Service life prediction system based on 5G low-light base station

Abstract

The invention discloses a service life prediction system based on a 5G low-light base station, and relates to the technical field of charging management, a data acquisition module obtains image data and charging assembly data of each part of a mobile base station low-light charging system, and a cloud processing platform receives the image data and the charging assembly data; and combining the image data with the weight of each part of the low-light charging system of the mobile base station to obtain wear data, carrying out life prediction on the low-light charging system based on the wear data and the charging assembly data, and obtaining and storing life data. Aiming at the difficulty in predicting the service life of the low-light charging system assembly of the mobile base station, a multi-parameter evaluation coding table is established based on wear data and charging assembly data, a deduction value is determined in combination with parameter degradation degrees and weights, and an evaluation rule table is formulated to predict the service life, so that operation and maintenance personnel can conveniently plan, maintain and replace the assembly in advance, and power supply interruption of the base station is avoided; the operation and maintenance efficiency is improved and the operation and maintenance cost is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of charge management, and more particularly to a life prediction system based on a 5G micro-light base station. BACKGROUND

[0002] In the current mobile communication field, the wide coverage of 5G network promotes a substantial increase in the number of mobile base stations. To solve the problem of power supply in some remote or complex terrain areas, a micro photovoltaic charging control system emerges as the times require. It can use solar energy to power mobile base stations and has micro-light charging capability. It can collect light energy and convert it into electric energy under continuous rainy weather, and has the advantages of convenient installation, no road and terrain restrictions, etc. However, the existing micro photovoltaic charging control system has many deficiencies in management and performance evaluation.

[0003] In terms of life prediction, the core components of the micro photovoltaic charging system, such as photovoltaic components, are affected by many factors such as ultraviolet radiation, thermal stress, thermal cycling, and hygrothermal degradation. The life of these components has great uncertainty. At present, there is a lack of effective models and algorithms, and it is difficult to accurately predict the remaining life of photovoltaic components by considering these complex factors. Energy storage devices are also affected by factors such as charge-discharge cycle times and operating temperatures. The existing technology cannot accurately estimate their life, making system operation and maintenance blind and possibly causing base station power outages due to unexpected component failures.

[0004] Therefore, how to predict the life of the micro-light charging system of the base station is a problem that needs to be solved by those skilled in the art. SUMMARY

[0005] Therefore, the present application provides a life prediction system based on a 5G micro-light base station to solve the problems in the background art.

[0006] In order to achieve the above purpose, the present application adopts the following technical solutions:

[0007] A life prediction system based on a 5G micro-light base station, comprising: a cloud processing platform, a mobile base station, and a data acquisition module. The data acquisition module acquires image data and charging component data of each part of the micro-light charging system of the mobile base station. The cloud processing platform receives the image data and charging component data, combines the image data with the weights of each part of the micro-light charging system of the mobile base station, acquires wear data, and performs life prediction of the micro-light charging system based on the wear data and charging component data. Life data is acquired and saved.

[0008] Preferably, the mobile base station comprises: a micro photovoltaic assembly, a charging circuit, a voltage boosting circuit, a local control module, a 5G base station, a storage battery; the micro photovoltaic assembly is connected with the charging circuit to convert solar energy into electric energy; the charging circuit is connected with the storage battery to charge the storage battery; the storage battery is connected with the voltage boosting circuit to store electric energy; the voltage boosting circuit is connected with the 5G base station to boost the direct current output by the storage battery for use by the 5G base station; and the local control module is in communication connection with the charging circuit, the photovoltaic assembly and the voltage boosting circuit to control the charging strategy of the micro light charging system.

[0009] Preferably, the cloud processing platform comprises:

[0010] A weight evaluation module combines the semantics of entities and relationships in the safe operation knowledge graph of the micro light charging system to generate a safe operation knowledge perception path, maps each entity, entity corresponding type and relationship in the knowledge graph to a vector representation using a graph convolution network, sequentially encodes the elements using an LSTM to capture the combined semantics of the entities conditioned on the relationships, merges multiple paths using a pooling layer, and outputs the most significant weight value of the interaction between the given operating condition and the component state;

[0011] A wear calculation module detects the wear area using a YOLOV3 network, then applies a UNet++ network to finely segment the wear area, acquires the precise area and calculates the area pixel area, combines the most significant weight value, and calculates and acquires the wear data;

[0012] A life prediction module establishes a multi-parameter quantity evaluation coding table according to the wear data and the charging assembly data, and performs life evaluation based on the multi-parameter quantity evaluation coding table to acquire life data.

