A Method and System for Predicting the State of a Power Terminal Device

Abstract

The present invention relates to the technical field of data processing, and in particular to a method and system for predicting the state of a power terminal device. The method includes the steps of: collecting output current data and output voltage data, obtaining a first weight coefficient for each output current data according to the inverse proportional relationship between the output current data and the output voltage data; obtaining a second weight coefficient for each output current data according to the surrounding data segments of each output voltage data and the first weight coefficient, obtaining a prediction error of the current output current data according to the second weight coefficient of each output current data; obtaining a predicted value of the current output current data according to the prediction error of the current output current data, judging the abnormal situation of the current output current data according to the predicted value of the current output current data, and further judging the abnormality of the transformer. The present invention improves the accuracy of anomaly detection and avoids the influence of noise data.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method and system for predicting the state of a power terminal device. Background Art

[0002] By predicting the state of power terminal devices, potential faults can be detected in advance, reducing the risk of sudden device failures. As one of the terminal devices in the power system, a transformer is mainly used for voltage conversion and adjustment, and its health status directly affects the stability and safety of power supply. The output current of the transformer is an important indicator for monitoring its health status. The output current reflects the load condition of the transformer. By monitoring the change in current, potential faults of the transformer can be predicted. However, when collecting the output current data of the transformer, the transformer itself generates electromagnetic fields during operation, and these electromagnetic fields can interfere with sensors through wires and other means, resulting in noise in the output current data. Therefore, the accuracy of fault warning for the transformer by analyzing the directly collected output current data is reduced. Thus, the present invention predicts the output current data at each sampling moment, and comprehensively determines whether a transformer has a fault based on the predicted value and the actual value of the output current data.

[0003] Zheng Junbao, Zhang Xu, Research on the Time Series Prediction Method Based on Historical Trend and Prediction Error, Computer Era, No. 9, 2023. The principle and process of the HTFE algorithm are proposed in this paper.

[0004] When using the HTFE algorithm to predict the current output current data, first, the error between the predicted value and the true value of the output current data at the previous sampling moment is multiplied by the error factor to obtain the prediction error of the current output current data. Then, based on the prediction error of the current output current data and in combination with the HTFE algorithm, the predicted value of the current output current data is obtained to judge the fault of the current transformer. However, due to the existence of electromagnetic interference, there are some noise data in the historical data. Then, when there is potential noise interference in the historical data, the method of only relying on the traditional HTFE algorithm to obtain the prediction error of the current output current data and then obtaining the predicted value of the current output current data is not accurate enough, thus affecting the accuracy of the fault judgment for the transformer. Summary of the Invention

[0005] In order to solve the technical problem that electromagnetic interference causes noise data in historical data, resulting in the inaccuracy of the method of using the traditional HTFE algorithm to obtain the prediction error of the current output current data and then obtaining the predicted value of the current output current data, the present invention provides a method and system for predicting the state of a power terminal device.

[0006] In a first aspect, the present invention provides a method for predicting the state of a power terminal device, adopting the following technical solution:

[0007] A method for predicting the state of a power terminal device, comprising the steps of:

[0008] Collect output current data and output voltage data; obtain the second weight coefficient of each output current data , representing the second weight coefficient of the nth output current data; representing the first weight coefficient of the nth output current data; exp() represents the exponential function with the natural constant as the base; representing the number of extreme points in the surrounding data segment of the output voltage data corresponding to the sampling moment of the nth output current data; representing the range of the output voltage data corresponding to the sampling moment of the nth output current data in the surrounding data segment; norm() represents the normalization function;

[0009] Obtain the prediction error of the current output current data , M represents the number of historical data of the current output current data; representing the error value of the mth historical data of the current output current data; representing the second weight coefficient of the mth historical data of the current output current data; representing the sum of the second weight coefficients of all historical data of the current output current data;

[0010] Use the HTFE algorithm, combined with the prediction error of the current output current data, to obtain the predicted value of the current output current data; according to the predicted value, judge the abnormality of the current output current data, and then judge whether the transformer is abnormal.

