Partial discharge raw data filtering method and system

Abstract

The present application relates to a partial discharge raw data filtering method, aiming at effectively eliminating the inherent bias of partial discharge acquisition board, improving the accuracy and reliability of partial discharge signal monitoring. The method includes three main steps: first, by converting the collected raw data from code value to millivolt value, and extracting several cycles of data in a specific format, a data matrix to be filtered is constructed; second, the baseline threshold and fluctuation threshold are determined by using the baseline calculation function and fluctuation coefficient, providing necessary parameter support for data filtering; finally, according to the data to be filtered and the calculated threshold, the filtering process is implemented, and the partial discharge data without inherent bias is obtained. The present application has strong adaptability and universality, can be applied in various power equipment, improves the accuracy of partial discharge monitoring, reduces the risk of misjudgment, provides scientific basis for the health management and maintenance of power equipment, has significant economic benefits and safety guarantee.

Description

Technical Field

[0001] This invention relates to a method, system, storage medium, and computing device for filtering raw partial discharge data, specifically belonging to the field of power equipment fault assessment technology. Background Technology

[0002] Partial discharge is a crucial indicator of the degradation of insulation materials in electrical equipment, and its monitoring and analysis are essential for ensuring the safe and stable operation of power equipment. Partial discharge typically occurs inside or on the surface of the insulation materials of electrical equipment, especially in high-voltage equipment, and can reflect the health status of the insulation system and its potential failure risks. Therefore, accurately capturing and analyzing partial discharge signals is of great significance for maintenance management in the power industry.

[0003] However, the inherent bias of the partial discharge acquisition board has become a critical technical challenge in partial discharge detection. These inherent biases can originate from various sources, including but not limited to the electrical characteristics of the equipment itself, environmental factors (such as temperature and humidity), noise in the measurement circuitry, and interference during signal transmission. These factors often lead to significant errors in the raw data, affecting subsequent data analysis and judgment, potentially resulting in misjudgments or omissions of partial discharge activity, thus posing a potential risk to the safe operation of the equipment.

[0004] Therefore, there is an urgent need for an effective filtering method to systematically identify and remove the inherent bias of the partial discharge acquisition board. Summary of the Invention

[0005] Purpose of the invention: This invention provides a method, system, storage medium, and electronic device for filtering raw partial discharge data, in order to improve the accuracy of partial discharge assessment of power equipment and identify and remove the inherent bias of the partial discharge acquisition board.

[0006] Technical solution:

[0007] This invention proposes a method for filtering raw partial discharge data, comprising:

[0008] The raw partial discharge data in code value format is collected, converted into calculable data in millivolts, the calculation period is set, and calculable data of several periods are extracted in a fixed format to construct the data to be filtered.

[0009] The preset deviation coefficient and the data to be filtered are fed into the baseline calculation function. Elements in the data to be filtered whose volatility is greater than the deviation coefficient are ignored. The average value of the elements that are not ignored is output as the baseline threshold. The volatility is the ratio of the absolute value of the difference between the current element and the mean of all elements in the data to be filtered to the standard deviation of all elements in the data to be filtered.

[0010] Multiply the baseline threshold by a preset fluctuation coefficient to obtain the fluctuation threshold; calculate the deviation rate of each element in the data to be filtered, compare the deviation rate with the fluctuation threshold, remove elements in the data to be filtered whose deviation rate is less than or equal to the fluctuation threshold, and update the values ​​of all remaining elements to the absolute value of the difference between the current element and the fluctuation threshold, as the filtered partial discharge data.

[0011] Furthermore, the raw data is converted into computable data in millivolts using the following formula:

[0012]

[0013] in, For computable data, The code value corresponding to the original data. As the base code value, This is the base millivolt value.

[0014] Furthermore, the construction of the data to be filtered includes:

[0015] The data to be filtered is stored in a data matrix to be filtered, represented as follows:

[0016]

[0017] in, The data matrix to be filtered. For the corresponding element, Number the elements. The number of elements , The total number of phases within one period. This represents the total number of cycles.

