WO2020215748A1 - 电能表任务执行能力检测方法及装置 - Google Patents

电能表任务执行能力检测方法及装置 Download PDF

Info

Publication number
WO2020215748A1
WO2020215748A1 PCT/CN2019/125093 CN2019125093W WO2020215748A1 WO 2020215748 A1 WO2020215748 A1 WO 2020215748A1 CN 2019125093 W CN2019125093 W CN 2019125093W WO 2020215748 A1 WO2020215748 A1 WO 2020215748A1
Authority
WO
WIPO (PCT)
Prior art keywords
matrix
electric energy
task execution
tested
energy meter
Prior art date
Application number
PCT/CN2019/125093
Other languages
English (en)
French (fr)
Inventor
郑思达
刘岩
袁瑞铭
刘科学
易忠林
刘影
杨晓坤
李文文
庞富宽
彭鑫霞
赵思翔
郭皎
张烁
刘晶
戚成飞
张威
魏彤珈
王皓
Original Assignee
国网冀北电力有限公司计量中心
国网冀北电力有限公司电力科学研究院
国家电网有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 国网冀北电力有限公司计量中心, 国网冀北电力有限公司电力科学研究院, 国家电网有限公司 filed Critical 国网冀北电力有限公司计量中心
Publication of WO2020215748A1 publication Critical patent/WO2020215748A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R35/00—Testing or calibrating of apparatus covered by the other groups of this subclass
    • G01R35/04—Testing or calibrating of apparatus covered by the other groups of this subclass of instruments for measuring time integral of power or current

