A power supply and distribution fault diagnosis analysis method and system for a port device

Abstract

This invention relates to a method and system for diagnosing and analyzing power supply and distribution faults in port equipment. The method includes the following steps: acquiring historical power data for multiple consecutive time periods of the port equipment's power supply network; constructing a model training dataset based on the historical power data; constructing a harmonic classification model using a computer deep learning method and the model training dataset; acquiring real-time power data for the current time period of the port equipment's power supply network; predicting the types of harmonics in the port equipment's power supply network in the next stage by inputting the real-time power data into the harmonic classification model; implementing a harmonic suppression strategy based on the predicted harmonic types; collecting power data after implementing the harmonic suppression strategy; and performing power fault diagnosis and analysis using the power data after implementing the harmonic suppression strategy. This invention improves the accuracy of harmonic prediction; simultaneously, by using power data preprocessed for harmonic analysis, the accuracy of the diagnostic analysis is further improved.

Description

Technical Field

[0001] This invention relates to the field of power supply and distribution fault diagnosis and analysis method and system for port equipment. Background Technology

[0002] Harmonics in a power system originate from electrical equipment, specifically from power generation and consumption equipment. Since the magnetic field generated by the rotor of a generator cannot be a perfect sine wave, the voltage waveform emitted by the generator cannot be a completely distortion-free sine wave. When a sinusoidal voltage is applied to a nonlinear load, the fundamental current is distorted, generating harmonics.

[0003] Meanwhile, power harmonics can also affect the accuracy of power metering equipment and electrical testing equipment, leading to inaccurate diagnostic results in some power diagnostic and analysis systems. Summary of the Invention

[0004] To address the technical problems in existing technologies, such as the reduced accuracy of power metering and electrical testing equipment due to the influence of power harmonics, this invention provides a method and system for diagnosing and analyzing power supply and distribution faults in port equipment.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] A method for diagnosing and analyzing power supply and distribution faults in port equipment includes the following steps:

[0007] Acquire historical power data for multiple consecutive time periods from the port equipment power supply network;

[0008] A training dataset for the model was constructed based on the historical power data.

[0009] A harmonic classification model is constructed based on computer deep learning methods and using the model training dataset.

[0010] Obtain real-time power data of the port equipment power supply network for the current time period;

[0011] By inputting the real-time power data into the harmonic classification model, the types of harmonics in the port equipment power supply network in the next stage are predicted.

[0012] Based on the predicted harmonic types, implement harmonic suppression strategies;

[0013] Collect power data after implementing harmonic suppression strategies;

[0014] Power fault diagnosis and analysis are performed using the power data after implementing the harmonic suppression strategy.

[0015] The beneficial effects of this invention are as follows: by using computer deep learning methods to construct a computer model, using the computer model to predict stage harmonics, and then using harmonic suppression technology to perform harmonic suppression processing at this stage, the accuracy of harmonic prediction is improved by preprocessing the harmonics; at the same time, the accuracy of the diagnostic analysis is improved by using the power data after harmonic preprocessing to perform diagnostic analysis of the power system.

[0016] Furthermore, historical power data for multiple consecutive time periods of the port equipment power supply network is obtained, including the following steps:

[0017] Obtain multiple historical datasets of the port equipment power supply network; wherein each historical dataset is historical data of the port equipment power supply network within one day;

[0018] The data in each of the historical datasets is divided into multiple sub-datasets; wherein, the multiple sub-datasets correspond to a collection of data within multiple consecutive time periods within a day; the historical power data includes multiple sub-datasets.

[0019] Furthermore, a model training dataset is constructed based on the historical electricity data, including the following steps:

[0020] Data from the same time period on the same day in multiple historical datasets are integrated into one dataset to obtain multiple mixed datasets; the model training dataset includes multiple mixed datasets.

[0021] Furthermore, based on computer deep learning methods and using the model training dataset, a harmonic classification model is constructed, including the following steps:

[0022] Construct a convolutional neural network model;

[0023] The convolutional neural network model is trained using data from the model training dataset to obtain the trained convolutional neural network model;

[0024] The trained convolutional neural network model is transferred to learn using the knowledge distillation method to obtain the harmonic classification model.

