A distribution network early warning method considering distributed processing information

Abstract

The present invention discloses a distribution network early warning method that considers distributed processing information, which includes the following steps: uniformly collecting and preprocessing power grid terminal data; detecting whether the collected information is qualified, making an abnormality judgment if qualified, and deleting it if unqualified; extracting and filtering the qualified information based on recursive deep learning measures; transmitting and presenting the finally obtained information to a control center, and then uniformly organizing and storing the data in the background, using input fixed values ​​to distinguish information; the present invention has high economic benefits and a safe distribution network environment; can monitor in real time; reduces the risk of data blocking and reduces interference data.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network early warning method optimization, and in particular to a distribution network early warning method considering distributed processing information. Background Art

[0002] Due to the complex structure of distribution networks and the need to meet the power supply requirements of a wide range of end users, and the continuous increase in per capita electricity consumption, the distribution network's power load has also increased accordingly, increasing the probability of abnormalities in the distribution network. Given the rapidly advancing technological landscape, the widespread adoption of intelligent distribution networks is a current development goal. Currently, most distribution networks have installed anomaly monitoring equipment, which can provide early warning signals of abnormal phenomena to a certain extent. If multiple and complex anomalies occur in the distribution network, a large amount of abnormal data will flow simultaneously into the distribution network operation center, causing momentary data congestion. Furthermore, the generated data is often mixed with noise, leaving only a small amount of data to determine the location of the anomaly and its type. Summary of the Invention

[0003] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0004] In view of the problems existing in the prior art, the present invention is proposed.

[0005] Therefore, the technical problem to be solved by the present invention is that if multiple and complex abnormal phenomena occur in the distribution network, a lot of abnormal data will flow into the distribution network operation center at the same time, causing instantaneous data congestion. In addition, most of the data generated are mixed with interference content, and only a small amount of data can be used to determine the location of the abnormal point and judge the type of abnormality.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a distribution network early warning method considering distributed processing information, which includes unified collection and preprocessing of power grid terminal data; detecting whether the collected information is qualified, and making an abnormality judgment if qualified, and deleting it if unqualified; extracting and filtering qualified information based on recursive deep learning measures; transmitting and presenting the finally obtained information to the control center, and then uniformly organizing and storing the data in the background, and using input values ​​to judge information.

[0007] As a preferred solution of the distribution network early warning method considering distributed processing information of the present invention, wherein: the unified collection of power grid terminal data includes:

[0008] Information collection and processing uses the detection circuit in the collection device to collect initial information about the power grid and convert the analog signal quantity for centralized processing by the core control chip of the collection device.

[0009] As a preferred solution of the distribution network early warning method considering distributed processing information of the present invention, wherein: the preprocessing includes:

[0010] The initial information is uniformly transmitted to the monitoring unit, which identifies abnormal data and interference data and records the distribution of information in real time; if unreasonable interference information is detected, it is deleted; if abnormal data is detected, all data are classified and abnormality judgment is performed.

[0011] As a preferred solution of the distribution network early warning method considering distributed information processing described in the present invention, the abnormality judgment includes: the distribution network information processed by the information monitoring unit is transmitted to the abnormality judgment unit in an interface intercommunication manner, and the abnormality judgment unit extracts and filters qualified information based on recursive deep learning measures; the extraction level obtains key factors of the input information, and the filtering level screens the key factors and filters out redundant content.

[0012] As a preferred solution of the distribution network early warning method considering distributed processing information described in the present invention, the abnormality determination unit adopts a distributed processing method to achieve multi-execution route coordination and distribution network analysis.

[0013] As a preferred solution of the distribution network early warning method considering distributed processing information of the present invention, the extraction level and the filtering level cooperate to enhance the parameter fitting level of the distributed information processing device, and finally obtain the corresponding relationship of abnormal data:

[0014] j(r)=(i×θ)r

[0015] In the formula, j(r) is the terminal information that has completed the classification operation; i is the front-end information that has completed the impurity removal and preliminary processing; θ is the mapping of the recursive deep learning measure; r is the terminal information.

