Transformer active protection fault characteristic quantity extraction and wave recording method and device

Abstract

The invention discloses a transformer active protection fault characteristic quantity extraction and wave recording method and device, and the method comprises the steps: collecting transformer data, converting the transformer data into digital sampling values, and organizing the digital sampling values into a sampling message; analyzing the sampling message in real time, extracting sampling data from the sampling message, and performing dynamic null drift correction on the sampling data; detecting whether a discharge fault occurs or not according to the sampling data after the dynamic null drift correction; and when a discharge fault is detected, performing fault characteristic quantity extraction on the sampling data, and determining fault recording data according to the extracted fault characteristic value data. Real-time preprocessing and fault characteristic quantity extraction can be carried out on the physical quantity information data of the transformer, the sampling value data volume input by a post-stage protection algorithm is reduced, and the data communication bandwidth and the data processing capability of a protection device are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of relay protection, and particularly relates to a method and device for extracting fault characteristic quantities and recording waves of transformer active protection. Background Art

[0002] As one of the most important components in the power system, a transformer is a key device for power transmission and voltage transformation, undertaking the task of transforming electrical energy. Its stable operation directly affects the continuity and stability of system power supply. For extra-high voltage large transformers, due to their high field strength, rapid fault development and great destructiveness, once a fault occurs, such as a deflagration accident caused by arc discharge, it will not only cause huge economic losses, but may even lead to casualties. Ensuring the safe and stable operation of these core devices has become an important issue in the power industry.

[0003] Existing transformer on-line monitoring technologies focus on detecting early defects, judging fault conditions by measuring the change trends and statistical laws of characteristic quantities, with slow response speed and difficulty in timely warning of serious defects that develop rapidly; differential protection relies on the large current characteristics at the transformer ends, and even if the protection acts quickly, the accident of oil tank bursting into flames cannot be completely avoided; light gas protection relies on the volume of accumulated gas for discrimination, the gas production and gas accumulation processes under discharge are slow, and are easily affected by non-fault factors such as trapped gas. Even if the strategy is changed from alarm to tripping, deflagration is still inevitable, and there is a risk of misoperation.

[0004] To solve the above problems, transformer active protection that reflects the rapid development of serious discharge defects has become a research hotspot. This method monitors the changes in the characteristics of multiple physical field parameters such as ultra-high frequency, high-frequency pulse current, and ultrasound in real time, and cuts off the transformer before fault breakdown to reduce the damage degree. This transformation not only improves the performance of the relay protection system, but also effectively prevents the possibility of the transformer evolving from a serious defect to a malignant fault, and improves the level of coordinated and safe operation of the transformer and the power grid. However, this also poses higher requirements for the processing performance of the transformer active protection system. To achieve real-time monitoring and rapid response, the data acquisition system must have high bandwidth and low latency characteristics, and at the same time, the data processing algorithm needs to be efficient and accurate to ensure that useful fault information can be quickly extracted from a large amount of data. In addition, powerful computing resources are required to support complex multi-dimensional data analysis to ensure that the protection system can make correct decisions in the shortest time, thereby maximizing the safety and stable operation of the transformer.

[0005] The prior art document 1 (CN202110207379) discloses a method and device for active protection of large power transformers. The prior art document 2 (CN202311580314) discloses a method and device for active protection of transformers based on multi - parameter protection analysis. Both of these methods collect and obtain several physical quantity information of the transformer through sensors. The physical quantity information includes: UHF, pulse current and ultrasonic signal of the power transformer. A criterion is constructed based on the above physical quantity information, and active protection of the transformer is carried out according to the criterion.

[0006] However, the disadvantages of the prior art document 1 and the prior art document 2 are both that the amount of original collected data they generate is extremely large, and there are many application parameters involved. The effect of the subsequent protection algorithm will be greatly reduced, which poses extremely high requirements for the subsequent data - processing ability of the protection host. For traditional architecture protection devices with a processor and DSP as the data - processing center, due to the limitations of data - transmission bandwidth, data - processing computing power, and data - storage methods, it is difficult to meet the requirements of the "four properties" of protection (selectivity, rapidity, sensitivity, and reliability), and at the same time, it is impossible to ensure the integrity and accuracy of fault data recording. Summary of the Invention

[0007] To solve the deficiencies in the prior art, the present invention provides a method and device for extracting fault characteristic quantities and recording waves for active protection of transformers. It adopts a software - hardware collaborative processing hybrid architecture platform based on FPGA (Field Programmable Gate Array) and a processor to perform real - time pre - processing and extraction of fault characteristic quantities on the physical quantity information data of the transformer, reducing the amount of sampled - value data input to the subsequent protection algorithm, thereby improving the data - communication bandwidth and data - processing ability of the protection device, and ensuring the accurate execution of the active protection algorithm and the "four properties" of protection.

