A mutual inductor digital sampling method and system

Abstract

The present application relates to the technical field of power system signal measurement, and particularly relates to a mutual inductor digital sampling method and system. The method comprises: acquiring a signal of a secondary side output of a mutual inductor for a power system in real time; determining a load fluctuation degree at a current time; determining a dynamic sensitivity coefficient at the current time; acquiring a dynamic reference mean value, a dynamic relaxation parameter and a dynamic decision threshold value in a CUSUM algorithm at the current time; determining a cumulative sum between the signal and the dynamic reference mean value in the CUSUM algorithm at the current time; and determining a sampling frequency of the mutual inductor according to the cumulative sum between the signal and the dynamic reference mean value in the CUSUM algorithm at the current time and the size of the dynamic decision threshold value, so as to realize digital sampling of the mutual inductor. Through adaptive adjustment of the reference mean value, the relaxation parameter and the decision threshold value, the present application reduces the false negative and false positive rates and improves the identification ability for initial faults such as mutual inductor inter-turn short circuit.

Description

Technical Field

[0001] This invention relates to the field of power system signal measurement technology, and in particular to a digital sampling method and system for instrument transformers. Background Technology

[0002] Instrument transformers are key devices in power systems used to measure high voltage and high current and proportionally transmit primary-side electrical quantities to secondary-side measurement and protection devices. With the rapid development of smart power systems and digital substations, high-precision and high-efficiency digital sampling of instrument transformer output signals has become a core technical aspect. To effectively save network bandwidth and storage resources while accurately capturing high-frequency transient information generated during power system faults or disturbances, a variable-rate sampling method can be used for instrument transformers. This method employs a lower sampling rate during steady-state operation of the power system and instantly switches to a higher sampling rate when a transient event is detected.

[0003] The Cumulative Sum Change-Point Detection (CUSUM) algorithm has become a commonly used technique for implementing variable-rate sampling methods due to its low computational cost and sensitivity to subtle but continuous changes in the statistical characteristics of the signal. The basic principle of the CUSUM algorithm is: by analyzing the deviation between the accumulated signal characteristics and its steady-state expected value, when the accumulated sum exceeds a preset threshold, it is determined that a transient event has occurred and high-speed sampling is triggered.

[0004] However, the triggering method of the traditional CUSUM algorithm has a core flaw: its core control parameters, such as the baseline average value of the transformer secondary output signal, the relaxation coefficient used to suppress normal fluctuations, and the threshold used for triggering decisions, are all fixed values ​​set once based on experience during the system commissioning phase. But as a complex dynamic system, the power system's load level, operating mode, and background noise changes are directly reflected in the transformer secondary output signal. For example, the normal fluctuation range of the transformer's "steady-state" signal differs significantly between daytime heavy load periods and nighttime light load periods. If a fixed set of parameters is used, and is set too sluggish to adapt to heavy load conditions, it may fail to capture early, hidden fault characteristics such as inter-turn short circuits in the transformer signal under light load conditions, resulting in missed detections. Conversely, if the parameters are set too sensitive to adapt to light load conditions, normal fluctuations in the transformer output signal during normal load increases or switching operations may be misjudged as transient events, leading to frequent high-speed sampling triggers and failing to achieve the goal of saving resources. Summary of the Invention

[0005] To address the problem that traditional CUSUM algorithms, which use fixed core control parameters due to the complex dynamic nature of power systems, fail to detect early, hidden faults such as inter-turn short circuits in transformer signals under light load conditions if the parameters are set too sluggishly, leading to missed detections. Conversely, if the parameters are set too sensitively, normal fluctuations in the transformer output signal during load increases or switching operations may be misinterpreted as transient events, causing frequent high-speed sampling and wasting resources. This invention provides a digital sampling method and system for transformers.