[0016] As a preferred solution of the distribution network early warning method considering distributed processing information of the present invention, the detailed detection of the property conditions of the terminal information result is carried out through a metric analysis method:

[0017]

[0018] In the formula: ln j(r) is the metric analysis value of the terminal result; m is the total number of abnormal types; x is the abnormal type number; f x is the probability of the xth type of exception occurring; i is the front-end information that has completed impurity removal and preliminary processing.

[0019] As a preferred solution of the distribution network early warning method considering distributed processing information of the present invention, wherein: according to the four mapping situations existing in the abnormality analysis method, (1,0,0,0), (0,1,0,0), (0,0,1,0), (0,0,0,1) are obtained;

[0020] The four mapping levels of the recursive deep learning method are matched with the first and second round parsing contents of (1,0,0), (0,1,0), (0,0,1), (0,0,0);

[0021] Get the mapping content of all levels according to the standard. The mapping differences can explain various types of exceptions.

[0022] As a preferred solution of the distribution network early warning method considering distributed processing information described in the present invention, assuming that the metric analysis value ln j(r) of the terminal result is used as the classification standard, the front-end abnormal information group I can be further divided into multiple information groups. The initial classification method of abnormal information is:

[0023]

[0024] In the formula: ln I(r) is the metric analysis value of the front-end information; m is the total number of abnormal types; x is the abnormal type number; i x is the front-end information of the xth abnormal type; I is the front-end abnormal information group; and ln j(r) is the metric analysis value of the terminal result.

[0025] As a preferred solution of the distribution network early warning method considering distributed processing information of the present invention, wherein: content gain is obtained according to the measurement analysis method and the classification method of abnormal information;

[0026] The content gain is a value obtained by subtracting In I(r) from In j(r), and the type of abnormality is determined by the difference in the content gain.

[0027] The beneficial effects of the present invention are as follows: the present invention has high economic benefits and a safe distribution network environment; it can monitor in real time; it reduces the risk of data blocking and reduces interference data. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0029] Figure 1 This is a system flow chart in the first embodiment.

[0030] Figure 2 This is a diagram showing the voltage data fluctuation in the second embodiment.

[0031] Figure 3 This is a diagram showing the current data fluctuation in the second embodiment. DETAILED DESCRIPTION

[0032] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0033] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0034] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0035] Example 1

[0036] Reference Figure 1 , which is the first embodiment of the present invention, provides a distribution network early warning method considering distributed processing information, including the following four steps:

[0037] Collect and pre-process the power grid terminal data in a unified manner;

[0038] Check whether the collected information is qualified. If qualified, make an abnormal judgment; if unqualified, delete it;

[0039] Extracting and filtering qualified information based on recursive deep learning measures;

[0040] The final information is transmitted to the control center, and then the data is uniformly organized and stored in the background, and the input value is used to judge the information.

[0041] Furthermore, information collection and processing utilizes the detection circuit in the collection device 100 to collect the initial information of the power grid and convert the analog signal quantity for centralized processing by the core control chip of the collection device 100; the collection device 100 is installed at the distribution network terminal and connected to the distribution network. The collection device 100 may include a voltage signal collection device 101, a current signal collection device 102 and a collection device that can be provided with other power grid signals, which contains a signal detection circuit, a signal converter and a signal processing chip; first, the detection circuit collects the voltage, current or other signal values ​​of the distribution network terminal, transmits them to the signal converter to convert the analog quantity into digital quantity, and then uniformly enters the signal processing chip for storage and integration.

[0042] Preprocessing involves transmitting initial information acquired by wireless communication devices throughout the distribution network to a monitoring unit 200, typically a remote control platform. This unit identifies abnormal and interfering data and records its distribution in real time. Programs within the processing center control platform perform further checks, deleting any unreasonable interference information. If abnormal data is detected, all data is categorized and anomaly determinations are made. This unit immediately performs preprocessing to remove unnecessary content and then performs simple categorization. Any significantly unreasonable information detected is promptly deleted.