[0018] Furthermore, the baseline calculation function includes:

[0019] Calculate the data matrix to be filtered Mean of all elements Standard deviation ;

[0020] Define a count (count) and a sum (sum), and set their initial values ​​to 0; iterate through the matrix of data to be filtered. The control variable for iterating is i, initialized to 1; check the i-th element. Check if the volatility is greater than the deviation coefficient; if so, skip the element; if not, update the sum value to the sum of the original sum value and the sum of the deviation coefficient. The sum of the numbers, count increments by 1; when i is greater than 1... End the traversal at this time;

[0021] Calculate baseline threshold The baseline threshold is the ratio of the summation statistic (sum) to the count statistic (count);

[0022] The i-th element volatility Represented as:

[0023] .

[0024] Furthermore, the process of removing elements from the data to be filtered whose volatility is less than or equal to the volatility threshold includes:

[0025] Based on baseline threshold and preset fluctuation coefficient Calculate the fluctuation threshold The formula is:

[0026]

[0027] Iterate through the data to be filtered and calculate the deviation rate of the corresponding element i from the baseline threshold. The formula is:

[0028]

[0029] Determine the deviation rate Is it greater than the fluctuation threshold? ;like If the value of the corresponding element to be filtered is subtracted from the baseline threshold; if If the value is 0, then the corresponding data to be filtered will be replaced with 0; this is represented as:

[0030]

[0031] The result of the traversal This is the filtered partial discharge data.

[0032] This invention also proposes a partial discharge raw data filtering system, comprising:

[0033] The data construction module is used to collect partial discharge raw data in code value format, convert the raw data into calculable data in millivolts, set the calculation period, and extract calculable data of several periods in a fixed format to construct the data to be filtered.

[0034] The volatility threshold module is used to calculate the volatility of each element in the data to be filtered, compare the volatility with a preset deviation coefficient, ignore elements in the data to be filtered whose volatility is greater than the deviation coefficient, and use the average value of all the elements that are not ignored as the volatility threshold; the volatility is the ratio of the absolute value of the difference between the current element and the mean of all elements in the data to be filtered to the standard deviation of all elements in the data to be filtered.

[0035] The filtering module is used to calculate the deviation rate of each element in the data to be filtered, compare the deviation rate with the fluctuation threshold, remove elements in the data to be filtered whose deviation rate is less than or equal to the fluctuation threshold, and update the value of all retained elements to the absolute value of the difference between the current element and the fluctuation threshold, as the filtered partial discharge data.

[0036] Furthermore, the conversion of raw data into computable data involves converting code values ​​into millivolt values, calculated using the following formula:

[0037]

[0038] in, For computable data, The code value corresponding to the original data. As the base code value, This is the base millivolt value.

[0039] Furthermore, the construction of the data to be filtered includes:

[0040] The data to be filtered is stored in a data matrix to be filtered, represented as follows:

[0041]

[0042] in, The data matrix to be filtered. For the corresponding element, Number the elements. The number of elements , The total number of phases within one period. This represents the total number of cycles.

[0043] Furthermore, the baseline calculation function includes:

[0044] Calculate the data matrix to be filtered mean Standard deviation ;

[0045] Define a count (count) and a sum (sum), and set their initial values ​​to 0; iterate through the matrix of data to be filtered. The control variable for iterating is i, initialized to 1; check the i-th element. Check if the volatility is greater than the deviation coefficient; if so, skip the element; if not, update the sum value to the sum of the original sum value and the sum of the deviation coefficient. The sum of the numbers, count increments by 1; when i is greater than 1... End the traversal at this time;

[0046] Calculate baseline threshold The baseline threshold is the ratio of the summation statistic (sum) to the count statistic (count);

[0047] The i-th element volatility Represented as:

[0048] .

[0049] Furthermore, the filtering function includes:

[0050] Based on baseline threshold and preset fluctuation coefficient Calculate the fluctuation threshold The formula is:

[0051]

[0052] Iterate through the data to be filtered and calculate the deviation rate of the corresponding element i from the baseline threshold. The formula is:

[0053]

[0054] Determine if the deviation rate is greater than the fluctuation threshold. If the value is greater than the baseline threshold, the corresponding data to be filtered is subtracted from the baseline threshold; if the value is less than or equal to the baseline threshold, the corresponding data to be filtered is replaced with 0. This is represented as:

[0055]

[0056] The result of the traversal This is the filtered partial discharge data.

