Method, system and device for determining credibility based on multi-source non-synchronous measurement data

Abstract

The application provides a credibility determination method, system and device based on multi-source non-synchronous measurement data, comprising: obtaining measurement data and the state of the measurement data; selecting a corresponding pre-constructed credibility model based on the state of the measurement data; inputting the measurement data into the pre-constructed credibility model corresponding to the state of the measurement data to determine the credibility of the measurement data; wherein the credibility model is a typical daily load curve determined by correlation analysis of historical multi-source non-synchronous measurement data, and the credibility model is constructed based on the slope of the typical daily load curve and the credibility of real-time measurement data obtained by simulation in each period. The application obtains a typical daily load curve through correlation analysis, divides the period on this basis, can more finely reflect the size and change of the credibility, and improves the accuracy of the credibility of each period.

Description

Technical Field

[0001] This invention relates to the field of data processing and analysis, and specifically to a method, system, and device for determining the reliability of multi-source asynchronous measurement data. Background Technology

[0002] With the intelligent upgrading and construction of distribution networks, massive amounts of multi-source heterogeneous data have been acquired. This data originates from Supervisory Control and Data Acquisition (SCADA) systems, Advanced Measurement Infrastructure (AMI) systems, asset management systems, etc., exhibiting characteristics such as high capacity and diversity. Multi-source measurement data includes voltage, current, active power, reactive power, and other measurement data jointly provided by SCADA, AMI, and other systems. The introduction of multi-source measurement data has effectively solved the long-standing problem of incomplete or missing distribution network observations. The reliability of multi-source asynchronous measurements refers to their relative error with the corresponding real-time measurement, which is caused by factors such as measurement equipment error and measurement delay. To establish a reliability model for multi-source asynchronous measurement data, it is necessary to acquire real-time multi-source measurement data, as well as multi-source measurement data with various degrees of delay. Existing smart meters can only collect multi-source measurements with delays, and cannot acquire real-time measurements. Furthermore, the same meter can only acquire multi-source asynchronous measurement data with one degree of delay within a certain reading interval.

[0003] The reliability of multi-source asynchronous measurements is closely related to the degree of load change. During periods of slow load change, the error range of multi-source asynchronous measurements is smaller and the reliability is higher. Conversely, during periods of drastic load change, the error range of multi-source asynchronous measurements is larger and the reliability is lower.

[0004] In recent years, many scholars have studied the reliability error of measurement data, mainly including a method for calculating the reliability of distribution transformer load measurements, used to correct the variance of measurement equipment and improve the accuracy of state estimation in medium-voltage distribution networks. This method is based on a large amount of measured meter data, is simple to operate, and has significant effects, but there is still room for further research. First, this method is based on transformer load data with a 15-minute acquisition cycle, and linearly estimates the reliability of asynchronous measurements with delays in seconds using load change distribution parameters at adjacent acquisition times; the model is relatively coarse. Second, this method has not studied low-voltage distribution networks, which have numerous nodes, and the problem of asynchronous measurements is more pronounced due to limitations in communication channel capacity. Finally, although this method considers the uncertainty of asynchronous AMI measurements, it still describes the state estimation results with deterministic numerical values, without discussing the uncertainty of state estimation results based on the aforementioned uncertain measurements, and lacks exploration of the reliability of the results. Summary of the Invention

[0005] To address the problem that existing technologies are relatively crude in their study of measurement data reliability errors and cannot accurately determine the reliability of measurement data, this invention proposes a reliability determination method based on multi-source asynchronous measurement data, including:

[0006] Acquire measurement data and the status of the measurement data;

[0007] Select the corresponding pre-built credibility model based on the state of the measurement data;

[0008] The confidence level of the measurement data is determined by inputting the measurement data into the pre-built confidence model corresponding to the state of the measurement data.

[0009] The credibility model is constructed by dividing the typical daily load curve based on the slope of the typical daily load curve into time periods, and using the credibility determined by the historical multi-source asynchronous measurement data and the simulation data obtained in each time period.

[0010] Optionally, the credibility model includes an offline credibility model and an online credibility model; the construction of the credibility model includes:

[0011] Acquire historical multi-source asynchronous measurement data of the distribution transformer within a set time period, and fit the daily multi-source asynchronous measurement data within the set time period into a daily load curve;

[0012] The daily load curves are normalized, and the correlation between any two daily load curves is calculated using the Pearson correlation coefficient formula.

[0013] The correlation matrix is ​​constructed from the correlations among all daily load curves;

[0014] Calculate the average correlation coefficient between each normalized daily load curve and other normalized daily load curves, and select the maximum value as the typical daily load curve;

[0015] The typical daily load curve is divided into time periods based on its slope.

[0016] Simulate the reliability of multi-source asynchronous measurement data and real-time measurement in different offline and online scenarios;

[0017] Different credibility induction methods were used to summarize and organize the credibility in offline and online scenarios, resulting in offline credibility models and online credibility models.

[0018] Optionally, the reliability of the simulation calculation of multi-source asynchronous measurement data and real-time measurements in different offline and online scenarios includes:

[0019] In offline scenarios, real-time measurements and multi-source asynchronous measurements with varying degrees of latency are simulated under the collection cycle of conventional smart meters. The corresponding reliability in offline scenarios is obtained by combining the real-time measurements and multi-source asynchronous measurements with varying degrees of latency with a reliability calculation formula.

[0020] In online scenarios, the system simulates real-time measurements at the time of data acquisition, as well as multi-source asynchronous measurements with latency in the one or two reading intervals preceding the acquisition time. The system then combines multi-source asynchronous measurement data with varying degrees of latency with real-time measurements using a reliability calculation formula to obtain the corresponding reliability in the online scenario.

[0021] Optionally, different credibility induction methods are used to summarize and organize the credibility in offline and online scenarios respectively, to obtain offline credibility models and online credibility models, including:

[0022] For credibility in offline scenarios, the credibility of the collection times contained in each time period is grouped into the same set, and the set is used as the offline credibility model;

[0023] For credibility in online scenarios, credibility with the same latency is grouped into the same credibility set;

[0024] The credibility of data points with the same degree of delay at each collection time point within each time period is combined to obtain the credibility set for each time period.

[0025] The credibility set of each time period is used as the online credibility model.

[0026] Optionally, the credibility is calculated using the following formula:

[0027]

[0028] In the formula, Let the i-th meter be on day d, from t to t+2T. A Electrical energy within a time period, W i,d,t Let the i-th meter be on day d, from t to t+T. A Electrical energy within a time period For SCADA active power measurement P at the distribution transformer substation area S d,t Numerical value To superimpose active power, T A This refers to the acquisition cycle for multi-source asynchronous measurements.

[0029] Optionally, the step of selecting a pre-built credibility model based on the state of the measurement data includes:

[0030] When the measurement data is offline, a pre-built offline reliability model is selected;

[0031] When the measurement data is online, a pre-built online credibility model is selected.

[0032] Optionally, it also includes: performing a probability distribution test on the credibility set to obtain the distribution type of the credibility set.

