CN114678962B - Distributed Array Temperature Measurement Abnormal Data Transmission Monitoring System Based on Power Internet of Things - Google Patents

Abstract

The invention relates to a distributed array temperature measurement abnormal data transmission monitoring system based on an electric power internet of things, and belongs to the technical field of magnetic variables. The system can acquire the magnetic induction intensity corresponding to the transformer; according to the magnetic induction intensity and the temperature difference value, a temperature difference value sequence and a magnetic induction intensity sequence corresponding to each circle of the detection surface are constructed and obtained; calculating to obtain a corresponding local structural index in the operation process of the transformer according to the temperature difference sequence and the magnetic induction intensity sequence corresponding to each circle of the detection surface; and inputting the local structural index into a target support vector machine classifier, and judging whether the transformer is abnormal in operation or not by using the target support vector machine classifier. The invention realizes the real-time judgment of whether the running state of the transformer is abnormal in the running process of the transformer according to the magnetic induction intensity.

Description

Distributed Array Temperature Measurement Abnormal Data Transmission Monitoring System Based on Power Internet of Things

technical field

The invention relates to the technical field of magnetic variables, in particular to a distributed array temperature measurement abnormal data transmission monitoring system based on the Internet of Things.

Background technique

With the acceleration of urbanization and the continuous expansion of the city, the demand for electricity in different industries and fields is increasing. It is necessary to build more substations to meet the demand for electricity consumption. However, the urban land is very tight and it is difficult to acquire it. Reasonable substation building land, so in the face of this situation, it is necessary to build an underground substation; mainly because the main buildings of the underground substation are underground, such as the main transformer or other major electrical equipment are underground, and only a small part is arranged on the ground ; Since the underground substation is in a relatively closed space, the heat dissipation problem of the underground substation is an important issue affecting the safety of the underground substation. As the main equipment in the underground substation, the transformer needs to pay more attention to the operation status of the transformer; because there are Part of the electromagnetic energy is converted into heat energy, and part of this heat energy will be dissipated into the environment. Therefore, when the transformer fails or operates abnormally, the ambient temperature around the transformer will be abnormal, which may affect the safety of the underground substation. , so it is very important to monitor the ambient temperature during the operation of the transformer in the underground substation, that is, it is very important to monitor whether there is an abnormal operation during the operation of the transformer.

Contents of the invention

The present invention provides a distributed array temperature measurement abnormal data transmission monitoring system based on the Internet of Things, which is used to solve the problem that the existing method cannot reliably monitor the ambient temperature of the transformer in the underground substation when it is running. The technical scheme adopted is specific as follows:

In the first aspect, an embodiment of the present invention provides a distributed array temperature measurement abnormal data transmission monitoring system based on the Internet of Things, including a memory and a processor, and the processor executes the computer program stored in the memory to realize Follow the steps below:

Acquiring the temperature difference of each first target position and the magnetic induction intensity of each second target position during the operation of the transformer to be detected;

According to the position of each first target position on the detection surface and the temperature difference, construct and obtain the temperature difference sequence corresponding to each circle of the detection surface; according to the position of each second target position on the detection surface and the magnetic induction intensity to construct the magnetic induction intensity sequence corresponding to each circle of the detection surface; according to the temperature difference sequence and magnetic induction intensity sequence corresponding to each circle of the detection surface, calculate the corresponding local structure during the operation of the transformer sex index;

The local structural index is input into the target support vector machine classifier, and the target support vector machine classifier is used to judge whether there is abnormal operation during the operation of the transformer.

Beneficial effects: the present invention uses the temperature differences of each first target position and the magnetic induction intensity of each second target position during the operation of the transformer to be detected as the basis for obtaining the temperature difference sequence and magnetic induction intensity sequence corresponding to each circle of the detection surface ; Use the temperature difference sequence and the magnetic induction intensity sequence corresponding to each circle of the detection surface as the basis for calculating the corresponding local structural index during the operation of the transformer; use the local structural index and the target support vector machine classifier as the basis for judging the transformer during operation Whether abnormal operation occurs; the present invention realizes the real-time judgment on whether the transformer operating state is abnormal during the operation process of the transformer according to the magnetic induction intensity, and the present invention can relatively reliably monitor the operating state of the transformer in the underground substation.