[0006] In a first aspect, the present invention provides a digital sampling method for a current transformer, which adopts the following technical solution:

[0007] A digital sampling method for instrument transformers includes: acquiring signals output from the secondary side of the instrument transformer for a power system in real time; determining the load fluctuation level at the current moment based on the difference in effective values ​​of signals between all two adjacent moments within the current observation period and the average of the effective values ​​of signals at all moments within a preset observation period; determining the dynamic sensitivity coefficient at the current moment based on the load fluctuation level and the load fluctuation level at the moment corresponding to the current moment under ideal operating conditions; adjusting the benchmark mean of the CUSUM algorithm based on the load fluctuation level to obtain the dynamic benchmark mean at the current moment; and combining the relaxation parameters of the CUSUM algorithm with... The ratio of the dynamic sensitivity coefficients is used as the dynamic relaxation parameter at the current moment; the ratio of the decision threshold of the CUSUM algorithm to the dynamic sensitivity coefficients is used as the dynamic decision threshold at the current moment; based on the cumulative sum of the deviations between the signal and the dynamic reference value in the CUSUM algorithm at the previous moment, the first-order difference value and the mean value of the dynamic reference at the current moment, and the dynamic relaxation parameter, the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm at the current moment is obtained; according to the magnitude of the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm at the current moment and the dynamic decision threshold, the sampling frequency of the current transformer is determined to realize the digital sampling of the current transformer.

[0008] The beneficial effects are as follows: By dynamically adjusting the key parameters (baseline mean, relaxation parameter, and decision threshold) of the CUSUM algorithm through real-time evaluation of the load fluctuation of the secondary output signal of the instrument transformer, the sampling system can automatically adapt to different operating conditions according to the load changes of the power system, improving the applicability and flexibility of the parameters; dynamic sensitivity adjustment enables the system to maintain high sensitivity under light load conditions, thereby more accurately capturing the characteristics of initial hidden faults such as inter-turn short circuits and reducing the possibility of missed detections; under heavy load or normal fluctuations, the system reduces sensitivity to avoid misjudging normal load changes as transient faults, reducing unnecessary high-speed sampling triggers, thereby saving data acquisition and processing resources and effectively reducing resource waste caused by misjudgment; based on the dynamic update mechanism of CUSUM cumulative sum, the system can respond to signal changes more agilely, ensuring that digital sampling can accurately capture anomalies while also adjusting the sampling frequency in a timely manner, enhancing real-time response capability and stability.

[0009] Furthermore, the signal is a current signal.

[0010] Furthermore, the signal is a voltage signal.

[0011] Furthermore, the degree of load fluctuation satisfies:

[0012] In the formula, The degree of load fluctuation at the current moment. This represents the number of times within the preset observation period at the current time. The first observation within the preset observation period at the current time. The effective value of the signal at each moment. The first observation within the current observation period The effective value of the signal at each moment. It is the absolute value symbol.

[0013] The beneficial effects are as follows: by calculating the average of the absolute differences of the signal effective values ​​at adjacent times and dividing it by the average of the signal effective values ​​during the observation period, a normalized representation of the load fluctuation magnitude is achieved, avoiding the problem that the absolute fluctuation value is amplified or reduced with the load magnitude; accurate quantification of the load fluctuation degree can help the system better distinguish between normal load fluctuations and fault signals, providing a more reliable dynamic benchmark for fault detection, improving detection accuracy and reliability, and performing better, especially in the early identification of hidden faults.

[0014] Furthermore, the dynamic sensitivity coefficient satisfies:

[0015] In the formula, This represents the dynamic sensitivity coefficient at the current moment. The degree of load fluctuation at the current moment. This represents the degree of load fluctuation at the current moment under ideal operating conditions. This is a preset adjustment factor used to regulate the degree to which sensitivity changes with load. It is a natural exponential function.

[0016] The beneficial effects are as follows: By constructing a natural exponential function, the sensitivity coefficient exhibits a non-linear response to changes in the ratio of load fluctuation, allowing for more flexible and precise adjustment of the sensitivity of the CUSUM algorithm parameters, avoiding the shortcomings of simple linear adjustment; when the current load fluctuation is higher than the reference load fluctuation under ideal operating conditions, the sensitivity is reduced to avoid misjudgments caused by large fluctuations; conversely, when the load fluctuation is small, the sensitivity is increased, which is more conducive to the capture of early fault characteristics; the reasonable adjustment of the dynamic sensitivity coefficient enables the CUSUM algorithm to take into account both the sensitive capture of minor fault signals and the suppression of normal load fluctuations, thereby effectively improving the accuracy of fault detection and the stability of system operation.

[0017] Furthermore, the ideal operating condition refers to the actual operating state of the power system at each moment when no faults occur.

[0018] Furthermore, the dynamic benchmark mean satisfies:

[0019] In the formula, This is the dynamic baseline mean at the current moment. This is the baseline mean of CUSUM. The degree of load fluctuation at the current moment. This represents the degree of load fluctuation at the current moment under ideal operating conditions. This is a preset limiting factor used to prevent the benchmark average from deviating from the actual reasonable range due to excessive load fluctuations.