[0033] Furthermore, this invention also provides a method for constructing a credibility model, comprising:

[0034] Acquire historical multi-source asynchronous measurement data of the distribution transformer within a set time period, and fit the daily multi-source asynchronous measurement data within the set time period into a daily load curve;

[0035] The daily load curves are normalized, and the correlation between any two daily load curves is calculated using the Pearson correlation coefficient formula.

[0036] The correlation matrix is ​​constructed from the correlations among all daily load curves;

[0037] Calculate the average correlation coefficient between each normalized daily load curve and other normalized daily load curves, and select the maximum value as the typical daily load curve;

[0038] The typical daily load curve is divided into time periods based on its slope.

[0039] Simulate the reliability of multi-source asynchronous measurement data and real-time measurement in different offline and online scenarios;

[0040] Different credibility induction methods were used to summarize and organize the credibility in offline and online scenarios, resulting in offline credibility models and online credibility models.

[0041] Optionally, the reliability of the simulation calculation of multi-source asynchronous measurement data and real-time measurements in different offline and online scenarios includes:

[0042] In offline scenarios, real-time measurements and multi-source asynchronous measurements with varying degrees of latency are simulated under the collection cycle of conventional smart meters. The corresponding reliability in offline scenarios is obtained by combining the real-time measurements and multi-source asynchronous measurements with varying degrees of latency with a reliability calculation formula.

[0043] In online scenarios, the system simulates real-time measurements at the time of data acquisition, as well as multi-source asynchronous measurements with latency in the one or two reading intervals preceding the acquisition time. The system then combines multi-source asynchronous measurement data with varying degrees of latency with real-time measurements using a reliability calculation formula to obtain the corresponding reliability in the online scenario.

[0044] Optionally, the credibility in offline and online scenarios are summarized and organized using different credibility induction methods to obtain offline credibility models and online credibility models, including:

[0045] For credibility in offline scenarios, the credibility of the collection times contained in each time period is grouped into the same set, and the set is used as the offline credibility model;

[0046] For credibility in online scenarios, credibility with the same latency is grouped into the same credibility set;

[0047] The credibility of data points with the same degree of delay at each collection time point within each time period is combined to obtain the credibility set for each time period.

[0048] The credibility set of each time period is used as the online credibility model.

[0049] Furthermore, the present invention also provides a reliability determination system based on multi-source asynchronous measurement data, comprising:

[0050] The acquisition module is used to acquire time-stamped measurement data and the status of the measurement data;

[0051] The selection module is used to select the corresponding pre-built credibility model based on the state of the measurement data;

[0052] A credibility calculation module is used to input the measurement data into the pre-built credibility model corresponding to the state of the measurement data to determine the credibility of the measurement data;

[0053] The credibility model is constructed by dividing the typical daily load curve based on the slope of the typical daily load curve into time periods, and using the credibility determined by the historical multi-source asynchronous measurement data and the simulation data obtained in each time period.

[0054] Optionally, it also includes a model building module for building a credibility model. Optionally, the model building module includes: a parameter processing submodule for acquiring historical multi-source asynchronous measurement data of the distribution transformer within a set time period, and fitting the daily multi-source asynchronous measurement data within the set time period into a daily load curve;

[0055] The correlation calculation submodule is used to normalize the daily load curves and calculate the correlation between any two daily load curves according to the Pearson correlation coefficient formula, and the correlation matrix is ​​formed by the correlation between all daily load curves.

[0056] The typical load curve determination submodule is used to calculate the average value of the correlation coefficient between each normalized daily load curve and other normalized daily load curves, and select the maximum value as the typical daily load curve.

[0057] The time period segmentation submodule is used to segment the typical daily load curve into time periods based on the slope of the typical daily load curve;

[0058] The simulation calculation submodule is used to simulate the reliability of multi-source asynchronous measurement data and real-time measurement in different offline and online scenarios.

[0059] The induction and organization submodule is used to summarize and organize the credibility in different offline and online scenarios using different credibility induction methods, and obtain offline credibility models and online credibility models respectively.

[0060] Optionally, the simulation calculation submodule is specifically used for:

[0061] In offline scenarios, real-time measurements and multi-source asynchronous measurements with varying degrees of latency are simulated under the collection cycle of conventional smart meters. The corresponding reliability in offline scenarios is obtained by combining the real-time measurements and multi-source asynchronous measurements with varying degrees of latency with a reliability calculation formula.

[0062] In online scenarios, the system simulates real-time measurements at the time of data acquisition, as well as multi-source asynchronous measurements with latency in the one or two reading intervals preceding the acquisition time. The system then combines multi-source asynchronous measurement data with varying degrees of latency with real-time measurements using a reliability calculation formula to obtain the corresponding reliability in the online scenario.

[0063] Optionally, the summarizing and organizing submodule is specifically used for:

[0064] For the credibility obtained in the offline scenario, the credibility of the collection time included in each time period is grouped into the same set, and the set is used as the offline credibility model;

[0065] For credibility obtained in online scenarios, credibility with the same latency is grouped into the same credibility set;

[0066] Based on the time period, the credibility of the data collection times within each time period with the same degree of delay is merged to obtain the credibility set of each time period;

[0067] The credibility set of each time period is used as the online credibility model.

[0068] Furthermore, the present invention also provides a model building module for a credibility model, comprising:

[0069] The parameter processing submodule is used to acquire historical multi-source asynchronous measurement data of the distribution transformer within a set time period, and to fit the daily multi-source asynchronous measurement data within the set time period into a load curve.

[0070] The correlation calculation submodule is used to normalize the load curves and calculate the correlation between any two load curves according to the Pearson correlation coefficient formula, and the correlation matrix is ​​formed by the correlation between all load curves.

[0071] The typical load curve determination submodule is used to calculate the average value of the correlation coefficient between each normalized daily load curve and other normalized daily load curves, and select the maximum value as the typical daily load curve.

[0072] The time period segmentation submodule is used to segment the typical daily load curve into time periods based on the slope of the typical daily load curve;

[0073] The simulation calculation submodule is used to simulate the reliability of multi-source asynchronous measurement data and real-time measurement in different offline and online scenarios.

[0074] The induction and organization submodule is used to summarize and organize the credibility in different offline and online scenarios using different credibility induction methods, and obtain offline credibility models and online credibility models respectively.

[0075] Optionally, the simulation calculation submodule specifically includes:

[0076] In offline scenarios, real-time measurements and multi-source asynchronous measurements with varying degrees of latency are simulated under the collection cycle of conventional smart meters. The corresponding reliability in offline scenarios is obtained by combining the real-time measurements and multi-source asynchronous measurements with varying degrees of latency with a reliability calculation formula.

[0077] In online scenarios, the system simulates real-time measurements at the time of data acquisition, as well as multi-source asynchronous measurements with latency in the one or two reading intervals preceding the acquisition time. The system then combines multi-source asynchronous measurement data with varying degrees of latency with real-time measurements using a reliability calculation formula to obtain the corresponding reliability in the online scenario.

[0078] In another aspect, the present invention also provides a computing device, comprising: one or more processors;

[0079] A processor is used to execute one or more programs;

[0080] When the one or more programs are executed by the one or more processors, the confidence determination method based on multi-source asynchronous measurement data or the confidence model construction method described above are implemented.

