A smart data acquisition system and method for electricity consumption information of a dedicated transformer data acquisition terminal

Abstract

This invention discloses an intelligent data acquisition system and method for electricity consumption information from dedicated transformer acquisition terminals, relating to the field of power data acquisition technology. The method includes the following steps: collecting historical data on the operating status of dedicated transformer acquisition terminals and constructing a set; dividing the data by time period and constructing feature vectors; generating training samples using the reporting success rate as a label; training a random forest regression model to output the communication quality coefficient for each time period; marking abnormal time periods with coefficients below a threshold; allocating candidate upload windows based on normal time periods; removing windows that intersect with abnormal time periods to obtain usable windows; calculating the deviation between the usable window and the baseline upload time; selecting the window with the smallest deviation as the optimized upload time for a single terminal and calculating the scheduling adaptability; selecting the globally optimized upload time based on the adaptability; generating an adaptive data acquisition schedule; and adjusting the reporting time of electricity consumption information for each terminal. This invention achieves intelligent data acquisition of electricity consumption information.

Description

Technical Field

[0001] This invention relates to the field of power data acquisition technology, specifically to an intelligent acquisition system and method for electricity consumption information of a dedicated transformer acquisition terminal. Background Technology

[0002] With the continuous increase in the demand for power Internet of Things construction and electricity management, the dedicated transformer data acquisition terminal, as the core equipment for electricity information collection, directly affects the accuracy and effectiveness of power grid electricity monitoring, energy consumption analysis and dispatch management due to the stability, real-time performance and reporting success rate of its data acquisition.

[0003] However, traditional methods for collecting electricity consumption information from dedicated transformer data acquisition terminals often suffer from the following problems when dealing with concurrent reporting from multiple terminals, dynamic channel changes, and data acquisition quality control: First, the acquisition efficiency is low. Massive numbers of terminals reporting at fixed times can easily cause communication channel congestion, resulting in signal interference, large communication delays, and numerous retransmissions, leading to low data reporting success rates and untimely acquisition. Second, there is a lack of communication quality prediction. Without combining historical operating data to assess the communication status of different time periods, blindly selecting upload time periods can easily trigger communication anomalies, resulting in insufficient data acquisition reliability. In addition, there is a lack of intelligent time scheduling capabilities. The upload time is not optimized for differences in terminal communication, resulting in poor global acquisition adaptability. Existing acquisition systems mostly adopt fixed reporting mechanisms and lack communication quality assessment and adaptive time scheduling functions, which cannot meet the needs of intelligent acquisition. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent data acquisition system and method for electricity consumption information of a dedicated transformer data acquisition terminal, so as to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent method for collecting electricity consumption information from a dedicated transformer data acquisition terminal, the method comprising the following steps: Collect historical data on the operating status of the dedicated transformer acquisition terminal and construct a historical data set; Divide the data into time periods and construct feature vectors for each time period; generate a training sample set using the actual reporting success rate of the dedicated transformer data acquisition terminal as a label, complete the training through a random forest regression model, and output the communication quality coefficients for the corresponding time periods; The time period when the communication quality coefficient is lower than the preset threshold is marked as the communication quality abnormal period; candidate upload time windows are allocated to the terminal based on the normal communication period, and candidate windows that intersect with the abnormal period are removed to obtain the available upload time window; the deviation value between each available window and the benchmark upload time is calculated, and the window with the smallest deviation is selected as the optimized upload time of the single dedicated transformer acquisition terminal, and the corresponding scheduling adaptability is calculated. Based on the scheduling adaptability, a globally optimized upload time is selected, and an adaptive data acquisition schedule is generated to adjust the power consumption information reporting time of each dedicated transformer acquisition terminal.