[0011] The innovation of the present invention lies in that according to the numerical characteristics of the output current data and the output voltage data, the second weight coefficient of each output current data is obtained, and the error value of each output current data is corrected according to the second weight coefficient of each output current data, so as to obtain the prediction error of the current output current data. This step considers more errors of the output current data, avoids being easily interfered by noise when relying on the error of a single output current data, makes the prediction result more stable and has higher robustness; and this step can automatically give more reliable output current data greater influence during calculation by dynamically adjusting the second weight coefficient of each output current data, while the noise data points will reduce their contribution to the final prediction due to their greater untruthfulness, thereby reducing the influence of noise on the prediction result, enhancing the accuracy and stability of the fault diagnosis system, being able to identify potential transformer faults earlier, and reducing the risks of misjudgment and missed judgment.

[0012] Preferably, the obtaining of the first weight coefficient of the output current data includes:

[0013] ;

[0014] wherein, represents the first weight coefficient of the nth output current data; R represents the number of reference data; represents the absolute value of the difference between the value of the nth output current data and the value of its rth reference data; represents the number of data whose values of all reference data of the nth output current data are greater than the value of the nth output current data; represents the number of data whose values of all reference data of the output voltage data corresponding to the sampling moment of the nth output current data are greater than the value of the output voltage data; exp() represents the exponential function with the natural constant as the base; represents the hyperparameter; || represents the absolute value symbol.

[0015] The smaller the first weight coefficient of the output current data, the greater the probability that the output current data is more noise data, and the less its error will be referred to subsequently.

[0016] Preferably, the obtaining of the reference data includes:

[0017] Preset the number of reference data R, and use the R output current data before each output current data as the reference data for each output current data; use the R output current data before each output voltage data as the reference data for each output voltage data.

[0018] Preferably, the obtaining of the surrounding data segment of the output voltage data includes:

[0019] Record the data segment formed by the reference data of each output voltage data and each output voltage data as the surrounding data segment of each output voltage data.

[0020] Preferably, the obtaining of the historical data of the current output current data includes:

[0021] Record the M output current data before the sampling moment of the current output current data as the historical data of the current output current data.

[0022] Preferably, the obtaining of the error value of the historical data includes:

[0023] Obtain the predicted value of each historical data of the current output current data according to the HTFE algorithm, and use the absolute value of the difference between the predicted value and the actual value of each historical data of the current output current data as the error value of each historical data of the current output current data.

[0024] It is convenient to obtain the prediction error of the current output current data according to the error values of each historical data of the current output current data subsequently.

[0025] Preferably, obtaining the predicted value of the current output current data includes:

[0026] The error factor of the preset prediction interval is T. According to the prediction error of the current output current data, the predicted value of the current output current data is calculated using the HTFE algorithm to obtain the predicted value of the current output current data.

[0027] The obtained predicted value of the current output current data is more accurate.

[0028] Preferably, judging the abnormal situation of the current output current data according to the predicted value, and then judging whether the transformer is abnormal, includes:

[0029] Preset , if the absolute value of the difference between the predicted value and the true value of the current output current data is greater than the product of the predicted value of the current output current data and , it is considered that the current output current data is abnormal data. If the number of abnormal data exceeds 3 within one minute after the sampling moment corresponding to the current output current data, the transformer issues a warning.

[0030] The abnormal detection result is improved.

[0031] In a second aspect, the present invention provides a state prediction system for a power terminal device, adopting the following technical solution:

[0032] A state prediction system for a power terminal device includes: a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned state prediction method for a power terminal device is implemented.

[0033] By adopting the above technical solution, the above-mentioned state prediction method for a power terminal device is generated into a computer program and stored in the memory to be loaded and executed by the processor, so as to manufacture a terminal device according to the memory and the processor, which is convenient to use.

