Online detection data processing method for distribution transformer

Abstract

The present invention discloses a method for processing online detection data of a distribution transformer, comprising the following steps: S1, synchronizing the time state of a host and a slave after starting, so that the host and the slave have the same time scale; S2, respectively collecting data of several groups of input and output ends of the distribution transformer through the host and the slave, and making the collected data associated with the time scale; S3, synchronizing several groups of data collected by the host and the slave to the same terminal, cleaning, supplementing and matching the data based on the correlation between the data and the time scale, and calculating to obtain the final result. The present invention generates a priori value that can be used to assist in evaluating the consistency of the overall data through the correlation between all collected data and the time scale. By judging the priori value, it can assist in judging whether the data collected by the host and the slave have relative consistency in time, so that the performance parameters characterizing the distribution transformer can be calculated more accurately.

Description

Technical Field

[0001] The invention belongs to the technical field of transformer detection, and in particular relates to a method for processing online detection data of a distribution transformer. Background Art

[0002] In the prior art, transformer online detection, also known as non-invasive performance detection, has the core idea of ​​monitoring the operating status of the distribution transformer and detecting its technical parameters by collecting and analyzing data information at the inlet and outlet of the distribution transformer without entering the equipment.

[0003] A Chinese patent application with prior art application number 201910640236.4 discloses an online detection method for the rated capacity of a distribution transformer. The method collects the input and output voltage u1(t), u2(t) and current i1(t), i2(t) data of the distribution transformer online through voltage and current transformers, and then calculates the collected data according to a preset data processing method to obtain the performance parameters of the distribution transformer.

[0004] In terms of hardware, the implementation of the online detection method for the distribution transformer requires the configuration of at least one host, one slave, and a current clamp and voltage connector connected to the host, as well as a current clamp and voltage connector connected to the slave. The current and voltage data of the input and output ends of the distribution transformer are collected simultaneously by the host and the slave respectively and processed after data transmission.

[0005] Since the transformer usually has a certain volume, in order to facilitate the detection operation, the host and the slave are usually separated, and data is transmitted between the host and the slave through wireless transmission methods such as Bluetooth and 4G.

[0006] However, after receiving the test data, the current test equipment can only calculate the collected data according to the preset data processing method, but cannot verify or judge the time synchronization of the received test data, which leads to the possibility of time differences and lags in the data collected by the master and slave machines. When the tested equipment and the test equipment are basically stable, the performance parameters of the transformer can be calculated and characterized to a certain extent, but when the tested equipment and the test equipment are in an unstable state, this method of using data without time synchronization for calculation will also affect the accuracy of the calculation results. Summary of the invention

[0007] In view of the above technical problems, this application proposes a method for processing online detection data of distribution transformers. The specific technical solution is as follows:

[0008] The method for processing online detection data of a distribution transformer comprises the following steps:

[0009] S1, after starting up the master and slave, synchronize the time status so that the master and slave have the same time scale;

[0010] S2, respectively collecting data of the input and output ends of several groups of distribution transformers through the host and the slave, and associating the collected data with the time scale;

[0011] S3. Synchronize several groups of data collected by the host and the slave to the same terminal, clean, supplement and match the data based on the correlation between the data and the time scale, and calculate to obtain the final result.

[0012] Furthermore, in step S1, time state synchronization mainly includes:

[0013] S1-1, after power-on, the host and slave establish communication through matching and synchronize time;

[0014] S1-2, the host generates a time scale sequence, the time scale sequence includes at least an initial value P, a filling step H and a scanning step ΔT, and the time scale sequence can be automatically extended and filled by the scanning step ΔT and the filling step H;

[0015] S1-3. The host transmits the generated time scale sequence to the slave, which stores and calls it.

[0016] Furthermore, the master uses the self-determined time as a time scale to synchronize the time to the slave.

[0017] Furthermore, step S1-1 and step S1-2 are executed in a swapped order or simultaneously.

[0018] Further, in step S2, the master and slave call their time scale sequences each time they collect a data, and add the current sequence value to the data to form a set A1 = {(K 11 ,P 11 ),(K 12 ,P 12 ),(K 13 ,P 13 )......(K 1n ,P 1n )} and A2={(K 21 ,P 21 ),(K 22 ,P 22 ),(K 23 ,P 23 )......(K 2n ,P 2n)}, where K represents the collected data, P represents the sequence value corresponding to the data, the subscript value "1" corresponds to the host, the subscript value "2" corresponds to the slave, and n represents the amount of data in the data set.

