, if the voltage at the new energy generator terminal is... If the voltage exceeds the low-voltage ride-through threshold for new energy sources and the new energy source lacks a post-ride-through ramp-up function, it transitions from the ride-through state to the normal state; if the new energy source's terminal voltage is higher than the low-voltage ride-through judgment threshold, and the new energy source lacks a post-ride-through ramp-up function, it transitions from the ride-through state to the normal state; if the new energy source's terminal voltage... If the voltage is higher than the low voltage ride-through threshold of new energy sources and the new energy source has a ramp-up function after the ride-through, the state will be transferred from the ride-through state to the ramp-up starting point.

[0032] When the distributed photovoltaic system is at the starting point of the ramp-up process, it transitions to the ramp-up process after a calculation step.

[0033] When the distributed photovoltaic system is in the ramp-up process, if the ramp-up ends and the power returns to the level before the ramp-up, it transitions from the ramp-up process to the normal state; if the generator terminal voltage If the voltage drop threshold for new energy low-voltage ride-through is below the threshold, the process transitions from the climbing phase to the ride-through phase.

[0034] For Type A photovoltaic models, the control methods during high and low voltage ride-through periods can be either specified current control or specified power control. The calculation formula for the specified current control method during the low voltage ride-through period is as follows:

[0035] (4)

[0036] In the formula, These are the active current setpoints and reactive current setpoints for distributed photovoltaic systems during low voltage ride-through, respectively. These are the steady-state values ​​of active current and reactive current of distributed photovoltaic systems before the fault occurs; These are the active current reference setting value and reactive current reference setting value for distributed photovoltaic power during low voltage ride-through, respectively. These are the active current coefficient 1 and the reactive current coefficient 1, respectively, representing the degree of influence of the terminal voltage on the active current and reactive current. These are the active current coefficient 2 and the reactive current coefficient 2, respectively, representing the degree of influence of the active current and reactive current before the fault. These are the voltage amplitude at the distributed renewable energy terminal and the threshold for entering the low voltage ride-through process, respectively.

[0037] The formula for calculating the specified power control method during low voltage ride-through is as follows:

[0038] (5)

[0039] In the formula, These are the active and reactive power commands during the renewable energy transit period; , These are the active and reactive power before the new energy source passes through; , These are the active power coefficient and reactive power coefficient of the underpass, respectively. , These are the active and reactive power setting values ​​for low-voltage operation, respectively.

[0040] The formula for calculating the specified current control method during high-voltage ride-through is as follows:

[0041] (6)

[0042] In the formula, These are the active current setpoints and reactive current setpoints for distributed photovoltaic systems during high voltage ride-through, respectively. These are the steady-state values ​​of active current and reactive current of distributed photovoltaic systems before the fault occurs; These are the active current reference setting value and reactive current reference setting value of distributed photovoltaic power generation during high voltage ride-through, respectively. These are related to the power level and reactive power support capability of distributed photovoltaic power generation. These are the voltage coefficients of active current and reactive current, respectively, indicating the degree of influence of terminal voltage on active current and reactive current. These are the initial current coefficients of the active current and the initial current coefficients of the reactive current, respectively, representing the degree of influence of the active and reactive currents before the fault. These are the voltage amplitude at the distributed renewable energy terminal and the threshold for entering the high-voltage ride-through process, respectively.

[0043] The formula for calculating the specified power control method during high-voltage ride-through is as follows:

[0044] (7)

[0045] In the formula, These are the active power setpoints and reactive power setpoints for distributed photovoltaic systems during high-voltage ride-through, respectively. These represent the steady-state values ​​of active power and reactive power of distributed photovoltaic power before the fault occurs. These are the active power control coefficient and the reactive power control coefficient, respectively, which represent the ratio of power during the fault to power before the fault. These are the active power reference setting value and the reactive power reference setting value, respectively.

[0046] The Class B photovoltaic model represents a distributed photovoltaic system without voltage ride-through capability. When the grid connection voltage of the photovoltaic power station is lower than the rated voltage threshold, wave blocking control is used instead of low voltage ride-through control. The Class B photovoltaic model starts wave blocking control at time t=0, and the power calculation formula during the wave blocking period is:

[0047] (8)

[0048] In the formula, , These represent the active and reactive power during the wave blocking period, respectively. The active power before entering the wave blocking phase; The duration of the wave-blocking state; This refers to the recovery speed of active power.

[0049] Furthermore, the static load equivalent unit adopts a static load model. Based on the law of load power variation with voltage, the static load model uses a polynomial model to equivalently describe the static characteristics of the load. The calculation formula is as follows:

[0050] (9)

[0051] In the formula, , , , These represent the load active power, reactive power, load bus voltage amplitude, and frequency during the system's steady-state power flow. , , , These are the actual values ​​of the load's active power, reactive power, load bus voltage amplitude, and frequency, respectively. For frequency deviation, ; These are the frequency deviation influencing factors for active and reactive power, respectively. , , These represent the proportions of active power to total load active power under constant impedance, constant current, and constant power conditions, respectively. ; These represent the proportions of reactive power to total load reactive power under constant impedance, constant current, and constant power conditions, respectively. .

[0052] Furthermore, the dynamic load equivalent unit adopts a dynamic load model, which is a third-order electromechanical transient model, and the calculation formula is as follows:

[0053] (10)

[0054] In the formula, These are the transient electromotive forces along the d-axis and q-axis of the induction motor, respectively. The inertial time constant; For rotor resistance, The rotor circuit time constant when the stator is open-circuited; , These are the rotor open-circuit reactance and short-circuit reactance, respectively. These are the rotor speed and the rated speed, respectively. These are the d-axis and q-axis components of the stator-side transient current, respectively. This refers to the rotor's electromagnetic torque. This represents the rotor's mechanical torque.

[0055] Furthermore, the equivalent capacitive reactance of the capacitor equivalent unit for:

[0056] (11)

[0057] In the formula, Equivalent capacitive reactance of the distribution network; This represents the total number of feeders in the distribution network. The capacitance of the distribution network feeder f is calculated to the 110kV voltage level. Equivalent capacitance of the distribution network; For the complex frequency domain factor, where The imaginary unit, ω is the angular frequency.

