A sampling control device and system for power systems

Abstract

The application relates to the technical field of electric power, in particular to a sampling control device and system for an electric power system, which comprises a multi-source signal acquisition module, which is used for acquiring voltage analog signals and current analog signals, aligning the time stamps of the voltage analog signals and the current analog signals by using a Beidou and GPS dual-mode time alignment component, analyzing communication message headers to match standard protocols, and merging the voltage analog signals and the current analog signals after alignment. In the application, when the signal-to-noise ratio is low, the compression gate width is automatically increased to filter background noise redundancy, and when the signal-to-noise ratio is high, the gate width is reduced to retain the evolution of small features. The adaptive adjustment strategy can reduce the data transmission bandwidth pressure, maximize the waveform details before and after the fault occurs, filter out abnormal branches exceeding the setting value of the instantaneous trip protection, and correlate the geographic coordinates, so as to replace the traditional manual experience-based troubleshooting mode, and improve the accuracy and response speed of fault analysis.

Description

Technical Field

[0001] This invention relates to the field of power technology, and in particular to a sampling control device and system for power systems. Background Technology

[0002] Existing technologies typically employ a deployment architecture that physically separates ring main units and terminal units. This distributed layout necessitates the use of lengthy external cables for signal transmission between devices. During long-term operation, these external cables are highly susceptible to environmental temperature variations, mechanical vibrations, and electromagnetic interference, leading to poor connector contact or signal attenuation, severely weakening the real-time performance and integrity of data interaction. Furthermore, devices from different manufacturers exhibit significant differences in hardware interface definitions and electrical parameter standards, often resulting in interface incompatibility during on-site installation. This necessitates substantial time for customized conversion and debugging, increasing the complexity and potential risks of project construction. Due to the lack of a unified communication protocol stack, the data formats generated by various secondary devices vary considerably, making it difficult for the main station system to efficiently clean and correlate multi-source data, thus creating information silos between subsystems.

[0003] Meanwhile, when a power grid failure occurs, maintenance personnel must rush to the site to troubleshoot section by section. This passive and inefficient emergency repair mode is severely constrained by traffic and weather, resulting in a longer power restoration time and making it difficult to meet the maintenance requirements of modern smart grids for high-reliability power supply. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and to propose a sampling control device and system for power systems.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a sampling control system for a power system includes:

[0006] The multi-source signal acquisition module is used to acquire voltage analog signals and current analog signals. It uses the Beidou and GPS dual-mode time synchronization components to align the timestamps of the voltage analog signals and current analog signals, parses the communication message header to match the standard protocol, merges the aligned voltage analog signals and current analog signals, and generates a multi-source raw sampling data stream.

[0007] The signal-to-noise ratio (SNR) threshold adjustment module is used to extract the signal amplitude sequence within a sliding time window based on the multi-source raw sampling data stream, calculate the variance of the signal amplitude sequence to estimate the noise floor level, calculate the ratio of the square of the effective signal value to the noise floor level to determine the real-time SNR index, compare the real-time SNR index with a preset steady-state noise baseline, increase the compression gate width parameter when the real-time SNR index is lower than the steady-state noise baseline, and decrease the compression gate width parameter when the real-time SNR index is higher than the steady-state noise baseline, thereby generating an adaptive compression deviation threshold.

[0008] Preferably, the system further includes:

[0009] The trajectory node compression module is used to construct upper and lower boundary slope limits at the starting sampling point of the multi-source original sampling data stream according to the adaptive compression deviation threshold, calculate the slope of the line connecting the subsequent sampling points and the starting sampling point, determine whether the line slope exceeds the range defined by the upper and lower boundary slope limits, mark the previous sampling point as a turning point when the line slope exceeds the range, reset the starting sampling point to the current sampling point and reconstruct the boundary slope limits, record the time coordinates and amplitude coordinates of the turning point, and generate a key trajectory node set.

[0010] The topology anomaly location module is used to map time coordinates and amplitude coordinates to the corresponding node positions in the power grid topology model based on the key trajectory node set, calculate the current change rate and voltage sag amplitude between adjacent nodes in the key trajectory node set, compare the current change rate with the instantaneous overcurrent protection setting value, compare the voltage sag amplitude with the transient recording start threshold, filter branch sections whose comparison results exceed the limit, associate the branch sections with the geographic information system coordinates, and generate a power grid topology anomaly location set.

[0011] Preferably, the multi-source signal acquisition module includes:

[0012] The synchronous trigger acquisition submodule, based on the ARM Cortex-A53 processor and FPGA logic unit, configures the parallel signal acquisition interface, uses the FPGA logic unit to send hardware trigger commands to the analog-to-digital conversion component, reads the voltage analog signal value and the current analog signal value in parallel, calls the Beidou and GPS dual-mode time synchronization component to obtain the current time information, and attaches the time information as a timestamp tag to the voltage analog signal value and the current analog signal value to generate a time-aligned analog quantity set.

[0013] The protocol message parsing submodule extracts the data packet header information according to the time-aligned analog quantity set and decomposes it according to the bit width. It compares the decomposed header fields with the standard register address mapping table bit by bit to identify the communication protocol type corresponding to the data packet. According to the communication protocol type, it calls the corresponding decoding logic to extract the payload data, converts the payload data into a unified floating-point format, and generates a protocol adaptation parsing frame.

[0014] The data stream merging and alignment submodule adapts and parses frames according to the protocol, retrieves timestamp tags of data frames from different channels, sorts data frames based on timestamp tags, fills in missing sampling points on the time axis, and splices the sorted voltage analog signal data frames and current analog signal data frames in channel order to establish a continuous data stream including a complete time series and channel identifiers, generating a multi-source raw sampling data stream.