Preferably, the method for obtaining the temperature difference of each first target position and the magnetic induction intensity of each second target position during the operation of the transformer to be detected includes:

According to the axis of the primary winding and the axis of the secondary winding of the transformer core, a first temperature detection surface and a second temperature detection surface corresponding to the transformer are obtained; a plurality of temperature sensors are distributed on the temperature detection surface;

The positions of the temperature sensors on the first temperature detection surface corresponding to the transformer are recorded as the first target positions;

Acquire the position corresponding to the first target position on the second temperature detection surface corresponding to the transformer, and record it as the matching position of each first target position on the first temperature detection surface corresponding to the transformer;

calculating the absolute value of the difference between the temperature value of each first target position on the first temperature detection surface corresponding to the transformer and the temperature value of the corresponding matching position;

The absolute value of the difference is recorded as the temperature difference corresponding to each first target position during the operation of the transformer;

Hall sensors are arranged on the housing surface of the transformer close to the N-level transformer magnet to collect the magnetic induction intensity, and the housing surface is recorded as the magnetic induction intensity detection surface corresponding to the transformer, and the magnetic induction intensity detection surface is provided with a Hall sensor;

The positions of the Hall sensors on the magnetic induction detection surface are recorded as the second target positions, and the magnetic induction of each second target position during the operation of the transformer is obtained.

Preferably, the method for obtaining the first temperature detection surface and the second temperature detection surface corresponding to the transformer according to the axis of the primary winding and the axis of the secondary winding of the transformer core includes:

According to the axis of the primary winding and the axis of the secondary winding, a plane is formed, and a temperature detection surface is respectively set at the same distance on both sides of the plane, and two temperature detection surfaces are set on the outside of the winding coil. Temperature sensors are respectively arranged on the two temperature detection surfaces, and the positions and numbers of the temperature sensors on the two temperature detection surfaces are the same and have a one-to-one correspondence relationship;

The two temperature detection surfaces corresponding to the transformer are respectively recorded as the first temperature detection surface and the second temperature detection surface corresponding to the transformer.

Preferably, the training process of the target support vector machine classifier includes: using the phase space analysis method to judge whether the local structural index of the training samples corresponding to the operation process of each transformer is abnormal, and marking the training samples corresponding to the operation process of each transformer; The corresponding training samples during the operation of the marked transformer are used to train the support vector machine classifier.

Preferably, the method of using the phase space analysis method to judge whether the local structural index of the corresponding training samples during the operation of each transformer is abnormal includes:

For the training samples corresponding to any transformer running process:

Taking the transformer target operating time period as a period of observation time in the phase space, calculate the standard deviation of the tracking index corresponding to each period of observation time during the operation of the transformer;

Calculate the structural separation index corresponding to each observation time according to the standard deviation of the tracking index corresponding to each observation time;

It is determined whether the structural separation index increases for more than two consecutive times during the operation of the transformer, and if so, it is determined that the local structural index of the transformer is abnormal during operation.

Preferably, according to the temperature difference sequence and the magnetic induction intensity sequence corresponding to each circle of the detection surface, the method for calculating the corresponding local structural index during the operation of the transformer includes:

Use the following formula to calculate the local structural index of each circle:

为第

圈的局部结构性指数，

为第

圈对应的温度差值序列，

为第

圈对应的磁感应强度序列，STD为标准差，range为极差，F为对角采样函数，

表示第

圈对应的对称元素对的数量；in,

for the first

The local structural index of the circle,

for the first

The temperature difference sequence corresponding to the circle,

for the first

The magnetic induction intensity sequence corresponding to the circle, STD is the standard deviation, range is the extreme difference, F is the diagonal sampling function,

Indicates the first

The number of symmetric element pairs corresponding to the circle;

According to the local structural index of each coil, the corresponding local structural index during operation of the transformer is obtained.

Description of drawings

In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or in the prior art, the following

A brief introduction to the accompanying drawings required in the description of the embodiments or the prior art, obviously, the accompanying drawings in the following description are only some embodiments of the present invention, for those of ordinary skill in the art, without paying Under the premise of creative work, other drawings can also be obtained based on these drawings.

FIG. 1 is a flow chart of a method for judging abnormal operation during the operation of a transformer in a distributed array temperature measurement abnormal data transmission monitoring system based on the Internet of Electric Power Things according to the present invention.

detailed description

The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, not all of them. Based on The embodiments of the present invention and all other embodiments obtained by persons of ordinary skill in the art belong to the protection scope of the embodiments of the present invention.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of skill in the art to which this invention belongs.