[0081] In another aspect, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed, it implements the reliability determination method based on multi-source asynchronous measurement data or the reliability model construction method as described above.

[0082] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0083] (1) This invention provides a method for determining the reliability of multi-source asynchronous measurement data, comprising: acquiring measurement data and the state of the measurement data; selecting a pre-constructed reliability model corresponding to the state of the measurement data; inputting the measurement data into the pre-constructed reliability model corresponding to the state of the measurement data to determine the reliability of the measurement data; wherein, the reliability model is a typical daily load curve determined by correlation analysis of historical multi-source asynchronous measurement data, the time period is divided based on the slope of the typical daily load curve, and the reliability is constructed by determining the reliability of historical multi-source asynchronous measurement and simulation-derived real-time measurement data in each time period. This invention obtains a typical daily load curve through correlation analysis, and divides the time period based on this curve, which can more precisely reflect the magnitude and changes of reliability, and improve the accuracy of the reliability of each time period.

[0084] (2) This invention effectively processes and analyzes offline and online measurement data to verify the accuracy of the multi-source asynchronous measurement reliability model. Attached Figure Description

[0085] Figure 1 This is a flowchart of the reliability determination method based on multi-source asynchronous measurement data of the present invention;

[0086] Figure 2 This is a schematic diagram of the credibility index set of the present invention;

[0087] Figure 3 This is a schematic diagram of the daily load curve of the pole-mounted transformer of the present invention;

[0088] Figure 4 This is a schematic diagram of the reliability curve of the present invention based on historical measurement data of real-time electricity meters;

[0089] Figure 5 This is a typical daily load curve according to an embodiment of the present invention;

[0090] Figure 6 (a) is the frequency distribution histogram and normal distribution probability density function of the absolute value of the credibility index in time period 1 of the present invention;

[0091] Figure 6(b) is the frequency distribution histogram of the absolute value of the credibility index and the normal distribution probability density function for time period 2 of the present invention, which conforms to the normal distribution model.

[0092] Figure 6 (c) is the frequency distribution histogram of the absolute value of the credibility index and the normal distribution probability density function for time period 3, which conforms to the normal distribution model of the present invention;

[0093] Figure 6 (d) is the frequency distribution histogram of the absolute value of the credibility index and the normal distribution probability density function for time period 4, which conforms to the normal distribution model of the present invention;

[0094] Figure 7 (a) is the frequency distribution histogram of the absolute value of the credibility index in time period 1 and the exponential distribution probability density function in the exponential distribution model of the present invention.

[0095] Figure 7 (b) is the frequency distribution histogram of the absolute value of the credibility index in time period 2 and the exponential distribution probability density function in the exponential distribution model of the present invention.

[0096] Figure 7 (c) is the frequency distribution histogram of the absolute value of the credibility index in time period 3 and the exponential distribution probability density function in the exponential distribution model of the present invention.

[0097] Figure 7 (d) is the frequency distribution histogram of the absolute value of the credibility index in time period 4 and the exponential distribution probability density function in the exponential distribution model of the present invention.

[0098] Figure 8 (a) is a schematic diagram of the standard deviation of the normal distribution of each set in each time period of the present invention;

[0099] Figure 8 (b) is a schematic diagram of the standard deviation of the index distribution of each set in each time period of the present invention. Detailed Implementation

[0100] To establish a refined reliability model for multi-source asynchronous measurements, time-segmented models are needed. With sufficient measurement data, smart meters can provide hourly measurement information with timestamps, theoretically allowing for the establishment of an hourly segmented model. However, currently, most regions only provide smart meter measurement data at midnight daily, making it difficult to obtain hourly or shorter interval data, thus lacking the conditions for establishing an hourly measurement reliability model. Power companies, on the other hand, can provide 15-minute interval measurement data for almost all transformers, whose values ​​can reflect the comprehensive superposition of the load of all users within the distribution area. This reflects the overall electricity consumption pattern of all users in the area while mitigating the impact of the randomness of individual user electricity consumption.

[0101] Therefore, this invention adopts a time-period division method based on distribution transformer load measurement. However, since distribution transformer load reflects the superposition of user electricity consumption behavior and is difficult to reflect the random electricity consumption of individual users, it is not advisable to divide the time periods too finely. Different types of users exhibit significantly different electricity consumption behaviors, thus resulting in different time-period divisions. Due to limitations, this invention only takes residential user load as an example to analyze a reliable time-period division and refined modeling method applicable to their typical daily load characteristics. Provided that measurement information is available, this method can also be extended to other types of users such as industrial and commercial users.

[0102] This invention proposes a reliability determination method based on multi-source asynchronous measurement data to improve the accuracy of distribution network status analysis. Using multi-source asynchronous measurement data, a time-segmented reliability model is established. This model has strong applicability; compared to a single-day model, time-segmented modeling can more precisely reflect the magnitude and variation of parameters in the reliability model. The accuracy of the multi-source asynchronous measurement reliability model is verified through effective processing and analysis of offline and online status measurement data.

[0103] Example 1:

[0104] Reliability determination methods based on multi-source asynchronous measurement data, such as Figure 1 As shown, it includes:

[0105] Step 1: Obtain the measurement data and the status of the measurement data;

[0106] Step 2: Select the corresponding pre-built credibility model based on the state of the measurement data;

[0107] Step 3: Input the measurement data into the pre-built credibility model corresponding to the state of the measurement data to determine the credibility of the measurement data;

[0108] The credibility model is constructed by dividing the typical daily load curve based on the slope of the typical daily load curve into time periods, and using the credibility determined by the historical multi-source asynchronous measurement data and the simulation data obtained in each time period.

[0109] The present invention will now be described in detail:

[0110] Before step 1, a method for constructing a credibility model is also included, the specific content of which is as follows: For low-voltage distribution networks, in view of the current situation of delay of multi-source asynchronous measurement data, a credibility determination method based on multi-source asynchronous measurement data is proposed, and credibility models are established under two application scenarios.

[0111] A method for constructing a credibility model, comprising:

[0112] Acquire historical multi-source asynchronous measurement data of the distribution transformer within a set time period, and fit the daily multi-source asynchronous measurement data within the set time period into a daily load curve;

[0113] The daily load curves are normalized, and the correlation between any two daily load curves is calculated using the Pearson correlation coefficient formula.

[0114] The correlation matrix is ​​constructed from the correlations among all daily load curves;

[0115] Calculate the average correlation coefficient between each normalized daily load curve and other normalized daily load curves, and select the maximum value as the typical daily load curve;

[0116] The typical daily load curve is divided into time periods based on its slope.

[0117] Simulate the reliability of multi-source asynchronous measurement data and real-time measurement in different offline and online scenarios;

[0118] Different credibility induction methods were used to summarize and organize the credibility in offline and online scenarios, resulting in offline credibility models and online credibility models.

[0119] The reliability of multi-source asynchronous measurement data and real-time measurements in different offline and online scenarios is simulated and calculated, including:

[0120] In offline scenarios, real-time measurements and multi-source asynchronous measurements with varying degrees of latency are simulated under the collection cycle of conventional smart meters. The corresponding reliability in offline scenarios is obtained by combining the real-time measurements and multi-source asynchronous measurements with varying degrees of latency with a reliability calculation formula.