[0015] Preferably, the signal-to-noise ratio threshold adjustment module includes:

[0016] The noise level estimation submodule extracts a fixed length of sampling points from the current moment to form a sliding time window based on the multi-source raw sampling data stream. It then iterates through all sampling amplitude values ​​within the sliding time window, discretizes the sampling amplitude values, calculates the average value, calculates the sum of squares of the deviations of each sampling point relative to the average value, divides the sum of squares of the deviations by the total number of sampling points to obtain the statistical variance value, and generates the background noise estimation value.

[0017] The quality index calculation submodule calculates the root mean square value of the signal amplitude within the sliding time window based on the estimated background noise value, squares the root mean square value to obtain the signal power value, divides the signal power value by the estimated background noise value to obtain the dimensionless ratio value, converts the ratio value to decibels, and generates real-time signal quality index.

[0018] The gate width parameter adjustment submodule reads the steady-state noise baseline value stored in the register according to the real-time signal quality index, calculates the difference between the real-time signal quality index and the steady-state noise baseline value, determines the positive or negative polarity of the difference, increases the value of the current compression gate width parameter by a preset step size when the difference is negative, and decreases the value of the current compression gate width parameter by a preset step size when the difference is positive, thereby generating an adaptive compression deviation threshold.

[0019] Preferably, the trajectory node compression module includes:

[0020] The slope boundary construction submodule reads the coordinates of the starting sampling point of the current compressed segment in the multi-source original sampling data stream according to the adaptive compression deviation threshold, adds and subtracts the adaptive compression deviation threshold to the amplitude value of the starting sampling point to obtain the upper boundary point and the lower boundary point, connects the starting sampling point and the upper boundary point to form the upper boundary slope, connects the starting sampling point and the lower boundary point to form the lower boundary slope, and generates a dynamic slope boundary pair.

[0021] The limit-out determination triggering submodule reads the time coordinate and amplitude coordinate of the next sampling point according to the dynamic slope boundary pair, calculates the real-time slope value of the line connecting the next sampling point and the starting sampling point, compares the real-time slope value with the upper boundary slope value and the lower boundary slope value in the dynamic slope boundary pair, determines whether the real-time slope value is outside the angle region formed by the upper boundary slope and the lower boundary slope, and generates a limit-out determination signal.

[0022] Preferably, the trajectory node compression module further includes:

[0023] The node record reset submodule, based on the over-limit trigger judgment signal, locks the previous sampling point of the current sampling point as the retention object when the judgment result is yes, extracts the timestamp and amplitude value of the retention object and stores it in the memory, sets the current sampling point as the new starting sampling point, clears the processed sampling point cache and initializes the calculation state of the next compression segment, and generates a key trajectory node set.

[0024] Preferably, the topology anomaly localization module includes:

[0025] The grid node mapping submodule, based on the key trajectory node set, calls the power system grid topology model database, parses the node connection relationship matrix in the power system grid topology model database, spatially matches the time coordinates and amplitude coordinates in the key trajectory node set with the electrical node index in the matrix, assigns a unique topology location identifier to each trajectory node, establishes the correspondence between trajectory data and power grid physical architecture, and generates a mapped grid node sequence.

[0026] The transient parameter extraction submodule retrieves the amplitude coordinate values ​​corresponding to adjacent topological location identifiers based on the mapped grid node sequence, calculates the difference in current amplitude between adjacent nodes and divides it by the time difference to obtain the current change rate, calculates the decrease in voltage amplitude and divides it by the rated voltage value to obtain the voltage sag percentage, and combines the current change rate and voltage sag percentage into a feature vector to generate an electrical transient parameter set.

[0027] Preferably, the topology anomaly localization module further includes:

[0028] The abnormal section screening submodule reads the instantaneous overcurrent protection setting value and transient recording start threshold of the protection device according to the electrical transient parameter group. It compares the current change rate in the feature vector with the instantaneous overcurrent protection setting value and the voltage sag percentage with the transient recording start threshold. It filters the topological location identifiers corresponding to feature vectors whose comparison results all exceed the threshold, queries the geographic information system database to match latitude and longitude coordinates, and generates a set of abnormal power grid topology locations.

[0029] The present invention also provides a device, which is a 12 kV primary and secondary integrated ring main unit, comprising physically isolated primary unit compartments and secondary unit compartments.

[0030] The secondary unit compartment integrates and encapsulates a power distribution automation terminal, a power management module, and a communication gateway. The power distribution automation terminal serves as the hardware carrier of the sampling control system.

[0031] The primary unit compartment is encapsulated with a sulfur hexafluoride fully insulated switch assembly or a solid insulated switch assembly.

[0032] The primary unit compartment and the secondary unit compartment are connected via an aviation plug-type prefabricated cable interface;

[0033] Wireless temperature sensors and measurement sensors are respectively installed at the contact and insulation positions of the primary unit compartment. The wireless temperature sensors and the measurement sensors are connected to the power distribution automation terminal and transmit temperature monitoring signals and partial discharge monitoring signals to the multi-source signal acquisition module.