This embodiment provides a distributed array temperature measurement abnormal data transmission monitoring system based on the Internet of Things, and the details are as follows:

As shown in Figure 1, the distributed array temperature measurement abnormal data transmission monitoring system based on the Internet of Power Internet of Things includes a memory and a processor, and the processor executes the computer program stored in the memory to realize the judgment of the abnormal operation of the transformer, including The following steps:

Step S001, acquiring the temperature difference of each first target position and the magnetic induction intensity of each second target position during the operation of the transformer to be detected.

The specific process of obtaining the temperature differences of each first target position and the magnetic induction intensity of each second target position during the operation of the transformer is as follows:

In this embodiment, since part of the electromagnetic energy is converted into heat energy during the operation of the transformer, part of the heat energy will be dissipated into the environment, and when the transformer operates abnormally, the temperature around the transformer may be abnormal. Therefore, this embodiment obtains the temperature during the operation of the transformer as one of the basis for reflecting the operation state of the transformer, and the transformer is a single-phase transformer, the mechanical skeleton of the single-phase transformer is an iron core, and the iron core of the transformer is a rectangular iron core. The primary winding is set on one side of the core, and the secondary winding is set on the other side of the iron core. The axis of the primary winding and the axis of the secondary winding form a plane, and a temperature detector is respectively set at the same distance on both sides of the plane. The two temperature detection surfaces are set on the outer side of the winding coil, and each temperature detection surface is respectively provided with a temperature sensor. The positions and numbers of the temperature sensors on the two temperature detection surfaces are the same, showing a one-to-one correspondence; therefore, the transformer Corresponding to two temperature detection surfaces, and record the two temperature detection surfaces corresponding to the transformer as the first temperature detection surface and the second temperature detection surface corresponding to the transformer; record the positions of the temperature sensors on the first temperature detection surface corresponding to the transformer is each first target position; and the position corresponding to the first target position is obtained on the second temperature detection surface corresponding to the transformer, and recorded as the matching position of each first target position on the first temperature detection surface corresponding to the transformer; and Obtain the temperature value of the matching position; calculate the absolute value of the difference between the temperature value of each first target position on the first temperature detection surface corresponding to the transformer and the temperature value of the corresponding matching position; record the absolute value of the difference as The temperature difference corresponding to each first target position on each first temperature detection surface in the transformer; therefore, the temperature difference value of each first target position during the operation of the transformer can be obtained through the above process.

Since the magnetic induction intensity is a physical quantity describing the strength and direction of the magnetic field, based on the Hall effect, it can be known that the magnetic induction intensity and the excitation current are positively correlated. Therefore, when the single-phase transformer is working, if the transformer operates abnormally, the excitation current will be Changes or copper loss and iron loss in the transformer can affect the magnetic induction intensity. Therefore, changes in the magnetic induction intensity of the transformer can reflect the operating state of the transformer, and the temperature around the transformer will also be abnormal; The housing surface of the transformer is arranged with a Hall sensor to collect the magnetic induction intensity, and the housing surface is recorded as the magnetic induction intensity detection surface corresponding to the transformer. The magnetic induction intensity detection surface is provided with a Hall sensor, and the Hall sensor is used in the magnetic induction intensity detection The surface presents a uniform array distribution, the number of sensors on the magnetic induction detection surface is the same as the number of sensors on the corresponding first temperature detection surface, and the arrangement of sensors on each detection surface is also similar; in this embodiment, each Hall on the magnetic induction detection surface The position of the sensor is recorded as each second target position; therefore, the magnetic induction intensity of each second target position during the operation of the transformer can be obtained.

Step S002, according to the position of each first target position on the detection surface and the temperature difference, construct the temperature difference sequence corresponding to each circle of the detection surface; according to the position of each second target position on the detection surface According to the position on the detection surface and the magnetic induction intensity, the magnetic induction intensity sequence corresponding to each circle of the detection surface is constructed; according to the temperature difference sequence and magnetic induction intensity sequence corresponding to each circle of the detection surface, the corresponding magnetic induction intensity sequence during the operation of the transformer is calculated. The local structural index of .

