阅读说明：本技术 一种单相电能表电能质量分析的方法及单相电能表 (Method for analyzing electric energy quality of single-phase electric energy meter and single-phase electric energy meter ) 是由 杨舟 蒋雯倩 李刚 陈珏羽 周毅波 周政雷 江革力 于 2019-07-29 设计创作，主要内容包括：本发明实施例公开了一种单相电能表电能质量分析的方法及单相电能表,其方法包括：电能表在上电后,电能表中电能计量芯片的模数转换器输出波形数据,所述输出波形数据包括：电压波形数据和电流波形数据；电能表管理模块基于所述波形数据计算出每周波梯度,并对所述每周波梯度进行录波形成录波数据；基于录波数据绘制成电压-电流曲线图；将电压-电流曲线图变换为有权重的像素化图像；采用训练好的卷积神经网络将所述像素化图像作为输入完成波形检测输出。实施本发明实施例具有较高的识别准确度,能识别出不同类型的电能质量故障类型。(The embodiment of the invention discloses a method for analyzing the electric energy quality of a single-phase electric energy meter and the single-phase electric energy meter, wherein the method comprises the following steps: after the electric energy meter is powered on, an analog-to-digital converter of an electric energy metering chip in the electric energy meter outputs waveform data, wherein the output waveform data comprises: voltage waveform data and current waveform data; the electric energy meter management module calculates the gradient of each cycle based on the waveform data and records the gradient of each cycle to form recording data; drawing a voltage-current curve graph based on the wave recording data; transforming the voltage-current graph into a weighted pixelated image; and adopting a trained convolutional neural network to take the pixilated image as input to finish waveform detection and output. The embodiment of the invention has higher identification accuracy and can identify different types of power quality fault types.)

1. A method for analyzing the electric energy quality of a single-phase electric energy meter is characterized by comprising the following steps:

After the electric energy meter is powered on, an analog-to-digital converter of an electric energy metering chip in the electric energy meter outputs waveform data, wherein the output waveform data comprises: voltage waveform data and current waveform data;

the electric energy meter management module calculates the gradient of each cycle based on the waveform data and records the gradient of each cycle to form recording data;

Drawing a voltage-current curve graph based on the wave recording data;

Transforming the voltage-current graph into a weighted pixelated image;

and adopting a trained convolutional neural network to take the pixilated image as input to finish waveform detection and output.

2. The method for analyzing the electric energy quality of the single-phase electric energy meter according to claim 1, wherein the electric energy meter management module calculates a weekly wave gradient based on the waveform data, and recording the weekly wave gradient to form recording data comprises:

Calculating the waveform gradient on each cycle on the waveform data, wherein the waveform gradient is the waveform change rate of the current cycle sampling point and the previous waveform sampling point;

And identifying the condition of the maximum waveform gradient change, and recording the first four cycles and the last four cycles of the maximum waveform gradient change.

3. The method for analyzing the electric energy quality of the single-phase electric energy meter according to claim 2, wherein the step of plotting a voltage-current curve graph based on the recording data comprises the following steps:

Based on the waveform of the recording wave, the voltage is taken as the vertical axis, the current is taken as the horizontal axis, and a voltage-current curve is drawn.

4. the method for the electric energy quality analysis of the single-phase electric energy meter according to claim 3, wherein the transforming the voltage-current graph into the weighted pixilated image comprises:

converting the voltage-current curve waveform into a weighted pixelized image, and dividing the image into n if the number of pixels on the horizontal axis and the vertical axis is respectively n²A plurality of pixels;

to n²each pixel of the pixels is provided with different weight, the maximum value of the weight is [ (n +1)²-1]The minimum weight is 0, and the weight is assigned according to the number of times the pixel is crossed by the full line.

5. The method for analyzing the electric energy quality of the single-phase electric energy meter according to claim 4, wherein the step of using the trained convolutional neural network to take the pixilated image as input to complete the waveform detection output comprises the following steps:

Adopting weighted pixilated waveforms of the electric energy quality types as model input quantity of a convolutional neural network, and training the convolutional neural network to obtain fully-connected weights;

and identifying and classifying the weighted pixilated images based on the trained convolutional neural network to obtain the type of the power quality problem to be identified.

6. the method for analyzing the power quality of the single-phase electric energy meter according to claim 5, wherein the power quality categories comprise: voltage sag, voltage fluctuation, harmonics, transient overvoltage.

7. a single-phase electric energy meter, comprising:

The analog-to-digital converter is used for outputting waveform data after the electric energy meter is powered on, and the outputting waveform data comprises the following steps: voltage waveform data and current waveform data;

The management module is used for calculating the gradient of each cycle based on the waveform data and recording the gradient of each cycle to form recording data; drawing a voltage-current curve graph based on the wave recording data; transforming the voltage-current graph into a weighted pixelated image; and adopting a trained convolutional neural network to take the pixilated image as input to finish waveform detection and output.