[0034] The present invention has the following technical effects: The object of the present invention is to obtain the second weight coefficient of each output current data according to the numerical characteristics of the output current data and the output voltage data, and correct the error value of each output current data according to the second weight coefficient of each output current data, so as to obtain the prediction error of the current output current data. This step takes into account more errors of the output current data, avoids being easily interfered by noise when relying on the error of a single output current data, makes the prediction result more stable and has higher robustness, and by dynamically adjusting the second weight coefficient of each output current data, more reliable output current data can be automatically given greater influence during calculation, while noise data points will reduce their contribution to the final prediction due to their greater untruthfulness, thereby reducing the influence of noise on the prediction result, enhancing the accuracy and stability of the fault diagnosis system, being able to identify potential transformer faults earlier, and reducing the risks of misjudgment and missed judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] By referring to the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown by way of example and not limitation, and the same or corresponding reference numerals represent the same or corresponding parts.

[0036] Figure 1 It is a flowchart of a method for predicting the state of a power terminal device in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0038] It should be understood that when the claims, specifications and drawings of the present invention use terms such as "first" and "second", they are only used to distinguish different objects and not to describe a specific order. The terms "comprising" and "including" used in the specifications and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their combinations.

[0039] An embodiment of the present invention discloses a method for predicting the state of a power terminal device, referring to Figure 1 , including steps S1 - S3:

[0040] S1: Collect the output current data and the output voltage data.

[0041] In the embodiment of the present invention, the preset sampling time is 2 seconds / time. An current sensor and a voltage sensor are installed on the transformer to collect the output current data and the output voltage data simultaneously for one hour, obtaining the output current data and the output voltage data at each sampling time. An analog-to-digital conversion device is used to perform analog-to-digital conversion on the output current data and the output voltage data at each sampling time, obtaining the digital representations of the two types of data.

[0042] S2: According to the inverse proportional relationship between the output current data and the output voltage data, obtain the first weight coefficient of each output current data; according to the surrounding data segments of each output voltage data and the first weight coefficient of each output current data, obtain the second weight coefficient of each output current data.

[0043] It should be noted that when using the HTFE algorithm to predict the current output current data, first, the error between the predicted value and the true value of the output current data at the previous sampling time is multiplied by the error factor to obtain the prediction error of the current output current data. Then, according to the prediction error of the current output current data and in combination with the HTFE algorithm, the predicted value of the current output current data is obtained to judge the fault of the current transformer. However, due to the existence of electromagnetic interference, there will be some noise data in the historical data. Then, when there is potential noise interference in the historical data, the method of only relying on the traditional HTFE algorithm to obtain the prediction error of the current output current data and then obtaining the predicted value of the current output current data becomes inaccurate, thus affecting the accuracy of the fault judgment of the transformer.

[0044] Therefore, the present invention proposes a method and system for predicting the state of a power terminal device. To avoid the influence of noise data, first, analyze the numerical change characteristics of the output current data before the current output current data to obtain the second weight coefficient of each output current data. Then, according to the second weight coefficient of each output current data, correct the error between the predicted value and the true value of each output current data to obtain the prediction error of the current output current data. Then, according to the prediction error of the current output current data and in combination with the HTFE algorithm, the predicted value of the current output current data is obtained to judge the fault of the current transformer, avoiding the influence of noise data on the prediction result.

[0045] Step S2 includes steps S20 - S21, which are specifically as follows:

[0046] S20: According to the inverse proportional relationship between the output current data and the output voltage data, obtain the first weight coefficient of each output current data.

[0047] It should be noted that by analyzing the numerical change characteristics of the output current data and the output voltage data, the first weight coefficient of each output current data is obtained. Since in the transformer working scenario, when the power at the input end is equal to the power at the output end, it can be known that if there is a change between the input current data and the output voltage data, the two can only be in an inverse proportional relationship. This is because power = voltage × current. Then when analyzing the first weight coefficient of the output current data, the more prominent the value of the output current data and the less tight the inverse proportional relationship between the output current data and the output voltage data, it can be explained that the probability of the output current data belonging to noise data is greater, then the reference value of the error of the output current data is lower, and its corresponding first weight coefficient is smaller.

