Search

CN-122020565-A - New energy charging pile operation data analysis processing method

CN122020565ACN 122020565 ACN122020565 ACN 122020565ACN-122020565-A

Abstract

The invention discloses a new energy charging pile operation data analysis processing method, which relates to the technical field of new energy charging pile data analysis and comprises the following steps that S1, multi-source operation data such as charging pile charging power, current and voltage and the like are collected and stored according to time stamps; the method comprises the steps of S2 preprocessing data, S3 extracting core characteristics of instantaneous operation, time dimension and equipment state, S4 clustering analysis of user charging behaviors to generate a classification tag library, S5 analysis of operation load situation to identify load fluctuation factors, S6 evaluation of equipment health state classification and early warning, S7 generation of optimization strategies such as load scheduling, and S8 acquisition of new data iteration optimization model parameters and algorithm logic. The method realizes multi-source data fusion and accurate analysis, improves fault recognition accuracy and load balance degree, generates a floor-type optimization strategy, and provides scientific support for fine management of the charging pile by iteratively optimizing and adapting to different scenes.

Inventors

  • YAO QINQIN
  • ZHENG LEI
  • ZHENG SHUAI

Assignees

  • 山东君策标准化服务有限公司

Dates

Publication Date
20260512
Application Date
20260226

Claims (10)

