An active power distribution network fault discrimination method, device and storage medium

Abstract

The application discloses an active power distribution network fault discrimination method and device and a storage medium, and relates to the technical field of electric power. The application comprises the following steps: acquiring the topological structure, system parameters, converter position and fault ride-through control strategy of an active power distribution network, building a mathematical simulation model, acquiring existing fault data and correcting the parameters of the mathematical simulation model; acquiring the transient current data of the converter and the system under fault state and non-fault state under different fault ride-through controls, and constructing the correlation relationship between the transient current peak value, trigger time, distortion rate and the power distribution network fault and fault type; learning the transient current data by using a deep feedforward neural network, and establishing a fault discrimination model; acquiring real-time operation data, inputting the real-time operation data into the fault discrimination model, and obtaining the fault discrimination diagnosis result of the active power distribution network. The application can improve the fault discrimination accuracy, and provides a feasible technical solution for realizing the fault diagnosis of the power distribution network under the background of high-proportion distributed power access.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of electric power, and particularly relates to an active power distribution network fault discrimination method, device and storage medium BACKGROUND

[0002] Distributed Generators (DGs) are connected to a power distribution network through Voltage Source Converters (VSCs, hereinafter referred to as converters), which causes the power distribution network to change from a passive network to an active network, and locally presents a high proportion of distributed generator access state. Single-phase short-circuit faults frequently occur in the power distribution network, and a fast and effective fault diagnosis method is crucial to ensure the safe and stable operation of the active power distribution network. However, the topology of the active power distribution network is variable, the DGs adopt various ride-through control, and the transient current characteristics are complex. These characteristics make it difficult for traditional fault diagnosis methods to effectively evaluate the fault state of the new power distribution network, and it is urgent to propose a fault diagnosis method suitable for a power distribution network with multiple DGs. SUMMARY

[0003] The application aims to provide an active power distribution network fault discrimination method, device and storage medium, which can effectively improve the fault discrimination accuracy of the power distribution network and provide a feasible technical solution for fault diagnosis of the power distribution network under the background of high proportion of distributed generator access.

[0004] To achieve the above-mentioned purpose, the application provides the following technical solution: an active power distribution network fault discrimination method, comprising the following steps:

[0005] Obtaining the topology structure, system parameters, virtual synchronous control converter position and fault ride-through control strategy of the active power distribution network, building a mathematical simulation model, obtaining existing fault data and correcting the parameters of the mathematical simulation model;

[0006] Based on the mathematical simulation model, obtaining the transient current data of the converter and the system under fault state and non-fault state under different fault ride-through control, and constructing the correlation relationship between the transient current peak value, trigger time, distortion rate and the power distribution network fault and fault type;

[0007] Using a deep feedforward neural network to learn the transient current data, and establishing a fault discrimination model of the active power distribution network;

[0008] Obtaining real-time operation data and inputting the real-time operation data into the fault discrimination model to obtain the fault discrimination diagnosis result of the active power distribution network.

[0009] Further, the transient characteristics of the virtual synchronous control converter are described as follows:

[0010]

[0011] where T0 and T em represent the reference torque and electromagnetic torque, J and D p represent the virtual inertia and virtual damping coefficient, respectively, and ω0 is the rated angular frequency.

[0012] Further, the fault ride-through control includes a converter virtual impedance current limiting control, a converter mode switching current limiting control, and a converter current limiter passive current limiting.

[0013] Further, the expression of the converter output current under the virtual impedance current limiting control is as follows:

[0014] i abc = (V abc - V g ) / (Z eq + Z v )

[0015] where i abc is the converter output current, V abc and V g are the internal voltage of the converter and the grid voltage, Z eq represents the equivalent impedance between the converter and the grid, and Z v is the added virtual impedance value.

[0016] Further, the converter output current under the mode switching current limiting control can be determined by the reference current value, as shown below:

[0017]

[0018] where I*d and I*q are the dq-axis reference output currents, which can be set by the grid code, I abc and δ are the current amplitude and phase angle, respectively, and I dref and I qref are the actual output current reference values of the converter.

[0019] Further, under the current limiter passive current limiting, the converter output current can be determined by the limiting value set by the current limiter, as shown below:

[0020]

[0021]

[0022] where I max and I ref are the maximum output current amplitude and the current limiter reference current value, respectively, and I*dref and I*qref represent the active current reference value and the reactive current reference value.

[0023] Further, the deep feedforward neural network comprises an input layer, a hidden layer and an output layer.

[0024] According to another aspect of the present application, the present application provides an apparatus comprising one or more processors, memory for storing one or more programs;

[0025] When the one or more programs are executed by the one or more processors, the one or more processors perform the active power distribution network fault discrimination method.