[0058] Furthermore, the equivalent impedance of the distribution network is calculated based on the principle of consistent network loss, and the calculation formula is as follows:

[0059] (12)

[0060] In the formula, , These are the distribution network resistance and reactance, respectively. , , These are the active power, reactive power, and voltage at PCC, respectively. , These represent the active power and reactive power at the virtual bus POC, respectively.

[0061] Furthermore, the trajectory sensitivity analysis method compares the trajectory sensitivity of the parameter to be determined by using average trajectory sensitivity. The formula for calculating the average trajectory sensitivity is as follows:

[0062] (13)

[0063] In the formula, For average trajectory sensitivity, This is the starting time for the calculation; To calculate the duration; The formula for calculating trajectory sensitivity is:

[0064] (14)

[0065] In the formula, The number of parameters to be identified; For parameters Initial value; For the first The change in each parameter This represents the active and reactive power response curves of the detailed model at the point of common coupling (PCC) when all parameters are given values. This indicates that other parameters remain unchanged, the first The parameters are from Become The detailed model's power response curve at PCC is shown.

[0066] The further proposed approach employs trajectory sensitivity analysis to screen key parameters. By calculating the impact of parameter changes on the active and reactive power of the PCC, the parameters with the greatest influence on the model output are identified. This method of determining model parameters significantly improves the computational efficiency and convergence speed of parameter identification while ensuring identifiability.

[0067] Furthermore, the step of using simulated annealing algorithm to optimize and identify it to obtain its optimal value includes:

[0068] S1: Obtain the initial voltage at the common connection point PCC Initial active power Initial reactive power ;

[0069] S2: Construct the steady-state equations of the equivalent model, using the initial voltage obtained in step S1. Initial active power Initial reactive power Substituting into the steady-state equations, the initial values ​​of the state variables and the values ​​of the intermediate variables of the equivalent model are obtained by inverse solution;

[0070] S3: Set the initial voltage As input, the key parameters are assigned values ​​based on the simulated annealing algorithm to obtain the output response of the equivalent model;

[0071] S4: Solve for the response power of the photovoltaic integrated load model based on the power balance equation. The calculation formula is as follows:

[0072] (15)

[0073] In the formula, These are the equivalent impedance of the distribution network, the static load, and the active power consumed by the induction motor, respectively. Active power injected into the equivalent distribution network for the equivalent photovoltaic system; These are the equivalent impedance of the distribution network, the static load, and the reactive power consumed by the induction motor, respectively. These refer to the reactive power injected into the equivalent distribution network by the equivalent capacitor and the equivalent photovoltaic system, respectively. , These represent the active power and reactive power at PCC, respectively.

[0074] S5: The root mean square error between the power measurement values ​​of the detailed model and the power output values ​​of the equivalent model is used as the optimization objective to achieve dynamic fitting of the equivalent model to the power response of the target model. The calculation formula is as follows:

[0075] (16)

[0076] In the formula, This is the root mean square error; , The first Detailed model active power and reactive power at each sampling time; , The first The equivalent model active power and reactive power at each sampling time; This represents the total number of sampling points during the entire simulation period.

[0077] S6: Solve for the objective function value of the current parameter set, return to S3, and continue iterating until the maximum number of iterations is reached, then output the parameter set that satisfies the optimal optimization objective.

[0078] Secondly, the present invention proposes an electromechanical equivalent modeling system for active power distribution networks, the system comprising: an equivalent unit construction module, a power distribution network equivalent model construction module, and a parameter identification and optimization module;

[0079] The equivalent unit construction module is used to cluster elements with similar characteristics in the model using the K-means clustering algorithm;

[0080] The power distribution network equivalent model construction module is used to construct a comprehensive equivalent model structure;

[0081] The parameter identification and optimization module is used to determine the parameters of the equivalent model of the power distribution network.

[0082] Thirdly, the present invention proposes a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of an active power distribution network electromechanical equivalent modeling method.

[0083] Fourthly, the present invention provides a computer device comprising:

[0084] Memory, used to store computer programs;

[0085] A processor for executing the computer program to implement the steps of the electromechanical equivalent modeling method for active power distribution networks.

[0086] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0087] (1) High equivalence accuracy, which can more realistically reflect the dynamic characteristics of active distribution networks. Due to the adoption of the multi-machine equivalence concept and refined clustering index, this model can effectively characterize the differences in dynamic response caused by different photovoltaic types and locations within the distribution network. Compared with traditional models, it can more accurately simulate complex transient phenomena such as partial photovoltaic grid disconnection and voltage delay recovery, thereby providing more reliable simulation results for the transient stability analysis of the power grid and reducing safety risks.

[0088] (2) It achieves a good balance between computational efficiency and accuracy, and is highly practical. This invention proposes selecting the inverter's outer loop control parameters as the basis for clustering and using the K-means clustering algorithm for clustering. Furthermore, the complexity of parameter identification is significantly reduced through key parameter screening. This enables the constructed model to maintain high accuracy while having a fast simulation speed, meeting the stringent requirements of computational efficiency for large power grid safety and stability analysis, and demonstrating good prospects for engineering applications.

[0089] (3) Wide applicability and adaptability, suitable for various scenarios. The method proposed in this invention is not only applicable to different types of distributed photovoltaic (Type A / Type B), but its clustering and parameter identification ideas are also universal. By adjusting the clustering index and parameter set, the model can adapt to the simulation requirements under different penetration levels, different grid strengths (short-circuit capacity), and different operating conditions (such as power feedback), showing strong robustness and wide applicability. Attached Figure Description

[0090] Figure 1 This is the state transition diagram for Class A new energy sources in this embodiment of the invention;

[0091] Figure 2 This is the equivalent model structure of the power distribution network in this embodiment of the invention. Detailed Implementation

[0092] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the present invention or its application or use.

[0093] Example 1:

[0094] This embodiment proposes a method for electromechanical equivalent modeling of active power distribution networks, including:

[0095] To obtain the parameters to be determined in the equivalent model of the distribution network, the influence of the parameters to be determined on the external power characteristics of the model is determined based on the trajectory sensitivity analysis method. Based on the influence, all parameters to be determined are divided into key parameters and minor parameters. For the key parameters, the simulated annealing algorithm is used to optimize and identify them to obtain their optimal values, and the optimal values ​​are used as the parameters of the equivalent model of the distribution network. For the minor parameters, the statistical synthesis method is used to take typical values ​​and fix them.

