Power supply fault detection method for intelligent electric energy meter
Patent Information
- Application Number
- CN202411054145.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-02
- Publication Date
- 2024-11-19
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
[0003]随着科学的发展和技术的进步,现代工业过程的规模和复杂程度都在日益提高,在对供电计量电表使用过程中,电表都会存在缓慢的变化过程,比如老化、催化剂性能的缓慢变化等,而且这些过程潜在缓变、故障,而且这些故障如果不能得到及时检测的话,会导致产品质量的下降,严重时可能发生事故
[0045] 1. A power supply fault detection method of a smart electric energy meter of the present invention aims at extracting the changing characteristics of the fault process data for minor and slowly changing faults that occur during the use of the electric energy meter, and constructing the statistics of the corresponding electric meter operating data and its statistical values. , giving the running trajectories of data variables at different times, based on which the automatic detection of power supply meter faults is completed, and the fault detection accuracy is high.
Smart Images

Figure CN118980985A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of smart electric energy meters, and in particular to a power supply fault detection method for a smart electric energy meter. Background Art
[0002] Smart meter is a new type of intelligent power metering equipment that realizes power metering and information collection through digital technology. It can not only measure electricity, but also measure multiple parameters such as power factor, active / reactive power, voltage, current, etc., effectively improving the accuracy and precision of power metering, and can also realize functions such as remote meter reading and remote control, providing a more effective means for the operation and management of the power grid. Smart meters require external power supply to use. It can be said that the power supply system is the life of smart meters. Therefore, the timely discovery of power supply failures of smart meters is also the basis for ensuring the reliable operation of the power system.
[0003] With the development of science and the advancement of technology, the scale and complexity of modern industrial processes are increasing day by day. In the process of using power supply metering meters, the meters will have slow changes, such as aging, slow changes in catalyst performance, etc., and these processes are potentially slow changes and failures. If these failures cannot be detected in time, it will lead to a decline in product quality, and accidents may occur in serious cases. Compared with most fault detection methods, it is difficult to extract meter fault features, which makes the implementation of power supply metering meter fault detection fall into a bottleneck. For this reason, the present invention proposes a power supply fault detection method for a smart electric energy meter. Summary of the invention
[0004] In view of the deficiencies in the prior art, an object of the present invention is to provide a power supply fault detection method for a smart electric energy meter to solve the above technical problems.
[0005] The above technical objectives of the present invention are achieved through the following technical solutions:
[0006] A method for detecting power supply failure of a smart electric energy meter comprises the following steps:
[0007] S1. Obtain meter measurement data and provide an independent metadata matrix under normal conditions of the meter;
[0008] S2. Based on the independent metadata matrix of the electric meter in a normal state, an electric meter fault detection model is established;
[0009] S3. Build a multi-parameter electric meter fault automatic detection experimental platform in the matlab environment to simulate the electric meter fault detection model;
[0010] S4. The staff actually inspects the power supply failure of the electric meter and compares and verifies it with the simulation data to optimize the electric meter fault detection model.
[0011] By adopting the above technical scheme, for small and slow-changing faults that occur during the use of the electric energy meter, the changing characteristics of the fault process data are extracted, the statistics of the corresponding meter operation data and their statistical values are constructed, and the operation trajectory of the data variables at different times is given. Based on this, the automatic detection of power supply meter faults is completed with high fault detection accuracy.
[0012] In a preferred example, the present invention can be further configured as follows: in the step S1, the meter measurement data is obtained, and an independent metadata matrix of the meter in a normal state is given. The specific method is: a normal meter measurement data matrix is given by using a kernel independent element method, and the matrix is mapped to a high-dimensional space to obtain a covariance matrix of the meter measurement data, a statistical monitoring master of the meter measurement data is established, and a normal independent metadata matrix of the meter is given;
[0013] The kernel independent element method is a nonlinear extension of kernel principal component analysis and independent component analysis. The basic idea is to nonlinearly map the data to a feature space with a better linear structure, and then use ICA to extract nonlinear independent elements in the feature space, which provides a new idea for fault diagnosis of multiple instruments in industrial production processes.
