Distributed power supply series-parallel power distribution network fault identification model construction method and system

Abstract

The invention provides a distributed power supply series-parallel power distribution network fault identification model construction method and system, and the method comprises the steps: constructing a series-parallel simulation power distribution network based on preset power distribution network simulation information, and constructing a distributed energy simulation system based on preset distributed energy structure information; connecting the distributed energy simulation system to the series-parallel simulation power distribution network, performing single-phase earth fault simulation on the series-parallel simulation power distribution network by using the arc model, and collecting a fault transient current simulation signal of the series-parallel simulation power distribution network; performing feature extraction on the fault transient current simulation signal by combining a wavelet transform method to obtain a fault signal training feature, determining a fault recognition result based on the fault signal training feature, and further taking the fault signal training feature as an input and the fault recognition result corresponding to the fault signal training feature as an output to obtain a fault recognition result; and training the fault identification model to obtain a trained fault identification model, thereby realizing accurate identification of the fault of the hybrid power distribution network containing the distributed power supply through the fault identification model.

Description

Technical Field

[0001] The present invention relates to the field of power distribution networks, and in particular to a method and system for constructing a fault identification model for a hybrid power distribution network. Background Art

[0002] Currently, traditional overhead distribution lines in urban medium-voltage power transmission and distribution projects are gradually being replaced by hybrid systems combining overhead and cable lines. However, medium-voltage distribution networks suffer from high failure rates and inaccurate fault location. According to statistics, the most common fault type in distribution systems is single-phase grounding faults, accounting for over 80%, and most are accompanied by arcing. Prolonged grounding faults can damage equipment in the power network, impacting the stable operation of the power system and even potentially causing a cascade of faults, resulting in widespread power outages. Therefore, when a system fault occurs, accurately identifying the type of single-phase grounding fault is essential for troubleshooting potential problems. Fault identification methods in power systems include threshold discrimination, transition resistance, and neural networks.

[0003] The threshold discrimination method mainly identifies the fault type by comparing the size of the characteristic value threshold. However, this method has uncertainty. When the model topology or fault conditions change, the selected feature quantity and feature threshold must also change accordingly. The transition resistance method is to reversely construct an equivalent model when a certain ground fault occurs. The size and linear properties of the fault transition resistance can be determined based on the calculated ground resistance characteristic curve. However, the transition state mechanism of a single-phase ground fault is complex, and its characteristic curve is affected by multiple factors such as the fault phase angle, transition resistance size, and fault location. Through training, neural networks can construct a complex relationship between input and output, and ultimately achieve prediction of unknown samples. They have good fault tolerance and have been applied and developed in many fields.

[0004] Furthermore, with the growing demand for energy, traditional fossil fuel projects are no longer able to meet societal needs. Consequently, new energy sources are being vigorously developed, with an increasing number of distributed power sources being connected to distribution networks, and traditional distribution networks are gradually undergoing transformation. Photovoltaic power generation systems and battery energy storage systems, in particular, are widely used worldwide due to their unique advantages. Compared with other renewable energy generation technologies, photovoltaic power generation offers low pollution, requires minimal land occupation, has low investment costs, is not geographically restricted, offers high power reliability, and is easy to install, making it widely used in the power generation industry. Batteries, on the other hand, can achieve power balance within the grid by absorbing or releasing electricity. However, the integration of distributed power sources increases the complexity of distribution networks and affects operating parameters such as the direction of power flow and the magnitude of short-circuit currents. These factors can pose greater challenges to fault identification.

[0005] Therefore, how to accurately identify faults in distribution networks containing distributed power systems is of great significance. Summary of the Invention

[0006] In order to solve the problem that the existing technology cannot accurately identify the fault type of the distribution network containing a distributed power system, the present invention proposes a method and system for constructing a fault identification model for a hybrid distribution network with distributed power sources.

[0007] In a first aspect, a method for constructing a fault identification model for a hybrid distribution network of distributed power sources is provided, comprising:

[0008] Build a hybrid simulation distribution network based on the preset distribution network simulation information, and build a distributed energy simulation system based on the preset distributed energy structure information;

[0009] Connecting the distributed energy simulation system to the hybrid simulation distribution network, performing single-phase grounding fault simulation on the hybrid simulation distribution network using an arc model, and collecting fault transient current simulation signals of the hybrid simulation distribution network;

[0010] Extracting features of the fault transient current simulation signal using a wavelet transform method to obtain a fault signal training feature, and determining a fault identification result based on the fault signal training feature;

[0011] The fault recognition model is trained with the fault signal training feature as input and the fault recognition result corresponding to the fault signal training feature as output to obtain a trained fault recognition model.

[0012] In a second aspect, the present invention provides a system for constructing a fault identification model for a hybrid distribution network of distributed power sources, comprising:

[0013] A construction module is used to build a hybrid simulation distribution network based on preset distribution network simulation information, and to build a distributed energy simulation system based on preset distributed energy structure information;

[0014] A simulation module, configured to connect the distributed energy simulation system to the hybrid simulation distribution network, perform single-phase grounding fault simulation on the hybrid simulation distribution network using an arc model, and collect fault transient current simulation signals of the hybrid simulation distribution network;

[0015] an extraction module, configured to extract features of the fault transient current simulation signal in combination with a wavelet transformation method to obtain a fault signal training feature, and determine a fault identification result based on the fault signal training feature;

[0016] The training module is used to train the fault recognition model with the fault signal training feature as input and the fault recognition result corresponding to the fault signal training feature as output to obtain the trained fault recognition model.

[0017] In a third aspect, the present invention provides a method for identifying faults in a hybrid distribution network of distributed power sources, comprising:

[0018] Acquiring a fault transient current signal of a target hybrid distribution network, wherein the target hybrid distribution network is a distribution network connected to a distributed energy system;

[0019] Extracting the characteristics of the fault transient current signal using a wavelet transform method to obtain the fault signal characteristics;

[0020] Inputting the fault signal characteristics into a pre-trained fault identification model to obtain a fault identification result corresponding to the target hybrid distribution network;

[0021] The fault identification model is constructed by the aforementioned method for constructing a fault identification model of a distributed power hybrid distribution network.