[0034] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0035] This invention eliminates data timing deviations caused by dispersed device deployment by parallel acquisition of analog voltage and current signals and using a dual-mode BeiDou and GPS time synchronization component for timestamp alignment, ensuring strict synchronization of multi-source data in the time dimension. The operation of parsing communication message headers to adapt to standard protocols and merging data streams breaks down communication barriers between heterogeneous devices, constructing a unified data foundation. A mechanism based on a sliding time window to extract signal amplitude sequences and calculate variance to estimate noise floor, and then calculating the real-time signal-to-noise ratio (SNR) and comparing it with a steady-state baseline, achieves dynamic perception of sampled data quality. When the signal-to-noise ratio is low, the gate width is automatically increased to filter background noise redundancy; when the signal-to-noise ratio is high, the gate width is decreased to retain subtle feature evolution. This adaptive adjustment strategy reduces the data transmission bandwidth pressure while maximally restoring waveform details before and after the fault. Combined with slope over-limit judgment to extract key trajectory nodes, and mapping the nodes to the power grid topology model for multi-dimensional feature analysis of current change rate and voltage sag amplitude, it can screen out abnormal branches that exceed the instantaneous overcurrent protection setting value and associate them with geographical coordinates. This replaces the traditional troubleshooting mode that relies on manual experience, improving the accuracy and response speed of fault diagnosis. Attached Figure Description

[0036] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0038] Please see Figure 1 The present invention provides a technical solution: a sampling control system for a power system includes:

[0039] The multi-source signal acquisition module is used to acquire voltage analog signals and current analog signals. It uses the Beidou and GPS dual-mode time synchronization components to align the timestamps of the voltage analog signals and current analog signals, parses the communication message header to match the standard protocol, merges the aligned voltage analog signals and current analog signals, and generates a multi-source raw sampling data stream.

[0040] The signal-to-noise ratio (SNR) threshold adjustment module is used to extract the signal amplitude sequence within the sliding time window based on the multi-source raw sampling data stream, calculate the variance of the signal amplitude sequence to estimate the noise floor level, calculate the ratio of the square of the effective signal value to the noise floor level to determine the real-time SNR index, compare the real-time SNR index with the preset steady-state noise baseline, increase the compression gate width parameter when the real-time SNR index is lower than the steady-state noise baseline, and decrease the compression gate width parameter when the real-time SNR index is higher than the steady-state noise baseline, thereby generating an adaptive compression deviation threshold.

[0041] The trajectory node compression module is used to construct upper and lower boundary slope limits at the starting sampling point of the multi-source original sampling data stream based on an adaptive compression deviation threshold, calculate the slope of the line connecting subsequent sampling points and the starting sampling point, determine whether the line slope exceeds the range defined by the upper and lower boundary slope limits, mark the previous sampling point as a turning point when the line slope exceeds the range, reset the starting sampling point to the current sampling point and reconstruct the boundary slope limits, record the time coordinates and amplitude coordinates of the turning point, and generate a set of key trajectory nodes.

[0042] The topology anomaly location module is used to map time coordinates and amplitude coordinates to the corresponding node positions in the power grid topology model based on the key trajectory node set. It calculates the current change rate and voltage sag amplitude between adjacent nodes in the key trajectory node set, compares the current change rate with the instantaneous overcurrent protection setting value, compares the voltage sag amplitude with the transient recording start threshold, filters branch sections whose comparison results exceed the limit, associates the branch sections with the geographic information system coordinates, and generates a set of power grid topology anomaly locations.

[0043] The multi-source signal acquisition module includes:

[0044] The synchronous trigger acquisition submodule, based on the ARM Cortex-A53 processor and FPGA logic unit, configures the parallel signal acquisition interface, uses the FPGA logic unit to send hardware trigger commands to the analog-to-digital conversion component, reads the voltage analog signal value and the current analog signal value in parallel, calls the Beidou and GPS dual-mode time synchronization component to obtain the current time information, and attaches the time information as a timestamp tag to the voltage analog signal value and the current analog signal value to generate a time-aligned analog quantity set.

[0045] The protocol message parsing submodule extracts the data packet header information and decomposes it according to the bit width based on the timing-aligned analog quantity set. It compares the decomposed header fields bit by bit with the standard register address mapping table to identify the communication protocol type corresponding to the data packet. Based on the communication protocol type, it calls the corresponding decoding logic to extract the payload data, converts the payload data into a unified floating-point format, and generates a protocol adaptation parsing frame.

[0046] The data stream merging and alignment submodule adapts and parses frames according to the protocol, retrieves timestamp tags of data frames from different channels, sorts data frames based on timestamp tags, fills in missing sampling points on the time axis, and splices the sorted voltage analog signal data frames and current analog signal data frames in channel order to establish a continuous data stream including complete time series and channel identifiers, generating a multi-source raw sampling data stream.

[0047] Specifically, the synchronous trigger acquisition submodule, based on the ARM Cortex-A53 processor and FPGA logic unit, configures the parallel signal acquisition interface. The ARM processor writes configuration instructions to the FPGA's control register, setting the sampling frequency to 10,000 times per second. Upon receiving the configuration instructions, the FPGA logic unit starts its internal high-precision counter. When the counter reaches the preset sampling period count value, it simultaneously sends high-level pulses to the 128-channel analog-to-digital converters (ADCs) as a hardware trigger instruction via the FPGA logic unit's parallel output pins. Upon receiving the hardware trigger instruction, the ADCs immediately freeze the current analog signal voltage value and begin the analog-to-digital conversion process. The FPGA logic unit monitors the status pins of the ADCs. When a conversion completion signal is detected, it sequentially reads the 16-bit binary values ​​output by the ADCs through the 128-channel parallel data bus, and combines the read voltage analog signal value with the current analog signal value. The numerical values ​​are temporarily stored in the FIFO queue inside the FPGA. At the same time, the BeiDou and GPS dual-mode time synchronization components are called. The GPRMC format message data output by the time synchronization components is read through the serial port. The general coordinated time information in the message, including the year, month, day, hour, minute, and second fields, is parsed. The rising edge of the second pulse signal output by the time synchronization components is captured. The microsecond-level counter inside the FPGA is used to record the microsecond offset after the rising edge of the second pulse. The general coordinated time information is converted into a second-level timestamp. The microsecond offset is combined with the second-level timestamp to construct a 64-bit absolute time value accurate to the microsecond. This absolute time value is used as a timestamp label and is appended to the header of each group of voltage analog signal values ​​and current analog signal values ​​temporarily stored in the FIFO queue in the order of data acquisition, forming a data packet with high-precision time stamps and generating a time-aligned analog quantity set.