[0014] In a preferred example, the present invention can be further configured as follows: the independent metadata matrix of the electric meter in a normal state is provided, as described in detail as follows:
[0015] Assume that X is the meter measurement data under normal conditions, N represents the number of samples, and s represents the element value of the kernel function. Then the calculation method of X is:
[0016]
[0017] in, represents a nonlinear function;
[0018] application Map X to Representing a high-dimensional space, the covariance matrix A between normal data types is calculated by the following formula:
[0019]
[0020] Among them, x' tpo Represents an independent metadata matrix;
[0021] The meter measurement data statistics monitoring dominant z is established by the following formula:
[0022]
[0023] Where g represents the dominant independent data source of meter measurement, we represents the weight space of the normal independent metadata matrix, ω represents the number of iterations, l represents the whitened data matrix, and x' opiu represents a Gaussian probability density function with a variance of 1;
[0024] The normal independent metadata matrix Z of the electric meter is given by the following formula:
[0025]
[0026] in, is the Gaussian intensity distinction threshold, x' hkp is the principal component space for detecting the weak non-Gaussian part, m is the residual non-Gaussian information in the current whitened data, E represents the residual space of the weak non-Gaussian part, c uip represents the high-order statistical information of the meter data variable, and e represents the residual space of the non-Gaussian part.
[0027] By adopting the above technical scheme, the kernel independent element method is used to give the normal meter measurement data matrix, and the matrix is mapped to a high-dimensional space to obtain the covariance matrix of the meter measurement data. The meter measurement data statistics monitoring dominant is established, and the independent metadata matrix of the normal meter measurement is given, which lays the foundation for realizing the optimized detection of power supply meter faults.
[0028] In a preferred example, the present invention can be further configured as follows: in the step S2, an electric meter fault detection model is established based on the independent metadata matrix of the electric meter in a normal state. The specific method is: select key variables of the electric energy meter measurement data, calculate the average trajectory of the key variables in the entire historical operating period, extract the key variable information in the real-time online sample, give the operating trajectory of the variables of the data at different times, and finally complete the optimized detection of power supply meter faults.
[0029] In a preferred example, the present invention can be further configured as follows: the key variables of the electric energy meter measurement data are selected by the following specific method:
[0030] Let δ represent the total historical sampling samples, λ represent the total number of variables, Represents the number of historical data samples, then the key variable E of the meter measurement data is selected as shown in the following formula:
[0031]
[0032] Among them, b represents the value obtained after differential calculation, and p represents the label of the current data variable.
[0033] In a preferred example, the present invention can be further configured as follows: a multi-parameter electric meter fault automatic detection experimental platform is built under a matlab environment to simulate an electric meter fault detection model, and the specific method is as follows:
[0034] Select a suitable smart energy meter, collect historical power supply fault data during the last six months of its operation, import the historical power supply data into the simulation software, determine whether the smart meter will issue an alarm, and then use the accuracy index to measure the early warning performance of the smart meter communication fault area.
[0035] In a preferred example, the present invention can be further configured as follows: the staff actually inspects the power supply failure of the electric meter and compares and verifies it with the simulation data, thereby optimizing the electric meter fault detection model. The specific method is:
[0036] The staff organized the inaccurately predicted data into an error set, reduced the calculation weight of the data in the error set in the meter fault detection model, and then imported new data to train the meter fault detection model and optimized the parameters of the meter fault detection model until the prediction accuracy of the meter fault detection model reached more than 95%;
[0037] When staff find a smart meter with a high probability of power supply failure, they need to go to the site for inspection and maintenance. If they find that a power supply failure has indeed occurred or that there is a problem with the power supply system that will cause the power supply failure, the prediction result is considered accurate. If no power supply problem is found on site, the prediction result is considered inaccurate, and feedback data is filled in on the details page of the smart meter list based on the actual inspection situation.
[0038] In a preferred example, the present invention can be further configured as follows: after step S4, the optimized electric meter fault detection model is verified, and the specific method is as follows:
[0039] In the optimized meter fault detection model, 10 sets of data are used as input. The 10 sets of data include 5 sets of normal data and 5 sets of fault data. Each set of data includes 1000 sample points. The sampling period is 3 minutes. The fault is introduced at the 200th sample point.
[0040] The detection experiment of multi-parameter electric meter fault detection model is carried out based on multivariable trajectory analysis method, and the missed detection rate and false alarm rate of the electric meter fault detection model are obtained;
[0041] The working principle of smart meter is to collect the user's power supply voltage and current in real time, and then process the collected voltage and current through integrated circuits, and convert them into pulses proportional to electric energy. Finally, the single chip microcomputer completes the processing and control of the pulses to realize the display and output of power consumption. Therefore, the DTZY876 model smart meter is selected as the experimental component.