[0048] In the embodiment of the present invention, a preset number of reference data R is set. The R output current data before each output current data are used as the reference data for each output current data; the R output current data before each output voltage data are used as the reference data for each output voltage data; in the embodiment of the present invention, the preset number of reference data R = 100. In other embodiments, the implementer can preset the value of the number of reference data R according to the specific implementation manner.

[0049] Obtain the first weight coefficient of each output current data:

[0050] ;

[0051] In the formula, represents the first weight coefficient of the nth output current data; R represents the number of reference data; represents the absolute value of the difference between the value of the nth output current data and the value of its rth reference data; represents the number of data among all the reference data of the nth output current data whose values are greater than the value of the nth output current data; represents the number of data among all the reference data of the output voltage data corresponding to the sampling moment of the nth output current data whose values are greater than the value of the output voltage data; exp() represents the exponential function with the natural constant as the base; represents a hyperparameter to avoid the value of being equal to 0; represents the sum of the differences between the nth output current data and the values of its reference output current data. The larger this value is, the more prominent the numerical performance of the nth output current data is, then the probability of the output current data belonging to noise data is greater, the reference value of the error value at the corresponding output current data will be lower, and its corresponding first weight coefficient will be smaller; It reflects the tightness of the inverse proportional relationship between the nth output current data and its corresponding output voltage data. The smaller this value, the greater the tightness of the inverse proportional relationship between the nth output current data and the output voltage data. Then, the larger this value, the more it can further prove that the probability of the nth output current data belonging to noise data is greater, and the reference value of the error value corresponding to this output current data will be lower, and its corresponding first weight coefficient will be smaller.

[0052] S21: Obtain the second weight coefficient of each output current data according to the surrounding data segments of each output voltage data and the first weight coefficient of each output current data.

[0053] It should be noted that the first weight coefficient of each output current data has been obtained. The analysis of this index is based on the numerical change characteristics of the output current data and the output voltage data at the same sampling moment. However, in the actual scenario, the output voltage data will also generate noise data due to electromagnetic interference. Then, when any output voltage data belongs to noise data, the first weight coefficient of the output current data corresponding to the sampling moment of this output voltage data will become inaccurate, and it cannot accurately reflect the reference value of the error value of the output current data; Therefore, it is necessary to obtain the surrounding data segments of each output voltage data, analyze the data change characteristics in the surrounding data segments of each output voltage data, and optimize the first weight coefficient of each output current data to obtain the second weight coefficient of each output current data. Among them, if the data change characteristics in the surrounding data segments of any output voltage data are smoother, it can indicate that the credibility of this output voltage data belonging to normal data rather than noise data is greater, then the credibility of the first weight coefficient of the output current data corresponding to the sampling moment of this output voltage data will also be greater, and its corresponding second weight coefficient will also be greater.

[0054] In the embodiment of the present invention, the reference data of each output voltage data and the data segment formed by each output voltage data are denoted as the surrounding data segments of each output voltage data.

[0055] Obtain the second weight coefficient of each output current data:

[0056] ;

[0057] In the formula, represents the second weight coefficient of the nth output current data; represents the first weight coefficient of the nth output current data; exp() represents the exponential function with the natural constant as the base; norm() represents the normalization function; represents the number of extreme points in the surrounding data segment of the output voltage data corresponding to the sampling moment of the nth output current data; The range in the surrounding data segment of the output voltage data corresponding to the sampling moment of the nth output current data; The larger the value of, the greater the reference value of the error value of the nth output current data, and then the corresponding second weight coefficient will also be larger; The smaller the value of, the less the degree of numerical change in the surrounding data segment of the output voltage data corresponding to the sampling moment of the nth output current data, and the greater the degree of smoothness of the surrounding data segment of the output voltage data; The smaller the value of, it can also indicate that the degree of smoothness of the surrounding data segment of the output voltage data corresponding to the sampling moment of the nth output current data is greater; Represents the degree of smoothness of the surrounding data segment of the output voltage data corresponding to the sampling moment of the nth output current data. The smaller its value, the greater the credibility that the corresponding output voltage data belongs to normal data rather than noise data. Then the credibility of the first weight coefficient of this output current data will also be greater, and the corresponding second weight coefficient will also be greater.

