A novel grid-connected microgrid fault detection device and detection method

Abstract

The present invention discloses a novel grid-connected microgrid fault detection device and detection method. The detection device includes a main grid, a microgrid, a voltage transformer, a current transformer, a wireless communication device, a microprocessor, an alarm and a display. The main grid is connected to the microgrid, the microgrid is connected to the voltage transformer and the current transformer, the voltage transformer and the current transformer are connected to the wireless communication device, the wireless communication device is connected to the microprocessor, and the microprocessor is connected to the alarm and the display. The detection method of the present invention is characterized in that a novel grid-connected microgrid fault detection model is established, so as to quickly and accurately detect the grid-connected microgrid fault, and can detect the location of the fault.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a novel grid-connected microgrid fault detection device and detection method. Background Art

[0002] As a low-voltage small power electronic power system, the fault characteristics of microgrids have undergone tremendous changes compared to those of traditional power grids. The grid structure has changed from a single-power radial network to a complex structure of multiple power sources. The magnitude, direction and duration of the fault current have changed, the action characteristics of the relay will be affected, and the protection device will malfunction or refuse to operate. Because the structure and operation mode of microgrids are different from those of large power grids, the corresponding fault detection, location and fault recovery methods are also different from those of traditional large power grids. Microgrid fault detection technology helps to achieve fault isolation and recovery, and plays an important role in microgrid control and operation. At present, the distributed power sources connected to microgrids are mainly inverter-type. Due to their output current limiting, traditional fault detection that relies on current characteristics is no longer applicable. Data-driven microgrid fault detection technology has become a research hotspot in recent years. Summary of the invention

[0003] The purpose of the present invention is to provide a novel grid-connected microgrid fault detection device and detection method to solve the problem of the prior art that it is seriously dependent on the network structure of the microgrid. It can directly detect the fault of the grid-connected microgrid based on the data of electrical quantities and detect the location of the fault.

[0004] Technical solution of the present invention:

[0005] A novel grid-connected microgrid fault detection device, wherein the main grid (1) is connected to a microgrid (2), the microgrid is connected to a voltage transformer (3) and a current transformer (4), the voltage transformer and the current transformer are connected to a wireless communication device (5), the wireless communication device is connected to a microprocessor (6), and the microprocessor is connected to an alarm (7) and a display (8).

[0006] Further: the main grid is a 20kV power grid.

[0007] Further: the microgrid is composed of circuit breakers CB1, CB2, CB3, CB4, CB5, CB6, CB7, CB8, CB9, CB10, CB11, distribution lines DL1, DL2, DL3, DL4, DL5, loads L1, L2, L3, L4, L5 and distributed power sources DG1, DG2.

[0008] Further: the distributed power source DG1 is a droop control power source, and the distributed power source DG2 is a constant power control power source.

[0009] Further: The microprocessor is STM32, the alarm is a buzzer, and the display is an LCD.

[0010] The detection method of a new grid-connected microgrid fault detection device includes the following steps:

[0011] Step 1: Use a voltage transformer and a current transformer to collect the three-phase voltage and three-phase current of the distributed power source DG1 under normal and short-circuit faults. The short-circuit faults include single-phase ground short-circuit, two-phase interphase short-circuit, two-phase ground short-circuit, three-phase interphase short-circuit, and three-phase ground short-circuit. The short-circuit fault positions are F1, F2, F3, F4, where F1 is at the point of common coupling PCC, F2 is on the line where the droop control power source DG1 is located, F3 is on the line where the constant power control power source DG2 is located, and F4 is on the line where the load L3 is located.

[0012] Step 2: Use a wireless communication device to transmit the collected three-phase voltage and three-phase current to the microprocessor. By processing the voltage and current with the microprocessor, the three-phase voltage, three-phase current, zero-sequence current, positive-sequence current, negative-sequence current, active power, reactive power, and frequency can be obtained, and a feature library of electrical quantities is established.

[0013] Step 3: Use the obtained historical electrical quantity data as training samples to establish a classifier.

[0014] Step 4: Use the newly obtained electrical quantity data as input and use the classifier to judge whether there is a fault and determine the location where the fault occurs.

[0015] Step 5: When the classifier determines that there is a fault, the alarm will send an alarm message, the display will show the fault information, and the microprocessor will transmit a tripping command to the circuit breaker where the faulty component is located through the wireless communication device, and the circuit breaker will trip.