Definitions

  • This application relates to the technical field of electric energy meters, for example, to a method and device for detecting task performance of electric energy meters.
  • the performance of electric energy meters is often experimental values measured under laboratory conditions through various experimental methods. These experimental values are all a specific number. On this basis, a certain amount of samples can be obtained through a large number of experiments.
  • Data the relevant performance of the electric energy meter can be detected by analyzing the sample data.
  • a typical sample data analysis method is principal component analysis and k-means clustering method.
  • the object of this method is numerical sample data, and the data of the task performed by the energy meter is usually whether the task is executed successfully, that is, the task execution data is only success or failure rather than a specific quantitative number. Therefore, the analysis method based on numerical sample data in the related art cannot be used to detect the task execution ability of the electric energy meter.
  • the main method used to analyze the performance of electric energy meters is mainly the traditional principal component analysis method.
  • the object of the traditional principal component analysis method is non-sequential data.
  • the sample data obtained through multiple experiments in related technologies is non-sequential Therefore, traditional principal component analysis methods can be used for this analysis.
  • the task execution of the electric energy meter is sequential, and the task is sent before the task receipt, and whether the task receipt is timely also reflects the ability of the energy meter to perform the task. Therefore, the traditional principal component analysis method for analyzing the performance of the electric energy meter cannot be used to detect the task execution ability of the electric energy meter.
  • the embodiment of the present application provides a method for detecting the task execution ability of an electric energy meter to detect the task execution ability of the electric energy meter.
  • the method includes: The task execution data is expanded in the time domain to obtain the time domain expanded sampling data of each electric energy meter to be tested; the time domain expanded sampling data retains the time sequence of the electric energy meter execution task, and indicates whether the task execution of the electric energy meter to be tested is successful ;
  • the original data matrix is constructed, and the covariance matrix corresponding to the original data matrix is determined according to the original data matrix;
  • the eigenvalues of the covariance matrix are determined according to the characteristics of the covariance matrix Value, determine the eigenvector matrix corresponding to the covariance matrix; according to the number of principal component components, select a preset number of eigenvectors from the eigenvector matrix to form an intercepted eigenvector matrix; according to the intercepted eigenvector matrix and original data
  • the embodiment of the present application also provides an electric energy meter task execution capability detection device for detecting the task execution capability of the electric energy meter.
  • the device includes: a time-domain extended sampling data determination unit configured to perform tasks according to the type of the electric energy meter to be detected And time, the task execution data of each electric energy meter to be detected is expanded in the time domain to obtain the time domain extended sampling data of each electric energy meter to be detected; the time domain extended sampling data retains the time sequence of the execution of the electric energy meter, It also indicates whether the task execution of the electric energy meter to be detected is successful; the original data matrix and covariance matrix determination unit is set to construct the original data matrix according to the time-domain extended sampling data of all the electric energy meters to be detected, and determine the original data according to the original data matrix The covariance matrix corresponding to the matrix; the eigenvector matrix determining unit is configured to determine the eigenvalues of the covariance matrix, and determine the eigenvector matrix corresponding to the covariance matrix according to the e
  • An embodiment of the present application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor executes the method for detecting the task execution capability of the electric energy meter.
  • the embodiment of the present application also provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program for executing the method for detecting the task execution capability of the electric energy meter.
  • FIG. 1 is a schematic flowchart of a method for detecting task execution capability of an electric energy meter in an embodiment of the present application
  • FIG. 2 is a schematic flowchart of a method for detecting task execution capability of an electric energy meter in another embodiment of the present application
  • Fig. 3 is a schematic diagram of the structure of a device for detecting task execution capability of an electric energy meter in an embodiment of the present application.
  • the detection method of the electric energy meter in the related technology cannot detect the task execution ability of the electric energy meter.
  • the object of the traditional principal component analysis method is a numerical value instead of a Boolean quantity, and it cannot analyze and detect whether the task is successfully executed or not.
  • the sample data processed by the traditional principal component analysis method is non-sequential data, and the task execution ability detection needs to consider the time sequence of the task data, so the traditional principle component analysis method is not suitable for the detection of the task execution ability of the electric energy meter.
  • the scheme first expands the task execution data in the time domain according to the time of task execution, and then conducts the test on this basis.
  • the object of the principal component analysis method is the data of the success or failure of the task execution with time sequence, and then the task execution ability of the electric energy meter is detected according to the results of the time domain extended principal component analysis method.
  • FIG. 1 is a schematic flowchart of a method for detecting task execution capability of an electric energy meter in an embodiment of the present application. As shown in FIG. 1, the method includes steps 101 to 105.
  • step 101 according to the task execution type and time of the electric energy meter to be detected, the task execution data of each electric energy meter to be detected is expanded in the time domain to obtain the time domain extended sampling data of each electric energy meter to be detected;
  • the extended sampling data in the time domain retains the time sequence of the task execution of the electric energy meter, and indicates whether the task execution of the electric energy meter to be detected is successful.
  • step 102 the original data matrix is constructed according to the time-domain extended sampling data of all the electric energy meters to be detected, and the covariance matrix corresponding to the original data matrix is determined according to the original data matrix.
  • step 103 the eigenvalues of the covariance matrix are determined, and the eigenvector matrix corresponding to the covariance matrix is determined according to the eigenvalues of the covariance matrix.
  • step 104 according to the number of principal component components, a preset number of eigenvectors are selected from the eigenvector matrix to form an intercepted eigenvector matrix; the principal component matrix is determined according to the intercepted eigenvector matrix and the original data matrix; The principal component matrix indicates the common characteristics of task execution capabilities of all the electric energy meters to be tested.
  • step 105 the task execution capability of each electric energy meter to be tested is tested according to the principal component matrix.
  • the processing object is numerical value, and the data whose task execution success or failure is Boolean cannot be analyzed, and thus the ability of the electric energy meter to perform the task cannot be detected.
  • the embodiment of this application provides The processing object of the technical solution is the task execution data of the electric energy meter, that is, the task execution data is only the data of success or failure, rather than a specific quantified number. Therefore, the method of this application is a method that can process the task execution success or failure. The method of Boolean sample data analysis can then detect the task execution ability of the electric energy meter.
  • the processing object has no time sequence and cannot analyze the execution tasks of the electric energy meter in a sequential order, and thus cannot detect the ability of the electric energy meter to perform tasks
  • the embodiment of the application provides The technical solution is to expand in the time domain, which is suitable for the situation where the task execution of the energy meter is sequential in the actual application environment, because the task of the energy meter is sent before the task receipt, and whether the task receipt is timely also reflects the energy meter Ability to perform tasks is strong or weak. Therefore, the method of the present application is an analysis method that can consider the timing of sample data to realize the detection of the ability of the electric energy meter to perform tasks.
  • the technical solutions provided by the embodiments of the present application realize the detection of the ability of the electric energy meter to perform tasks.
  • the type of task execution of the electric energy meter includes: the type of sending task, the type of successful task execution, and the type of failed task execution.
  • the tasks of the electric energy meter may include tasks such as collection, cost control, time calibration, and price adjustment.
  • step 101 After step 101, a step of normalization processing is performed (see S2 in FIG. 2).
  • the method for detecting task execution capability of electric energy meters based on time-domain extended principal component analysis may further include: normalizing the time-domain extended sampling data of each electric energy meter to be detected to obtain all The normalized time-domain extended sampling data of the electric energy meter.
  • normalizing the time-domain extended sampling data of each electric energy meter to be detected improves the efficiency and accuracy of data processing.
  • the original data matrix is constructed according to the time-domain extended sampling data of all the electric energy meters to be detected, and the covariance matrix corresponding to the original data matrix is determined according to the original data matrix, which may include: The time-domain extended sampling data after a transformation is used to construct the original data matrix, and the covariance matrix corresponding to the original data matrix is determined according to the original data matrix.
  • an original data matrix X with M max rows and N columns is constructed, Among them, N is the number of electric energy meters, and M max is the sampling data of all electric energy meters arranged in sequence according to the execution time, max ⁇ M 1 ,M 2 ,...,M N ⁇ M max /2 ⁇ sum ⁇ M 1 ,M 2 ,...,M N ⁇ , where the task execution data of the j-th electric energy meter has a total of M j groups, 1 ⁇ j ⁇ N, and X is a complex matrix of M max ⁇ N.
  • Calculate X according to the original data matrix X Covariance matrix C, C X H X, C is an N ⁇ N complex matrix.
  • the above step 104 includes the step of determining the number (quantity) of the principal component components (see S5 in FIG. 2).
  • the above step 104 includes the step of determining the principal component matrix (see S6 in FIG. 2). After the number of principal components is determined, according to the number of principal component components, a preset number of eigenvectors (the first k eigenvectors) are selected from the eigenvector matrix to form the intercepted eigenvector matrix, and then the principal component matrix is determined.
  • the common characteristics of the situation (therefore, the principal component matrix can also be called the P common characteristics matrix P), complete the time-domain extended principal component analysis process.
  • step 105 see "S7-S10" in FIG. 2.
  • detecting the task execution capability of each electric energy meter to be detected according to the principal component matrix may include at least one of the following steps: detecting all the tasks of the electric energy meter to be detected according to the number of principal components The difference of execution ability; according to the mean value of all the elements in the first column of the principal component matrix, detect the difference of the task execution ability of all the energy meters to be tested; according to the variance of all the elements in the first column of the principal component matrix, detect all the energy to be tested The task execution ability difference of the common characteristics is shown; according to the modulus of the first row element of the conjugate transpose matrix of the intercepting eigenvector matrix, the task execution ability of all the electric energy meters to be tested is detected.
  • the task execution capability of the electric energy meter is detected according to one or any combination of the above methods, which improves the flexibility and accuracy of detecting the task execution capability of the electric energy meter.
  • detecting the difference in the task execution capability of all the electric energy meters to be detected according to the number of principal component components may include: determining all the tasks of the electric energy meter to be detected when the number of principal components is less than a first threshold The execution ability difference is small, and when the number of principal components is greater than or equal to the first threshold, it is determined that the task execution ability difference of all the electric energy meters to be detected is large.