[0121] In online scenarios, the system simulates real-time measurements at the time of data acquisition, as well as multi-source asynchronous measurements with latency in the one or two reading intervals preceding the acquisition time. The system then combines multi-source asynchronous measurement data with varying degrees of latency with real-time measurements using a reliability calculation formula to obtain the corresponding reliability in the online scenario.

[0122] Different credibility induction methods were used to summarize and organize credibility in offline and online scenarios, resulting in offline credibility models and online credibility models, including:

[0123] For credibility in offline scenarios, the credibility of the collection times contained in each time period is grouped into the same set, and the set is used as the offline credibility model;

[0124] For credibility in online scenarios, credibility with the same latency is grouped into the same credibility set;

[0125] The credibility of data points with the same degree of delay at each collection time point within each time period is combined to obtain the credibility set for each time period.

[0126] The credibility set of each time period is used as the online credibility model.

[0127] Data foundation for the reliability of multi-source asynchronous measurements:

[0128] Real-time smart meter measurement data. In urban areas, real-time smart meters are installed at multiple user power supply points. These meters are smart meters with high-frequency waveform recording capabilities, capable of collecting high-frequency voltage, current, and other measurement information. This data is used for non-intrusive load monitoring research to extract steady-state and transient characteristics of the load. Real-time smart meters can provide measurement data with almost no latency, which can be used to simulate real-time measurements under the collection cycle of conventional smart meters and asynchronous measurements with varying degrees of latency, thereby establishing a reliability model for multi-source asynchronous measurement data.

[0129] Distribution transformer measurement data. In urban areas, intelligent terminals are installed at the distribution transformers. These terminals are intelligent data acquisition and control terminals for low-voltage distribution areas, meeting the needs of high-performance concurrency, large-capacity storage, multiple data acquisition objects, localized analysis and decision-making, and collaborative computing. They have functions such as data acquisition, equipment operation status monitoring, and energy metering. The intelligent terminals collect data such as three-phase voltage, current, active power, and reactive power from the distribution transformers, enabling real-time, 15-minute, hourly, and daily data acquisition and extraction.

[0130] Typical daily load calculation and segmentation of distribution transformers. Sufficient multi-source asynchronous measurement data allows for the establishment of a multi-source asynchronous measurement reliability model based on dates. This embodiment uses a monthly multi-source asynchronous measurement reliability model as an example to illustrate the modeling method. Correspondingly, to assist in lean modeling, it is necessary to filter the distribution transformer data within a month, selecting typical daily load curves to characterize user electricity consumption behavior and characteristics within a month, for time period segmentation.

[0131] Based on the time-segmentation method of the typical daily load curve of a distribution transformer, the following is a detailed introduction to this invention using data collected every 15 minutes from a smart terminal of a certain distribution network as an example:

[0132] The steps are as follows:

[0133] (1) If the data collection interval of the smart terminal installed on the distribution transformer is set to 15 minutes, then each daily load curve is composed of 96 data collection points.

[0134] (2) Normalize multiple daily load curves within a month and calculate their correlation using the Pearson correlation coefficient formula, as shown in equation (1), to form a correlation coefficient with r.xy Let r be the 12th-order correlation matrix of the elements. xy This represents the correlation coefficient between the x-th superimposed load curve and the y-th superimposed load curve.

[0135]

[0136] In the formula, x t y t Let x and y represent the active power of the x-th and y-th normalized daily load curves at time t, respectively. As shown in equation (1), the correlation coefficient between the x-th normalized daily load curve and itself is 1. Therefore, the correlation matrix r is a symmetric matrix with diagonal elements of 1.

[0137] (3) Calculate the average correlation coefficient between each normalized daily load curve and all other daily load curves, i.e., calculate the average value after removing the diagonal elements from each row of the correlation matrix r. Select the curve with the highest average value, i.e., the curve with the highest correlation with other daily load curves, as the typical daily load curve for that month.

[0138] (4) The division of time periods is closely related to user type, data conditions, and application requirements. This embodiment takes residential users as an example, dividing residential load into four time periods: off-peak, morning peak, flat, and evening peak: Off-peak: 1:00-7:00; Morning peak: 8:00-12:00; Flat: 13:00-17:00; Evening peak: 18:00-24:00. The time period division varies slightly depending on the load curve; therefore, adjustments should be made based on the above segmentation, such as... Figure 5 As shown, the boundaries of each segment are adjusted based on the slope of the daily load curve of a typical distribution transformer. The adjusted time period will serve as the basis for the reliability modeling of multi-source asynchronous measurements.

[0139] Multi-source asynchronous measurement reliability modeling. Using historical measurement data from real-time smart meters, this invention simulates real-time measurements under the conventional smart meter acquisition cycle and multi-source asynchronous measurement data with varying degrees of delay, and calculates the relative error between the two, i.e., the reliability of the multi-source asynchronous measurement data. The acquisition cycle of a conventional smart meter is determined by the power company's business needs; this invention sets its acquisition cycle T. A The time delay for multi-source measurements is 15 minutes, and the time delay varies randomly between 0 and 30 minutes. Within one reading interval, the moment when the system first issues a call command is the acquisition moment, and both offline and online status processing are performed at the acquisition moment.

[0140] This embodiment takes the active power of residential users as an example to establish a reliability model for multi-source asynchronous active power measurement. This embodiment uses multi-source asynchronous active power measurement P... A With real-time active power measurement P R percentage of relative error λP To measure its credibility, as shown in equation (2). λ P The smaller the value, the higher the credibility; conversely, the larger the value, the lower the credibility.

[0141]

[0142] In the formula, These represent the real-time active power measurement of the i-th household at time t on day d and the multi-source asynchronous active power measurement actually read back within the reading interval, t∈{0:00,0:15,…23:45}.

[0143] The reliability of all users at all data collection points over a period of time is calculated. Based on the typical daily load curve of the distribution transformer on the corresponding date, the reliability is segmented and organized into different sets, ensuring that each set has an appropriate sample size. Probability distribution verification is performed on each set, and a corresponding reliability set is established based on the sample distribution type.

[0144] The methods for summarizing credibility differ between offline and online scenarios. Therefore, this invention describes the credibility model establishment process for both application scenarios.

[0145] Credibility modeling for offline state processing:

[0146] In offline scenarios, multi-source asynchronous active power measurements are obtained by correcting the average active power within the corresponding reading interval using SCADA measurements.

[0147]

[0148] In the formula, Let t represent the time from t to t+2T for the i-th meter on day d. A Electrical energy within a time period, W i,d,t Let t ~ t+T represent the time of the i-th meter on day d. A Electrical energy within a time period For SCADA active power measurement P at the distribution transformer substation area S d,t Numerical value To superimpose active power, the multi-source asynchronous active power measurements from various smart meters in the same distribution transformer area are superimposed from bottom to top to obtain the superimposed active power. T A The acquisition period T for multi-source asynchronous measurement A .

[0149] The confidence calculation process for multi-source asynchronous measurements is as follows:

[0150] (1) Calculate the multi-source asynchronous active power measurement of the i-th household at time t on day d, simulate the corresponding real-time measurement, and calculate the reliability.