[0026] According to another aspect of the present application, the present application provides a storage medium, wherein a computer program is stored in the storage medium, and the computer program is run to perform the active power distribution network fault discrimination method.

[0027] The present application has at least the following advantages:

[0028] Firstly, the correlation between the transient current of the converter and the system level and the fault type and the fault location is established to obtain the transient current data with strong representation ability, then the deep feedforward neural network is used to learn and train the transient current data to obtain the active power distribution network fault diagnosis network model, finally the measured transient current data is input into the fault diagnosis network model to realize the fault diagnosis of the power distribution network, which can effectively improve the fault discrimination accuracy and provide a feasible technical solution for realizing the fault diagnosis of the power distribution network under the background of high proportion of distributed power access.

[0029] Of course, implementing any product of the present application does not necessarily require achieving all the advantages described above at the same time. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 It is an IEEE33 node DG containing power distribution network topology structure diagram;

[0031] Figure 2 It is a DG topology structure and control block diagram under different fault ride-through control;

[0032] Figure 3 It is a transient equivalent circuit diagram of DG under different fault ride-through control;

[0033] Figure 4 It is a transient current numerical simulation result diagram of DG under different fault ride-through control;

[0034] Figure 5 It is a feedforward neural network learning algorithm diagram;

[0035] Figure 6 It is an embodiment step flow chart of the present application. DETAILED DESCRIPTION

[0036] With reference to the drawings of the embodiments of the present disclosure, the technical solutions in the embodiments of the present disclosure will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present disclosure.

[0037] For the convenience of understanding, the present disclosure is described taking the IEEE 33-node distribution network with DG as an object, and the topology structure diagram of the IEEE 33-node distribution network with DG is as shown in Figure 1 , wherein the solid line represents the actual connection line and the network, and the dashed line represents the reconfigurable / connection line and the network. The distributed power sources DG1 and DG2 are connected to node 13 and node 29 respectively, the distributed power source is connected to the distribution network through a DC / AC converter, and the topology structure is as shown in Figure 2 , wherein the DC micro source includes but is not limited to energy storage batteries, photovoltaic panels and other DC form power sources. The DC power input is converted into AC power through a DC / AC converter and connected to a common bus. The output end of the DC / AC converter is connected to an LC filter inductance and capacitance. The control system of the DC / AC converter mainly includes a phase-locked loop synchronization module / power synchronization control template, a voltage / current control module and a pulse width modulation / drive module. The control phase angle is obtained by collecting the converter port voltage and inputting the phase-locked loop synchronization module / power synchronization control module. The voltage / current control module controls the output current amplitude of the converter, and the given current reference value is realized through PI control loop to realize zero-error tracking. The pulse width modulation and / or drive module generates a modulation signal to drive the IGBT and other switching devices.

[0038] The active distribution network mainly includes a converter, and there are many types of converters. The present application mainly studies a virtual synchronous control converter (hereinafter referred to as a converter) and proposes a fault discrimination method for an active distribution network.

[0039] Please refer to Figures 1-6 , the present application provides a technical solution: a fault discrimination method for an active distribution network, which is as follows:

[0040] 1. Obtain the topology structure, system parameters, virtual synchronous control converter position and fault ride-through control strategy of the active distribution network, and establish a system mathematical model in PSCAD / EMTDC. The control structure of the virtual synchronous control converter is as shown in Figure 2 , and the transient characteristics can be described by the following formula:

[0041]

[0042] , wherein T0 and T em represent the reference torque and the electromagnetic torque, J and D pω0is the rated angular frequency, respectively represent virtual inertia and virtual damping coefficient;

[0043] Take three typical fault ride-through control as an example, the mathematical model is established respectively:

[0044] (1) Virtual impedance current limiting control

[0045] As shown in Figure 2 , when the switch S1-S3 is in 1, the converter adopts virtual impedance current limiting control, the virtual impedance current limiting control adds additional virtual impedance Z v , so as to reduce the reference voltage value of the converter, realize fault current limiting, at this time, the output current of the converter can be calculated by the following formula:

[0046] i abc =(V abc -V g ) / (Z eq +Z v ) (2)

[0047] Where i abc is the output current of the converter, V abc and V g are the internal potential and grid voltage of the converter, Z eq represents the equivalent impedance between the converter and the grid, and Z v is the added virtual impedance value. By observation, after adding the virtual impedance, the virtual synchronous control converter will remain in the controlled voltage source state, and the added virtual impedance value will have a non-negligible impact on the transient output current.