[0022] In a fourth aspect, the present invention further provides a distributed power hybrid distribution network fault identification system, comprising:

[0023] an acquisition module, configured to acquire a fault transient current signal of a target hybrid distribution network, wherein the target hybrid distribution network is a distribution network connected to a distributed energy system;

[0024] A feature extraction module, configured to extract features of the fault transient current signal in combination with a wavelet transform method to obtain fault signal features;

[0025] An input module, configured to input the fault signal characteristics into a pre-trained fault identification model to obtain a fault identification result corresponding to the target hybrid distribution network;

[0026] The fault identification model is constructed by the aforementioned method for constructing a fault identification model of a distributed power hybrid distribution network.

[0027] In another aspect, the present application further provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0028] The memory is used to store one or more programs;

[0029] When the one or more programs are executed by the at least one processor, a method for constructing a distributed power hybrid distribution network fault identification model and a method for identifying a distributed power hybrid distribution network fault as described above are implemented.

[0030] On the other hand, the present application also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed, it implements a method for constructing a distributed power hybrid distribution network fault identification model and a method for identifying a distributed power hybrid distribution network fault as described above.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] The present invention provides a method and system for constructing a fault identification model of a hybrid distribution network of distributed power sources. The method builds a hybrid simulation distribution network based on preset distribution network simulation information, and constructs a distributed energy simulation system based on preset distributed energy structure information. The distributed energy simulation system is then connected to the hybrid simulation distribution network, and a single-phase grounding fault simulation is performed on the hybrid simulation distribution network using an arc model. The fault transient current simulation signal of the hybrid simulation distribution network is collected, and a wavelet transformation method is used to extract features of the fault transient current simulation signal to obtain a fault signal training feature. The fault identification result is determined based on the fault signal training feature. The fault identification model is then trained with the fault signal training feature as input and the fault identification result corresponding to the fault signal training feature as output to obtain a trained fault identification model, thereby realizing accurate identification of hybrid distribution network faults containing distributed power sources through the fault identification model. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a flow chart of the method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0034] Figure 2 This is a specific flow chart of the method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0035] Figure 3 A schematic diagram of a distribution network simulation for the method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0036] Figure 4 A schematic diagram of a photovoltaic power generation system for the method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0037] Figure 5 A schematic diagram of an energy storage system for the method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0038] Figure 6 A schematic diagram of a battery constant current charging control strategy for a method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0039] Figure 7A schematic diagram of a battery constant voltage charging control strategy for a method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0040] Figure 8 A schematic diagram of a battery constant power discharge control strategy for a method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0041] Figure 9 A schematic diagram of a three-phase inverter control system for the method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0042] Figure 10 A PQ control schematic diagram of the method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0043] Figure 11 A schematic diagram of a transient waveform of a fault current in a method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0044] Figure 12 A schematic diagram of a low-resistance ground fault current in the method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0045] Figure 13 A schematic diagram of a high-resistance ground fault current in a method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0046] Figure 14 A schematic diagram of a continuous arc grounding fault current in the method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0047] Figure 15 A schematic diagram of an indirect arc grounding fault current in the method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0048] Figure 16 Schematic diagram of time-frequency wavelet transform of different ground faults in the method for constructing a fault identification model of a hybrid distribution network of distributed power sources of the present invention;

[0049] Figure 17 A schematic diagram of a neural network training process for a method for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0050] Figure 18 A schematic diagram of the system structure for constructing a fault identification model for a hybrid distribution network of distributed power sources according to the present invention;

[0051] Figure 19 This is a flow chart of the method for identifying faults in a hybrid distribution network of distributed power sources according to the present invention;

[0052] Figure 20This is a schematic diagram of the structure of the distributed power hybrid distribution network fault identification system of the present invention;

[0053] Figure 21 The figure is a schematic structural diagram of an electronic device of the present invention. DETAILED DESCRIPTION

[0054] The present invention proposes a method and system for constructing a fault identification model for a hybrid distribution network of distributed power sources. An overhead-cable hybrid distribution network system is constructed using EMTP software, and different distributed power supply modules are constructed. The distributed power sources are connected to a common node of the distribution network system via a three-phase active inverter and an LC filter, thereby simulating fault characteristic waveforms under different types of ground faults. Transient zero-sequence current characteristics under different faults are collected and extracted, and accurate fault identification in the distributed power system is achieved by training a neural network on the extracted fault characteristics.

[0055] In order to better understand the present invention, the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0056] Example 1:

[0057] A method for constructing a fault identification model for a hybrid distribution network with distributed generation, such as Figure 1 As shown, including:

[0058] Step 1: Build a hybrid simulation distribution network based on the preset distribution network simulation information, and build a distributed energy simulation system based on the preset distributed energy structure information;

[0059] Step 2: Connecting the distributed energy simulation system to the hybrid simulation distribution network, performing single-phase grounding fault simulation on the hybrid simulation distribution network using an arc model, and collecting fault transient current simulation signals of the hybrid simulation distribution network;

[0060] Step 3: extracting features of the fault transient current simulation signal using a wavelet transform method to obtain fault signal training features, and determining a fault identification result based on the fault signal training features;

[0061] Step 4: Using the fault signal training features as input and the fault recognition results corresponding to the fault signal training features as output, the fault recognition model is trained to obtain a trained fault recognition model.

[0062] In this embodiment, in the process of building a hybrid simulated distribution network based on the preset distribution network simulation information in step 1, a single-phase continuous arc grounding fault simulation is performed by selecting a typical 10kV single radial distribution network simulation line; preferably, the three-core cable can be modeled according to the geometric dimensions and electrical parameters of the YJV22-3*400 model three-core armored cable, and a grounding transformer is used as the neutral point to adjust the grounding method of the neutral point.

[0063] In this embodiment, in step 1, the process of constructing a distributed energy simulation system based on preset distributed energy structure information includes:

[0064] A photovoltaic power generation system including a photovoltaic array and a DC-DC boost module is constructed based on preset distributed energy structure information, and a distributed energy control strategy of the photovoltaic power generation system is set to apply a disturbance to the photovoltaic power generation system using a disturbance observation method;

[0065] Building a battery system including a battery and a bidirectional DC-DC conversion module based on preset distributed energy structure information, and setting the distributed energy control strategy of the battery system to adopt a constant current and constant voltage charging strategy and a constant power discharging strategy;

[0066] Building an inverter grid including a three-phase voltage source inverter and a filter circuit based on preset distributed energy structure information, and setting the distributed energy control strategy of the inverter grid to adopt a constant power control strategy;

[0067] A distributed energy system is constructed based on the photovoltaic power generation system, the battery system and the inverter grid.