[0020] The beneficial effects are as follows: By linearly adjusting the baseline mean of CUSUM according to the relative change in load fluctuation, the baseline mean can dynamically change with load fluctuations, fully reflecting the current true operating status of the system; the limiting factor prevents the baseline mean from deviating from a reasonable range due to excessive load fluctuations, ensuring that the baseline mean changes moderately, maintaining the stability and reliability of the system, and avoiding misjudgments or omissions caused by excessive baseline adjustment; when the load is low, fine-tuning the baseline mean makes it easier for the algorithm to detect weak abnormal signals, enhancing the ability to identify faults early; when the load is high, an appropriate increase in the baseline mean can reduce false alarms caused by normal system fluctuations, and better distinguish between normal fluctuations and fault characteristics; the dynamic baseline mean adjustment method can combine the background characteristics of different monitoring points and different time periods to achieve personalized and real-time parameter optimization, improving the overall sensitivity and applicability of power system monitoring.

[0021] Furthermore, the cumulative sum between the signal and the dynamic reference value at the current moment in the CUSUM algorithm satisfies:

[0022] In the formula, This is the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm at the current moment. This is the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm at the previous time step. The first difference value at the current time. This is the dynamic baseline mean at the current moment. is the dynamic relaxation coefficient at the current moment.

[0023] The beneficial effects are as follows: By combining the first-order difference of the signal at the current moment, the dynamic benchmark mean, and the dynamic relaxation coefficient, the cumulative sum is dynamically adjusted to ensure that the CUSUM algorithm can keenly capture the cumulative trend of abnormal deviations in the signal, thereby improving the accuracy and timeliness of fault detection; the use of a maximum value function to limit the cumulative sum to a non-negative range effectively avoids the impact of negative value accumulation on the detection effect, ensuring that the algorithm focuses on the cumulative abnormal changes and strengthening the response to positive abnormal signals; the use of a dynamic benchmark mean and dynamic relaxation parameters allows the update process of the cumulative sum to adapt to the current load fluctuation, avoiding missed and false alarms caused by the failure of fixed parameter settings, thus enhancing the robustness of CUSUM; the accumulation mechanism enhances the detection capability of weak but continuous abnormal signals, enabling early features of hidden faults such as inter-turn short circuits to be effectively captured, facilitating early fault warning and timely handling.

[0024] Furthermore, determining the sampling frequency of the current transformer includes: in response to the current time when the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm is less than the dynamic decision threshold, determining that the power system is in a steady state, and maintaining the sampling frequency of the current transformer at a preset low sampling rate; otherwise, determining that the power system is in a transient event, and adjusting the sampling frequency of the current transformer to a preset high sampling rate.

[0025] Secondly, the present invention provides a digital sampling system for a current transformer, which adopts the following technical solution:

[0026] A digital sampling system for a current transformer includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned digital sampling method for a current transformer is implemented.

[0027] By adopting the above technical solution, a computer program is generated from the above-mentioned digital sampling method for current transformers and stored in a memory so that it can be loaded and executed by a processor. This allows for the creation of terminal devices based on the memory and processor, making them convenient to use.

[0028] The present invention has the following technical effects:

[0029] (1) Breaking through the limitations of fixed parameters in the traditional CUSUM algorithm, this algorithm calculates the load fluctuation level in real time and generates a dynamic sensitivity coefficient, thereby achieving adaptive adjustment of the benchmark mean, relaxation parameters, and decision threshold. It automatically reduces sensitivity during heavy load periods to avoid missed alarms and increases sensitivity during light load periods to capture hidden fault characteristics, effectively solving the adaptability contradiction of fixed parameters under different operating conditions, significantly reducing the missed alarm and false alarm rates, and improving the ability to identify initial faults such as inter-turn short circuits in instrument transformers.

[0030] (2) By using dynamic decision thresholds to reasonably trigger high-speed sampling, we can avoid false triggering due to normal load fluctuations caused by overly sensitive parameters, reduce invalid sampling operations, and reduce the waste of data storage and processing resources. At the same time, we can ensure that high-speed sampling is accurately started when a real fault occurs, realize the on-demand allocation of sampling resources, and improve system operating efficiency.