[0151] (2) Since the number of smart meters installed is limited, if a corresponding credibility model is established for each collection time, the sample size of the model will be too low, which will not be representative enough and it will be difficult to accurately determine the distribution type of credibility.

[0152] Therefore, it is necessary to summarize and classify the reliability to ensure that the sample size is appropriate and reasonable. Based on the segmentation principle of the typical daily load curve of the distribution transformer, the degree of load change is similar within the same time period. Therefore, the relative errors of multi-source asynchronous measurements and real-time measurements within each acquisition cycle are similar, i.e., the reliability λ... P Since they are similar in size, the reliability λ of the collection times contained in each time period can be determined. P They are grouped into the same set.

[0153] (3) Due to different users and different times, λ P Due to its inherent uncertainty and randomness, this embodiment analyzes the λ values ​​for the four time periods in Table 2 to explore statistical patterns. P The set is subjected to probability distribution tests, that is, to verify λ when the confidence level is 95%. P Whether the set follows common distribution types such as normal distribution, exponential distribution, gamma distribution, Rayleigh distribution, etc.

[0154] Credibility modeling for online state processing:

[0155] In online scenarios, multi-source asynchronous active power measurements are simulated by multi-source measurements with time delays from the one or two reading intervals preceding the acquisition time. The reliability calculation process for multi-source asynchronous measurements is as follows:

[0156] (1) Calculate the multi-source asynchronous active power measurement of the i-th household at time t on day d, simulate the corresponding real-time measurement, and calculate the reliability. If the time delay is measured in minutes and varies randomly within 0–30 minutes, then there are 31 different multi-source asynchronous active power measurements at each acquisition time. Therefore, 31 different confidence levels can be calculated and assigned to their respective confidence level sets, such as… Figure 2 As shown:

[0157] In online mode, this embodiment sets the calculation cycle to equal the acquisition cycle of multi-source asynchronous measurements, calculating once every 15 minutes. In online mode, only measurement data prior to the calculation time can be acquired; therefore, it is necessary to use data as real-time as possible, rather than waiting for all multi-source asynchronous measurements from a certain reading interval to arrive before calculation. Thus, it is necessary to use the multi-source asynchronous measurement values ​​from the previous one or even the previous two reading intervals as the multi-source asynchronous measurements. (Based on active power...) For example, in Figure 2 In China, for The simulation method and the selection method for multi-source asynchronous active power measurement are shown in the following formula:

[0158]

[0159] P represents the multi-source asynchronous active power measurement collected by the smart meter; t represents the calculation time in the online state, t∈{0:00,0:15,…23:45}; t d1 t d2 These represent the timestamps of the multi-source asynchronous measurements acquired at the previous and two previous reading intervals before time t, respectively, for the i-th smart meter on day d. In this embodiment, the delay time is set in minutes, ranging from 0 to 2T. A Internal random variation. Therefore, t d1 t d2 The range of values ​​for t is d1 ∈{tT A ,tT A +1,…,t+T A}, t d2 ∈{t-2T A ,t-2T A +1,…,t}. Based on this setting, the two reading intervals before time t contain multi-source asynchronous active power measurements. All can arrive before time t, while the active power measurement time t of the previous reading interval is... d1 When the time delay is greater than t, the instantaneous delay time is greater than T. A ,but It cannot arrive before time t; in this case, only one method can be used. State calculation at time t; if t d1 ≤t, meaning the delay time is less than or equal to T. A ,but All can arrive at or before time t. At this time, the active power measurement that is closest to time t, i.e., has a larger scale, is selected to participate in the state calculation at time t. (2) Calculate λ for the i-th household at each of the 96 collection times on day d. PSimilar to offline processing, considering the similar load change trends within the same time period, the reliability of data points with the same degree of delay at each collection time within each time period can be merged. That is, the data points with the same λ number within each time period can be merged. P The datasets were merged to increase the sample size. After merging, the user had 31 lambda values ​​for each time period within that date. P gather.

[0160] (3) Calculate λ for all users P And divided according to time periods. Therefore, all users have 31 λ values ​​in time periods 1 through 4. P gather.

[0161] (4) For each time period in (3), λ P The set is subjected to probability distribution tests, that is, to verify λ when the confidence level is 95%. P Determine whether the set follows common distribution types such as normal distribution, exponential distribution, gamma distribution, Rayleigh distribution, etc., and calculate the distribution parameters. If the parameter range is reasonable for each time period, then the distribution model can be used as a credibility model.

[0162] Step 1: Obtain the measurement data and the status of the measurement data;

[0163] Step 2, which involves selecting a pre-built credibility model based on the state of the measurement data, specifically includes:

[0164] When the measurement data is offline, a pre-built offline reliability model is selected;

[0165] When the measurement data is online, a pre-built online credibility model is selected.

[0166] Step 3, which involves inputting the measurement data into the pre-built credibility model corresponding to the state of the measurement data to determine the credibility of the measurement data, specifically includes:

[0167] The offline measurement data is input into a pre-built offline credibility model to obtain the credibility of the offline measurement data.

[0168] The online status measurement data is input into a pre-built online credibility model to obtain the credibility of the online status measurement data.

[0169] The reliability determination method based on multi-source asynchronous measurement data provided by the present invention further includes: performing a probability distribution test on the reliability set to obtain the distribution type of the reliability set, as follows:

[0170] To verify the accuracy of the latency processing strategy for multi-source asynchronous measurement data, a reliability modeling method based on multi-source asynchronous measurement data is proposed to address the latency problem. To refine the reliability model, a time-segmented normal distribution model is established and validated, demonstrating its strong applicability. Compared to a single-day model, time-segmented modeling can more precisely reflect the parameter magnitudes and changes in the reliability model. By analyzing data results in offline and online application scenarios, the advantages of this method—easy data acquisition and easy method promotion—are demonstrated.

[0171] Example 2:

[0172] This embodiment provides a reliability determination method based on multi-source asynchronous measurement data. The reliability model establishment process is divided into two application scenarios: online and offline. The specific steps are as follows:

[0173] Historical measurement data basis for the reliability of multi-source asynchronous measurements:

[0174] (1) Real-time historical measurement data of smart meters:

[0175] This embodiment acquired historical measurement data from six smart meters installed at the main power supply inlet for residential users. The measurement data was time-stamped, collected at a frequency of 10Hz, and included measurements of electrical energy, active power, voltage amplitude, and current amplitude. The data collection period was concentrated between June 1st and June 30th, 2021. The method described in this paper was used to simulate real-time measurements at different collection times under different application scenarios and multi-source asynchronous measurements with varying degrees of time delay.

[0176] When the sampling frequency is 10Hz, theoretically, each household can acquire 864,000 data points per day. Among the 53 sets of daily household data collected, 41 sets have basically complete data with a data missing rate of about 2%, while 12 sets have multiple hours of missing data with a data missing rate of more than 10%. This embodiment only cleans the former set to simulate multi-source asynchronous measurement and real-time measurement.