[0025] Furthermore, the historical power data includes historical harmonic voltage and historical harmonic current; the real-time power data includes real-time harmonic voltage and real-time harmonic current.

[0026] Furthermore, predicting the harmonic types of the port equipment power supply network in the next stage by inputting the real-time power data into the harmonic classification model includes the following steps:

[0027] The real-time power data is input into the harmonic classification model to obtain the frequency of the harmonics in the next stage.

[0028] When there is only one frequency of the harmonic in the next stage, the frequency of the harmonic is output.

[0029] When there are multiple harmonic frequencies in the next stage, the harmonic frequencies with frequencies lower than a preset threshold will be filtered out, and one or more of the remaining harmonic frequencies will be output.

[0030] Furthermore, based on the predicted harmonic types, a harmonic suppression strategy is implemented, including the following steps:

[0031] When there is only one frequency of the output harmonic, a harmonic suppressor is used to suppress the harmonic.

[0032] When there are multiple frequencies of the output harmonics, multiple harmonic suppressors are used to suppress the harmonics.

[0033] To address the aforementioned technical problems, the present invention also provides a power supply and distribution fault diagnosis and analysis system for port equipment, the specific technical content of which is as follows:

[0034] A power supply and distribution fault diagnosis and analysis system for port equipment includes a data acquisition device, a data analysis device, and a harmonic suppression device;

[0035] The data acquisition device is used to acquire historical power data of the port equipment power supply network for multiple consecutive time periods;

[0036] The data analysis device is used to construct a model training dataset based on the historical power data; and to construct a harmonic classification model based on computer deep learning methods and using the model training dataset.

[0037] The data acquisition device is also used to acquire real-time power data of the port equipment power supply network for the current time period;

[0038] The data analysis device is also used to predict the type of harmonics in the port equipment power supply network in the next stage by inputting the real-time power data into the harmonic classification model.

[0039] Harmonic suppression devices are used to implement harmonic suppression strategies based on the predicted harmonic types.

[0040] The data acquisition device is also used to collect power data after the implementation of the harmonic suppression strategy;

[0041] The data analysis device is also used to perform power fault diagnosis analysis using the power data after implementing the harmonic suppression strategy. Attached Figure Description

[0042] Figure 1This is a flowchart of a power supply and distribution fault diagnosis and analysis method for port equipment according to an embodiment of the present invention;

[0043] Figure 2 This is a schematic diagram of the structure of a power supply and distribution fault diagnosis and analysis system for port equipment according to an embodiment of the present invention. Detailed Implementation

[0044] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0045] like Figure 1 As shown in the figure, this embodiment provides a method for diagnosing and analyzing power supply and distribution faults in port equipment, including the following steps:

[0046] S1. Obtain historical power data for multiple consecutive time periods from the port equipment power supply network;

[0047] Obtaining historical power data for multiple consecutive time periods from the port equipment power supply network includes the following steps:

[0048] S101. Obtain multiple historical datasets of the port equipment power supply network; wherein, each historical dataset is the historical data of the port equipment power supply network within one day; the historical data or historical power data specifically includes harmonic voltage and / or harmonic current;

[0049] S102. Divide the data in each of the historical datasets into multiple sub-datasets; wherein, the multiple sub-datasets correspond to a collection of data within multiple consecutive time periods within a day; the historical power data includes multiple sub-datasets. In this embodiment of the invention, the time period can be different lengths such as one hour, two hours, or three hours, but the time period in this invention cannot exceed twelve hours.

[0050] S2. Construct a model training dataset based on the historical power data; specifically, integrate the data from the sub-datasets within the same time period on the same day in multiple historical datasets into a single dataset to obtain multiple mixed datasets; the model training dataset includes multiple mixed datasets.

[0051] S3. Construct a harmonic classification model based on computer deep learning methods and using the model training dataset;

[0052] The harmonic classification model is constructed using computer deep learning methods and the training dataset of the model, including the following steps:

[0053] S301. Construct a convolutional neural network model;

[0054] S302. Train the convolutional neural network model using the data in the model training dataset to obtain the trained convolutional neural network model;

[0055] S303. Based on the knowledge distillation method, the trained convolutional neural network model is transferred to obtain the harmonic classification model.