The specific process of calculating the corresponding local structural index during the operation of the transformer is as follows:

The temperature difference matrix and the magnetic induction intensity matrix are constructed according to the temperature difference of each first target position and the magnetic induction intensity of each second target position during the operation of the transformer; and because the temperature sensors are distributed in a uniform array on the temperature detection surface, Hall The sensors are also distributed in a uniform array on the magnetic induction detection surface, and the sensors on the temperature detection surface are similar to the sensors on the magnetic induction detection surface, so the corresponding magnetic induction matrix and the corresponding temperature difference matrix during the operation of the transformer. The number of rows and the number of columns are the same; therefore, in this embodiment, the magnetic induction detection surface is recorded as the detection surface, and the position of the sensor corresponding to the i-th row and the j-th column in the corresponding temperature difference matrix during the operation of the transformer is recorded as the corresponding The position of the sensor corresponding to the i-th row and the j-th column in the magnetic induction matrix; therefore, it can be considered that each second target position on the magnetic induction detection surface corresponding to the transformer will correspond to two sensors, namely a temperature sensor and a Hall sensor , that is, the magnetic induction matrix and temperature difference matrix corresponding to the detection surface of the transformer.

Based on the above matrix, the sensor readings are accessed in the shape of a zigzag. The way of accessing the zigzag is: from the smallest unit in the center, such as 2✕2, expand a circle outward to obtain samples of the outer circle (including 4*4-2* 2=12 samples), and so on, to get the samples of each circle. When the width or height of a circle cannot continue to expand, only the number of samples existing in the expanded circle is counted.

Since the magnetic induction intensity of each position on the magnetic induction intensity detection surface is similar during normal operation of the transformer, the temperature difference at the corresponding position on the first temperature detection surface and the second temperature detection surface is similar, but when the transformer is used frequently, it may cause the transformer There are some abnormal phenomena in the components of the transformer, which lead to abnormalities in the operation of the transformer, that is, the abnormal heat energy generated during the operation of the transformer or the abnormal magnetic induction intensity during the operation of the transformer. Therefore, the distributed array temperature measurement abnormal data transmission monitoring system is for Abnormal data transmission runs are monitored.

和温度差值

的读数，其中a为每圈的索引。Based on one lap per visit, the magnetic induction intensity of one lap can be obtained

and temperature difference

The readings of , where a is the index of each revolution.

，计算公式为：For readings in one lap, calculate the local structural index for each lap

, the calculation formula is:

为第

圈的局部结构性指数，

为第

圈对应的温度差值序列，

为第

圈对应的磁感应强度序列，STD为标准差，range为极差，F为对角采样函数，F每次采集一个元素以及该元素绕中心对称的对角方向的对向元素；由于变压器运行环境应当是统一的，因此磁感应强度所反映的进程也应当是尽可能的统一。

代表了第

圈所有两个对称的磁感应强度相差的均值，

表示第

圈对应的对称元素对的数量。当该均值较大时，意味着进程不统一，从而增大温度差值不统一的标准差，从而体现变压器运行过程中的局部结构性指数；当该指数过大时，意味着检测面的一圈无法很好均一运行环境，体现局部的差异较大的现象。in,

for the first

The local structural index of the circle,

for the first

The temperature difference sequence corresponding to the circle,

for the first

The magnetic induction intensity sequence corresponding to the circle, STD is the standard deviation, range is the extreme difference, F is the diagonal sampling function, F collects one element each time and the opposite element in the diagonal direction of the element symmetrical around the center; because the transformer operating environment should is uniform, so the process reflected by the magnetic induction intensity should also be as uniform as possible.

represents the first

The average value of the difference between all two symmetrical magnetic induction intensities of the circle,

Indicates the first

The number of pairs of symmetric elements to which the circle corresponds. When the average value is large, it means that the process is not uniform, thereby increasing the standard deviation of the temperature difference, which reflects the local structural index during the operation of the transformer; when the index is too large, it means that the detection surface is uniform The circle cannot be well uniformed in the operating environment, which reflects the phenomenon of large local differences.

，

，N为变压器运行过程中的检测面能够被采样圈数。So far, based on the local structural index of each circle, the corresponding local structural index vector during the operation of the transformer is obtained

,

, N is the number of turns that the detection surface can be sampled during the operation of the transformer.

Step S003, input the local structural index into the target support vector machine classifier, and use the target support vector machine classifier to judge whether there is abnormal operation during the operation of the transformer.