[0008] The present invention adopts the following technical solutions.

[0009] The first aspect of the present invention provides a method for extracting fault characteristic quantities and recording waves for active protection of transformers, including:

[0010] S1: Collect transformer data, convert the transformer data into digital sampled values, and organize the digital sampled values into sampling messages;

[0011] S2: Parse the sampling messages in S1 in real - time, extract sampling data from the sampling messages, and perform dynamic zero - drift correction on the sampling data;

[0012] S3: Detect whether a discharge fault occurs according to the sampling data after dynamic zero - drift correction in S2;

[0013] S4: When a discharge fault is detected, extract the fault feature quantities from the sampled data after dynamic zero-drift correction in S2, and determine the fault recording data based on the extracted fault feature value data and perform real-time recording.

[0014] Optionally, the dynamic zero-drift correction of the sampled data in S2 is performed according to the following formula:

[0015] f rev (i) = ε(i) - ZeroDrift,

[0016] where f rev (i) is the i-th sampled data after zero-drift correction, f(i) is the i-th sampled data, and ZeroDrift is the zero-drift value.

[0017] Optionally, the zero-drift value ZeroDrift is calculated according to the following formula:

[0018] ZeroDrift = Avg N + α * Δ,

[0019] where ZeroDrift is the zero-drift value, Δ is the difference between the average value of the sampled data within the preset time of this stage and the average value of the sampled data within the preset time of the previous stage, α is the correction coefficient, and Avg N is the average value of the sampled data within the preset time of this stage;

[0020]

[0021] where f(i) is the i-th sampled data, N is the number of sampled data, and Avg N is the average value of the sampled data within the preset time of this stage;

[0022] When the preset time of this stage is the preset time of the first stage or the preset time of the second stage, Δ = 0.

[0023] Optionally, when |Δ| < 0.5% Avg N , α = 0.8 is adopted; and / or,

[0024] when 0.5% Avg N ≤ |Δ| ≤ 2% Avg N , α = 1.0 is adopted; and / or,

[0025] when |Δ| > 2% Avg N , α = 0.0 is adopted.

[0026] Optionally, in S3, detecting whether a discharge fault occurs based on the sampled data after dynamic zero-drift correction includes:

[0027] Compare and judge the sampled data after dynamic zero-drift correction with the discharge start threshold value;

[0028] If the sampled data is greater than the discharge start threshold value, it is determined that the discharge fault starts due to over-limit, and record the over-limit start time;

[0029] Compare the sampled data with the discharge return threshold value. If the sampled data is less than the discharge return threshold value, it is determined that the discharge fault ends, and record the discharge fault end time.

[0030] Optionally, when a discharge fault is detected in S4, extract the fault characteristic quantities from the sampled data, including:

[0031] Determine the discharge fault duration according to the over-limit start time and the discharge fault end time;

[0032] Extract the maximum sampled value of the fault and the fault noise energy from the sampled data;

[0033] Form the fault characteristic value data with the over-limit start time, the discharge fault duration, the maximum sampled value of the fault and the fault noise energy.

[0034] Optionally, extract the fault noise energy according to the following formula:

[0035]

[0036] Among them, Noise is the sampling data noise energy, f rev (i) is the i-th sampled data after zero-drift correction, and N is the number of sampled data.