[0096] Based on the K-means clustering algorithm, similar components in the target active distribution network are grouped according to their characteristic similarity to form equivalent units; the equivalent units include distributed photovoltaic equivalent units, capacitor equivalent units, static load equivalent units, and dynamic load equivalent units;

[0097] Based on the topology and power flow data of the target active distribution network and its equivalent units, an equivalent model of the distribution network is constructed, such as... Figure 2 As shown, the equivalent model of the distribution network includes a virtual bus POC, equivalent impedance, and equivalent unit. The equivalent unit is connected to the virtual bus POC and is connected to the common coupling point PCC through the equivalent impedance.

[0098] To obtain the parameters to be determined in the equivalent model of the distribution network, the influence of the parameters on the external power characteristics of the model is determined based on the trajectory sensitivity analysis method. Based on the degree of influence, all parameters to be determined are divided into key parameters and minor parameters. For key parameters, the simulated annealing algorithm is used to optimize and identify them to obtain their optimal values, which are then used as the parameters of the equivalent model of the distribution network. For minor parameters, the statistical synthesis method is used to obtain typical values ​​and fix them.

[0099] In this embodiment, the external power characteristics of the model specifically refer to the active and reactive power characteristics exhibited by the equivalent model of the distribution network at the common coupling point (PCC).

[0100] In this embodiment, the optimal values ​​of all parameters to be determined must be obtained through the parameter identification process described above. This process needs to be adapted to specific factors such as the actual topology of the target distribution network, the type and capacity of photovoltaic access, the load composition, and the operating conditions. Therefore, the parameter values ​​have a significant scenario dependence. By constructing a systematic parameter identification and optimization framework, rather than providing a fixed set of parameters, the method can be flexibly and effectively applied to different actual distribution network scenarios.

[0101] In this embodiment, similar components include distributed photovoltaic (PV) systems. Distributed PV systems are clustered using the K-means clustering algorithm, and several distributed PV systems with similar operating characteristics are considered as typical distributed PV power plants, simplifying the modeling scale. This includes:

[0102] Distributed photovoltaic (PV) systems are categorized based on the PV capacity connected to the distribution network at various voltage levels, including current-controlled mode (Category A), power-controlled mode (Category A), and Category B distributed PV capacities. The proportion of current-controlled mode (Category A) distributed PV capacity connected is [percentage missing]. The proportion of Class A distributed photovoltaic grid connection capacity under power control mode is: The proportion of Class B distributed photovoltaic grid connection capacity is ;

[0103] Based on the distributed photovoltaic (PV) cluster feature vector, the K-means clustering algorithm is used to group distributed PV power plants with similar characteristics in both current-controlled (Type A) and power-controlled (Type A) distributed PV systems into the same sub-cluster, making the dynamic characteristics of the distributed PV power plants within the sub-cluster similar. The distributed PV cluster feature vector is determined by selecting the inverter's outer loop control parameters. , The actual dynamic characteristics of distributed photovoltaic power generation are represented as follows:

[0104] (1)

[0105] In the formula, The feature vector of the distributed photovoltaic cluster; , These are the integral parameters and proportional coefficients of the outer loop PI control of the photovoltaic inverter, respectively.

[0106] After K-means clustering, the photovoltaic arrays of various types are evaluated as single-unit equivalents to determine the output power, photovoltaic array parameters, inverter normal operation control parameters, and ride-through control parameters of each type of single-unit equivalent photovoltaic array.

[0107] In this embodiment, the K-means clustering algorithm is used to divide distributed photovoltaic power stations with similar characteristics in the current-controlled A-type distributed photovoltaic system into the same sub-cluster, including:

[0108] S1: Obtain a sample dataset of Class A distributed photovoltaic systems with current control mode within the target area. ,in The number of Class A distributed photovoltaic systems under current control mode. This is for the i-th current control mode, Class A distributed photovoltaic system.

[0109] S2: Determining the number of clusters in Class A distributed photovoltaic systems using the elbow method for current control modes. This includes: letting the cluster number k be in [2, ... Traverse the interval, calculate the sum of squared errors corresponding to different k values, and use the k corresponding to the inflection point of the sum of squared errors curve as the number of clusters of Class A distributed photovoltaic systems in the current control mode. The formula for calculating the sum of squared errors is:

[0110] (2)

[0111] In the formula, The initial cluster centers of the i-th class are randomly selected; For the i-th distributed photovoltaic sub-cluster; This represents the sum of squared errors when the number of clusters is k.

[0112] S3: Random selection These samples serve as the initial cluster centers for Class A distributed photovoltaic systems under current control mode;

[0113] S4: Calculate the distance between each sample index and the randomly selected initial cluster centers. The expression is:

[0114] (3)

[0115] In the formula, This represents the v-th attribute of the i-th distributed photovoltaic system. The v-th attribute represents the initial cluster center of the i-th class selected randomly. The dimension of the feature vector, i.e., the number of clustering indicators;

[0116] S5: Based on distance Each sample is assigned to the nearest cluster, forming a sub-cluster. The mean of the samples within each sub-cluster is used as the new cluster center, calculated using the following formula:

[0117] (4)

[0118] In the formula, Indicates the first Cluster centers of distributed photovoltaic sub-clusters; Indicates the first The attributes of the photovoltaic power stations contained in a distributed photovoltaic sub-cluster; Indicates the first The number of photovoltaic power stations in a distributed photovoltaic sub-cluster.

[0119] S6: Repeat S4 and S5 until the cluster centers no longer change or the maximum number of iterations is reached, and output the results of the clustering of the distributed photovoltaic cluster.

[0120] Perform the same clustering process as steps S1-S6 on the power control mode A distributed photovoltaic system to complete the subgroup division of the power control mode A distributed photovoltaic system.