[0016] Preferably, step 3 above is replaced by the following steps:

[0017] Step 1: If there are N training samples, randomly select N training samples with replacement, and use the selected N training samples to train a basic classifier as the samples at the root node of the classifier;

[0018] Step 2: Define the Gini coefficient:

[0019] Step 3: When each training sample has M features, randomly select m features from these M features, where m << M. Traverse the selected m features and all their splitting points, calculate the Gini coefficient, and select the feature with the smallest Gini coefficient and its corresponding splitting point as the optimal feature and the optimal splitting point;

[0020] Step 4: Split according to the optimal features and the optimal splitting point, and distribute the training samples to two child nodes according to the features;

[0021] Step 5: Recursively call steps 3 and 4 for the two child nodes until the labels of all samples in the child nodes are unique, and a basic classifier is obtained;

[0022] Step 6: Repeat all the above steps to generate a large number of basic classifiers. Each basic classifier makes a judgment on the training samples, and the final classifier is generated according to the majority rule.

[0023] Beneficial effects of the present invention:

[0024] The present invention collects historical normal and short-circuit fault voltage and current data, processes the collected data into various electrical quantities, and uses the electrical quantity data to establish a classifier, thereby detecting faults in the grid-connected microgrid and detecting the location of the fault, solving the problem of the prior art that is heavily dependent on the network structure of the microgrid, and has strong practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A schematic diagram of a novel grid-connected microgrid fault detection device of the present invention

[0026] Figure 2 A schematic diagram of a novel grid-connected microgrid fault detection method of the present invention DETAILED DESCRIPTION

[0027] The present invention is further described below in conjunction with the accompanying drawings and specific implementation methods.

[0028] like Figure 1As shown, it is a schematic diagram of a novel grid-connected microgrid fault detection device. The main grid (1) is connected to the microgrid (2), the microgrid is connected to the voltage transformer (3) and the current transformer (4), the voltage transformer and the current transformer are connected to the wireless communication device (5), the wireless communication device is connected to the microprocessor (6), and the microprocessor is connected to the alarm (7) and the display (8). The main grid is a 20kV power grid, the microgrid is composed of circuit breakers CB1, CB2, CB3, CB4, CB5, CB6, CB7, CB8, CB9, CB10, CB11, distribution lines DL1, DL2, DL3, DL4, DL5, loads L1, L2, L3, L4, L5 and distributed power sources DG1 and DG2, the distributed power source DG1 is a droop control power source, the distributed power source DG2 is a constant power control power source, the microprocessor is STM32, the alarm is a buzzer, and the display is an LCD. The voltage transformer and current transformer collect the three-phase voltage and three-phase current of the distributed power source DG1 in the microgrid under normal and short-circuit fault conditions. The three-phase voltage and three-phase current are transmitted to the microprocessor using wireless communication equipment. The microprocessor can obtain the feature library of various electrical quantities, and the classifier is established using the feature library. The classifier can detect the fault of the parallel microgrid. When a fault occurs, the alarm will issue an alarm message and the display will display the fault information.

[0029] The detection method of the novel grid-connected microgrid fault detection device is as follows: Figure 2 As shown, the following steps are included:

[0030] Step 1: Use voltage transformers and current transformers to collect the three-phase voltage and three-phase current of the distributed power supply DG1 under normal and short-circuit faults. The short-circuit faults include single-phase ground short circuit, two-phase phase short circuit, two-phase ground short circuit, three-phase phase short circuit, and three-phase ground short circuit. The short-circuit fault locations are F1, F2, F3, and F4. F1 is the common coupling point PCC, F2 is the line where the droop control power supply DG1 is located, F3 is the line where the constant power control power supply DG2 is located, and F4 is the line where the load L3 is located.

[0031] Step 2: Use wireless communication equipment to transmit the collected three-phase voltage and three-phase current to the microprocessor. Use the microprocessor to process the voltage and current to obtain three-phase voltage, three-phase current, zero-sequence current, positive-sequence current, negative-sequence current, active power, reactive power, and frequency, and establish a characteristic library of electrical quantities.

[0032] Step 3: Use the historical electrical quantity data obtained as training samples to establish a classifier.

[0033] Step 4: Use the newly obtained electrical quantity data as input, and use a classifier to determine whether there is a fault and the location where the fault occurs.

[0034] Step 5: When the classifier determines that there is a fault, the alarm will send an alarm message, the display will show the fault information, and the microprocessor will transmit a tripping command to the circuit breaker where the faulty component is located through a wireless communication device, and the circuit breaker will trip.