[0013] Preferably, the life prediction module specifically comprises:

[0014] A coding table acquisition unit acquires component parameter classifications corresponding to each charging assembly according to the charging assembly data to form parameter indicators, determines charging assembly parameter evaluation quantities according to the parameter indicators, constructs a multi-parameter quantity evaluation coding table, obtains the degradation degree of each parameter, and further refines the degradation degree according to the wear data;

[0015] A standard acquisition unit obtains single parameter indicator deduction values according to the most significant weight values and the degradation degree in combination with corresponding standard rules, acquires clear parameter evaluation standards, and formulates a multi-parameter indicator rule table;

[0016] An overall evaluation unit then formulates a charging assembly state evaluation rule table according to the deduction of each parameter to perform state evaluation on the charging assembly as a whole and acquire life data.

[0017] Preferably, the component parameters specifically include any or all of micro photovoltaic component power generation efficiency, battery state of health, charging current charging conversion efficiency, and boost circuit output voltage stability.

[0018] Preferably, the multi-parameter evaluation coding table includes charging component parameter classification, parameter state name, and judgment basis.

[0019] Preferably, the overall evaluation unit specifically includes:

[0020]

[0021] wherein W s is the overall score weight coefficient of the charging component under different parameters, E s is the life data of the charging component, s represents the overall score weight coefficient of the charging component under different parameters, Y represents the total deduction score value of each parameter of the entire charging component, Y max represents the total deduction score value of all parameters of the entire charging component, N c represents the fuzzy evaluation set corresponding evaluation score.

[0022] Compared with the prior art, the life prediction system based on a 5G micro light base station provided by the present application can solve the problem of predicting the life of the mobile base station micro light charging system component, establish a multi-parameter evaluation coding table based on wear data and charging component data, determine the deduction score value in combination with the parameter degradation degree and weight, formulate an evaluation rule table to predict the life, and can understand the component state in advance, predict the failure time, facilitate the maintenance personnel to plan maintenance and replace components in advance, avoid power supply interruption of the base station, improve the operation and maintenance efficiency, and reduce the operation and maintenance cost. Key parameters such as micro photovoltaic component power generation efficiency and battery state of health directly affect the system performance and life. The multi-parameter evaluation coding table and the multi-parameter index rule table are constructed, the parameter classification, state name, judgment basis, and deduction score value are clear, a perfect evaluation system is formed, and a standardized and normalized method is provided for system state evaluation and life prediction. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only embodiments of the present application, and those skilled in the art can obtain other drawings according to the provided drawings without creative labor.

[0024] Figure 1 The accompanying drawings provide structural schematic diagrams of the present application. DETAILED DESCRIPTION

[0025] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be clearly and completely described below, obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0026] The embodiment of the present application discloses a life prediction system based on a 5G micro-light base station, as shown in the figure, comprising a cloud processing platform, a mobile base station and a data acquisition module, the data acquisition module acquires image data and charging component data of each part of the micro-light charging system of the mobile base station, the cloud processing platform receives the image data and the charging component data, combines the image data with the weights of each part of the micro-light charging system of the mobile base station, acquires wear data, and performs life prediction of the micro-light charging system based on the wear data and the charging component data, acquires life data and performs saving processing. Figure 1

[0027] In one specific embodiment, the mobile base station comprises a micro photovoltaic component, a charging circuit, a boost circuit, a local control module, a 5G base station and a storage battery; the micro photovoltaic component is connected with the charging circuit to convert solar energy into electric energy; the charging circuit is connected with the storage battery to charge the storage battery; the storage battery is connected with the boost circuit to store electric energy; the boost circuit is connected with the 5G base station to boost the direct current output by the storage battery for use by the 5G base station; and the local control module is in communication connection with the charging circuit, the photovoltaic component and the boost circuit to control the charging strategy of the micro-light charging system.