[0121] In this embodiment, single-unit equivalents are performed on various photovoltaic systems after K-means clustering, including:

[0122] (1) Equivalent power requirements for distributed photovoltaic power

[0123] The single-unit equivalent power of distributed photovoltaic (PV) systems must first ensure that the output power of the distributed PV system at the grid connection point remains constant before and after the equivalent power is calculated. In other words, the output power of a distributed PV sub-cluster is equal to the output power of all distributed PV power stations within that sub-cluster. The calculation formula is as follows:

[0124] (5)

[0125] In the formula, Let be the equivalent active power of the k-th distributed photovoltaic sub-cluster; Let be the active power of the b-th distributed photovoltaic power station in the k-th distributed photovoltaic sub-cluster; For the first The equivalent reactive power of a distributed photovoltaic sub-cluster; For the k-th distributed photovoltaic sub-cluster The reactive power of a distributed photovoltaic power station.

[0126] 2) Equivalent requirements for distributed photovoltaic array parameters

[0127] According to the national standard, the capacity range of distributed power sources connected to the 110kV side of the power grid is 50-100MW. The capacity and number of photovoltaic modules for Class A photovoltaic under current control mode, Class A photovoltaic under power control mode, and Class B photovoltaic modules connected to the 110kV side are set. The photovoltaic cell model is determined with reference to the photovoltaic cell model with a large proportion of photovoltaic capacity in the distribution network to be equivalent.

[0128] In the equivalent modeling of distributed photovoltaic (PV) systems, the output power and voltage of the PV array remain constant before and after equivalence. This is a crucial prerequisite for consistent external characteristics of the distributed PV cluster model. According to the power conservation principle, under the condition of constant terminal voltage, the output current of the equivalent array should be equal to the sum of the output currents of all PV power plants within the cluster. This current value can be achieved by adjusting the number of parallel PV panels in the equivalent array. The number of series and parallel connections in the distributed PV array is calculated as follows:

[0129] (6)

[0130] In the formula, These represent the number of parallel photovoltaic panels in the photovoltaic arrays before and after the equivalent values, respectively. These represent the number of photovoltaic panels connected in series in the front and rear photovoltaic arrays, respectively.

[0131] In this embodiment, the distributed photovoltaic equivalent unit adopts a photovoltaic power generation system model, including a type A photovoltaic model and a type B photovoltaic model; wherein the type A photovoltaic model represents a distributed photovoltaic system with low / high voltage ride-through capability to ensure that it can provide necessary reactive power support and maintain grid-connected operation during grid faults, and its low voltage state transition relationship is as follows: Figure 1 As shown, it is represented as:

[0132] 1→2: < ;

[0133] 2→1: > Furthermore, the new energy vehicle lacks the ability to climb hills after crossing obstacles;

[0134] 2→3: > Furthermore, the new energy source has the ability to climb slopes after crossing obstacles;

[0135] 3→4: Experiencing one calculation step;

[0136] 4→1: Climb ends (power restored to pre-crossing level);

[0137] 4→2: < .

[0138] in, This indicates the terminal voltage of the new energy source; This indicates the low-voltage ride-through (LVRT) threshold for new energy sources, typically 0.9 pu. According to the "Technical Requirements for Distributed Power Generation Grid Connection," the LVRT control logic for distributed photovoltaic systems usually employs a strict threshold-based judgment condition. In practical engineering applications, when the voltage is exactly equal to 0.9 pu, the system will maintain normal operation and will not trigger LVRT control. This avoids frequent switching of the control state due to measurement noise near the threshold.

[0139] For Type A photovoltaic models, the control methods during high and low voltage ride-through periods can be either specified current control or specified power control. The calculation formula for the specified current control method during the low voltage ride-through period is as follows:

[0140] (7)

[0141] In the formula, These are the active current setpoints and reactive current setpoints for distributed photovoltaic systems during low voltage ride-through, respectively. These are the steady-state values ​​of active current and reactive current of distributed photovoltaic systems before the fault occurs; These are the active current reference setting value and reactive current reference setting value for distributed photovoltaic power during low voltage ride-through, respectively. These are the active current coefficient 1 and the reactive current coefficient 1, respectively, representing the degree of influence of the terminal voltage on the active current and reactive current. These are the active current coefficient 2 and the reactive current coefficient 2, respectively, representing the degree of influence of the active current and reactive current before the fault. These are the voltage amplitude at the distributed renewable energy terminal and the threshold for entering the low-voltage ride-through process, respectively; the calculation formula for the specified power control mode during the low-voltage ride-through period is as follows:

[0142] The formula for calculating the specified power control method during low voltage ride-through is as follows:

[0143] (8)

[0144] In the formula, These are the active and reactive power commands during the renewable energy transit period; , These are the active and reactive power before the new energy source passes through; , These are the active power coefficient and reactive power coefficient of the underpass, respectively. , These are the active and reactive power setting values ​​for low-voltage operation, respectively.

[0145] The formula for calculating the specified current control method during high-voltage ride-through is as follows:

[0146] (9)

[0147] In the formula, These are the active current setpoints and reactive current setpoints for distributed photovoltaic systems during high voltage ride-through, respectively. These are the steady-state values ​​of active current and reactive current of distributed photovoltaic systems before the fault occurs; These are the active current reference setting value and reactive current reference setting value of distributed photovoltaic power generation during high voltage ride-through, respectively. These are related to the power level and reactive power support capability of distributed photovoltaic power generation. These are the voltage coefficients of active current and reactive current, respectively, indicating the degree of influence of terminal voltage on active current and reactive current. These are the initial current coefficients of the active current and the initial current coefficients of the reactive current, respectively, representing the degree of influence of the active and reactive currents before the fault. These are the voltage amplitude at the distributed renewable energy terminal and the threshold for entering the high-voltage ride-through process, respectively.

[0148] The formula for calculating the specified power control method during high-voltage ride-through is as follows:

[0149] (10)

[0150] In the formula, These are the active power setpoints and reactive power setpoints for distributed photovoltaic systems during high-voltage ride-through, respectively. These represent the steady-state values ​​of active power and reactive power of distributed photovoltaic power before the fault occurs. These are the active power control coefficient and the reactive power control coefficient, respectively, which represent the ratio of power during the fault to power before the fault. These are the active power reference setting value and the reactive power reference setting value, respectively.

