5G network maintenance methods

Abstract

The present invention discloses a 5G network maintenance method, comprising: S1: obtaining operational data of each 5G base station in a target area; S2: inputting the operational data into a pre-established fault prediction model to output a fault probability; S3: grouping the 5G base stations according to their fault probability ranges; and S4: obtaining component information for components in the 5G base stations in each group, identifying and marking abnormal components. The present invention can more accurately identify abnormal components, enabling targeted maintenance.

Description

Technical Field

[0001] The present invention relates to the technical field related to network maintenance. More specifically, the present invention relates to a 5G network maintenance method. Background Art

[0002] During long-term operation, 5G network base stations may experience various faults, requiring real-time maintenance to ensure network stability. Existing technologies typically rely on regular manual inspections, including cleaning equipment, checking circuits, and replacing components. This approach relies heavily on human experience and responsibility, is inherently unfocused, lacks targeted maintenance, and is costly. Therefore, it is necessary to design a technical solution that can, to some extent, overcome these shortcomings. Summary of the Invention

[0003] One object of the present invention is to provide a 5G network maintenance method that can more accurately determine abnormal components and achieve targeted maintenance.

[0004] In order to achieve these objects and other advantages of the present invention, according to one aspect of the present invention, the present invention provides a 5G network maintenance method, including: S1: obtaining operating data of each 5G base station in the target area; S2: inputting the operating data into a pre-established fault prediction model and outputting the fault probability; S3: grouping each of the 5G base stations according to the fault probability range; S4: obtaining component information of the components in the 5G base stations in each group, determining abnormal components, and marking them.

[0005] Furthermore, the method for establishing the fault prediction model includes: obtaining the historical operation data and label value of the 5G base station, inputting the data into a neural network model for training, and obtaining the fault prediction model; wherein, if the 5G base station fails, the label value is 0, and if the 5G base station is normal, the label value is 1.

[0006] Furthermore, the operating data includes at least signal strength, temperature, receiving power, transmitting power, and voltage.

[0007] Furthermore, the component information includes at least the manufacturer's name, production batch, installer's name, and usage time range.

[0008] Furthermore, each of the 5G base stations is divided into a high probability group and a low probability group according to the fault probability range; the fault score of each component in the 5G base station in the high probability group is calculated, and the fault score is the weighted sum of the fault scores of each component information, and the fault score is the quotient of the proportion of the component information in the high probability group and the proportion of the component information in the low probability group; components with fault scores greater than a predetermined threshold are selected and marked as abnormal components.

[0009] Furthermore, the components that actually fail and the components that actually do not fail in the high-probability group within a predetermined time period are obtained, and the weights of the component information corresponding to the components are corrected.

[0010] According to another aspect of the present invention, a 5G network maintenance device is also provided, including: an acquisition module for acquiring operating data of each 5G base station in the target area; a fault probability calculation module for inputting the operating data into a pre-established fault prediction model and outputting a fault probability; a grouping module for grouping each of the 5G base stations according to a fault probability range; and a marking module for acquiring component information of components in the 5G base stations in each group, determining abnormal components, and marking them.

[0011] Furthermore, the method for establishing the fault prediction model includes: obtaining the historical operation data and label value of the 5G base station, inputting the data into a neural network model for training, and obtaining the fault prediction model; wherein, if the 5G base station fails, the label value is 0, and if the 5G base station is normal, the label value is 1.

[0012] Furthermore, the operating data includes at least signal strength, temperature, receiving power, transmitting power, and voltage; and the component information includes at least manufacturer name, production batch, installer name, and usage time range.

[0013] Furthermore, the grouping module is used to divide each of the 5G base stations into a high-probability group and a low-probability group according to the fault probability range; the marking module is used to calculate the fault score of each component in the 5G base station in the high-probability group, and the fault score is the weighted sum of the fault scores of each component information, and the fault score is the quotient of the proportion of the component information in the high-probability group and the proportion of the component information in the low-probability group. The components with the fault score greater than the predetermined threshold are selected and marked as abnormal components; the marking module is also used to obtain the components that actually failed and the components that actually did not fail in the high-probability group within a predetermined time period, and correct the weights of the component information corresponding to the components.