[0037] The second aspect of the present invention provides a device for extracting fault characteristic quantities and recording waves for transformer active protection. The device includes:

[0038] An intelligent sensing unit for collecting transformer data and generating a sampling message from the collected transformer data;

[0039] A calculation unit connected to the intelligent sensing unit through an optical fiber for extracting fault characteristic quantities from the sampling message and recording wave control;

[0040] A protection unit connected to the calculation unit for performing transformer active protection according to the channel characteristic value data formed after the calculation unit extracts the fault characteristic quantities and the protection algorithm;

[0041] A recording wave storage unit connected to the calculation unit for storing the channel fault recording wave data obtained by the calculation unit's recording wave control;

[0042] The calculation unit includes:

[0043] A message parsing module for real-time parsing of sampling messages and extracting sampling data from the sampling messages;

[0044] A correction module for dynamically correcting the zero drift of the sampling data;

[0045] A fault detection module for detecting whether a discharge fault occurs according to the sampling data after dynamic zero drift correction; A feature quantity extraction module for extracting fault feature quantities from the sampling data when a discharge fault is detected;

[0046] A waveform recording control module for determining fault waveform recording data according to the extracted fault characteristic value data and sending the fault waveform recording data to a waveform recording storage unit.

[0047] Optionally, the device adopts a collaborative processing architecture platform based on FPGA and CPU. Among them, the CPU is set in the protection unit to execute protection algorithms, and the FPGA is set in the intelligent sensing unit, the computing unit and the waveform recording storage unit to execute the tasks corresponding to the intelligent sensing unit, the computing unit and the waveform recording storage unit.

[0048] Optionally, the waveform recording storage unit includes a DDR storage module and an SSD storage module; after receiving the fault waveform recording data of multiple computing units, the waveform recording storage unit writes the fault waveform recording data into the DDR storage module according to channel partitioning, and writes the fault waveform recording data in the DDR storage module into the SSD storage module.

[0049] The third aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is loaded into the processor, the method for extracting fault feature quantities and recording waveforms of the above-mentioned transformer active protection is implemented.

[0050] The fourth aspect of the present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the method for extracting fault feature quantities and recording waveforms of the above-mentioned transformer active protection is implemented.

[0051] Compared with the prior art, the beneficial effects of the present invention at least include:

[0052] The present invention can improve the accuracy of sampling data by dynamically correcting the zero drift of the sampled data, and extract features from the original acquisition data before the subsequent protection algorithm, which can further reduce the amount of data processed by the protection algorithm and improve the processing speed of the system.

[0053] In the present invention, heterogeneous collaborative computing is achieved between an FPGA and a CPU processor. The FPGA is responsible for sampling data network transmission, sampling message parsing, sampling data preprocessing, discharge fault detection, extraction of discharge fault characteristic values, and discharge fault data recording. The CPU processor focuses on the implementation of protection algorithms and the operation of the host management system. By reasonably allocating the above tasks to the FPGA and the CPU processor, criteria are constructed for the protection algorithm by extracting fault characteristic quantities, and the integrity of fault data recording is ensured through secondary storage of fault recording data. This hybrid architecture with dual-core processing enhances the data processing ability and communication bandwidth efficiency of the system, ensures the accurate execution of the active protection algorithm and the "four properties" of protection, and guarantees the integrity and accuracy of fault data recording. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings. Among them:

[0055] Figure 1 is a schematic diagram of the principle of a device for extracting fault characteristic quantities and recording waveforms for active protection of a transformer provided by an embodiment of the present invention;

[0056] Figure 2 is a schematic diagram of the principle of the function of a computing unit provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in this application are only a part of the embodiments of the present invention, not all of them. Based on the spirit of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present invention.

[0058] Embodiment 1 of the present invention provides a method for extracting fault characteristic quantities and recording waveforms for active protection of a transformer, which adopts a software and hardware collaborative processing hybrid architecture platform based on an FPGA and a processor. The method includes the following steps:

[0059] S1: Collect transformer data, convert the transformer data into digital sampling values, and organize the digital sampling values into sampling messages.

[0060] Specifically, the transformer data includes high-frequency pulse current signal quantities, ultra-high frequency signal quantities, and ultrasonic signal quantities.

[0061] Install various intelligent sensing units at relevant positions on the transformer body to collect various parameters such as high-frequency pulse current, UHF, and ultrasonic. The FPGA of the intelligent sensing unit caches the digital sampling values in the internal RAM.

[0062] Optionally, when the cached digital sampling values reach the set number of points, the FPGA of the intelligent sensing unit organizes the header field, sampling values, and CRC check into a sampling message. The header field includes transmission auxiliary information such as address, sequence number, type, etc., as well as monitoring data such as temperature, power supply, reference voltage, and clock status.

