Search

CN-121980492-A - Multi-source heterogeneous data fusion processing method

CN121980492ACN 121980492 ACN121980492 ACN 121980492ACN-121980492-A

Abstract

The invention relates to the technical field of virtual power plants, in particular to a multi-source heterogeneous data fusion processing method which comprises the steps of obtaining parameter data of VPP equipment from a plurality of data sources, respectively carrying out standardized processing on the parameter data corresponding to each data source to respectively obtain standardized parameter data corresponding to each data source, using an evidence theory to fuse the standardized parameter data corresponding to all the data sources to obtain fusion parameters and a trusted interval of the fusion parameters, carrying out parameter availability evaluation according to the fusion parameters and the trusted interval of the fusion parameters to obtain an evaluation result, and adjusting a parameter use strategy and a weight vector according to the evaluation result, thereby solving the problems of inaccurate aggregate capacity evaluation and VPP scheduling failure caused by inconsistent accuracy of a plurality of data sources and sensors in the prior art.

Inventors

  • HUANG CHUZHI
  • Wang Zita
  • ZHUANG SHAOYANG
  • HUANG ZONGWEI
  • CAI SHANSHAN
  • CHEN WEIPENG

Assignees

  • 泉州亿兴电力工程建设有限公司鲤城自动化分公司

Dates

Publication Date
20260505
Application Date
20251223

Claims (10)