[0057] Beneficial effects:

[0058] This invention effectively eliminates the inherent bias of the partial discharge acquisition board through systematic data conversion and threshold calculation, accurately extracts the true characteristics of the partial discharge signal, ensures the authenticity and validity of the acquired data, and improves the accuracy of judging the health status of power equipment.

[0059] This invention also allows for the dynamic adjustment of filtering strategies based on the partial discharge characteristics of different devices and operating conditions through the calculation of baseline thresholds and fluctuation thresholds. This maximizes the retention of effective signals and eliminates interference noise, providing important data support for the long-term health management of equipment, thereby effectively extending the service life of the equipment and reducing maintenance costs. Attached Figure Description

[0060] Figure 1 Flowchart of the partial discharge raw data filtering method;

[0061] Figure 2 A waveform data obtained from sampling by a partial discharge sampling board;

[0062] Figure 3 This is a waveform data obtained by sampling using the method of the present invention. Detailed Implementation

[0063] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0064] Example 1

[0065] like Figure 1 As shown, a method for filtering raw partial discharge data includes the following steps:

[0066] Step 1, data conversion.

[0067] The raw partial discharge data (code values) acquired by the partial discharge acquisition board is converted into computable data (millivolt values). Several cycles of computable data are extracted according to a given format to construct the data to be filtered, which is then stored in a data matrix. The cycle is the alternating current cycle; for a 50Hz alternating current, the cycle is 20 milliseconds. One cycle consists of 360 points, meaning one partial discharge data point is taken from each phase to form one cycle. Several cycles are then used to construct the required filtering data. The conversion formula for the raw partial discharge data into computable data is as follows:

[0068]

[0069] in, For computable data, For the code value corresponding to the computable data, As the base code value, As the reference millivolt value, in this embodiment, the sampling board AD bit depth is 14 bits, and the range is -2V to 2V, so the reference code value is 2. 14 / 2=8192, the reference millivolt value is 2000mV.

[0070] The data matrix to be filtered can be represented as:

[0071]

[0072] in, The data matrix to be filtered. For the corresponding element, Number the elements. The number of elements , The total number of phases within one period. This represents the total number of cycles.

[0073] Step 2, threshold calculation.

[0074] Combine the deviation coefficient, the data matrix C to be filtered, and the length of the data to be filtered (i.e., the number of elements in the data matrix to be filtered in step 1). The data is fed into a baseline calculation function to obtain a baseline threshold. Based on the baseline threshold and a preset fluctuation coefficient, a fluctuation threshold is calculated. The deviation coefficient is a predefined threshold; points exceeding the deviation coefficient are filtered out, thus achieving the filtering purpose.

[0075] The deviation coefficient, the data to be filtered, and the length of the data to be filtered are fed into the baseline calculation function to obtain the baseline threshold. The calculation process of the baseline calculation function is as follows: first, the mean and standard deviation of the data to be filtered are calculated; then, the data to be filtered is traversed, points whose volatility exceeds the deviation coefficient are removed, and the mean is recalculated to obtain the baseline threshold.

[0076] The calculation method for the mean and standard deviation of the data to be filtered can be expressed as follows:

[0077]

[0078] in, The mean of the data to be filtered is... denoted as the standard deviation of the data to be filtered.

[0079] The process of iterating through the data to be filtered, ignoring points where the volatility exceeds the deviation coefficient, is generally a one-dimensional loop structure. First, a count (count=0) and a summation count (sum=0) are defined. Then the loop begins, with the control variable i, ranging from 1 to... The value increments by 1 each time. Then, a judgment process is initiated: it checks if the volatility corresponding to the element is greater than the deviation coefficient. If it is, no action is taken. If it is less than or equal to the deviation coefficient, then `sum+=` is executed. The count is incremented by 1. The judgment ends, and the loop ends. Finally, the baseline threshold is... The calculation formula is expressed as:

[0080]

[0081] The method for calculating the volatility of the corresponding element is expressed as follows:

[0082]

[0083] in, This represents the volatility of the corresponding element.