[0151] Type B photovoltaic (PV) models represent distributed PV systems that lack low-voltage ride-through capability. These are typically early-built, small-capacity distributed PV systems or those that do not meet the latest grid connection standards. When the grid connection voltage of the PV power station is lower than the rated voltage threshold, wave blocking control is used instead of low-voltage ride-through control. Type B PV models initiate wave blocking control at time t=0. The power calculation formula during the wave blocking period is as follows:

[0152] (11)

[0153] In the formula, , These represent the active and reactive power during the wave blocking period, respectively. The active power before entering the wave blocking phase; The duration of the wave-blocking state; This refers to the recovery speed of active power.

[0154] In this embodiment, the static load equivalent unit adopts a static load model. Based on the law of load power variation with voltage, the static load model uses a polynomial model to equivalently describe the static characteristics of the load. The calculation formula is as follows:

[0155] (12)

[0156] In the formula, , , , These represent the load active power, reactive power, load bus voltage amplitude, and frequency during the system's steady-state power flow. , , , These are the actual values ​​of the load's active power, reactive power, load bus voltage amplitude, and frequency, respectively. For frequency deviation, ; , These are the frequency deviation influencing factors for active and reactive power, respectively. , , These represent the proportions of active power to total load active power under constant impedance, constant current, and constant power conditions, respectively. ; , , These represent the proportions of reactive power to total load reactive power under constant impedance, constant current, and constant power conditions, respectively. Generally, the proportions of active and reactive load types are equal, i.e. , , .

[0157] In this embodiment, the dynamic load equivalent unit adopts a dynamic load model, which in turn adopts a third-order electromechanical transient model. The calculation formula is as follows:

[0158] (13)

[0159] In the formula, These are the transient electromotive forces along the d-axis and q-axis of the induction motor, respectively. The inertial time constant; Rotor resistance; These are stator reactance, rotor reactance, and magnetizing reactance, respectively. When the stator is open-circuited, the rotor circuit time constant is... ; , These are the rotor open-circuit reactance and short-circuit reactance, respectively. , ; These are the rotor speed and the rated speed, respectively. , For slippage, , As the reference frequency, ; , These are the transient currents on the stator side. axis, Axial components; For rotor electromagnetic torque, ; For rotor mechanical torque, , , ,in The rotor rotation speed is the steady-state speed of the system. Let be the rotational speed of the system in steady state. A, B, and C represent the proportions of the load components proportional to the square of the rotational speed, the proportions of the load components proportional to the first power of the rotational speed, and the proportions of the load components independent of the rotational speed, respectively. The proportion of the induction motor in the load is denoted by . To express.

[0160] The core of equivalent modeling for dynamic loads lies in using the capacity-weighted averaging method to aggregate multiple motor models into a single equivalent model. The key step in this method is to comprehensively calculate specific parameters from the original models into single parameters in the equivalent model. These single parameters encompass all electrical, mechanical, and excitation reactance parameters in the dynamic load model that require equivalent averaging, such as the inertia time constant H and stator resistance R. s Stator leakage impedance X s And parameters such as the proportion of induction motors in the load. To avoid the redundancy of listing all parameters one by one, the following text will use a single parameter to refer to each parameter, and its comprehensive calculation formula is as follows:

[0161] (14)

[0162] In the formula, The per-unit value of a parameter based on its own capacity; This is the per-unit value of the parameter after being equalized; For the first The proportion of the rated capacity of a type of motor in the total rated capacity of all motors participating in the equivalent value.

[0163] In this embodiment, for the capacitors in the distribution network, in order to reflect their active control capability, their capacity is reduced to the 110kV voltage level and then superimposed to obtain the equivalent capacitance of the distribution network. The equivalent capacitive reactance of the capacitor equivalent unit is... for:

[0164] (15)

[0165] In the formula, Equivalent capacitive reactance of the distribution network; This represents the total number of feeders in the distribution network. The capacitance of the distribution network feeder f is calculated to the 110kV voltage level. Equivalent capacitance of the distribution network; For the complex frequency domain factor, where The imaginary unit, ω is the angular frequency.

[0166] In this embodiment, the equivalent impedance of the distribution network is calculated based on the principle of consistent network loss, and the calculation formula is as follows:

[0167] (16)

[0168] In the formula, , These are the distribution network resistance and reactance, respectively. , , These are the active power, reactive power, and voltage at PCC, respectively. , These represent the active power and reactive power at the virtual bus POC, respectively.

[0169] In this embodiment, the trajectory sensitivity analysis method compares the magnitudes of the trajectory sensitivity of the parameter to be determined by using the average trajectory sensitivity. The formula for calculating the average trajectory sensitivity is as follows:

[0170] (17)

[0171] In the formula, For average trajectory sensitivity, This is the starting time for the calculation; To calculate the duration; The formula for calculating trajectory sensitivity is:

[0172] (18)

[0173] In the formula, The number of parameters to be identified; For parameters Initial value; For the first The change in each parameter This represents the active and reactive power response curves of the detailed model at the point of common coupling (PCC) when all parameters are given values. This indicates that other parameters remain unchanged, the first The parameters are from Become The detailed model's power response curve at PCC is shown.

[0174] All parameters to be determined are assigned to their average trajectory sensitivity values. The parameters are sorted from largest to smallest to form a sensitivity sequence. This embodiment uses a proportional threshold method for partitioning: the top 30% of parameters in the sequence are selected as critical parameters, and the rest as secondary parameters. This proportional threshold can be flexibly adjusted according to the actual application requirements for model accuracy and computational efficiency. For higher accuracy requirements, the proportion of critical parameters can be appropriately increased, such as to 40%; for greater emphasis on computational efficiency, the proportion can be decreased, such as to 20%. For the critical parameters obtained from the partitioning, simulated annealing is used for optimization and identification to obtain their optimal values, which are then used as parameters in the equivalent model. For secondary parameters, a statistical synthesis method is used to obtain typical values ​​for fixation. After completing the parameter identification and construction of the equivalent model, it is embedded into the electromechanical transient analysis of a large power grid to replace the original complex detailed distribution network model, thereby significantly improving the computational efficiency of large-scale power grid safety and stability simulation while ensuring the accuracy of key dynamic characteristics.

[0175] In this embodiment, a simulated annealing algorithm is used to optimize and identify it in order to obtain its optimal value, including:

[0176] S1: Obtain the initial voltage at the common connection point PCC Initial active power Initial reactive power ;

[0177] S2: Construct the steady-state equations of the equivalent model, using the initial voltage obtained in step S1. Initial active power Initial reactive power Substitute the equations into the steady-state equations and solve them in reverse to obtain the initial and intermediate values ​​of the state variables of the equivalent model.