[0006] Historical operational status data of the dedicated transformer data acquisition terminals are collected to construct a historical dataset, which is then divided into time periods, and feature vectors are constructed for each time period. A training sample set is generated using the actual reporting success rate of the dedicated transformer data acquisition terminals as labels. This set is trained using a random forest regression model, and the communication quality coefficients for the corresponding time periods are output. Specific steps include: Acquire the historical operating status data of each dedicated transformer acquisition terminal and generate a historical data set T, T = {T1, T2, ..., Ti, ..., TI}, where Ti represents the historical operating status data of the i-th dedicated transformer acquisition terminal and I represents the total number of dedicated transformer acquisition terminals; For the i-th dedicated transformer data acquisition terminal, its historical operating status data Ti = {D} i1 D i2 , ..., D ij , ..., D iJ}; where D ij This represents the original operating status record of the i-th dedicated transformer data acquisition terminal in the j-th time period, where j represents the time period number, j∈{1,2,...,J}, and J represents the total number of time periods divided per unit time. The original operating status record D ij This includes the online rate, original signal strength value, average communication latency, and number of retransmissions during the time period. A feature vector is constructed for each time period. The feature vector includes the online rate, normalized signal strength, average historical communication delay, cumulative retransmission count, and time period number for the corresponding time period. Using the actual reported success rate within the same time period as the label, a training sample pair of the feature vector and the actual reported success rate is constructed to form a training sample set; A random forest regression model is used to train the training sample set to learn the mapping relationship between feature vectors and reporting success rate; For the a-th dedicated transformer data acquisition terminal to be evaluated in the current j-th time period, construct its corresponding feature vector, input it into the trained random forest regression model, and the predicted value output by the model is the communication quality coefficient for that time period.

[0007] Periods with communication quality coefficients below a preset threshold are marked as periods of abnormal communication quality. Specific steps include: The time periods when the communication quality coefficient is less than the preset communication quality threshold are marked, forming a set of time periods with abnormal communication quality for each dedicated transformer data acquisition terminal, denoted as Q. i Q i ={Q i1 Q i2 Q ik QiK}, Q ik Q represents the communication quality anomaly of the i-th dedicated transformer data acquisition terminal during the k-th time period, where K represents the total number of time periods during which the dedicated transformer data acquisition terminal detected communication quality anomalies. i The set of time segments representing the communication quality anomalies of the i-th dedicated transformer data acquisition terminal is defined by statistically analyzing the distribution of communication quality coefficients of all dedicated transformer data acquisition terminals in each time period over a period of time, and setting a specific quantile as the communication quality threshold for the current period.

[0008] Candidate upload time windows are allocated to terminals based on normal communication periods. Candidate windows that intersect with abnormal periods are then removed to obtain available upload time windows. The specific steps include: Based on the periods of abnormal communication quality for each dedicated transformer data acquisition terminal, a candidate upload time window is allocated to each dedicated transformer data acquisition terminal. One dedicated transformer data acquisition terminal is arbitrarily selected and denoted as the *a*-th dedicated transformer data acquisition terminal, and its allocated candidate upload time window is denoted as W. a ={W a1 W a2 , ..., W am , ..., W aM}, where W am This represents the m-th candidate upload time window of the a-th dedicated transformer data acquisition terminal, where M represents the total number of candidate upload time windows. The candidate upload time window is selected from a time period with normal communication quality. When the a-th dedicated transformer data acquisition terminal is assigned a candidate upload time window W am The set of time segments Q with abnormal communication quality a If there is an overlap, then mark the candidate upload time window as unavailable and start from W. a Remove from the middle; iterate through the candidate upload time windows of the a-th dedicated transformer data acquisition terminal and filter out all those with Q. a The available upload time windows that do not overlap are denoted as W'. a ={W' a1 W' a2 , ..., W' an , ..., W' aN}, where W' an This represents the nth available upload time window after the a-th dedicated transformer data acquisition terminal has been filtered, where N represents the total number of available upload time windows.

[0009] Calculate the deviation between each available window and the baseline upload time, and select the window with the smallest deviation as the optimized upload time for a single dedicated transformer data acquisition terminal. Specific steps include: Calculate the deviation between the a-th dedicated transformer data acquisition terminal and the baseline upload time in the nth available upload time window, defined as follows: ΔT an =|Wan '-T a0 |;Among them, T a0 ΔT represents the reference upload time of the a-th dedicated transformer data acquisition terminal. an This represents the deviation between the a-th dedicated transformer data acquisition terminal and the baseline upload time in the nth available upload time window; the available upload time window with the smallest deviation is selected as the optimized upload time for the a-th dedicated transformer data acquisition terminal.