[0063] S2: Parse the sampling message in real time, extract the sampling data from the sampling message, and perform dynamic zero-drift correction on the sampling data.

[0064] Specifically, S2 specifically includes:

[0065] S2.1: The FPGA of the calculation unit parses the received sampling message in real time and extracts the sampling data from the sampling message.

[0066] Specifically, the FPGA of the calculation unit integrates multiple Ethernet interface modules, can receive sampling messages from multiple intelligent sensing units simultaneously, parse the sampling messages in real time, extract the sampling data, and obtain status information such as temperature, power supply, and reference voltage.

[0067] S2.2: The FPGA of the calculation unit performs dynamic zero-drift correction on the extracted sampling data. The zero-drift correction is used to correct the DC component in the sampling loop, calculate the noise energy of the sampling data after zero-drift correction, and simultaneously count the maximum sampling value.

[0068] For the sampling data parsed by the FPGA of the calculation unit, first perform zero-drift correction to correct the DC component in the sampling loop. The specific process of zero-drift correction: The FPGA of the calculation unit calculates the cumulative sum of the signed sampling values over a period of time and calculates the average value; compares the average value with the set mean threshold. If it is less than the set mean threshold, it is considered that the average value calculated this time is valid and used as the zero-drift value for the next stage; if it is greater than the set mean threshold, it is considered that the average value calculated this time is invalid and the original zero-drift value is continued to be used. The process of zero-drift correction is to subtract the zero-drift value from the sampling value.

[0069] The dynamic zero-drift correction of the sampling data in S2 is performed according to the following formula:

[0070] f rev (i) = f(i) - ZeroDrift,

[0071] where, f rev (i) is the i-th sampling data after zero-drift correction, f(i) is the i-th sampling data, and ZeroDrift is the zero-drift value.

[0072] Optionally, the zero drift value ZeroDrift is calculated according to the following formula:

[0073] ZeroDrift = Avg N + α * Δ,

[0074] where ZeroDrift is the zero drift value, Δ is the difference between the average value of the sampled data within the preset time of this stage and the average value of the sampled data within the preset time of the previous stage, α is the correction coefficient, and Avg N is the average value of the sampled data within the preset time of this stage;

[0075]

[0076] where f(i) is the i-th sampled data, n is the number of sampled data, and Avg N is the average value of the sampled data within the preset time of this stage;

[0077] When the preset time of this stage is the preset time of the first stage or the preset time of the second stage, Δ = 0.

[0078] Optionally, when |Δ| < 0.5% Avg N , α = 0.8 is adopted; and / or when 0.5% Avg N ≤ |Δ| ≤ 2% Avg N , α = 1.0 is adopted; and / or when |Δ| > 2% Avg N , α = 0.0 is adopted.

[0079] Specifically, when |Δ| < 0.5% Avg N , it is the low drift interval, and a conservative correction coefficient of α = 0.8 is adopted; when 0.5% Avg N ≤ |Δ| ≤ 2% Avg N , it is the conventional correction interval, and a complete correction coefficient of α = 1.0 is adopted; when |Δ| > 2% Avg N , it is the abnormal detection interval, judged as a sensor failure, the original zero drift value is kept unchanged, and α = 0.0 is adopted.

[0080] Optionally, the correction coefficient α is calculated according to the following formula:

[0081]

[0082] where Δ is the difference between the average value of the sampled data within the preset time of this stage and the average value of the sampled data within the preset time of the previous stage, α is the correction coefficient, and Avg N is the average value of the sampled data within the preset time of this stage.

[0083] In this way, within the regular correction range, the correction parameters transition smoothly, which can reduce system instability or abnormal data caused by parameter mutations, thereby improving the correctness of the data and providing strong data support for subsequent protection algorithms.

[0084] It can be understood that those skilled in the art can modify the specific values of the low-drift range, regular correction range, anomaly detection range, and α according to actual application requirements. This embodiment only gives an exemplary method.

[0085] In this embodiment, by dynamically zero-drift correcting the sampled data, the influence of environmental temperature, sensor aging, power supply fluctuations, etc. on the accuracy of the sampled data is reduced, the smooth transition between the old and new zero-drift values is achieved, over-correction is suppressed, the accuracy of the sampled data is improved, and further the accuracy and integrity of the subsequent fault recording data are improved.

