Charging pile production test data collection and analysis method and system

Abstract

The present invention relates to the field of data analysis technology, and in particular to a charging pile production test data acquisition and analysis method and system. The method comprises the following steps: based on multiple sensors, real-time production monitoring sampling of charging piles is performed to obtain real-time status monitoring parameters of multiple parts; time sequence segmentation and fitting of real-time status monitoring parameters of multiple parts is performed to construct a time series of status monitoring parameters of multiple parts; parameter change association mining is performed on the time series of status monitoring parameters of multiple parts, and multi-time point deep learning is performed to construct a charging pile state trend evolution map; real-time production test detection is performed on the charging pile state trend evolution map, and abnormal fault identification is performed to obtain abnormal fault points of the charging pile; fault propagation path tracing is performed on the abnormal fault points of the charging pile, and then fault attribution analysis is performed to obtain the fault cause of the abnormal part. The present invention realizes efficient test data analysis and accurate fault prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a method and system for collecting and analyzing production test data of a charging pile. Background Art

[0002] As the core component of electric vehicle charging equipment, the effectiveness of the production test data collection and analysis method is crucial to ensure the normal operation of the charging equipment. With the rapid development of the electric vehicle industry and the continuous expansion of the production scale of charging piles, the requirements for the production quality and stability of charging piles are also increasing.

[0003] In the traditional charging pile production test data collection and analysis, there are often problems such as incomplete data collection, low analysis efficiency, and limited fault prediction capabilities. These problems lead to low production line efficiency and long fault handling cycles, which seriously affect the production operation and service quality of charging piles. In response to these challenges, an intelligent charging pile production test data analysis method is needed. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a charging pile production test data collection and analysis method and system to solve at least one of the above technical problems.

[0005] To achieve the above object, the present invention provides a charging pile production test data collection and analysis method, comprising the following steps: Step S1: Based on multiple sensors, real-time production monitoring sampling is performed on the charging pile to obtain real-time status monitoring parameters of multiple parts, and time sequence segmentation and fitting of the real-time status monitoring parameters of multiple parts are performed to construct a time series sequence of the status monitoring parameters of multiple parts; Step S2: Mining the correlation of parameter changes in the time series of multi-part status monitoring parameters, and conducting multi-time point deep learning to construct a charging pile status trend evolution map; Step S3: Perform real-time production test on the charging pile status trend evolution map and identify abnormal faults, thereby obtaining abnormal fault points of the charging pile; Step S4: tracing the fault propagation path of the abnormal fault point of the charging pile, and then performing fault attribution analysis to obtain the cause of the fault at the abnormal part; Step S5: predict the future periodic fault evolution of the abnormal fault point of the charging pile according to the fault cause of the abnormal part, make a fault repair diagnosis decision, and build a fault repair diagnosis strategy; Step S6: Based on the fault repair diagnosis strategy, the production process parameter adjustment simulation is performed on the charging pile to generate a plurality of parameter adjustment production response data; based on the plurality of parameter adjustment production response data, the production parameters are fine-tuned and optimized to construct an intelligent production parameter control model.

[0006] The present invention performs real-time production monitoring sampling of charging piles through multiple sensors to obtain real-time status monitoring parameters of multiple parts, constructs a time series of the status monitoring parameters of multiple parts, and understands the evolution process of the status of each part of the charging pile, providing basic data for subsequent analysis. The parameter change association mining and deep learning of the time series of the status monitoring parameters of multiple parts are performed to construct a charging pile status trend evolution map. The construction of the charging pile status trend evolution map reveals the correlation between parameters, providing a deeper analysis basis for abnormal fault identification, and performs real-time production test detection and abnormal fault identification on the charging pile status trend evolution map to find the abnormal fault point of the charging pile, detect and identify the abnormality in time, and improve production efficiency and product quality. quantity, reduce potential losses, trace the fault propagation path and analyze the fault attribution of abnormal fault points, find out the cause of the fault in the abnormal part, understand the cause of the fault and formulate a more effective repair strategy, improve the stability and reliability of the production line, predict the evolution of future cycle faults and make fault repair diagnosis decisions based on the cause of the fault in the abnormal part, build a fault repair diagnosis strategy, predict fault evolution and formulate a diagnosis strategy to reduce future failure risks, improve the stability of the production line, simulate the adjustment of production process parameters based on the fault repair diagnosis strategy, generate production response data, and perform parameter fine-tuning optimization, build an intelligent production parameter control model, adjust and optimize parameters through simulation, improve production efficiency and product quality, and realize intelligent production control.

[0007] Preferably, step S1 comprises the following steps: Step S11: Perform real-time production monitoring sampling on the charging pile based on multiple sensors to obtain real-time status monitoring parameters of multiple parts; Step S12: performing multi-point parameter mean calculation on the real-time status monitoring parameters of multiple parts to generate standard monitoring ranges for different parts; Step S13: performing outlier detection on the real-time status monitoring parameters of multiple parts based on the standard monitoring ranges of different parts, and extracting abnormal outlier parameters; Step S14: performing outlier filtering processing based on the abnormal outlier parameters to generate outlier filtering state monitoring parameters; Step S15: performing low-frequency filtering and denoising on the outlier filtering state monitoring parameters to obtain filtered denoised state monitoring parameters; Step S16: Perform time-sequential segmentation fitting on the filtered and denoised state monitoring parameters to construct a time series of multi-part state monitoring parameters.

[0008] The present invention performs real-time production monitoring sampling on charging piles through multiple sensors to obtain real-time status monitoring parameters of multiple parts, provides original data for subsequent data processing, performs multi-time point parameter mean calculation on the real-time status monitoring parameters of multiple parts, generates standard monitoring ranges for different parts, and determines the normal working range by generating the standard monitoring range, provides a reference range for subsequent abnormal detection, performs abnormal value detection on the real-time monitoring parameters based on the standard monitoring range, extracts abnormal outlier parameters, and detects and extracts abnormal values ​​to help timely discover potential problems and reduce failure risks, performs outlier filtering processing based on the abnormal outlier parameters, generates outlier filtering status monitoring parameters, and reduces noise interference and improves data accuracy and reliability, performs low-frequency filtering and denoising on the outlier filtering status monitoring parameters to obtain filtered denoising status monitoring parameters, removes noise to improve data quality, and makes the data more interpretable and valuable for reference, performs time sequence segmentation and fitting on the filtered denoising status monitoring parameters, and constructs a time series sequence of status monitoring parameters of multiple parts. The construction of the time series sequence understands the trend of state changes and provides a basis for further analysis and prediction.

[0009] Preferably, step S2 comprises the following steps: Step S21: performing time series parameter change analysis on the time series of multi-part state monitoring parameters to generate parameter change trends of different parts; Step S22: performing component topology dependency analysis on multiple parts of the charging pile to extract topology dependency relationships between the parts; Step S23: performing parameter change association mining on parameter change trends of different parts according to the topological dependency relationship between the parts, and extracting the parameter change association rules of different parts; Step S24: performing dynamic co-evolution analysis on the correlation rules of parameter changes of different parts to generate state co-evolution data of different parts; Step S25: Perform multi-time-point deep learning on the state collaborative evolution data of different parts to construct a charging pile state trend evolution map.