[0010] Calculate the scheduling adaptability of the target transformer data acquisition terminal, select the globally optimized upload time based on the scheduling adaptability, and generate an adaptive data acquisition schedule to adjust the power consumption information reporting time of each transformer data acquisition terminal. Specific steps include: According to the formula: R a =1-[|T a1 -T a0 | / T a0 ], calculate the scheduling adaptability of the a-th dedicated transformer data acquisition terminal; where R a T represents the scheduling adaptability of the a-th dedicated transformer data acquisition terminal. a1 |T represents the optimized upload time of the a-th dedicated transformer data acquisition terminal. a1 -T a0 | represents the absolute value of the time difference between the optimized upload time and the baseline upload time; By iterating through all dedicated transformer data acquisition terminals, the scheduling adaptability of each dedicated transformer data acquisition terminal is obtained as R. i , i = 1, 2, ..., I, where I represents the total number of dedicated transformer data acquisition terminals; Based on the scheduling adaptability R of each dedicated transformer data acquisition terminal i The candidate upload time window with the highest scheduling adaptability is selected as the final optimized upload time, and an adaptive data acquisition schedule {T'1, T'2, ..., T'i, ..., T'I} is generated, where T'i represents the optimized upload time of the i-th dedicated transformer data acquisition terminal; The dedicated transformer data acquisition terminal adjusts the reporting time of the collected electricity consumption information according to the adaptive data acquisition schedule.

[0011] A smart data acquisition system for electricity consumption information of a dedicated transformer data acquisition terminal, the system comprising: Data acquisition module, model training module, quality assessment module, window filtering module, time optimization module, scheduling and execution module; The output of the data acquisition module is connected to the input of the model training module; the output of the model training module is connected to the input of the quality assessment module; the output of the quality assessment module is connected to the input of the window filtering module; the output of the window filtering module is connected to the input of the time optimization module; and the output of the time optimization module is connected to the input of the scheduling execution module.

[0012] The data acquisition module includes a historical acquisition unit and a feature construction unit; the historical acquisition unit is used to collect historical data of terminal operation and construct a dataset; the feature construction unit is used to generate feature vectors according to time periods. The model training module includes a sample generation unit and a model training unit; the sample generation unit is used to generate training samples using the reported success rate as the label; the model training unit is used to train the random forest regression model. The quality assessment module includes a coefficient calculation unit and an anomaly marking unit; the coefficient calculation unit is used to output the communication quality coefficient for a time period; the anomaly marking unit is used to mark the time period with abnormal communication quality.

[0013] The window filtering module includes a candidate allocation unit and an available filtering unit; the candidate allocation unit is used to allocate candidate upload time windows; the available filtering unit is used to remove windows that intersect with abnormal time periods; The time optimization module includes a single-terminal optimization unit and an adaptation calculation unit; the single-terminal optimization unit is used to determine the optimal upload time for a single terminal; the adaptation calculation unit is used to calculate the terminal scheduling adaptation. The scheduling execution module includes a global optimization unit and a table generation unit; the global optimization unit selects the globally optimal upload time; the table generation unit is used to generate an adaptive data collection schedule.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. Based on the normal communication period, the available upload time window is selected, and the optimal upload time for a single terminal is determined according to the principle of minimum deviation. Unlike the existing technology of fixed centralized reporting, which is prone to channel congestion, this invention can match the optimal reporting time for a single terminal. 2. Based on the scheduling adaptability, a globally optimized upload time is selected, and an adaptive collection schedule is generated to dynamically adjust the reporting time of each terminal. Unlike the existing technology of static reporting and lack of global scheduling capability, this invention can realize the intelligent and balanced allocation of reporting time for multiple terminals. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating an intelligent method for collecting electricity consumption information from a dedicated transformer data acquisition terminal according to the present invention. Detailed Implementation

[0016] 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 embodiments of the present invention, and not all embodiments. 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.