[0068] Specifically, a photovoltaic power generation system is constructed, including a photovoltaic array and a DC / DC boost module, and an MPPT algorithm (i.e., the aforementioned perturbation observation method) is used to obtain a pulse signal for controlling the switching of a DC / DC direct current chopper circuit to ensure that the photovoltaic system always maintains maximum power generation; a battery energy storage system is constructed, including a battery and a bidirectional DC / DC conversion module to realize the charging and discharging process of the battery, and the charging and discharging conversion is realized by controlling the switching state of the DC / DC circuit; further, in order to make full use of energy, constant current charging is used at the beginning of battery charging, and when the constant current charging makes the battery capacity reach a certain value, it is converted to constant voltage charging; and for the battery discharge process, a constant power discharge mode is used; further, an inverter and a filter circuit are constructed to realize the conversion of direct current to alternating current. This method selects a voltage-active inverter as the system inverter circuit and adopts a constant power control strategy to keep the output power of the distributed power supply consistent with the reference power.

[0069] In this embodiment, the constant current and constant voltage charging strategy includes: when the power of the battery system meets the preset charging conditions, extracting the difference between the actual working current of the battery in the battery system and the given battery charging current, and performing proportional integral control on the output switch duty cycle based on the difference, performing pulse width modulation on the switch duty cycle, and generating a control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC conversion module; when the real-time collected battery voltage value meets the preset voltage rating, converting the current constant current charging mode to a constant voltage charging mode.

[0070] Among them, the constant voltage charging method adopts a dual closed-loop control structure. The outer loop controls the difference between the actual battery voltage value and the set battery voltage value through proportional-integral control to output the current reference signal, and the inner loop controls the difference between the current reference signal and the actual battery charging current through proportional-integral control to output the switch duty cycle. The switch duty cycle is pulse-width modulated to generate the control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC conversion module.

[0071] In this embodiment, the constant power discharge method includes: when the battery system power meets the preset discharge conditions, a dual closed-loop control structure is adopted, wherein the outer loop controls the difference between the actual battery voltage and the set battery voltage through proportional-integral control to output a current reference signal; and the inner loop controls the difference between the current reference signal and the actual battery charging current through proportional-integral control to output a switch duty cycle, and pulse-width modulates the switch duty cycle to generate a control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC converter module.

[0072] In this embodiment, the constant power control strategy includes: converting the voltage and current in the inverter grid from the ABC coordinate system to the DQ coordinate system through the ABC / DQ transformation matrix to obtain the actual value of active power and the actual value of reactive power; controlling the difference between the active power reference value and the actual value of active power and the difference between the reactive power reference value and the actual value of reactive power through proportional integral control, and outputting the D-axis current component reference value and the Q-axis current component reference value respectively; obtaining the D-axis voltage reference component and the Q-axis voltage reference component based on the D-axis current component reference value and the Q-axis current component reference value by using the current inner loop control; inversely transforming the D-axis voltage reference component and the Q-axis voltage reference component to the ABC coordinate system, and obtaining the inverter control signal through pulse width modulation, wherein the inverter control signal is used to control the three-phase voltage source inverter to output constant power.

[0073] In this embodiment, after the hybrid simulated distribution network and the distributed energy simulation system are completed in step 1, the distributed energy simulation system can be connected to the hybrid simulated distribution network to perform single-phase grounding fault simulation on the hybrid simulated distribution network using an arc model, and collect the fault transient current simulation signal of the hybrid simulated distribution network, which specifically includes:

[0074] The distributed energy simulation system is connected to different locations of the hybrid simulation distribution network, and the Mayer arc model is used to simulate a single-phase grounding fault on the hybrid simulation distribution network connected to the distributed energy simulation system; and fault transient current simulation signals are collected at different access locations and under different operating conditions of the hybrid simulation distribution network.

[0075] The different positions of the hybrid simulated distribution network include but are not limited to the front end position, the adjacent feeder end position and / or the distribution network end position of the hybrid simulated distribution network.

[0076] Specifically, a typical arc model is used to simulate a single-phase continuous arc grounding fault. Existing arc models generally view arc behavior with a black box concept. Preferably, the Mayr "black box" model can be used to simulate the arc, which meets the accuracy requirements of simulating arc faults.

[0077] In this embodiment, in step 3, the process of extracting features from the fault transient current simulation signal by combining the wavelet transformation method to obtain the fault signal training features includes:

[0078] performing standardization processing on the fault transient current simulation signal to obtain a processed fault transient current training signal;

[0079] Performing continuous wavelet transform on the processed fault transient current signal based on the db6 wavelet basis function to obtain local characteristics of the signal;

[0080] After scaling and translating the local features of the signal, a fault signal training feature including a two-dimensional time-frequency graph is obtained.

[0081] Example 2:

[0082] See also Figure 2 , specific embodiments of the present invention include:

[0083] Step 1: Build a typical 10kV single-radiation hybrid distribution network simulation circuit;

[0084] The rated parameters of the main transformer of the designed simulation line are 110kV / 20MVA, and it includes three different types of feeders. Figure 3The loads on these feeders are: 10 km of overhead lines (e.g., L1, L2, L3, L4, and L5), 20 km of cable lines, and 20 km of hybrid overhead and cable lines. The loads on all feeders are 0.92 MW in active power and 0.392 MVAR in reactive power. The three-core armored cable model is based on the geometric dimensions and electrical parameters of the YJV22-3*400 model, as shown in Table 1.