  • the first threshold can be determined according to the on-site situation when the electric energy meter performs the task.
  • the number k of the principal component components represents the difference in the task execution ability of all electric energy meters.
  • the smaller k represents the smaller difference in the task execution ability of the electric energy meters participating in the evaluation.
  • the larger k represents the participation in the evaluation.
  • the task execution ability of the electric energy meter is very different, and the k value is used to detect the difference in the task execution ability of all N electric energy meters.
  • the principal component component is the result of principal component analysis and calculation, which represents the result in the embodiment of the present application.
  • the detection of the difference in the task performance of all the electric energy meters to be tested according to the average of all elements in the first column of the principal component matrix may include: the average and complex plane of all elements in the first column of the principal component matrix When the difference between the origin point is less than the second threshold, it is determined that all the energy meters to be tested have good performance capabilities, the proportion of successful execution receipts when the task sends data is large, and the time required to receive the receipt is short; in the principal component matrix When the difference between the mean value of all elements in the first column and the origin of the complex plane is greater than or equal to the second threshold, it is determined that the task execution capability of all the energy meters to be tested is poor, and the proportion of successful execution receipts when the task sends data is small, and It takes a long time to receive the receipt.
  • the second threshold value can be determined according to the on-site situation when the electric energy meter performs the task.
  • the first column p 1 of the common feature matrix P is the dominant feature of the task performance of all electric energy meters.
  • the average value of all elements in p 1 is more than the origin of the complex plane (the origin of the complex plane is the origin of the complex number domain, and the horizontal
  • the coordinate represents the real part
  • the ordinate represents the imaginary part
  • the origin is a complex number with both real and imaginary parts being 0).
  • detecting the difference in task performance of all the common features of the electric energy meter to be detected according to the variance of all elements in the first column of the principal component matrix may include: the variance of all elements in the first column of the principal component matrix is less than In the case of the third threshold, the task performance difference of determining the common characteristics of all the electric energy meters to be tested is small; when the variance of all elements in the first column of the principal component matrix is greater than or equal to the third threshold, all the electric energy meters to be tested are determined The task execution ability of common characteristics varies greatly. Among them, the third threshold can be determined according to the on-site situation when the electric energy meter performs the task.
  • the variance of all elements of p 1 indicates the difference in task execution ability of all electric energy meters, and when the variance of all elements of p 1 is less than the third threshold, the tasks of all electric energy meters are executed The ability difference is small. On the contrary, when the variance of all elements of p 1 is greater than or equal to the third threshold, the task execution ability of the common characteristics of all electric energy meters is different.
  • the task performance capability of all the electric energy meters to be detected is detected, which may include: the first row element of the conjugate transpose matrix In the case that the modulus of any element in is greater than the fourth threshold, it is determined that the task performance of the electric energy meter to be detected corresponding to any element is good; the modulus of any element in the first row of the conjugate transpose matrix If it is less than or equal to the fourth threshold, it is determined that the task execution capability of the electric energy meter to be detected corresponding to any element is poor.
  • the fourth threshold can be determined according to the on-site situation when the electric energy meter performs the task.
  • U H be the conjugate transpose matrix of the truncated eigenvector matrix U k
  • the relationship between the element u 1j in the first row and the jth column of the matrix U H is the corresponding relationship between the j-th electric energy meter and the dominant characteristic p 1 respectively.
  • the greater the modulus of u 1j the task execution capability of the j-th electric energy meter The closer to the dominant feature, the better the task performance of the energy meter.
  • the smaller the modulus of u 1j is the more the task performance of the j-th energy meter deviates from the dominant feature, the worse the task performance of the energy meter is.
  • an embodiment of the present application also provides a device for detecting the task execution capability of an electric energy meter, as shown in the following embodiment. Since the principle of the task execution capability detection device of the electric energy meter is similar to the above-mentioned task execution ability detection method of the electric energy meter, the implementation of the task execution ability detection device of the electric energy meter can refer to the implementation of the above-mentioned task execution ability detection method of the electric energy meter, and the repetition will not be repeated. .
  • the term "module” or "unit” can be a combination of software and/or hardware that can realize predetermined functions.
  • the devices described in the following embodiments can be implemented by software, implementation by hardware or a combination of software and hardware is also possible and conceived.
  • FIG. 3 is a schematic structural diagram of a device for detecting task execution capability of an electric energy meter in an embodiment of the present application.
  • the device includes: a time-domain extended sampling data determining unit 10, which is set to execute according to the task execution type of the electric energy meter to be detected And time, the task execution data of each electric energy meter to be detected is expanded in the time domain to obtain the time domain extended sampling data of each electric energy meter to be detected; the time domain extended sampling data retains the time sequence of the execution of the electric energy meter, And indicate whether the task execution of the electric energy meter to be detected is successful; the original data matrix and covariance matrix determining unit 20 is configured to construct the original data matrix according to the time-domain extended sampling data of all the electric energy meters to be detected, and determine the original data matrix according to the original data matrix.
  • the covariance matrix corresponding to the data matrix; the eigenvector matrix determining unit 30 is configured to determine the eigenvalues of the covariance matrix, and determine the eigenvector matrix corresponding to the covariance matrix according to the eigenvalues of the covariance matrix; intercept the eigenvector matrix and
  • the principal component matrix determining unit 40 is configured to select a preset number of eigenvectors from the eigenvector matrix to form an intercepted eigenvector matrix according to the number of principal component components; determine according to the intercepted eigenvector matrix and the original data matrix
  • the above-mentioned detection unit is configured to detect the task execution capability of each electric energy meter to be detected according to at least one of the following steps: according to the number of principal component components, detect all the task execution capabilities of the electric energy meter to be detected Variability; based on the mean value of all elements in the first column of the principal component matrix, detect the difference in task performance of all the energy meters to be tested; based on the variance of all elements in the first column of the principal component matrix, detect the common characteristics of all energy meters to be tested The task execution ability of the difference; according to the modulus of the first row element of the conjugate transpose matrix of the intercepted eigenvector matrix, the task execution ability of all the electric energy meters to be tested is detected.
  • the detection unit is configured to: in the case that the number of principal component components is less than the first threshold, determine that the task performance difference of all the electric energy meters to be tested is small, and when the number of principal component components is greater than or equal to the first threshold, In the case of threshold value, it is determined that the task execution ability of all the electric energy meters to be tested has a large difference; when the difference between the mean value of all elements in the first column of the principal component matrix and the origin of the complex plane is less than the second threshold value, all the energy meters to be tested are determined The task execution ability is good, the proportion of successful execution receipt when the task is sent is large, and the time required to receive the receipt is short; the difference between the mean value of all elements in the first column of the principal component matrix and the origin of the complex plane is greater than or equal to In the case of the second threshold, it is determined that the task execution ability of all the electric energy meters to be tested is poor, the proportion of successful execution receipts when the task sends data is small, and the time required to receive the receipt is long;
  • the above-mentioned energy meter task execution capability detection device may further include: a normalization processing unit configured to perform normalization processing on the time-domain extended sampling data of each electric energy meter to be detected to obtain all the electric energy to be detected
  • the normalized time-domain extended sampling data of the table The normalized time-domain extended sampling data of the table; the original data matrix and the covariance matrix determining unit are set to: construct the original data according to the normalized time-domain extended sampling data of all the electric energy meters to be tested Matrix, according to the original data matrix, determine the covariance matrix corresponding to the original data matrix.
  • the object of traditional principal component analysis methods is numerical values rather than Boolean quantities, and cannot be analyzed and tested for the success or failure of task execution.
  • the sample data processed by traditional principal component analysis methods are non-sequential data, and task execution ability
  • the detection needs to consider the time sequence of task data, so the traditional principal component analysis method is not suitable for the detection of the task execution ability of the electric energy meter.
  • the method of this application expands the task execution data in the time domain according to the time of task execution.
  • the principal component analysis method on this basis is based on the sequential task execution success or failure data, which improves the traditional principal component analysis. The method cannot be used to test the task execution ability of the electric energy meter.
  • An embodiment of the present application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor executes the method for detecting the task execution capability of the electric energy meter.
  • the embodiment of the present application also provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program for executing the method for detecting the task execution capability of the electric energy meter.
  • the task execution capability detection method of the electric energy meter based on the time-domain extended principal component analysis in this embodiment is the data of the task execution of the electric energy meter, that is, the task execution data is only success or failure rather than a specific Quantitative numbers, so the method of this application is a method that can process the data analysis of the Boolean sample composed of task execution success or failure, and then can detect the problem of the task execution ability of the electric energy meter; at the same time, this method is effective in principal component analysis.
  • Expansion of the time domain is suitable for the situation where the task execution of the energy meter is sequential in the actual application environment, because the task of the energy meter is sent before the task receipt, and whether the task receipt is timely also reflects the ability of the energy meter to perform tasks. weak. Therefore, the method of the present application may consider the sequential analysis method of sample data to realize the detection of the ability of the electric energy meter to perform tasks.
  • modules or steps of the above-mentioned embodiments of the application can be implemented by a general computing device, and they can be concentrated on a single computing device or distributed among multiple computing devices.
  • they can be implemented with program codes executable by a computing device, so that they can be stored in a storage device for execution by the computing device, and in some cases, can be executed in a different order than here.
  • the embodiments of the present application are not limited to any specific hardware and software combination.

