Power distribution terminal based on genetic algorithm and fault judgment method

Abstract

The invention discloses a power distribution terminal and a fault discrimination method based on a genetic algorithm, and relates to the technical field of power distribution automation of a power system, and the method comprises the following steps: collecting real-time state data of each feeder line in a power distribution network, and extracting fault features; generating a fault probability for the fault features based on a pre-trained fault prediction model, and judging whether a fault occurs based on the fault probability; if a fault occurs, a fault equipment combination is obtained based on a genetic algorithm, and fault positioning is realized; according to the obtained fault equipment combination, driving automatic control equipment of the power distribution network to send a fault isolation instruction to realize fault isolation; and after fault isolation, the power distribution terminal generates a power supply recovery scheme according to the topological structure and the load condition of the power distribution network. According to the method, the fault position can be quickly and accurately determined and effectively processed through the genetic algorithm under the conditions of a complex power distribution network structure and a fault.

Description

Technical Field

[0001] This invention relates to the field of power system distribution automation technology, specifically to a distribution terminal and fault diagnosis method based on a genetic algorithm. Background Technology

[0002] With the continuous development of power systems, the scale of distribution networks is becoming increasingly large and the structure increasingly complex. As an important component of the distribution network, the rapid and accurate fault diagnosis of feeders is crucial for ensuring the reliable operation of the power system.

[0003] For example, patent CN119598422A discloses a method, device, equipment, and medium for predicting faults in flexible DC distribution networks. The method involves collecting grid operation data of the flexible DC distribution network; extracting features from the grid operation data to obtain feature variables; optimizing a preset initial weight vector using a preset optimization target model to adjust the feature weight allocation; and calculating the fault probability based on the feature weight allocation, the feature vector, and a pre-constructed fault probability prediction model.

[0004] For example, patent CN115112997A discloses a method, device, and terminal equipment for locating fault intervals in a distribution network. After a fault occurs, a multi-bus node and an N-order fault information description matrix are established; the fault current of the multi-branch bus node is judged; the judgment matrix P of the fault switching node in each bus in the distribution network is determined; a logical AND operation is performed on each Pij and Pji element in the judgment matrix P to determine the fault occurrence point; a multi-population genetic algorithm is used to further determine the fault interval, and the nodes between two bus nodes are further divided into nodes; then a switching function and a fitness function are constructed, and the minimum solution is obtained to obtain the final fault interval.

[0005] Existing technologies have the following drawbacks: firstly, feeder automation terminals cannot predict equipment failures in advance; secondly, when failures are predicted, traditional rule-based fault identification methods are not adaptable enough to complex distribution networks. Existing feeder automation terminals suffer from slow calculation speed and insufficient accuracy in fault location and handling, especially in complex distribution network structures and fault conditions, making it difficult to quickly and accurately determine the fault location and effectively handle it, thus failing to meet the needs of modern power systems. Therefore, a more efficient and accurate fault identification method is needed. Summary of the Invention

[0006] To address the aforementioned technical problems, the first aspect of this invention provides a power distribution terminal based on a genetic algorithm, comprising the following steps:

[0007] Step 1: Collect real-time status data of each feeder in the distribution network and extract fault characteristics;

[0008] Step 2: Based on the fault features and a pre-trained fault prediction model, generate fault probabilities and determine whether a fault will occur based on these probabilities;

[0009] Step 3: If a fault occurs, the faulty equipment combination is obtained based on the genetic algorithm to achieve fault location;

[0010] Step 4: Based on the obtained combination of faulty devices, drive the power distribution network automation control equipment to issue a fault isolation command to achieve fault isolation;

[0011] Step 5: After fault isolation, the power distribution terminal generates a power restoration plan based on the topology and load conditions of the power distribution network.

[0012] In a preferred embodiment, the fault characteristics include: current, voltage, power, temperature, vibration frequency, and switching status.

[0013] In a preferred embodiment, the process of pre-training the fault prediction model is as follows:

[0014] Historical state data is acquired and divided into training and test sets. Fault features are extracted, and a classifier is constructed. The fault features in the historical state data of the training set are used as input data for the classifier, which outputs the fault probability. The parameters of the classifier are continuously adjusted to train the classifier and obtain an initial classifier. The initial classifier is tested using a test set, and a classifier that meets the preset accuracy is output as the fault prediction model. The classifier includes one of the following: Support Vector Machine, Naive Bayes, Temporal Neural Network (RNN) or CNN, and Long Short-Term Memory Network (LSTM).