  1. 1. The multi-source heterogeneous data fusion processing method is characterized by comprising the following steps of: Parameter data of VPP equipment from a plurality of data sources are obtained, and standardized processing is carried out on the parameter data corresponding to each data source respectively to obtain standardized parameter data corresponding to each data source respectively; fusing all the standardized parameter data corresponding to the data sources by using an evidence theory to obtain fusion parameters and a trusted interval of the fusion parameters; and carrying out parameter availability evaluation according to the fusion parameters and the trusted intervals of the fusion parameters to obtain evaluation results, and adjusting parameter use strategies and weight vectors according to the evaluation results.
  2. 2. The method for processing multi-source heterogeneous data fusion according to claim 1, wherein the steps of obtaining parameter data of VPP devices from a plurality of data sources, and respectively performing standardization processing on the parameter data corresponding to each data source to obtain standardized parameter data corresponding to each data source respectively, specifically include: the data sources include manufacturer specifications, sensors, historical databases, and similar equipment libraries; acquiring first parameter data of VPP equipment with a data source of manufacturer specification, wherein the data type of the first parameter data is a static value, and the first confidence coefficient corresponding to the first parameter data is 60%; Acquiring second parameter data of VPP equipment with a data source being a sensor, wherein the second parameter data is obtained by real-time running and measurement of the sensor, the sampling frequency is 1 Hz-10 Hz, and a second confidence coefficient corresponding to the second parameter data is determined according to the level of the sensor; acquiring third parameter data of the VPP equipment with a data source being a historical database, wherein the third parameter data is an operation record of approximately 3-6 months in the historical database, and the third confidence coefficient corresponding to the third parameter data is 80%; Obtaining fourth parameter data of VPP equipment with a data source being a similar equipment library, wherein the similar equipment library comprises other grid-connected equipment from the same manufacturer with the same model; Respectively carrying out standardization processing on the first parameter data, the second parameter data, the third parameter data and the fourth parameter data according to a data standardization mapping formula to obtain the first standardization parameter data, the second standardization parameter data, the third standardization parameter data and the fourth standardization parameter data, wherein the data standardization mapping formula is shown in the following formula: ; Wherein, the Any one of the normalized parameter values representing normalized parameter data, Represents any one parameter value of the parameter data, And Representing the maximum and minimum values, respectively, of the specification range of the VPP device.
  3. 3. The multi-source heterogeneous data fusion processing method according to claim 2, wherein the acquiring the second parameter data of the VPP device with the data source being the sensor specifically comprises: Acquiring the precision of at least one sensor and the initial measurement data of the VPP equipment obtained by real-time running measurement of the at least one sensor; An initial weight of the initial measurement data is calculated based on the accuracy of each sensor, as shown in the following equation: ; Wherein, the Representing the initial weight of the initial measurement data corresponding to the f-th sensor, Indicating the accuracy of the f-th sensor, Representing the summation; For each time t, the statistical features of the first n initial measurements of each sensor are calculated according to the following equation: ; ; Wherein, the The mean value is represented as such, The standard deviation is indicated as such, Indicating the time instant at which the time difference between before time instant t and time instant t is h, Representing initial measurement data of the f-th sensor at time t-h; Initial measurement data of each sensor at time t Comparing the statistical features of the first n initial measured data corresponding to the statistical features, when At the time, the initial measurement data of the f-th sensor at the time t Marking as abnormal, and performing weight de-weighting processing on the initial weight of the initial measurement data of the f-th sensor to obtain updated weight, wherein the updated weight is shown in the following formula: ; Wherein, the Representing the updated weights; Representing a weight reduction coefficient; Fusing the initial measurement data of all the sensors according to the initial measurement data corresponding to each sensor and the updated weight to obtain second parameter data corresponding to the sensors as a data source, wherein the fusion formula is shown as follows: ; Wherein, the And the data source is represented as second parameter data corresponding to the sensor.
  4. 4. The method for processing multi-source heterogeneous data according to claim 2, wherein the fusing of all the standardized parameter data corresponding to the data sources using the evidence theory to obtain the fused parameters and the trusted intervals of the fused parameters specifically comprises: defining parameter estimation as a plurality of hypotheses as an identification framework, as shown in the following formula: ; Wherein, the The representation assumes that the first normalized parameter data is fully trusted, The representation assumes that the second normalized parameter data is fully trusted, The representation assumes that the third normalized parameter data is fully trusted, The representation assumes that the fourth normalized parameter data is fully trusted, The parameter estimation is represented by a set of parameters, Representing the first normalized parameter data, Representing the second normalized parameter data, Representing the data of a third standardized parameter, Representing fourth normalized parameter data; Converting the standardized parameter data corresponding to each data source into basic probability distribution estimated for each parameter in the identification framework, wherein the basic probability distribution is shown as the following formula: ; ; ; ; ; ; ; ; ; Wherein, the Representing the corresponding base probability distribution of the first normalized parameter data, Representing a confidence level of the first normalized parameter data; representing the corresponding basic probability distribution of the second normalized parameter data, Indicating the quality parameter of the sensor(s), Representing recent data; represents the basic probability distribution corresponding to the third normalized parameter data, Representing the completeness of the data; representing the basic probability