[0017] like Figure 1 As shown, the present invention provides a technical solution: an intelligent method for collecting electricity consumption information from a dedicated transformer data acquisition terminal. This method includes the following steps: Collect historical data on the operating status of the dedicated transformer acquisition terminal and construct a historical data set; Divide the data into time periods and construct feature vectors for each time period; generate a training sample set using the actual reporting success rate of the dedicated transformer data acquisition terminal as a label, complete the training through a random forest regression model, and output the communication quality coefficients for the corresponding time periods; The time period when the communication quality coefficient is lower than the preset threshold is marked as the communication quality abnormal period; candidate upload time windows are allocated to the terminal based on the normal communication period, and candidate windows that intersect with the abnormal period are removed to obtain the available upload time window; the deviation value between each available window and the benchmark upload time is calculated, and the window with the smallest deviation is selected as the optimized upload time of the single dedicated transformer acquisition terminal, and the corresponding scheduling adaptability is calculated. Based on the scheduling adaptability, a globally optimized upload time is selected, and an adaptive data acquisition schedule is generated to adjust the power consumption information reporting time of each dedicated transformer acquisition terminal.

[0018] Historical operational status data of the dedicated transformer data acquisition terminals are collected to construct a historical dataset, which is then divided into time periods, and feature vectors are constructed for each time period. A training sample set is generated using the actual reporting success rate of the dedicated transformer data acquisition terminals as labels. This set is trained using a random forest regression model, and the communication quality coefficients for the corresponding time periods are output. Specific steps include: Acquire the historical operating status data of each dedicated transformer acquisition terminal and generate a historical data set T, T = {T1, T2, ..., Ti, ..., TI}, where Ti represents the historical operating status data of the i-th dedicated transformer acquisition terminal and I represents the total number of dedicated transformer acquisition terminals; For the i-th dedicated transformer data acquisition terminal, its historical operating status data Ti = {D} i1 D i2 , ..., D ij , ..., D iJ}; where D ij This represents the original operating status record of the i-th dedicated transformer data acquisition terminal in the j-th time period, where j represents the time period number, j∈{1,2,...,J}, and J represents the total number of time periods divided per unit time. The original operating status record D ij This includes the online rate, original signal strength value, average communication latency, and number of retransmissions during the time period. A feature vector is constructed for each time period. The feature vector includes the online rate, normalized signal strength, average historical communication delay, cumulative retransmission count, and time period number for the corresponding time period. Using the actual reported success rate within the same time period as the label, a training sample pair of the feature vector and the actual reported success rate is constructed to form a training sample set; A random forest regression model is used to train the training sample set to learn the mapping relationship between feature vectors and reporting success rate; For the a-th dedicated transformer data acquisition terminal to be evaluated in the current j-th time period, construct its corresponding feature vector, input it into the trained random forest regression model, and the predicted value output by the model is the communication quality coefficient for that time period.

[0019] Periods with communication quality coefficients below a preset threshold are marked as periods of abnormal communication quality. Specific steps include: The time periods when the communication quality coefficient is less than the preset communication quality threshold are marked, forming a set of time periods with abnormal communication quality for each dedicated transformer data acquisition terminal, denoted as Q. i Q i ={Q i1 Q i2 Q ik Q iK}, Q ik Q represents the communication quality anomaly of the i-th dedicated transformer data acquisition terminal during the k-th time period, where K represents the total number of time periods during which the dedicated transformer data acquisition terminal detected communication quality anomalies. i The set of time segments representing the communication quality anomalies of the i-th dedicated transformer data acquisition terminal is defined by statistically analyzing the distribution of communication quality coefficients of all dedicated transformer data acquisition terminals in each time period over a period of time, and setting a specific quantile as the communication quality threshold for the current period.