[0010] The present invention generates parameter change trends of different parts through time series parameter change analysis, which helps to understand the changing laws of the states of different parts. The generation of parameter change trends provides basic data for subsequent steps and provides important clues for revealing the association between parts. Component topological dependency analysis is performed to extract the topological dependency between parts, and the correlation and dependency between the parts are understood. The mining of topological dependency provides a structured information basis for subsequent parameter association analysis. Based on the topological dependency, the parameter change trends of different parts are associated and mined, the parameter change association law is extracted, and the correlation between parts is revealed. The mining of the parameter change association law deeply understands the mutual influence and change trend between parts, and the dynamic collaborative evolution analysis is performed on the parameter change association law to generate state collaborative evolution data of different parts, which helps to understand the dynamic evolution process between parts. The dynamic collaborative evolution analysis provides in-depth insights into revealing the complex collaborative relationship between the states of various parts of the charging pile. Based on multi-time point deep learning, a charging pile state trend evolution map is constructed to help comprehensively understand the evolution law and trend of the charging pile state. The application of deep learning methods makes the analysis of charging pile state data more intelligent and comprehensive, and provides more in-depth insights for the collection and analysis of production test data.

[0011] Preferably, step S3 specifically comprises the following steps: Step S31: Perform normal evolution situation analysis on the charging pile state trend evolution map to extract the normal evolution trajectory; Step S32: Performing real-time production testing on the charging pile based on the normalized evolution trajectory to identify evolution trajectory deviation data; Step S33: Calculate the deviation frequency of the evolving trajectory deviation data to obtain the trajectory deviation parameter frequency; Step S34: abnormal fault identification is performed based on the trajectory deviation parameter frequency, so as to obtain the abnormal fault point of the charging pile.

[0012] The present invention extracts the normal evolution trajectory of the charging pile state through normal evolution situation analysis, understands the normal evolution law of the charging pile state, and the extraction of the normal evolution trajectory provides a benchmark for subsequent steps for comparing and identifying abnormal situations. Based on the normal evolution trajectory, real-time production test and detection are performed on the charging pile, evolution trajectory deviation data is identified, and abnormal situations are captured in time. Real-time detection can help to discover existing problems in time, ensure the normal operation of the charging pile and the stability of the production process, calculate the deviation frequency of the evolution trajectory deviation data, obtain the frequency of trajectory deviation parameters, quantify the frequency of occurrence of abnormal situations, and calculate the deviation frequency. Quantify the severity of the abnormal situation and provide data support for the identification and processing of abnormal faults. Abnormal fault identification is performed based on the trajectory deviation parameter frequency, so as to find the abnormal fault point of the charging pile, help locate and solve the problem, and the identification of abnormal fault points helps the production test team to respond quickly and take necessary corrective measures to ensure the quality and reliability of charging pile production.

[0013] Preferably, the specific steps of step S31 are: Define sliding time windows; Based on the sliding time window, the charging pile status trend evolution map is divided into time windows and multiple time window segment samples are extracted; Performing time window state trend feature analysis on multiple time window samples, thereby obtaining the state trend feature of each time window; Quantify the state trend changes of the state trend characteristics of each time window segment and identify the state trend change values ​​in different time periods; Based on the state trend change values ​​in different time periods, the state value of the charging pile state trend evolution map is sampled to obtain the state point of each time period; Perform iterative moving average operation on the state points in each time period to obtain multiple smooth state points; Connect multiple smooth state points by trend evolution to obtain the state trend change path; Conduct normal evolution situation analysis on the state trend change path and extract the normal evolution trajectory.

[0014] The present invention uses a sliding time window, which is a time interval of fixed length, and is often used to capture local features and trends of data in time series data analysis. By defining a sliding time window, the time series data is divided into multiple overlapping time periods so that the data can be analyzed in more detail. By sliding the time window, multiple time window segment samples are extracted to capture the changing characteristics of the charging pile state in different time periods, making the analysis of the state trend more detailed and comprehensive, and providing a basis for subsequent feature extraction and change quantification. The time window segment state trend feature analysis understands the performance of the charging pile state in each time period, provides a basis for the quantification of state changes, obtains key state trend features, and provides key data support for subsequent state value sampling and path analysis. By quantifying the state trend change, the change value of the state trend in different time periods is identified, thereby quantifying the degree of state change, revealing the evolution law of the charging pile state, and providing support for subsequent state point extraction and path analysis. The analysis provides data support. By sampling the state value, the state points of the charging pile in each time period are obtained, which is conducive to a more detailed analysis of the state changes and provides basic data for subsequent smoothing and path connection, making the changes in the state trend clearer and more visual. The iterative moving average operation smoothes the state point data, reduces the noise and mutations in the data, makes the state changes more stable and continuous, provides a more stable data basis, and makes the path evolution more coherent. The trend evolution connection connects the smoothed state points in time to form a state trend change path, showing the overall evolution process of the state. This connection operation understands the overall change trend of the charging pile state and provides rich data support for the normal evolution situation analysis. The normal evolution situation analysis extracts the normal change trajectory of the charging pile state, reveals the general laws and trends of the state, and provides an in-depth understanding of the collection and analysis of the charging pile production test data, providing an important basis for subsequent decision-making and optimization.

[0015] Preferably, the specific steps of step S4 are: Step S41: analyzing the time point of abnormal parameter changes on the evolution trajectory deviation data to generate the time point of fault occurrence; Step S42: tracing the fault propagation path of the abnormal fault point of the charging pile based on the fault occurrence time point, and extracting the fault point propagation path; Step S43: reversely trace the fault point propagation path to locate the fault, thereby marking the fault location point; Step S44: Perform fault attribution analysis on the fault location point to obtain the fault cause of the abnormal part.

[0016] The present invention analyzes the evolution trajectory deviation data, identifies the time point of parameter anomaly changes, reveals the abnormal conditions existing in the process of charging pile state evolution, generates the time point of fault occurrence, accurately locates the time when the fault occurs, provides time clues for subsequent fault analysis, traces the fault propagation path, identifies the propagation relationship and path between the fault points, understands how the fault propagates and diffuses in the charging pile system, extracts the fault point propagation path to form an overall fault propagation map, provides basic data for subsequent fault tracing and positioning, and reverse tracing and positioning can track the source of the fault point propagation path, that is, find the starting point of the fault, so as to more accurately locate the existing fault part, mark the fault part positioning point to accurately locate the specific location of the fault, and provide accurate data support for subsequent fault attribution analysis. Fault attribution analysis aims to determine the cause of the fault at the abnormal part, help identify the root problem that causes the abnormality, and provide a basis for solving the fault. By analyzing the cause of the fault at the abnormal part, corresponding measures are taken to repair the fault, thereby improving the reliability and stability of the charging pile system.