  1. 1. The new energy charging pile operation data analysis processing method is characterized by comprising the following steps of: s1, multi-source operation data acquisition, namely synchronously acquiring charging parameters, equipment states, environment information and power grid data through a metering module, a communication module, an environment sensor and a power grid side acquisition device which are arranged in a charging pile, and structurally storing the acquisition data to form an original operation data set; S2, preprocessing data quality, namely performing outlier rejection, missing value complementation and noise smoothing on an original operation data set, and normalizing data to generate a normalized operation data set; S3, extracting basic operation characteristics, extracting instantaneous operation, time dimension and equipment state characteristics based on a standardized data set, removing redundant characteristic parameters through a characteristic dimension reduction algorithm, and reserving a core characteristic set which is strongly related to the operation state and the health degree of the charging pile; S4, clustering the charging behavior characteristics, selecting a core characteristic set as input, performing cluster analysis on the charging behavior of the user by adopting a density clustering algorithm, dividing the user types and generating a charging behavior clustering model and a user classification tag library; s5, analyzing an operation load situation, constructing a charging pile operation load time sequence based on a standardized operation data set, calculating a multi-time scale load parameter, and identifying load influence factors by combining a clustering model to quantitatively evaluate load balance and fluctuation characteristics; S6, evaluating the health state of the equipment, constructing an evaluation index system by combining the related data of the equipment operation, determining index weight, calculating health indexes, dividing the health state of the equipment into four grades of health, sub-health, abnormality and fault, automatically triggering early warning prompt aiming at the abnormal state, and generating an evaluation report and tracing information; S7, generating an operation optimization strategy, and generating a load scheduling, equipment maintenance and user charging guide optimization strategy set aiming at the problems of uneven load, equipment hidden danger and demand matching by combining geographic position and peripheral distribution data of the charging piles based on charging behavior clustering, load situation analysis and equipment health assessment results; and S8, carrying out iterative optimization on the data model, repeating the steps S1-S7, evaluating implementation effects, comparing key indexes, optimizing feature extraction through a gradient descent algorithm, clustering the algorithm and evaluating index weight parameters, and iteratively improving the accuracy and applicability of the model to generate an optimized model and a strategy library.
  2. 2. The method for analyzing and processing operation data of a new energy charging pile according to claim 1, further comprising a load balance quantization calculating step, executed in S5, of quantizing the balance degree of the operation load of the charging pile by a formula, wherein the specific formula is: Wherein the method comprises the steps of In order to achieve a degree of load balancing, To count the number of samples in the time window, Is the first The charge power of the individual sampling points is, Is the average charging power over a statistical time window.
  3. 3. The method for analyzing and processing operation data of a new energy charging pile according to claim 1, further comprising a step of dynamically correcting the health degree of the equipment, wherein the step is executed in S6, the health index of the equipment is dynamically adjusted by a formula, and the specific formula is as follows: Wherein the method comprises the steps of For a dynamically revised device health index, For the initially calculated device health index, The running time period is accumulated for the device, Is the difference between the ambient temperature and the standard operating temperature, In order for the run-time to affect the coefficients, Is the ambient temperature influence coefficient.
  4. 4. The method for analyzing and processing the operation data of the new energy charging pile according to claim 1, wherein in the step of S1 multi-source operation data acquisition, a high-precision electric energy metering chip is adopted by the metering module, the measurement precision level is not lower than 0.5 level, 4G/5G, ethernet and LoRa communication protocols are supported by the communication module, a temperature and humidity compound sensor is adopted by the environment sensor, and geographic position information is acquired by the Beidou positioning module.
  5. 5. The new energy charging pile operation data analysis processing method according to claim 1, wherein in the step of S2 data quality preprocessing, the threshold value setting for abnormal data identification is based on the factory parameters and the historical normal operation data of the equipment, the linear interpolation method with the complete missing value is used for continuous data, the window size of a moving average filtering method is set to 5 sampling points, the data normalization adopts a Z-score normalization method, and the formula is that Wherein As the raw data is to be processed, As a mean value of the data, Data standard deviation.
  6. 6. The method for analyzing and processing the operation data of the new energy charging pile according to claim 1, wherein in the step of extracting the basic operation characteristics, the window size of the time window sliding method is set to 10 minutes, the charging amount change rate per unit time is calculated through the charging amount difference value and the time interval of the adjacent sampling points, the characteristic dimension reduction algorithm adopts a principal component analysis algorithm, and principal components with the accumulated variance contribution rate not lower than 85% are reserved as core characteristic parameters.
  7. 7. The method for analyzing and processing the operation data of the new energy charging pile according to claim 1, wherein in the step of clustering the charging behavior characteristics, the neighborhood radius of the density clustering algorithm is set to 0.5, the minimum neighborhood sample number is set to 5, the clustering density is calculated by the sample number in a unit time window, the user classification tag library is updated in real time, and the newly added user behavior data is automatically classified into the corresponding clustering category or forms the new category.
  8. 8. The method for analyzing and processing the operation data of the new energy charging pile according to claim 1, wherein in the step of analyzing the operation load situation, the calculation period of the load average value, the fluctuation rate and the peak value is respectively set to be 1 hour, 1 day and 1 month, the load situation thermodynamic diagram is drawn by adopting the color depth to represent the load intensity, the abscissa of a load change trend curve is the time, the ordinate is the charging power, and the load change trend is shown by a curve fitting algorithm.
  9. 9. The method for analyzing and processing the operation data of the new energy charging pile according to claim 1, wherein in the step of evaluating the health state of the equipment, the health evaluation index system comprises an equipment operation state index, a fault frequency index, an off-line time index and a maintenance record index, a judgment matrix of the analytic hierarchy process is constructed according to expert experience and historical data, and fault risk tracing information is associated with fault occurrence time, corresponding operation characteristics and environmental conditions to realize fault cause positioning.
  10. 10. The new energy charging pile operation data analysis processing method according to claim 1, wherein in the step of generating the operation optimization strategy, the load scheduling strategy comprises peak-shifting charging guidance, multi-pile power distribution and peak-shifting and valley-filling strategy interacted with a power grid, the equipment maintenance optimization strategy comprises preventive maintenance planning based on health indexes and rapid overhaul and replacement strategy of fault equipment, the user charging guidance strategy comprises recommending optimal charging piles, guiding users to idle charging piles and providing charging preferential information, and the operation optimization strategy set is displayed through a visual interface and supports direct checking and execution of operators; In the step of iterative optimization of the data model, indexes for quantitatively evaluating implementation effects comprise a load balance degree lifting rate, a device health index lifting rate, a user average waiting charging time reducing rate and a charging fault occurrence rate reducing rate, a learning rate initial value of a gradient descent algorithm is set to be 0.01, an iterative stopping condition is that the lifting rate of an evaluation index is continuously lower than 1% for 3 times or reaches the maximum iteration number of 1000 times, and an operation strategy library after iterative optimization is stored in a cloud server to support sharing and synchronous updating of multiple charging piles.