[0177] (2) Data measured by the matching variable:

[0178] Since this embodiment could not obtain the measurement data of the upstream distribution transformer of the real-time smart meter, the historical load measurement data of 12 10kV pole-mounted distribution transformers within the same date range were selected as a substitute. Following the method described in the text, a daily load curve with the highest correlation to other load curves was selected as the typical daily load curve for June. Figure 3 As shown. The measurement data acquisition cycle for the pole-mounted distribution transformer is 15 minutes, therefore a typical daily load curve contains 96 load data points. According to... Figure 3The slope of the curve indicates the severity of load changes, while the trend of the standard deviation of the reliability calculated based on historical measurement data from real-time smart meters over different time periods assists in the judgment. Figure 4 As shown.

[0179] Depend on Figure 3 and Figure 4 It can be seen that the changes in residents' load all follow the trend of low-low-morning-slow-evening-peak. Therefore, this embodiment divides the day into four time periods and establishes a reliability model for each of the four time periods. The segmented results are as follows: Figure 4 As shown in Table 1.

[0180] Table 1. Segmentation of typical daily load curves in mid-to-late June

[0181]

[0182] Multi-source asynchronous measurement reliability model:

[0183] (1) Reliability model of multi-source asynchronous measurement in offline state processing

[0184] Using real-time measurement data, confidence models for four time periods were established according to the method described in the paper. In the offline scenario, there is only one confidence set for each time period, and its sample size is shown in Table 2.

[0185] Table 2 Sample size of all user credibility indicators

[0186]

[0187] For each of the four time periods, λ P The sets were subjected to probability distribution tests. The tests showed that all the sets followed both normal and exponential distributions. The probability density functions and parameters for different distribution types are shown below:

[0188] 1) Normal distribution model

[0189] The frequency distribution histograms of the normal distribution for each time period and the corresponding fitted normal distribution probability density function curves are shown below. Figure 6 As shown in Table 3, the parameters of the normal distribution are as follows.

[0190] Table 3. Parameters of the normal distribution model for each time period

[0191]

[0192] Depend on Figure 6 (a) Figure 6 (b) Figure 6 (c) Figure 6 (d) and Table 3 show that λ in time period 1 and time period 3 PThe mean and standard deviation of the set are smaller than those of time periods 2 and 4. The smaller the normal distribution parameter for a given time period, the greater the λ within that period. P The smaller the value, the more concentrated the distribution, meaning the difference between multi-source asynchronous active power and real-time active power is smaller. This conclusion is consistent with the time period division. Therefore, the normal distribution model shown in Table 3 can be used to describe the reliability of multi-source asynchronous measurements in offline scenarios.

[0193] 2) Exponential distribution model:

[0194] Taking the absolute value of the credibility index for each time period, each set follows an exponential distribution. The frequency distribution histogram and the corresponding fitted exponential distribution probability density function curve are shown below. Figure 7 (a) Figure 7 (b) Figure 7 (c) and Figure 7 As shown in (d), the parameters of the exponential distribution are shown in Table 4.

[0195] Table 4. Index distribution parameters for each time period

[0196]

[0197] The parameter θ of the exponential distribution represents the expected value of the distribution, and can also represent the standard deviation. As shown in Tables 4 and 5, the parameters of the exponential distribution also conform to the results of the time period division. However, unlike the normal distribution, the standard deviation of the exponential distribution is not equal to the standard deviation of the samples in the set; instead, it is obtained by maximum likelihood estimation, and its value is smaller than the standard deviation of the samples themselves.

[0198] (2) Reliability model of multi-source asynchronous measurement for online status processing.

[0199] Using real-time measurement data, and following the method of this embodiment, a confidence model is established for different levels of latency in four time periods. In the online scenario, there are 31 confidence sets in each time period, and each confidence set represents a latency situation. The sample size of each set in each time period is the same, as shown in Table 5.

[0200] Table 5 Sample size of all user credibility indicators

[0201]

[0202] For all 124 λ values ​​in the four time periods P The set was subjected to probability distribution testing. The test showed that the set also follows both a normal distribution and an exponential distribution. The parameters for different distribution types are shown below:

[0203] 1) Normal distribution model:

[0204] All sets in each time period follow a normal distribution, and the standard deviation of each set is as follows: Figure 8As shown in (a).

[0205] Depend on Figure 8 (a) It can be seen that within the same time period, the larger the set number, the longer the time-scale distance state estimation of multi-source asynchronous active power measurements, the larger the standard deviation of its normal distribution, and the more dispersed the distribution. The λ within this set... P The larger the maximum value, the greater the difference between multi-source asynchronous active power and real-time active power, and the lower the reliability. This conclusion is consistent with objective laws. Within time periods 1 and 3, λ... P The standard deviation of the set is smaller than that of time periods 2 and 4, a conclusion consistent with the time period division. Therefore, Figure 8 The normal distribution model shown in (a) can be used to describe the reliability of multi-source asynchronous measurements.

[0206] 2) Exponential distribution model:

[0207] All sets in each time period also follow an exponential distribution, and the exponential distribution parameters of each set are as follows: Figure 8 As shown in (b). Figure 8 (b) It can be seen that the trend of the exponential distribution parameter is the same as that of the normal distribution, but the standard deviation parameter of each set is smaller than that of the normal distribution. This is also because the parameters of the exponential distribution are obtained through maximum likelihood estimation rather than statistical acquisition.

[0208] Example 3:

[0209] Based on the same inventive concept, this invention also provides a reliability determination system based on multi-source asynchronous measurement data, including:

[0210] The acquisition module is used to acquire time-stamped measurement data and the status of the measurement data;

[0211] The credibility calculation module is used to select a pre-built credibility model based on the state of the measurement data to determine the credibility of the measurement data;

[0212] The credibility model is constructed by dividing the typical daily load curve based on the slope of the typical daily load curve into time periods, and using the credibility determined by the historical multi-source asynchronous measurement data and the simulation data obtained in each time period.

[0213] The reliability determination system based on multi-source asynchronous measurement data also includes a model building module, specifically used for:

[0214] The parameter processing submodule is used to acquire historical multi-source asynchronous measurement data of the distribution transformer within a set time period, and to fit the daily multi-source asynchronous measurement data within the set time period into a load curve.

[0215] The correlation calculation submodule is used to normalize the load curves and calculate the correlation between any two load curves according to the Pearson correlation coefficient formula, and the correlation matrix is ​​formed by the correlation between all load curves.

[0216] The typical load curve determination submodule is used to calculate the average value of the correlation coefficient between each normalized daily load curve and other normalized daily load curves, and select the maximum value as the typical daily load curve.

[0217] The time period segmentation submodule is used to segment the typical daily load curve into time periods based on the slope of the typical daily load curve;

[0218] The simulation calculation submodule is used to simulate the reliability of multi-source asynchronous measurement data and real-time measurement in different offline and online scenarios.

[0219] The induction and organization submodule is used to summarize and organize the credibility in different offline and online scenarios using different credibility induction methods, and obtain offline credibility models and online credibility models respectively.

[0220] The simulation calculation submodule is specifically used for:

[0221] In offline scenarios, we simulate real-time measurements under the conventional smart meter data collection cycle and multi-source asynchronous measurement data with varying degrees of latency.

[0222] In online scenarios, simulate real-time measurement at the acquisition moment, as well as multi-source asynchronous measurement with latency in the one or two reading intervals before the acquisition moment;

[0223] The credibility is obtained by combining multi-source asynchronous measurement data with varying degrees of time delay with real-time measurement using a credibility calculation formula.