Abstract

一种电能表任务执行能力检测方法及装置，其中，该方法包括：根据待检测电能表的任务执行的类型和时间，对每个待检测电能表的任务执行数据在时域进行扩展，得到每个待检测电能表的时域扩展采样数据（101）；根据所有待检测电能表的时域扩展采样数据，构建原始数据矩阵，及确定原始数据矩阵对应的协方差矩阵（102）；确定协方差矩阵的特征值和对应的特征向量矩阵（103）；根据主成分分量的数量，从所述特征向量矩阵中选取出预设数量的特征向量组成截取特征向量矩阵；根据截取特征向量矩阵和原始数据矩阵，确定主成分矩阵（104）；根据主成分矩阵，对每个待检测电能表的任务执行能力进行检测（105）。

Description

电能表任务执行能力检测方法及装置
本申请要求在2019年04月23日提交中国专利局、申请号为201910329419.4的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本申请涉及电能表技术领域，例如涉及一种电能表任务执行能力检测方法及装置。
背景技术
为了进一步发挥智能电能表资产效益，减少电网运营成本，提高供电可靠性和用户满意度，目前在智能电能表非计量功能上做了大量研究和应用。智能电能表高级应用已成为配网运营管理的重要手段，拓展计量采集系统的非计量功能开发和应用，建立完善的数据共享机制，制定智能电能表数据支撑各专业应用的工作规则，全面有效支撑运检、发展、安质等专业数据需求。为了支撑相关工作，需要对相关技术中台区全载波、半载波、宽带载波等通信方式下电能表的采集任务执行能力和本地网络通信能力进行检测。目前对采集台区电能表的采集能力和本地网络通信没有进行量化的检测方式，因此需综合主站多数据采集、费控、校时、调价等任务，结合本地通信信道通信监控，实现电能表任务执行能力的可量化和可操作的检测。相关技术中对电能表任务执行能力检测方案的缺点如下。
在相关技术中，电能表的性能往往通过各种实验手段在实验室条件下测量得到的实验值，这些实验值均为一个具体的数字，在此基础上通过大量实验即可获得一定量的样本数据，通过对样本数据进行分析即可检测电能表的相关性能，一种典型的样本数据分析方法为主成分分析和k均值聚类法。但是这种方法针对的对象是数值样本数据，而电能表执行任务的数据通常为任务执行是否成功，即任务执行数据仅为成功或失败而不是一个具体量化的数字。因此，相关技术中基于数值样本数据的分析方法不能用于检测电能表任务执行能力。
目前主要用于分析电能表性能的方法主要是传统的主成分分析法，传统的主成分分析法处理的对象是无时序性的数据，相关技术中通过多次实验取得的样本数据即为无时序性的数据，因此传统主成分分析方法可用于这种分析。但 实际应用环境中，电能表执行任务是具有时序性的，任务发送是先于任务回执的，而且任务回执是否及时也反映了电能表执行任务的能力强弱。因此，传统的用于分析电能表性能的主成分分析方法不能用于检测电能表任务执行能力。
发明内容
本申请实施例提供了一种电能表任务执行能力检测方法，用以检测电能表任务执行能力，该方法包括：根据待检测电能表的任务执行的类型和时间，对每个待检测电能表的任务执行数据在时域进行扩展，得到每个待检测电能表的时域扩展采样数据；所述时域扩展采样数据保留电能表执行任务的时序性，且指示待检测电能表的任务执行是否成功；根据所有待检测电能表的时域扩展采样数据，构建原始数据矩阵，根据原始数据矩阵，确定原始数据矩阵对应的协方差矩阵；确定所述协方差矩阵的特征值，根据协方差矩阵的特征值，确定协方差矩阵对应的特征向量矩阵；根据主成分分量的数量，从所述特征向量矩阵中选取出预设数量的特征向量组成截取特征向量矩阵；根据所述截取特征向量矩阵和原始数据矩阵，确定主成分矩阵；所述主成分矩阵指示所有待检测电能表任务执行能力的共性特征；根据所述主成分矩阵，对每个待检测电能表的任务执行能力进行检测。
本申请实施例还提供了一种电能表任务执行能力检测装置，用以检测电能表任务执行能力，该装置包括：时域扩展采样数据确定单元，设置为根据待检测电能表的任务执行的类型和时间，对每个待检测电能表的任务执行数据在时域进行扩展，得到每个待检测电能表的时域扩展采样数据；所述时域扩展采样数据保留电能表执行任务的时序性，且指示待检测电能表的任务执行是否成功；原始数据矩阵及协方差矩阵确定单元，设置为根据所有待检测电能表的时域扩展采样数据，构建原始数据矩阵，根据原始数据矩阵，确定原始数据矩阵对应的协方差矩阵；特征向量矩阵确定单元，设置为确定所述协方差矩阵的特征值，根据协方差矩阵的特征值，确定协方差矩阵对应的特征向量矩阵；截取特征向量矩阵及主成分矩阵确定单元，设置为根据主成分分量的数量，从所述特征向量矩阵中选取出预设数量的特征向量组成截取特征向量矩阵；根据所述截取特征向量矩阵和原始数据矩阵，确定主成分矩阵；所述主成分矩阵指示所有待检测电能表任务执行能力的共性特征；检测单元，设置为根据所述主成分矩阵，对每个待检测电能表的任务执行能力进行检测。
本申请实施例还提供了一种计算机设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，所述处理器执行所述电能表任务执行能力检测方法。
本申请实施例还提供了一种计算机可读存储介质，所述计算机可读存储介质存储有执行电能表任务执行能力检测方法的计算机程序。
附图说明
此处所说明的附图用来提供对本申请的进一步理解，构成本申请的一部分，并不构成对本申请的限定。在附图中：
图1是本申请实施例中电能表任务执行能力检测方法的流程示意图；
图2是本申请又一实施例中电能表任务执行能力检测方法的流程示意图；
图3是本申请实施例中电能表任务执行能力检测装置的结构示意图。
具体实施方式
下面结合实施方式和附图，对本申请做进一步详细说明。在此，本申请的示意性实施方式及其说明用于解释本申请，但并不作为对本申请的限定。
申请人发现，相关技术中对电能表检测方法无法检测电能表任务执行能力，存在的情况是：传统主成分分析方法的对象是数值而不是布尔量，不能针对任务执行成功与否进行分析和检测，同时传统主成分分析方法处理的样本数据为无时序性的数据，而任务执行能力检测需要考虑任务数据的时序性，因此传统主成分分析方法并不适用于电能表任务执行能力的检测。
申请人提出了一种基于时域扩展主成分分析的电能表任务执行能力检测(评估)方案，该方案首先根据任务执行的时间对任务执行数据在时域进行扩展，然后在此基础上进行的主成分分析方法的对象是具有时序性的任务执行成功与否的数据，进而根据时域扩展主成分分析方法的结果检测电能表任务执行能力。下面对该基于时域扩展分析的电能表任务执行能力检测的方案进行详细介绍如下。
图1是本申请实施例中电能表任务执行能力检测方法的流程示意图，如图1所示，该方法包括步骤101至步骤105。
在步骤101中，根据待检测电能表的任务执行的类型和时间，对每个待检测电能表的任务执行数据在时域进行扩展，得到每个待检测电能表的时域扩展 采样数据；所述时域扩展采样数据保留电能表执行任务的时序性，且指示待检测电能表的任务执行是否成功。
在步骤102中，根据所有待检测电能表的时域扩展采样数据，构建原始数据矩阵，根据原始数据矩阵，确定原始数据矩阵对应的协方差矩阵。
在步骤103中，确定所述协方差矩阵的特征值，根据协方差矩阵的特征值，确定协方差矩阵对应的特征向量矩阵。
在步骤104中，根据主成分分量的数量，从所述特征向量矩阵中选取出预设数量的特征向量组成截取特征向量矩阵；根据所述截取特征向量矩阵和原始数据矩阵，确定主成分矩阵；所述主成分矩阵指示所有待检测电能表任务执行能力的共性特征。
在步骤105中，根据所述主成分矩阵，对每个待检测电能表的任务执行能力进行检测。
本申请实施例提供的技术方案：
与相关技术中传统主成分分析方法的处理对象是数值，不能对任务执行成功与否为布尔量的数据进行分析，进而无法对电能表执行任务能力进行检测的方案相比较，本申请实施例提供的技术方案的处理对象是电能表执行任务的数据，即任务执行数据仅为成功或失败的数据，而不是一个具体量化的数字，因此本申请方法是一种能处理由任务执行成功或失败组成的布尔量样本数据分析的方法，进而可以检测电能表任务执行能力。与相关技术中传统的主成分分析法的处理对象无时序性，不能对具有先后顺序的电能表执行任务进行分析，进而无法对电能表执行任务能力进行检测的方案相比较，本申请实施例提供的技术方案是在时域进行扩展，适用于实际应用环境中电能表执行任务是具有时序性的情况，由于电能表的任务发送是先于任务回执的，而且任务回执是否及时也反映了电能表执行任务的能力强弱。因此，本申请方法为可考虑样本数据时序性的分析方法以实现对电能表执行任务能力的检测。本申请实施例提供的技术方案实现了对电能表执行任务能力的检测。