Because the present embodiment judges whether the abnormal operation phenomenon occurs in the transformer in the running period by the support vector machine classifier; and the support vector machine classifier needs to be trained before it can be used; therefore, the training samples need to be marked, and for To improve the accuracy of the support vector machine classifier, this embodiment uses the phase space analysis method as the basis for marking the training samples, and divides the marking types into two categories, one is normal and the other is abnormal; although based on phase space The spatial method can realize the judgment of whether the transformer is abnormal, but the abnormality analysis based on the phase space analysis method needs to ensure that the change of the phase space is continuous. Real-time judgment of whether it is abnormal, so this embodiment uses the phase space method to mark the training samples in the training process of the support vector machine classifier, uses the training samples to carry out supervised training on the support vector machine classifier, and uses the trained support vector The machine classifier realizes the real-time judgment on whether abnormal operation occurs in the operation process of the transformer. Next, the related process will be explained:

Since the general phase space reconstruction generally uses a uniform embedding method, but it cannot be applied to the array data of the transformer, so this embodiment analyzes the edge effect based on the local structure, so that the data inhomogeneity in the array can be continuously expressed and its system Therefore, for any transformer training sample, this embodiment characterizes the local structural index of the corresponding sample, and then the temperature difference and magnetic induction intensity of each position of the corresponding detection surface of the transformer during the operation of the transformer The evolution of the space is expanded according to the polar coordinate method, so as to realize an improved space bending effect.

，其中N为局部结构的个数，即采样圈数；利用互信息法选取延迟时间参数

，利用虚假邻近点法选取嵌入维数参数m，相空间重构的方式为：The phase space is constructed so that the local structural index can be used as the space of all possible states during the operation of the transformer. Under the observation of manual participation, after ensuring that the samples are normally loaded during a transformer operation, set the starting time t=0, and establish its vector through local structure

, where N is the number of local structures, that is, the number of sampling circles; use the mutual information method to select the delay time parameter

, using the false neighbor method to select the embedding dimension parameter m, the phase space reconstruction method is:

。So far, reconstruct the phase space of the local structural exponential change of N local structures, and use this phase space as the reference phase space

.

，更新数据点

，其中N为局部结构的个数，即采样圈数。最小分析间隔

是人为指定的，本实施例可以设置为10秒或者20秒；采用与参考相空间

相同的延迟时间

和嵌入维数

重构

时刻局部结构性指数的相空间，采用和上述相同方式重构当前的相空间：Each time the readings of the magnetic induction intensity matrix and the temperature difference matrix are updated, the local structural index of each substructure can be obtained

, update the data point

, where N is the number of local structures, that is, the number of sampling cycles. Minimum Analysis Interval

It is artificially specified, and this embodiment can be set to 10 seconds or 20 seconds; using the reference phase space

same delay time

and embedding dimension

refactor

For the phase space of the local structural index at any time, the current phase space is reconstructed in the same way as above:

...

So far, the phase space U during the operation of the transformer can be obtained.

时刻相空间中的某一个向量

，在参考相空间中寻找

个与其距离最近的向量

，其中

。for

A certain vector in the phase space of time

, in the reference phase space to find

the closest vector to

,in

.

，以T为观察时间间隔，得到观察时间间隔T内的结构分离指数

。具体的，对于相空间，有：Based on the above processing method, all data points are updated for each sampling

, taking T as the observation time interval, the structure separation index in the observation time interval T is obtained

. Specifically, for the phase space, there are:

的跟踪函数为：

The tracking function for is:

与向量

的距离，最远的距离为

,设计相空间权值

。令

自增1，继续计算

，直到

。然后，利用

时刻相空间中所有向量对应的所有跟踪函数，并计算

时刻相空间的跟踪指标：calculate

with vector

distance, the farthest distance is

, the design phase space weights

. make

Increment by 1, continue to calculate

,until

. Then, use

All tracking functions corresponding to all vectors in the phase space at time, and compute

Tracking metrics in time-phase space:

，

为

时刻相空间的关联维数。以一段时间为相空间的观测时间，例如60s，共观测6次P值，T=6，每个P值对应10s，计算

,共T个跟踪指标的平均值

及标准差

。Among them, q(n) is the weight function,

,

for

Correlation dimension of the phase space at time. Take a period of time as the observation time of the phase space, such as 60s, observe 6 P values in total, T=6, each P value corresponds to 10s, calculate

, the average value of a total of T tracking indicators

and standard deviation

.

变化过程在一段观察时间间隔T的相空间状态的指标。当一段时间内状态指标发生较大变化时，意味着的变压器所依赖的环境发生了较明显的变化。一般情况下，当误差导致各区域发生缓慢的差异变化时，变压器运行过程中局部结构性指数会发生波动；因此，标准差

会较小，从而可以使用3西格玛准则来估计变压器运行质量时可容忍的变化。本实施例基于变压器运行过程的相空间分析，得到结构分离指数

。Tracking metrics reveal the transformer operating process per smallest unit

The change process is an indicator of the phase space state over a period of observation time interval T. When the state index changes greatly within a period of time, it means that the environment on which the transformer depends has changed significantly. In general, local structural indices fluctuate during transformer operation when errors cause slow differential changes across regions; therefore, the standard deviation

will be small so that the 3 sigma criterion can be used to estimate the tolerable variation in transformer operating quality. This embodiment is based on the phase space analysis of the transformer operation process to obtain the structural separation index

.