[0017] Preferably, the specific steps of step S5 are: Step S51: predicting the future periodic fault evolution of the abnormal fault point of the charging pile according to the fault cause of the abnormal part, thereby obtaining fault evolution prediction data of multiple periods; Step S52: performing fault risk calculation on the fault evolution prediction data of multiple cycles to obtain a fault prediction risk value of each cycle; Step S53: quantifying the fault repair cost at the current time point for the fault prediction risk value of each cycle, thereby obtaining the fault repair costs at multiple time points; Step S54: performing optimal repair analysis based on the fault repair costs at multiple time points to extract the optimal repair time point; Step S55: Make a fault repair diagnosis decision based on the optimal repair time point and build a fault repair diagnosis strategy.

[0018] The present invention predicts the cause of failure of abnormal parts, generates fault evolution prediction data for multiple cycles, identifies potential fault development trends in advance, and provides a basis for formulating preventive measures and maintenance plans. It calculates the fault risk value, evaluates the probability of failure and the degree of impact of each cycle, helps prioritize high-risk fault problems, improves production efficiency and product reliability, quantifies the fault repair cost and evaluates the cost-effectiveness of fault repair, and helps decision makers make wise decisions in resource allocation and maintenance plan formulation. It determines the best fault repair time point through optimal repair analysis, balances maintenance costs and production losses, improves maintenance efficiency and reduces the risk of production interruption, constructs a fault repair diagnosis strategy to guide actual maintenance work, improves the accuracy and efficiency of fault repair, reduces maintenance time and cost, and ensures the normal operation of charging piles.

[0019] Preferably, the specific steps of step S6 are: Step S61: dynamically adjusting the production process parameters based on the fault repair diagnosis strategy to obtain multiple parameter adjustment schemes; Step S62: performing production simulation on the charging pile based on multiple parameter adjustment schemes to generate multiple parameter adjustment production response data; Step S63: multiple parameters are adjusted to produce response data to perform production performance evaluation, and obtain the production performance evaluation result of the adjustment scheme; Step S64: fine-tune and optimize the production parameters of the charging pile according to the production performance evaluation results of the adjustment plan, and build an intelligent production parameter control model.

[0020] The present invention improves the adaptability and flexibility of the production line by dynamically adjusting the production process parameters and responding to the results of the fault repair diagnosis strategy, generates multiple parameter adjustment schemes to explore different optimization paths, and provides diversified options for optimizing the production process. The impact of different parameter adjustment schemes on the production of charging piles is evaluated through production simulation to provide reference and prediction for actual production. Multiple parameter adjustment production response data are generated to understand the potential results of each scheme and provide a basis for further decision-making. The production performance evaluation quantifies the effects of different parameter adjustment schemes and helps determine the best adjustment scheme to improve production efficiency and quality. The production performance evaluation results of the adjustment schemes are obtained to compare the advantages and disadvantages of each scheme and provide guidance for further optimization. Through fine-tuning and optimization of production parameters, the production process is finely regulated according to the evaluation results to improve production efficiency and product quality. An intelligent production parameter regulation model is constructed to realize automated production parameter adjustment and enhance the intelligence level of the production line.

[0021] In this specification, a charging pile production test data collection and analysis system is provided, which is used to execute the charging pile production test data collection and analysis method as described above, including: The parameter processing module is used to perform real-time production monitoring sampling of charging piles based on multiple sensors, obtain real-time status monitoring parameters of multiple parts, perform time sequence segmentation and fitting on the real-time status monitoring parameters of multiple parts, and construct a time series sequence of status monitoring parameters of multiple parts; The state trend evolution module is used to mine the parameter change correlation of the time series of state monitoring parameters of multiple parts, and conduct multi-time point deep learning to build a charging pile state trend evolution map; The fault analysis module is used to perform real-time production test and detection on the charging pile status trend evolution map, and identify abnormal faults to obtain the abnormal fault points of the charging pile; The fault location module is used to trace the fault propagation path of the abnormal fault point of the charging pile, and then perform fault attribution analysis to obtain the cause of the fault at the abnormal part; The fault repair decision module is used to predict the future cycle fault evolution of the abnormal fault point of the charging pile according to the fault cause of the abnormal part, make fault repair diagnosis decisions, and build a fault repair diagnosis strategy; The fine-tuning optimization module is used to simulate the adjustment of production process parameters of charging piles based on the fault repair diagnosis strategy, generate multiple parameter adjustment production response data; perform production parameter fine-tuning optimization based on multiple parameter adjustment production response data, and build an intelligent production parameter control model.

[0022] The present invention constructs a time series of multi-part state monitoring parameters through real-time production monitoring sampling and time sequence segmentation and fitting of real-time state monitoring parameters of multiple parts, monitors the state changes of various parts of the charging pile in real time, provides a data basis for subsequent state analysis and fault diagnosis, and improves the visualization and monitoring level of the production process. Through the time series of multi-part state monitoring parameters, parameter change association mining and deep learning are carried out to construct a charging pile state trend evolution map, monitor the state changes of the charging pile in real time, discover potential problems in advance, optimize equipment maintenance and production plans, improve production efficiency and equipment reliability, and perform real-time production test detection and abnormal fault identification on the state trend evolution map, so as to quickly locate the abnormal fault point of the charging pile, provide High fault diagnosis speed and accuracy can reduce production interruption time and ensure production stability. Through fault propagation path tracing and fault attribution analysis, it can deeply understand the cause of fault in abnormal parts, help accurately locate the source of fault, and provide important basis for subsequent fault repair, reducing fault repair time and cost. Through future cycle fault evolution prediction and fault repair diagnosis decision-making, a fault repair diagnosis strategy can be built, which can effectively plan fault repair work, reduce maintenance costs, and improve the reliability and stability of production equipment. Through production process parameter adjustment simulation and intelligent production parameter control model construction, production parameter settings can be optimized, production efficiency and product quality can be improved, energy consumption and scrap rate can be reduced, and continuous optimization and improvement of the production process can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic diagram of the steps of a charging pile production test data collection and analysis method of the present invention; Figure 2 Detailed implementation flow chart of step S1; Figure 3 Detailed implementation flow chart of step S2; Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0024] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0025] The present application example provides a charging pile production test data collection and analysis method and system. The execution subject of the charging pile production test data collection and analysis method and system includes but is not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio and image management system, an information management system, and a cloud data management system.