[0014] The present invention has at least the following beneficial effects:

[0015] The present invention obtains the operating data of each 5G base station in the target area, inputs the operating data into a fault prediction model, outputs the fault probability, groups each 5G base station according to the fault probability range, obtains the component information of the components in the 5G base station in each group, determines the abnormal components, and marks them; the present invention predicts the failure probability of 5G base stations and groups the 5G base stations, which can more accurately determine the abnormal components, realize targeted maintenance, improve maintenance effects, and reduce maintenance costs.

[0016] Other advantages, objectives and features of the present invention will be reflected in part from the following description and will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flowchart of an embodiment of the present application. DETAILED DESCRIPTION

[0018] The present invention will be described in further detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.

[0019] It should be understood that terms such as "having," "comprising," and "including" used in the embodiments of this application do not exclude the presence or addition of one or more other elements or combinations thereof. All directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of this application are intended only to explain the relative positional relationships and movement of components in a specific posture. If the specific posture changes, the directional indications will also change accordingly. When an element is referred to as being "fixed to" or "disposed on" another element, it can be directly on the other element or there may be an intervening element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or indirectly connected to the other element through an intervening element. References to "first," "second," etc. in the embodiments of this application are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features designated as "first" or "second" may explicitly or implicitly include at least one of such features.

[0020] It should be noted that the technical solutions between the various embodiments of the present application can be combined with each other, but it must be based on the fact that ordinary technicians in this field can implement it. When the combination of technical solutions is mutually contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0021] like Figure 1 As shown, the embodiment of the present application provides a 5G network maintenance method, including:

[0022] S1: Obtain the operating data of each 5G base station in the target area;

[0023] Exemplarily, the operation data includes various internal and external data generated during the operation of the 5G base station, such as power, current, temperature, etc.;

[0024] For example, the external environments of the 5G base stations in the target area are similar, and are in similar temperature and humidity ranges, to ensure the accuracy of subsequent steps;

[0025] S2: Inputting the operating data into a pre-established fault prediction model and outputting a fault probability;

[0026] Exemplarily, the fault prediction model is determined in advance based on historical data and can be established using statistical methods, machine learning methods, or deep learning methods. After inputting the operating data of the current time or the current time period, the calculated fault probability is output;

[0027] S3: Grouping the 5G base stations according to their failure probability ranges;

[0028] For example, the failure probabilities of 5G base stations are sorted by size and grouped according to the order of size;

[0029] S4: Obtain component information of the components in the 5G base station in each group, determine abnormal components, and mark them;

[0030] In this step, since the component information such as manufacturer, production batch, and usage time used by each 5G base station is different, and abnormalities in these components usually lead to 5G base station failures, this step calculates the frequency of component information in different failure probability groups, obtains the key component information that affects component abnormalities, infers abnormal components, and marks them;

[0031] For example, after marking abnormal components, maintenance personnel are informed to carry out targeted maintenance and monitoring, and the reuse of abnormal components is restricted, thereby reducing the failure probability of 5G base stations;

[0032] By predicting the failure probability of 5G base stations and grouping them, the present invention can more accurately identify abnormal components and related information, making it easier for maintenance personnel to perform targeted maintenance, improve maintenance efficiency, reduce maintenance costs, reduce reliance on manual experience and sense of responsibility, and ensure network stability.

[0033] In another embodiment, the method for establishing the fault prediction model includes: obtaining historical operation data and label values ​​of the 5G base station, inputting the data into a neural network model for training, and obtaining the fault prediction model; wherein, if the 5G base station is faulty, the label value is 0, and if the 5G base station is normal, the label value is 1;

[0034] For example, historical operating data and its fault conditions are collected to establish training and test sets. In the training and test sets, the operating data is normalized to form vectors, and the fault conditions are assigned label values. The training set is input into the neural network, and the trained model is tested with the test set. When the test accuracy is higher than 90%, it is put into use.

[0035] Exemplarily, the neural network adopts a BP neural network, which outputs a fault probability value of [0, 1].

[0036] In another embodiment, the operating data includes at least signal strength, temperature, received power, transmitted power, and voltage; the above operating data can fully reflect the operating status of the 5G base station;

[0037] For example, the signal strength is tested within the same distance, or the signal strength is tested between adjacent 5G base stations, the temperature is measured by a temperature sensor, and the power and voltage are measured by an electric meter.