[0035] Preferably, step 3 described above is replaced by the following steps:

[0036] Step 1: If there are N training samples, randomly select N training samples with replacement, and use the selected N training samples to train a basic classifier as the samples at the root node of the classifier;

[0037] Step 2: Define the Gini coefficient:

[0038] Step 3: When each training sample has M features, randomly select m features from these M features, where m << M. Traverse the selected m features and all their splitting points, calculate the Gini coefficient, and select the feature with the smallest Gini coefficient and its corresponding tangent point as the optimal feature and the optimal splitting point;

[0039] Step 4: Split according to the optimal feature and the optimal splitting point, and distribute the training samples to two child nodes according to the feature;

[0040] Step 5: Recursively call steps 3 and 4 for the two child nodes until the labels of all samples in the child nodes are unique, and a basic classifier is obtained;

[0041] Step 6: Repeat all the above steps to generate a large number of basic classifiers. Each basic classifier makes a judgment on the training samples, and according to the principle that the minority obeys the majority, a final classifier is generated.

[0042] The specific implementation manners of the present invention have been described in detail above in conjunction with the accompanying drawings. However, the above content is only the preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. Any content related to the present invention and its related scope is regarded as plagiarism.

Claims

1. A novel grid-connected microgrid fault detection device. The main grid (1) is connected to the microgrid (2). The microgrid is connected to a voltage transformer (3) and a current transformer (4). The voltage transformer and the current transformer are connected to a wireless communication device (5). The wireless communication device is connected to a microprocessor (6). The microprocessor is connected to an alarm (7) and a display (8); The microgrid consists of circuit breakers CB1, CB2, CB3, CB4, CB5, CB6, CB7, CB8, CB9, CB10, CB11, distribution lines DL1, DL2, DL3, DL4, DL5, loads L1, L2, L3, L4, L5, and distributed power sources DG1, DG2; The control steps are as follows: (6-1) Use the voltage transformer and the current transformer to collect the three-phase voltage and three-phase current of the distributed power source DG1 under normal and short-circuit faults. The short-circuit faults include single-phase ground short circuit, two-phase interphase short circuit, two-phase ground short circuit, three-phase interphase short circuit, and three-phase ground short circuit. The short-circuit fault locations are F1, F2, F3, F4. F1 is at the point of common coupling (PCC), F2 is on the line where the droop control power source DG1 is located, F3 is on the line where the constant power control power source DG2 is located, and F4 is on the line where the load L3 is located; (6-2) Use the wireless communication device to transmit the collected three-phase voltage and three-phase current to the microprocessor. The microprocessor processes the voltage and current to obtain the three-phase voltage, three-phase current, zero-sequence current, positive-sequence current, negative-sequence current, active power, reactive power, and frequency, and establish a feature library of electrical quantities; (6-3) Use the obtained historical electrical quantity data as training samples to establish a classifier; (6-4) Use the newly obtained electrical quantity data as input and use the classifier to judge whether there is a fault and determine the location where the fault occurs; (6-5) When the classifier determines that there is a fault, the alarm will send an alarm message, the display will display the fault information, and the microprocessor will transmit a tripping command to the circuit breaker where the fault element is located through the wireless communication device, and the circuit breaker will trip; The specific steps of (6-3) are as follows: (7-1) If there are N training samples, randomly select N training samples with replacement, and use the selected N training samples to train a basic classifier as the samples at the root node of the classifier; (7-2)Define the Gini coefficient: (7-3) When each training sample has M features, randomly select m features from these M features, where m << M. Traverse the selected m features and all their splitting points, calculate the Gini coefficient, and select the feature with the smallest Gini coefficient and its corresponding splitting point as the optimal feature and the optimal splitting point; (7-4) Split according to the optimal feature and the optimal splitting point, and distribute the training samples to two child nodes according to the feature; (7-5) Recursively call steps (7-3) and (7-4) for the two child nodes until the labels of all samples in the child nodes are unique, and a basic classifier is obtained; (7-6) Repeat all the above steps to generate a large number of basic classifiers. Each basic classifier makes a judgment on the training samples, and the final classifier is generated according to the majority rule.

2. A novel grid-connected microgrid fault detection device according to claim 1, characterized in that: The main grid is a 20kV power grid.

3. A novel grid-connected microgrid fault detection device according to claim 1, characterized in that: The distributed power source DG1 is a droop control power source, and the distributed power source DG2 is a constant power control power source.

4. A novel grid-connected microgrid fault detection device according to claim 1, characterized in that: The microprocessor is STM32, the alarm is a buzzer, and the display is an LCD.

Images