下面对本申请实施例涉及的各个步骤进行详细介绍如下。
在上述步骤101(参见图2中的S1)中，设参与电能表任务执行能力检测的电能表(待检测电能表)个数为N，原始数据的起始时间为t start，结束时间为t end，则测试时间长度Δt为Δt＝t end-t start，设第j个电能表的任务执行数据共有M j组，其中第i组任务发送指令对应的时间为t 1,ij，任务执行回执对应的时间为t 2,ij。如果 第j个电能表的第i组任务为发送指令，则对应的任务执行数据在时域进行扩展后的时域扩展采样数据为V 1,ij∠α 1,ij，其幅值为V 1,ij＝1，其相角为α 1,ij，α 1,ij＝(t 1,ij-t start)π/Δt。执行任务回执存在两种情况，执行成功或者执行失败，如果第j个电能表的第i组任务为任务执行成功回执，则对应的任务执行数据在时域进行扩展后的时域扩展采样数据为V 2,ij∠α 2,ij，其幅值为V 2,ij＝1，其相角为α 2,ij，α 2,ij＝(t 2,ij-t start)π/Δt+π；如果第j个电能表的第i组任务为任务执行失败回执，则对应的任务执行数据在时域进行扩展后的时域扩展采样数据为V 2,ij∠α 2,ij，其幅值为V 2,ij＝0.5，其相角为α 2,ij，α 2,ij＝(t 2,ij-t start)π/Δt+π。
在一实施例中，电能表的任务执行的类型包括：发送任务的类型、任务执行成功的类型、任务执行失败的类型。
在一实施例中，电能表的任务可以包括：采集、费控、校时、调价等任务。
在步骤101之后，进行归一化处理的步骤(参见图2中S2)。
在一个实施例中，基于时域扩展的主成分分析的电能表任务执行能力检测方法，还可以包括：对每一待检测电能表的时域扩展采样数据进行归一化处理，得到所有待检测电能表的归一化处理后的时域扩展采样数据。
在一实施例中，对每一待检测电能表的时域扩展采样数据进行归一化处理提高了数据处理的效率和精度，归一化处理方法可以包括：对第j个电能表任务执行数据的时域扩展采样数据
Figure PCTCN2019125093-appb-000001
进行归一化，，V ij∠α ij＝[V 1,ij∠α 1,ij,V 2,ij∠α 2,ij] T，i＝1,2,…,M j，j＝1,2,…,N，归一化后得到的第j个电能表任务执行数据的归一化时域扩展采样数据为
Figure PCTCN2019125093-appb-000002
的均值为0，标准差为1。
在一个实施例中，根据所有待检测电能表的时域扩展采样数据，构建原始数据矩阵，根据原始数据矩阵，确定原始数据矩阵对应的协方差矩阵，可以包括：根据所有待检测电能表的归一化处理后的时域扩展采样数据，构建原始数据矩阵，根据原始数据矩阵，确定原始数据矩阵对应的协方差矩阵。
在一个实施例中，在上述步骤102(参见图2中S3)中，构建M max行N列的原始数据矩阵X，
Figure PCTCN2019125093-appb-000003
其中，N为电能表的个数，M max为所有电能表采样数据按执行任务时间先后排列得到的，max{M 1,M 2,…,M N}≤M max/2≤sum{M 1,M 2,…,M N}，其中，第j个电能表的任务执行数据共有M j组，1≤j≤N，X为M max×N的复数矩阵，根据原始数据矩阵X计算X的协方差矩阵C，C＝X HX，C为N×N的复数矩阵。
在一个实施例中，在上述步骤103(参见图2中S4)中，计算矩阵C的全部特征值λ 1,λ 2,…λ N，且λ 1≥λ 2≥…≥λ N≥0，所有特征值均为实数；对于特征值λ j，j＝1,2,…,N，求出齐次线性方程组(λ jI-C)＝0的基础解系，得到C对于λ j的一组特征向量u j，则特征向量矩阵U＝[u 1,u 2,…u N]，U为N×N的复数矩阵，且满足U HCU＝Λ，其中Λ＝diag(λ 1,λ 2,…,λ N)。
上述步骤104包括：确定主成分分量的个数(数量)的步骤(参见图2中的S5)。
在一实施例中，根据格特曼量表(Guttman)准则选择主成分分量的个数k，即k＝max{k|λ k≥1}。
上述步骤104包括：确定主成分矩阵的步骤(参见图2中的S6)。在确定了主成分个数之后，根据主成分分量的数量，从所述特征向量矩阵中选取出预设数量的特征向量(前k个特征向量)组成截取特征向量矩阵，进而确定主成分矩阵。
在一实施例中，根据主成分分量的个数k从所述特征向量矩阵U中选取出第1至k个特征向量，即特征向量矩阵U的第1至k列组成的截取特征向量矩阵U k＝[u 1,u 2,…u k]，U k为N×k的复数矩阵，并对原始数据矩阵X进行变换，得到主成分矩阵P，则P代表了所有N个电能表任务执行情况的共性特征(因此，主成分矩阵也可以称作P共性特征矩阵P)，完成时域扩展主成分分析过程。
在一实施例中，对原始数据矩阵X进行变换，得到主成分矩阵P，可以包括：通过P＝XU k计算主成分矩阵P，P为M max×k的复数矩阵。
上述步骤105，参见图2中的“S7-S10”。
在一个实施例中，根据所述主成分矩阵，对每个待检测电能表的任务执行能力进行检测，可以包括以下至少之一的步骤：根据主成分分量的数量，检测所有待检测电能表任务执行能力的差异性；根据主成分矩阵的第一列所有元素的均值，检测所有待检测电能表任务执行能力的差异性；根据主成分矩阵的第一列所有元素的方差，检测所有待检测电能表共性特征的任务执行能力差异性；根据截取特征向量矩阵的共轭转置矩阵的第一行元素的模，检测所有待检测电能表的任务执行能力。
在一实施例中，按照如上所述方法的其中之一或任意组合，检测电能表任务执行能力，提高了检测电能表任务执行能力的灵活性和精确度。
首先，介绍根据主成分分量的数量，检测所有待检测电能表任务执行能力的差异性的步骤，参见图2中的S7。
在一个实施例中，根据主成分分量的数量，检测所有待检测电能表任务执行能力的差异性，可以包括：在主成分分量的数量小于第一阈值的情况下，确定所有待检测电能表任务执行能力差异小，在主成分分量的数量大于或等于第一阈值的情况下，确定所有待检测电能表任务执行能力差异大。其中，第一阈值可以根据电能表执行任务时现场的情况确定。
在一实施例中，主成分分量的个数k代表了所有电能表任务执行能力的差异性，k越小代表了参与评价的电能表任务执行能力差异小，反之，k越大代表了参与评价的电能表任务执行能力差异大，k值用于检测所有N个电能表任务执行能力的差异性。
在一实施例中，主成分分量是主成分分析计算的结果，在本申请实施例中代表结果。
其次，介绍根据主成分矩阵的第一列所有元素的均值，检测所有待检测电能表任务执行能力的差异性的步骤，参见图2中的S8。
在一个实施例中，根据主成分矩阵的第一列所有元素的均值，检测所有待检测电能表任务执行能力的差异性，可以包括：在主成分矩阵的第一列所有元素的均值与复平面原点的差值小于第二阈值的情况下，确定所有待检测电能表任务执行能力好，在任务发送数据时得到执行成功回执的比重大，且收到回执所需的时间短；在主成分矩阵的第一列所有元素的均值与复平面原点的差值大于或等于第二阈值的情况下，确定所有待检测电能表任务执行能力差，在任务发送数据时得到执行成功回执的比重小，且收到回执所需的时间长。其中，第二阈值可以根据电能表执行任务时现场的情况确定。
在一实施例中，共性特征矩阵P的第一列p 1为所有电能表任务执行能力的主导特征，在p 1所有元素的均值越与复平面原点(复平面原点是复数域的原点，横坐标代表实部，纵坐标代表虚部，原点是实部虚部都为0的复数)的差值小于第二阈值的情况下，确定所有电能表任务执行能力好，在任务发送数据时得到执行成功回执的比重大，且收到回执所需的时间短，反之，在p 1所有元素的均值与复平面原点的差值大于或等于第二阈值的情况下，确定所有电能表任务执行能力差，在任务发送数据时得到执行成功回执的比重(比例)小，且收到回执所需的时间长。