[0031] (3) The algorithm parameters are dynamically adjusted based on real-time signal characteristics, so that the sampling method can keep up with the changes in power system load level, operation mode and background noise. No manual re-adjustment of parameters is required. Whether it is the alternation of heavy load during the day and light load at night or sudden load switching operation, the detection performance can be maintained stably, which greatly improves the robustness of digital sampling of transformers in complex power environments.

[0032] (4) Precise sampling triggering and fault feature capture ensure that the acquired secondary side output signal of the instrument transformer can truly reflect the operating status of the equipment. Based on the sampling data optimized by dynamic parameters, it can provide a high-quality data foundation for the health status assessment, life prediction and power system fault diagnosis of the instrument transformer, and help realize the intelligent operation and maintenance of power equipment. Attached Figure Description

[0033] Figure 1 This is a flowchart of a digital sampling method for a current transformer according to an embodiment of the present invention. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] This invention discloses a digital sampling method for current transformers, referring to... Figure 1 This includes steps S1-S6:

[0036] S1: Real-time acquisition of signals output from the secondary side of the current transformer used in the power system.

[0037] It should be noted that the system continuously samples the analog voltage or current signal output from the secondary side of the transformer (transformers are divided into voltage transformers and current transformers, and the voltage or current signal is obtained according to the different types of transformers) at a preset low sampling rate (e.g., 1kHz).

[0038] Specifically, the signal is a current signal.

[0039] Specifically, the signal is a voltage signal.

[0040] S2: Determine the degree of load fluctuation at the current moment.

[0041] It is important to note that the "steady state" of a power system is not absolutely static, but rather exhibits normal fluctuations related to load levels. For example, the effective value of the signal during periods of heavy load may vary slowly over a wide range, while the fluctuation range of the effective value of the signal during periods of light load is relatively small. Using only a fixed threshold to determine whether fluctuations are normal is insufficient to adapt to actual operating conditions under different loads. Therefore, to quantify the current level of calm or activity in the power grid, this step constructs a load fluctuation index. This index obtains the dynamic stability state of the power grid at the current moment by analyzing the relationship between the magnitude of changes in the effective value of the signal and its average value during the observation period.

[0042] The load fluctuation level at the current moment is determined based on the difference in the effective values ​​of the signals at all two adjacent moments within the observation period at the current moment, and the average of the effective values ​​of the signals at all moments within the preset observation period at the current moment.

[0043] Specifically, the load fluctuation level satisfies:

[0044] ;

[0045] In the formula, The degree of load fluctuation at the current moment. This represents the number of times within the preset observation period at the current time. The first observation within the preset observation period at the current time. The effective value of a signal at a given moment (the effective value of a signal is the square root of the average of the squares of the signal values ​​of a periodically changing signal within one power frequency cycle, which is equivalent to the value of a DC signal that generates the same energy in the same time period; where the power frequency cycle can be calculated theoretically, such as 20ms at a standard frequency of 50Hz, or converted after measuring the real-time frequency with tools such as an oscilloscope and frequency counter, or obtained by directly measuring the waveform period). The first observation within the current observation period The effective value of the signal at each moment. It is the absolute value symbol.

[0046] The implementers can set the length of the observation period according to the specific implementation situation, ensuring that the length of the observation period is greater than the length of the power frequency cycle. The effective value of the signal at each moment within the observation period is not limited to the power frequency cycle when it is acquired. For example, the power frequency cycle of the first moment within the observation period is the power frequency cycle composed of multiple moments before the observation period.

[0047] in, This represents the cumulative change in the effective value of the signal during the observation period. A larger value indicates more frequent and greater fluctuations in the effective value of the signal during the observation period, and a more active power grid operation. The larger the value, the more stable the signal; conversely, the smaller the value, the more gradual the change in the effective signal value, and the smoother the power grid operation. The smaller it is. This represents the average effective value of the signal during the observation period. A larger value indicates a higher overall load level during the observation period, meaning the signal itself experiences greater normal fluctuations. Assuming the cumulative change remains constant, then... The smaller the value, the higher the overall load level during the observation period; conversely, the smaller the value, the lower the overall load level during the observation period. In this case, even minor signal anomalies may be early signs of hidden faults. If the cumulative change remains constant, then... The larger it is, the more significant it becomes. and The purpose of multiplication is to scale the average of the signal's effective values ​​over the observation period to match the calculation dimension of the cumulative change (the cumulative change is...). The sum of adjacent differences makes the numerator (cumulative change) and denominator (adjusted average) mathematically proportional, and ultimately the ratio objectively reflects the relative magnitude of the fluctuation relative to the load level.