[0043] The abnormality judgment includes an abnormality determination unit 300. The distribution network information processed by the information monitoring unit 200 is transmitted to the abnormality determination unit 300 in an interface intercommunication manner. The abnormality determination unit 300 extracts and filters qualified information based on recursive deep learning measures;

[0044] The extraction stage extracts key factors from the input information, while the filtering stage screens these key factors and removes redundant content. The anomaly identification unit 300 utilizes a distributed processing approach to coordinate multiple execution routes and analyze the distribution network. The interface utilizes a conventional 32-channel mode and introduces public information concepts into the distribution network and monitoring unit 200, acquiring all essential parameters for anomaly identification and analysis, as well as the actual distribution network status.

[0045] Considering the complex layout of the distribution network, the monitoring unit 200 will obtain a considerable amount of information, so a large number of parameter analysis tools must be used to achieve the goal. The distributed information processing device is developed based on Apache server software, combining a distributed structure and a large-scale parallel analysis unit. When any device uses a distributed structure for recording, there are not many requirements for its internal structure, which can effectively reduce the cost of purchasing distribution network devices. The large-scale parallel analysis unit is the key point of the distributed information processing device, which follows the node project arrangement rules. After distributed information processing, the original structure analysis, grid anomaly information acquisition, parameter recording and calculation results are achieved by the collaboration of multiple tools. The specific arrangement can change the accuracy of the device's abnormal information testing and early warning.

[0046] Furthermore, the monitoring unit 200 and the abnormality identification unit 300 are respectively connected to the user and device communication unit 400 for communication connection. The information monitoring inside the user and device communication unit 400 is connected to the monitoring unit 200 and the abnormality identification unit 300, which plays the role of circulating information in each stage of the analysis operation of the monitoring unit 200. If the device issues an early warning signal, the user and device communication unit 400 will also output an early warning message to inform the back-end dispatch workers. When the early warning signal is issued, the terminal communication equipment of the distribution network at the abnormal location transmits a large amount of abnormal data to the background, causing instantaneous information congestion. The back-end dispatch workers sort all the information through the user and device communication unit, and then combine it with the information processing function of the large-scale parallel analysis unit to reduce invalid content and determine the specific location of the distribution network failure problem.

[0047] The information transmission and presentation to the control center includes: the user and device communication unit 400 transmits the abnormal data to the result presentation unit 500 at the field end for projecting the abnormal information of the distribution network data.

[0048] This unit plays the role of presenting the final abnormal location and type. After classification inspection and obtaining key factors, the distribution network abnormality information library can determine the distribution network area where abnormal phenomena exist. The phenomenon that the device will produce an error accident cannot be avoided. In order to prevent the distributed data analysis content from being affected by the device being turned on again during a period without abnormal phenomena, based on the user and device communication unit 400, it is possible to pre-set abnormal judgment. Now, fixed distributed information analysis content is preset according to different data, and the abnormal judgment effect of the device is determined at the same time to prevent the phenomenon of multiple or erroneous issuance of emergency signals.

[0049] Example 2

[0050] Reference Figure 2 、 3 This is the second embodiment of the present invention. Based on the previous embodiment, in an environment with a large number of parameters, it is necessary to consider the key factors of a large amount of initial information, analyze the abnormal content, and obtain meaningful abnormal data in the abnormal information database. When using a distributed information processing device, the hierarchical part of the recursive deep learning method is mainly the extraction level and the filtering level. The extraction level and the filtering level have different effects during the operation of abnormality. The role of the extraction level is to obtain key factors of the input information, and the role of the filtering level is to screen the key factors and filter out redundant content.

[0051] The extraction stage and the filtering stage operate in coordination to enhance the parameter fitting level of the distributed information processing device.

[0052] Finally, the corresponding relationship of abnormal data can be obtained:

[0053] j(r)=(i×θ)r

[0054] In the formula, j(r) is the terminal information that has completed the classification operation; i is the front-end information that has completed the impurity removal and preliminary processing; θ is the mapping of the recursive deep learning measure; r is the terminal information.