[0026] See also Figures 1 to 4 The present invention provides a charging pile production test data collection and analysis method, the charging pile production test data collection and analysis method comprising the following steps: Step S1: Based on multiple sensors, real-time production monitoring sampling is performed on the charging pile to obtain real-time status monitoring parameters of multiple parts, and time sequence segmentation and fitting of the real-time status monitoring parameters of multiple parts are performed to construct a time series sequence of the status monitoring parameters of multiple parts; Step S2: Mining the correlation of parameter changes in the time series of multi-part status monitoring parameters, and conducting multi-time point deep learning to construct a charging pile status trend evolution map; Step S3: Perform real-time production test on the charging pile status trend evolution map and identify abnormal faults, thereby obtaining abnormal fault points of the charging pile; Step S4: tracing the fault propagation path of the abnormal fault point of the charging pile, and then performing fault attribution analysis to obtain the cause of the fault at the abnormal part; Step S5: predict the future periodic fault evolution of the abnormal fault point of the charging pile according to the fault cause of the abnormal part, make a fault repair diagnosis decision, and build a fault repair diagnosis strategy; Step S6: Based on the fault repair diagnosis strategy, the production process parameter adjustment simulation is performed on the charging pile to generate a plurality of parameter adjustment production response data; based on the plurality of parameter adjustment production response data, the production parameters are fine-tuned and optimized to construct an intelligent production parameter control model.

[0027] The present invention performs real-time production monitoring sampling of charging piles through multiple sensors to obtain real-time status monitoring parameters of multiple parts, constructs a time series of the status monitoring parameters of multiple parts, and understands the evolution process of the status of each part of the charging pile, providing basic data for subsequent analysis. The parameter change association mining and deep learning of the time series of the status monitoring parameters of multiple parts are performed to construct a charging pile status trend evolution map. The construction of the charging pile status trend evolution map reveals the correlation between parameters, providing a deeper analysis basis for abnormal fault identification, and performs real-time production test detection and abnormal fault identification on the charging pile status trend evolution map to find the abnormal fault point of the charging pile, detect and identify the abnormality in time, and improve production efficiency and product quality. quantity, reduce potential losses, trace the fault propagation path and analyze the fault attribution of abnormal fault points, find out the cause of the fault in the abnormal part, understand the cause of the fault and formulate a more effective repair strategy, improve the stability and reliability of the production line, predict the evolution of future cycle faults and make fault repair diagnosis decisions based on the cause of the fault in the abnormal part, build a fault repair diagnosis strategy, predict fault evolution and formulate a diagnosis strategy to reduce future failure risks, improve the stability of the production line, simulate the adjustment of production process parameters based on the fault repair diagnosis strategy, generate production response data, and perform parameter fine-tuning optimization, build an intelligent production parameter control model, adjust and optimize parameters through simulation, improve production efficiency and product quality, and realize intelligent production control.

[0028] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of a charging pile production test data collection and analysis method of the present invention. In this example, the steps of the charging pile production test data collection and analysis method include: Step S1: Based on multiple sensors, real-time production monitoring sampling is performed on the charging pile to obtain real-time status monitoring parameters of multiple parts, and time sequence segmentation and fitting of the real-time status monitoring parameters of multiple parts are performed to construct a time series sequence of the status monitoring parameters of multiple parts; In this embodiment, a series of various types of sensors are deployed at key parts of the charging pile, such as the power module, the control unit, the charging connector, etc., including temperature sensors, current sensors, voltage sensors, etc. These sensors can collect process parameter data of various parts in real time, such as temperature, current, voltage and other physical quantities. Through an integrated data acquisition system, these real-time monitoring data are uniformly collected and stored at a high sampling frequency (such as 10 times per second). After sensor acquisition and processing by the data acquisition system, real-time process parameter data sequences of key parts of the charging pile are obtained. These data sequences reflect the real-time state changes of various parts of the charging pile during the production process. For example, the input voltage sequence of the power module, the control unit, and the charging connector are obtained. The operating temperature sequence of the control unit, the output current sequence of the charging connector, etc., are arranged and aligned in chronological order. For each data sequence, the time series analysis method (such as Fourier transform, autoregressive model, etc.) is used to perform curve fitting and smoothing. In this way, a smooth curve of the state parameters of each part changing with time is obtained, which eliminates the noise interference in the sampling process. The processed state parameter curves of each part are combined into a multidimensional time series data set in chronological order. This time series sequence records the state change trajectory of each key part of the charging pile during the entire production process. Subsequent fault analysis and root cause location will be based on this multidimensional time series data.

[0029] Step S2: Mining the correlation of parameter changes in the time series of multi-part status monitoring parameters, and conducting multi-time point deep learning to construct a charging pile status trend evolution map; In this embodiment, correlation analysis, causal reasoning and other methods are used to deeply explore the inherent correlation between the monitoring parameters, such as analyzing the correlation between the input voltage change of the power module and the temperature change of the control unit, or the causal relationship between the current change of the charging connector and the overall efficiency. In this way, a correlation model between the process parameters of various parts of the charging pile is established to reflect the mutual influence mechanism between them. Deep learning technology is used to take time series parameter data as input to establish a prediction model that can describe the law of charging pile state evolution. Specifically, a time series deep learning model such as a long short-term memory (LSTM) network is used to learn the changing characteristics of the charging pile parameters in the time dimension. By training the state data of multiple time nodes, the model can capture the complex time series dependency between the parameters. Based on the predictive ability of the deep learning model, the changing trends of the parameters of various parts of the charging pile in the future time series are simulated and predicted. These predicted parameter change trends are intuitively presented in a graphical form to form a time series map of the charging pile state evolution. This map not only shows the state of the process parameters at the current moment, but also predicts the state changes that will occur in the future.

[0030] Step S3: Perform real-time production test on the charging pile status trend evolution map and identify abnormal faults, thereby obtaining abnormal fault points of the charging pile; In this embodiment, the established charging pile status trend evolution map is integrated into the real-time production monitoring system of the charging pile. The system will receive process parameter data from multiple sensors in real time, and compare and analyze it with the predicted status trend map. Through real-time comparison of time series data, the system can monitor in real time whether the current operating status of the charging pile is consistent with the predicted trend. If the current process parameter changes of the charging pile deviate significantly from the predicted status trend, the system will identify this abnormal situation. The system will conduct an in-depth analysis of this deviation in combination with the previously established parameter association model. If the output current suddenly drops, but the predicted trend shows that it should be stable, then the control unit has a fault. By comparing and analyzing the real-time monitoring data with the predicted trend, the system can quickly locate the specific part of the charging pile where the abnormal fault occurs. The fault point information will be marked in the digital twin model of the charging pile to form an intuitive fault diagnosis result. Subsequent fault tracing and root cause analysis will all conduct in-depth exploration around these abnormal fault points.