The above-mentioned phase space analysis results are obtained during one transformer operation, but not for multiple times. Since the tracking indicators are in the end period, the indicators of the system are different from the samples corresponding to the operation process of the brand new transformer. The value may be high, so it is necessary to ensure that the change of the phase space is continuous, so the data in the next transformer operation process is processed in reverse order in time, so that the two transformer operation processes form a cycle to ensure that the phase space analysis is continuous and infinitely long, so that more Accurately analyze the structural separation index to avoid the impact caused by restarting the transformer operation process. Therefore, according to the above method, when the transformer is running next time, all records need to be processed in reverse order, and the last sample is taken as the time t=0 to ensure that the change process is continuous. cycle.

Although the phase space can track the operation process of the transformer and find abnormalities, in view of the need to construct a cyclic process, not every process can be performed in real time. Therefore, this embodiment constructs SVM samples based on two categories of structural discrete codes, and initializes the SVM. The specific process is as follows:

Determine whether there is abnormal operation data that affects the operation of the transformer, and construct the corresponding local structural index w.

The structural discrete code is the vector of the local structural index represented by the detection surface at each moment, and the structural separation index L corresponding to the detection surface can be calculated based on the vector of the local structural index represented by the detection surface at each moment.

在两次测量和更新后都比之前的L大，则意味着相临的共三个结构离散码都能够体现变压器运行过程中发生运行异常的现象，对其标注为B。反之对其它样本标注为A，代表该运行过程正常。Two levels of structured discrete codes are obtained based on L, and the high discrete structured discrete codes and low discrete structured discrete codes are marked. The marking process is: when the transformer is running, the structural separation index

After two measurements and updates, it is larger than the previous L, which means that the adjacent three structural discrete codes can reflect the phenomenon of abnormal operation during the operation of the transformer, which is marked as B. On the contrary, the other samples are marked as A, which means that the operation process is normal.

This sample includes all the data encountered during the operation of the transformer. It is approximately complete data, and the classification of the sample is determined. Based on this classification, the SVM classifier is trained. This classifier divides the quality effect of the transformer operation process into two categories: Classes, including transformer normal operation type A and transformer operation minor abnormal type B.

Train the support vector machine classifier to get the trained target support vector machine classifier:

This embodiment is based on the two types of automatic analysis, adopts a supervised classification method and uses a support vector machine (SVM) classifier to perform classification based on the above-mentioned characteristic parameters during the operation of the transformer. The advantage of the SVM classifier is that it can use the hyperplane and kernel function to perform linear or nonlinear classification, and the result is relatively accurate. Therefore, this embodiment uses a fuzzy multi-class SVM-based transformer abnormality detection and classification method.

The calculation process of the SVM in this embodiment is: dividing hyperplanes with different anomalies, calculating the interval between different anomalies, analyzing the condition of the hyperplane when the interval is the largest, and summarizing the optimal hyperplane. Classify the composite rubber data completed during the operation of the transformer. The detailed calculation steps belong to the prior art and will not be repeated here.

Eighty percent of the transformer operation data sample obtained above is used as a training sample, and the remaining 20 percent is used as a test sample, and the training sample data corresponding to the transformer operation process is used to perform a support vector machine SVM classifier, and the classifier is changed to The value of the parameters, calculate the corresponding parameter values when the classification performance of the transformer operation effect classifier reaches the best, so far the classifier completes the parameter training, and puts the remaining 20% of the transformer operation data test samples into the classifier , to test the classification effect, and then judge whether the classification is correct and whether the correct rate meets the requirements, otherwise continue to modify the parameters of the classifier until the correct rate meets the requirements for use.

So far, the trained support vector machine SVM classifier can be obtained, and it is recorded as the target support vector machine classifier. Based on the target support vector machine classifier, two classifications of transformer operation quality can be realized, one is normal conditions, and the other is It is a slight abnormality. Since severe abnormalities are generally impossible to occur, this embodiment does not consider serious abnormalities.

Therefore, in this embodiment, based on the obtained target support vector machine classifier, it can be judged whether abnormal operation phenomenon occurs in the operation process of the transformer to be detected; therefore, the local structural index corresponding to the operation process of the transformer to be detected is input into the target support vector machine In the classifier, the target support vector machine classifier is used to judge whether there is an abnormal operation phenomenon in the operation process of the transformer to be detected.