[0056] S4. Obtain the real-time power data of the port equipment power supply network for the current time period; the historical power data includes historical harmonic voltage and historical harmonic current; the real-time power data includes real-time harmonic voltage and real-time harmonic current.

[0057] S5. Predict the types of harmonics in the port equipment power supply network in the next stage by inputting the real-time power data into the harmonic classification model.

[0058] Predicting the harmonic types of the port equipment power supply network in the next stage by inputting the real-time power data into the harmonic classification model includes the following steps:

[0059] S501. Input the real-time power data into the harmonic classification model to obtain the frequency of the harmonics in the next stage;

[0060] S502. When there is only one frequency of the harmonic in the next stage, the frequency of the harmonic is output.

[0061] S503. When there are multiple frequencies of the harmonics in the next stage, the harmonics with frequencies less than a preset threshold are filtered out, and one or more of the remaining harmonic frequencies are output.

[0062] S6. Implement harmonic suppression strategies based on the predicted harmonic types.

[0063] Based on the predicted harmonic types, a harmonic suppression strategy is implemented, including the following steps:

[0064] S601. When there is only one frequency of the output harmonic, a harmonic suppressor is used to suppress the harmonic.

[0065] S602. When there are multiple frequencies of the output harmonics, multiple harmonic suppressors are used to suppress the harmonics.

[0066] S7. Collect power data after implementing harmonic suppression strategies; for example, when smart meters or smart power monitoring devices collect current and voltage data, they can collect power data from the power grid after implementing harmonic suppression strategies.

[0067] S8. Perform power fault diagnosis and analysis using the power data after implementing the harmonic suppression strategy. Utilize current and voltage data collected by smart meters or smart power monitoring devices to perform power fault analysis. For example, when voltage or current surges occur, analyze and display the corresponding line fault conditions.

[0068] like Figure 2 As shown, in some embodiments, a power supply and distribution fault diagnosis and analysis system for port equipment includes a data acquisition device, a data analysis device, and a harmonic suppression device; wherein, the data acquisition device can preferably be a voltage and current meter or a power quality detection instrument, the data analysis device can be a computer connected to the data acquisition device via a network, and the harmonic suppression device can be an LC filter or a pulse carrier modulator, etc.

[0069] The data acquisition device is used to acquire historical power data for multiple consecutive time periods of the port equipment power supply network; the port equipment power supply network is... Figure 2 China Port Power Supply System.

[0070] The data analysis device is used to construct a model training dataset based on the historical power data and to construct a harmonic classification model based on computer deep learning methods.

[0071] The data acquisition device is also used to acquire real-time power data of the port equipment power supply network for the current time period.

[0072] The data analysis device is also used to predict the type of harmonics in the port equipment power supply network in the next stage by inputting the real-time power data into the harmonic classification model.

[0073] Harmonic suppression devices are used to implement harmonic suppression strategies based on the predicted harmonic types.

[0074] The data acquisition device is also used to collect power data after the implementation of the harmonic suppression strategy.

[0075] The data analysis device is also used to perform power fault diagnosis analysis using the power data after implementing the harmonic suppression strategy.

[0076] This invention utilizes deep learning to construct a computer model, uses the computer model to predict phased harmonics, and then uses harmonic suppression technology to suppress harmonics at this stage, thus achieving harmonic preprocessing. At the same time, it uses the power data after harmonic preprocessing to perform diagnostic analysis of the power system, thereby improving the accuracy of the diagnostic analysis.