8. The single-phase electric energy meter according to claim 7, wherein the management module is configured to calculate a waveform gradient on each cycle of the waveform data, the waveform gradient being a waveform rate of change of a current cycle sampling point and a previous waveform sampling point; and identifying the condition of the maximum waveform gradient change, and recording the first four cycles and the last four cycles of the maximum waveform gradient change.

9. the single-phase electric energy meter of claim 8, wherein the management module is configured to plot a voltage-current curve based on the recording waveform with voltage as a vertical axis and current as a horizontal axis; and converting the voltage-current curve waveform into a weighted pixelized image, and dividing the image into n if the number of pixels on the horizontal axis and the vertical axis is respectively n²A plurality of pixels; to n²each pixel of the pixels is provided with different weight, the maximum value of the weight is [ (n +1)²-1]the minimum weight is 0, and the weight is assigned according to the number of times the pixel is crossed by the full line.

10. the single-phase electric energy meter according to claim 9, wherein the management module is configured to train the convolutional neural network to obtain fully-connected weights by using weighted pixilated waveforms of the power quality category as model input quantities of the convolutional neural network; and identifying and classifying the weighted pixilated images based on the trained convolutional neural network to obtain the type of the power quality problem to be identified.

Technical Field

the invention relates to the technical field of power detection, in particular to a method for analyzing the power quality of a single-phase electric energy meter and the single-phase electric energy meter.

background

with the continuous development of science and technology, the large-scale application of wind power generation and photovoltaic power generation and the increase of the application of power electronic equipment at the user side, the problems of power supply reliability, electromagnetic compatibility and the like caused by the reduction of the power quality of a power grid are increased day by day. The electric energy quality needs to be monitored and analyzed on the low-voltage user side urgently, on one hand, situation perception of the tail end of the smart power grid and development of a transparent power grid are facilitated, on the other hand, electric energy pollution on the user side is observed at any time, and power supply service quality is facilitated to be improved. At present, a single-phase electric energy meter of a low-voltage user has no electric energy quality detection and analysis function, and a harmonic electric energy meter with an independent function has a single function and is only used for monitoring the harmonic condition. The invention detects and analyzes part of low-frequency conducted noise and identifies the specific power quality problem.

Disclosure of Invention

The invention aims to overcome the defects of the prior art, provides the data characteristics of a single-phase electric energy meter, provides a method for analyzing the electric energy quality of the single-phase electric energy meter and the single-phase electric energy meter, and improves the identification accuracy.

in order to solve the above problems, the present invention provides a method for analyzing the electric energy quality of a single-phase electric energy meter, comprising the following steps:

After the electric energy meter is powered on, an analog-to-digital converter of an electric energy metering chip in the electric energy meter outputs waveform data, wherein the output waveform data comprises: voltage waveform data and current waveform data;

the electric energy meter management module calculates the gradient of each cycle based on the waveform data and records the gradient of each cycle to form recording data;

Drawing a voltage-current curve graph based on the wave recording data;

transforming the voltage-current graph into a weighted pixelated image;

And adopting a trained convolutional neural network to take the pixilated image as input to finish waveform detection and output.

the electric energy meter management module calculates the gradient of each cycle based on the waveform data, and records the gradient of each cycle to form recording data, wherein the recording data comprises the following steps:

calculating the waveform gradient on each cycle on the waveform data, wherein the waveform gradient is the waveform change rate of the current cycle sampling point and the previous waveform sampling point;

And identifying the condition of the maximum waveform gradient change, and recording the first four cycles and the last four cycles of the maximum waveform gradient change.

The step of drawing a voltage-current curve graph based on the wave recording data comprises the following steps:

Based on the waveform of the recording wave, the voltage is taken as the vertical axis, the current is taken as the horizontal axis, and a voltage-current curve is drawn.

The transforming the voltage-current graph into a weighted pixelated image includes:

Converting the voltage-current curve waveform into a weighted pixelized image, and dividing the image into n if the number of pixels on the horizontal axis and the vertical axis is respectively n²A plurality of pixels;

To n²Each pixel of the pixels is provided with different weight, the maximum value of the weight is [ (n +1)²-1]The minimum weight is 0, and the weight is assigned according to the number of times the pixel is crossed by the full line.

The step of finishing waveform detection and output by using the trained convolutional neural network to take the pixilated image as input comprises the following steps:

Adopting weighted pixilated waveforms of the electric energy quality types as model input quantity of a convolutional neural network, and training the convolutional neural network to obtain fully-connected weights;

and identifying and classifying the weighted pixilated images based on the trained convolutional neural network to obtain the type of the power quality problem to be identified.