[0015] In a preferred embodiment, the process of determining whether a fault will occur is as follows:

[0016] If the failure probability output by the failure prediction model is greater than or equal to the preset failure probability threshold, it indicates that a failure has occurred.

[0017] If the fault probability output by the fault prediction model is less than the preset fault probability threshold, it means that no fault has occurred, and the process returns to step 1 to continue collecting real-time status data of each feeder in the distribution network.

[0018] In a preferred embodiment, the process of obtaining the faulty device combination based on the genetic algorithm is as follows:

[0019] The equipment on each feeder in the distribution network is mapped to a binary chromosome representation that can be processed by a genetic algorithm. A number of chromosomes are randomly generated to form an initial population. According to a pre-designed fitness function, the fitness value of each chromosome is calculated. The chromosomes with fitness values ​​greater than or equal to a pre-set fitness threshold are selected as parent chromosomes. Crossover is performed on the selected parent chromosomes to generate new offspring chromosomes. Mutation is performed on the offspring chromosomes to increase the diversity of the population. The parent and offspring chromosomes are merged to form a new generation population. It is determined whether the termination condition is met. If it is met, the iteration stops, the chromosome corresponding to the highest fitness value is output, and the binary encoded chromosome is decoded to identify the faulty equipment and obtain the faulty equipment combination. Otherwise, the iteration continues.

[0020] In a preferred embodiment, the fault isolation command includes: circuit breaker control command, disconnector control command, sectionalizer control command, and grounding switch control command.

[0021] A second aspect of this invention provides a fault diagnosis method for power distribution terminals based on genetic algorithms, comprising the following steps:

[0022] Step 1: Collect real-time status data of each feeder in the distribution network and extract fault characteristics;

[0023] Step 2: Based on the fault features and a pre-trained fault prediction model, generate fault probabilities and determine whether a fault will occur based on these probabilities;

[0024] Step 3: If a fault occurs, the faulty equipment combination is obtained based on the genetic algorithm to achieve fault location;

[0025] Step 4: Based on the obtained combination of faulty devices, drive the power distribution network automation control equipment to issue a fault isolation command to achieve fault isolation;

[0026] Step 5: After fault isolation, the power distribution terminal generates a power restoration plan based on the topology and load conditions of the power distribution network.

[0027] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention collects real-time status data of each feeder in the distribution network and extracts fault features; based on a pre-trained fault prediction model, it generates fault probabilities for the fault features and determines whether a fault will occur based on the fault probabilities; if a fault occurs, it obtains faulty equipment combinations based on a genetic algorithm to achieve fault location; based on the obtained faulty equipment combinations, it drives the distribution network automation control equipment to issue fault isolation commands to achieve fault isolation; after fault isolation, the distribution terminal generates a power restoration plan based on the distribution network topology and load conditions. By using a fault prediction model to predict the probability of equipment fault occurrence in advance, and obtaining faulty equipment combinations through a genetic algorithm when a fault is predicted to occur, fault location is achieved; based on the obtained faulty equipment combinations, the distribution network automation control equipment is driven to issue fault isolation commands to achieve fault isolation. The genetic algorithm solves the problems of insufficient adaptability of traditional rule-based fault identification methods in complex distribution networks, slow calculation speed and insufficient accuracy of feeder automation terminals in fault location and processing; especially the difficulty in quickly and accurately determining fault locations and effectively handling faults under complex distribution network structures and fault conditions. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating the steps of the power distribution terminal and fault diagnosis method based on genetic algorithm according to an embodiment of this application. Detailed Implementation

[0029] Please see Figure 1 As shown, the power distribution terminal based on the genetic algorithm includes the following steps:

[0030] Step 1: Collect real-time status data of each feeder in the distribution network and extract fault characteristics; the fault characteristics include: current, voltage, power, temperature, vibration frequency, switch status, etc.

[0031] Based on the above embodiments, it should be added that the real-time status data includes: current data, voltage data, power data, temperature data, vibration frequency data, and switch status data; these are acquired by current transformers, voltage transformers, three-phase power transmitters, thermocouple temperature sensors, vibration sensors, circuit breakers, etc.; fault features can be extracted using signal processing methods such as Fourier transform and wavelet transform.

[0032] Distribution terminals are an important component of distribution automation systems. They are installed at key nodes in the distribution network, such as feeder switches and distribution transformers. Their core function is to monitor the operating status of distribution equipment in real time (such as parameters like voltage, current, and power factor), execute control commands issued by the distribution automation system (such as switch opening and closing operations), and interact with the distribution automation system to achieve remote monitoring and automated management of the distribution network.

