Cross-platform sharing early warning method for health degree of charging pile

Abstract

The invention relates to the technical field of charging pile health assessment, and discloses a charging pile health degree cross-platform sharing early warning method, which comprises the following steps: sending preprocessed electrical data of an ith charging pile to a constructed prediction model, obtaining a first prediction coefficient of the ith charging pile, obtaining an external fault factor of the ith charging pile, and obtaining a second prediction coefficient of the ith charging pile; and constructing a second prediction coefficient of the ith charging pile, and comprehensively constructing an early warning coefficient of the ith charging pile. According to the method, the real-time performance and systematicness of fault monitoring are ensured through a complete closed loop from data acquisition, preprocessing, model prediction to multi-factor fusion early warning and maintenance scheduling, the electrical data processing and multi-dimensional fault factor modeling of different charging piles are determined step by step, the prediction accuracy and pertinence are improved, and the fault monitoring efficiency is improved. Deviation possibly caused by a single data source is avoided, it is ensured that maintenance measures can be rapidly implemented after early warning, and the execution force of operation and maintenance management and the resource utilization efficiency are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of charging pile health assessment, and in particular to a cross-platform shared early warning method for charging pile health. Background Art

[0002] It provides a practical solution to improve environmental problems and alleviate energy conflicts, and charging piles as charging tools for new energy vehicles need special attention;

[0003] Existing charging stations have a large number of charging piles, which are widely distributed, and most charging stations are basically unmanned. In order to reduce the operation and maintenance costs of charging stations and improve the reliability of charging piles, it is necessary to adopt more reliable charging pile fault prediction technology to enable them to have a higher level of automation so that maintenance can be carried out more conveniently. Although there are many fault prediction methods for charging piles, most of them use traditional recurrent neural networks or their variants LSTM, which have problems such as insufficient generalization ability and low prediction accuracy. Summary of the Invention

[0004] In response to the shortcomings of the existing technology, the present invention provides a cross-platform shared warning method for charging pile health, which has the advantages of combining different platforms to provide warnings for charging pile failures, thereby solving the above-mentioned technical problems.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for sharing early warning of charging pile health across platforms, comprising the following steps:

[0006] S1: Obtain electrical data of charging states of different platforms and pre-process them separately;

[0007] S2: Send the pre-processed electrical data of the i-th charging pile to the constructed prediction model to obtain the first prediction coefficient of the i-th charging pile;

[0008] S3: Obtain the external fault factor of the i-th charging pile and construct the second prediction coefficient of the i-th charging pile;

[0009] S4: comprehensively constructing the warning coefficient of the i-th charging pile based on the first prediction coefficient of the i-th charging pile in S2 and the second prediction coefficient of the i-th charging pile in S3;

[0010] S5: Determine whether to issue an early warning based on the early warning coefficient of the i-th charging pile. If an early warning is required, execute S6. If not, increase i+1 and jump to S2 until all platforms are traversed.

[0011] S6: Issue an early warning to the manager of the i-th charging pile, evaluate the failure coefficient of the i-th charging pile, and dispatch staff to carry out maintenance. Then, after i+1, jump to S2 until all platforms are traversed.

[0012] As a preferred technical solution of the present invention, the electrical data of the charging status of different platforms obtained in S1 include: charging pile voltage, charging pile current, charging pile power and charging pile temperature. The preprocessing in S1 includes the following steps:

[0013] S1.1: Standardize the electrical data of the i-th charging pile. The specific expression is as follows:

[0014]

[0015] in, represents the initial values ​​of the charging pile voltage, charging pile current, charging pile power and charging pile temperature at the i-th charging pile at time t, It represents the processing value of the charging pile voltage, charging pile current, charging pile power and charging pile temperature at the input i-th charging pile at time t, and Norm represents normalization;

[0016] S1.2: Process the charging pile voltage, charging pile current, charging pile power and charging pile temperature at the i-th moment t Make a judgment, when Any element in or less than Then remove the element and obtain the pre-processed electrical data of the i-th charging pile;

[0017] in, express The mean of express The standard deviation of .