[0048] When the switch S1-S3 is in 2, the converter adopts mode switching current limiting control, which switches the converter from voltage control mode to current control mode after detecting that the output current exceeds the threshold value, directly controls the transient output current, so as to realize the suppression of fault current, at this time, the output current of the converter can be determined by the reference current value, as shown in the following formula:

[0049]

[0050] Where I*d and I*q are the dq-axis reference output currents, which can be set by the grid rule, I abc and δ are the current amplitude and phase angle, respectively, and I dref and I qref are the actual output current reference values of the converter.

[0051] When the switches S1-S3 are at 3, the current limiter is used to limit the current passively, and when the output current exceeds the threshold value set by the limiter, the converter will be passively switched from the direct control voltage mode to the direct control output current mode, so as to limit the fault output current, at this time, the output current of the converter can be determined by the amplitude value set by the limiter, as shown in the following formula:

[0052]

[0053] Wherein I max and I ref are the maximum output current amplitude and the current limiter reference current value respectively, and I*dref and I*qref represent the active current reference value and the reactive current reference value. The selection of the current reference value can be set according to the characteristics of the converter itself.

[0054] On this basis, the mathematical simulation model is simulated under various fault conditions, and the historical measured data is used to correct and verify the simulation model and parameters, so as to ensure that the simulation data of the established mathematical simulation model under specific conditions is highly consistent with the historical measured data, and the established mathematical simulation model can accurately reflect the operating characteristics of the actual distribution network.

[0055] 2. According to the established mathematical simulation model, different DG ride-through control strategies determine the transient current characteristics of the converter, so in the present disclosure, the correlation between the transient current data of the converter and the system and the fault characteristics of the distribution network is established, so as to obtain transient current data with strong representation ability.

[0056] When the switch is at 1, the equivalent circuit diagram of the virtual synchronous control converter is as shown in Figure 3 , which can be equivalent to a controlled voltage source form, at this time, according to the basic circuit theorem, the transient current of the converter is mainly composed of the alternating current periodic component i p and the direct current decay component i ap , as shown in the following formula:

[0057]

[0058]

[0059] Wherein i p is the alternating current periodic component, i ap is the direct current decay component, E, V g and V’g represent the voltage phase of the converter, the voltage phase of the grid in normal state and the voltage phase of the grid during the fault, Z eq and Z’eq are the line equivalent impedance before the fault occurs and during the fault, T a = L’ / R’, T ais the decay time constant, L' and R' are the reactance and resistance, respectively.

[0060] When virtual impedance current limiting control is adopted, the converter still keeps voltage control mode, which can be equivalent to a controlled voltage source as shown in Fig. 3(a). At this time, the increased virtual impedance can effectively suppress the AC periodic component, but has little effect on the DC decay component as shown in Fig. 3(b). Figure 3 Figure 4 When mode switching current limiting control is adopted, the converter is actively switched from voltage control mode to current control mode, which can be equivalent to a controlled current source as shown in Fig. 3(b). At this time, the fault current is directly given by the reference current value of the converter as shown in Fig. 3(c). It is worth noting that these two controls depend on the fault detection signal, and the fault detection delay has a non-negligible effect on the transient current characteristics. Figure 4 Figure 3 When limiter passive current limiting control is adopted, the converter is passively switched from voltage control mode to current control mode, which can be equivalent to a controlled current source as shown in Fig. 3(c). At this time, the fault current is directly given by the limiter amplitude value of the converter, but the control switching process will bring obvious high-frequency harmonic signals as shown in Fig. 3(d). Limiter passive current limiting control does not require detection of fault signals, so the detection delay has little effect on the transient output current. Figure 4

[0061] The transient current characteristics of the converter under different ride-through controls can be summarized as shown in Table 1 below:

[0062] Table 1 Transient current characteristics of the converter under different ride-through controls

[0063]

[0064] According to the above summary, taking the transient current peak value, trigger time, and distortion rate as the main strong characteristic information, and based on the simulation model established in step 1, the strong characteristic information vector is obtained by simulating different nodes (nodes 4, 9, 19, 23, etc.), different fault depths (drop depth 90%, 70%), and different fault types (single-phase, three-phase, two-phase grounding, etc.), which is used as the source of learning data for the deep learning network.

[0065] 3. According to the simulation data obtained in step 2, a deep feedforward neural network is used for learning, and the neural network structure is as shown in Fig. 4, wherein the input layer inputs the strong characteristic information vector, and the output layer gives the results such as whether a fault occurs and the fault type. After multiple iterations and calculations, a distribution network fault diagnosis model considering different fault ride-through controls of DG is formed, which is used as the network model for determining faults. Figure 5

[0066] ​​​​4. The fault diagnosis model obtained according to step 3, real-time input of the measured data of the power distribution network, obtaining a fault diagnosis decision, on the basis of which, multiple experiments are carried out and summarized as shown in Table 2. As can be seen from Table 2, the fault diagnosis accuracy of the present disclosure is 100%, and the fault type discrimination accuracy reaches 98.1%, verifying the effectiveness of the present disclosure.