[0031] Step S4: tracing the fault propagation path of the abnormal fault point of the charging pile, and then performing fault attribution analysis to obtain the cause of the fault at the abnormal part; In this embodiment, a causal network model of fault propagation is established in combination with the overall structure and process flow of the charging pile. The model describes the logical connection between the functional modules of the charging pile and the path of fault propagation between these modules. Using this fault propagation model, starting from the determined abnormal fault point, the paths along which the fault propagates are traced back. Through probability reasoning and path analysis algorithms, the key propagation nodes and paths experienced by the fault are obtained. For the fault propagation path determined in the previous step, the key parts along the path are deeply analyzed. The multi-dimensional process parameter data collected in step S1 is used, combined with the experience and knowledge of industry experts, to explore the root causes of the faults in these parts. For example, by analyzing the time series of the input voltage fluctuations of the power module and combining the circuit topology, the fault is located to be caused by the aging of the power chip, or by identifying the abnormal pattern of the current data of the charging connector, it is found that the fault is caused by the wear of the contact parts. After the above-mentioned fault propagation tracing and fault attribution analysis, the root cause of the abnormal fault in the charging pile can be finally determined. These root cause information will be fed back to the production management personnel to provide decision support for subsequent fault repair and process optimization.

[0032] Step S5: predict the future periodic fault evolution of the abnormal fault point of the charging pile according to the fault cause of the abnormal part, make a fault repair diagnosis decision, and build a fault repair diagnosis strategy; In this embodiment, combined with the fault root cause analysis results obtained in the previous step, a prediction model based on the fault mechanism is constructed. The model uses the process parameter change characteristics of the fault root cause, such as aging of the power module, wear of the contact parts, etc., to predict the development trend of the fault in the future. Through time series analysis and life prediction methods, the trajectory of the fault evolution over time is simulated to predict when the fault will occur again in the future. These fault evolution prediction results provide an important time window for subsequent fault repair. Combined with the fault root cause analysis and fault evolution prediction, targeted fault repair diagnosis strategies are formulated. For example, for the aging fault of the power module, a repair measure of replacing key parts is proposed, and its impact on production is evaluated. For the wear fault of the contact parts, it is recommended to increase the regular inspection and maintenance cycle and arrange the maintenance time reasonably. At the same time, the cost-effectiveness of different repair plans will be evaluated to provide the best decision support for production management personnel. The above-mentioned fault analysis, prediction and repair decisions are integrated into a set of standardized fault repair diagnosis strategies. This strategy not only includes fault location, cause analysis, prediction and early warning, but also covers specific repair measures and time arrangements.

[0033] Step S6: Based on the fault repair diagnosis strategy, the production process parameter adjustment simulation is performed on the charging pile to generate a plurality of parameter adjustment production response data; based on the plurality of parameter adjustment production response data, the production parameters are fine-tuned and optimized to construct an intelligent production parameter control model.

[0034] In this embodiment, the fault repair diagnosis strategy established in the previous step is used to specifically adjust the key process parameters of the charging pile, such as power supply voltage, control frequency, cooling temperature, etc. This parameter adjustment simulation is intended to verify the impact of different fault repair measures on the production process. The digital twin technology based on the physical model is used to synchronize the virtual model of the charging pile with the actual production process in real time to achieve simulation testing of parameter adjustment. By analyzing the simulation results, the production response brought about by various repair measures is predicted to provide a basis for the final process optimization. In the parameter adjustment simulation process, different process parameter combinations are systematically tried and the corresponding production response data are recorded. These data include various key performance indicators such as product quality indicators, production efficiency, and energy consumption levels. By adjusting a large number of parameters, the simulation results can be used to predict the production response brought about by various repair measures. The testing of the adjustment plan has accumulated a rich production response data set, providing a basis for subsequent optimization modeling. Using the above production response data set, an intelligent production parameter optimization model was constructed using machine learning technology. This model can learn the complex relationship between process parameters and production performance, and find the optimal parameter configuration scheme. It uses an optimization algorithm based on reinforcement learning, and continuously approaches the optimal production process parameters through repeated trial and error and model updating. The optimal parameter configuration obtained by the above optimization is summarized into a standardized intelligent production parameter control model. This model can automatically adjust the process parameters according to the real-time production status to keep the production performance of the charging pile in the best state. This control model is embedded in the charging pile manufacturing execution system to realize intelligent closed-loop optimization of the production process.

[0035] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Step S11: Perform real-time production monitoring sampling on the charging pile based on multiple sensors to obtain real-time status monitoring parameters of multiple parts; Step S12: performing multi-point parameter mean calculation on the real-time status monitoring parameters of multiple parts to generate standard monitoring ranges for different parts; Step S13: performing outlier detection on the real-time status monitoring parameters of multiple parts based on the standard monitoring ranges of different parts, and extracting abnormal outlier parameters; Step S14: performing outlier filtering processing based on the abnormal outlier parameters to generate outlier filtering state monitoring parameters; Step S15: performing low-frequency filtering and denoising on the outlier filtering state monitoring parameters to obtain filtered denoised state monitoring parameters; Step S16: Perform time-sequential segmentation fitting on the filtered and denoised state monitoring parameters to construct a time series of multi-part state monitoring parameters.

[0036] In this embodiment, a variety of sensors (such as temperature sensors, current sensors, pressure sensors, etc.) are installed at key parts of the charging pile (such as the battery management system, charging interface, heat dissipation system, etc.), and the sensors are set to collect status monitoring parameters in real time, such as current, voltage, temperature, etc., record sampling time and parameter values, and store the collected real-time status monitoring parameters of multiple parts in a database for subsequent analysis and processing. The real-time status monitoring parameters of different parts are extracted from the database, grouped and sorted by time, and the parameters of each part are averaged at multiple time points to generate a standard monitoring range for each part (such as mean value ± standard deviation). The standard monitoring range of each part is recorded, and a report is generated. A suitable outlier detection algorithm (such as Z-score, IQR method) is selected for outlier identification. Based on the defined standard monitoring range, outlier detection is performed on the real-time status monitoring parameters of multiple parts to extract outliers that exceed the standard range. Outlier parameters, record their specific information (such as timestamps and parameter values), determine the standards and methods for outlier filtering (such as filtering based on the median or mean), analyze the extracted abnormal outlier parameters, apply the set filtering standards to remove outliers, generate state monitoring parameters after outlier filtering, ensure the accuracy of the data, select a suitable low-frequency filtering algorithm (such as low-pass filtering or moving average filtering) for denoising, perform low-frequency filtering on the generated outlier filtered state monitoring parameters, remove high-frequency noise, generate filtered denoised state monitoring parameters, ensure the smoothness of the data, save the filtered denoised state monitoring parameters as a data set, divide the filtered denoised state monitoring parameters in chronological order, determine the parameter value of each time period, perform fitting analysis on the parameters of each time period, generate a time series of multi-part state monitoring parameters, ensure the continuity and accuracy of the sequence data, and record the parameter change trend of each time period.

[0037] In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Step S21: performing time series parameter change analysis on the time series of multi-part state monitoring parameters to generate parameter change trends of different parts; Step S22: performing component topology dependency analysis on multiple parts of the charging pile to extract topology dependency relationships between the parts; Step S23: performing parameter change association mining on parameter change trends of different parts according to the topological dependency relationship between the parts, and extracting the parameter change association rules of different parts; Step S24: performing dynamic co-evolution analysis on the correlation rules of parameter changes of different parts to generate state co-evolution data of different parts; Step S25: Perform multi-time-point deep learning on the state collaborative evolution data of different parts to construct a charging pile state trend evolution map.