[0018] As a preferred technical solution of the present invention, in S2, the preprocessed electrical data of the i-th charging pile is sent to the constructed prediction model. Specifically, the prediction model is constructed by training the preprocessed electrical data of the charging piles of each platform through the Transformer model independently deployed on each platform, and then the final prediction model is constructed through federated learning. The preprocessed electrical data of the i-th charging pile is input into the final prediction model in real time to obtain the fault probability, and the fault probability is used as the first prediction coefficient of the i-th charging pile.

[0019] As a preferred technical solution of the present invention, the specific expression for obtaining the external fault factor of the i-th charging pile and constructing the second prediction coefficient of the i-th charging pile in S3 is as follows:

[0020]

[0021] Among them, DEYC i Represents the second prediction coefficient of the i-th charging pile, HJYXXSi represents the environmental impact coefficient of the i-th charging pile, FHYXXS i Indicates the protection impact coefficient of the i-th charging pile, WXXS i Represents the maintenance impact coefficient of the i-th charging pile.

[0022] As the preferred technical solution of the present invention, the environmental impact coefficient of the i-th charging pile is HJYXXS i The specific expression is as follows:

[0023]

[0024] Among them, CCSC i It represents the total time that the dust concentration in the environment of the i-th charging pile exceeds the safety value, ZSBL i Indicates the total duration of the i-th charging pile cable being exposed to direct sunlight.

[0025] As a preferred technical solution of the present invention, the protection influence coefficient FHYXXS of the i-th charging pile i The specific expression is as follows:

[0026] FHYXXS i =α*ZSBL i +β*DSBL i

[0027] Among them, ZSBL i represents the total duration of the i-th charging pile cable being exposed to direct sunlight, α and β represent weight coefficients that sum to 1, and DSBL i Indicates the blockage ratio of the heat dissipation area of ​​the i-th charging pile.

[0028] As a preferred technical solution of the present invention, the maintenance impact coefficient WXXS of the i-th charging pile i The specific expression is as follows:

[0029]

[0030] Among them, XLMJ i represents the proportion of maintenance area during the maintenance process of the i-th charging pile, WXCS i Indicates the maintenance overtime rate.

[0031] As a preferred technical solution of the present invention, the specific expression for comprehensively constructing the warning coefficient of the i-th charging pile based on the first prediction coefficient of the i-th charging pile in S2 and the second prediction coefficient of the i-th charging pile in S3 in S4 is as follows:

[0032] YJXS i =DYYC i *(1+DEYCi )

[0033] Among them, DEYC i represents the second prediction coefficient of the i-th charging pile, YJXS i represents the warning coefficient of the i-th charging pile, DYYC i Represents the first prediction coefficient of the i-th charging pile.

[0034] As a preferred technical solution of the present invention, the judgment of whether to issue an early warning based on the early warning coefficient of the i-th charging pile in S5 is specifically as follows: when the early warning coefficient YJXS of the i-th charging pile is i When the pre-set warning threshold δ is exceeded, a warning is issued. When the warning coefficient YJXS of the i-th charging pile is i If the set warning threshold δ is not exceeded, no warning will be issued. When a warning is issued, a list of all charging piles that issue warnings is constructed based on the failure coefficient of the i-th charging pile, and the maintenance personnel scheduling weight is calculated, and maintenance personnel are selected for maintenance based on the weight.

[0035] As a preferred technical solution of the present invention, the specific expression of the failure coefficient of the i-th charging pile is as follows:

[0036]

[0037] Among them, δ represents the set warning threshold, YJXS i represents the warning coefficient of the i-th charging pile, GZXS i represents the failure coefficient of the i-th charging pile;

[0038] The specific expression of the maintenance personnel scheduling weight is as follows:

[0039]

[0040] Among them, GZXS j→i represents the scheduling weight of the jth maintenance personnel going to the i-th charging pile for maintenance, LJCD j→i represents the straight-line distance from the jth maintenance worker to the ith charging pile, SJJL j→i represents the actual distance traveled by the j-th maintenance worker to the i-th charging station.