[0224] The summarization and organization submodule is specifically used for:

[0225] For the credibility obtained in the offline scenario, the credibility of the collection time included in each time period is grouped into the same set, and the set is used as the offline credibility model;

[0226] For credibility obtained in online scenarios, credibility with the same latency is grouped into the same credibility set;

[0227] Based on the time period, the credibility of the data collection times within each time period with the same degree of delay is merged to obtain the credibility set of each time period;

[0228] The credibility set of each time period is used as the online credibility model.

[0229] For ease of description, the various parts of the above device are described separately as modules or units based on their functions. Of course, in implementing this invention, the functions of each module or unit can be implemented in one or more software or hardware components.

[0230] Example 4:

[0231] A model building module for a credibility model includes:

[0232] The parameter processing submodule is used to acquire historical multi-source asynchronous measurement data of the distribution transformer within a set time period, and to fit the daily multi-source asynchronous measurement data within the set time period into a load curve.

[0233] The correlation calculation submodule is used to normalize the load curves and calculate the correlation between any two load curves according to the Pearson correlation coefficient formula, and the correlation matrix is ​​formed by the correlation between all load curves.

[0234] The typical load curve determination submodule is used to calculate the average value of the correlation coefficient between each normalized daily load curve and other normalized daily load curves, and select the maximum value as the typical daily load curve.

[0235] The time period segmentation submodule is used to segment the typical daily load curve into time periods based on the slope of the typical daily load curve;

[0236] The simulation calculation submodule is used to simulate the reliability of multi-source asynchronous measurement data and real-time measurement in different offline and online scenarios.

[0237] The induction and organization submodule is used to summarize and organize the credibility in different offline and online scenarios using different credibility induction methods, and obtain offline credibility models and online credibility models respectively.

[0238] The simulation calculation submodule specifically includes:

[0239] In offline scenarios, real-time measurements and multi-source asynchronous measurements with varying degrees of latency are simulated under the collection cycle of conventional smart meters. The corresponding reliability in offline scenarios is obtained by combining the real-time measurements and multi-source asynchronous measurements with varying degrees of latency with a reliability calculation formula.

[0240] In online scenarios, the system simulates real-time measurements at the time of data acquisition, as well as multi-source asynchronous measurements with latency in the one or two reading intervals preceding the acquisition time. The system then combines multi-source asynchronous measurement data with varying degrees of latency with real-time measurements using a reliability calculation formula to obtain the corresponding reliability in the online scenario.

[0241] The summarization and organization submodule is specifically used for:

[0242] For credibility in offline scenarios, the credibility of the collection times contained in each time period is grouped into the same set, and the set is used as the offline credibility model;

[0243] For credibility in online scenarios, credibility with the same latency is grouped into the same credibility set;

[0244] The credibility of data points with the same degree of delay at each collection time point within each time period is combined to obtain the credibility set for each time period.

[0245] The credibility set of each time period is used as the online credibility model.

[0246] Example 5:

[0247] Based on the same inventive concept, in another embodiment of the present invention, a computing device is provided. This computing device includes a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used to execute the steps of a reliability determination method based on multi-source asynchronous measurement data or a reliability model construction method.

[0248] Example 6:

[0249] Based on the same inventive concept, in another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory). This computer-readable storage medium is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the reliability determination method or the reliability model construction method based on multi-source asynchronous measurement data in the above embodiments.

[0250] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0251] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0252] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0253] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0254] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of the claims of the present invention pending approval.

Claims

1. A method for determining the reliability of multi-source asynchronous measurement data, characterized in that, include: Acquire measurement data and the status of the measurement data; Select the corresponding pre-built credibility model based on the state of the measurement data; The confidence level of the measurement data is determined by inputting the measurement data into the pre-built confidence model corresponding to the state of the measurement data. The credibility model is constructed by dividing the typical daily load curve based on the slope of the typical daily load curve into time periods and using the credibility determined by the historical multi-source asynchronous measurement and simulation data in each time period. The credibility model includes an offline credibility model and an online credibility model; the construction of the credibility model includes: Acquire historical multi-source asynchronous measurement data of the distribution transformer within a set time period, and fit the daily multi-source asynchronous measurement data within the set time period into a daily load curve; The daily load curves are normalized, and the correlation between any two daily load curves is calculated using the Pearson correlation coefficient formula. The correlation matrix is ​​constructed from the correlations among all daily load curves; Calculate the average correlation coefficient between each normalized daily load curve and other normalized daily load curves, and select the maximum value as the typical daily load curve; The typical daily load curve is divided into time periods based on its slope. Simulate the reliability of multi-source asynchronous measurement data and real-time measurement in different offline and online scenarios; Different credibility induction methods are used to summarize and organize the credibility in offline and online scenarios respectively, so as to obtain the offline credibility model and the online credibility model; The reliability of the simulated calculations of multi-source asynchronous measurement data and real-time measurements in different offline and online scenarios includes: In offline scenarios, real-time measurements and multi-source asynchronous measurements with varying degrees of latency are simulated under the collection cycle of conventional smart meters. The corresponding reliability in offline scenarios is obtained by combining the real-time measurements and multi-source asynchronous measurements with varying degrees of latency with a reliability calculation formula. In online scenarios, the real-time measurement at the time of collection is simulated, as well as the multi-source asynchronous measurement with time delay in the one or two reading intervals before the time of collection. The credibility in the online scenario is obtained by combining the multi-source asynchronous measurement data with different degrees of time delay and the real-time measurement with the credibility calculation formula. The credibility calculation formula is as follows: ； In the formula, For the i-th meter on day d Electrical energy within a time period For the i-th meter on day d Electrical energy within a time period For SCADA active power measurement at the distribution transformer area Numerical value To superimpose active power, The acquisition cycle for multi-source asynchronous measurements; The measurement data is multi-source asynchronous measurement data, which includes: real-time smart meter measurement data and distribution variable measurement data.

2. The method as described in claim 1, characterized in that, The credibility of offline and online scenarios is summarized and organized using different credibility induction methods to obtain offline credibility models and online credibility models, including: For credibility in offline scenarios, the credibility of the collection times contained in each time period is grouped into the same set, and the set is used as the offline credibility model; For credibility in online scenarios, credibility with the same latency is grouped into the same credibility set; The credibility of data points with the same degree of delay at each collection time point within each time period is combined to obtain the credibility set for each time period. The credibility set of each time period is used as the online credibility model.

3. The method as described in claim 1, characterized in that, The state selection based on the measurement data, corresponding to the pre-built credibility model, includes: When the measurement data is offline, a pre-built offline reliability model is selected; When the measurement data is online, a pre-built online credibility model is selected.

4. The method as described in claim 2, characterized in that, Also includes: Perform a probability distribution test on the credibility set to obtain the distribution type of the credibility set.