接着，介绍根据主成分矩阵的第一列所有元素的方差，检测所有待检测电能表共性特征的任务执行能力差异性的步骤，参见图2中的S9。
在一个实施例中，根据主成分矩阵的第一列所有元素的方差，检测所有待检测电能表共性特征的任务执行能力差异性，可以包括：在主成分矩阵的第一列所有元素的方差小于第三阈值的情况下，确定所有待检测电能表共性特征的任务执行能力差异小；在主成分矩阵的第一列所有元素的方差大于或等于第三阈值的情况下，确定所有待检测电能表共性特征的任务执行能力差异大。其中，第三阈值可以根据电能表执行任务时现场的情况确定。
在一实施例中，p 1所有元素的方差指示了所有电能表共性特征的任务执行能力差异性，在p 1所有元素的方差小于第三阈值的情况下，则所有电能表共性特征的任务执行能力差异小，反之，在p 1所有元素的方差大于或等于第三阈值的情况下，则所有电能表共性特征的任务执行能力差异大。
最后，介绍根据截取特征向量矩阵的共轭转置矩阵的第一行元素的模，检测所有待检测电能表的任务执行能力的步骤，参见图2中的S10。
在一个实施例中，根据截取特征向量矩阵的共轭转置矩阵的第一行元素的模，检测所有待检测电能表的任务执行能力，可以包括：在共轭转置矩阵的第一行元素中的任一元素的模大于第四阈值的情况下，确定该任一元素对应的待检测电能表的任务执行能力好；在共轭转置矩阵的第一行元素中的任一元素的模小于或等于第四阈值的情况下，确定该任一元素对应的待检测电能表的任务执行能力差。其中，第四阈值可以根据电能表执行任务时现场的情况确定。
在一实施例中，设U H为截取特征向量矩阵U k的共轭转置矩阵，
Figure PCTCN2019125093-appb-000004
则矩阵U H的第j列为第j个电能表的任务执行情况与所有电能表的共性特征P之间的关系，矩阵U H的第1行为N个电能表分别与主导特征p 1的对应关系，其中，矩阵U H的第1行第j列元素u 1j为第j个电能表分别与主导特征p 1的对应关系，u 1j的模越大，代表第j个电能表的任务执行能力越接近主导特征，该电能表的任务执行能力越好，反之，u 1j的模越小，代表第j个电能表的任务执行能力越偏离主导特征，该电能表的任务执行能力越差。
基于同一发明构思，本申请实施例中还提供了一种电能表任务执行能力检测装置，如下面的实施例。由于电能表任务执行能力检测装置的原理与上述电能表任务执行能力检测方法相似，因此电能表任务执行能力检测装置的实施可以参考上述电能表任务执行能力检测方法的实施，重复之处不再赘述。以下所 使用的，术语“模块”或者“单元”可以实现预定功能的软件和/或硬件的组合。尽管以下实施例所描述的装置可以以软件来实现，但是硬件，或者软件和硬件的组合的实现也是可能并被构想的。
图3是本申请实施例中电能表任务执行能力检测装置的结构示意图，如图3所示，该装置包括：时域扩展采样数据确定单元10，设置为根据待检测电能表的任务执行的类型和时间，对每个待检测电能表的任务执行数据在时域进行扩展，得到每个待检测电能表的时域扩展采样数据；所述时域扩展采样数据保留电能表执行任务的时序性，且指示待检测电能表的任务执行是否成功；原始数据矩阵及协方差矩阵确定单元20，设置为根据所有待检测电能表的时域扩展采样数据，构建原始数据矩阵，根据原始数据矩阵，确定原始数据矩阵对应的协方差矩阵；特征向量矩阵确定单元30，设置为确定所述协方差矩阵的特征值，根据协方差矩阵的特征值，确定协方差矩阵对应的特征向量矩阵；截取特征向量矩阵及主成分矩阵确定单元40，设置为根据主成分分量的数量，从所述特征向量矩阵中选取出预设数量的特征向量组成截取特征向量矩阵；根据所述截取特征向量矩阵和原始数据矩阵，确定主成分矩阵；所述主成分矩阵指示所有待检测电能表任务执行能力的共性特征；检测单元50，设置为根据所述主成分矩阵，对每个待检测电能表的任务执行能力进行检测。
在一个实施例中，上述检测单元设置为按照以下至少之一的步骤，对每个待检测电能表的任务执行能力进行检测：根据主成分分量的数量，检测所有待检测电能表任务执行能力的差异性；根据主成分矩阵的第一列所有元素的均值，检测所有待检测电能表任务执行能力的差异性；根据主成分矩阵的第一列所有元素的方差，检测所有待检测电能表共性特征的任务执行能力差异性；根据截取特征向量矩阵的共轭转置矩阵的第一行元素的模，检测所有待检测电能表的任务执行能力。
在一个实施例中，所述检测单元设置为：在主成分分量的数量小于第一阈值的情况下，确定所有待检测电能表任务执行能力差异小，在主成分分量的数量大于或等于第一阈值的情况下，确定所有待检测电能表任务执行能力差异大；在主成分矩阵的第一列所有元素的均值与复平面原点的差值小于第二阈值的情况下，确定所有待检测电能表任务执行能力好，在任务发送数据时得到执行成功回执的比重大，且收到回执所需的时间短；在主成分矩阵的第一列所有元素的均值与复平面原点的差值大于或等于第二阈值的情况下，确定所有待检测电 能表任务执行能力差，在任务发送数据时得到执行成功回执的比重小，且收到回执所需的时间长；在主成分矩阵的第一列所有元素的方差小于第三阈值的情况下，确定所有待检测电能表共性特征的任务执行能力差异小；在主成分矩阵的第一列所有元素的方差大于或等于第三阈值的情况下，确定所有待检测电能表共性特征的任务执行能力差异大；在共轭转置矩阵的第一行元素中的任一元素的模大于第四阈值的情况下，确定该任一元素对应的待检测电能表的任务执行能力好；在共轭转置矩阵的第一行元素中的任一元素的模小于或等于第四阈值的情况下，确定该任一元素对应的待检测电能表的任务执行能力差。
在一个实施例中，上述电能表任务执行能力检测装置还可以包括：归一化处理单元，设置为对每一待检测电能表的时域扩展采样数据进行归一化处理，得到所有待检测电能表的归一化处理后的时域扩展采样数据；所述原始数据矩阵及协方差矩阵确定单元设置为：根据所有待检测电能表的归一化处理后的时域扩展采样数据，构建原始数据矩阵，根据原始数据矩阵，确定原始数据矩阵对应的协方差矩阵。
综上，传统主成分分析方法的对象是数值而不是布尔量，不能针对任务执行成功与否进行分析和检测，同时传统主成分分析方法处理的样本数据为无时序性的数据，而任务执行能力检测需要考虑任务数据的时序性，因此传统主成分分析方法并不适用于电能表任务执行能力的检测。而本申请方法根据任务执行的时间对任务执行数据在时域进行扩展，在此基础上进行的主成分分析方法的对象是具有时序性的任务执行成功与否的数据，改善了传统主成分分析方法不能进行电能表任务执行能力检测的情况。
本申请实施例还提供了一种计算机设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，所述处理器执行所述电能表任务执行能力检测方法。
本申请实施例还提供了一种计算机可读存储介质，所述计算机可读存储介质存储有执行电能表任务执行能力检测方法的计算机程序。
由以上技术方案可以看出，本实施例基于时域扩展主成分分析的电能表任务执行能力检测方法，其对象是电能表执行任务的数据，即任务执行数据仅为成功或失败而不是一个具体量化的数字，因此本申请方法是一种能处理由任务执行成功或失败组成的布尔量样本数据分析的方法，进而可以检测电能表任务执行能力的问题；同时，这一方法对主成分分析在时域进行扩展，适用于实际 应用环境中电能表执行任务是具有时序性的情况，由于电能表的任务发送是先于任务回执的，而且任务回执是否及时也反映了电能表执行任务的能力强弱。因此，本申请方法可考虑样本数据时序性的分析方法以实现对电能表执行任务能力的检测。
显然，本领域的技术人员应该明白，上述的本申请实施例的各模块或各步骤可以用通用的计算装置来实现，它们可以集中在单个的计算装置上，或者分布在多个计算装置所组成的网络上，例如，它们可以用计算装置可执行的程序代码来实现，从而，可以将它们存储在存储装置中由计算装置来执行，并且在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤，或者将它们分别制作成各个集成电路模块，或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。这样，本申请实施例不限制于任何特定的硬件和软件结合。