[0084] The fluctuation threshold is calculated based on the baseline threshold and a preset fluctuation coefficient. The calculation method for the fluctuation threshold is as follows: ,in, For fluctuation threshold, The fluctuation coefficient is 0.05 in this embodiment.

[0085] Step 3, filtering.

[0086] The data matrix C to be filtered and the length of the data to be filtered from step 1 are... The baseline threshold B and fluctuation threshold T are fed into the filtering function to finally obtain the filtered partial discharge data.

[0087] The calculation process in this step is generally a one-dimensional loop structure. First, it iterates through the data to be filtered, calculating the deviation rate of the corresponding element from the baseline threshold. Then, it determines whether the deviation rate is greater than the fluctuation threshold. If it is greater, the baseline threshold is subtracted from the corresponding data to be filtered. If it is less than or equal to the baseline threshold, the corresponding data to be filtered is replaced with 0. The resulting new data is the filtered partial discharge data.

[0088] The formula for calculating the deviation rate of the corresponding element is expressed as follows:

[0089]

[0090] in, This represents the deviation rate of the corresponding element.

[0091] The method for constructing the filtered partial discharge data elements is expressed as follows:

[0092]

[0093] in, This corresponds to the filtered partial discharge data.

[0094] To verify the improvement of the method of the present invention compared with the prior art, this embodiment processes the raw data of the same set of partial discharge signals under the same experimental conditions, and the results are as follows: Figure 2 , Figure 3 As shown. Among them, Figure 2 This is a waveform data graph collected by a partial discharge sampling board used in existing technology. Each point in the graph represents a set of partial discharge data and a waveform location. A noticeable interfering horizontal line is observed in the collected waveform data graph. This is an inherent bias of the sampling board and will negatively impact the diagnosis of partial discharge types. Figure 3 The image shows the waveform data obtained after processing the original partial discharge data using the method of this invention. As can be seen from the image, compared with the prior art, the method of this invention filters out interference and eliminates inherent bias through algorithms, resulting in more accurate waveform data.

[0095] Example 2

[0096] Based on the technical solution of Embodiment 1, the present invention also discloses a computer software system for the above method, namely a partial discharge raw data filtering system, comprising:

[0097] Data conversion module: Converts the raw partial discharge data (code value) collected by the partial discharge acquisition board into computable data (millivolt value), and extracts computable data from several cycles according to a given format to construct the data to be filtered.

[0098] Threshold calculation module: The deviation coefficient, the data to be filtered, and the length of the data to be filtered are fed into the baseline calculation function to obtain the baseline threshold, and the fluctuation threshold is calculated based on the baseline threshold and the preset fluctuation coefficient.

[0099] Filtering module: Calculates the fluctuation threshold, removes elements in the data to be filtered whose fluctuation rate is less than or equal to the fluctuation threshold, and updates the values ​​of all retained elements to the corresponding deviation rate, which are then used as the filtered partial discharge data.

Claims

1. A method for filtering raw partial discharge data, characterized in that, include: The raw partial discharge data in code value format is collected, converted into calculable data in millivolts, the calculation period is set, and calculable data of several periods are extracted in a fixed format to construct the data to be filtered. The preset deviation coefficient and the data to be filtered are fed into the baseline calculation function. Elements in the data to be filtered whose volatility is greater than the deviation coefficient are ignored. The average value of the elements that are not ignored is output as the baseline threshold. The volatility is the ratio of the absolute value of the difference between the current element and the mean of all elements in the data to be filtered to the standard deviation of all elements in the data to be filtered. The baseline calculation function includes: Calculate the data matrix to be filtered Mean of all elements Standard deviation ; Define a count (count) and a sum (sum), and set their initial values ​​to 0; iterate through the matrix of data to be filtered. The control variable being iterated is i, with an initial value of 1; the data matrix to be filtered is determined. The i-th element Check if the volatility is greater than the deviation coefficient; if so, skip the element; if not, update the sum value to the sum of the original sum value and the sum of the deviation coefficient. The sum of the numbers, count increments by 1; when i is greater than 1... End the traversal at that time. The number of elements; Calculate baseline threshold The baseline threshold is the ratio of the summation statistic (sum) to the count statistic (count); The i-th element volatility Represented as: Multiply the baseline threshold by a preset fluctuation coefficient to obtain the fluctuation threshold; calculate the deviation rate of each element in the data to be filtered, compare the deviation rate with the fluctuation threshold, and remove elements in the data to be filtered whose deviation rate is less than or equal to the fluctuation threshold, including: Based on baseline threshold and preset fluctuation coefficient Calculate the fluctuation threshold The formula is: Iterate through the data to be filtered and calculate the deviation rate of the corresponding element i from the baseline threshold. The formula is: Determine the deviation rate Is it greater than the fluctuation threshold? ;like If the value of the corresponding element to be filtered is subtracted from the baseline threshold; if If the value is 0, then the corresponding data to be filtered will be replaced with 0; this is represented as: The result of the traversal This is the filtered partial discharge data.