[0048] S3: Determine the dynamic sensitivity coefficient at the current moment.

[0049] It is important to note that in power system condition monitoring and fault detection, the sensitivity of the detection system to abnormal signals needs to be adapted to the actual operating state of the power grid. If a fixed sensitivity is used, normal large fluctuations may be misinterpreted as fault signals during periods of drastic load fluctuations (active grid), leading to false alarms. Conversely, during periods of moderate load fluctuations (calm grid), insufficient sensitivity may prevent the timely detection of weak early fault signals, resulting in missed alarms. Therefore, this step aims to construct a sensitivity coefficient that can dynamically adjust according to the current load fluctuation level, achieving adaptive adjustment of detection sensitivity. When grid fluctuations are more severe than ideal conditions, the sensitivity is reduced to decrease false alarms; when grid fluctuations are more moderate than ideal conditions, the sensitivity is increased to enhance the ability to detect weak fault signals.

[0050] Based on the load fluctuation level and the load fluctuation level at the current moment corresponding to the current moment under ideal operating conditions, the dynamic sensitivity coefficient at the current moment is determined.

[0051] Specifically, the dynamic sensitivity coefficient satisfies:

[0052] ;

[0053] In the formula, This represents the dynamic sensitivity coefficient at the current moment. The degree of load fluctuation at the current moment. This refers to the time corresponding to the current moment under ideal working conditions (for example, if the current moment is 2 PM on a weekday, then...). The load fluctuation level at 2 PM on a weekday under ideal working conditions. This is a preset adjustment factor used to regulate the degree to which sensitivity changes with load. It is a natural exponential function.

[0054] Specifically, the ideal operating condition refers to the actual operating state of the power system at each moment when no faults occur.

[0055] Implementers can set adjustment factors according to the specific implementation situation, for example, adjustment factors .

[0056] in, The larger the value, the more severe the actual fluctuations in the current power grid are compared to ideal operating conditions. In this case, it is necessary to reduce the detection sensitivity to avoid false triggering. The smaller the value, the better; conversely, the smaller the value, the smoother the current fluctuation compared to ideal operating conditions, requiring increased detection sensitivity to capture weak fault signals. The larger it is. This indicates that a dynamic sensitivity coefficient is obtained by performing a nonlinear transformation on the fluctuation ratio using an exponential function, and then... Performing exponential operations makes the change in sensitivity coefficient exhibit a smooth nonlinear characteristic, avoiding the impact of sudden parameter changes on system stability. equal When the exponent term is 0, Equal to 1, meaning the sensitivity remains at the baseline level; when Greater than When the exponent term is negative, When the value is less than 1, the sensitivity decreases; when Less than When the exponent term is positive, A value greater than 1 indicates increased sensitivity.

[0057] S4: Obtain the dynamic baseline mean, dynamic relaxation parameters, and dynamic decision threshold used in the CUSUM algorithm at the current time.

[0058] It should be noted that in the CUSUM algorithm, the baseline mean is the reference benchmark for determining whether the signal deviates from the steady state. Since the mean of the first-order difference value (i.e., the baseline mean) of the signal varies under different load levels (the mean difference value is larger under heavy load), using a fixed baseline mean would lead to systematic errors in deviation calculation. Therefore, this step dynamically adjusts the baseline mean by introducing the degree of load fluctuation, making it match the actual mean difference value under the current operating conditions.

[0059] The baseline mean of the CUSUM algorithm is adjusted based on the load fluctuation level to obtain the dynamic baseline mean at the current moment.

[0060] Specifically, the dynamic benchmark mean satisfies:

[0061] ;

[0062] In the formula, This is the dynamic baseline mean at the current moment. This is the baseline mean of CUSUM. The degree of load fluctuation at the current moment. This represents the degree of load fluctuation at the current moment under ideal operating conditions. This is a preset limiting factor used to prevent the benchmark average from deviating from the actual reasonable range due to excessive load fluctuations.

[0063] Implementers can set limiting factors according to the specific implementation situation, for example, limiting factors .

[0064] in, This represents the average level of the initial baseline mean. A larger value indicates more significant fluctuations in the signal of the baseline mean. The larger the value, the greater the value; conversely, the smaller the value, the greater the value. The lower the baseline level, the better. This represents the adjustment factor based on the baseline mean of load fluctuations. A larger value indicates that the current load fluctuations are more severe than under ideal operating conditions. The larger the value, the more stable the load fluctuation; conversely, the smaller the value, the more stable the current load fluctuation. The smaller it is.