[0042] In a preferred example, the present invention can be further configured as follows: in step S1, the meter measurement data is obtained, and the specific method is: connecting the acquisition line in the power supply circuit of the smart meter. There are two power supply modes for smart meters, one is directly connected to the AC power supply for power supply, and the other is powered by a battery. The two power supply modes determine that the power supply data collection method is also different. For the power supply power meter, the power line, that is, the phase line, needs to be connected to the acquisition line; for the battery-powered meter, it is a passive circuit, and the resistors, capacitors, inductors and other components need to be connected in series one by one in a certain order using the acquisition line.
[0043] Then set the collection configuration information and use the collection line to remotely obtain the meter power supply data. The collection line is connected to the collection repeater and forwarded to the remote server by the router. The remote server needs to set the collection configuration information according to the power supply method of the smart meter, including: collection address, check code, wiring scheme, etc. After the configuration is completed, the power supply data collected by the collection line is remotely received and stored.
[0044] In summary, the present invention includes at least one of the following beneficial technical effects:
[0045] 1. The power supply fault detection method of a smart electric energy meter of the present invention extracts the change characteristics of the fault process data for small and slow-changing faults that occur during the use of the electric energy meter, constructs the statistics of the corresponding electric meter operation data and its statistical values, and gives the operation trajectory of the data variables at different times. Based on this, the automatic detection of power supply meter faults is completed, and the fault detection accuracy is high.
[0046] 2. The power supply fault detection method of a smart electric energy meter of the present invention uses the kernel independent element method to give a normal electric meter measurement data matrix, and maps the matrix to a high-dimensional space to obtain the covariance matrix of the electric meter measurement data, establishes the statistical monitoring of the electric meter measurement data, and gives an independent metadata matrix of the normal measurement of the electric meter, laying a foundation for realizing the optimized detection of power supply meter faults. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0048] Figure 1 The present invention is a flow chart of a method for detecting power supply failure of a smart electric energy meter. DETAILED DESCRIPTION
[0049] The present invention is further described in detail below in conjunction with the accompanying drawings.
[0050] Example:
[0051] Reference Figure 1 The present invention discloses a method for detecting power supply faults of intelligent electric energy meters, the principle of which is as follows: in the process of automatic detection of minor faults of multi-parameter high-precision electric energy meters, an observation vector that can reflect the current operating state of the electric energy meter is given, a statistical monitoring model of the electric energy meter operating data is established, the variable values of the multivariate exponentially weighted moving average of the electric energy meter operating at different times are calculated, the projection matrix of the electric energy meter data on the principal element space is calculated, and the minor fault detection features of the multi-parameter high-precision electric energy meter are extracted, and the minor fault detection features of the multi-parameter high-precision electric energy meter are automatically detected based on this.
[0052] The specific method includes the following steps:
[0053] S1. Obtain meter measurement data and provide an independent metadata matrix under normal conditions of the meter;
[0054] S2. Based on the independent metadata matrix of the electric meter in a normal state, an electric meter fault detection model is established;
[0055] S3. Build a multi-parameter electric meter fault automatic detection experimental platform in the matlab environment to simulate the electric meter fault detection model;
[0056] S4. The staff actually inspects the power supply failure of the electric meter and compares and verifies it with the simulation data to optimize the electric meter fault detection model.
[0057] In the step S1, the meter measurement data is obtained, and an independent metadata matrix of the meter in a normal state is given. The specific method is: a normal meter measurement data matrix is given by using a kernel independent element method, and the matrix is mapped to a high-dimensional space to obtain a covariance matrix of the meter measurement data, a statistical monitoring matrix of the meter measurement data is established, and a normal independent metadata matrix of the meter is given.