5. A method for constructing a credibility model, characterized in that, include: Acquire historical multi-source asynchronous measurement data of the distribution transformer within a set time period, and fit the daily multi-source asynchronous measurement data within the set time period into a daily load curve; The daily load curves are normalized, and the correlation between any two daily load curves is calculated using the Pearson correlation coefficient formula. The correlation matrix is ​​constructed from the correlations among all daily load curves; Calculate the average correlation coefficient between each normalized daily load curve and other normalized daily load curves, and select the maximum value as the typical daily load curve; The typical daily load curve is divided into time periods based on its slope. Simulate the reliability of multi-source asynchronous measurement data and real-time measurement in different offline and online scenarios; Different credibility induction methods were used to summarize and organize the credibility in offline and online scenarios, resulting in offline credibility models and online credibility models; The reliability of the simulated calculations of multi-source asynchronous measurement data and real-time measurements in different offline and online scenarios includes: In offline scenarios, real-time measurements and multi-source asynchronous measurements with varying degrees of latency are simulated under the collection cycle of conventional smart meters. The corresponding reliability in offline scenarios is obtained by combining the real-time measurements and multi-source asynchronous measurements with varying degrees of latency with a reliability calculation formula. In online scenarios, the real-time measurement at the time of collection is simulated, as well as the multi-source asynchronous measurement with time delay in the one or two reading intervals before the time of collection. The credibility in the online scenario is obtained by combining the multi-source asynchronous measurement data with different degrees of time delay and the real-time measurement with the credibility calculation formula. The credibility calculation formula is as follows: ； In the formula, For the i-th meter on day d Electrical energy within a time period For the i-th meter on day d Electrical energy within a time period For SCADA active power measurement at the distribution transformer area Numerical value To superimpose active power, This refers to the acquisition cycle for multi-source asynchronous measurements.

6. The method as described in claim 5, characterized in that, The credibility of offline and online scenarios is summarized and organized using different credibility induction methods to obtain offline credibility models and online credibility models, including: For credibility in offline scenarios, the credibility of the collection times contained in each time period is grouped into the same set, and the set is used as the offline credibility model; For credibility in online scenarios, credibility with the same latency is grouped into the same credibility set; The credibility of data points with the same degree of delay at each collection time point within each time period is combined to obtain the credibility set for each time period. The credibility set of each time period is used as the online credibility model.

7. A system for implementing the reliability determination method based on multi-source asynchronous measurement data as described in any one of claims 1-4, characterized in that, include: The acquisition module is used to acquire time-stamped measurement data and the status of the measurement data; The selection module is used to select the corresponding pre-built credibility model based on the state of the measurement data; A credibility calculation module is used to input the measurement data into the pre-built credibility model corresponding to the state of the measurement data to determine the credibility of the measurement data; The credibility model is constructed by dividing the typical daily load curve based on the slope of the typical daily load curve into time periods, and using the credibility determined by the historical multi-source asynchronous measurement data and the simulation data obtained in each time period.

8. The system as described in claim 7, characterized in that, It also includes a model building module for building credibility models.

9. The system as described in claim 8, characterized in that, The model building module includes: The parameter processing submodule is used to acquire historical multi-source asynchronous measurement data of the distribution transformer within a set time period, and to fit the daily multi-source asynchronous measurement data within the set time period into a load curve. The correlation calculation submodule is used to normalize the load curves and calculate the correlation between any two load curves according to the Pearson correlation coefficient formula, and the correlation matrix is ​​formed by the correlation between all load curves. The typical load curve determination submodule is used to calculate the average value of the correlation coefficient between each normalized daily load curve and other normalized daily load curves, and select the maximum value as the typical daily load curve. The time period segmentation submodule is used to segment the typical daily load curve into time periods based on the slope of the typical daily load curve; The simulation calculation submodule is used to simulate the reliability of multi-source asynchronous measurement data and real-time measurement in different offline and online scenarios. The induction and organization submodule is used to summarize and organize the credibility in different offline and online scenarios using different credibility induction methods, and obtain offline credibility models and online credibility models respectively.

10. The system as described in claim 9, characterized in that, The simulation calculation submodule is specifically used for: In offline scenarios, real-time measurements and multi-source asynchronous measurements with varying degrees of latency are simulated under the collection cycle of conventional smart meters. The corresponding reliability in offline scenarios is obtained by combining the real-time measurements and multi-source asynchronous measurements with varying degrees of latency with a reliability calculation formula. In online scenarios, the system simulates real-time measurements at the time of data acquisition, as well as multi-source asynchronous measurements with latency in the one or two reading intervals preceding the acquisition time. The system then combines multi-source asynchronous measurement data with varying degrees of latency with real-time measurements using a reliability calculation formula to obtain the corresponding reliability in the online scenario.

11. The system as described in claim 9, characterized in that, The summarization and organization submodule is specifically used for: For credibility in offline scenarios, the credibility of the collection times contained in each time period is grouped into the same set, and the set is used as the offline credibility model; For credibility in online scenarios, credibility with the same latency is grouped into the same credibility set; The credibility of data points with the same degree of delay at each collection time point within each time period is combined to obtain the credibility set for each time period. The credibility set of each time period is used as the online credibility model.

12. A model building module for implementing the method for building a credibility model as described in any one of claims 5-6, characterized in that, include: The parameter processing submodule is used to acquire historical multi-source asynchronous measurement data of the distribution transformer within a set time period, and to fit the daily multi-source asynchronous measurement data within the set time period into a load curve. The correlation calculation submodule is used to normalize the load curves and calculate the correlation between any two load curves according to the Pearson correlation coefficient formula, and the correlation matrix is ​​formed by the correlation between all load curves. The typical load curve determination submodule is used to calculate the average value of the correlation coefficient between each normalized daily load curve and other normalized daily load curves, and select the maximum value as the typical daily load curve. The time period segmentation submodule is used to segment the typical daily load curve into time periods based on the slope of the typical daily load curve; The simulation calculation submodule is used to simulate the reliability of multi-source asynchronous measurement data and real-time measurement in different offline and online scenarios. The induction and organization submodule is used to summarize and organize the credibility in different offline and online scenarios using different credibility induction methods, and obtain offline credibility models and online credibility models respectively.

13. The model building module as described in claim 12, characterized in that, The simulation calculation submodule specifically includes: In offline scenarios, real-time measurements and multi-source asynchronous measurements with varying degrees of latency are simulated under the collection cycle of conventional smart meters. The corresponding reliability in offline scenarios is obtained by combining the real-time measurements and multi-source asynchronous measurements with varying degrees of latency with a reliability calculation formula. In online scenarios, the system simulates real-time measurements at the time of data acquisition, as well as multi-source asynchronous measurements with latency in the one or two reading intervals preceding the acquisition time. The system then combines multi-source asynchronous measurement data with varying degrees of latency with real-time measurements using a reliability calculation formula to obtain the corresponding reliability in the online scenario.

14. A computer device, characterized in that, include: One or more processors; A processor is used to execute one or more programs; When the one or more programs are executed by the one or more processors, they implement the reliability determination method based on multi-source asynchronous measurement data as described in any one of claims 1-4 or the reliability model construction method as described in any one of claims 5-6.

15. A computer-readable storage medium, characterized in that, It contains a computer program, which, when executed, implements the reliability determination method based on multi-source asynchronous measurement data as described in any one of claims 1-4 or the reliability model construction method as described in any one of claims 5-6.

Images