[0086] S2.3: The computing unit FPGA writes the zero-drift corrected sampled data into the internal cache for reading during subsequent fault recording.

[0087] Optionally, the cache is a RAM cache.

[0088] Specifically, the zero-drift corrected sampled data is written into the RAM cache inside the computing unit FPGA. The RAM address accumulates circularly to store the sampled data cyclically, and the RAM capacity can store the sampled data before and after the fault point. In this way, no matter when the data is read, the sampled data of a previous period can be traced, which is convenient for subsequent fault recording.

[0089] S3: Detect whether a discharge fault occurs according to the sampled data after dynamic zero-drift correction. S3 specifically includes:

[0090] S3.1: The computing unit FPGA compares and judges the sampled data with the discharge start threshold value in real time. If the sampled data is greater than the discharge start threshold value, it is determined that the discharge fault has exceeded the limit and started, and the time of over-limit start is recorded.

[0091] Each sampling channel has a discharge start threshold value ε configured by the application program pulse , and the discharge start threshold value is the absolute value of the sampled value. The computing unit FPGA compares and judges the absolute value of the sampled value with the start threshold value in real time. If it is greater than the start threshold value, it is determined that the discharge fault has exceeded the limit and started, and the time of over-limit start is recorded.

[0092] |f(i)| > ε pulse ,

[0093] where f(i) is the i-th sampled data, and ε pulseis the discharge start threshold value. S3.2: The computing unit FPGA compares and judges the sampled data and the discharge return threshold value in real time. If the sampled data is less than the discharge return threshold value, it is determined that the discharge fault ends, and the end time of the discharge fault is recorded.

[0094] Each sampling channel has a discharge return threshold value ε configured by the application program return , and the return threshold value is the absolute value of the sampled value. The computing unit FPGA compares and judges the absolute value of the sampled value and the return threshold value in real time. If it is less than the return threshold value, it is determined that the discharge fault ends.

[0095] |f(i)|>ε return ,

[0096] where f(i) is the i-th sampled data, and ε return is the discharge return threshold value.

[0097] S4: When a discharge fault is detected, extract the fault characteristic quantity from the sampled data, and determine the fault recording data according to the extracted fault characteristic value data. S4 specifically includes:

[0098] S4.1: Determine the discharge fault duration according to the over-limit start time and the end time of the discharge fault.

[0099] S4.2: Extract the maximum fault sampled value and the fault noise energy from the sampled data.

[0100] S4.3: Compose the over-limit start time, the discharge fault duration, the maximum fault sampled value and the fault noise energy into the fault characteristic value data.

[0101] Specifically, the computing unit FPGA calculates and records the fault characteristic value data, and the fault characteristic value data includes the over-limit start time, the discharge fault duration, the maximum fault sampled value, and the fault noise energy during the discharge fault duration.

[0102] From the start of the discharge fault over-limit to the end of the discharge fault, this period is the discharge fault duration. The computing unit FPGA calculates and records the over-limit start time, the discharge fault duration, the maximum fault sampled value, and the fault noise energy during this period as the fault characteristic value record.

[0103] The computing unit FPGA sends the characteristic value message to the protection unit at a fixed period. The characteristic value message includes the sampling zero drift, the noise energy, and the sampling maximum value information of each sampling channel; when a discharge fault is detected, the characteristic value message also includes the fault characteristic value data.

[0104] Optionally, the computing unit FPGA calculates the noise energy for the sampled data after zero drift correction:

[0105]

[0106] Among them, Noise is the noise energy of the sampled data, f rev (i) is the sampled data after the i-th zero-drift correction, and N is the number of sampled data.

[0107] Optionally, the computing unit FPGA performs real-time comparison on the sampled data after zero-drift correction and statistically obtains the maximum sampled value:

[0108] f max = Max(|f(i)|),

[0109] Among them, f max is the maximum sampled value, and f(i) is the i-th sampled data.