[0058] The independent metadata matrix of the electric meter under normal condition is described as follows:
[0059] Assume that X is the meter measurement data under normal conditions, N represents the number of samples, and s represents the element value of the kernel function. Then the calculation method of X is:
[0060]
[0061] in, represents a nonlinear function;
[0062] application Map X to Representing a high-dimensional space, the covariance matrix A between normal data types is calculated by the following formula:
[0063]
[0064] Among them, x' tpo Represents an independent metadata matrix;
[0065] The meter measurement data statistics monitoring dominant z is established by the following formula:
[0066]
[0067] Where g represents the dominant independent data source of meter measurement, we represents the weight space of the normal independent metadata matrix, ω represents the number of iterations, l represents the whitened data matrix, and x' opiu represents a Gaussian probability density function with a variance of 1;
[0068] The normal independent metadata matrix Z of the electric meter is given by the following formula:
[0069]
[0070] in, is the Gaussian intensity distinction threshold, x' hkp is the principal component space for detecting the weak non-Gaussian part, m is the residual non-Gaussian information in the current whitened data, E represents the residual space of the weak non-Gaussian part, c uip represents the high-order statistical information of the meter data variable, and e represents the residual space of the non-Gaussian part.
[0071] The above method uses the kernel independent element method to give the normal meter measurement data matrix, and maps the matrix to a high-dimensional space to obtain the covariance matrix of the meter measurement data, establishes the meter measurement data statistics monitoring leader, and gives the independent metadata matrix of the normal meter measurement, laying the foundation for realizing the optimized detection of power supply meter faults.
[0072] The kernel independent element method is a nonlinear extension of kernel principal component analysis and independent component analysis. The basic idea is to nonlinearly map the data to a feature space with a better linear structure, and then use ICA to extract nonlinear independent elements in the feature space, which provides a new idea for fault diagnosis of multiple instruments in industrial production processes.
[0073] In the step S2, an electric meter fault detection model is established based on the independent metadata matrix of the electric meter in a normal state. The specific method is: select key variables of the electric energy meter measurement data, calculate the average trajectory of the key variables in the entire historical operating period, extract the key variable information in the real-time online sample, give the operating trajectory of the variables of the data at different times, and finally complete the optimized detection of the power supply meter fault.
[0074] The specific method for selecting the key variables of the electric energy meter measurement data is as follows:
[0075] Let δ represent the total historical sampling samples, λ represent the total number of variables, Represents the number of historical data samples, then the key variable E of the meter measurement data is selected as shown in the following formula:
[0076]
[0077] Among them, b represents the value obtained after differential calculation, and p represents the label of the current data variable.
[0078] The multi-parameter electric meter fault automatic detection experimental platform is built in the matlab environment to simulate the electric meter fault detection model. The specific method is:
[0079] Select a suitable smart energy meter, collect historical power supply fault data during the last six months of its operation, import the historical power supply data into the simulation software, determine whether the smart meter will issue an alarm, and then use the accuracy index to measure the early warning performance of the smart meter communication fault area.
[0080] The staff actually inspected the power supply failure of the electric meter and compared and verified it with the simulation data, so as to optimize the electric meter fault detection model. The specific method is as follows:
[0081] The staff organized the inaccurately predicted data into an error set, reduced the calculation weight of the data in the error set in the meter fault detection model, and then imported new data to train the meter fault detection model and optimized the parameters of the meter fault detection model until the prediction accuracy of the meter fault detection model reached more than 95%.
[0082] When staff find a smart meter with a high probability of power supply failure, they need to go to the site for inspection and maintenance. If they find that a power supply failure has indeed occurred or that there is a problem with the power supply system that will cause the power supply failure, the prediction result is considered accurate. If no power supply problem is found on site, the prediction result is considered inaccurate, and feedback data is filled in on the details page of the smart meter list based on the actual inspection situation.
[0083] After step S4, the optimized electric meter fault detection model is verified, and the specific method is as follows:
[0084] In the optimized meter fault detection model, 10 sets of data are used as input. The 10 sets of data include 5 sets of normal data and 5 sets of fault data. Each set of data includes 1000 sample points. The sampling period is 3 minutes. The fault is introduced at the 200th sample point.
[0085] The detection experiment of multi-parameter electric meter fault detection model was carried out based on multivariable trajectory analysis method, and the missed detection rate and false alarm rate of the electric meter fault detection model were obtained.
[0086] Specifically, the working principle of the smart meter is to collect the user's power supply voltage and current in real time, and then process the collected voltage and current through the integrated circuit and convert them into pulses proportional to the electric energy. Finally, the single-chip microcomputer completes the processing and control of the pulses to realize the display and output of power consumption. Therefore, this embodiment selects the DTZY876 model smart meter as the experimental element.