[0019] Furthermore, step S3 includes the following steps:

[0020] S3-1, extract the sequence values ​​in the data sets A1 and A2 respectively to form two priori sets, respectively F1 = {P 11 ,P 12 ,P 13 ......P 1n} and F2={P 21 ,P 22 ,P 23 ......P 2n};

[0021] S3-2, based on all elements in the two sets, calculate the prior values ​​M1 and M2;

[0022] S3-3. Take any element P in M1 and the prior set F1 1i Pack and convert any element P in M2 and the prior set F2 2i Pack and then synchronize to the same terminal for comparison to determine whether M1 and M2 are equal, and P 1i and P 2i whether they are equal;

[0023] If they are equal, the data collected in data sets A1 and A2 are packaged and synchronized to the same terminal, K of the collected data is extracted, and the data with equal sequence values ​​are grouped for fitting calculation;

[0024] If P 1i and P 2i If M1 and M2 are not equal, then mark P 1i and P 2i The confidence level is 1, and step S3-4 is executed;

[0025] If P 1i and P 2i Not equal, M1 and M2 are not equal, then mark P 1i and P 2i The confidence level is 0, and step S3-5 is executed;

[0026] S3-4, P 1i As an element to construct the confidence set Z1, any element P of the remaining elements in the set F1 1j Add to set Z1, and add P 2i As an element to construct the confidence set Z2, any element P of the remaining elements in the set F2 2jAdd to set Z2, calculate the prior values ​​Y1 and Y2, synchronize Y1 and Y2 to the same terminal for comparison, and determine whether Y1 and Y2 are equal;

[0027] If yes, mark P 1j and P 2j The confidence level is 1, and steps S3-4 are repeated;

[0028] If otherwise, mark P 1j and P 2j The confidence level is 0, and step S3-5 is executed;

[0029] S3-5, discard the elements with a confidence level of 0, recalculate the prior values ​​M1 and M2, and replace any element P of the remaining elements in the set F1 with 1j and M1, and any element P of the remaining elements in the set F2 2j and M2 are packaged, and then synchronized to the same terminal for comparison to determine whether M1 and M2 are equal, and P 1j and P 2j whether they are equal;

[0030] If they are equal, the data collected in the data sets A1 and A2 corresponding to the updated F1 and F2 are packaged and synchronized to the same terminal, K of the collected data is extracted, and the data with equal sequence values ​​are grouped for fitting calculation;

[0031] If P 1j and P 2j If M1 and M2 are not equal, then mark P 1j and P 2j The confidence level is 1, and step S3-4 is executed;

[0032] If P 1j and P 2j Not equal, M1 and M2 are not equal, then mark P 1j and P 2j The confidence level is 0, and steps S3-5 are repeated.

[0033] Furthermore, in step S3-1, the sequence values ​​in the data sets A1 and A2 are extracted and arranged in ascending or descending order to form two prior sets.

[0034] Further, in step S3-2, M1=H(g(F1)), where g(F1)=P 11 P 12 ...P 1n +P 11 P 12 ...P 1(n-1) +...+P 11 P 12 +P 11; M2=H(g(F2)), where g(F2)=P 21 P 22 ...P 2n +P 21 P 22 ...P 2(n-1) +...+P 21 P 22 +P 21 ; In the above formula, M is the hash value obtained by calculating the hash function.

[0035] Furthermore, in step S3-4, the number of elements in confidence sets Z1 and Z2 is a preset value. When the number of elements in confidence sets Z1 and Z2 reaches the preset value, step S3-4 is terminated, and sets A1 and A2 are updated based on sets Z1 and Z2. The data collected in sets A1 and A2 are packaged and synchronized to the same terminal, K of the collected data is extracted, and data with equal sequence values ​​are grouped for fitting calculations.

[0036] Furthermore, after the confidence of the elements in sets F1 and F2 is set to 1 or 0 in step S3, a discard operation is performed, and new data is collected and supplemented by the master and slave to update sets A1 and A2; the number of elements in sets F1 and F2 is the same as that in sets Z1 and Z2.