[0038] In another embodiment, the component information includes at least the manufacturer's name, production batch, installer's name, and usage time range;

[0039] This embodiment can use the above information to find out the factors affecting the failure in the production, installation, use and other links of the component, so as to facilitate targeted processing after the abnormal component is determined.

[0040] In another embodiment, each of the 5G base stations is divided into a high-probability group and a low-probability group according to a fault probability range; a fault score of each component in the 5G base station in the high-probability group is calculated, where the fault score is a weighted sum of the fault scores of each component information, and the fault score is the quotient of the proportion of the component information in the high-probability group and the proportion of the component information in the low-probability group; components with fault scores greater than a predetermined threshold are selected and marked as abnormal components;

[0041] This embodiment calculates the fault score of each component information to obtain the component fault score. The fault score can reflect the abnormality level of all components. Components with higher fault scores are marked as abnormal components to provide a reference for subsequent detection, maintenance, and replacement of components.

[0042] Exemplarily, the predetermined threshold is determined based on a ratio or statistics to ensure that components with a value greater than the predetermined threshold are more likely to be abnormal;

[0043] For example, the manufacturer name, production batch, installer name, and usage time range of the marked abnormal components are monitored closely, and the use of components from that manufacturer and production batch is restricted. The installer's installation process is investigated, and components within that usage time range are considered for replacement.

[0044] For example, the top 20% and bottom 20% 5G base stations are classified into high-probability and low-probability groups, respectively. All components of base stations in the high-probability group require attention, and specific abnormal components are identified by statistical component information.

[0045] Exemplarily, p=λ1*p1+λ2*p2+λ3*p3…, λ1, λ2, and λ3 are weights, p1, p2, and p3 are fault scores, and p is the fault score. For example, if the components of a certain production batch account for 80% in the high-probability group and 10% in the low-probability group, then the fault score p1 corresponding to the component information of the production batch is 8. Similarly, p2 and p3 are calculated to be 3 and 4, respectively, that is, the fault score is 15.

[0046] In another embodiment, components that actually fail and components that actually do not fail in the high probability group within a predetermined time period are obtained, and weights of the component information corresponding to the components are corrected;

[0047] Exemplarily, the initial weights are all equal, such as the initial weights of λ1, λ2, and λ3 are all 1 / 3, or are assigned based on actual investigations. After obtaining the components that actually failed and the components that actually did not fail, the weights of the component information related to the component are adjusted, such as by multiplying by a coefficient. The coefficient can be determined by expert scoring or experience to optimize the weights.

[0048] The embodiment of the present application further provides a 5G network maintenance device, comprising: an acquisition module for acquiring operating data of each 5G base station in a target area; a fault probability calculation module for inputting the operating data into a pre-established fault prediction model and outputting a fault probability; a grouping module for grouping each of the 5G base stations according to a fault probability range; a marking module for acquiring component information of components in the 5G base stations in each group, determining abnormal components, and marking them;

[0049] This embodiment constructs an acquisition module, a fault probability calculation module, a grouping module and a marking module through a software program. The fault probability calculation module is used to predict the failure probability of a 5G base station, the grouping module is used to group the 5G base stations, and the marking module is used to accurately determine abnormal components and related information, so as to facilitate maintenance personnel to implement targeted maintenance, improve maintenance efficiency, reduce maintenance costs, reduce reliance on manual experience and responsibility, and ensure network stability.

[0050] In another embodiment, the method for establishing the fault prediction model includes: obtaining historical operation data and label values ​​of the 5G base station, inputting the data into a neural network model for training, and obtaining the fault prediction model; wherein, if the 5G base station is faulty, the label value is 0, and if the 5G base station is normal, the label value is 1;

[0051] For example, historical operating data and its fault conditions are collected to establish training sets and test sets. In the training sets and test sets, the operating data are normalized to form vectors, and the fault conditions are assigned label values. The training set is input into the neural network, and the trained model is tested with the test set. When the test accuracy is higher than 90%, it is put into use; the neural network uses a BP neural network and outputs a fault probability value of [0,1].