[0048] The protocol message parsing submodule extracts the header information of the data packet based on the timing-aligned analog quantity set and decomposes it according to the bit width. It establishes a matching rule base containing feature codes of various standard communication protocols. This rule base pre-stores the Ethernet type field (0x88BA) of the IEC 61850 protocol, the start character (0x68) of the IEC 104 protocol, and the slave address range (0x01 to 0xF7) of the MODBUS-RTU protocol. The first two bytes of the data packet are read as the header identification code. The extracted header identification code is XORed with the feature codes in the matching rule base one by one. A successful match is determined when the XOR result is zero. If the header identification code is 0x68, the next byte is read as the length field, and the total length of the data packet is checked against this length field. If they match, it is confirmed as the IEC 104 protocol. If the header identification code corresponds to the Ethernet type (0x88BA), it is confirmed as the IEC 61850 sampled value protocol. After identifying the communication protocol type corresponding to the data packet, the corresponding decoding logic is called according to the communication protocol type. For IEC... According to the IEC 61850 protocol, following the structural definition of the application service data unit, the control field and type identifier field are skipped, and the information object address segment is located. The subsequent four bytes are read as the information body data. For the IEC 61850 protocol, the application protocol data unit is parsed according to the ASN.1 encoding rules, and the analog quantity value area under the dataset field is located. The extracted payload data is uniformly converted into a single-precision floating-point format conforming to the IEEE 754 standard, according to the data type defined by the protocol, such as a 16-bit signed integer or a 32-bit unsigned integer. Based on the preset transformer ratio coefficient, such as a voltage transformer ratio of 10,000 to 100 and a current transformer ratio of 600 to 5, the floating-point number is multiplied to restore the true physical quantity value of the primary side, and a protocol adaptation parsing frame is generated.

[0049] The data stream merging and alignment submodule adapts and parses frames according to the protocol, retrieves timestamp tags from data frames of different channels, and creates a time axis reference in microseconds. The starting point of this reference is set to the first timestamp of the current processing cycle. It iterates through all channel data frames, reads the 64-bit absolute time value from the header of each data frame, subtracts this value from the time axis reference, and calculates the time offset of each data frame relative to the reference. Based on the magnitude of the time offset, it uses a quicksort algorithm to sort the voltage analog signal data frames and current analog signal data frames in ascending order. It then checks the timestamp difference between two adjacent data frames in the sorted data stream, calculated using the following formula: ,in For the timestamp of the next frame, The timestamp of the previous frame is used to calculate the difference. Compared with the system's standard sampling interval of one hundred microseconds, if If the time interval is greater than 100 microseconds but less than 500 microseconds, it is determined that there are missing sampling points. The value of the missing points is calculated using linear interpolation. The calculation formula is as follows: ,in The simulated values ​​for the missing points. and These are the analog values ​​for the preceding and following frames, respectively. The theoretical timestamps for missing points are used to construct new data frames using the calculated values ​​and insert them into the corresponding positions in the data stream to fill in the missing sampling points on the time axis. Then, in the order of the channel identifiers, i.e. from channel one to channel one hundred and twenty-eight, the voltage analog signal data frames and current analog signal data frames at the same time are horizontally spliced ​​together to form a wide-byte data line containing all channel data, establishing a continuous data stream including the complete time series and channel identifiers, and generating a multi-source raw sampling data stream.

[0050] The signal-to-noise ratio threshold adjustment module includes:

[0051] The noise level estimation submodule extracts a fixed length of sampling points from the current moment to form a sliding time window based on the multi-source raw sampling data stream. It then iterates through all sampling amplitude values ​​within the sliding time window, discretizes the sampling amplitude values, calculates the average value, calculates the sum of squares of the deviations of each sampling point relative to the average value, divides the sum of squares of the deviations by the total number of sampling points to obtain the statistical variance value, and generates the background noise estimation value.

[0052] The quality index calculation submodule calculates the root mean square value of the signal amplitude within the sliding time window based on the estimated background noise value, squares the root mean square value to obtain the signal power value, divides the signal power value by the estimated background noise value to obtain the dimensionless ratio value, converts the ratio value to decibels, and generates real-time signal quality index.

[0053] The gate width parameter adjustment submodule reads the steady-state noise baseline value stored in the register according to the real-time signal quality index, calculates the difference between the real-time signal quality index and the steady-state noise baseline value, determines the positive or negative polarity of the difference, increases the value of the current compression gate width parameter by a preset step size when the difference is negative, and decreases the value of the current compression gate width parameter by a preset step size when the difference is positive, thereby generating an adaptive compression deviation threshold.