[0067] Table 2 Fault diagnosis accuracy of the power distribution network based on the deep learning algorithm

[0068]

[0069] According to another aspect of the present application, the present application provides an apparatus comprising one or more processors, a memory for storing one or more programs;

[0070] When the one or more programs are executed by the one or more processors, the one or more processors execute the active power distribution network fault discrimination method.

[0071] According to another aspect of the present application, the present application provides a storage medium, in which a computer program is stored, and the computer program is executed to execute the active power distribution network fault discrimination method.

[0072] It should be noted that, in this document, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0073] For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances. When an element is referred to as "assembled", "mounted", "fixed" or "disposed" on another element, it can be directly on the other element or there can be a middle element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there can be a middle element. The terms "vertical", "horizontal", "up", "down", "left", "right" and similar expressions used herein are only for illustrative purposes and are not the only implementation.

[0074] While embodiments of the application have been shown and described, it is to be understood that the application is not limited to the details of the embodiments described, since numerous changes, modifications, substitutions, and alterations can be made thereto without departing from the spirit and scope of the application as defined in the appended claims and their equivalents.

[0075] In the description of the specification, reference to "one embodiment", "an example", "a specific example", and so on, means that a particular feature, structure, material, or characteristic being described is included in at least one embodiment or example of the disclosure. The appearances of the above expressions in various places in the specification are not necessarily referred to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

Claims

1. A method of active power distribution network fault discrimination, characterized by, The method comprises the following steps: obtaining the topology structure, system parameters, virtual synchronous control converter position and fault ride-through control strategy of the active power distribution network, building a mathematical simulation model, obtaining existing fault data and correcting the parameters of the mathematical simulation model; based on the mathematical simulation model, obtaining the transient current data of the converter and the system in the fault state and the non-fault state under different fault ride-through controls, and constructing the correlation relationship between the transient current peak value, trigger time, distortion rate and the power distribution network fault and fault type; using a deep feedforward neural network to learn the transient current data and establishing a fault discrimination model of the active power distribution network; obtaining real-time operation data and inputting the real-time operation data into the fault discrimination model to obtain the fault discrimination and diagnosis result of the active power distribution network; the fault ride-through control comprises virtual impedance current limiting control of the converter, mode switching current limiting control of the converter and passive current limiting control of the current limiter of the converter; the expression of the converter output current under the virtual impedance current limiting control is as follows: i abc = (V abc - V g ) / (Z eq + Z v ) where V abc and V g are the internal potential of the converter and the grid voltage, Z eq represents the equivalent impedance between the converter and the grid, Z v is an added virtual impedance value; the converter output current under the mode switching current limiting control can be determined by the reference current value, and is as follows: where I*dand I*qare the dq-axis reference output currents, which can be set by grid codes, I abc and δ are the current amplitude and phase angle, respectively; under the passive current limiting control of the current limiter, the converter output current can be determined by the set limiting value of the limiter, and is as follows: where I max and I ref are the maximum output current amplitude and the current limiter reference current value, respectively, and I*dref and I*qref represent the active and reactive current reference values.

2. The method of claim 1, wherein, the transient characteristics of the virtual synchronous control converter are described as follows: where T0and T em represent the reference torque and the electromagnetic torque, J and D p represent the virtual inertia and the virtual damping coefficient, respectively, and ω0is the rated angular frequency.

3. The method of claim 2, wherein: when the virtual synchronous control converter is equivalent to a controlled voltage source, the transient current of the virtual synchronous control converter is represented as follows: where i p is the ac periodic component, i ap is the dc decaying component, E, V g and V'g represent the converter voltage phasor, the grid voltage phasor in normal conditions and the grid voltage phasor during fault, Z eq and Z'eq are the line equivalent impedances before and during fault, T a = L ’ / R ’ , T a is the decaying time constant, L ’ and R ’ are the reactance and resistance respectively.

4. The method of claim 3, wherein: the deep feedforward neural network comprises an input layer, a hidden layer and an output layer.

5. An apparatus, comprising: comprise: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors execute the active power distribution network fault discrimination method in any one of claims 1 to 4.

6. A storage medium having stored therein a computer program, characterized in that, running the computer program can execute the active power distribution network fault discrimination method in any one of claims 1 to 4.

Images