[0087] This embodiment extracts the changing characteristics of the fault process data for small and slow-changing faults that occur during the use of the electric energy meter, constructs the statistics of the corresponding meter operation data and their statistical values, and gives the operation trajectory of the data variables at different times. Based on this, automatic detection of power supply meter faults is completed with high fault detection accuracy.
[0088] Specifically, the electric meter measurement data in step S1 may be obtained by using the method described in the patent publication number CN118068254A, namely:
[0089] Connect the collection line to the power supply circuit of the smart meter. There are two power supply modes for smart meters: one is to directly connect to the AC power supply for power supply, and the other is to power supply through batteries. The two power supply modes determine different methods of collecting power supply data. For power supply-powered meters, the collection line needs to be connected to the power line, that is, the phase line; for battery-powered meters, which are passive circuits, the resistors, capacitors, inductors and other components need to be connected in series one by one in a certain order using the collection line. Then set the collection configuration information and use the collection line to remotely obtain the meter power supply data. The collection line is connected to the collection repeater and forwarded to the remote server by the router. The remote server needs to set the collection configuration information according to the power supply mode of the smart meter, including: collection address, check code, wiring scheme, etc. After the configuration is completed, the power supply data collected by the collection line is remotely received and stored.
[0090] The implementation principle of the above embodiment is:
[0091] In the process of automatic detection of minor faults of multi-parameter high-precision electric meters, an observation vector that can reflect the current operating status of the electric meter is given, a statistical monitoring model of the electric meter operation data is established, the variable values of the multivariate exponentially weighted moving average of the electric meter operation at different times are calculated, the projection matrix of the electric meter data on the principal element space is calculated, and the minor fault detection features of the multi-parameter high-precision electric meter are extracted. Based on this, the automatic detection of minor faults of the multi-parameter high-precision electric meter is completed, which specifically includes the following steps:
[0092] S1. Obtain the meter measurement data and give the independent metadata matrix of the meter under normal conditions. Use the kernel independent element method to give the normal meter measurement data matrix, and map the matrix to a high-dimensional space to obtain the covariance matrix of the meter measurement data, establish the meter measurement data statistics monitoring master, and give the normal independent metadata matrix of the meter.
[0093] S2. Based on the independent metadata matrix of the electric meter in normal state, establish the electric meter fault detection model. Select the key variables of the electric meter measurement data, calculate the average trajectory of the key variables in the entire historical operation period, extract the key variable information in the real-time online sample, give the operation trajectory of the variables of the data at different times, and finally complete the optimization detection of the power supply meter fault.
[0094] S3. Build a multi-parameter electric meter fault automatic detection experimental platform in the matlab environment to simulate the electric meter fault detection model;
[0095] S4. The staff actually inspects the power supply failure of the electric meter and compares and verifies it with the simulation data to optimize the electric meter fault detection model.
[0096] The present invention aims at small and slow-changing faults occurring during the use of electric energy meters, extracts the changing characteristics of fault process data, constructs the statistics of corresponding meter operation data and its statistical limits, and gives the operation trajectories of data variables at different times. Based on this, automatic fault detection of power supply meter is completed with high fault detection accuracy.
[0097] The above description is combined with the accompanying drawings to fully illustrate the specific embodiments of this article so that those skilled in the art can practice them. Parts and features of some embodiments may be included in or replace parts and features of other embodiments. The scope of the embodiments of the present invention includes the entire scope of the claims, and all available equivalents of the claims. In the present invention, the terms "first", "second", etc. are only used to distinguish one element from another, without requiring or implying any actual relationship or order between these elements. In fact, the first element can also be called the second element, and vice versa. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that the structure, device or equipment including a series of elements includes not only those elements, but also other elements that are not explicitly listed, or also include elements inherent to such structure, device or equipment. In the absence of more restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the structure, device or equipment including the elements. Each embodiment is described in a progressive manner herein, and each embodiment focuses on the differences from other embodiments, and the same and similar parts between the embodiments can be referred to each other.
[0098] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims
1. A method for detecting power supply failure of a smart electric energy meter, characterized in that: The following steps are involved: S1. Obtain meter measurement data and provide an independent metadata matrix under normal conditions of the meter; S2. Based on the independent metadata matrix of the electric meter in a normal state, an electric meter fault detection model is established; S3. Build a multi-parameter electric meter fault automatic detection experimental platform in the matlab environment to simulate the electric meter fault detection model; S4. The staff actually inspects the power supply failure of the electric meter and compares and verifies it with the simulation data to optimize the electric meter fault detection model.