[0055] According to the four mapping situations of the exception analysis method, (1,0,0,0), (0,1,0,0), (0,0,1,0), and (0,0,0,1) are obtained; the four mapping levels of the recursive deep learning method, the matching first-level and second-round analysis contents are (1,0,0), (0,1,0), (0,0,1), and (0,0,0); according to the standard, the mapping contents of all levels are obtained, and the different mappings can explain various types of exceptions.

[0056] The abnormal types in the distribution network are set to include circuit break abnormalities, short circuit abnormalities, substation abnormalities, etc. The abnormality mapping results can be obtained through recursive deep learning methods. The terminal information results of all levels are shown in the following table.

[0057] Table 1. Recursive deep learning abnormal information results

[0058]

[0059] The optimal property selection measure is used as the classification condition for the abnormal type. The most effective property classification rule only applies to one type. The detailed detection of the property conditions of the terminal information results is carried out through the measurement analysis method:

[0060]

[0061] Where: lnj(r) is the metric analysis value of the terminal result; m is the total number of abnormal types; x is the abnormal type number; f x is the probability of the xth type of exception occurring; i is the front-end information that has completed impurity removal and preliminary processing.

[0062] Assuming that the metric analysis value ln j(r) of the terminal result is used as the classification standard, the front-end abnormal information group I can be further divided into multiple information groups. The initial classification method of abnormal information is:

[0063]

[0064] In the formula: ln I(r) is the metric analysis value of the front-end information; m is the total number of abnormal types; x is the abnormal type number; i x is the front-end information of the xth abnormal type; I is the front-end abnormal information group; and ln j(r) is the metric analysis value of the terminal result.

[0065] Furthermore, the value obtained by subtracting ln I(r) from ln j(r) is the content gain. If the difference between these values ​​reaches its minimum, optimal classification is achieved. Assuming that a set of information contains multiple types of anomalies, combining the content gain allows for highly accurate identification and selection. Distributed information processing devices are used to perform initial anomaly removal and preliminary processing, preventing errors caused by excessive amounts of information. Multi-level recursive deep learning methods significantly reduce the time required to determine mapping conditions and modify method data. The final anomaly type is determined by combining metric analysis values, and the difference in content gain indicates the type of anomaly, which helps improve the accuracy of information selection.

[0066] During the device performance test, the monitoring device conceived by the above method will first randomly select 3 sets of voltage and current data, and the time intervals of the data selection will be different, as shown in the following example: Figure 2 、 3 .

[0067] like Figure 2 and 3 The three time intervals are inconsistent: x is 0.1s, y is 0.3s, and z is 0.5s. Observation shows that the voltage and current signal fluctuations are very regular, with no significant difference in peak value and no sudden change, indicating that this device has a good filtering level.

[0068] Considering that the recursive deep learning method used by the distributed information processing device is of great help in solving the problem of removing and filtering abnormal voltage and current data in the distribution network, and can comprehensively and accurately obtain the key factors of abnormal information, within the distribution network abnormal voltage and current information group, a random selection of points is used to determine the differences between the actual data and the monitored data, and a comparison is made in combination with traditional information processing measures, as shown in the following table:

[0069] Table 2 Abnormal voltage data comparison results

[0070]

[0071]

[0072] Table 3 Abnormal current data comparison results

[0073]

[0074] The table above shows that the deviation of abnormal data from distributed processing is significantly smaller than that from traditional processing measures. This indicates that the distributed processing information distribution network early warning device described in this article is more stable in handling abnormal information and more efficient in issuing alarms than traditional devices.