[0041] Compared with the existing technology, the present invention provides a cross-platform shared early warning method for charging pile health, which has the following beneficial effects:

[0042] The present invention ensures the real-time and systematic nature of fault monitoring through a complete closed loop from data collection, preprocessing, model prediction to multi-factor fusion early warning and maintenance scheduling; secondly, by clarifying the electrical data processing and multi-dimensional fault factor modeling of different charging piles in steps, the accuracy and pertinence of the prediction are improved, avoiding the deviation that may be caused by a single data source; furthermore, the first prediction coefficient and the second prediction coefficient are comprehensively utilized to construct the early warning coefficient, realizing a multi-angle evaluation of the equipment status, making the early warning more scientific and reasonable; finally, the introduction of fault coefficient evaluation and maintenance personnel scheduling ensures that maintenance measures can be quickly implemented after the early warning, significantly improving the execution of operation and maintenance management and the efficiency of resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0045] See also Figure 1 ,The cross-platform sharing early warning method for charging pile health includes the following steps:

[0046] S1: Acquires electrical data from different platforms at different charging states and pre-processes them separately. This complete closed loop from data acquisition, pre-processing, model prediction, to multi-factor fusion early warning and maintenance scheduling ensures real-time and systematic fault monitoring.

[0047] The electrical data of the charging status of different platforms obtained in S1 include: charging pile voltage, charging pile current, charging pile power and charging pile temperature. The preprocessing in S1 includes the following steps:

[0048] S1.1: Standardize the electrical data of the i-th charging pile. The specific expression is as follows:

[0049]

[0050] in, represents the initial values ​​of the charging pile voltage, charging pile current, charging pile power and charging pile temperature at the i-th charging pile at time t, Indicates the processed values ​​of the charging pile voltage, charging pile current, charging pile power, and charging pile temperature at the i-th charging pile at time t. Norm indicates normalization. Normalization is a common operation and will not be described in detail here. It can be Z standard or other standards, but the normalization method for all different electrical data must be consistent;

[0051] S1.2: Process the charging pile voltage, charging pile current, charging pile power and charging pile temperature at the i-th moment t Make a judgment, when Any element in or less than Then remove the element and obtain the pre-processed electrical data of the i-th charging pile;

[0052] in, express The mean of express The standard deviation of

[0053] S2: The pre-processed electrical data of the i-th charging pile is sent to the constructed prediction model to obtain the first prediction coefficient of the i-th charging pile. By clarifying the electrical data processing and multi-dimensional fault factor modeling of different charging piles in steps, the accuracy and pertinence of the prediction are improved, and the deviation that may be caused by a single data source is avoided. In S2, the pre-processed electrical data of the i-th charging pile is sent to the constructed prediction model. Specifically, the prediction model is to train the pre-processed electrical data of the charging piles of each platform through the Transformer model independently deployed on each platform, and then build the final prediction model through federated learning. The pre-processed electrical data of the i-th charging pile is input into the final prediction model in real time to obtain the fault probability, and the fault probability is used as the first prediction coefficient of the i-th charging pile.

[0054] The core of the Transformer model lies in its self-attention mechanism, which can effectively capture long-range dependencies in sequence data. In the health assessment and fault diagnosis of charging piles, the operating data of the health assessment of charging piles (such as current, voltage, temperature, etc.) can be regarded as time series data, and the Transformer model is used to extract features and perform classification prediction on these data. The loss function in this article: Binary Cross-Entropy loss function is used, which is suitable for binary classification fault probability prediction. The optimization algorithm: AdamW optimizer, which is stable and efficient, is suitable for Transformer model training. Since the Transformer model has good application effects in many power fields and can realize the prediction of charging pile faults, it will not be elaborated here.