[0020] Candidate upload time windows are allocated to terminals based on normal communication periods. Candidate windows that intersect with abnormal periods are then removed to obtain available upload time windows. The specific steps include: Based on the periods of abnormal communication quality for each dedicated transformer data acquisition terminal, a candidate upload time window is allocated to each dedicated transformer data acquisition terminal. One dedicated transformer data acquisition terminal is arbitrarily selected and denoted as the *a*-th dedicated transformer data acquisition terminal, and its allocated candidate upload time window is denoted as W. a ={W a1 W a2 , ..., W am , ..., W aM}, where Wam This represents the m-th candidate upload time window of the a-th dedicated transformer data acquisition terminal, where M represents the total number of candidate upload time windows. The candidate upload time window is selected from a time period with normal communication quality. When the a-th dedicated transformer data acquisition terminal is assigned a candidate upload time window W am The set of time segments Q with abnormal communication quality a If there is an overlap, then mark the candidate upload time window as unavailable and start from W. a Remove from the middle; iterate through the candidate upload time windows of the a-th dedicated transformer data acquisition terminal and filter out all those with Q. a The available upload time windows that do not overlap are denoted as W'. a ={W' a1 W' a2 , ..., W' an , ..., W' aN}, where W' an This represents the nth available upload time window after the a-th dedicated transformer data acquisition terminal has been filtered, where N represents the total number of available upload time windows.

[0021] Calculate the deviation between each available window and the baseline upload time, and select the window with the smallest deviation as the optimized upload time for a single dedicated transformer data acquisition terminal. Specific steps include: Calculate the deviation between the a-th dedicated transformer data acquisition terminal and the baseline upload time in the nth available upload time window, defined as follows: ΔT an =|W an '-T a0 |;Among them, T a0 ΔT represents the reference upload time of the a-th dedicated transformer data acquisition terminal. an This represents the deviation between the a-th dedicated transformer data acquisition terminal and the baseline upload time in the nth available upload time window; the available upload time window with the smallest deviation is selected as the optimized upload time for the a-th dedicated transformer data acquisition terminal.

[0022] Calculate the scheduling adaptability of the target transformer data acquisition terminal, select the globally optimized upload time based on the scheduling adaptability, and generate an adaptive data acquisition schedule to adjust the power consumption information reporting time of each transformer data acquisition terminal. Specific steps include: According to the formula: R a =1-[|T a1 -T a0 | / T a0 ], calculate the scheduling adaptability of the a-th dedicated transformer data acquisition terminal; where R a T represents the scheduling adaptability of the a-th dedicated transformer data acquisition terminal. a1 |T represents the optimized upload time of the a-th dedicated transformer data acquisition terminal. a1 -Ta0 | represents the absolute value of the time difference between the optimized upload time and the baseline upload time; By iterating through all dedicated transformer data acquisition terminals, the scheduling adaptability of each dedicated transformer data acquisition terminal is obtained as R. i , i = 1, 2, ..., I, where I represents the total number of dedicated transformer data acquisition terminals; Based on the scheduling adaptability R of each dedicated transformer data acquisition terminal i The candidate upload time window with the highest scheduling adaptability is selected as the final optimized upload time, and an adaptive data acquisition schedule {T'1, T'2, ..., T'i, ..., T'I} is generated, where T'i represents the optimized upload time of the i-th dedicated transformer data acquisition terminal; The dedicated transformer data acquisition terminal adjusts the reporting time of the collected electricity consumption information according to the adaptive data acquisition schedule.

[0023] In Example 1: Historical data on the operating status of the dedicated transformer acquisition terminal is collected, and the historical operating records of all terminals are integrated to construct a complete historical data set. The set contains operating information such as the online status, signal strength, communication latency, and number of data retransmissions of each terminal during different operating periods. The overall operation cycle is divided into continuous time periods according to the time dimension. For each divided time period, the corresponding operation parameters are extracted to construct a feature vector. The feature vector covers the online status, normalized signal strength, average communication latency, cumulative retransmission count and time period attribute information within the time period. The actual reporting success rate of each terminal in the corresponding time period is used as the label. The feature vector and the label are combined to form a training sample set. The training sample set is input into the random forest regression model to carry out model training, allowing the model to fully learn the mapping relationship between the feature vector and the reporting success rate. After the model training is completed, the feature vector of the time period to be evaluated is input, and the communication quality coefficient of the corresponding time period can be output. Based on the preset communication quality judgment criteria, the time periods when the communication quality coefficient does not meet the standard are marked as communication quality abnormal time periods. The abnormal communication quality time periods of each terminal are compiled. Based on the normal time periods with stable communication quality, candidate upload time windows are allocated to each special transformer data acquisition terminal. The overlap between the candidate windows and the abnormal time periods is checked one by one. Unusable windows that have intersection with the abnormal time periods are eliminated, and unconflicted usable upload time windows are selected. Determine the baseline upload time for each terminal, calculate the time deviation between each available upload time window and the baseline upload time, select the window with the smallest deviation as the optimized upload time for a single terminal, and calculate the scheduling adaptability corresponding to the optimized upload time. Iterate through the scheduling adaptability of all dedicated transformer data acquisition terminals, select the upload time with the best adaptability as the global optimized upload time, generate an adaptive data acquisition schedule based on the global optimization results, and synchronize the schedule to each dedicated transformer data acquisition terminal.