[0037] The beneficial effects of the present invention are:

[0038] (1) After synchronizing the time states of the host and the slave, the present invention associates and marks the collected data based on the same time scale, uses the time difference of data collection to perform outlierization on the relatively stable parameters of the transformer working performance, and generates a priori value that can be used to assist in evaluating the consistency of the overall data through the correlation between all collected data and the time scale. By judging the priori value, it is possible to assist in judging whether the data collected by the host and the slave are relatively consistent in time, so that the performance parameters characterizing the distribution transformer can be calculated more accurately;

[0039] (2) Based on the time state synchronization and the setting of the time ruler, the automatic extension and filling feature of the time ruler can be used to collect data without restriction as needed. The scanning step length ΔT used for automatic extension and filling makes the data collected within a more reasonable time frequency band be considered as valid data that meets the time relative consistency, thereby taking into account both the detection efficiency and the time relative consistency of the data.

[0040] (3) When judging by a priori value, the sequence value corresponding to the data is added. On the one hand, it can eliminate the accidental defect of using a single priori value to assist in judging whether the data collected by the host and the slave are relatively consistent in time. On the other hand, when the data does not meet the relative consistency in time, the collected data group can be further cleaned, supplemented and matched based on the characteristics of automatic extension and filling of the time ruler and the process of data synchronization processing until the data volume required for detection is obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 What is shown is a schematic diagram of the structure of an online detection device;

[0042] Figure 2 What is shown is a schematic diagram of the process of performing time state synchronization in step S1;

[0043] Figure 3 The figure shows a flow chart of synchronizing several groups of data collected by the host and the slave to the same terminal in step S3. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments.

[0045] Figure 1 The figure shows a flow chart of a method for processing online detection data of a distribution transformer, which method comprises the following steps:

[0046] S1. Start the master and slave and synchronize their time status.

[0047] S2. Collect data from the input and output terminals of several groups of distribution transformers through the host and slave devices respectively.

[0048] S3. Synchronize several groups of data collected by the host and the slave to the same terminal, and perform calculations to obtain the final result.

[0049] Figure 2 The figure shows a flow chart of performing time state synchronization in step S1. In step S1, performing time state synchronization mainly includes:

[0050] S1-1. After power-on, the host and the slave establish communication through matching and synchronize time. The synchronized time in this process can be natural time such as 17:13:49 on June 12, 2024, indicating that the host synchronizes its own natural time 17:13:49 on June 12, 2024 to the slave, or it can be a self-determined time such as ST00:02:26 calculated from the moment the host is powered on, indicating that the host completes communication with the slave at 2 minutes and 26 seconds after power-on and synchronizes the time to the slave. From then on, the host and the slave have the same time scale.

[0051] Since the functions of the master and slave are to detect the performance parameters of the distribution transformer, the core module may not have the function of network access, so it is impossible to automatically update the time through the NTP protocol. Based on this, in this application, it is preferred to use the self-determined time as the time scale for time synchronization.

[0052] S1-2. The host generates a time scale sequence, which includes at least an initial value P, a filling step H and a scanning step ΔT. The time scale sequence is a sequence that can be automatically extended and filled by the scanning step ΔT and the filling step H.

[0053] Here, there is no difference in the execution order of steps S1-1 and S1-2, and they can be performed simultaneously, that is, after the host is powered on, the time scale sequence is generated at the same time, and communication with the slave is established through matching; or, after the host is powered on, the time scale sequence is generated first, and communication with the slave is established through matching; or, after the host is powered on, communication with the slave is established through matching, and then the time scale sequence is generated.

[0054] S1-3. The host transmits the generated time scale sequence to the slave, which stores and calls it.

[0055] In this application, the host and the slave are not absolute. The host usually refers to a device that, in addition to collecting basic data, can also process, calculate and fit the basic data to obtain the final performance parameters of the transformer, while the slave relatively lacks the function of processing, calculating and fitting the basic data to obtain the final performance parameters of the transformer; when both devices have the function of processing, calculating and fitting the basic data to obtain the final performance parameters of the transformer, the host and the slave are no longer clearly distinguished. Instead, the host is determined by which device is responsible for processing, calculating and fitting the basic data to obtain the final performance parameters of the transformer during the detection process, and the other device is the slave accordingly.