Description

New energy charging pile operation data analysis processing method Technical Field The invention relates to the technical field of new energy charging pile data analysis, in particular to a new energy charging pile operation data analysis processing method. Background Along with the continuous rising of the holding quantity of the new energy automobile, the charging pile is used as a core supporting facility, and the running stability and the service efficiency of the charging pile directly influence the popularization and the use experience of the new energy automobile. In the long-term high-frequency operation process of the charging pile, the problems of large load fluctuation, quick equipment aging, various user demands and the like are faced, the conditions of charging faults, unbalanced load, untimely maintenance and the like are easy to occur, the charging experience of the user is affected, and potential safety hazards are also possibly caused. Therefore, the real-time monitoring, fault early warning and optimal scheduling of the running state of the charging pile are realized through accurate data analysis and processing, and the method becomes the key for improving the operation and maintenance management level and the service quality of the charging pile and is also an important research direction in the intelligent energy field. In the prior art, although a certain progress has been made in related data analysis and processing schemes, a plurality of technical short boards still exist. In patent application documents with patent publication number of CN118885703B and patent name of "a running state detection analysis method suitable for a high-power device of a charging pile", detection data of a charging pile switch, an electromagnetic lock, a resistor and other components are collected, a least square method is adopted to fit an optimal curve after preprocessing, and the optimal curve is compared with current data to judge faults. However, the technology only focuses on fault detection of the high-power device, does not relate to integral load situation analysis of the charging pile, mining of user behavior characteristics and operation strategy optimization, has single data acquisition dimension, is limited in fault diagnosis in application scene, and cannot meet the requirements of full life cycle operation and maintenance and efficient scheduling of the charging pile. In patent application documents with patent publication number CN118132956A and patent name of 'a new energy charging pile operation data analysis processing method based on artificial intelligence', periodic data monitoring of charging piles in different charging areas is mentioned, and abnormal identification and area management dynamic prompt of the charging piles are realized by calculating indexes such as daily charging duration, area charging saturation and the like. However, the overall state analysis of the technical emphasis area is not accurate enough in the evaluation of the health state of the individual charging pile, a dynamic optimization mechanism is not introduced, the influence of environmental factors and power grid parameters on the operation of the charging pile is not considered, the comprehensiveness and adaptability of an analysis result are not enough, and the generation of an individual operation optimization strategy is difficult to support. In patent application documents with patent publication number of CN117892155B and patent name of "a charging pile data processing method and a system thereof", a prediction model is constructed based on a multi-view shape learning mechanism, charging pile charging and operation data are predicted, and matching recommendation of the charging pile and a vehicle is realized by combining vehicle characteristic information. However, the technology focuses on user matching recommendation, focuses on the defect of insufficient attention on operation and maintenance core requirements such as health state evaluation, load balance scheduling and the like of the charging pile, does not construct a complete health evaluation index system and an iterative optimization mechanism, and data processing results are difficult to directly convert into an actual operation strategy of equipment maintenance and load scheduling, so that the practicability and the floor property are limited. In summary, the prior art has three core limitations that firstly, the data acquisition dimension is on one side, single scene or single index is focused in multiple ways, multi-source data fusion of equipment operation, user behavior, environmental parameters and power grid states is not realized, secondly, the analysis depth is insufficient, the analysis depth is stopped at the fault recognition or state statistics level, the accurate quantitative evaluation and dynamic optimization on the load situation and the health state are lacking, thirdly, the application direction is undefined, the analysi