distribution corresponding to the fourth normalized parameter data, The similarity between the currently used VPP equipment and other grid-connected equipment with the same manufacturer and model in a similar equipment library is represented; And fusing the basic probability distribution of each parameter estimation in the identification framework pair by pair sequentially by adopting a Dempster synthesis rule to obtain the credibility of the parameters, wherein the expression of the credibility of the parameters is shown as follows: ; Wherein, the M1 and m2 respectively represent basic probability distribution of each parameter estimation in any two identification frames, and B and C respectively represent subsets of the basic probability distribution from each parameter estimation in the identification frames; the trusted interval of the fusion parameters is calculated as follows: ; ; Wherein, the Representing a trust function, which is the lower bound of a parameter; Representing likelihood functions as upper bounds of parameters, the trusted interval of the parameters being represented as ; And adopting PIGNISTIC probability to calculate fusion parameters, wherein the calculation formula is shown as follows: ; Wherein, the The probability of being PIGNISTIC is determined, 。
  5. 5. The method for processing multi-source heterogeneous data fusion according to claim 4, wherein the estimating the availability of parameters according to the fusion parameters and the trusted intervals of the fusion parameters to obtain an estimation result, and adjusting the parameters to use policies and weight vectors according to the estimation result, specifically comprises: And calculating a parameter availability evaluation value according to the fusion parameters and the trusted interval of the fusion parameters, wherein the calculation formula is shown as follows: ; ; ; ; Wherein, the Representing a fusion parameter availability evaluation value; representing the average credibility of the fusion parameters; Representing a fusion parameter stability evaluation value; representing the standard deviation of the parameter estimates, Representing the mean of the parameter estimates; represents the fusion parameter consistency evaluation value, Standard deviation representing cross validation errors; judging the sizes of the parameter usability evaluation value and the first evaluation threshold value, if Greater than the first evaluation threshold, then the fusion parameters are used, otherwise, judging the sizes of the parameter usability evaluation value and the second evaluation threshold value, if If the data source is larger than the second evaluation threshold, using the fusion parameter and the weighted confidence interval, otherwise using the data source as the first parameter data of the VPP equipment with the manufacturer specification, sending out an alarm signal, and updating the dynamic weight, wherein the updating formula is shown in the following formula: ; Wherein, the Representing the weight corresponding to the standardized parameter data corresponding to the kth data source; Representing the latest estimation error of the kth data source.
  6. 6. The method for processing multi-source heterogeneous data fusion according to claim 5, wherein the calculating process of the fusion parameter average credibility specifically comprises: The cross-validation error is evaluated, and the standard deviation of the cross-validation error is calculated according to the third standardized parameter data by the following calculation process: ; ; ; Wherein, the Representing the relative error of the historical data, Representing the historical estimate data of the object, Representing historical measured data; Representing the average relative error; Representing a mean function; Representing a standard deviation function; parameter variance assessment, standard deviation of the parameter estimation The calculation process of (2) is as follows: ; Calculating a coefficient of variation of the parameter based on the standard deviation of the parameter estimation The following formula is shown: ; And evaluating the data sufficiency, wherein the data sufficiency comprises sample number sufficiency and time span sufficiency, and the sample number sufficiency evaluation process specifically comprises the following steps: if the number of data samples in the parameter data corresponding to each data source is greater than 100, the number sufficiency evaluation value is 1.0, otherwise, if the number of data samples is greater than 10, the number sufficiency evaluation value is expressed as: ; Wherein, the The quantity sufficiency evaluation value is represented, If the number of the data samples is not more than 10, the number sufficiency evaluation value is 0.2; The time span sufficiency evaluation process specifically includes: If the time span of the parameter data corresponding to each data source is more than 30 days, the time coverage rate is 1.0, otherwise, if the time span is more than 3 days, the time coverage rate is expressed as: ; Wherein, the The time coverage is indicated as being the time coverage, Representing a time span, wherein if the time span is not more than 3 days, the time coverage rate is 0.2; Calculating the average credibility of the fusion parameters according to the average relative error, the parameter variation coefficient, the quantity sufficiency evaluation value and the time coverage rate, wherein the average credibility is shown as the following formula: ; The 95% confidence interval for the fusion parameter is expressed as: ; Wherein, the ; If it is Greater than a first evaluation threshold and Greater than the second evaluation threshold, then fusion parameters are used And its 95% confidence interval.
  7. 7. A multi-source heterogeneous data fusion processing device, comprising: The data acquisition and standardization processing module is configured to acquire parameter data of the VPP equipment from a plurality of data sources, and respectively carry out standardization processing on the parameter data corresponding to each data source to respectively acquire standardized parameter data corresponding to each data source; the data fusion module is configured to fuse all the standardized parameter data corresponding to the data sources by using the evidence theory to obtain fusion parameters and a trusted interval of the fusion parameters; And the fusion parameter evaluation module is configured to evaluate the availability of the parameters according to the fusion parameters and the trusted interval of the fusion parameters, obtain an evaluation result, and adjust the parameter use strategy and the weight vector according to the evaluation result.
  8. 8. An electronic device, comprising: One or more processors; A memory for storing one or more programs, When executed by the one or more processors, causes the one or more processors to implement the method of any of claims 1-6.
  9. 9. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the method according to any of claims 1-6.
  10. 10. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any of claims 1-6.