[0065] It should be noted that in the CUSUM algorithm, the relaxation parameter filters out normal fluctuations in the signal; the cumulative sum is only counted when the deviation exceeds the relaxation parameter. The amplitude of normal fluctuations varies under different load conditions: under heavy loads, normal fluctuations are large, requiring a larger relaxation parameter to tolerate these fluctuations; under light loads, normal fluctuations are small, requiring a smaller relaxation parameter to capture minor deviations. Therefore, the relaxation parameter under the reference condition is adjusted using a dynamic sensitivity coefficient.

[0066] The ratio of the relaxation parameter of the CUSUM algorithm to the dynamic sensitivity coefficient is used as the dynamic relaxation parameter at the current moment. Under light load and high sensitivity requirements (i.e., when the dynamic sensitivity coefficient is greater than 1), the dynamic relaxation parameter decreases, making it more sensitive to small deviations. Under heavy load and low sensitivity requirements (i.e., when the dynamic sensitivity coefficient is less than 1), the dynamic relaxation parameter increases, thus tolerating larger normal fluctuations.

[0067] It should be noted that in the CUSUM algorithm, the decision threshold is the critical value used to determine whether the accumulated deviation has reached the standard for event occurrence. In high-sensitivity scenarios, the threshold needs to be lowered to ensure that weak deviations accumulate to the trigger condition; in low-sensitivity scenarios, the threshold needs to be raised to prevent normal fluctuations from accumulating to false triggers. Therefore, by associating the decision threshold under the reference operating condition with the dynamic sensitivity coefficient, the threshold can be dynamically adjusted according to sensitivity requirements.

[0068] The ratio of the decision threshold of the CUSUM algorithm to the dynamic sensitivity coefficient is used as the dynamic decision threshold at the current moment. Under light load and high sensitivity requirements (i.e., when the dynamic sensitivity coefficient is greater than 1), the dynamic decision threshold is lowered, making it easier for anomalies to accumulate and trigger. Under heavy load and low sensitivity requirements (i.e., when the dynamic sensitivity coefficient is less than 1), the dynamic decision threshold is increased, enhancing the anti-interference capability of the CUSUM algorithm and reducing false triggers caused by normal fluctuations.

[0069] S5: Determine the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm at the current moment.

[0070] It should be noted that the cumulative summation is the core of the CUSUM algorithm. Its function is to continuously accumulate the deviation between the observed value and the benchmark value. When the accumulated deviation exceeds a threshold, an event response is triggered. Since the power grid operating conditions are dynamically changing, the calculation of the accumulated deviation must be based on real-time updated dynamic benchmark averages, dynamic relaxation parameters, etc., to accurately reflect the degree of signal deviation under the current operating conditions. Therefore, dynamic parameters are introduced into the cumulative summation step in this step, enabling the deviation accumulation process to adapt to the power grid operating status in real time.

[0071] Based on the cumulative sum of the deviations between the signal and the dynamic reference value in the CUSUM algorithm at the previous time step, the first-order difference value and the mean value of the dynamic reference at the current time step, and the dynamic relaxation parameter, the cumulative sum between the signal and the dynamic reference value at the current time step is obtained.

[0072] Specifically, the cumulative sum between the signal and the dynamic reference value at the current moment in the CUSUM algorithm satisfies:

[0073] ;

[0074] In the formula, This is the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm at the current moment. This is the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm at the previous time step. The first difference value at the current time. This is the dynamic baseline mean at the current moment. is the dynamic relaxation coefficient at the current moment.

[0075] in, This represents the cumulative sum (initially 0) of the deviation between the first-order difference of the signal at the current moment and the dynamic benchmark mean and dynamic relaxation parameters, plus the deviation from the previous moment. A larger value indicates a higher degree of accumulation of current and historical deviations, and a greater likelihood that the signal deviates from its steady state. Therefore, when this value is greater than 0... The larger the value, the higher the probability of the event occurring; if the value is less than 0, it means that the deviation has not exceeded the relaxation parameter, the cumulative sum is reset to 0, that is, it is considered that there is no significant deviation at present.

[0076] S6: Determine the sampling frequency of the current transformer based on the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm at the current moment and the magnitude of the dynamic decision threshold, so as to realize the digital sampling of the current transformer.