[0024] A smart data acquisition system for electricity consumption information of a dedicated transformer data acquisition terminal, the system comprising: Data acquisition module, model training module, quality assessment module, window filtering module, time optimization module, scheduling and execution module; The output of the data acquisition module is connected to the input of the model training module; the output of the model training module is connected to the input of the quality assessment module; the output of the quality assessment module is connected to the input of the window filtering module; the output of the window filtering module is connected to the input of the time optimization module; and the output of the time optimization module is connected to the input of the scheduling execution module.

[0025] The data acquisition module includes a historical acquisition unit and a feature construction unit; the historical acquisition unit is used to collect historical data of terminal operation and construct a dataset; the feature construction unit is used to generate feature vectors according to time periods. The model training module includes a sample generation unit and a model training unit; the sample generation unit is used to generate training samples using the reported success rate as the label; the model training unit is used to train the random forest regression model. The quality assessment module includes a coefficient calculation unit and an anomaly marking unit; the coefficient calculation unit is used to output the communication quality coefficient for a time period; the anomaly marking unit is used to mark the time period with abnormal communication quality.

[0026] The window filtering module includes a candidate allocation unit and an available filtering unit; the candidate allocation unit is used to allocate candidate upload time windows; the available filtering unit is used to remove windows that intersect with abnormal time periods; The time optimization module includes a single-terminal optimization unit and an adaptation calculation unit; the single-terminal optimization unit is used to determine the optimal upload time for a single terminal; the adaptation calculation unit is used to calculate the terminal scheduling adaptation. The scheduling execution module includes a global optimization unit and a table generation unit; the global optimization unit selects the globally optimal upload time; the table generation unit is used to generate an adaptive data collection schedule.

[0027] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for intelligently collecting electricity consumption information from a dedicated transformer data acquisition terminal, characterized in that: The method includes the following steps: Collect historical data on the operating status of the dedicated transformer acquisition terminal and construct a historical data set; Divide the data into time periods and construct feature vectors for each time period; generate a training sample set using the actual reporting success rate of the dedicated transformer data acquisition terminal as a label, complete the training through a random forest regression model, and output the communication quality coefficients for the corresponding time periods; The time period when the communication quality coefficient is lower than the preset threshold is marked as the communication quality abnormal period; candidate upload time windows are allocated to the terminal based on the normal communication period, and candidate windows that intersect with the abnormal period are removed to obtain the available upload time window; the deviation value between each available window and the benchmark upload time is calculated, and the window with the smallest deviation is selected as the optimized upload time of the single dedicated transformer acquisition terminal, and the corresponding scheduling adaptability is calculated. Based on the scheduling adaptability, a globally optimized upload time is selected, and an adaptive data acquisition schedule is generated to adjust the power consumption information reporting time of each dedicated transformer acquisition terminal.