[0178] S3: Initial voltage As input, key parameters are assigned values ​​based on the simulated annealing algorithm to obtain the output response of the equivalent model;

[0179] S4: Solve for the response power of the photovoltaic integrated load model based on the power balance equation. The calculation formula is as follows:

[0180] (19)

[0181] In the formula, These are the equivalent impedance of the distribution network, the static load, and the active power consumed by the induction motor, respectively. Active power injected into the equivalent distribution network for the equivalent photovoltaic system; These are the equivalent impedance of the distribution network, the static load, and the reactive power consumed by the induction motor, respectively. These refer to the reactive power injected into the equivalent distribution network by the equivalent capacitor and the equivalent photovoltaic system, respectively. , These represent the active power and reactive power at PCC, respectively.

[0182] S5: The root mean square error between the power measurement values ​​of the detailed model and the power output values ​​of the equivalent model is used as the optimization objective to achieve dynamic fitting of the equivalent model to the power response of the target model. The calculation formula is as follows:

[0183] (20)

[0184] In the formula, This is the root mean square error; , The first Detailed model active power and reactive power at each sampling time; , The first The equivalent model active power and reactive power at each sampling time; This represents the total number of sampling points during the entire simulation period.

[0185] S6: Solve for the objective function value of the current parameter set, return to S3, and continue iterating until the maximum number of iterations is reached, then output the parameter set that satisfies the optimal optimization objective.

[0186] Example 2

[0187] Based on Example 1, this embodiment proposes an application based on the constructed electromechanical equivalent model of the active power distribution network.

[0188] After completing the identification and construction of the electromechanical equivalent model parameters of the active distribution network, it is embedded as a whole module into the electromechanical transient analysis of the large power grid, so as to directly replace the original detailed distribution network model with complex structure and numerous nodes. This significantly improves the computational efficiency of large-scale power grid safety and stability simulation while ensuring the accuracy of key dynamic characteristics.

[0189] Example 3

[0190] This embodiment proposes an electromechanical equivalent modeling system for active power distribution networks. The system includes: an equivalent unit construction module, a power distribution network equivalent model construction module, and a parameter identification and optimization module.

[0191] The equivalent unit building module is used to cluster elements with similar characteristics in the model using the K-means clustering algorithm;

[0192] The power distribution network equivalent model construction module is used to construct a comprehensive equivalent model structure;

[0193] The parameter identification and optimization module is used to determine the parameters of the equivalent model of the distribution network.

[0194] Example 4:

[0195] This embodiment proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of an active power distribution network electromechanical equivalent modeling method.

[0196] Example 5:

[0197] This embodiment proposes a computer device, including:

[0198] Memory, used to store computer programs;

[0199] A processor is used to execute computer programs to implement a method for electromechanical equivalent modeling of active power distribution networks.

[0200] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0201] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0202] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0203] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0204] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. An active power distribution network electromechanical equivalent modeling method, characterized in that, include: Based on the K-means clustering algorithm, similar components in the target active distribution network are grouped according to their characteristic similarity to form equivalent units; the equivalent units include distributed photovoltaic equivalent units, capacitor equivalent units, static load equivalent units, and dynamic load equivalent units; Based on the topology and power flow data of the target active distribution network and the equivalent unit, an equivalent model of the distribution network including the virtual bus POC, the equivalent impedance and the equivalent unit is constructed. The equivalent unit is connected to the virtual bus POC and connected to the common coupling point PCC through the equivalent impedance. The parameters to be determined in the equivalent model of the distribution network are obtained. Based on the trajectory sensitivity analysis method, the influence of the parameters to be determined on the external power characteristics of the model is determined. Based on the influence, all parameters to be determined are divided into key parameters and minor parameters. For the key parameters, the simulated annealing algorithm is used to optimize and identify them to obtain their optimal values, and the optimal values ​​are used as the parameters of the equivalent model of the distribution network. For the minor parameters, the statistical synthesis method is used to take typical values ​​and fix them.

2. The active power distribution network electromechanical equivalent modeling method of claim 1, wherein, The similar components include distributed photovoltaic (PV) systems. These distributed PV systems are clustered, and several distributed PV systems with similar operating characteristics are considered equivalent to typical distributed PV power plants, simplifying the modeling scale. This includes: The types of the distributed photovoltaic are divided based on the capacities of various types of photovoltaic accessed at various voltage levels of the to-be-equivalent power distribution network, including capacities of current control mode A type distributed photovoltaic, power control mode A type distributed photovoltaic and B type distributed photovoltaic, wherein the capacity access proportion of the current control mode A type distributed photovoltaic is , the capacity access proportion of the power control mode A type distributed photovoltaic is , and the capacity access proportion of the B type distributed photovoltaic is . Based on the distributed photovoltaic clustering feature vector, the current control mode A type distributed photovoltaic and the power control mode A type distributed photovoltaic with similar characteristics are divided into the same sub-cluster by using the K-means clustering algorithm, so that the dynamic characteristics of the distributed photovoltaic power stations in the sub-cluster are close, wherein the distributed photovoltaic clustering feature vector is obtained by selecting the inverter outer loop control parameters 、 to characterize the actual operation dynamic characteristics of the distributed photovoltaic, which is expressed as: (1) In the formula, is a feature vector of the distributed photovoltaic cluster; , is the integral parameter and the proportional coefficient of the outer loop PI control of the photovoltaic inverter, respectively. After K-means clustering, the photovoltaic arrays of various types are evaluated as single-unit equivalents to determine the output power, photovoltaic array parameters, inverter normal operation control parameters, and ride-through control parameters of each type of single-unit equivalent photovoltaic array.