[0056] The main purpose of the above steps is to establish synchronization between the master and the slave, and to further synchronize the time scale sequence between the master and the slave as a verification basis in the subsequent data synchronization process.

[0057] Normally, the initial value P can be a natural number 1, and the filling step H can also be a natural number 1. The scanning step ΔT is determined based on the detection method's requirement for temporal consistency of the data collected by the host and the slave, and can be set according to the actual detection environment. Usually, the value of ΔT is small, and an exemplary value of 0.5S can be taken.

[0058] Taking the above exemplary values ​​as an example, after the host generates a time scale sequence, it will automatically extend and fill its own time scale sequence from the current time. For example, if the host starts to generate sequence 1 at time ST00:02:30, then sequence 2 will be generated at ST00:02:30.5, and so on. When the host transmits the generated time scale sequence to the slave, it will also send its own sequence status to the slave, so that the slave can update its own time scale sequence based on the sequence status. For example, if the slave receives the sequence status time of the host as ST00:02:40, and the host's sequence has been extended to 21, the slave will directly extend and fill from 21 to generate its own time scale sequence.

[0059] In step S2, the host and the slave use mutual inductors to collect current and voltage data at the input and output ends of the distribution transformer, respectively. In order to obtain the performance parameters of the distribution transformer through fitting, the host and the slave need to collect multiple sets of data to implement the data fitting process. In this process, each time the host collects and obtains a piece of data, it will call the time scale sequence and add the current sequence value to the data. Similarly, each time the slave collects and obtains a piece of data, it will also call its time scale sequence and add the current sequence value to the data, thereby forming the data set of the host:

[0060] A1={(K 11 ,P 11 ),(K 12 ,P 12 ),(K 13 ,P 13 )......(K 1n ,P 1n )}, where K represents the collected data, P represents the sequence value corresponding to the data, the subscript value "1" corresponds to the host, and n represents the amount of data in the data set;

[0061] At the same time, a data set of the slave can also be formed:

[0062] A2={(K 21 ,P 21 ),(K 22 ,P 22 ),(K 23 ,P 23 )......(K 2n,P 2n )}, where K represents the collected data, P represents the sequence value corresponding to the data, the subscript constant "2" corresponds to the slave, and n represents the amount of data in the data set.

[0063] After completing data collection and labeling in step S2, Figure 3 The flowchart of step S3 is shown as follows: synchronizing several groups of data collected by the host and the slave to the same terminal. The method of step S3 is as follows:

[0064] S3-1, extract the sequence values ​​in the data sets A1 and A2 respectively and arrange them in ascending or descending order to form two prior sets, respectively F1 = {P 11 ,P 12 ,P 13 ......P 1n} and F2={P 21 ,P 22 ,P 23 ......P 2n};

[0065] S3-2, based on all elements in the two sets, calculate the prior values ​​M1 and M2;

[0066] M1=H(g(F1)), where g(F1)=P 11 P 12 ...P 1n +P 11 P 12 ...P 1(n-1) +...+P 11 P 12 +P 11 ;

[0067] M2=H(g(F2)), where g(F2)=P 21 P 22 ...P 2n +P 21 P 22 ...P 2(n-1) +...+P 21 P 22 +P 21 ;

[0068] In the above formula, M is the hash value obtained by calculating the hash function, such as the commonly used MD5 and SHA-1.

[0069] S3-3. Take any element P in M1 and the prior set F1 1i Pack and convert any element P in M2 and the prior set F2 2iPack and then synchronize to the same terminal for comparison to determine whether M1 and M2 are equal, and P 1i and P 2i whether they are equal;

[0070] If they are equal, the data collected in data sets A1 and A2 are packaged and synchronized to the same terminal, K of the collected data is extracted, and the data with equal sequence values ​​are grouped for fitting calculation;

[0071] If P 1i and P 2i If M1 and M2 are not equal, then mark P 1i and P 2i The confidence level is 1, and step S3-4 is executed;

[0072] If P 1i and P 2i Not equal, M1 and M2 are not equal, then mark P 1i and P 2i The confidence level is 0, and step S3-5 is executed;