[0085] Table 1 Dimensional parameters of three-core cables

[0086]

[0087]

[0088] Step 2: Use Mayr arc model to simulate single-phase continuous arc grounding fault;

[0089] The simulation uses a Mayr arc connected in series with a constant resistor to simulate an arc-induced ground fault. The Mayr model assumes a constant arc column diameter and a gradually decreasing arc temperature along the radial direction. Energy dissipation relies on heat conduction and radial diffusion, and the dissipated energy is assumed to be constant. The arc heat balance equation based on the Mayr model is as follows:

[0090]

[0091] Among them, g M is the arc conductance, τ M is the Mayr arc time constant, P0 is the arc power constant, u arc is the arc voltage, i arc is the arc current, Represents the conductivity g M The derivative of the Mayr arc element with respect to time t. The essence of the Mayr arc element is composed of arc voltage, arc current, and time constant τ m The arc behavior during an arc-to-ground fault is described by controlling the arc variable resistor using four inputs: the power dissipation P0 and the time logic module. By adjusting and changing the appropriate arc parameters, a typical arc behavior can be generated during a fault.

[0092] Step 3: Build a photovoltaic power generation system;

[0093] Including photovoltaic array, DC / DC boost module, see Figure 4 , the MPPT algorithm is used to obtain the pulse signal that controls the switching of the DC / DC chopper circuit switch to ensure that the photovoltaic system always maintains the maximum power generation.

[0094] The photovoltaic system control process is as follows:

[0095] S301, build photovoltaic array and DC / DC boost module;

[0096] S302, using the MPPT algorithm to control the switching of the DC / DC chopper circuit switch to adjust the maximum output power of the photovoltaic system;

[0097] This method uses the perturbation observation method to achieve the maximum power tracking of photovoltaic cells. The voltage of the photovoltaic cell is disturbed, and the disturbance variation is a constant voltage △U. The power P output by the photovoltaic power generation module is collected in real time. k and the power P at the previous moment (k-1) By comparison, we can clearly identify the direction of the applied disturbance voltage and adjust the voltage change direction by adjusting the duty cycle of the IGBT switch cut-off signal in the DC / DC circuit:

[0098]

[0099] Among them, T on is the switch on time, T off is the switch closing time, D is the duty cycle, if P k >P (k-1) , then the working voltage of the photovoltaic array changes in the opposite direction.

[0100] Step 4: Build a battery energy storage system;

[0101] Including batteries and bidirectional DC / DC conversion modules, the battery charge and discharge conversion is realized through corresponding control methods in different situations. Figure 5 ;

[0102] The battery system control process is as follows:

[0103] S401, build a battery and a bidirectional DC / DC conversion module;

[0104] The on / off states of the two IGBT electronic switches in the bidirectional DC / DC conversion module are always opposite;

[0105] S402, using a combination of constant current charging and constant voltage charging to achieve reasonable utilization of battery energy, using constant current charging at the beginning of charging, such as Figure 6 As shown in the figure, when the battery capacity reaches a certain value, it switches to constant voltage charging, as shown in the figure. Figure 7 As shown;

[0106] If the system power is excessive, the battery is charged. This method adopts the battery constant current charging control strategy to extract the actual working current i of the battery and the given battery charging current i refThe difference is made, and the obtained deviation is input into the proportional integral link. The output switch duty cycle D is used to generate a PWM signal, thereby controlling the IGBT adjustment system output; if the battery voltage has reached 95% of the rated value, the charging mode is switched to the constant voltage mode, and a double closed-loop control structure is adopted. The outer loop compares the actual value of the battery voltage u with the given value u. ref The difference enters the proportional integral link to output the current reference signal i ref , the inner loop then converts i ref The difference between the actual charging current i and the current is fed into the proportional-integral link, and the PWM modulation is used to generate the control signal for controlling the switch.

[0107] S403, using a constant power discharge method to discharge the battery, such as Figure 8 As shown;

[0108] If the system power is insufficient, the battery needs to be discharged to supplement the power shortage. The discharge also adopts a double closed-loop control structure. The outer loop compares the actual power value P with the given value P. ref The difference enters the proportional integral link to output the current reference value i ref , the inner loop then compares it with the actual current i, the difference enters the proportional integral link, and the output value generates the switch control signal through PWM modulation;

[0109] Step 5: Build a voltage-active inverter circuit and filter circuit to connect distributed power to the grid;

[0110] S501, build a three-phase voltage source inverter and filter circuit, refer to Figure 9 ;

[0111] The three-phase voltage source inverter consists of six IGBTs and diodes connected in reverse parallel, outputting SPWM pulse voltage. After the LC filter removes the high-order harmonics, it outputs three-phase sinusoidal current and is directly connected to the distribution network.

[0112] S502, use constant power (PQ) control strategy to control the system inverter, realize the independent control of active power and reactive power, refer to Figure 10 ;

[0113] First, the voltage and current in the power system are converted from the ABC coordinate system to the DQ coordinate system through the ABC / DQ module, thereby obtaining the d-axis and q-axis components of the current and voltage, realizing the decoupling of active power P and reactive power Q, and defining the ABC / DQ transformation matrix T abc·dqo as follows:

[0114]

[0115] Where ωt represents the rotation angle.

[0116] Secondly, the active power and reactive power reference values ​​Pref , Q ref The difference between the actual value P and Q is made, and the current component reference value i is output through the proportional integral link. dref 、i qref Next, the voltage reference component u is obtained by using the current inner loop control dref 、u qref , its control expression is:

[0117]

[0118] Among them, i dref Indicates the d-axis current reference value, i qref Indicates the q-axis current reference value, u d and u q Represents the feedforward compensation voltage or other disturbance compensation term, Ri d and Ri q Represents the resistance voltage drop compensation of d-axis and q-axis respectively, ωLi q and ωLi d Represents the back electromotive force decoupling of the d-axis and q-axis, i d represents the actual d-axis current measurement value obtained through coordinate transformation, i q represents the actual q-axis current measurement value obtained by coordinate transformation, s is the complex frequency variable, k p and k i is the proportional integral parameter of the current inner loop controller; the obtained u dref 、u qref The inverse transformation to the ABC coordinate system is used to control the PWM to generate trigger pulses; ultimately, the drive signal of the inverter is obtained to achieve the purpose of controlling the inverter to output constant power.

[0119] Step 6: Connect the distributed power sources to different locations of the distribution network, collect transient current signals of different ground faults, extract signal features and realize fault classification.

[0120] S601, connect the distributed power supply to different locations of the distribution network and extract the fault transient current signal, such as Figure 11 As shown;

[0121] The different locations where distributed power sources are connected to the fault have a direct impact on the fault current waveform. Distributed power sources are connected to the front end of the distribution network, the adjacent feeder end, and the end of the distribution network. The transient zero-sequence current of the ground fault under different connection locations and operating conditions is extracted to establish a fault data set.