Claims (10)

  1. 一种电能表任务执行能力检测方法，包括：
    根据待检测电能表的任务执行的类型和时间，对每个待检测电能表的任务执行数据在时域进行扩展，得到每个待检测电能表的时域扩展采样数据；所述时域扩展采样数据保留电能表执行任务的时序性，且指示待检测电能表的任务执行是否成功；
    根据所有待检测电能表的时域扩展采样数据，构建原始数据矩阵，并根据所述原始数据矩阵，确定所述原始数据矩阵对应的协方差矩阵；
    确定所述协方差矩阵的特征值，根据所述协方差矩阵的特征值，确定所述协方差矩阵对应的特征向量矩阵；
    根据主成分分量的数量，从所述特征向量矩阵中选取出预设数量的特征向量组成截取特征向量矩阵；根据所述截取特征向量矩阵和所述原始数据矩阵，确定主成分矩阵；所述主成分矩阵指示所有待检测电能表的任务执行能力的共性特征；
    根据所述主成分矩阵，对每个待检测电能表的任务执行能力进行检测。
  2. 如权利要求1所述的电能表任务执行能力检测方法，其中，根据所述主成分矩阵，对每个待检测电能表的任务执行能力进行检测，包括以下至少之一的步骤：
    根据主成分分量的数量，检测所有待检测电能表的任务执行能力的差异性；
    根据主成分矩阵的第一列所有元素的均值，检测所有待检测电能表的任务执行能力的差异性；
    根据主成分矩阵的第一列所有元素的方差，检测所有待检测电能表共性特征的任务执行能力的差异性；
    根据截取特征向量矩阵的共轭转置矩阵的第一行元素的模，检测所有待检测电能表的任务执行能力。
  3. 如权利要求2所述的电能表任务执行能力检测方法，其中，
    根据主成分分量的数量，检测所有待检测电能表的任务执行能力的差异性，包括：在主成分分量的数量小于第一阈值的情况下，确定所有待检测电能表的任务执行能力的差异小，在主成分分量的数量大于或等于第一阈值的情况下，确定所有待检测电能表的任务执行能力的差异大；
    根据主成分矩阵的第一列所有元素的均值，检测所有待检测电能表任务执行能力的差异性，包括：在主成分矩阵的第一列所有元素的均值与复平面原点 的差值小于第二阈值的情况下，确定所有待检测电能表任务执行能力好，在任务发送数据时得到执行成功回执的比重大，且收到回执所需的时间短；在主成分矩阵的第一列所有元素的均值与复平面原点的差值大于或等于第二阈值的情况下，确定所有待检测电能表任务执行能力差，在任务发送数据时得到执行成功回执的比重小，且收到回执所需的时间长；
    根据主成分矩阵的第一列所有元素的方差，检测所有待检测电能表共性特征的任务执行能力差异性，包括：在主成分矩阵的第一列所有元素的方差小于第三阈值的情况下，确定所有待检测电能表共性特征的任务执行能力差异小；在主成分矩阵的第一列所有元素的方差大于或等于第三阈值的情况下，确定所有待检测电能表共性特征的任务执行能力差异大；
    根据截取特征向量矩阵的共轭转置矩阵的第一行元素的模，检测所有待检测电能表的任务执行能力，包括：在共轭转置矩阵的第一行元素中的任一元素的模大于第四阈值的情况下，确定所述任一元素对应的待检测电能表的任务执行能力好；在共轭转置矩阵的第一行元素中的任一元素的模小于或等于第四阈值的情况下，确定所述任一元素对应的待检测电能表的任务执行能力差。
  4. 如权利要求1所述的电能表任务执行能力检测方法，还包括：对每个待检测电能表的时域扩展采样数据进行归一化处理，得到所有待检测电能表归一化处理后的时域扩展采样数据；
    根据所有待检测电能表的时域扩展采样数据，构建原始数据矩阵，并根据原始数据矩阵，确定所述原始数据矩阵对应的协方差矩阵，包括：
    根据所有待检测电能表归一化处理后的时域扩展采样数据，构建原始数据矩阵，并根据所述原始数据矩阵，确定所述原始数据矩阵对应的协方差矩阵。
  5. 一种电能表任务执行能力检测装置，包括：
    时域扩展采样数据确定单元，设置为根据待检测电能表的任务执行的类型和时间，对每个待检测电能表的任务执行数据在时域进行扩展，得到每个待检测电能表的时域扩展采样数据；所述时域扩展采样数据保留电能表执行任务的时序性，且指示待检测电能表的任务执行是否成功；
    原始数据矩阵及协方差矩阵确定单元，设置为根据所有待检测电能表的时域扩展采样数据，构建原始数据矩阵，并根据所述原始数据矩阵，确定所述原始数据矩阵对应的协方差矩阵；
    特征向量矩阵确定单元，设置为确定所述协方差矩阵的特征值，根据所述 协方差矩阵的特征值，确定所述协方差矩阵对应的特征向量矩阵；
    截取特征向量矩阵及主成分矩阵确定单元，设置为根据主成分分量的数量，从所述特征向量矩阵中选取出预设数量的特征向量组成截取特征向量矩阵；根据所述截取特征向量矩阵和所述原始数据矩阵，确定主成分矩阵；所述主成分矩阵指示所有待检测电能表的任务执行能力的共性特征；
    检测单元，设置为根据所述主成分矩阵，对每个待检测电能表的任务执行能力进行检测。
  6. 如权利要求5所述的电能表任务执行能力检测装置，其中，所述检测单元设置为按照以下至少之一的步骤，对每个待检测电能表的任务执行能力进行检测：
    根据主成分分量的数量，检测所有待检测电能表的任务执行能力的差异性；
    根据主成分矩阵的第一列所有元素的均值，检测所有待检测电能表的任务执行能力的差异性；
    根据主成分矩阵的第一列所有元素的方差，检测所有待检测电能表共性特征的任务执行能力的差异性；
    根据截取特征向量矩阵的共轭转置矩阵的第一行元素的模，检测所有待检测电能表的任务执行能力。
  7. 如权利要求6所述的电能表任务执行能力检测装置，其中，所述检测单元设置为：
    在主成分分量的数量小于第一阈值的情况下，确定所有待检测电能表的任务执行能力的差异小，在主成分分量的数量大于或等于第一阈值的情况下，确定所有待检测电能表的任务执行能力的差异大；
    在主成分矩阵的第一列所有元素的均值与复平面原点的差值小于第二阈值的情况下，确定所有待检测电能表任务执行能力好，在任务发送数据时得到执行成功回执的比重大，且收到回执所需的时间短；在主成分矩阵的第一列所有元素的均值与复平面原点的差值大于或等于第二阈值的情况下，确定所有待检测电能表任务执行能力差，在任务发送数据时得到执行成功回执的比重小，且收到回执所需的时间长；
    在主成分矩阵的第一列所有元素的方差小于第三阈值的情况下，确定所有待检测电能表共性特征的任务执行能力差异小；在主成分矩阵的第一列所有元素的方差大于或等于第三阈值的情况下，确定所有待检测电能表共性特征的任 务执行能力差异大；
    在共轭转置矩阵的第一行元素中的任一元素的模大于第四阈值的情况下，确定所述任一元素对应的待检测电能表的任务执行能力好；在共轭转置矩阵的第一行元素中的任一元素的模小于或等于第四阈值的情况下，确定所述任一元素对应的待检测电能表的任务执行能力差。
  8. 如权利要求5所述的电能表任务执行能力检测装置，还包括：归一化处理单元，设置为对每个待检测电能表的时域扩展采样数据进行归一化处理，得到所有待检测电能表归一化处理后的时域扩展采样数据；
    所述原始数据矩阵及协方差矩阵确定单元设置为：根据所有待检测电能表的归一化处理后的时域扩展采样数据，构建原始数据矩阵，并根据所述原始数据矩阵，确定所述原始数据矩阵对应的协方差矩阵。
  9. 一种计算机设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现权利要求1至4中任一项所述的方法。
  10. 一种计算机可读存储介质，所述计算机可读存储介质存储有执行权利要求1至4中任一项所述方法的计算机程序。
PCT/CN2019/125093 2019-04-23 2019-12-13 电能表任务执行能力检测方法及装置 WO2020215748A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910329419.4A CN110031790A (zh) 2019-04-23 2019-04-23 电能表任务执行能力检测方法及装置
CN201910329419.4 2019-04-23