[0073] S3-4, P 1i As an element to construct the confidence set Z1, any element P of the remaining elements in the set F1 1j Add to set Z1, and add P 2i As an element to construct the confidence set Z2, any element P of the remaining elements in the set F2 2j Add to set Z2, calculate the prior values ​​Y1 = H(g(Z1)) and Y2 = H(g(Z2)), synchronize Y1 and Y2 to the same terminal for comparison, and determine whether Y1 and Y2 are equal;

[0074] If yes, mark P 1j and P 2j The confidence level is 1, and steps S3-4 are repeated;

[0075] If otherwise, mark P 1j and P 2j The confidence level is 0, and step S3-5 is executed;

[0076] S3-5, discard the elements with a confidence level of 0, recalculate the prior values ​​M1 and M2, and replace any element P of the remaining elements in the set F1 with 1j and M1, and any element P of the remaining elements in the set F2 2j and M2 are packaged, and then synchronized to the same terminal for comparison to determine whether M1 and M2 are equal, and P 1j and P 2j whether they are equal;

[0077] If they are equal, the data collected in the data sets A1 and A2 corresponding to the updated F1 and F2 are packaged and synchronized to the same terminal, K of the collected data is extracted, and the data with equal sequence values ​​are grouped for fitting calculation;

[0078] If P 1j and P 2j If M1 and M2 are not equal, then mark P 1j and P 2j The confidence level is 1, and step S3-4 is executed;

[0079] If P 1j and P 2j Not equal, M1 and M2 are not equal, then mark P 1j and P 2j The confidence level is 0, and steps S3-5 are repeated;

[0080] In the above steps, until the number of elements in the constructed confidence sets Z1 and Z2 meets the data volume requirements of the fitting calculation, then the data sets A1 and A2 are updated based on the sets Z1 and Z2, the data collected in the data sets A1 and A2 are packaged and synchronized to the same terminal, K of the collected data is extracted, and the data with equal sequence values ​​are grouped for fitting calculation; usually the fitting calculation of online monitoring of distribution transformers requires 20 to 50 data groups, so the number of elements of Z1 and Z2 can be set to a saturation value.

[0081] On the other hand, when the confidence of elements in sets F1 and F2 is set to 1, they will be added to the confidence sets Z1 and Z2, and will also be discarded from the original sets F1 and F2. Correspondingly, when the confidence of elements in sets F1 and F2 is set to 0, they will also be discarded. Therefore, whenever an element is missing from the original sets F1 and F2, new data needs to be collected and supplemented by the host and slave to update the data sets A1 and A2.

[0082] On the other hand, since S3-3 will also judge the sets F1 and F2, and the original sets F1 and F2 also have P 1i and P 2i When M1 and M2 are equal, the corresponding data collected in sets A1 and A2 in this state also meet the requirements of the fitting calculation. Therefore, sets F1 and F2 are preferably set to the same saturation values ​​as Z1 and Z2.

[0083] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them.