[0033] Step 2: Generate fault probabilities based on the fault features using a pre-trained fault prediction model, and determine whether a fault will occur based on the fault probabilities.

[0034] The process of pre-training the fault prediction model is as follows: acquire historical state data, divide it into training and test sets, extract fault features, construct a classifier, use the fault features in the historical state data of the training set as input data for the classifier, the classifier outputs the fault probability, continuously adjust the parameters of the classifier, train the classifier to obtain an initial classifier; use the test set to test the initial classifier, output a classifier that meets the preset accuracy, and use it as the fault prediction model; the classifier includes, but is not limited to, one of: support vector machine, Naive Bayes, temporal neural network RNN ​​or CNN, and long short-term memory network LSTM;

[0035] The process of determining whether a malfunction will occur is as follows:

[0036] If the failure probability output by the failure prediction model is greater than or equal to the preset failure probability threshold, it indicates that a failure has occurred.

[0037] If the fault probability output by the fault prediction model is less than the preset fault probability threshold, it means that no fault has occurred, and the process returns to step 1 to continue collecting real-time status data of each feeder in the distribution network.

[0038] Based on the above embodiments, it should be noted that the pre-set fault probability threshold was obtained by technical researchers through extensive research. The fault probability threshold may vary depending on the scenario; for example, the threshold for shopping malls, industrial parks, and residential communities is relatively higher to ensure that the fault probability output by the fault prediction model reflects the actual risk.

[0039] Step 3: If a fault occurs, obtain the faulty equipment combination based on the genetic algorithm to achieve fault location.

[0040] The process of obtaining faulty equipment combinations based on genetic algorithms is as follows: The equipment on each feeder in the distribution network is mapped to a binary chromosome representation that can be processed by the genetic algorithm. A number of chromosomes are randomly generated to form an initial population. Based on a pre-designed fitness function, the fitness value of each chromosome is calculated. A parent chromosome with a fitness value greater than or equal to a pre-set fitness threshold is selected. Crossover is performed on the selected parent chromosomes to generate new child chromosomes. Mutation is performed on the child chromosomes to increase population diversity. The parent and child chromosomes are merged to form a new generation population. The process is then checked to determine if the termination condition is met. If it is, the iteration stops, and the chromosome corresponding to the highest fitness value is output. The binary encoded chromosome is decoded to identify the faulty equipment, thus obtaining the faulty equipment combination. Otherwise, the iteration continues.

[0041] Based on the above embodiments, it is necessary to further explain that the equipment of each feeder in the distribution network is mapped to a binary bit, and a binary string of length n is used to represent each possible fault combination. Each bit corresponds to a device, such as "1" indicating that the corresponding device has failed and "0" indicating that the device is normal. For example, if a feeder in the distribution network has 3 devices a, b, and c, then the binary bit [1, 0, 1] indicates that devices a and c are faulty and device b is normal.

[0042] The termination condition is the maximum number of iterations of the genetic algorithm. For example, if the number of iterations is set to 100, the iteration stops when this number is reached. Alternatively, if the change in the highest fitness value of the population is less than a threshold for several consecutive generations, the genetic algorithm is considered to have converged and the iteration stops. For example, if the change in fitness value is less than 0.0001 after 90 iterations, the genetic algorithm is considered to have converged and the iteration stops.

[0043] Step 4: Based on the obtained combination of faulty devices, drive the power distribution network automation control equipment to issue a fault isolation command to achieve fault isolation.

[0044] The fault isolation commands include: circuit breaker control commands, disconnector switch control commands, sectionalizing switch control commands, grounding switch control commands, etc.

[0045] Based on the above embodiments, it should be understood that, for example, when a fault occurs in a section of a feeder line in the distribution network due to abnormal voltage, the distribution network automation control equipment is driven to send a circuit breaker control command, such as a tripping command, to the circuit breaker so that the circuit breaker can quickly disconnect the faulty line from the distribution network and prevent the fault from further expanding.

[0046] Step 5: After fault isolation, the power distribution terminal generates a power restoration plan based on the topology and load conditions of the power distribution network.

[0047] A fault detection method for power distribution terminals based on genetic algorithms, comprising the following steps:

[0048] Step 1: Collect real-time status data of each feeder in the distribution network and extract fault characteristics;

[0049] Step 2: Based on the fault features and a pre-trained fault prediction model, generate fault probabilities and determine whether a fault will occur based on these probabilities;

[0050] Step 3: If a fault occurs, the faulty equipment combination is obtained based on the genetic algorithm to achieve fault location;

[0051] Step 4: Based on the obtained combination of faulty devices, drive the power distribution network automation control equipment to issue a fault isolation command to achieve fault isolation;

[0052] Step 5: After fault isolation, the power distribution terminal generates a power restoration plan based on the topology and load conditions of the power distribution network.