[0028] In one specific embodiment, the cloud processing platform comprises:

[0029] A weight evaluation module combines the semantics of entities and relationships in the safe operation knowledge graph of the micro-light charging system to generate a safe operation knowledge perception path, maps each entity, entity corresponding type and relationship in the knowledge graph to a vector representation using a graph convolution network, sequentially encodes the elements using LSTM to capture the combined semantics of the entities conditioned on the relationship, merges multiple paths using a pooling layer, and outputs the most weight value of the given operation condition and component state interaction;

[0030] A wear calculation module detects the wear area using a YOLOV3 network on the image data, then applies a UNet++ network to finely segment the wear area, acquires the precise area and calculates the area pixel area, combines the most weight value, and calculates and acquires the wear data;

[0031] A life prediction module establishes a multi-parameter quantity evaluation coding table according to the wear data and the charging component data, and performs life evaluation based on the multi-parameter quantity evaluation coding table to acquire life data. ​

[0032] In one specific embodiment, in the weight evaluation module,

[0033] The semantics of entities and relationships in the knowledge graph for the safe operation of low-light charging systems are combined to generate a knowledge perception path for safe operation:

[0034] The knowledge graph for the low-light charging system is defined as KG = {(h,r,t)|h,t∈ε,r∈R}, where each triple (h,r,t) represents a fact, indicating that there is a relation r between the head entity h and the tail entity t. User interaction data is represented as a bipartite graph. and Let represent the user set and the item set, respectively. The interaction between users and items is represented by a triple τ = (u, interaction, i). t This refers to the operating conditions of the low-light charging system; item i includes all the component state variables of the low-light charging system; M and N refer to the number of operating conditions and component state variables, respectively; and interaction represents the interaction between u and i, that is, there is a relationship between the two.

[0035] Among them, the micro-photovoltaic charging system refers to micro-photovoltaic modules, charging circuits, boost circuits, local control modules, and batteries.

[0036] Triples in a knowledge graph describe the direct or indirect relationship attributes of items. These attributes constitute one or more paths between a given user and item pair, i.e., the knowledge perception path of the micro-light charging system, defined as... Where e1 = u, e L =i, (e1,r1,e l+1 ) is the l-th triple in p, where l represents the number of triples in the path. Under low light intensity conditions, there may be a path between the power generation efficiency of a low-light power generation device and the charging status of an energy storage device, reflecting the influence of light intensity on power generation efficiency, and thus affecting the charging status of the energy storage device.

[0037] Using graph convolutional networks, each entity, its corresponding type, and its relation in the knowledge graph are mapped to a vector representation as follows:

[0038] Given a path p k The graph convolutional network projects the type of each entity (such as operating condition type, component type, state variable type, etc.) and specific value (such as specific light intensity value, battery capacity value at a certain moment, etc.) onto two independent low-dimensional embedding vectors. and In this approach, the semantics of relationships are integrated into path representation learning, where d is the size of the embedding vector. This transforms the complex information in the knowledge graph into a vector form that is easier to compute and process.

[0039] Using LSTM to sequentially encode elements, the combinatorial semantics of entities conditioned on relations are captured as follows:

[0040] When the path step size is l-1, the current entity e is connected in series. l-1 、e′ l-1 and relation r l-1 The embedding vector generates the vector x l-1 (Right now in (It's a chain operation), then x l-1 As the input vector to the LSTM, the output hidden state vector h is... l-1 and use the final state h l Represents the entire path p k .

[0041] Establish path p k After representing the vector state, the final state is input into two fully connected layers to obtain the path p. k The predicted score is calculated using the following formula: Where W1 and W2 are the coefficient weights of the first and second layers, respectively, ReLU is the activation function, and τ is the interaction representation between the user and the item. Through sequential encoding of LSTM, the combinatorial semantics of entities conditioned on relationships can be effectively captured, and the potential connections between different components and state variables can be mined.

[0042] By using a pooling layer to merge multiple paths, the weight values ​​for the interaction between a given operating condition and component state are output as follows:

[0043] In the knowledge perception path of the low-light charging system, different paths contribute differently to the user preference model. Therefore, the pooling layer needs to perform a weighted pooling operation on the total score of all paths and output the final score of the interaction between the given user and the target item, i.e., the final weight value of the component or state variable. First, the scores of all paths are aggregated through a weighted pooling operation, and then the sigmoid function is used to transform the scores into the [0,1] interval, as shown in the formula. Where s k is the predicted score of the Kth path, γ is a hyperparameter controlling the weight of each exponent, and σ is the sigmoid function. This is the final weight value.

[0044] In one specific embodiment, the lifetime prediction module specifically includes:

[0045] The coding table acquisition unit obtains the component parameter classification of each charging component based on the charging component data to form parameter indicators, determines the evaluation quantity of each parameter of the charging component based on the parameter indicators, constructs a multi-parameter evaluation coding table, obtains the degree of degradation of each parameter, and further combines wear data to refine the degree of degradation based on the degree of wear.