The power quality categories include: voltage sag, voltage fluctuation, harmonics, transient overvoltage.

correspondingly, the embodiment of the invention also provides a single-phase electric energy meter, which comprises:

The analog-to-digital converter is used for outputting waveform data after the electric energy meter is powered on, and the outputting waveform data comprises the following steps: voltage waveform data and current waveform data;

the management module is used for calculating the gradient of each cycle based on the waveform data and recording the gradient of each cycle to form recording data; drawing a voltage-current curve graph based on the wave recording data; transforming the voltage-current graph into a weighted pixelated image; and adopting a trained convolutional neural network to take the pixilated image as input to finish waveform detection and output.

The management module is used for calculating the waveform gradient on each cycle on the waveform data, and the waveform gradient is the waveform change rate of the current cycle sampling point and the previous waveform sampling point; and identifying the condition of the maximum waveform gradient change, and recording the first four cycles and the last four cycles of the maximum waveform gradient change.

The management module is used for drawing a voltage-current curve by taking the voltage as a longitudinal axis and the current as a transverse axis based on the wave recording waveform; and converting the voltage-current curve waveform into a weighted pixelized image, and dividing the image into n if the number of pixels on the horizontal axis and the vertical axis is respectively n²a plurality of pixels; to n²Each pixel of the pixels is provided with different weight, the maximum value of the weight is [ (n +1)²-1]The minimum weight is 0, and the weight is assigned according to the number of times the pixel is crossed by the whole line

The management module is used for training the convolutional neural network to obtain fully-connected weights by adopting weighted pixilated waveforms of electric energy quality types as model input quantity of the convolutional neural network; and identifying and classifying the weighted pixilated images based on the trained convolutional neural network to obtain the type of the power quality problem to be identified.

In the embodiment of the invention, data acquisition of the electric energy meter is realized in the power-on process, the convolutional neural network model is trained by adopting the original data, the training data can be marked as the electric energy quality problems listed in the invention, and can also be marked as the electric energy quality problems with more definite reasons according to needs or scientific and technological development, such as voltage sag generated by the operation of power electronic devices, transient voltage change caused by lightning, transient voltage change caused by operation and the like, and the training model can be output as a classification result according to user-defined data, so that the flexibility is higher. Meanwhile, the method has higher identification accuracy and can identify different types of power quality fault types.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 shows a flow chart of a method for analyzing the electric energy quality of a single-phase electric energy meter in an embodiment of the invention;

fig. 2 is a schematic structural diagram of a single-phase electric energy meter in an embodiment of the invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

the method for analyzing the electric energy quality of the single-phase electric energy meter provided by the embodiment of the invention comprises the following steps: after the electric energy meter is powered on, an analog-to-digital converter of an electric energy metering chip in the electric energy meter outputs waveform data, wherein the output waveform data comprises: voltage waveform data and current waveform data; the electric energy meter management module calculates the gradient of each cycle based on the waveform data and records the gradient of each cycle to form recording data; drawing a voltage-current curve graph based on the wave recording data; transforming the voltage-current graph into a weighted pixelated image; and adopting a trained convolutional neural network to take the pixilated image as input to finish waveform detection and output.

Specifically, fig. 1 shows a flowchart of a method for detecting a post-fault of an electric energy meter in an embodiment of the present invention, which includes the following specific steps:

s101, after the electric energy meter is electrified, an analog-to-digital converter of an electric energy metering chip in the electric energy meter outputs waveform data;

Here, the output waveform data includes: voltage waveform data, and/or current waveform data.

After the electric energy meter is electrified, the metering chip ADC outputs voltage waveform data, current waveform data and the like, and the analog-to-digital converter ADC is a device for converting continuously-changed analog signals into discrete digital signals and needs to be converted into a digital form which is easier to store, process and transmit.

s102, calculating a waveform gradient on each cycle on waveform data, wherein the waveform gradient is the waveform change rate of a current cycle sampling point and a previous waveform sampling point;

And calculating the change rate of each sampling point of the voltage waveform and the current waveform. Taking the voltage waveform as an example, assuming that the voltage waveform of N cycles has N sampling points, the change rate of the m-th sampling point relative to m-1 is obtained, and the change rate of each sampling point of the voltage waveform is as follows (as formula 1):

similarly, the gradient of the current waveform is required to be obtained, N current waveform sampling points of N cycles are set, the change rate of the m-th sampling point relative to m-1 is obtained, and the change rate of each sampling point of the current waveform is as follows (as formula 2):

S103, identifying the condition of the maximum waveform gradient change, and recording the first four cycles and the last four cycles under the condition of the maximum waveform gradient change;

the method includes calculating the maximum change of a voltage gradient (i.e., the change rate of each sampling point of a voltage waveform) and a current gradient (i.e., the change rate of each sampling point of a current waveform) occurring in N cycles, and recording the four cycles before and after the change of the gradient, where N is an arbitrary natural number, and where N is a periodic calculation or a calculation pattern is referred to for the calculation of N cycles.