[0053] Example 2

[0054] The distribution terminal and fault diagnosis method based on genetic algorithms described in this embodiment has been successfully applied to 10kV distribution substation terminals (DTUs) and feeder terminals (FTUs), achieving significant economic benefits. When the fault prediction model predicts a fault, the genetic algorithm is used for fault location and isolation, and a power restoration plan is formulated based on the distribution network topology and load conditions.

[0055] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0056] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

[0057] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A power distribution terminal based on a genetic algorithm, characterized in that, Includes the following steps: Step 1: Collect real-time status data of each feeder in the distribution network and extract fault characteristics; Step 2: Based on the fault features and a pre-trained fault prediction model, generate fault probabilities and determine whether a fault will occur based on these probabilities; Step 3: If a fault occurs, the faulty equipment combination is obtained based on the genetic algorithm to achieve fault location; Step 4: Based on the obtained combination of faulty devices, drive the power distribution network automation control equipment to issue a fault isolation command to achieve fault isolation; Step 5: After fault isolation, the power distribution terminal generates a power restoration plan based on the topology and load conditions of the power distribution network.

2. The power distribution terminal based on genetic algorithm according to claim 1, characterized in that, The fault characteristics include: current, voltage, power, temperature, vibration frequency, and switching status.

3. The power distribution terminal based on genetic algorithm according to claim 1, characterized in that, The process of pre-training a fault prediction model is as follows: Historical state data is acquired, and training and testing sets are divided. Fault features are extracted, and a classifier is constructed. The fault features in the historical state data of the training set are used as the input data of the classifier. The classifier outputs the fault probability. The parameters of the classifier are continuously adjusted to train the classifier and obtain the initial classifier. The initial classifier is tested using a test set, and the output classifier that meets the preset accuracy is used as the fault prediction model; the classifier includes one of the following: support vector machine, naive Bayes, temporal neural network RNN ​​or CNN, and long short-term memory network LSTM.

4. The power distribution terminal based on genetic algorithm according to claim 1, characterized in that, The process of determining whether a malfunction will occur is as follows: If the failure probability output by the failure prediction model is greater than or equal to the preset failure probability threshold, it indicates that a failure has occurred. If the fault probability output by the fault prediction model is less than the preset fault probability threshold, it means that no fault has occurred, and the process returns to step 1 to continue collecting real-time status data of each feeder in the distribution network.

5. The power distribution terminal based on genetic algorithm according to claim 1, characterized in that, The process of obtaining faulty equipment combinations based on genetic algorithms is as follows: The equipment on each feeder in the distribution network is mapped to a binary chromosome representation that can be processed by a genetic algorithm. A number of chromosomes are randomly generated to form an initial population. According to a pre-designed fitness function, the fitness value of each chromosome is calculated. The chromosomes with fitness values ​​greater than or equal to a pre-set fitness threshold are selected as parent chromosomes. Crossover is performed on the selected parent chromosomes to generate new offspring chromosomes. Mutation is performed on the offspring chromosomes to increase the diversity of the population. The parent and offspring chromosomes are merged to form a new generation population. It is determined whether the termination condition is met. If it is met, the iteration stops, the chromosome corresponding to the highest fitness value is output, and the binary encoded chromosome is decoded to identify the faulty equipment and obtain the faulty equipment combination. Otherwise, the iteration continues.

6. The power distribution terminal based on genetic algorithm according to claim 1, characterized in that, The fault isolation commands include: circuit breaker control commands, disconnector control commands, sectionalizing switch control commands, and grounding switch control commands.

7. The fault diagnosis method for power distribution terminals based on genetic algorithms as described in any one of claims 1-6, characterized in that, Includes the following steps: Step 1: Collect real-time status data of each feeder in the distribution network and extract fault characteristics; Step 2: Based on the fault features and a pre-trained fault prediction model, generate fault probabilities and determine whether a fault will occur based on these probabilities; Step 3: If a fault occurs, the faulty equipment combination is obtained based on the genetic algorithm to achieve fault location; Step 4: Based on the obtained combination of faulty devices, drive the power distribution network automation control equipment to issue a fault isolation command to achieve fault isolation; Step 5: After fault isolation, the power distribution terminal generates a power restoration plan based on the topology and load conditions of the power distribution network.

Images