，轻微异常别标记为

，而对于分配到了

组的变压器运行过程，此类中变压器运行过程发生了一些轻微变压器运行过程异常，可能是变压器元件性能出现异常。本实施例减少了人工观测时间成本、且解决多个变压器运行过程的样本在温度、变压器运行过程的空间中的混沌系统的所有可能性中如何合理判断运行异常的问题。Put the local structural index corresponding to the operation process of the transformer to be detected into the SVM support vector machine to perform two classifications through calculation, and obtain the result of abnormal operation in the operation process of the transformer to be detected; specifically, if the situation is normal, it will be marked as

, minor anomalies are not marked as

, and for the assigned

During the operation of the transformer of the group, some slight abnormalities in the operation of the transformer occurred during the operation of the transformer, which may be due to the abnormal performance of the transformer components. This embodiment reduces the time cost of manual observation, and solves the problem of how to reasonably judge abnormal operation among all the possibilities of the temperature and chaotic system in the space of the transformer operating process in the samples of the operating process of multiple transformers.

Beneficial effects: In this embodiment, the temperature differences of each first target position and the magnetic induction intensity of each second target position during the operation of the transformer to be detected are used as the key to obtain the temperature difference sequence and magnetic induction intensity sequence corresponding to each circle of the detection surface Basis; use the temperature difference sequence and magnetic induction intensity sequence corresponding to each circle of the detection surface as the basis for calculating the corresponding local structural index during the operation of the transformer; use the local structural index and the target support vector machine classifier as the basis for judging the operation process of the transformer The basis of whether there is abnormal operation in the transformer; this embodiment realizes the real-time judgment on whether the transformer operating state is abnormal during the operation of the transformer according to the magnetic induction intensity, and this embodiment can relatively reliably monitor the operating state of the transformer in the underground substation.

The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still implement the foregoing embodiments Modifications to the technical solutions described in the examples, or equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the application, and should be included in the Within the protection scope of this application.

Claims (4)

1. A distributed array temperature measurement abnormal data transmission monitoring system based on an electric power Internet of things comprises a memory and a processor, and is characterized in that the processor executes a computer program stored in the memory to realize the following steps:

acquiring the temperature difference value of each first target position and the magnetic induction intensity of each second target position in the operation process of the transformer to be detected;

according to the position of each first target position on a first temperature detection surface and the temperature difference value, constructing and obtaining a temperature difference value sequence corresponding to each circle of the first temperature detection surface; according to the position of each second target position on the magnetic induction intensity detection surface and the magnetic induction intensity, constructing and obtaining a magnetic induction intensity sequence corresponding to each circle of the magnetic induction intensity detection surface; calculating to obtain a corresponding local structural index in the operation process of the transformer according to the temperature difference sequence corresponding to each circle of the first temperature detection surface and the magnetic induction intensity sequence corresponding to each circle of the magnetic induction intensity detection surface;

inputting the local structural index into a target support vector machine classifier, and judging whether the transformer is abnormal in operation by using the target support vector machine classifier;

the method for acquiring the temperature difference of each first target position and the magnetic induction intensity of each second target position in the operation process of the transformer to be detected comprises the following steps:

obtaining a first temperature detection surface and a second temperature detection surface corresponding to the transformer according to the axis of the primary winding and the axis of the secondary winding of the transformer core; a plurality of temperature sensors are distributed on the temperature detection surface;

recording the position of each temperature sensor on a first temperature detection surface corresponding to the transformer as each first target position;

acquiring a position corresponding to the first target position on a second temperature detection surface corresponding to the transformer, and recording the position as a matching position of each first target position on a first temperature detection surface corresponding to the transformer;

calculating the absolute value of the difference between the temperature value of each first target position on the first temperature detection surface corresponding to the transformer and the temperature value of the corresponding matching position;

recording the absolute value of the difference as the temperature difference corresponding to each first target position in the running process of the transformer;

arranging a Hall sensor on a shell surface of the transformer close to the N-level of the transformer magnet to acquire magnetic induction intensity, marking the shell surface as a magnetic induction intensity detection surface corresponding to the transformer, and arranging the Hall sensor on the magnetic induction intensity detection surface;

recording the position of each Hall sensor on the magnetic induction intensity detection surface as each second target position, and obtaining the magnetic induction intensity of each second target position in the running process of the transformer;