[0122] S602, data preprocessing;

[0123] In order to eliminate the impact of differences in zero-sequence current amplitudes in different fault lines and ensure the comparability, stability, and accuracy of the analysis results, the collected time series data is standardized:

[0124]

[0125] Among them, max|X| represents the maximum absolute value of the sampled current, Represents the data obtained after normalization. i Represents the i-th sample value.

[0126] S603, extracting fault signal features using wavelet transform method;

[0127] The db6 wavelet is selected as the wavelet basis function to perform continuous wavelet transform on the original signal, decomposing and capturing its detail components to highlight the local features of the original signal. The specific transformation formula is as follows:

[0128]

[0129] Among them, WT(a,τ) represents the wavelet transform coefficient, x(t) represents the original signal, a represents the scaling parameter, τ represents the scale parameter, and ψ(t) represents the wavelet basis function, also known as the mother wavelet. After scaling and translation, the wavelet sequence ψ is obtained. a,τ (t):

[0130]

[0131] Where a represents the scaling parameter, τ represents the scale parameter, and R represents a set of natural numbers. Finally, a two-dimensional time-frequency graph is obtained, which carries both time-domain and frequency-domain information.

[0132] S604: Implement fault classification using a convolutional neural network.

[0133] A convolutional neural network model was built, and 80% of the samples were randomly selected for training. The remaining 20% ​​of the samples were used as a test set to evaluate model performance. This method ultimately set the model batch size to 32 and the initial learning rate to 0.0001. To address slow iterative convergence, the Adam optimizer was used to dynamically adjust the learning rate. Categorical cross entropy was used as the loss function for the multi-layer CNN, and training was performed 100 times.

[0134] In one embodiment, a typical medium voltage distribution network structure is constructed as follows: Figure 3 As shown in Figure 1, a single-phase arc grounding fault occurs at point F of line L2, and the transient zero-sequence current waveform of the fault is measured at point 1. The distributed power supply is connected to the front end A of the distribution network, the adjacent feeder end B, and the end C of the distribution network. The transient zero-sequence current waveform obtained by simulation is as follows: Figure 12-15 shown.

[0135] Regarding the fault transient waveform, when distributed generation (DGs) are connected to the front end of the distribution network, they supply power in parallel with the distribution network, increasing system capacity. After a single-phase arc fault occurs, the PV system and the distribution network jointly supply short-circuit current to the fault point, increasing the fault phase current. When DGs are located on feeders adjacent to the fault point, reverse current flows toward the fault point. In other words, compared to the case without DGs, this scenario also provides additional current, similarly increasing the fault phase current. However, when DGs are connected to the end of the distribution network, after a fault occurs, the short-circuit current flows from the fault point to the ground, and the current generated by the DGs does not contribute to the short-circuit current. Therefore, in this case, the short-circuit current is solely supplied by the system current, and the DGs' presence has little impact on the short-circuit current. This demonstrates that the presence of DGs influences the fault transient waveform and warrants consideration.

[0136] In order to include as many operating conditions as possible in the dataset, the simulation model parameters were adjusted in various ways, mainly considering various fault conditions such as random fault location, transition resistance, initial fault phase angle, and grounding operation mode of the distribution network:

[0137] (1) Fault line: Fault settings are performed on different lines, different locations on the same line, and at different fault points relative to the distributed power source;

[0138] (2) Fault initial phase angle: The time when the fault occurs is random, and the fault initial phase angle can be set arbitrarily within one cycle (0.02s);

[0139] (3) Transition resistance: set different linear transition resistances;

[0140] (4) Distribution network neutral point operation mode: two different grounding modes are set: neutral point ungrounded and neutral point resonant grounded;

[0141] Finally, the simulation results show four typical distribution network faults: low resistance grounding (label 1), high resistance grounding (label 2), stable arc grounding (label 3) and intermittent arc grounding (label 4). Figure 12-15 As shown. The above waveform is transformed by stationary wavelet to obtain a two-dimensional time-frequency diagram as shown Figure 16 shown.

[0142] Finally, the convolutional network is trained. Figure 17As can be seen from the figure, the accuracy gradually increases with the number of training rounds, exceeding 90% after 20 rounds. After 60 rounds of training, the convergence rate of accuracy and loss function gradually slows down, plateauing with little improvement. The loss function drops below 0.5%, and the accuracy is around 98% after the 100th round of training. The above shows that the proposed fault diagnosis method can accurately identify different fault types including single-phase grounding in distributed power systems. Accuracy over Epochs represents the change in accuracy with training rounds, Loss over Epochs represents the change in loss with training rounds, Epochs represents the number of training rounds, and Train Accuracy represents the training accuracy. Test Accuracy represents the test accuracy, Train Loss represents the training loss, and Test Loss represents the training loss.

[0143] In summary, the present invention provides a method for constructing a fault identification model for a hybrid distribution network of distributed power sources, which has the following characteristics:

[0144] (1) The EMTP software was used to construct an overhead-cable hybrid medium-voltage distribution network, in which the three-core cable was modeled according to the geometric dimensions and electrical parameters of the actual specific model of three-core armored cable to better simulate the actual situation;

[0145] (2) Build a photovoltaic power generation system and use the MPPT disturbance observation algorithm to track the maximum power of the photovoltaic system. The control principle is simple, the required measurement parameters are few, and the calculation cost is saved;

[0146] (3) Build a battery energy storage system, use a combination of constant current and constant voltage charging to realize battery charging, and use constant power control to realize battery discharge, so as to ensure the battery life from a practical perspective;

[0147] (4) Distributed power sources are connected to different locations of the distribution network, and transient current waveforms under different working conditions are extracted. Combined with existing signal analysis methods, transient signal features are extracted. On this basis, accurate fault identification in the distributed power system is achieved through neural network training.