Publications (1)

Publication Number Publication Date
WO2020215748A1 true WO2020215748A1 (zh) 2020-10-29

Family

ID=67239869

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/125093 WO2020215748A1 (zh) 2019-04-23 2019-12-13 电能表任务执行能力检测方法及装置

Country Status (2)

Country Link
CN (1) CN110031790A (zh)
WO (1) WO2020215748A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110031790A (zh) * 2019-04-23 2019-07-19 华北电力科学研究院有限责任公司 电能表任务执行能力检测方法及装置
CN110502518A (zh) * 2019-08-26 2019-11-26 国网重庆市电力公司电力科学研究院 一种智能电能表时序数据处理方法和处理装置
CN110488218B (zh) * 2019-08-26 2021-11-30 国网重庆市电力公司电力科学研究院 一种电能表运行状态评估方法和评估装置

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1578968A (zh) * 2001-08-29 2005-02-09 霍尼韦尔国际公司 用于受监督的神经网络学习的组合方法
CN101788656A (zh) * 2010-01-29 2010-07-28 东南大学 功能核磁共振扫描下功能响应信号的甄别方法
CN109307854A (zh) * 2018-10-29 2019-02-05 中国电力科学研究院有限公司 一种电能计量装置综合评估方法及系统
US20190066021A1 (en) * 2017-08-28 2019-02-28 Clari Inc. Method and system for summarizing user activities of tasks into a single activity score using machine learning to predict probabilities of completeness of the tasks
CN109447394A (zh) * 2018-09-14 2019-03-08 中国科学院空间应用工程与技术中心 一种对地观测卫星任务效能评价指标体系及分析方法
CN110031790A (zh) * 2019-04-23 2019-07-19 华北电力科学研究院有限责任公司 电能表任务执行能力检测方法及装置
CN209961899U (zh) * 2019-04-23 2020-01-17 华北电力科学研究院有限责任公司 电能表任务执行能力检测装置

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1578968A (zh) * 2001-08-29 2005-02-09 霍尼韦尔国际公司 用于受监督的神经网络学习的组合方法
CN101788656A (zh) * 2010-01-29 2010-07-28 东南大学 功能核磁共振扫描下功能响应信号的甄别方法
US20190066021A1 (en) * 2017-08-28 2019-02-28 Clari Inc. Method and system for summarizing user activities of tasks into a single activity score using machine learning to predict probabilities of completeness of the tasks
CN109447394A (zh) * 2018-09-14 2019-03-08 中国科学院空间应用工程与技术中心 一种对地观测卫星任务效能评价指标体系及分析方法
CN109307854A (zh) * 2018-10-29 2019-02-05 中国电力科学研究院有限公司 一种电能计量装置综合评估方法及系统
CN110031790A (zh) * 2019-04-23 2019-07-19 华北电力科学研究院有限责任公司 电能表任务执行能力检测方法及装置
CN209961899U (zh) * 2019-04-23 2020-01-17 华北电力科学研究院有限责任公司 电能表任务执行能力检测装置

Also Published As

Publication number Publication date
CN110031790A (zh) 2019-07-19

Similar Documents

Publication Publication Date Title
WO2020215748A1 (zh) 电能表任务执行能力检测方法及装置
Charrad et al. NbClust: an R package for determining the relevant number of clusters in a data set
Kim The relation among fit indexes, power, and sample size in structural equation modeling
West et al. Model fit and model selection in structural equation modeling
Chen et al. Detecting big structural breaks in large factor models
Pothisiri et al. Structural damage detection and assessment from modal response
An et al. Are the macroeconomic effects of oil price shock symmetric?: A Factor-Augmented Vector Autoregressive approach
Nielsen Nonparametric cointegration analysis of fractional systems with unknown integration orders
Zhang et al. Efficient factor garch models and factor-dcc models
Korovaitsev et al. Evaluation of metrological reliability of measuring instruments under the conditions of incomplete data
Dette et al. Detecting structural breaks in eigensystems of functional time series
Song et al. An improved sufficient dimension reduction-based Kriging modeling method for high-dimensional evaluation-expensive problems
CN105242236B (zh) 宽带信号超分辨测向中的阵元位置误差校正方法
Jun et al. Tracy-Widom limit for the largest eigenvalue of high-dimensional covariance matrices in elliptical distributions
US20220166702A1 (en) Fuzzing preprocessing apparatus and method for automating smart network fuzzing
Ellingson et al. Nonparametric bootstrap of sample means of positive-definite matrices with an application to diffusion-tensor-imaging data analysis
Lee et al. Robust distribution-free multivariate CUSUM charts for spatiotemporal biosurveillance in the presence of spatial correlation
CN209961899U (zh) 电能表任务执行能力检测装置
CN111027190A (zh) 一种用于设备模型数值相似度的评估方法及装置
Hwang et al. Estimating the number of true null hypotheses in multiple hypothesis testing
Ozturk et al. A new method for assessing multivariate normality with graphical applications
Wang et al. A necessary test for complete independence in high dimensions using rank-correlations
CN107506824B (zh) 一种配电网的不良观测数据检测方法及装置
CN110738248A (zh) 状态感知数据特征提取方法及装置、系统性能评估方法
Glimm et al. Multivariate tests of normal mean vectors with restricted alternatives

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19926160

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19926160

Country of ref document: EP

Kind code of ref document: A1