[0110] The computing unit FPGA sends eigenvalue messages to the protection unit at a fixed period. When no discharge fault is detected, the eigenvalue messages only contain information such as the sampling zero-drift, noise energy, and sampling maximum value of each medium- and high-speed sampling channel; when a discharge fault is detected, in addition to the sampling zero-drift, noise energy, and sampling maximum value, the eigenvalue messages also include the over-limit start time during the fault of each channel, the discharge fault duration, the maximum sampled value of the fault, and the fault noise energy, which are these fault eigenvalue records. The protection unit receives the eigenvalue periodic message and uses it as the input quantity for the transformer active protection logic judgment.

[0111] When a discharge fault is detected, the computing unit FPGA reads the sampled data before the fault, during the fault duration, and after the fault from the circular sampled value buffer RAM, and organizes these data into an Ethernet oscillogram message and sends it to the oscillogram storage unit. The source address segment of the oscillogram message contains the sampling channel number. If the data volume of a single discharge fault is large and exceeds the maximum length of one frame of Ethernet, the FPGA will perform packet-sending processing.

[0112] In this embodiment, through real-time analysis of the collected data, fault eigenvalue extraction is performed in real time and online, and fault oscillogram recording is synchronized, which improves the real-time performance of the system compared with traditional offline data fault analysis and oscillogram recording.

[0113] Combined with Figure 1 、 Figure 2 As shown, Embodiment 2 of the present invention provides a device for extracting and recording fault characteristic quantities of transformer active protection, including:

[0114] An intelligent sensing unit for collecting transformer data and generating a sampling message from the collected transformer data;

[0115] A computing unit, connected to the intelligent sensing unit through an optical fiber, for extracting fault characteristic quantities from the sampling message and controlling oscillogram recording;

[0116] The protection unit is connected to the calculation unit and performs active protection on the transformer according to the characteristic value data of each channel formed after extracting the fault characteristic quantity of the calculation unit and the protection algorithm.

[0117] The waveform recording and storage unit is connected to the calculation unit and stores the fault waveform recording data of each channel obtained by the waveform recording control of the calculation unit.

[0118] The calculation unit includes:

[0119] The message parsing module is used to parse the sampling message in real time and extract the sampling data from the sampling message.

[0120] The correction module is used to perform dynamic zero drift correction on the sampling data.

[0121] The fault detection module is used to detect whether a discharge fault occurs according to the sampling data after dynamic zero drift correction; the characteristic quantity extraction module is used to extract the fault characteristic quantity from the sampling data when a discharge fault is detected.

[0122] The waveform recording control module is used to determine the fault waveform recording data according to the extracted fault characteristic value data and send the fault waveform recording data to the waveform recording and storage unit.

[0123] Optionally, the device adopts a collaborative processing architecture platform based on FPGA and CPU. Among them, the CPU is set in the protection unit and is used to execute the protection algorithm. The FPGA is set in the intelligent sensing unit, the calculation unit and the waveform recording and storage unit and is used to execute the tasks corresponding to the intelligent sensing unit, the calculation unit and the waveform recording and storage unit.

[0124] Specifically, the intelligent sensing unit includes a sensor, an FPGA of the intelligent sensing unit and an Ethernet sending module. The sensor collects transformer data. The FPGA of the intelligent sensing unit generates a sampling message from the collected transformer data and sends it to the calculation unit through the Ethernet sending module.

[0125] The calculation unit includes an FPGA of the calculation unit. The FPGA of the calculation unit extracts the fault characteristic quantity and controls the waveform recording of the sampling message.

[0126] The waveform recording and storage unit includes an FPGA of the waveform recording and storage unit. The FPGA of the waveform recording and storage unit is used to write the fault waveform recording data of each channel obtained by the waveform recording control of the FPGA of the calculation unit into the corresponding storage area.

[0127] The protection unit includes a CPU of the protection unit. The CPU of the protection unit is used to run the protection algorithm.

[0128] In this embodiment, by constructing a heterogeneous computing system, physical isolation and logical collaboration of sensing acquisition, data transmission, data analysis, and fault recording storage are achieved, improving the system reliability compared with the traditional centralized architecture.

[0129] Optionally, the fault recording storage unit includes a DDR storage module and an SSD storage module; after receiving the fault recording data of multiple computing units, the fault recording storage unit writes the fault recording data into the DDR storage module according to channel partitions, and writes the fault recording data in the DDR storage module into the SSD storage module.