2. The intelligent acquisition method for electricity consumption information of a dedicated transformer acquisition terminal according to claim 1, characterized in that: Collect historical data on the operating status of the dedicated transformer acquisition terminal, construct a historical data set, divide it into time periods according to the time dimension, and construct feature vectors for each time period; A training sample set is generated using the actual reporting success rate of the dedicated transformer data acquisition terminal as a label. The set is then trained using a random forest regression model, outputting the communication quality coefficients for the corresponding time periods. Specific steps include: Acquire the historical operating status data of each dedicated transformer acquisition terminal and generate a historical data set T, T = {T1, T2, ..., Ti, ..., TI}, where Ti represents the historical operating status data of the i-th dedicated transformer acquisition terminal and I represents the total number of dedicated transformer acquisition terminals; For the i-th dedicated transformer data acquisition terminal, its historical operating status data Ti = {D} i1 D i2 , ..., D ij , ..., D iJ }; where D ij This represents the original operating status record of the i-th dedicated transformer data acquisition terminal in the j-th time period, where j represents the time period number, j∈{1,2,...,J}, and J represents the total number of time periods divided per unit time. A feature vector is constructed for each time period. The feature vector includes the online rate, normalized signal strength, average historical communication delay, cumulative retransmission count, and time period number for the corresponding time period. Using the actual reported success rate within the same time period as the label, a training sample pair of the feature vector and the actual reported success rate is constructed to form a training sample set; A random forest regression model is used to train the training sample set to learn the mapping relationship between feature vectors and reporting success rate; For the a-th dedicated transformer data acquisition terminal to be evaluated in the current j-th time period, construct its corresponding feature vector, input it into the trained random forest regression model, and the predicted value output by the model is the communication quality coefficient for that time period.

3. The intelligent acquisition method for electricity consumption information of a dedicated transformer acquisition terminal according to claim 2, characterized in that: Periods with communication quality coefficients below a preset threshold are marked as periods of abnormal communication quality. Specific steps include: The time periods when the communication quality coefficient is less than the preset communication quality threshold are marked, forming a set of time periods with abnormal communication quality for each dedicated transformer data acquisition terminal, denoted as Q. i Q i ={Q i1 Q i2 Q ik Q iK }, Q ik Q represents the communication quality anomaly of the i-th dedicated transformer data acquisition terminal during the k-th time period, where K represents the total number of time periods during which the dedicated transformer data acquisition terminal detected communication quality anomalies. i The set of time segments representing the communication quality abnormality of the i-th dedicated transformer data acquisition terminal is defined by statistically analyzing the distribution of the communication quality coefficients of all dedicated transformer data acquisition terminals in each time period over a period of time, and setting a specific quantile as the communication quality threshold for the current period.

4. The intelligent acquisition method for electricity consumption information of a dedicated transformer acquisition terminal according to claim 4, characterized in that: Candidate upload time windows are allocated to terminals based on normal communication periods. Candidate windows that intersect with abnormal periods are then removed to obtain available upload time windows. The specific steps include: Based on the periods of abnormal communication quality for each dedicated transformer data acquisition terminal, a candidate upload time window is allocated to each terminal. One terminal is arbitrarily selected and denoted as the *a*-th terminal, and its allocated candidate upload time window is denoted as W. a ={W a1 W a2 , ..., W am , ..., W aM }, where W am This represents the m-th candidate upload time window of the a-th dedicated transformer data acquisition terminal, where M represents the total number of candidate upload time windows. The candidate upload time window is selected from a time period with normal communication quality. When the a-th dedicated transformer data acquisition terminal is assigned a candidate upload time window W am The set of time segments Q with abnormal communication quality a If there is an overlap, then mark the candidate upload time window as unavailable and start from W. a Remove from the middle; iterate through the candidate upload time windows of the a-th dedicated transformer data acquisition terminal and filter out all those with Q. a The available upload time windows that do not overlap are denoted as W'. a ={W' a1 W' a2 , ..., W' an , ..., W' aN }, where W' an This represents the nth available upload time window after the a-th dedicated transformer data acquisition terminal has been filtered, where N represents the total number of available upload time windows.