[0077] 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 concept and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for diagnosing and analyzing power supply and distribution faults in port equipment, characterized in that, Includes the following steps: Acquire historical power data for multiple consecutive time periods from the port equipment power supply network; A training dataset for the model was constructed based on the historical power data. A harmonic classification model is constructed based on computer deep learning methods and using the model training dataset. Obtain real-time power data of the port equipment power supply network for the current time period; By inputting the real-time power data into the harmonic classification model, the types of harmonics in the port equipment power supply network in the next stage are predicted. Based on the predicted harmonic types, implement harmonic suppression strategies; Collect power data after implementing harmonic suppression strategies; Power fault diagnosis and analysis are performed using the power data after implementing the harmonic suppression strategy. Predicting the harmonic types of the port equipment power supply network in the next stage by inputting the real-time power data into the harmonic classification model includes the following steps: The real-time power data is input into the harmonic classification model to obtain the frequency of the harmonics in the next stage. When there is only one frequency of the harmonic in the next stage, the frequency of the harmonic is output. When there are multiple harmonic frequencies in the next stage, the harmonic frequencies with frequencies lower than a preset threshold will be filtered out, and one or more of the remaining harmonic frequencies will be output. Based on the predicted harmonic types, a harmonic suppression strategy is implemented, including the following steps: When there is only one frequency of the output harmonic, a harmonic suppressor is used to suppress the harmonic. When there are multiple frequencies of the output harmonics, multiple harmonic suppressors are used to suppress the harmonics.

2. The method for diagnosing and analyzing power supply and distribution faults in port equipment according to claim 1, characterized in that, Obtaining historical power data for multiple consecutive time periods from the port equipment power supply network includes the following steps: Obtain multiple historical datasets of the port equipment power supply network; wherein each historical dataset is historical data of the port equipment power supply network within one day; The data in each of the historical datasets is divided into multiple sub-datasets; wherein, the multiple sub-datasets correspond to a collection of data within multiple consecutive time periods within a day; the historical power data includes multiple sub-datasets.

3. The method for diagnosing and analyzing power supply and distribution faults in port equipment according to claim 2, comprising constructing a model training dataset based on the historical power data, includes the following steps: Data from the sub-datasets within the same time period on the same day in multiple historical datasets are integrated into a single dataset to obtain multiple hybrid datasets; the model training dataset includes multiple hybrid datasets.

4. The method for diagnosing and analyzing power supply and distribution faults in port equipment according to claim 2, characterized in that, The harmonic classification model is constructed using computer deep learning methods and the training dataset of the model, including the following steps: Construct a convolutional neural network model; The convolutional neural network model is trained using data from the model training dataset to obtain the trained convolutional neural network model; The trained convolutional neural network model is transferred to learn using the knowledge distillation method to obtain the harmonic classification model.

5. The method for diagnosing and analyzing power supply and distribution faults in port equipment according to claim 1, characterized in that, The historical power data includes historical harmonic voltage and historical harmonic current; the real-time power data includes real-time harmonic voltage and real-time harmonic current.

6. A power supply and distribution fault diagnosis and analysis system for port equipment, characterized in that, Includes data acquisition devices, data analysis devices, and harmonic suppression devices; The data acquisition device is used to acquire historical power data of the port equipment power supply network for multiple consecutive time periods; The data analysis device is used to construct a model training dataset based on the historical power data. A harmonic classification model is constructed based on computer deep learning methods and using the model training dataset. The data acquisition device is also used to acquire real-time power data of the port equipment power supply network for the current time period; The data analysis device is also used to predict the type of harmonics in the port equipment power supply network in the next stage by inputting the real-time power data into the harmonic classification model. Harmonic suppression devices are used to implement harmonic suppression strategies based on the predicted harmonic types. The data acquisition device is also used to collect power data after the implementation of the harmonic suppression strategy; The data analysis device is also used to perform power fault diagnosis analysis using the power data after implementing the harmonic suppression strategy. Predicting the harmonic types of the port equipment power supply network in the next stage by inputting the real-time power data into the harmonic classification model includes the following steps: The real-time power data is input into the harmonic classification model to obtain the frequency of the harmonics in the next stage. When there is only one frequency of the harmonic in the next stage, the frequency of the harmonic is output. When there are multiple harmonic frequencies in the next stage, the harmonic frequencies with frequencies lower than a preset threshold will be filtered out, and one or more of the remaining harmonic frequencies will be output. Based on the predicted harmonic types, a harmonic suppression strategy is implemented, including the following steps: When there is only one frequency of the output harmonic, a harmonic suppressor is used to suppress the harmonic. When there are multiple frequencies of the output harmonics, multiple harmonic suppressors are used to suppress the harmonics.

Images