[0148] Example 3:

[0149] The present invention based on the same inventive concept also provides a distributed power hybrid distribution network fault identification model construction system, such as Figure 18 As shown, including:

[0150] A construction module is used to build a hybrid simulation distribution network based on preset distribution network simulation information, and to build a distributed energy simulation system based on preset distributed energy structure information;

[0151] A simulation module, configured to connect the distributed energy simulation system to the hybrid simulation distribution network, perform single-phase grounding fault simulation on the hybrid simulation distribution network using an arc model, and collect fault transient current simulation signals of the hybrid simulation distribution network;

[0152] an extraction module, configured to extract features of the fault transient current simulation signal in combination with a wavelet transformation method to obtain a fault signal training feature, and determine a fault identification result based on the fault signal training feature;

[0153] The training module is used to train the fault recognition model with the fault signal training feature as input and the fault recognition result corresponding to the fault signal training feature as output to obtain the trained fault recognition model.

[0154] Preferably, the construction module constructs a distributed energy simulation system based on preset distributed energy structure information, including:

[0155] A photovoltaic power generation system including a photovoltaic array and a DC-DC boost module is constructed based on preset distributed energy structure information, and a distributed energy control strategy of the photovoltaic power generation system is set to apply a disturbance to the photovoltaic power generation system using a disturbance observation method;

[0156] Building a battery system including a battery and a bidirectional DC-DC conversion module based on preset distributed energy structure information, and setting the distributed energy control strategy of the battery system to adopt a constant current and constant voltage charging strategy and a constant power discharging strategy;

[0157] Building an inverter grid including a three-phase voltage source inverter and a filter circuit based on preset distributed energy structure information, and setting the distributed energy control strategy of the inverter grid to adopt a constant power control strategy;

[0158] A distributed energy system is constructed based on the photovoltaic power generation system, the battery system and the inverter grid.

[0159] Preferably, the constant current and constant voltage charging strategy in the building module includes:

[0160] When the power of the battery system meets the preset charging condition, the difference between the actual operating current of the battery in the battery system and the given battery charging current is extracted, and a proportional-integral control is performed based on the difference to output a switch duty cycle, and the switch duty cycle is pulse-width modulated to generate a control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC conversion module;

[0161] When the real-time collected battery voltage value meets the preset voltage rating, the current constant current charging mode is converted to constant voltage charging mode.

[0162] The constant voltage charging method adopts a dual closed-loop control structure, wherein the outer loop controls the difference between the actual battery voltage and the set battery voltage through proportional-integral control to output a current reference signal, and the inner loop controls the difference between the current reference signal and the actual battery charging current through proportional-integral control to output a switch duty cycle. The switch duty cycle is pulse-width modulated to generate a control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC conversion module.

[0163] Preferably, the constant power discharge mode in the building module includes:

[0164] When the battery system power meets the preset discharge condition, a double closed-loop control structure is adopted, and the difference between the actual battery voltage value and the given battery voltage value is controlled by the proportional integral control output current reference signal through the outer loop;

[0165] The difference between the current reference signal and the actual charging current of the battery is controlled by proportional-integral through the inner loop to output the switch duty cycle, and the switch duty cycle is pulse-width modulated to generate a control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC conversion module.

[0166] Preferably, the constant power control strategy in the building module includes:

[0167] The voltage and current in the inverter grid are converted from the ABC coordinate system to the DQ coordinate system through the ABC / DQ transformation matrix to obtain the actual value of active power and the actual value of reactive power;

[0168] The difference between the active power reference value and the actual active power value and the difference between the reactive power reference value and the actual reactive power value are controlled by proportional integral, and a D-axis current component reference value and a Q-axis current component reference value are output respectively;

[0169] Based on the D-axis current component reference value and the Q-axis current component reference value, a D-axis voltage reference component and a Q-axis voltage reference component are obtained by using a current inner loop control;

[0170] The D-axis voltage reference component and the Q-axis voltage reference component are inversely transformed into an ABC coordinate system and pulse-width modulated to obtain an inverter control signal, wherein the inverter control signal is used to control the three voltage source inverters to output constant power.

[0171] Preferably, the simulation module is further used to:

[0172] Connecting the distributed energy simulation system to different locations of the hybrid simulated distribution network, and using the Mayer arc model to perform single-phase grounding fault simulation on the hybrid simulated distribution network connected to the distributed energy simulation system, wherein the different locations of the hybrid simulated distribution network include the front end location of the hybrid simulated distribution network, the adjacent feeder end location and / or the distribution network end location;

[0173] Fault transient current simulation signals are collected at different access locations and under different operating conditions of the hybrid simulation distribution network.

[0174] Preferably, the extraction module extracts features of the fault transient current simulation signal in combination with a wavelet transform method to obtain fault signal training features, including:

[0175] performing standardization processing on the fault transient current simulation signal to obtain a processed fault transient current training signal;

[0176] Performing continuous wavelet transform on the processed fault transient current signal based on the db6 wavelet basis function to obtain local characteristics of the signal;

[0177] After scaling and translating the local features of the signal, a fault signal training feature including a two-dimensional time-frequency graph is obtained.

[0178] Example 4:

[0179] The present invention based on the same inventive concept also provides a method for identifying faults in a hybrid distribution network of distributed power sources, such as Figure 19 As shown, including:

[0180] Step S1: Acquire a fault transient current signal of a target hybrid distribution network, wherein the target hybrid distribution network is a distribution network connected to a distributed energy system;

[0181] Step S2: extracting the characteristics of the fault transient current signal by combining the wavelet transform method to obtain the fault signal characteristics;

[0182] Step S3: inputting the fault signal feature into a pre-trained fault identification model to obtain a fault identification result corresponding to the target hybrid distribution network;

[0183] The fault identification model is constructed by the aforementioned method for constructing a fault identification model of a distributed power hybrid distribution network.

[0184] Example 5:

[0185] The present invention based on the same inventive concept also provides a distributed power hybrid distribution network fault identification system, such as Figure 20 As shown, including:

[0186] an acquisition module, configured to acquire a fault transient current signal of a target hybrid distribution network, wherein the target hybrid distribution network is a distribution network connected to a distributed energy system;

[0187] A feature extraction module, configured to extract features of the fault transient current signal in combination with a wavelet transform method to obtain fault signal features;

[0188] An input module, configured to input the fault signal characteristics into a pre-trained fault identification model to obtain a fault identification result corresponding to the target hybrid distribution network;

[0189] The fault identification model is constructed by the aforementioned method for constructing a fault identification model of a distributed power hybrid distribution network.