[0038] In this embodiment, a suitable time series analysis method (such as trend analysis and seasonal decomposition) is selected to perform parameter change analysis, a trend analysis is performed on the parameter time series of each part, the change pattern and trend of each part are identified, a parameter change trend diagram of different parts is generated, key change points and trend characteristics are recorded, and a suitable topological analysis tool (such as graph theory or network analysis) is selected to analyze the dependency relationship between components. Based on the structure and function of the charging pile, a topological dependency analysis is performed on multiple parts, a dependency diagram between components is drawn, and the dependency relationship between each part and other parts is recorded, including direct dependency and indirect dependency. A suitable association rule mining algorithm (such as Apriori or FP-Growth) is selected to perform parameter change association analysis. Based on the extracted topological dependency, the parameter change trend of different parts is analyzed. The correlation mining is carried out based on the potential, the correlation laws of parameter changes between different parts are extracted, the influence and reaction relationship between them are analyzed, and the appropriate dynamic analysis method (such as dynamic system modeling or collaborative filtering algorithm) is selected for evolution analysis. Based on the extracted correlation laws of parameter changes, the state co-evolution of different parts is analyzed, and the state co-evolution data of different parts are generated. The mutual influence and changes of each part at different time points are recorded, and the appropriate deep learning model (such as LSTM, GRU or convolutional neural network) is selected for data modeling. The generated state co-evolution data is used as input to prepare training and test data sets, train the deep learning model, learn the time evolution law of the state of different parts of the charging pile, generate a charging pile state trend evolution map, and show the evolution process of the state of different parts and their relationship.

[0039] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Step S31: Perform normal evolution situation analysis on the charging pile state trend evolution map to extract the normal evolution trajectory; Step S32: Performing real-time production testing on the charging pile based on the normalized evolution trajectory to identify evolution trajectory deviation data; Step S33: Calculate the deviation frequency of the evolving trajectory deviation data to obtain the trajectory deviation parameter frequency; Step S34: abnormal fault identification is performed based on the trajectory deviation parameter frequency, so as to obtain the abnormal fault point of the charging pile.

[0040] In this embodiment, a state trend evolution map of the charging pile is extracted, including the time series changes of various monitoring parameters (such as current, voltage, temperature, etc.), and a suitable analysis method (such as time series analysis or moving average) is selected to identify the normal evolution trajectory. The extracted data is smoothed to eliminate noise, and the normal evolution trajectory is identified. The characteristics of the normal evolution trajectory, including the starting point, the end point and the change trend, are recorded. The current state parameter data is collected in real time from the sensor of the charging pile, and the state parameters collected in real time are compared with the extracted normal evolution trajectory. The data points with significant deviations from the normal trajectory are identified, and these deviation data are recorded. The method for calculating the deviation frequency (such as calculating the ratio of the number of deviation occurrences to the total number of monitoring times) is determined, the extracted evolution trajectory deviation data is analyzed, the frequency of occurrence of each deviation type is calculated, and trajectory deviation parameter frequency data is generated, including the frequency percentage of each deviation type. The identification standard of abnormal faults is determined (such as the deviation frequency exceeds a certain threshold). Based on the calculated trajectory deviation parameter frequency, it is determined whether there is an abnormal fault, and all fault points identified as abnormal are recorded, including the fault type, occurrence time and related parameters. The identified abnormal fault points of the charging pile are saved as a data set for subsequent processing and response.

[0041] In this embodiment, the specific steps of step S31 are: Define sliding time windows; Based on the sliding time window, the charging pile status trend evolution map is divided into time windows and multiple time window segment samples are extracted; Performing time window state trend feature analysis on multiple time window samples, thereby obtaining the state trend feature of each time window; Quantify the state trend changes of the state trend characteristics of each time window segment and identify the state trend change values ​​in different time periods; Based on the state trend change values ​​in different time periods, the state value of the charging pile state trend evolution map is sampled to obtain the state point of each time period; Perform iterative moving average operation on the state points in each time period to obtain multiple smooth state points; Connect multiple smooth state points by trend evolution to obtain the state trend change path; Conduct normal evolution situation analysis on the state trend change path and extract the normal evolution trajectory.

[0042] In this embodiment, the length of the sliding time window (such as 5 minutes, 10 minutes) and the sliding step size (such as 1 minute, 2 minutes) are determined, and the appropriate window type (such as a fixed window or a variable window) is selected according to the needs. Time series data, including various monitoring parameters (such as current, voltage, temperature, etc.) are extracted from the state monitoring data of the charging pile. According to the defined sliding time window, the charging pile state trend evolution map is divided, and multiple time window segment samples are extracted. Each sample corresponds to a sliding time window, and an appropriate time series feature extraction method (such as mean, standard deviation, maximum value, minimum value, etc.) is determined. The state trend feature analysis is performed on each time window segment sample, and various feature indicators are calculated. The state trend features of each time window segment are recorded, including the trend direction and intensity, and a quantitative method for the state trend change (such as a differential method or a percentage change) is determined. The state trend features of each time window segment are quantified, and the state trend change values ​​of different time periods are identified. Each change value and its corresponding time period are recorded to determine the state trend. The state value sampling standard (such as sampling frequency or sampling under specific conditions) is used to sample the state value of the charging pile state trend evolution map based on the state trend change values ​​in different time periods, record the state points in each time period to ensure the representativeness and accuracy of the data, determine the window size of the moving average (such as 3 points, 5 points) to smooth the state point data, perform iterative moving average operations on the state points in each time period, generate multiple smoothed state points, ensure that the smoothed data can reflect trend changes and reduce the impact of noise, determine the method of connecting state points (such as linear interpolation or spline interpolation), perform trend evolution connections on multiple smoothed state points, generate state trend change paths, ensure that the generated paths can accurately reflect the trends and characteristics of state changes, select appropriate normal evolution situation analysis methods (such as statistical analysis or machine learning models), perform normal evolution situation analysis on the generated state trend change paths, extract normal evolution trajectories, and record the characteristics of normal evolution trajectories, including change patterns and duration.

[0043] In this embodiment, the specific steps of step S4 are: Step S41: analyzing the time point of abnormal parameter changes on the evolution trajectory deviation data to generate the time point of fault occurrence; Step S42: tracing the fault propagation path of the abnormal fault point of the charging pile based on the fault occurrence time point, and extracting the fault point propagation path; Step S43: reversely trace the fault point propagation path to locate the fault, thereby marking the fault location point; Step S44: Perform fault attribution analysis on the fault location point to obtain the fault cause of the abnormal part.