[0130] Specifically, after the FPGA of the fault recording storage unit receives the fault recording messages of multiple computing units, it parses the source address of the fault recording message, obtains the corresponding sampling channel number, differentiates the fault recording messages according to the sampling channel number, and writes them into the fault recording data storage space of the corresponding channel in the DDR storage module; the fault recording storage unit writes the fault recording data in the DDR storage module into the large-capacity SSD storage module for fault recording storage and fault waveform display. In this way, by adopting a two-level storage mode, the integrity of the fault recording data is ensured.

[0131] Embodiment 3 of the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is loaded into the processor, the method described in Embodiment 1 is implemented.

[0132] Embodiment 4 of the present invention provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the method described in Embodiment 1 is implemented.

[0133] Compared with the prior art, the beneficial effects of the present invention at least include:

[0134] By dynamically correcting the zero drift of the sampled data, the present invention can improve the accuracy of the sampled data, and by extracting the features of the original acquired data before the post-stage protection algorithm, the amount of data processed by the protection algorithm can be further reduced, improving the processing speed of the system.

[0135] The FPGA and the CPU processor achieve heterogeneous collaborative computing. The FPGA is responsible for sampling data network transmission, sampling message parsing, sampling data preprocessing, discharge fault detection, extraction of discharge fault characteristic values, and discharge fault data recording; the CPU processor focuses on the implementation of protection algorithms and the operation of the host management system. By reasonably allocating the above tasks to the FPGA and the CPU processor, criteria are constructed for the protection algorithm by extracting fault characteristic quantities, and the integrity of fault data recording is ensured by means of secondary storage of fault recording data. This hybrid architecture with dual-core processing enhances the data processing capacity and communication bandwidth efficiency of the system, ensures the accurate execution of the active protection algorithm and the "four properties" of protection, and guarantees the integrity and accuracy of fault data recording.

[0136] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0137] The present disclosure may be a system, a method, and / or a computer program product. The computer program product may include a computer-readable storage medium having thereon computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0138] A computer-readable storage medium may be a tangible device that can retain and store instructions for use by an instruction execution device. A computer-readable storage medium may be, for example, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disc (DVD), a memory stick, a floppy disk, a mechanically encoded device such as a punch card or raised structures in a groove having instructions stored thereon, and any suitable combination of the foregoing. The computer-readable storage medium used herein is not construed as being a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagated through a waveguide or other transmission medium (e.g., an optical pulse through an optical fiber cable), or an electrical signal transmitted through a wire.

[0139] The computer-readable program instructions described herein can be downloaded to various computing / processing devices from a computer-readable storage medium or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network may include copper transmission cables, optical fiber transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing / processing device.

[0140] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state-setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as the "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, by using the state information of the computer-readable program instructions to customize an electronic circuit, such as a programmable logic circuit, a field-programmable gate array (FPGA), or a programmable logic array (PLA), the electronic circuit can execute the computer-readable program instructions to implement various aspects of the present disclosure.

[0141] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: it is still possible to modify the specific implementation manners of the present invention or make equivalent replacements, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A method for extracting and recording fault characteristics of transformer active protection, characterized in that: include: S1: Collect transformer data, convert the transformer data into digital sampling values, and organize the digital sampling values ​​into sampling messages; S2: parse the sampling message in S1 in real time, extract sampling data from the sampling message, and perform dynamic zero drift correction on the sampling data; S3: Detect whether a discharge fault occurs according to the sampling data after dynamic zero drift correction in S2; S4: When a discharge fault is detected, the fault characteristic value is extracted from the sampling data after dynamic zero drift correction in S2, and the fault recording data is determined according to the extracted fault characteristic value data and real-time recording is performed.

2. The method for extracting and recording fault characteristics of active transformer protection according to claim 1, characterized in that: In S2, the sampled data is dynamically corrected for zero drift according to the following formula: f rev (i)=f(i)-ZeroDrift, Among them, f rev (i) is the i-th sampling data after zero drift correction, f(i) is the i-th sampling data, and ZeroDrift is the zero drift value.