[0054] Specifically, the noise level estimation submodule extracts a fixed-length sampling point from the multi-source raw sampling data stream to form a sliding time window, and sets the length of the sliding time window. Two hundred sampling points, corresponding to one complete cycle of a 50 Hz power frequency signal, are read from the data stream. The amplitude values ​​of each sampling point are denoted as the sequence. ,in Indicates the first Given the amplitude values ​​at each sampling point, iterate through all sampling amplitude values ​​within the sliding time window, discretize the sampling amplitude values, and calculate the average value of the sequence. The calculation formula is: Then iterate through the sequence again. Each sampling point Compared with the average The deviation value is obtained by subtraction, the deviation value is squared to obtain the deviation square, and the deviation squares of all sampling points are summed to obtain the deviation sum. The calculation formula is: Sum of squared deviations Divide by the total number of sampling points To obtain statistical variance values The calculation formula is: The statistical variance value characterizes the degree of fluctuation and dispersion of the signal within the sliding window, and is used as the basis for evaluating the intensity of non-periodic random components in the current signal to generate an estimated background noise value.

[0055] The quality index calculation submodule calculates the root mean square value of the signal amplitude within the sliding time window based on the estimated background noise value, using the sampling point sequence within the aforementioned sliding window. The amplitude value of each sampling point Perform a squaring operation, sum all the squared values, and then divide by the total number of sampling points. Then, perform a square root operation on the quotient to obtain the root mean square value. The calculation formula is: The signal power value is obtained by squaring the root mean square value. ,Right now The background noise estimate generated in the previous step is read as the noise power value. , signal power value Divide by the estimated background noise value Obtain dimensionless ratio values The calculation formula is: In order to adapt to signal changes with a large dynamic range, the ratio value Take the logarithm to base 10, multiply the result by 10, and convert it to a value in decibels. The calculation formula is: This value directly reflects the relative strength of the effective signal energy compared to the background noise energy, generating a real-time signal quality index.

[0056] The gate width parameter adjustment submodule reads the steady-state noise baseline value pre-stored in the register based on the real-time signal quality index. This baseline value is set to 60 decibels and is based on the average signal-to-noise ratio data measured during long-term operation of the equipment in an ideal laboratory environment. It then calculates the real-time signal quality index. Compared with steady-state noise baseline values The difference The calculation formula is: Determine the difference The positive and negative polarities, if A value less than zero indicates that the current ambient noise is high or the signal is weak. To suppress noise interference, it is necessary to increase the compression gate width and set a base step size. The value is 0.01; read the current compression gate width parameter. According to the formula Calculate the new door width parameters, where The absolute value of the difference, if A value greater than or equal to zero indicates good signal quality. To retain more waveform details, the compression gate width needs to be reduced, according to the formula... Calculate the new gate width parameters, and then... A limiting process is applied, restricting the width to between a preset minimum door width of 0.001 and a maximum door width of 0.1. Then take ,like Then take Generate an adaptive compression deviation threshold.

[0057] The trajectory node compression module includes:

[0058] The slope boundary construction submodule reads the coordinates of the starting sampling point of the current compressed segment in the multi-source original sampled data stream according to the adaptive compression deviation threshold. It adds and subtracts the adaptive compression deviation threshold to the amplitude value of the starting sampling point to obtain the upper boundary point and the lower boundary point. It connects the starting sampling point and the upper boundary point to form the upper boundary slope, and connects the starting sampling point and the lower boundary point to form the lower boundary slope, thus generating a dynamic slope boundary pair.

[0059] The over-limit determination triggering submodule reads the time coordinate and amplitude coordinate of the next sampling point based on the dynamic slope boundary pair, calculates the real-time slope value of the line connecting the next sampling point and the starting sampling point, compares the real-time slope value with the upper boundary slope value and the lower boundary slope value in the dynamic slope boundary pair, determines whether the real-time slope value is outside the angle region formed by the upper boundary slope and the lower boundary slope, and generates an over-limit triggering determination signal.

[0060] The node record reset submodule, based on the over-limit trigger judgment signal, locks the previous sampling point of the current sampling point as the reserved object when the judgment result is yes, extracts the timestamp and amplitude value of the reserved object and stores it in the memory, sets the current sampling point as the new starting sampling point, clears the processed sampling point cache and initializes the calculation state of the next compression segment, and generates the key trajectory node set.

[0061] Specifically, the slope boundary construction submodule, based on the adaptive compression deviation threshold, reads the coordinates of the starting sampling point of the current compressed segment in the multi-source raw sampled data stream and marks the starting sampling point as... ,in A timestamp in microseconds representing the start time. This indicates the analog amplitude value at the start time; the corresponding adaptive compression deviation threshold for this channel is then read. This threshold has been dynamically adjusted to, for example, 0.02 based on the signal-to-noise ratio in the previous module, for each subsequent sampling point entering the sliding window. Calculate the upper limit amplitude of its tolerance range respectively. and lower bound amplitude Construct from the starting sampling point Point to the upper tolerance limit of the current sampling point The geometric connection is used to calculate the slope of the connection as a candidate value for the upper boundary slope. The calculation formula is: Similarly, construct a pointer to the lower bound. Connect the geometric lines and calculate the candidate values ​​of the lower boundary slope. The calculation formula is: To maintain an effective sector region that encompasses all historical data within the current compressed segment, the boundary slope needs to be converged and updated by reading the effective upper boundary slope maintained at the previous time step. and effective lower boundary slope , will the currently calculated and Compare and take the smaller value as the new ,Will and Compare and take the larger value as the new That is, execution and Through continuous iteration and updates, it has become possible to improve upon the previous version. and The angle region formed gradually narrows to accurately fit the evolution trend of the data flow and generate dynamic slope boundary pairs.