[0044] In this embodiment, the real-time monitoring data of the key process parameters (such as voltage, current, temperature, etc.) of the charging pile are analyzed by using the time series analysis technology. The time nodes where the various parameters change abnormally are identified by using the outlier detection algorithm. These abnormal change time points are summarized to determine the time point when the fault first occurs. In combination with the circuit topology and process flow of the charging pile, a causal network model of fault propagation is constructed. Using this model, starting from the time point when the fault occurs, the key parts where the fault exists are traced upstream and downstream. Through probability reasoning and path analysis, it is obtained along which key paths the fault propagates. For the fault propagation path determined in the previous step, the reverse tracing method is used to locate the specific parts where the fault occurs in turn. The multi-dimensional sensor data inside the charging pile is used in combination with the fault propagation model to accurately determine the specific component location where the fault occurs. These fault parts are marked in the overall structural model of the charging pile to form an intuitive fault location result. For the determined fault part, the historical change trend and influencing factors of its process parameters are deeply analyzed. In combination with the experience and knowledge of industry experts, methods such as fault tree analysis are used to dig out the root cause of the fault at this part. The analysis results are fed back to the production management personnel to provide decision support for subsequent fault repair and process optimization.

[0045] In this embodiment, the specific steps of step S5 are: Step S51: predicting the future periodic fault evolution of the abnormal fault point of the charging pile according to the fault cause of the abnormal part, thereby obtaining fault evolution prediction data of multiple periods; Step S52: performing fault risk calculation on the fault evolution prediction data of multiple cycles to obtain a fault prediction risk value of each cycle; Step S53: quantifying the fault repair cost at the current time point for the fault prediction risk value of each cycle, thereby obtaining the fault repair costs at multiple time points; Step S54: performing optimal repair analysis based on the fault repair costs at multiple time points to extract the optimal repair time point; Step S55: Make a fault repair diagnosis decision based on the optimal repair time point and build a fault repair diagnosis strategy.

[0046] In this embodiment, the historical fault data of the charging pile and the causes of faults in abnormal parts are collected, the pattern and periodicity of the fault occurrence are analyzed, and a suitable time series prediction model (such as ARIMA, LSTM or regression analysis) is selected to predict the fault evolution. The historical data and the causes of the faults are used to train the prediction model to predict the fault evolution in future cycles, generate fault evolution prediction data for multiple cycles, record the probability of fault occurrence and the degree of impact of each cycle, determine a suitable risk assessment method (such as fault tree analysis (FTA) or fuzzy logic) to calculate the fault risk, analyze the fault evolution prediction data of each cycle, calculate the fault prediction risk value of each cycle, record the risk value of each cycle, consider factors such as the probability of failure, the consequences of impact and the coping ability, and determine the quantitative method of the fault repair cost (such as based on historical repair costs, material costs and Comprehensive assessment of labor costs), analyze the fault prediction risk value of each cycle, quantify the fault repair cost at the current time point, record the fault repair cost at each time point, including direct cost and potential loss, select a suitable optimization analysis model (such as linear programming or dynamic programming) to find the optimal repair time point, based on the fault repair cost at multiple time points, use the optimization model to calculate the optimal repair time point to ensure the best balance between cost and risk, record the optimal repair time point and its corresponding cost and risk value, determine the appropriate decision analysis method (such as decision tree analysis, cost-benefit analysis) to construct a repair diagnosis strategy, based on the optimal repair time point, formulate a detailed fault repair diagnosis strategy, including repair steps, resource allocation and time arrangement, to ensure that the strategy can effectively respond to potential failures, reduce risks and reduce downtime.

[0047] In this embodiment, the specific steps of step S6 are: Step S61: dynamically adjusting the production process parameters based on the fault repair diagnosis strategy to obtain multiple parameter adjustment schemes; Step S62: performing production simulation on the charging pile based on multiple parameter adjustment schemes to generate multiple parameter adjustment production response data; Step S63: multiple parameters are adjusted to produce response data to perform production performance evaluation, and obtain the production performance evaluation result of the adjustment scheme; Step S64: fine-tune and optimize the production parameters of the charging pile according to the production performance evaluation results of the adjustment plan, and build an intelligent production parameter control model.

[0048] In this embodiment, based on the fault repair diagnosis strategy, key process parameters related to the production of charging piles (such as current, pressure, temperature, etc.) are identified, and multiple parameter adjustment plans are formulated, taking into account different process requirements and fault repair needs, such as increasing the charging current, improving the charging efficiency, adjusting the operating temperature range of the heat dissipation system, and preventing overheating. Suitable production simulation tools or software (such as MATLAB / Simulink, AnyLogic, etc.) are selected to simulate the production process, and the generated multiple parameter adjustment plans are input into the simulation tool. The production environment and operating conditions are set, and the simulation is run to observe the working status and performance of the charging pile under each parameter adjustment plan, and determine the standards and indicators for production performance evaluation (such as charging efficiency). The system analyzes the collected production response data of multiple parameter adjustments, applies statistical analysis methods (such as variance analysis, regression analysis, etc.) to evaluate the performance of each scheme, compares the effects of different parameter adjustment schemes, generates evaluation results and corresponding charts, determines the appropriate optimization algorithm (such as genetic algorithm, particle swarm optimization, simulated annealing, etc.) to fine-tune the production parameters, identifies the best performance parameter combination based on the evaluation results, performs further fine-tuning and optimization, generates an intelligent production parameter control model, and can automatically adjust parameters to cope with different production needs and failure situations. The effectiveness of the optimization model is verified in the actual production environment to ensure that it can improve production efficiency and reduce the failure rate.

[0049] In this embodiment, a charging pile production test data collection and analysis system is provided, which is used to execute the charging pile production test data collection and analysis method as described above, including: The parameter processing module is used to perform real-time production monitoring sampling of charging piles based on multiple sensors, obtain real-time status monitoring parameters of multiple parts, perform time sequence segmentation and fitting on the real-time status monitoring parameters of multiple parts, and construct a time series sequence of status monitoring parameters of multiple parts; The state trend evolution module is used to mine the parameter change correlation of the time series of state monitoring parameters of multiple parts, and conduct multi-time point deep learning to build a charging pile state trend evolution map; The fault analysis module is used to perform real-time production test and detection on the charging pile status trend evolution map, and identify abnormal faults to obtain the abnormal fault points of the charging pile; The fault location module is used to trace the fault propagation path of the abnormal fault point of the charging pile, and then perform fault attribution analysis to obtain the cause of the fault at the abnormal part; The fault repair decision module is used to predict the future cycle fault evolution of the abnormal fault point of the charging pile according to the fault cause of the abnormal part, make fault repair diagnosis decisions, and build a fault repair diagnosis strategy; The fine-tuning optimization module is used to simulate the adjustment of production process parameters of charging piles based on the fault repair diagnosis strategy, generate multiple parameter adjustment production response data; perform production parameter fine-tuning optimization based on multiple parameter adjustment production response data, and build an intelligent production parameter control model.