[0046] The standard acquisition unit obtains the single parameter index deduction value according to the respective maximum weight value and the deterioration degree combined with the corresponding standard rule, acquires the clear parameter evaluation standard, and formulates the multi-parameter index rule table.

[0047] The overall evaluation unit then formulates the charging assembly state evaluation rule table according to the deduction of each parameter, performs state evaluation on the overall charging assembly, and acquires the service life data.

[0048] In one specific embodiment, the component parameters specifically include any or all of the following: micro photovoltaic component power generation efficiency, battery health state, charging current charging conversion efficiency, and boost circuit output voltage stability.

[0049] In one specific embodiment, the multi-parameter quantity evaluation coding table includes: charging assembly parameter classification, parameter state name, and judgment basis.

[0050] In one specific embodiment, the overall evaluation unit specifically includes:

[0051]

[0052] Wherein, W s is the overall charging assembly score weight coefficient under different parameters, E s is the service life data of the charging assembly, s represents the overall charging assembly score weight coefficient under different parameters, Y represents the total deduction value of each parameter of the entire charging assembly, Y max represents the total deduction value of all parameters of the entire charging assembly, N c represents the fuzzy evaluation set corresponding to the evaluation score.

[0053] In one specific embodiment, the multi-parameter quantity evaluation coding table is shown in Table 1, and the multi-parameter index rule table is shown in Table 2.

[0054] Table 1 Multi-parameter quantity evaluation coding table

[0055]

[0056]

[0058] Table 2 Multi-parameter index rule table

[0059]

[0060] In the present specification, each embodiment is described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts between each embodiment can be referred to each other. For the device disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method part.

[0061] The foregoing description of the disclosed embodiments enables a person skilled in the art to make or use the application. Numerous modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other embodiments without the use of the inventive faculty. Therefore, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A life prediction system based on a 5G micro light base station, characterized by, The cloud processing platform, the mobile base station, and the data acquisition module are provided. The mobile base station comprises a micro photovoltaic module, a charging circuit, a voltage boosting circuit, a local control module, a 5G base station, and a storage battery. 2.The life prediction system based on 5G micro optical base station according to claim 1, wherein, The cloud processing platform comprises a weight evaluation module, a wear calculation module, and a life prediction module. 3.The life prediction system based on 5G micro optical base station of claim 1, wherein, The weight evaluation module combines the semantics of entities and relationships in the safe operation knowledge graph of the micro light charging system to generate a safe operation knowledge perception path, maps each entity, entity type, and relationship in the knowledge graph to a vector representation using a graph convolution network, sequentially encodes the elements using LSTM to capture the combined semantics of entities conditioned on relationships, merges multiple paths using a pooling layer, and outputs the most significant weight value for the interaction between the given operating condition and the component state. The wear calculation module detects the wear area using a YOLOV3 network, then applies a UNet++ network for fine segmentation of the wear area, obtains the precise area and calculates the area pixel area, combines the most significant weight value, and calculates the wear data. The life prediction module establishes a multi-parameter quantity evaluation coding table based on the wear data and the charging component data, and performs life evaluation based on the multi-parameter quantity evaluation coding table to obtain the life data. The life prediction module specifically comprises an encoding table acquisition unit, a standard acquisition unit, and a comprehensive evaluation unit.

4. The life prediction system based on 5G micro optical base station according to claim 3, characterized in that, The component parameters specifically include any or all of the micro photovoltaic module power generation efficiency, the storage battery health status, the charging current charging conversion efficiency, and the voltage boosting circuit output voltage stability. ​ ​ ​ 5. The life prediction system based on 5G micro optical base station according to claim 4, characterized in that, ​ 6. The life prediction system based on 5G micro optical base station according to claim 4, characterized in that, The multi-parameter quantitative evaluation coding table comprises a charging component parameter classification, a parameter state name, and a judgment basis.

7. The life prediction system based on 5G micro optical base station according to claim 4, characterized in that, The overall evaluation unit specifically comprises: Wherein, W s is the overall score weight coefficient of the charging assembly under different parameters, E s is the life data of the charging assembly, s represents the overall score weight coefficient of the charging assembly under different parameters, Y represents the total deduction score value of each parameter of the entire charging assembly, Y max represents the total deduction score value of all parameters of the entire charging assembly, N c represents the evaluation score value corresponding to the fuzzy evaluation set.

Images