3. The electromechanical equivalent modeling method for active power distribution networks according to claim 2, characterized in that, The K-means clustering algorithm is used to divide distributed photovoltaic power stations with similar characteristics in the current control mode A-type distributed photovoltaic system into the same sub-cluster, including: S1: Obtain a sample dataset of Class A distributed photovoltaic systems with current control mode within the target area. ,in The number of Class A distributed photovoltaic systems under current control mode. For the i-th current control mode, type A distributed photovoltaic system, i=1,2,… ; S2: The number of clusters of the current control mode A distributed photovoltaic system is determined using the elbow method. This includes: letting the cluster number k be in [2, ... Traverse the interval, calculate the sum of squared errors corresponding to different k values, and use the k corresponding to the inflection point of the sum of squared errors curve as the number of clusters of the current control mode A-type distributed photovoltaic system. Among them, the sum of squared errors The calculation formula is: (2) In the formula, The initial cluster centers of the i-th class are randomly selected; For the i-th distributed photovoltaic sub-cluster; This represents the sum of squared errors when the number of clusters is k. S3: Random selection One sample is used as the initial cluster center for the current control mode A-type distributed photovoltaic; S4: Calculate the distance between each sample index and the randomly selected initial cluster centers. The expression is: (3) In the formula, This represents the v-th attribute of the i-th distributed photovoltaic system. The v-th attribute represents the initial cluster center of the i-th class selected randomly. The dimension of the feature vector, i.e., the number of clustering indicators; S5: Based on the distance Each sample is assigned to the nearest cluster to form a sub-cluster, and the mean of the samples in each sub-cluster is used as the new cluster center. S6: Repeat S4 and S5 until the cluster centers no longer change or the maximum number of iterations is reached, and output the results of the clustering of the distributed photovoltaic cluster. Perform the same clustering process as steps S1-S6 on the power control mode A-type distributed photovoltaic system to complete the division of the power control mode A-type distributed photovoltaic subgroup.

4. The electromechanical equivalent modeling method for active power distribution networks according to claim 2, characterized in that, The process of performing single-unit equivalence on various photovoltaic systems after K-means clustering includes keeping the grid-connected output power unchanged before and after equivalence, so that the equivalent single-unit output power is equal to the sum of the output power of all distributed photovoltaic power stations in the sub-cluster; keeping the output power and output voltage of the photovoltaic array unchanged before and after equivalence, so that the output current of the equivalent array is equal to the sum of the output current of all photovoltaic power stations in the cluster, and adjusting the number of parallel photovoltaic panels in the equivalent array accordingly to complete the calculation of the number of series and parallel connections.

5. The electromechanical equivalent modeling method for active power distribution networks according to claim 1, characterized in that, The distributed photovoltaic equivalent unit adopts a photovoltaic power generation system model, including a type A photovoltaic model and a type B photovoltaic model; wherein the type A photovoltaic model represents a distributed photovoltaic system with low / high voltage ride-through capability, and its low voltage state transition relationship is as follows: When the distributed photovoltaic system is in normal condition, if the voltage at the new energy generator terminal is... If the voltage drop is below the low voltage ride-through threshold for new energy sources, the system transitions from the normal state to the ride-through state. When the distributed photovoltaic system is in a pass-through state, if the voltage at the new energy generator terminal is... If the voltage exceeds the low-voltage ride-through threshold for new energy sources and the new energy source lacks a post-ride-through ramp-up function, it transitions from the ride-through state to the normal state; if the new energy source's terminal voltage is higher than the low-voltage ride-through judgment threshold, and the new energy source lacks a post-ride-through ramp-up function, it transitions from the ride-through state to the normal state; if the new energy source's terminal voltage... If the voltage is higher than the low voltage ride-through threshold of new energy sources and the new energy source has a ramp-up function after the ride-through, the state will be transferred from the ride-through state to the ramp-up starting point. When the distributed photovoltaic system is at the starting point of the ramp-up process, it transitions to the ramp-up process after a calculation step. When the distributed photovoltaic system is in the ramp-up process, if the ramp-up ends and the power returns to the level before the ramp-up, it transitions from the ramp-up process to the normal state; if the generator terminal voltage If the voltage drop threshold for new energy low-voltage ride-through is below the threshold, the process transitions from the climbing phase to the ride-through phase. For Type A photovoltaic models, the control methods during high and low voltage ride-through periods can be either specified current control or specified power control. The calculation formula for the specified current control method during the low voltage ride-through period is as follows: (4) In the formula, These are the active current setpoints and reactive current setpoints for distributed photovoltaic systems during low voltage ride-through, respectively. These are the steady-state values ​​of active current and reactive current of distributed photovoltaic systems before the fault occurs; These are the active current reference setting value and reactive current reference setting value for distributed photovoltaic power during low voltage ride-through, respectively. These are the active current coefficient 1 and the reactive current coefficient 1, respectively, representing the degree of influence of the terminal voltage on the active current and reactive current. These are the active current coefficient 2 and the reactive current coefficient 2, respectively, representing the degree of influence of the active current and reactive current before the fault. These are the voltage amplitude at the distributed renewable energy terminal and the threshold for entering the low voltage ride-through process, respectively. The formula for calculating the specified power control method during low voltage ride-through is as follows: (5) In the formula, These are the active and reactive power commands during the renewable energy transit period; , These are the active and reactive power before the new energy source passes through; , These are the active power coefficient and reactive power coefficient of the underpass, respectively. , These are the active and reactive power setting values ​​for low-voltage operation, respectively. The formula for calculating the specified current control method during high-voltage ride-through is as follows: (6) In the formula, These are the active current setpoints and reactive current setpoints for distributed photovoltaic systems during high voltage ride-through, respectively. These are the steady-state values ​​of active current and reactive current of distributed photovoltaic systems before the fault occurs; These are the active current reference setting value and reactive current reference setting value of distributed photovoltaic power generation during high voltage ride-through, respectively. These are related to the power level and reactive power support capability of distributed photovoltaic power generation. These are the voltage coefficients of active current and reactive current, respectively, indicating the degree of influence of terminal voltage on active current and reactive current. These are the initial current coefficients of the active current and the initial current coefficients of the reactive current, respectively, representing the degree of influence of the active and reactive currents before the fault. These are the voltage amplitude at the distributed renewable energy terminal and the threshold for entering the high-voltage ride-through process, respectively. The formula for calculating the specified power control method during high-voltage ride-through is as follows: (7) In the formula, These are the active power setpoints and reactive power setpoints for distributed photovoltaic systems during high-voltage ride-through, respectively. These represent the steady-state values ​​of active power and reactive power of distributed photovoltaic power before the fault occurs. These are the active power control coefficient and the reactive power control coefficient, respectively, which represent the ratio of power during the fault to power before the fault. These are the active power reference setting value and the reactive power reference setting value, respectively. The Class B photovoltaic model represents a distributed photovoltaic system without voltage ride-through capability. When the grid connection voltage of the photovoltaic power station is lower than the rated voltage threshold, wave blocking control is used instead of low voltage ride-through control. The Class B photovoltaic model starts wave blocking control at time t=0, and the power calculation formula during the wave blocking period is: (8) In the formula, , These represent the active and reactive power during the wave blocking period, respectively. The active power before entering the wave blocking phase; The duration of the wave-blocking state; This refers to the recovery speed of active power.