[0062] The limit-out detection trigger submodule reads the time coordinates and amplitude coordinates of the next sampling point based on the dynamic slope boundary pair, and denotes them as follows. This point is a new data point immediately following the processed data points within the current compression window. Its coordinates are then compared to the starting sampling point of the current compression segment. Perform geometric calculations to determine the slope of the line connecting these two points in real time. The calculation formula is: The dynamic slope boundary pairs maintained in the preceding steps are invoked, including the upper boundary slope limit that is valid at the current moment. and lower boundary slope limit These two limits define the conditions for meeting the adaptive compression deviation threshold. Under the premise of data extension of a legal sector area, the real-time connection slope will be calculated. respectively with and Perform numerical range comparison, and the judgment condition is: or If this condition is met, it means a new sampling point The addition of this variable causes the data trajectory to deviate from the linear trend band originating from the starting point. This means the current linear segment cannot fit the point, and the error is within acceptable limits. At this point, a trend abrupt change or curvature exceeding the limit is determined, and the logical variable is adjusted accordingly. Set to true; if the condition is not met, it indicates that the new sampling point is still within the linear trend band, and the logical variable is... Set to false value to generate an over-limit trigger signal.

[0063] The node record reset submodule, based on the limit-crossing trigger judgment signal, performs archiving and reset operations on key nodes when the judgment result is yes, i.e., when a trend change is detected. First, it locks the previous sampling point of the current sampling point. This point is the last valid data point within the current linear compression segment that satisfies the error constraint. It is identified as a trajectory inflection point, and a 64-bit microsecond-level timestamp is extracted from this inflection point. and 32-bit floating-point amplitude value These two values ​​are combined into a standardized data structure and written into a specific address region of the on-chip static random access memory. Simultaneously, the sampling point that caused the limit to be exceeded is... Set as the new starting sampling point for the compressed section Update the start time register and start amplitude register, and set the upper boundary slope limit. Reset to positive infinity, limit the lower boundary slope value The value is reset to negative infinity to initialize the calculation state of the sector region of the next compression segment. The first-in-first-out queue used to temporarily store the intermediate data of the current segment in the FPGA is cleared to ensure that the subsequent slope calculation is based entirely on the new starting point. If the judgment result is negative, no storage operation is performed. Only the dynamic slope boundary pair is updated and the next sampling point is processed. Finally, all the turning point data written to the memory are summarized to generate the key trajectory node set.

[0064] The topology anomaly localization module includes:

[0065] The grid node mapping submodule, based on the key trajectory node set, calls the power system grid topology model database, parses the node connection relationship matrix in the power system grid topology model database, spatially matches the time coordinates and amplitude coordinates in the key trajectory node set with the electrical node index in the matrix, assigns a unique topology location identifier to each trajectory node, establishes the correspondence between trajectory data and power grid physical architecture, and generates a mapped grid node sequence.

[0066] The transient parameter extraction submodule retrieves the amplitude coordinate values ​​corresponding to adjacent topological location identifiers based on the mapped grid node sequence, calculates the difference in current amplitude between adjacent nodes and divides it by the time difference to obtain the current change rate, calculates the decrease in voltage amplitude and divides it by the rated voltage value to obtain the voltage sag percentage, and combines the current change rate and voltage sag percentage into a feature vector to generate an electrical transient parameter set.

[0067] The abnormal section screening submodule reads the instantaneous overcurrent protection setting value and transient recording start threshold of the protection device based on the electrical transient parameter group. It compares the current change rate in the feature vector with the instantaneous overcurrent protection setting value and the voltage sag percentage with the transient recording start threshold. It filters the topological location identifiers corresponding to feature vectors whose comparison results all exceed the threshold, queries the geographic information system database to match latitude and longitude coordinates, and generates a set of abnormal power grid topology locations.

[0068] Specifically, the grid node mapping submodule, based on the key trajectory node set, calls the power system grid topology model database. This database stores the physical connection relationships of the distribution network in the form of an adjacency matrix. The rows and columns in the matrix represent different electrical nodes, and the matrix elements represent the connection impedance or switching status between nodes. It reads the data of each node in the key trajectory node set, extracts its associated channel identifier information (e.g., channel ID 0x0A), and queries a pre-set channel-node mapping table. This mapping table defines the correspondence between physical acquisition channels and primary equipment electrical nodes; for example, it maps channel ID 0x0A to electrical node index Node_501. The time coordinates in the key trajectory node set are then mapped... and amplitude coordinates By binding the data to the electrical node index, multidimensional data involving three-phase current and three-phase voltage are mapped to the corresponding A-phase, B-phase, and C-phase sub-nodes, respectively. Based on the hierarchical relationship in the topology model, a unique topology location identifier is assigned to each trajectory node in the format of "substation ID_feeder ID_equipment ID", such as "Sub_01_Feeder_05_Switch_12". This anchors the originally isolated time series data to the spatial architecture of the power grid, establishes the correspondence between trajectory data and the physical architecture of the power grid, and generates a mapping grid node sequence.

[0069] The transient parameter extraction submodule retrieves the amplitude coordinate values ​​corresponding to adjacent topological location identifiers based on the mapped grid node sequence, traverses the node sequence under the same topological location identifier in chronological order, and selects two adjacent key nodes. and ,in Represents the current amplitude. Represents the voltage amplitude, and calculates the difference in current amplitude between adjacent nodes. Calculate the time difference Dividing the difference value by the time difference yields the rate of change of current. The calculation formula is: Simultaneously read the rated voltage value of the node. For example, with 10,000 volts, calculate the decrease in voltage amplitude. Divide the voltage drop by the rated voltage and multiply by 100% to obtain the voltage sag percentage. The calculation formula is: The calculated rate of change of current and voltage sag percentage Combined into a two-dimensional feature vector The feature vector is then labeled with a corresponding timestamp and topological location tag to generate an electrical transient parameter set.