[0058] S3: Obtain the prediction error of the current output current data according to the second weight coefficient of each output current data; obtain the predicted value of the current output current data according to the prediction error of the current output current data, and judge the abnormal situation of the current output current data according to the predicted value of the current output current data, and then judge the abnormality of the transformer.

[0059] Step S3 includes steps S30 - S31, specifically as follows:

[0060] S30: Obtain the prediction error of the current output current data according to the second weight coefficient of each output current data.

[0061] It should be noted that the second weight coefficient of each output current data is obtained. Therefore, the prediction error of the current output current data can be obtained according to the second weight coefficient of each output current data and the error value of each output current data. If the second weight coefficient of any output current data is larger, it means that the error value of this output current data should be more referenced.

[0062] In the embodiment of the present invention, the M output current data before the sampling moment of the current output current data are recorded as the historical data of the current output current data; the predicted value of each historical data of the current output current data is obtained according to the HTFE algorithm, and the absolute value of the difference between the predicted value and the actual value of each historical data of the current output current data is used as the error value of each historical data of the current output current data; in the embodiment of the present invention, the preset number of historical data M = 100. In other embodiments, the implementer can preset the value of the number of historical data M according to the specific implementation manner.

[0063] Obtain the prediction error of the current output current data:

[0064] ;

[0065] In the formula, represents the prediction error of the current output current data; M represents the number of historical data of the current output current data; represents the error value of the m-th historical data of the current output current data; represents the second weight coefficient of the m-th historical data of the current output current data; represents the sum of the second weight coefficients of all historical data of the current output current data.

[0066] S31: Obtain the predicted value of the current output current data according to the prediction error of the current output current data, and judge the abnormal situation of the current output current data according to the predicted value of the current output current data, so as to judge the abnormality of the transformer.

[0067] It should be noted that after obtaining the prediction error of the current output current data, it is necessary to obtain the predicted value of the current output current data according to the prediction error of the current output current data and in combination with the HTFE algorithm.

[0068] In the embodiment of the present invention, the error factor of the preset prediction interval is T. According to the prediction error of the current output current data, the HTFE algorithm is used to calculate the predicted value of the current output current data, and the predicted value of the current output current data is obtained; in the embodiment of the present invention, the error factor T of the preset prediction interval = 0.4. In other embodiments, the implementer can preset the value of T according to the specific implementation situation.

[0069] Preset = 0.05. In other embodiments, the implementer can preset the value according to the specific implementation situation. If the absolute value of the difference between the predicted value and the true value of the current output current data is greater than the product of the predicted value of the current output current data and , it is considered that the current output current data is abnormal data. If the number of abnormal data exceeds 3 within one minute after the sampling moment corresponding to the current output current data, the transformer issues a warning at this time.

[0070] The embodiment of the present invention also discloses a state prediction system for a power terminal device, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a state prediction method for a power terminal device according to the present invention is implemented.

[0071] The above system further includes other components well-known to those skilled in the art, such as a communication bus and a communication interface. Their settings and functions are known in the art, so they will not be elaborated here.

[0072] In the present invention, the aforementioned memory can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, device, or apparatus. For example, a computer-readable storage medium can be any suitable magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory, a dynamic random access memory, a static random access memory, an enhanced dynamic random access memory, a high-bandwidth memory, a hybrid memory cube, etc., or any other medium that can be used to store the required information and can be accessed by an application program, a module, or both. Any such computer storage medium can be part of the device or accessible or connectable to the device.

[0073] Although this specification has shown and described multiple embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and idea of the present invention. It should be understood that various alternatives to the embodiments of the present invention described herein can be adopted in the practice of the present invention.