[0055] In addition, the advantages of federated learning are mainly reflected in the following aspects:

[0056] Protect data privacy. Data does not leave the local device or platform, avoiding the risk of sensitive information exposure and data leakage. It complies with data privacy regulations and compliance requirements, reduces data transmission costs, and only transmits model parameters or gradients, reducing the need for large-scale original data transmission, saving bandwidth and storage resources, and improving model generalization capabilities. By aggregating data models from multiple heterogeneous clients and integrating more diverse data features, the adaptability and robustness of the model in different scenarios are enhanced. Distributed training has strong acceleration and scalability. Multiple clients can train in parallel, making full use of distributed computing resources, supporting large-scale systems and cross-regional collaborative learning, avoiding central single point failures, and distributing the training process across multiple nodes, reducing dependence on a single server, improving system reliability and stability, promoting multi-party collaboration, and supporting cross-institutional and cross-enterprise joint modeling while ensuring data privacy, realizing resource sharing and collaborative optimization;

[0057] S3: Obtain the external fault factor of the i-th charging pile and construct the second prediction coefficient of the i-th charging pile;

[0058] In S3, the external fault factor of the i-th charging pile is obtained and the specific expression of the second prediction coefficient of the i-th charging pile is constructed as follows:

[0059]

[0060] Among them, DEYC i Represents the second prediction coefficient of the i-th charging pile, HJYXXS i represents the environmental impact coefficient of the i-th charging pile, FHYXXS i Indicates the protection impact coefficient of the i-th charging pile, WXXS i Represents the maintenance impact coefficient of the i-th charging pile.

[0061] The environmental impact coefficient of the i-th charging pile HJYXXS i The specific expression is as follows:

[0062]

[0063] Among them, CCSC i It represents the total time that the dust concentration in the environment of the i-th charging pile exceeds the safety value, ZSBL i The total proportion of time that the i-th charging pile cable is exposed to direct sunlight uses the time when dust concentration exceeds the safety standard and the proportion of time in direct sunlight as indicators, which directly reflects the actual impact of the environment on the equipment. By averaging the two key environmental factors, it simplifies the environmental complexity and covers the main stress sources in the environment.

[0064] Protection influence coefficient FHYXXS of the i-th charging pile i The specific expression is as follows:

[0065] FHYXXS i =α*ZSBL i +β*DSBL i

[0066] Among them, ZSBL i represents the total duration of the i-th charging pile cable being exposed to direct sunlight, α and β represent weight coefficients that sum to 1, and DSBL i It represents the blockage ratio of the heat dissipation of the i-th charging pile. It measures the protection impact through two indicators: the direct sunlight ratio and the heat dissipation blockage ratio. It highlights the control effect on key damage factors, can quantitatively reflect the specific protection effect of the protection work, and help optimize the protection strategy.

[0067] Maintenance impact coefficient WXXS of the i-th charging pile i The specific expression is as follows:

[0068]

[0069] Among them, XLMJ i represents the proportion of maintenance area during the maintenance process of the i-th charging pile, WXCS i Indicates the maintenance overtime rate. It quantifies the maintenance quality and timeliness through the maintenance area ratio and the maintenance overtime rate. It is closely related to the actual operation and maintenance, covers both maintenance scope and delay, and can comprehensively evaluate the impact of maintenance on equipment reliability.

[0070] S4: comprehensively constructing the warning coefficient of the i-th charging pile based on the first prediction coefficient of the i-th charging pile in S2 and the second prediction coefficient of the i-th charging pile in S3;

[0071] The specific expression of the warning coefficient of the i-th charging pile constructed in S4 based on the first prediction coefficient of the i-th charging pile in S2 and the second prediction coefficient of the i-th charging pile in S3 is as follows:

[0072] YJXS i =DYYC i *(1+DWYC i )

[0073] Among them, DWYC i represents the second prediction coefficient of the i-th charging pile, YJXS i represents the warning coefficient of the i-th charging pile, DYYC iIt represents the first prediction coefficient of the i-th charging pile. The first prediction coefficient and the second prediction coefficient are comprehensively used to construct the early warning coefficient, which can realize the multi-angle evaluation of the equipment status and make the early warning more scientific and reasonable.

[0074] S5: Determine whether to issue an early warning based on the early warning coefficient of the i-th charging pile. If an early warning is required, execute S6. If not, increase i+1 and jump to S2 until all platforms are traversed.