5. The intelligent acquisition method for electricity consumption information of a dedicated transformer acquisition terminal according to claim 4, characterized in that: Calculate the deviation between each available window and the baseline upload time, and select the window with the smallest deviation as the optimized upload time for a single dedicated transformer data acquisition terminal. Specific steps include: Calculate the deviation between the a-th dedicated transformer data acquisition terminal and the baseline upload time in the nth available upload time window, defined as follows: ΔT an =|W an '-T a0 |;Among them, T a0 ΔT represents the reference upload time of the a-th dedicated transformer data acquisition terminal. an This represents the deviation between the a-th dedicated transformer data acquisition terminal and the baseline upload time in the nth available upload time window; the available upload time window with the smallest deviation is selected as the optimized upload time for the a-th dedicated transformer data acquisition terminal.

6. The intelligent acquisition method for electricity consumption information of a dedicated transformer acquisition terminal according to claim 5, characterized in that: Calculate the scheduling adaptability of the target transformer data acquisition terminal, select the globally optimized upload time based on the scheduling adaptability, and generate an adaptive data acquisition schedule to adjust the power consumption information reporting time of each transformer data acquisition terminal. Specific steps include: According to the formula: R a =1-[|T a1 -T a0 | / T a0 ], calculate the scheduling adaptability of the a-th dedicated transformer data acquisition terminal; where R a T represents the scheduling adaptability of the a-th dedicated transformer data acquisition terminal. a1 |T represents the optimized upload time of the a-th dedicated transformer data acquisition terminal. a1 -T a0 | represents the absolute value of the time difference between the optimized upload time and the baseline upload time; By iterating through all dedicated transformer data acquisition terminals, the scheduling adaptability of each dedicated transformer data acquisition terminal is obtained as R. i , i = 1, 2, ..., I, where I represents the total number of dedicated transformer data acquisition terminals; Based on the scheduling adaptability R of each dedicated transformer data acquisition terminal i The candidate upload time window with the highest scheduling adaptability is selected as the final optimized upload time, and an adaptive data acquisition schedule {T'1, T'2, ..., T'i, ..., T'I} is generated, where T'i represents the optimized upload time of the i-th dedicated transformer data acquisition terminal; The dedicated transformer data acquisition terminal adjusts the reporting time of the collected electricity consumption information according to the adaptive data acquisition schedule.

7. An intelligent data acquisition system for electricity consumption information of a dedicated transformer data acquisition terminal, applied to the intelligent data acquisition method for electricity consumption information of a dedicated transformer data acquisition terminal as described in any one of claims 1-6, characterized in that: The system includes: Data acquisition module, model training module, quality assessment module, window filtering module, time optimization module, scheduling and execution module; The output of the data acquisition module is connected to the input of the model training module; the output of the model training module is connected to the input of the quality assessment module; the output of the quality assessment module is connected to the input of the window filtering module; the output of the window filtering module is connected to the input of the time optimization module; and the output of the time optimization module is connected to the input of the scheduling execution module.

8. The intelligent data acquisition system for electricity consumption information of a dedicated transformer data acquisition terminal according to claim 6, characterized in that: The data acquisition module includes a historical acquisition unit and a feature construction unit; the historical acquisition unit is used to collect historical data of terminal operation and construct a dataset; the feature construction unit is used to generate feature vectors according to time periods. The model training module includes a sample generation unit and a model training unit; the sample generation unit is used to generate training samples using the reported success rate as the label; the model training unit is used to train the random forest regression model. The quality assessment module includes a coefficient calculation unit and an anomaly marking unit; the coefficient calculation unit is used to output the communication quality coefficient for a time period; the anomaly marking unit is used to mark the time period with abnormal communication quality.

9. The intelligent data acquisition system for electricity consumption information of a dedicated transformer data acquisition terminal according to claim 7, characterized in that: The window filtering module includes a candidate allocation unit and an available filtering unit; the candidate allocation unit is used to allocate candidate upload time windows; the available filtering unit is used to remove windows that intersect with abnormal time periods; The time optimization module includes a single-end optimization unit and an adaptation calculation unit; the single-end optimization unit is used to determine the optimal upload time for a single terminal. The adaptation calculation unit is used to calculate the terminal scheduling adaptation degree. The scheduling execution module includes a global optimization unit and a table generation unit; the global optimization unit selects the globally optimal upload time. The table generation unit is used to generate an adaptive data acquisition schedule.

Images