[0074] The above are all preferred embodiments of the present invention, and the protection scope of the present invention is not limited thereby. Therefore, all equivalent changes made according to the structure, shape, and principle of the present invention shall be covered within the protection scope of the present invention.

Claims

1. A method for predicting the state of a power terminal device, characterized in that, Including the steps: Collect the output current data and the output voltage data; obtain the second weight coefficient of each output current data , represents the second weight coefficient of the nth output current data; represents the first weight coefficient of the nth output current data; exp() represents the exponential function with the natural constant as the base; represents the number of extreme points in the surrounding data segment of the output voltage data corresponding to the sampling moment of the nth output current data; represents the range in the surrounding data segment of the output voltage data corresponding to the sampling moment of the nth output current data; norm() represents the normalization function; Obtain the prediction error of the current output current data , where M represents the number of historical data of the current output current data; represents the error value of the m-th historical data of the current output current data; represents the second weight coefficient of the m-th historical data of the current output current data; represents the sum of the second weight coefficients of all historical data of the current output current data; Using the HTFE algorithm, combining with the prediction error of the current output current data, to obtain the predicted value of the current output current data; according to the predicted value, to judge the abnormal situation of the current output current data, and further to judge whether the transformer is abnormal.

2. The state prediction method of a power terminal device according to claim 1, wherein, The obtaining of the first weight coefficient of the output current data includes: ； Wherein, represents the first weight coefficient of the n-th output current data; R represents the number of reference data; represents the absolute value of the difference between the value of the n-th output current data and the value of its r-th reference data; represents the number of data whose values of all reference data of the n-th output current data are greater than the value of the n-th output current data; represents the number of data whose values of all reference data of the output voltage data corresponding to the sampling moment of the n-th output current data are greater than the value of the output voltage data; exp() represents the exponential function with the natural constant as the base; represents the hyperparameter; || represents the absolute value symbol.

3. The state prediction method of a power terminal device according to claim 2, characterized in that, The obtaining of the reference data includes: Presetting the number of reference data as R, taking the R output current data before each output current data as the reference data of each output current data; taking the R output current data before each output voltage data as the reference data of each output voltage data.

4. A method for predicting the state of a power terminal device according to claim 1, characterized in that The obtaining of the surrounding data segment of the output voltage data includes: Denoting the data segment formed by the reference data of each output voltage data and each output voltage data as the surrounding data segment of each output voltage data.

5. A method for predicting the state of a power terminal device according to claim 1, characterized in that, The obtaining of the historical data of the current output current data includes: Taking the M output current data before the sampling moment of the current output current data as the historical data of the current output current data.

6. The state prediction method of a power terminal device according to claim 1, characterized in that, The obtaining of the error value of the historical data includes: According to the HTFE algorithm, obtaining the predicted value of each historical data of the current output current data, and taking the absolute value of the difference between the predicted value and the actual value of each historical data of the current output current data as the error value of each historical data of the current output current data.

7. A method for predicting the state of a power terminal device according to claim 1, characterized in that, The obtaining of the predicted value of the current output current data includes: Presetting the error factor of the prediction interval as T, and according to the prediction error of the current output current data, using the HTFE algorithm to calculate the predicted value of the current output current data, so as to obtain the predicted value of the current output current data.

8. A method for predicting the state of a power terminal device according to claim 1, characterized in that, The judging of the abnormal situation of the current output current data according to the predicted value, and further judging whether the transformer is abnormal, includes: Preset If the absolute value of the difference between the predicted value and the true value of the current output current data is greater than the product of the predicted value of the current output current data and , the current output current data is considered abnormal data. If the number of abnormal data within one minute after the sampling moment corresponding to the current output current data exceeds 3, the transformer issues a warning.

9. A state prediction system for a power terminal device, characterized in that, Including: A processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, it realizes a state prediction method of a power terminal device according to any one of claims 1-8.

Images