6. The electromechanical equivalent modeling method for active power distribution networks according to claim 1, characterized in that, The static load equivalent unit adopts a static load model. Based on the law of load power variation with voltage, the static load model uses a polynomial model to equivalently describe the static characteristics of the load. The calculation formula is as follows: (9) In the formula, , , , These represent the load active power, reactive power, load bus voltage amplitude, and frequency during the system's steady-state power flow. , , , These are the actual values ​​of the load's active power, reactive power, load bus voltage amplitude, and frequency, respectively. For frequency deviation, ; These are the frequency deviation influencing factors for active and reactive power, respectively. , , These represent the proportions of active power to total load active power under constant impedance, constant current, and constant power conditions, respectively. ; These represent the proportions of reactive power to total load reactive power under constant impedance, constant current, and constant power conditions, respectively. .

7. The electromechanical equivalent modeling method for active power distribution networks according to claim 1, characterized in that, The dynamic load equivalent unit adopts a dynamic load model, which in turn adopts a third-order electromechanical transient model. The calculation formula is as follows: (10) In the formula, These are the transient electromotive forces along the d-axis and q-axis of the induction motor, respectively. The inertial time constant; For rotor resistance, The rotor circuit time constant when the stator is open-circuited; , These are the rotor open-circuit reactance and short-circuit reactance, respectively. These are the rotor speed and the rated speed, respectively. These are the d-axis and q-axis components of the stator-side transient current, respectively. This refers to the rotor's electromagnetic torque. This represents the rotor's mechanical torque.

8. The method for electromechanical equivalent modeling of active power distribution networks according to claim 1, characterized in that, The equivalent capacitive reactance of the capacitor equivalent unit for: (11) In the formula, Equivalent capacitive reactance of the distribution network; This represents the total number of feeders in the distribution network. The capacitance of the distribution network feeder f is calculated to the 110kV voltage level. Equivalent capacitance of the distribution network; For the complex frequency domain factor, where The imaginary unit, ω is the angular frequency.

9. The electromechanical equivalent modeling method for active power distribution networks according to claim 1, characterized in that, The equivalent impedance of the distribution network is calculated based on the principle of consistent network loss, and the calculation formula is as follows: (12) In the formula, , These are the distribution network resistance and reactance, respectively. , , These are the active power, reactive power, and voltage at PCC, respectively. , These represent the active power and reactive power at the virtual bus POC, respectively.

10. The method for electromechanical equivalent modeling of active power distribution networks according to claim 1, characterized in that, The trajectory sensitivity analysis method compares the magnitudes of the trajectory sensitivity of the parameter to be determined by using average trajectory sensitivity. The formula for calculating the average trajectory sensitivity is as follows: (13) In the formula, For average trajectory sensitivity, This is the starting time for the calculation; To calculate the duration; The formula for calculating trajectory sensitivity is: (14) In the formula, The number of parameters to be identified; For parameters Initial value; For the first The change in each parameter This represents the active and reactive power response curves of the detailed model at the point of common coupling (PCC) when all parameters are given values. This indicates that other parameters remain unchanged, the first The parameters are from Become The detailed model's power response curve at PCC is shown.

11. The method for electromechanical equivalent modeling of active power distribution networks according to claim 1, characterized in that, The step of using simulated annealing algorithm to optimize and identify it in order to obtain its optimal value includes: S1: Obtain the initial voltage at the common connection point PCC Initial active power Initial reactive power ; S2: Construct the steady-state equations of the equivalent model, using the initial voltage obtained in step S1. Initial active power Initial reactive power Substituting into the steady-state equations, the initial values ​​of the state variables and the values ​​of the intermediate variables of the equivalent model are obtained by inverse solution; S3: Set the initial voltage As input, the key parameters are assigned values ​​based on the simulated annealing algorithm to obtain the output response of the equivalent model; S4: Solve for the response power of the photovoltaic integrated load model based on the power balance equation. The calculation formula is as follows: (15) In the formula, These are the equivalent impedance of the distribution network, the static load, and the active power consumed by the induction motor, respectively. Active power injected into the equivalent distribution network for the equivalent photovoltaic system; These are the equivalent impedance of the distribution network, the static load, and the reactive power consumed by the induction motor, respectively. These refer to the reactive power injected into the equivalent distribution network by the equivalent capacitor and the equivalent photovoltaic system, respectively. , These are the active power and reactive power at PCC, respectively. S5: The root mean square error between the power measurement values ​​of the detailed model and the power output values ​​of the equivalent model is used as the optimization objective to achieve dynamic fitting of the equivalent model to the power response of the target model. The calculation formula is as follows: (16) In the formula, This is the root mean square error; , The first Detailed model active power and reactive power at each sampling time; , The first The equivalent model active power and reactive power at each sampling time; This represents the total number of sampling points during the entire simulation period. S6: Solve for the objective function value of the current parameter set, return to S3, and continue iterating until the maximum number of iterations is reached, then output the parameter set that satisfies the optimal optimization objective.

12. An electromechanical equivalent modeling system for active power distribution networks, characterized in that, The system includes: an equivalent unit construction module, a distribution network equivalent model construction module, and a parameter identification and optimization module; The equivalent unit construction module is used to cluster elements with similar characteristics in the model using the K-means clustering algorithm; The power distribution network equivalent model construction module is used to construct a comprehensive equivalent model structure; The parameter identification and optimization module is used to determine the parameters of the equivalent model of the power distribution network.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the electromechanical equivalent modeling method for active power distribution networks as described in any one of claims 1 to 11.

14. A computer device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the steps of the active power distribution network electromechanical equivalent modeling method according to any one of claims 1 to 11.

Images