[0070] The abnormal section screening submodule reads the instantaneous overcurrent protection setting value and transient waveform recording start threshold of the protection device based on the electrical transient parameter group. The setting is based on 1.5 to 2 times the maximum load current of the line, for example, setting it to a current change rate threshold of 2,000 amperes per second, and a transient recording start threshold. Set to 20% of the rated voltage, meaning triggering when the voltage drops by more than 20%, iterates through each eigenvector in the electrical transient parameter set. The current rate of change component in the eigenvector With instantaneous overcurrent protection setting value Perform a numerical comparison to determine if the condition is met. At the same time, the voltage sag percentage component With transient recording start threshold Perform a numerical comparison to determine if the condition is met. Only when both comparison conditions are met simultaneously is it determined that a short-circuit fault or large disturbance has occurred at that location at that moment. The topological location identifier corresponding to the feature vector is extracted, for example, "Sub_01_Feeder_05_Switch_12". This identifier is then used as the primary key to query the geographic information system database to obtain the longitude coordinates corresponding to the electrical equipment. and latitude coordinates The abnormal device's ID, fault occurrence time, peak current rate of change, and latitude and longitude coordinates are packaged to generate a set of abnormal locations in the power grid topology.

[0071] The present invention also provides a device, which is a 12 kV primary and secondary integrated ring main unit, including physically isolated primary unit compartments and secondary unit compartments.

[0072] The secondary unit compartment integrates and encapsulates a power distribution automation terminal, a power management module, and a communication gateway. The power distribution automation terminal serves as the hardware carrier of the sampling control system.

[0073] The primary unit compartment is encapsulated with sulfur hexafluoride fully insulated switch assembly or solid insulated switch assembly;

[0074] The primary unit compartment and the secondary unit compartment are connected via an aviation plug-type prefabricated cable interface;

[0075] Wireless temperature sensors and measurement sensors are respectively installed at the contact and insulation positions of the primary unit compartment. The wireless temperature sensors and measurement sensors are connected to the power distribution automation terminal and transmit temperature monitoring signals and partial discharge monitoring signals to the multi-source signal acquisition module.

Claims

1. A sampling control system for a power system, characterized in that, The system comprises: A multi-source signal acquisition module is configured to acquire voltage analog signals and current analog signals, align time stamps of the voltage analog signals and the current analog signals by using a Beidou and GPS dual-mode time matching component, parse a communication message header to match a standard protocol, merge the voltage analog signals and the current analog signals after alignment, and generate a multi-source original sampling data stream; A signal-to-noise ratio threshold adjustment module is configured to extract a signal amplitude sequence in a sliding time window according to the multi-source original sampling data stream, calculate a variance value of the signal amplitude sequence to estimate a noise floor level, calculate a ratio of a signal effective value square to the noise floor level to determine a real-time signal-to-noise ratio index, compare the real-time signal-to-noise ratio index with a preset steady-state noise reference line in a numerical manner, increase a compression gate width parameter when the real-time signal-to-noise ratio index is lower than the steady-state noise reference line, decrease the compression gate width parameter when the real-time signal-to-noise ratio index is higher than the steady-state noise reference line, and generate an adaptive compression deviation threshold value; The system further comprises: A trajectory node compression module is configured to construct an upper boundary slope limit value and a lower boundary slope limit value at a starting sampling point of the multi-source original sampling data stream according to the adaptive compression deviation threshold value, calculate a slope of a line connecting a subsequent sampling point and the starting sampling point, determine whether the slope of the line exceeds a range defined by the upper boundary slope limit value and the lower boundary slope limit value, mark a previous sampling point as a turning point when the slope of the line exceeds the range, reset the starting sampling point as a current sampling point and reconstruct the boundary slope limit value, record a time coordinate and an amplitude coordinate of the turning point, and generate a key trajectory node set; A topology anomaly positioning module is configured to map the time coordinate and the amplitude coordinate to node positions corresponding to a power grid topology structure model according to the key trajectory node set, calculate a current change rate value and a voltage sag amplitude value between adjacent nodes in the key trajectory node set, compare the current change rate value with a speed-break protection setting value, compare the voltage sag amplitude value with a transient recording wave starting threshold value, screen branch sections whose comparison results exceed limit values, associate the branch sections with geographic information system coordinates, and generate a power grid topology anomaly position set.

2. The sampling control system for a power system according to claim 1, characterized by, The multi-source signal acquisition module comprises: A synchronous trigger acquisition sub-module is configured to configure a signal parallel acquisition interface according to an ARM Cortex-A53 processor and an FPGA logic unit, send a hardware trigger instruction to an analog-to-digital conversion component by using the FPGA logic unit, read voltage analog signal values and current analog signal values in parallel, call a Beidou and GPS dual-mode time matching component to obtain current time information, attach the time information as a time stamp label to the voltage analog signal values and the current analog signal values, and generate a time sequence aligned analog quantity set; A protocol message parsing sub-module is configured to extract packet header information from the time sequence aligned analog quantity set, disassemble the packet header information according to bit widths, compare disassembled header fields with a standard register address mapping table bit by bit, identify a communication protocol type corresponding to the packet, call corresponding decoding logic to extract payload data according to the communication protocol type, convert the payload data into a unified floating-point number format, and generate a protocol adaptation analysis frame. The data stream merging alignment sub-module parses a frame according to the protocol, retrieves a timestamp label of a data frame of different channels, sorts the data frame according to the timestamp label as an axis, fills in missing sample point positions on a time axis, splices the voltage analog signal data frame and the current analog signal data frame according to a channel sequence after sorting, establishes a continuous data stream including a complete time sequence and a channel identifier, and generates a multi-source original sampling data stream.