[0075] In S5, the judgment of whether to issue an early warning is based on the early warning coefficient of the i-th charging pile is as follows: when the early warning coefficient of the i-th charging pile YJXS i When the pre-set warning threshold δ is exceeded, a warning is issued. When the warning coefficient YJXS of the i-th charging pile is i If the set warning threshold δ is not exceeded, no warning will be issued. When a warning is issued, a list of all charging piles that issue warnings is constructed based on the failure coefficient of the i-th charging pile, and the maintenance personnel scheduling weight is calculated, and maintenance personnel are selected for maintenance based on the weight.

[0076] S6: Issue an early warning to the manager of the i-th charging pile, evaluate the failure coefficient of the i-th charging pile, and dispatch staff to perform maintenance. Then, after i+1, jump to S2 until all platforms are traversed. The specific expression of the failure coefficient of the i-th charging pile is as follows:

[0077]

[0078] Among them, δ represents the set warning threshold, YJXS i represents the warning coefficient of the i-th charging pile, GZXS i represents the failure coefficient of the i-th charging pile. The failure coefficient is calculated based on the degree of exceedance of the warning threshold, which realizes the quantitative assessment of the failure and is conducive to accurately determining the maintenance priority;

[0079] The specific expression of maintenance personnel scheduling weight is as follows:

[0080]

[0081] Among them, GZXS j→i represents the scheduling weight of the jth maintenance personnel going to the i-th charging pile for maintenance, LJCD j→i represents the straight-line distance from the jth maintenance worker to the ith charging pile, SJJL j→i The actual distance from the jth maintenance personnel to the ith charging pile is represented by the dispatch weight, which introduces two indicators: straight-line distance and actual distance. This is more in line with the actual situation on site, reflects the impact of the road and other environmental factors, ensures the rationality and effectiveness of the dispatch, and dispatches the i-th charging pile to GZXS. j→i The smallest one.

[0082] The threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by technicians in this field for each set of sample data. As long as it does not affect the proportional relationship between the parameter and the quantized value, the weight can be determined by technicians in this field based on each sample data and multiple rounds of experiments. The above formulas are all calculated by removing the dimension and taking the numerical value.

[0083] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A cross-platform shared early warning method for charging pile health, characterized by: The following steps are involved: S1: Obtain electrical data of charging states of different platforms and pre-process them separately; S2: Send the pre-processed electrical data of the i-th charging pile to the constructed prediction model to obtain the first prediction coefficient of the i-th charging pile; S3: Obtain the external fault factor of the i-th charging pile and construct the second prediction coefficient of the i-th charging pile; S4: comprehensively constructing the warning coefficient of the i-th charging pile based on the first prediction coefficient of the i-th charging pile in S2 and the second prediction coefficient of the i-th charging pile in S3; S5: Determine whether to issue an early warning based on the early warning coefficient of the i-th charging pile. If an early warning is required, execute S6. If not, increase i+1 and jump to S2 until all platforms are traversed. S6: Issue an early warning to the manager of the i-th charging pile, evaluate the failure coefficient of the i-th charging pile, and dispatch staff to carry out maintenance. Then, after i+1, jump to S2 until all platforms are traversed.

2. The cross-platform shared early warning method for charging pile health according to claim 1 is characterized by: The electrical data of the charging status of different platforms obtained in S1 include: charging pile voltage, charging pile current, charging pile power and charging pile temperature. The preprocessing in S1 includes the following steps: S1.1: Standardize the electrical data of the i-th charging pile. The specific expression is as follows: in, represents the initial values ​​of the charging pile voltage, charging pile current, charging pile power and charging pile temperature at the i-th charging pile at time t, It represents the processing value of the charging pile voltage, charging pile current, charging pile power and charging pile temperature at the input i-th charging pile at time t, and Norm represents normalization; S1.2: Process the charging pile voltage, charging pile current, charging pile power and charging pile temperature at the i-th moment t Make a judgment, when Any element in or less than Then remove the element and obtain the pre-processed electrical data of the i-th charging pile; in, express The mean of express The standard deviation of .