Claims

1. A method for processing online detection data of a distribution transformer, characterized in that: The steps include: S1, after starting up the master and slave, synchronize the time status so that the master and slave have the same time scale; S2, respectively collecting data of the input and output ends of several groups of distribution transformers through the host and the slave, and associating the collected data with the time scale; S3, synchronizing several groups of data collected by the host and the slave to the same terminal, cleaning, supplementing and matching the data based on the correlation between the data and the time scale, and performing calculations to obtain the final result; In step S1, time state synchronization mainly includes: S1-1, after power-on, the host and slave establish communication through matching and synchronize time; S1-2, the host generates a time scale sequence, the time scale sequence includes at least an initial value P, a filling step H and a scanning step ΔT, and the time scale sequence is automatically extended and filled by the scanning step ΔT and the filling step H; S1-3, the host transmits the generated time scale sequence to the slave, which stores and calls it; In step S2, the master and slave call their time scale sequences each time they collect a data, and add the current sequence value to the data to form a set A1 = {(K 11 ,P 11 ),(K 12 ,P 12 ),(K 13 ,P 13 )......(K 1n ,P 1n )} and A2={(K 21 ,P 21 ),(K 22 ,P 22 ),(K 23 ,P 23 )......(K 2n ,P 2n )}, where K represents the collected data, P represents the sequence value corresponding to the data, the subscript value "1" corresponds to the master, the subscript value "2" corresponds to the slave, and n represents the amount of data in the data set; Step S3 includes the following steps: S3-1, extract the sequence values ​​in the data sets A1 and A2 respectively to form two priori sets, respectively F1 = {P 11 ,P 12 ,P 13 ......P 1n } and F2={P 21 ,P 22 ,P 23 ......P 2n }; S3-2, based on all elements in the two sets, calculate the prior values ​​M1 and M2; S3-3. Take any element P in M1 and the prior set F1 1i Pack and convert any element P in M2 and the prior set F2 2i Pack and then synchronize to the same terminal for comparison to determine whether M1 and M2 are equal, and P 1i and P 2i whether they are equal; If they are equal, the data collected in data sets A1 and A2 are packaged and synchronized to the same terminal, K of the collected data is extracted, and the data with equal sequence values ​​are grouped for fitting calculation; If P 1i and P 2i If M1 and M2 are not equal, then mark P 1i and P 2i The confidence level is 1, and step S3-4 is executed; If P 1i and P 2i Not equal, M1 and M2 are not equal, then mark P 1i and P 2i The confidence level is 0, and step S3-5 is executed; S3-4, P 1i As an element to construct the confidence set Z1, any element P of the remaining elements in the set F1 1j Add to set Z1, and add P 2i As an element to construct the confidence set Z2, any element P of the remaining elements in the set F2 2j Add to set Z2, calculate the prior values ​​Y1 and Y2, synchronize Y1 and Y2 to the same terminal for comparison, and determine whether Y1 and Y2 are equal; If yes, mark P 1j and P 2j The confidence level is 1, and steps S3-4 are repeated; If otherwise, mark P 1j and P 2j The confidence level is 0, and step S3-5 is executed; S3-5, discard the elements with a confidence level of 0, recalculate the prior values ​​M1 and M2, and replace any element P of the remaining elements in the set F1 with 1j and M1, and any element P of the remaining elements in the set F2 2j and M2 are packaged, and then synchronized to the same terminal for comparison to determine whether M1 and M2 are equal, and P 1j and P 2j whether they are equal; If they are equal, the data collected in the data sets A1 and A2 corresponding to the updated F1 and F2 are packaged and synchronized to the same terminal, K of the collected data is extracted, and the data with equal sequence values ​​are grouped for fitting calculation; If P 1j and P 2j If M1 and M2 are not equal, then mark P 1j and P 2j The confidence level is 1, and step S3-4 is executed; If P 1j and P 2j Not equal, M1 and M2 are not equal, then mark P 1j and P 2j The confidence level is 0, and steps S3-5 are repeated.

2. The method for processing online detection data of distribution transformer according to claim 1, characterized in that: The master uses the self-determined time as the time scale to synchronize the time to the slave.

3. The method for processing online detection data of distribution transformer according to claim 1, characterized in that: Step S1-1 and step S1-2 are executed in a swapped order or simultaneously.

4. The method for processing online detection data of distribution transformer according to claim 1, characterized in that: In step S3-1, the sequence values ​​in the data sets A1 and A2 are extracted and arranged in ascending or descending order to form two prior sets.

5. The method for processing online detection data of distribution transformer according to claim 1, characterized in that: In step S3-2, M1=H(g(F1)), where g(F1)=P 11 P 12 ...P 1n +P 11 P 12 ...P 1(n-1) +...+P 11 P 12 +P 11 ; M2=H(g(F2)), where g(F2)=P 21 P 22 ...P 2n +P 21 P 22 ...P 2(n-1) +...+P 21 P 22 +P 21 ; In the above formula, M is the hash value obtained by calculating the hash function.

6. The method for processing online detection data of distribution transformer according to claim 1, characterized in that: In step S3-4, the number of elements in confidence sets Z1 and Z2 is a preset value. When the number of elements in confidence sets Z1 and Z2 reaches the preset value, step S3-4 is terminated, and sets A1 and A2 are updated based on sets Z1 and Z2. The data collected in sets A1 and A2 are packaged and synchronized to the same terminal, K of the collected data is extracted, and data with equal sequence values ​​are grouped for fitting calculations.

7. The method for processing online detection data of distribution transformer according to claim 1, characterized in that: After the confidence of the elements in sets F1 and F2 is set to 1 or 0 in step S3, a discard operation is performed, and new data is collected and supplemented by the master and slave to update sets A1 and A2; the number of elements in sets F1 and F2 is the same as that in sets Z1 and Z2.

Images