[0052] In another embodiment, the operating data includes at least signal strength, temperature, received power, transmitted power, and voltage; the component information includes at least manufacturer name, production batch, installer name, and usage time range;

[0053] For example, the signal strength is tested within the same distance, or the signal strength is tested between adjacent 5G base stations, the temperature is measured by a temperature sensor, and the power and voltage are measured by an electric meter.

[0054] In another embodiment, the grouping module is used to divide each of the 5G base stations into a high-probability group and a low-probability group according to the fault probability range; the marking module is used to calculate the fault score of each component in the 5G base station in the high-probability group, where the fault score is the weighted sum of the fault scores of each component information, and the fault score is the quotient of the proportion of the component information in the high-probability group and the proportion of the component information in the low-probability group. Components with fault scores greater than a predetermined threshold are selected and marked as abnormal components; the marking module is further used to obtain components that actually failed and components that actually did not fail in the high-probability group within a predetermined time period, and correct the weights of the component information corresponding to the components;

[0055] For example, the top 20% and bottom 20% 5G base stations are classified into a high probability group and a low probability group, respectively;

[0056] Exemplarily, p=λ1*p1+λ2*p2+λ3*p3…, λ1, λ2, and λ3 are weights, p1, p2, and p3 are fault scores, and p is the fault score.

[0057] Exemplarily, the initial values ​​of λ1, λ2, and λ3 are all equal, or are assigned according to actual investigation. After obtaining the components that actually failed and the components that actually did not fail, the weight of the component information related to the component is adjusted, such as multiplying by a coefficient.

[0058] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the description and implementation methods. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A 5G network maintenance method, characterized in that: include: S1: Obtain the operating data of each 5G base station in the target area; S2: Inputting the operating data into a pre-established fault prediction model and outputting a fault probability; S3: Divide the 5G base stations into a high probability group and a low probability group according to the failure probability range; S4: Obtain component information of the components in the 5G base station in each group, determine abnormal components, and mark them; The component information at least includes the manufacturer's name, production batch, installer's name, and usage time range; Calculating a fault score for each component in the 5G base station in the high probability group, where the fault score is a weighted sum of the fault scores of each component information, and the fault score is the quotient of the proportion of the component information in the high probability group and the proportion of the component information in the low probability group; Components with fault scores greater than a predetermined threshold are selected and marked as abnormal components.

2. The 5G network maintenance method according to claim 1, wherein: The method for establishing the fault prediction model includes: The historical operation data and label value of the 5G base station are obtained, and the data are input into the neural network model for training to obtain the fault prediction model; wherein, if the 5G base station fails, the label value is 0, and if the 5G base station is normal, the label value is 1.

3. The 5G network maintenance method according to claim 2, wherein: The operating data includes at least signal strength, temperature, receiving power, transmitting power, and voltage.

4. The 5G network maintenance method according to claim 1, wherein: The components that actually fail and the components that actually do not fail in the high probability group within a predetermined time period are obtained, and the weights of the component information corresponding to the components are corrected.

5. 5G network maintenance device, characterized in that: include: The acquisition module is used to obtain the operating data of each 5G base station in the target area; a fault probability calculation module, configured to input the operating data into a pre-established fault prediction model and output a fault probability; A grouping module is used to divide each of the 5G base stations into a high probability group and a low probability group according to the failure probability range; a marking module, configured to obtain component information of components in the 5G base station in each group, identify abnormal components, and mark them; The operating data includes at least signal strength, temperature, received power, transmitted power, and voltage; the component information includes at least manufacturer name, production batch, installer name, and usage time range; The marking module is used to calculate the fault score of each component in the 5G base station in the high probability group, where the fault score is the weighted sum of the fault scores of each component information, and the fault score is the quotient of the proportion of the component information in the high probability group and the proportion of the component information in the low probability group. The components with the fault score greater than a predetermined threshold are selected and marked as abnormal components; The marking module is further configured to obtain components that actually fail and components that actually do not fail in the high-probability group within a predetermined time period, and to correct weights of the component information corresponding to the components.

6. The 5G network maintenance device according to claim 5, wherein: The method for establishing the fault prediction model includes: The historical operation data and label value of the 5G base station are obtained, and the data are input into the neural network model for training to obtain the fault prediction model; wherein, if the 5G base station fails, the label value is 0, and if the 5G base station is normal, the label value is 1.

Images