[0190] Example 6

[0191] like Figure 21 As shown, the present invention also provides an electronic device, which may be a computer, a single-chip microcomputer, a smart mobile device, or the like. The electronic device in this embodiment may include a processor, a memory, a transceiver component, and the like. The memory, processor, and transceiver component are connected via a bus; the memory may be used to store an execution program, which may include instructions; and the processor may be used to execute the instructions stored in the memory. The memory may also be used to store data, which may be accessed and / or modified during the execution of the instructions.

[0192] The processor may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the storage medium to implement the corresponding method flow or corresponding function, so as to realize a distributed power hybrid distribution network fault identification model construction method and a distributed power hybrid distribution network fault identification method as described above.

[0193] Example 7

[0194] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It is understandable that the storage medium here can include both built-in storage media in the electronic device and, of course, extended storage media supported by the electronic device. The storage medium provides a storage space that stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space. These instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can realize the steps of a distributed power hybrid distribution network fault identification model construction method and a distributed power hybrid distribution network fault identification method as described above.

[0195] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0196] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts 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, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0197] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0198] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0199] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A method for constructing a fault identification model for a distributed power hybrid distribution network, characterized in that: include: Build a hybrid simulation distribution network based on the preset distribution network simulation information, and build a distributed energy simulation system based on the preset distributed energy structure information; Connecting the distributed energy simulation system to the hybrid simulation distribution network, performing single-phase grounding fault simulation on the hybrid simulation distribution network using an arc model, and collecting fault transient current simulation signals of the hybrid simulation distribution network; Extracting features of the fault transient current simulation signal using a wavelet transform method to obtain a fault signal training feature, and determining a fault identification result based on the fault signal training feature; The fault recognition model is trained with the fault signal training feature as input and the fault recognition result corresponding to the fault signal training feature as output to obtain a trained fault recognition model.

2. The method according to claim 1, characterized in that The distributed energy simulation system is constructed based on the preset distributed energy structure information, including: A photovoltaic power generation system including a photovoltaic array and a DC-DC boost module is constructed based on preset distributed energy structure information, and a distributed energy control strategy of the photovoltaic power generation system is set to apply a disturbance to the photovoltaic power generation system using a disturbance observation method; Building a battery system including a battery and a bidirectional DC-DC conversion module based on preset distributed energy structure information, and setting the distributed energy control strategy of the battery system to adopt a constant current and constant voltage charging strategy and a constant power discharging strategy; Building an inverter grid including a three-phase voltage source inverter and a filter circuit based on preset distributed energy structure information, and setting the distributed energy control strategy of the inverter grid to adopt a constant power control strategy; A distributed energy system is constructed based on the photovoltaic power generation system, the battery system and the inverter grid.

3. The method according to claim 2, characterized in that The constant current and constant voltage charging strategy includes: When the power of the battery system meets the preset charging condition, the difference between the actual operating current of the battery in the battery system and the given battery charging current is extracted, and a proportional-integral control is performed based on the difference to output a switch duty cycle, and the switch duty cycle is pulse-width modulated to generate a control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC conversion module; When the real-time collected battery voltage value meets the preset voltage rating, the current constant current charging mode is converted to constant voltage charging mode. The constant voltage charging method adopts a dual closed-loop control structure, wherein the outer loop controls the difference between the actual battery voltage and the set battery voltage through proportional-integral control to output a current reference signal, and the inner loop controls the difference between the current reference signal and the actual battery charging current through proportional-integral control to output a switch duty cycle. The switch duty cycle is pulse-width modulated to generate a control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC conversion module.

4. The method according to claim 2, characterized in that The constant power discharge mode includes: When the battery system power meets the preset discharge condition, a double closed-loop control structure is adopted, and the difference between the actual battery voltage value and the given battery voltage value is controlled by the proportional integral control output current reference signal through the outer loop; The difference between the current reference signal and the actual charging current of the battery is controlled by proportional-integral through the inner loop to output the switch duty cycle, and the switch duty cycle is pulse-width modulated to generate a control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC conversion module.

5. The method according to claim 2, characterized in that The constant power control strategy includes: The voltage and current in the inverter grid are converted from the ABC coordinate system to the DQ coordinate system through the ABC / DQ transformation matrix to obtain the actual value of active power and the actual value of reactive power; The difference between the active power reference value and the actual active power value and the difference between the reactive power reference value and the actual reactive power value are controlled by proportional integral, and a D-axis current component reference value and a Q-axis current component reference value are output respectively; Based on the D-axis current component reference value and the Q-axis current component reference value, a D-axis voltage reference component and a Q-axis voltage reference component are obtained by using a current inner loop control; The D-axis voltage reference component and the Q-axis voltage reference component are inversely transformed into an ABC coordinate system and pulse-width modulated to obtain an inverter control signal, wherein the inverter control signal is used to control the three voltage source inverters to output constant power.

6. The method according to claim 1, characterized in that The step of connecting the distributed energy simulation system to the hybrid simulation distribution network, performing single-phase grounding fault simulation on the hybrid simulation distribution network using an arc model, and collecting a fault transient current simulation signal of the hybrid simulation distribution network includes: Connecting the distributed energy simulation system to different locations of the hybrid simulated distribution network, and using the Mayer arc model to perform single-phase grounding fault simulation on the hybrid simulated distribution network connected to the distributed energy simulation system, wherein the different locations of the hybrid simulated distribution network include the front end location of the hybrid simulated distribution network, the adjacent feeder end location and / or the distribution network end location; Fault transient current simulation signals are collected at different access locations and under different operating conditions of the hybrid simulation distribution network.

7. The method according to claim 1, characterized in that The method of extracting features of the fault transient current simulation signal in combination with the wavelet transform method to obtain fault signal training features includes: performing standardization processing on the fault transient current simulation signal to obtain a processed fault transient current training signal; Performing continuous wavelet transform on the processed fault transient current signal based on the db6 wavelet basis function to obtain local characteristics of the signal; After scaling and translating the local features of the signal, a fault signal training feature including a two-dimensional time-frequency graph is obtained.