[0077] Specifically, determining the sampling frequency of the current transformer includes:

[0078] If the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm at the current moment is less than the dynamic decision threshold, the power system is considered to be in a steady state, and the sampling frequency of the instrument transformer is maintained at a preset low sampling rate; otherwise, the power system is considered to be in a transient event, and the sampling frequency of the instrument transformer is adjusted to a preset high sampling rate.

[0079] Implementers can set a high sampling rate according to the specific implementation situation, for example, a high sampling rate of 100kHz.

[0080] This invention also discloses a digital sampling system for a current transformer, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a digital sampling method for a current transformer according to the present invention.

[0081] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0082] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method of digital sampling of a transformer, characterized by, The method comprises: obtaining, in real time, a signal of a secondary side output of a mutual inductor for a power system; determining a load fluctuation degree of a current time according to a difference between effective values of signals of two adjacent times within an observation period of the current time and a mean value of effective values of signals of all times within a preset observation period of the current time; determining a dynamic sensitivity coefficient of the current time according to the load fluctuation degree and a load fluctuation degree of a time corresponding to the current time under an ideal working condition, wherein the dynamic sensitivity coefficient satisfies: ； is a dynamic sensitivity coefficient for the current time, is a load fluctuation degree for the current time, is a load fluctuation degree for the time corresponding to the current time under ideal working conditions, is a preset adjustment factor for adjusting the degree of change of the sensitivity with the load, is a natural exponential function; adjusting a reference mean value of a CUSUM algorithm based on the load fluctuation degree to obtain a dynamic reference mean value of the current time; taking a ratio of a relaxation parameter of the CUSUM algorithm to the dynamic sensitivity coefficient as a dynamic relaxation parameter of the current time; taking a ratio of a decision threshold of the CUSUM algorithm to the dynamic sensitivity coefficient as a dynamic decision threshold of the current time; and the dynamic reference mean value satisfies: ； a dynamic reference average value for the current time, a reference average value for the CUSUM, a preset limiting factor for avoiding the reference average value deviating from the actual reasonable range due to excessive load fluctuation. obtaining a cumulative sum of a deviation between the signal and the dynamic reference value in the CUSUM algorithm of a previous time, a first-order difference value of the current time and the dynamic reference mean value and the dynamic relaxation parameter, to obtain a cumulative sum of the deviation between the signal and the dynamic reference value in the CUSUM algorithm of the current time; determining a sampling frequency of the mutual inductor according to a size of the cumulative sum of the deviation between the signal and the dynamic reference value in the CUSUM algorithm of the current time and the dynamic decision threshold, to realize digital sampling of the mutual inductor.

2. A method of digital sampling of a transformer according to claim 1, characterized in that, The signal is a current signal.

3. The method of claim 1, wherein, The signal is a voltage signal.

4. The method of claim 1, wherein, The load fluctuation degree satisfies: ； In the formula, is the load fluctuation degree at the current time, is the number of times in the preset observation period at the current time, is the effective value of the signal at the first time in the preset observation period at the current time, is the effective value of the signal at the first time in the preset observation period at the current time, is the absolute value symbol.

5. The method of digital sampling of a transformer of claim 1, wherein, The ideal working condition is an actual operating state of each time when the power system has no fault.

6. The method of digital sampling of a transformer of claim 1, wherein, The cumulative sum of the deviation between the signal and the dynamic reference value in the CUSUM algorithm of the current time satisfies: ； wherein is the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm at the current time instant, is the cumulative sum between the signal and the dynamic reference value in the CUSUM algorithm at the previous time instant of the current time instant, is the first order difference value at the current time instant, is the dynamic reference mean value at the current time instant, is the dynamic relaxation coefficient at the current time instant.

7. The method of digital sampling of a transformer of claim 1, wherein, The determination of the sampling frequency of the mutual inductor comprises: in response to the cumulative sum of the deviation between the signal and the dynamic reference value in the CUSUM algorithm of the current time being less than the dynamic decision threshold, it is determined that the power system is in a steady state, and the sampling frequency of the mutual inductor is maintained at a preset low sampling rate; otherwise, it is determined that the power system is in a transient event, and the sampling frequency of the mutual inductor is adjusted to a preset high sampling rate.

8. A mutual inductor digital sampling system characterized by, The method comprises: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a mutual inductor digital sampling method according to any one of claims 1-7 is realized.

Images