3. The sampling control system for a power system according to claim 1, characterized by, The signal-to-noise ratio threshold adjustment module comprises: A noise level estimation sub-module intercepts a fixed length of sample points before a current time to form a sliding time window according to the multi-source original sampling data stream, traverses all sample amplitude values in the sliding time window, performs a discretization process on the sample amplitude values and calculates an average value, calculates a sum of squares of deviations of each sample point relative to the average value, divides the sum of squares of deviations by a total number of sample points to obtain a statistical variance value, and generates a background noise estimation value; A quality index calculation sub-module calculates a root mean square value of signal amplitude in the sliding time window according to the background noise estimation value, performs a square operation on the root mean square value to obtain a signal power value, divides the signal power value by the background noise estimation value to obtain a dimensionless ratio value, converts the ratio value into a decibel unit, and generates a real-time signal quality index; A gate width parameter adjustment sub-module reads a steady-state noise reference line value pre-stored in a register according to the real-time signal quality index, calculates a difference value between the real-time signal quality index and the steady-state noise reference line value, judges a positive or negative polarity of the difference value, increases a value of a current compression gate width parameter by a preset step size when the difference value is negative, decreases the value of the current compression gate width parameter by the preset step size when the difference value is positive, and generates an adaptive compression deviation threshold value.

4. The sampling control system for a power system according to claim 1, characterized by, The trajectory node compression module comprises: A slope boundary construction sub-module reads a starting sample point coordinate of a current compression segment in the multi-source original sampling data stream according to the adaptive compression deviation threshold value, adds and subtracts the adaptive compression deviation threshold value from the starting sample point amplitude value respectively to obtain an upper boundary point and a lower boundary point, connects the starting sample point with the upper boundary point to form an upper boundary slope, connects the starting sample point with the lower boundary point to form a lower boundary slope, and generates a dynamic slope boundary pair; An out-of-limit judgment trigger sub-module reads a time coordinate and an amplitude coordinate of a next sample point according to the dynamic slope boundary pair, calculates a real-time slope value of a connecting line between the next sample point and the starting sample point, compares the real-time slope value with upper boundary slope and lower boundary slope values in the dynamic slope boundary pair in a numerical interval, judges whether the real-time slope value is located outside an included angle region formed by the upper boundary slope and the lower boundary slope, and generates an out-of-limit trigger judgment signal.

5. The sampling control system for a power system according to claim 4, characterized by, The trajectory node compression module further comprises: A node record reset sub-module locks a previous sample point of a current sample point as a reserved object when the judgment result is yes according to the out-of-limit trigger judgment signal, extracts a timestamp and an amplitude value of the reserved object and stores them in a memory, sets the current sample point as a new starting sample point, clears a processed sample point cache and initializes a calculation state of a next compression segment, and generates a key trajectory node set.

6. The sampling control system for a power system according to claim 1, characterized by, The topological anomaly positioning module comprises: A grid node mapping submodule, according to the key trajectory node set, calls a power system grid topology structure model database, parses a node connection relationship matrix in the power system grid topology structure model database, spatially matches time coordinates and amplitude coordinates in the key trajectory node set with electrical node indexes in the matrix, allocates a unique topological position identifier to each trajectory node, establishes a corresponding relationship between trajectory data and a grid physical architecture, and generates a mapping grid node sequence; A transient parameter extraction submodule, according to the mapping grid node sequence, retrieves amplitude coordinate values corresponding to adjacent topological position identifiers, calculates a difference value of current amplitudes between adjacent nodes and divides the difference value by a time difference value to obtain a current change rate, calculates a drop value of voltage amplitudes and divides the drop value by a rated voltage value to obtain a voltage sag percentage, combines the current change rate and the voltage sag percentage into a feature vector, and generates an electrical transient parameter group.

7. The sampling control system for a power system according to claim 6, characterized by, The topological anomaly positioning module further comprises: An abnormal section screening submodule, according to the electrical transient parameter group, reads speed-break protection setting values and transient recording wave starting threshold values of a protection device, compares the current change rate in the feature vector with the speed-break protection setting values, compares the voltage sag percentage with the transient recording wave starting threshold values, screens topological position identifiers corresponding to feature vectors whose comparison results all exceed threshold values, queries geographic information system database to match longitude and latitude coordinates, and generates a grid topological anomaly position set.

8. The apparatus for a sample control system of a power system according to any one of claims 1 to 7, characterized by: The device is a 12-kilovolt primary and secondary fusion ring network cabinet, comprising a physically isolated primary unit compartment and a secondary unit compartment; The secondary unit compartment is integrated and packaged with a power distribution automation terminal, a power management module and a communication gateway, and the power distribution automation terminal serves as a hardware carrier of the sampling control system; The primary unit compartment is packaged with a sulfur hexafluoride fully insulated switch assembly or a solid insulation switch assembly; The primary unit compartment and the secondary unit compartment are connected through an aviation plug type prefabricated cable interface; Wireless temperature sensing sensors and measurement sensors are respectively arranged at contact positions and insulation positions of the primary unit compartment, and the wireless temperature sensing sensors and the measurement sensors are connected to the power distribution automation terminal and deliver temperature monitoring signals and partial discharge monitoring signals to the multi-source signal acquisition module.

Images