8. A distributed power hybrid distribution network fault identification model construction system, characterized in that: include: A construction module is used to build a hybrid simulation distribution network based on preset distribution network simulation information, and to build a distributed energy simulation system based on preset distributed energy structure information; A simulation module, configured to connect the distributed energy simulation system to the hybrid simulation distribution network, perform single-phase grounding fault simulation on the hybrid simulation distribution network using an arc model, and collect fault transient current simulation signals of the hybrid simulation distribution network; an extraction module, configured to extract features of the fault transient current simulation signal in combination with a wavelet transformation method to obtain a fault signal training feature, and determine a fault identification result based on the fault signal training feature; The training module is used to train the fault recognition model with the fault signal training feature as input and the fault recognition result corresponding to the fault signal training feature as output to obtain the trained fault recognition model.

9. The system according to claim 8, characterized in that The construction module constructs a distributed energy simulation system based on preset distributed energy structure information, including: A photovoltaic power generation system including a photovoltaic array and a DC-DC boost module is constructed based on preset distributed energy structure information, and a distributed energy control strategy of the photovoltaic power generation system is set to apply a disturbance to the photovoltaic power generation system using a disturbance observation method; Building a battery system including a battery and a bidirectional DC-DC conversion module based on preset distributed energy structure information, and setting the distributed energy control strategy of the battery system to adopt a constant current and constant voltage charging strategy and a constant power discharging strategy; Building an inverter grid including a three-phase voltage source inverter and a filter circuit based on preset distributed energy structure information, and setting the distributed energy control strategy of the inverter grid to adopt a constant power control strategy; A distributed energy system is constructed based on the photovoltaic power generation system, the battery system and the inverter grid.

10. The system according to claim 9, characterized in that The constant current and constant voltage charging strategy in the building block includes: When the power of the battery system meets the preset charging condition, the difference between the actual operating current of the battery in the battery system and the given battery charging current is extracted, and a proportional-integral control is performed based on the difference to output a switch duty cycle, and the switch duty cycle is pulse-width modulated to generate a control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC conversion module; When the real-time collected battery voltage value meets the preset voltage rating, the current constant current charging mode is converted to constant voltage charging mode. The constant voltage charging method adopts a dual closed-loop control structure, wherein the outer loop controls the difference between the actual battery voltage and the set battery voltage through proportional-integral control to output a current reference signal, and the inner loop controls the difference between the current reference signal and the actual battery charging current through proportional-integral control to output a switch duty cycle. The switch duty cycle is pulse-width modulated to generate a control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC conversion module.

11. The system according to claim 9, wherein: The constant power discharge mode in the building module includes: When the battery system power meets the preset discharge condition, a double closed-loop control structure is adopted, and the difference between the actual battery voltage value and the given battery voltage value is controlled by the proportional integral control output current reference signal through the outer loop; The difference between the current reference signal and the actual charging current of the battery is controlled by proportional-integral through the inner loop to output the switch duty cycle, and the switch duty cycle is pulse-width modulated to generate a control signal for turning on and off the IGBT electronic switch in the bidirectional DC-DC conversion module.

12. The system according to claim 9, wherein: The constant power control strategy in the building block includes: The voltage and current in the inverter grid are converted from the ABC coordinate system to the DQ coordinate system through the ABC / DQ transformation matrix to obtain the actual value of active power and the actual value of reactive power; The difference between the active power reference value and the actual active power value and the difference between the reactive power reference value and the actual reactive power value are controlled by proportional integral, and a D-axis current component reference value and a Q-axis current component reference value are output respectively; Based on the D-axis current component reference value and the Q-axis current component reference value, a D-axis voltage reference component and a Q-axis voltage reference component are obtained by using a current inner loop control; The D-axis voltage reference component and the Q-axis voltage reference component are inversely transformed into an ABC coordinate system and pulse-width modulated to obtain an inverter control signal, wherein the inverter control signal is used to control the three voltage source inverters to output constant power.

13. The system according to claim 8, wherein: The simulation module is further configured to: Connecting the distributed energy simulation system to different locations of the hybrid simulated distribution network, and using the Mayer arc model to perform single-phase grounding fault simulation on the hybrid simulated distribution network connected to the distributed energy simulation system, wherein the different locations of the hybrid simulated distribution network include the front end location of the hybrid simulated distribution network, the adjacent feeder end location and / or the distribution network end location; Fault transient current simulation signals are collected at different access locations and under different operating conditions of the hybrid simulation distribution network.

14. The system according to claim 8, wherein: The extraction module extracts features from the fault transient current simulation signal using a wavelet transform method to obtain fault signal training features, including: performing standardization processing on the fault transient current simulation signal to obtain a processed fault transient current training signal; Performing continuous wavelet transform on the processed fault transient current signal based on the db6 wavelet basis function to obtain local characteristics of the signal; After scaling and translating the local features of the signal, a fault signal training feature including a two-dimensional time-frequency graph is obtained.

15. A method for identifying faults in a hybrid distribution network of distributed power sources, characterized in that: include: Acquiring a fault transient current signal of a target hybrid distribution network, wherein the target hybrid distribution network is a distribution network connected to a distributed energy system; Extracting the characteristics of the fault transient current signal using a wavelet transform method to obtain the fault signal characteristics; Inputting the fault signal characteristics into a pre-trained fault identification model to obtain a fault identification result corresponding to the target hybrid distribution network; The fault identification model is constructed by a method for constructing a distributed power hybrid distribution network fault identification model according to any one of claims 1 to 7.

16. A distributed power hybrid distribution network fault identification system, characterized in that: include: an acquisition module, configured to acquire a fault transient current signal of a target hybrid distribution network, wherein the target hybrid distribution network is a distribution network connected to a distributed energy system; A feature extraction module, configured to extract features of the fault transient current signal in combination with a wavelet transform method to obtain fault signal features; An input module, configured to input the fault signal characteristics into a pre-trained fault identification model to obtain a fault identification result corresponding to the target hybrid distribution network; The fault identification model is constructed by a method for constructing a distributed power hybrid distribution network fault identification model according to any one of claims 1 to 7.

17. An electronic device, characterized in that: include: at least one processor and memory; The memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, the method for constructing a distributed power hybrid distribution network fault identification model according to any one of claims 1 to 7 and the method for identifying a distributed power hybrid distribution network fault according to claim 15 are implemented.

18. A readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, the distributed power hybrid distribution network fault identification model construction method according to any one of claims 1 to 7 and the distributed power hybrid distribution network fault identification method according to claim 15 are implemented.

Images