the method for obtaining the first temperature detection surface and the second temperature detection surface corresponding to the transformer according to the axis of the primary winding and the axis of the secondary winding of the transformer core comprises the following steps:

forming a plane according to the axis of the primary winding and the axis of the secondary winding, respectively arranging a temperature detection surface at the same distance on two sides of the plane, wherein the two temperature detection surfaces are arranged on the outer sides of the winding coils, the two temperature detection surfaces are respectively provided with a temperature sensor, and the positions and the number of the temperature sensors arranged on the two temperature detection surfaces are the same and are in one-to-one correspondence;

recording the two temperature detection surfaces corresponding to the transformer as a first temperature detection surface and a second temperature detection surface corresponding to the transformer respectively;

the method for constructing and obtaining the temperature difference value sequence corresponding to each circle of the first temperature detection surface according to the position of each first target position on the first temperature detection surface and the temperature difference value comprises the following steps:

reading the temperature difference value of each first target position on the first temperature detection surface in a zigzag mode, specifically: firstly, reading the temperature difference of each first target position on a minimum circle consisting of 2 multiplied by 2 first target positions at the central position of a first temperature detection surface to obtain a temperature difference sequence of the minimum circle, then sequentially expanding one circle outwards on the basis of the minimum circle, reading the temperature difference of each first target position on each outer expansion circle to obtain a temperature difference sequence of each outer expansion circle, and when the width or height of one circle cannot be expanded outwards, only reading the temperature difference of the first target position on the outer expansion circle, and finally constructing to obtain a temperature difference sequence corresponding to each circle of the first temperature detection surface;

the method for constructing and obtaining the magnetic induction intensity sequence corresponding to each circle of the magnetic induction intensity detection surface according to the position of each second target position on the magnetic induction intensity detection surface and the magnetic induction intensity comprises the following steps:

reading the magnetic induction intensity of each second target position on the magnetic induction intensity detection surface in a square-shaped mode, and specifically comprises the following steps: firstly, reading the magnetic induction intensity of each second target position on a minimum circle consisting of 2 multiplied by 2 second target positions at the central position of a magnetic induction intensity detection surface to obtain a magnetic induction intensity sequence of the minimum circle, then expanding the minimum circle outwards in sequence for one circle, reading the magnetic induction intensity of each second target position on each external expansion circle to obtain a magnetic induction intensity sequence of each external expansion circle, when the width or height of one circle cannot be expanded outwards continuously, only reading the magnetic induction intensity of the second target position existing on the external expansion circle, and finally constructing and obtaining the magnetic induction intensity sequence corresponding to each circle of the magnetic induction intensity detection surface.

2. The power internet of things-based distributed array thermometry anomaly data transmission monitoring system according to claim 1, wherein the training process of the target support vector machine classifier comprises: judging whether the local structural indexes of the training samples corresponding to the operation processes of the transformers are abnormal by using a phase space analysis method, and marking the training samples corresponding to the operation processes of the transformers; and training the support vector machine classifier based on the corresponding training samples in the marked transformer operation process.

3. The power internet of things-based distributed array temperature measurement abnormal data transmission monitoring system of claim 2, wherein the method for judging whether the local structural index of the corresponding training sample in the operation process of each transformer is abnormal by using a phase space analysis method comprises the following steps:

for a training sample corresponding to any transformer operation process:

calculating the standard deviation of the tracking index corresponding to each observation time in the running process of the transformer by taking the target running time period of the transformer as one observation time period of a phase space;

calculating the structural separation index corresponding to each observation time according to the standard deviation of the tracking index corresponding to each observation time;

and judging whether the structural separation index is increased for more than two times in the running process of the transformer, and if so, judging that the local structural index of the running process of the transformer is abnormal.

4. The power internet of things-based distributed array temperature measurement abnormal data transmission monitoring system of claim 1, wherein the method for obtaining the corresponding local structural index in the running process of the transformer through calculation according to the temperature difference sequence and the magnetic induction intensity sequence corresponding to each circle of the detection surface comprises the following steps:

the local structural index of each turn is calculated using the following formula:

wherein,

is as follows

The local structural index of the ring is,

is as follows

The temperature difference value sequence corresponding to the ring,

is as follows

The magnetic induction intensity sequence corresponding to the circle, STD is standard deviation, range is range, F is diagonal sampling function,

denotes the first

The number of pairs of symmetric elements corresponding to the circle;

and obtaining a corresponding local structural index in the running process of the transformer according to the local structural index of each circle.

Images