Description

Multi-source heterogeneous data fusion processing method Technical Field The invention relates to the technical field of virtual power plants, in particular to a multi-source heterogeneous data fusion processing method. Background Virtual power plants ((VirtualPowerPlant, VPP)) can provide important regulatory support for power systems by aggregating distributed power, energy storage, adjustable load, and other resources. The different types of equipment (wind power, photovoltaic, energy storage and controllable load) in the VPP come from different manufacturers, the parameter specification and the actual performance of the equipment have obvious differences, namely, the difference between the nominal parameter and the actual measured parameter is 20-40%, a plurality of parameter data sources (manufacturer specification, operation monitoring and historical data) lack of a unified fusion method, the sensor precision is inconsistent, the weight distribution is difficult, the aggregation capacity assessment is inaccurate, and the VPP scheduling is invalid. Disclosure of Invention The application aims to provide a multi-source heterogeneous data fusion processing method aiming at the technical problems. In a first aspect, the present invention provides a multi-source heterogeneous data fusion processing method, including the following steps: Parameter data of VPP equipment from a plurality of data sources are obtained, and standardized processing is carried out on the parameter data corresponding to each data source respectively to obtain standardized parameter data corresponding to each data source respectively; Fusing the standardized parameter data corresponding to all the data sources by using the evidence theory to obtain fusion parameters and a trusted interval of the fusion parameters; And carrying out parameter availability evaluation according to the fusion parameters and the trusted intervals of the fusion parameters to obtain evaluation results, and adjusting the parameter use strategy and the weight vector according to the evaluation results. Preferably, parameter data of VPP equipment from a plurality of data sources is obtained, and the parameter data corresponding to each data source is respectively subjected to standardization processing to obtain standardized parameter data corresponding to each data source, which specifically includes: Data sources include manufacturer specifications, sensors, historical databases, and similar equipment libraries; acquiring first parameter data of VPP equipment with a data source of manufacturer specification, wherein the data type of the first parameter data is a static value, and the first confidence coefficient corresponding to the first parameter data is 60%; Acquiring second parameter data of VPP equipment with a data source being a sensor, wherein the second parameter data is obtained by real-time operation measurement of the sensor, the sampling frequency is 1 Hz-10 Hz, and a second confidence coefficient corresponding to the second parameter data is determined according to the level of the sensor; Acquiring third parameter data of the VPP equipment with the data source being a historical database, wherein the third parameter data is an operation record of approximately 3-6 months in the historical database, and the third confidence coefficient corresponding to the third parameter data is 80%; Obtaining fourth parameter data of VPP equipment with a data source being a similar equipment library, wherein the similar equipment library comprises other grid-connected equipment from the same manufacturer with the same model; Respectively carrying out standardization processing on the first parameter data, the second parameter data, the third parameter data and the fourth parameter data according to a data standardization mapping formula to obtain the first standardization parameter data, the second standardization parameter data, the third standardization parameter data and the fourth standardization parameter data, wherein the data standardization mapping formula is shown as follows: ; Wherein, the Any one of the normalized parameter values representing normalized parameter data,Represents any one parameter value of the parameter data,AndRepresenting the maximum and minimum values, respectively, of the specification range of the VPP plant. Preferably, the acquiring the second parameter data of the VPP device whose data source is a sensor specifically includes: Acquiring the precision of at least one sensor and the initial measurement data of the VPP equipment obtained by real-time running measurement of the at least one sensor; An initial weight of the initial measurement data is calculated based on the accuracy of each sensor, as shown in the following equation: ; Wherein, the Representing the initial weight of the initial measurement data corresponding to the f-th sensor,Indicating the accuracy of the f-th sensor,Representing the summation; For each time