3. The method for extracting and recording fault characteristics of active transformer protection according to claim 2, characterized in that: The zero drift value ZeroDrift is calculated according to the following formula: ZeroDrift=Avg N +α*Δ, Among them, ZeroDrift is the zero drift value, Δ is the difference between the average value of the sampling data within the preset time of this stage and the average value of the sampling data within the preset time of the previous stage, α is the correction coefficient, Avg N It is the average value of the sampled data within the preset time of this stage; Among them, f(i) is the i-th sampling data, N is the number of sampling data, Avg N It is the average value of the sampled data within the preset time of this stage; When the preset time of this stage is the preset time of the first stage or the preset time of the second stage, Δ=0.

4. The method for extracting and recording fault characteristics of active transformer protection according to claim 3, characterized in that: When |Δ|<0.5%Avg N When α=0.8 is used; and / or, When 0.5%Avg N ≤|Δ|≤2%Avg N , using α=1.0; and / or, When |Δ|>2%Avg N When α=0.0 is used.

5. The method for extracting and recording fault characteristics of transformer active protection according to claim 1, characterized in that: In S3, the discharge fault is detected based on the sampled data after dynamic zero drift correction, including: Compare and judge the sampling data after dynamic zero drift correction with the discharge start threshold value; If the sampling data is greater than the discharge start threshold, it is determined as a discharge fault over-limit start, and the over-limit start time is recorded; The sampling data is compared with the discharge return threshold value. If the sampling data is less than the discharge return threshold value, it is determined that the discharge fault has ended, and the discharge fault end time is recorded.

6. The method for extracting and recording fault characteristics of active transformer protection according to claim 5, characterized in that: When a discharge fault is detected in S4, fault feature quantity is extracted from the sampled data, including: Determine the duration of the discharge fault according to the over-limit starting time and the discharge fault ending time; Extract the maximum fault sampling value and fault noise energy from the sampling data; The over-limit starting time, discharge fault duration, fault maximum sampling value and fault noise energy are combined into fault characteristic value data.

7. The method for extracting and recording fault characteristics of active transformer protection according to claim 6, characterized in that: The fault noise energy is extracted according to the following formula: Where Noise is the noise energy of the sampled data, f rev (i) is the sampling data after the i-th zero drift correction, and N is the number of sampling data.

8. A device for extracting and recording transformer active protection fault characteristics using the method for extracting and recording transformer active protection fault characteristics as claimed in any one of claims 1 to 7, characterized in that: The device comprises: An intelligent sensor unit is used to collect transformer data and generate sampling messages from the collected transformer data; The computing unit is connected to the intelligent sensing unit through optical fiber to extract fault feature quantities and perform wave recording control on the sampled messages; The protection unit is connected to the calculation unit and performs active protection of the transformer according to the characteristic value data of each channel formed after the fault characteristic quantity is extracted by the calculation unit and the protection algorithm; The recording storage unit is connected to the calculation unit and stores the fault recording data of each channel obtained by the recording control of the calculation unit; The computing unit comprises: The message parsing module is used to parse the sampled message in real time and extract the sampled data from the sampled message; A correction module is used to perform dynamic zero drift correction on the sampled data; A fault detection module is used to detect whether a discharge fault occurs based on the sampled data after dynamic zero drift correction; a feature extraction module is used to extract fault feature quantities from the sampled data when a discharge fault is detected; The recording control module is used to determine the fault recording data according to the extracted fault characteristic value data and send the fault recording data to the recording storage unit.

9. The device for extracting and recording fault characteristics of active transformer protection according to claim 8, characterized in that: The device adopts an FPGA and CPU collaborative processing architecture platform, wherein the CPU is set in the protection unit to execute the protection algorithm, and the FPGA is set in the intelligent sensing unit, the computing unit and the recording storage unit to execute the tasks corresponding to the intelligent sensing unit, the computing unit and the recording storage unit.

10. The device for extracting and recording fault characteristics of active transformer protection according to claim 8, characterized in that: The recording storage unit includes a DDR storage module and an SSD storage module; after receiving the fault recording data from multiple computing units, the recording storage unit writes the fault recording data into the DDR storage module according to channel partitions, and writes the fault recording data in the DDR storage module into the SSD storage module.

11. An electronic device, comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is used to operate according to the instruction to execute the steps of the method for extracting and recording fault characteristics of active transformer protection according to any one of claims 1 to 7.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method for extracting and recording fault characteristics of transformer active protection as described in any one of claims 1 to 7 are implemented.

Images