And S102 to S103, calculating the gradient of each cycle of wave based on the waveform data through the electric energy meter management module, and recording the gradient of each cycle of wave to form recording data. The waveform gradient on each cycle on waveform data is calculated, and the waveform gradient is the waveform change rate of a current cycle sampling point and a previous waveform sampling point; and identifying the condition of the maximum waveform gradient change, and recording the first four cycles and the last four cycles of the maximum waveform gradient change.

S104, drawing a voltage-current curve graph based on the wave recording data;

specifically, the voltage waveform, and the like in the recording data are extracted for the recording data formed in S103, so that the voltage-current curve can be plotted based on the recording waveform with the voltage as the vertical axis and the current as the horizontal axis.

S105, converting the voltage-current curve graph into a pixilated image with weight;

in particular, the voltage-current curve waveform is converted into a weighted onethe pixelized image is divided into n pixels by setting the number of pixels on the horizontal axis and the number of pixels on the vertical axis to n²A plurality of pixels; to n²Each pixel of the pixels is provided with different weight, the maximum value of the weight is [ (n +1)²-1]The minimum weight is 0, and the weight is assigned according to the number of times the pixel is crossed by the full line.

And S106, finishing waveform detection and output by using the trained convolutional neural network to take the pixilated image as input.

Specifically, a weighted pixilated waveform of the power quality type is used as a model input quantity of a convolutional neural network, and the convolutional neural network is trained to obtain a fully-connected weight; and identifying and classifying the weighted pixilated images based on the trained convolutional neural network to obtain the type of the power quality problem to be identified. Here, the power quality categories include: voltage sag, voltage fluctuation, harmonics, transient overvoltage.

Correspondingly, fig. 2 shows a schematic structural diagram of a single-phase electric energy meter in an embodiment of the present invention, including:

The analog-to-digital converter is used for outputting waveform data after the electric energy meter is powered on, and the outputting waveform data comprises the following steps: voltage waveform data and current waveform data;

The management module is used for calculating the gradient of each cycle based on the waveform data and recording the gradient of each cycle to form recording data; drawing a voltage-current curve graph based on the wave recording data; transforming the voltage-current graph into a weighted pixelated image; and adopting a trained convolutional neural network to take the pixilated image as input to finish waveform detection and output.

Specifically, the management module is configured to calculate a waveform gradient on each cycle of the waveform data, where the waveform gradient is a waveform change rate of a current cycle sampling point and a previous waveform sampling point; and identifying the condition of the maximum waveform gradient change, and recording the first four cycles and the last four cycles of the maximum waveform gradient change.

Specifically, the management module is used for taking the voltage as the longitudinal voltage based on the wave recording waveformthe axis, the current as the horizontal axis, plots the voltage-current curve; and converting the voltage-current curve waveform into a weighted pixelized image, and dividing the image into n if the number of pixels on the horizontal axis and the vertical axis is respectively n²a plurality of pixels; to n²Each pixel of the pixels is provided with different weight, the maximum value of the weight is [ (n +1)²-1]the minimum weight is 0, and the weight is assigned according to the number of times the pixel is crossed by the full line.

Specifically, the management module is used for training the convolutional neural network to obtain fully-connected weights by adopting weighted pixilated waveforms of electric energy quality types as model input quantities of the convolutional neural network; and identifying and classifying the weighted pixilated images based on the trained convolutional neural network to obtain the type of the power quality problem to be identified. Here, the power quality categories include: voltage sag, voltage fluctuation, harmonics, transient overvoltage.

To sum up, in the embodiment of the present invention, data acquisition of the electric energy meter is realized in the power-on process, the convolutional neural network model is trained by using the raw data, the training data may be labeled as the electric energy quality problem listed in the present invention, or may be labeled as the electric energy quality problem with a more definite reason according to needs or scientific and technological development, for example, voltage sag generated by operation of power electronic devices, transient voltage change caused by lightning, transient voltage change caused by operation, and the like, and the training model may be output as a classification result according to user-defined data, so that higher flexibility is achieved. Meanwhile, the method has higher identification accuracy and can identify different types of power quality fault types.

those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by associated hardware instructed by a program, which may be stored in a computer-readable storage medium, and the storage medium may include: a Read Only Memory (ROM), a Random Access Memory (RAM), a magnetic or optical disk, or the like.

In addition, the method for analyzing the electric energy quality of the single-phase electric energy meter and the single-phase electric energy meter provided by the embodiment of the invention are described in detail, a specific example is applied in the description to explain the principle and the implementation mode of the invention, and the description of the embodiment is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.