2. The method for detecting power supply failure of a smart electric energy meter according to claim 1, characterized in that: In the step S1, the meter measurement data is obtained, and an independent metadata matrix of the meter in a normal state is given. The specific method is: a normal meter measurement data matrix is given by using a kernel independent element method, and the matrix is mapped to a high-dimensional space to obtain a covariance matrix of the meter measurement data, a statistical monitoring matrix of the meter measurement data is established, and a normal independent metadata matrix of the meter is given.
3. The method for detecting power supply failure of a smart electric energy meter according to claim 2, characterized in that: The independent metadata matrix of the electric meter under normal condition is described as follows: Assume that X is the meter measurement data under normal conditions, N represents the number of samples, and s represents the element value of the kernel function. Then the calculation method of X is: in, represents a nonlinear function; application Map X to Representing a high-dimensional space, the covariance matrix A between normal data types is calculated by the following formula: Among them, x' tpo Represents an independent metadata matrix; The meter measurement data statistics monitoring dominant z is established by the following formula: Where g represents the dominant independent data source of meter measurement, we represents the weight space of the normal independent metadata matrix, ω represents the number of iterations, l represents the whitened data matrix, and x' opiu represents a Gaussian probability density function with a variance of 1; The normal independent metadata matrix Z of the electric meter is given by the following formula: in, is the Gaussian intensity distinction threshold, x' hkp is the principal component space for detecting the weak non-Gaussian part, m is the residual non-Gaussian information in the current whitened data, E represents the residual space of the weak non-Gaussian part, c uip represents the high-order statistical information of the meter data variable, and e represents the residual space of the non-Gaussian part.
4. The method for detecting power supply failure of a smart electric energy meter according to claim 3, characterized in that: In the step S2, an electric meter fault detection model is established based on the independent metadata matrix of the electric meter in a normal state. The specific method is: select key variables of the electric energy meter measurement data, calculate the average trajectory of the key variables in the entire historical operating period, extract the key variable information in the real-time online sample, give the operating trajectory of the variables of the data at different times, and finally complete the optimized detection of the power supply meter fault.
5. The method for detecting power supply failure of a smart electric energy meter according to claim 4, characterized in that: The specific method for selecting the key variables of the electric energy meter measurement data is as follows: Let δ represent the total historical sampling samples, λ represent the total number of variables, Represents the number of historical data samples, then the key variable E of the meter measurement data is selected as shown in the following formula: Among them, b represents the value obtained after differential calculation, and p represents the label of the current data variable.
6. The method for detecting power supply failure of a smart electric energy meter according to claim 5, characterized in that: The multi-parameter electric meter fault automatic detection experimental platform is built in the matlab environment to simulate the electric meter fault detection model. The specific method is: Select a suitable smart energy meter, collect historical power supply fault data during the last six months of its operation, import the historical power supply data into the simulation software, determine whether the smart meter will issue an alarm, and then use the accuracy index to measure the early warning performance of the smart meter communication fault area.
7. The method for detecting power supply failure of a smart electric energy meter according to claim 6, characterized in that: The staff actually inspected the power supply failure of the electric meter and compared and verified it with the simulation data, so as to optimize the electric meter fault detection model. The specific method is as follows: The staff organized the inaccurately predicted data into an error set, reduced the calculation weight of the data in the error set in the meter fault detection model, and then imported new data to train the meter fault detection model and optimized the parameters of the meter fault detection model until the prediction accuracy of the meter fault detection model reached more than 95%.
8. The method for detecting power supply failure of a smart electric energy meter according to claim 7, characterized in that: After step S4, the optimized electric meter fault detection model is verified, and the specific method is as follows: In the optimized meter fault detection model, 10 sets of data are used as input. The 10 sets of data include 5 sets of normal data and 5 sets of fault data. Each set of data includes 1000 sample points. The sampling period is 3 minutes. The fault is introduced at the 200th sample point. The detection experiment of multi-parameter electric meter fault detection model was carried out based on multivariable trajectory analysis method, and the missed detection rate and false alarm rate of the electric meter fault detection model were obtained.
Citation Information
Patent Citations
Main and auxiliary integrated fault locating method and system based on gradient descent algorithm
CN109782124A
Human body behavior recognition method based on scene priori knowledge
CN112560668A
Power supply fault detection method for intelligent electric energy meter
CN118068254A