3. The cross-platform shared early warning method for charging pile health according to claim 2 is characterized by: In S2, the pre-processed electrical data of the i-th charging pile is sent to the constructed prediction model. Specifically, the prediction model is constructed by training the pre-processed electrical data of the charging piles of each platform through the Transformer model independently deployed on each platform, and then the final prediction model is constructed through federated learning. The pre-processed electrical data of the i-th charging pile is input into the final prediction model in real time to obtain the fault probability, and the fault probability is used as the first prediction coefficient of the i-th charging pile.

4. The cross-platform shared early warning method for charging pile health according to claim 1 is characterized by: The specific expression for obtaining the external fault factor of the i-th charging pile and constructing the second prediction coefficient of the i-th charging pile in S3 is as follows: Among them, DEYC i Represents the second prediction coefficient of the i-th charging pile, HJYXXS i represents the environmental impact coefficient of the i-th charging pile, FHYXXS i Indicates the protection impact coefficient of the i-th charging pile, WXXS i Represents the maintenance impact coefficient of the i-th charging pile.

5. The cross-platform shared early warning method for charging pile health according to claim 4 is characterized by: The environmental impact coefficient of the i-th charging pile is HJYXXS i The specific expression is as follows: Among them, CCSC i It represents the total time that the dust concentration in the environment of the i-th charging pile exceeds the safety value, ZSBL i Indicates the total duration of the i-th charging pile cable being exposed to direct sunlight.

6. The cross-platform shared early warning method for charging pile health according to claim 4 is characterized by: The protection influence coefficient FHYXXS of the i-th charging pile i The specific expression is as follows: FHYXXS i =α*ZSBL i +β*DSBL i Among them, ZSBL i represents the total duration of the i-th charging pile cable being exposed to direct sunlight, α and β represent weight coefficients that sum to 1, and DSBL i Indicates the blockage ratio of the heat dissipation area of ​​the i-th charging pile.

7. The cross-platform shared warning method for charging pile health according to claim 4 is characterized by: The maintenance impact coefficient of the i-th charging pile WXXS i The specific expression is as follows: Among them, XLMJ i represents the proportion of maintenance area during the maintenance process of the i-th charging pile, WXCS i Indicates the maintenance overtime rate.

8. The method for sharing and warning the health status of charging piles across platforms according to claim 4 is characterized by: The specific expression for comprehensively constructing the warning coefficient of the i-th charging pile in S4 based on the first prediction coefficient of the i-th charging pile in S2 and the second prediction coefficient of the i-th charging pile in S3 is as follows: YJXS i =DYYC i *(1+DWYC i ) Among them, DWYC i represents the second prediction coefficient of the i-th charging pile, YJXS i represents the warning coefficient of the i-th charging pile, DYYC i Represents the first prediction coefficient of the i-th charging pile.

9. The cross-platform shared early warning method for charging pile health according to claim 4 is characterized by: The judgment of whether to issue an early warning based on the early warning coefficient of the i-th charging pile in S5 is specifically as follows: when the early warning coefficient of the i-th charging pile YJXS i When the pre-set warning threshold δ is exceeded, a warning is issued. When the warning coefficient YJXS of the i-th charging pile is i If the set warning threshold δ is not exceeded, no warning will be issued. When a warning is issued, a list of all charging piles that issue warnings is constructed based on the failure coefficient of the i-th charging pile, and the maintenance personnel scheduling weight is calculated, and maintenance personnel are selected for maintenance based on the weight.

10. The method for sharing and warning the health status of charging piles across platforms according to claim 1 is characterized by: The specific expression of the failure coefficient of the i-th charging pile is as follows: Among them, δ represents the set warning threshold, YJXS i represents the warning coefficient of the i-th charging pile, GZXS i represents the failure coefficient of the i-th charging pile; The specific expression of the maintenance personnel scheduling weight is as follows: Among them, GZXS j→i represents the scheduling weight of the jth maintenance personnel going to the i-th charging pile for maintenance, LJCD j→i represents the straight-line distance from the jth maintenance worker to the ith charging pile, SJJL j→i represents the actual distance traveled by the j-th maintenance worker to the i-th charging station.

Images