2. The partial discharge raw data filtering method according to claim 1, characterized in that, The process of converting the raw data into computable data in millivolts is described by the following formula: in, For computable data, The code value corresponding to the original data. As the base code value, This is the base millivolt value.

3. The partial discharge raw data filtering method according to claim 2, characterized in that, The construction of the data to be filtered includes: The data to be filtered is stored in a data matrix to be filtered, represented as follows: in, The data matrix to be filtered. For the corresponding element, Number the elements. The number of elements , The total number of phases within one period. This represents the total number of cycles.

4. A partial discharge raw data filtering system, characterized in that, include: The data construction module is used to collect partial discharge raw data in code value format, convert the raw data into calculable data in millivolts, set the calculation period, and extract calculable data of several periods in a fixed format to construct the data to be filtered. The volatility threshold module is used to input the preset deviation coefficient and the data to be filtered into the baseline calculation function, ignore elements in the data to be filtered whose volatility is greater than the deviation coefficient, and output the average value of the elements that are not ignored as the baseline threshold; the volatility is the ratio of the absolute value of the difference between the current element and the mean of all elements in the data to be filtered to the standard deviation of all elements in the data to be filtered. The baseline calculation function includes: Calculate the data matrix to be filtered Mean of all elements Standard deviation ; Define a count (count) and a sum (sum), and set their initial values ​​to 0; iterate through the matrix of data to be filtered. The control variable being iterated is i, with an initial value of 1; the data matrix to be filtered is determined. The i-th element Check if the volatility is greater than the deviation coefficient; if so, skip the element; if not, update the sum value to the sum of the original sum value and the sum of the deviation coefficient. The sum of the numbers, count increments by 1; when i is greater than 1... End the traversal at that time. The number of elements; Calculate baseline threshold The baseline threshold is the ratio of the summation statistic (sum) to the count statistic (count); The i-th element volatility Represented as: The baseline threshold is multiplied by a preset fluctuation coefficient to obtain the fluctuation threshold; The filtering module is used to calculate the deviation rate of each element in the data to be filtered, compare the deviation rate with the fluctuation threshold, and remove elements in the data to be filtered whose deviation rate is less than or equal to the fluctuation threshold, including: Based on baseline threshold and preset fluctuation coefficient Calculate the fluctuation threshold The formula is: Iterate through the data to be filtered and calculate the deviation rate of the corresponding element i from the baseline threshold. The formula is: Determine the deviation rate Is it greater than the fluctuation threshold? ;like If the value of the corresponding element to be filtered is subtracted from the baseline threshold; if If the value is 0, then the corresponding data to be filtered will be replaced with 0; this is represented as: The result of the traversal This is the filtered partial discharge data.

5. The partial discharge raw data filtering system according to claim 4, characterized in that, The process of converting raw data into computable data involves converting code values ​​into millivolt values, calculated using the following formula: in, For computable data, The code value corresponding to the original data. As the base code value, This is the base millivolt value.

6. The partial discharge raw data filtering system according to claim 5, characterized in that, The construction of the data to be filtered includes: The data to be filtered is stored in a data matrix to be filtered, represented as follows: in, The data matrix to be filtered. For the corresponding element, Number the elements. The number of elements , The total number of phases within one period. This represents the total number of cycles.

Images