[0075] It is important to note that the construction and arrangement of the present application shown in a number of different exemplary embodiments are merely illustrative. Although only a few embodiments are described in detail in this disclosure, it should be readily understood by those who refer to this disclosure that many modifications are possible (e.g., the size, scale, structure, shape and proportion of various elements, as well as parameter values ​​(e.g., temperature, pressure, etc.), mounting arrangements, use of materials, color, directional changes, etc.) without departing substantially from the novel teachings and advantages of the subject matter described in this application. For example, an element shown as integrally formed may be composed of multiple parts or elements, the position of the element may be inverted or otherwise changed, and the nature or number or position of the discrete elements may be altered or changed. Therefore, all such modifications are intended to be included within the scope of the present invention. The order or sequence of any process or method steps may be changed or reordered according to alternative embodiments. In the claims, any "means plus function" clause is intended to cover the structure described herein that performs the function, and is not only structurally equivalent but also equivalent structures. Other replacements, modifications, changes, and omissions may be made in the design, operating conditions, and arrangement of the exemplary embodiments without departing from the scope of the present invention. Therefore, the invention is not limited to the specific embodiments, but extends to various modifications that still fall within the scope of the appended claims.

[0076] Additionally, in order to provide a concise description of exemplary embodiments, all features of an actual embodiment (ie, those features that are not relevant to the best mode presently contemplated for carrying out the invention or those that are not relevant to implementing the invention) may not be described.

[0077] It will be appreciated that in the development of any actual embodiment, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but will, for those of ordinary skill having the benefit of this disclosure, be a routine undertaking of design, fabrication, and production without undue experimentation.

[0078] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A distribution network early warning method considering distributed processing information, characterized by: include, Collect and pre-process the power grid terminal data in a unified manner; Check whether the collected information is qualified. If qualified, make an abnormal judgment; if unqualified, delete it; The abnormality judgment includes: The distribution network information processed by the information monitoring unit (200) is transmitted to the abnormality determination unit (300) in an interface intercommunication manner, and the abnormality determination unit (300) extracts and filters qualified information based on a recursive deep learning measure; The extraction level obtains key factors from the input information, and the filtering level screens the key factors and filters out redundant content; The final information is transmitted to the control center, and then the data is uniformly organized and stored in the background, and the input value is used to judge the information; The extraction and filtering levels work together to enhance the parameter fitting level of the distributed information processing device, and ultimately obtain the corresponding relationship of the abnormal data: ， In the formula: Information about terminals that have completed classification operations; Front-end information that has been cleaned and preliminarily processed; Mapping of recursive deep learning measures; Terminal information; According to the four mapping situations of the exception analysis method, we can get (1,0,0,0), (0,1,0,0), (0,0,1,0), (0,0,0,1); The four mapping levels of the recursive deep learning method are matched with the first and second round parsing contents of (1,0,0), (0,1,0), (0,0,1), (0,0,0); Obtain mapping content at all levels according to the standard. Differences in mapping can explain various types of anomalies. The terminal result measurement analysis value As a classification standard, the front-end abnormal information group I can be further divided into multiple information groups. The initial classification method of abnormal information is: ， In the formula: ; m is the total number of exception types; x is the exception type number; Front-end information for the xth type of exception; I is the front-end abnormal information group; the measurement analysis value of the terminal result ; Obtaining content gain based on the measurement analysis method and the classification method of abnormal information; The content gain is minus The obtained value is used to determine the type of anomaly based on the difference in content gain; The metric analysis calculation formula of the terminal information result is: ， In the formula: ; m is the total number of exception types; x is the exception type number; is the probability of the xth abnormal type occurring; This is the front-end information that has been cleaned and preliminarily processed.

2. The distribution network early warning method considering distributed processing information according to claim 1, characterized in that: The pretreatment includes: The initial information is uniformly transmitted to the monitoring unit (200), and the monitoring unit (200) identifies abnormal data and interference data and records the distribution of the information in real time; If unreasonable interference information is detected, it will be deleted; If abnormal data is detected, all data will be classified and abnormal judgment will be made.

3. The distribution network early warning method considering distributed processing information according to claim 2, characterized in that: The unified collection of power grid terminal data includes: Information collection and processing: using the detection circuit in the collection device (100) to collect initial information of the power grid and converting the analog signal quantity for centralized processing by the core control chip of the collection device (100).

4. The distribution network early warning method considering distributed processing information according to claim 3, characterized in that: The abnormality determination unit (300) adopts a distributed processing method to achieve multi-execution route coordination and distribution network analysis.

Images