[0050] The present invention constructs a time series of multi-part state monitoring parameters through real-time production monitoring sampling and time sequence segmentation and fitting of real-time state monitoring parameters of multiple parts, monitors the state changes of various parts of the charging pile in real time, provides a data basis for subsequent state analysis and fault diagnosis, and improves the visualization and monitoring level of the production process. Through the time series of multi-part state monitoring parameters, parameter change association mining and deep learning are carried out to construct a charging pile state trend evolution map, monitor the state changes of the charging pile in real time, discover potential problems in advance, optimize equipment maintenance and production plans, improve production efficiency and equipment reliability, and perform real-time production test detection and abnormal fault identification on the state trend evolution map, so as to quickly locate the abnormal fault point of the charging pile, provide High fault diagnosis speed and accuracy can reduce production interruption time and ensure production stability. Through fault propagation path tracing and fault attribution analysis, it can deeply understand the cause of fault in abnormal parts, help accurately locate the source of fault, and provide important basis for subsequent fault repair, reducing fault repair time and cost. Through future cycle fault evolution prediction and fault repair diagnosis decision-making, a fault repair diagnosis strategy can be built, which can effectively plan fault repair work, reduce maintenance costs, and improve the reliability and stability of production equipment. Through production process parameter adjustment simulation and intelligent production parameter control model construction, production parameter settings can be optimized, production efficiency and product quality can be improved, energy consumption and scrap rate can be reduced, and continuous optimization and improvement of the production process can be achieved.

[0051] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0052] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A charging pile production test data collection and analysis method, characterized in that: The following steps are involved: Step S1: Based on multiple sensors, real-time production monitoring sampling is performed on the charging pile to obtain real-time status monitoring parameters of multiple parts, and time sequence segmentation and fitting of the real-time status monitoring parameters of multiple parts are performed to construct a time series sequence of the status monitoring parameters of multiple parts; Step S2: Mining the correlation of parameter changes in the time series of multi-part status monitoring parameters, and conducting multi-time point deep learning to construct a charging pile status trend evolution map; Step S3: Perform real-time production test on the charging pile status trend evolution map and identify abnormal faults, thereby obtaining abnormal fault points of the charging pile; Step S4: tracing the fault propagation path of the abnormal fault point of the charging pile, and then performing fault attribution analysis to obtain the cause of the fault at the abnormal part; Step S5: predict the future periodic fault evolution of the abnormal fault point of the charging pile according to the fault cause of the abnormal part, make a fault repair diagnosis decision, and build a fault repair diagnosis strategy; Step S6: simulating the production process parameter adjustment of the charging pile based on the fault repair diagnosis strategy to generate a plurality of parameter adjustment production response data; Adjust the production response data based on multiple parameters to fine-tune and optimize the production parameters and build an intelligent production parameter control model.

2. The charging pile production test data collection and analysis method according to claim 1 is characterized in that: The specific steps of step S1 are: Step S11: Perform real-time production monitoring sampling on the charging pile based on multiple sensors to obtain real-time status monitoring parameters of multiple parts; Step S12: performing multi-point parameter mean calculation on the real-time status monitoring parameters of multiple parts to generate standard monitoring ranges for different parts; Step S13: performing outlier detection on the real-time status monitoring parameters of multiple parts based on the standard monitoring ranges of different parts, and extracting abnormal outlier parameters; Step S14: performing outlier filtering processing based on the abnormal outlier parameters to generate outlier filtering state monitoring parameters; Step S15: performing low-frequency filtering and denoising on the outlier filtering state monitoring parameters to obtain filtered denoised state monitoring parameters; Step S16: Perform time-sequential segmentation fitting on the filtered and denoised state monitoring parameters to construct a time series of multi-part state monitoring parameters.

3. The charging pile production test data collection and analysis method according to claim 1 is characterized in that: The specific steps of step S2 are: Step S21: performing time series parameter change analysis on the time series of multi-part state monitoring parameters to generate parameter change trends of different parts; Step S22: performing component topology dependency analysis on multiple parts of the charging pile to extract topology dependency relationships between the parts; Step S23: performing parameter change association mining on parameter change trends of different parts according to the topological dependency relationship between the parts, and extracting the parameter change association rules of different parts; Step S24: performing dynamic co-evolution analysis on the correlation rules of parameter changes of different parts to generate state co-evolution data of different parts; Step S25: Perform multi-time-point deep learning on the state collaborative evolution data of different parts to construct a charging pile state trend evolution map.

4. The charging pile production test data collection and analysis method according to claim 1 is characterized in that: The specific steps of step S3 are: Step S31: Perform normal evolution situation analysis on the charging pile state trend evolution map to extract the normal evolution trajectory; Step S32: Performing real-time production testing on the charging pile based on the normalized evolution trajectory to identify evolution trajectory deviation data; Step S33: Calculate the deviation frequency of the evolving trajectory deviation data to obtain the trajectory deviation parameter frequency; Step S34: abnormal fault identification is performed based on the trajectory deviation parameter frequency, so as to obtain the abnormal fault point of the charging pile.

5. The charging pile production test data collection and analysis method according to claim 4 is characterized in that: The specific steps of step S31 are: Define sliding time windows; Based on the sliding time window, the charging pile status trend evolution map is divided into time windows and multiple time window segment samples are extracted; Performing time window state trend feature analysis on multiple time window samples, thereby obtaining the state trend feature of each time window; Quantify the state trend changes of the state trend characteristics of each time window segment and identify the state trend change values ​​in different time periods; Based on the state trend change values ​​in different time periods, the state value of the charging pile state trend evolution map is sampled to obtain the state point of each time period; Perform iterative moving average operation on the state points in each time period to obtain multiple smooth state points; Connect multiple smooth state points by trend evolution to obtain the state trend change path; Conduct normal evolution situation analysis on the state trend change path and extract the normal evolution trajectory.

6. The charging pile production test data collection and analysis method according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: analyzing the time point of abnormal parameter changes on the evolution trajectory deviation data to generate the time point of fault occurrence; Step S42: tracing the fault propagation path of the abnormal fault point of the charging pile based on the fault occurrence time point, and extracting the fault point propagation path; Step S43: reversely trace the fault point propagation path to locate the fault, thereby marking the fault location point; Step S44: Perform fault attribution analysis on the fault location point to obtain the fault cause of the abnormal part.

7. The charging pile production test data collection and analysis method according to claim 1 is characterized in that: The specific steps of step S5 are: Step S51: predicting the future periodic fault evolution of the abnormal fault point of the charging pile according to the fault cause of the abnormal part, thereby obtaining fault evolution prediction data of multiple periods; Step S52: performing fault risk calculation on the fault evolution prediction data of multiple cycles to obtain a fault prediction risk value of each cycle; Step S53: quantifying the fault repair cost at the current time point for the fault prediction risk value of each cycle, thereby obtaining the fault repair costs at multiple time points; Step S54: performing optimal repair analysis based on the fault repair costs at multiple time points to extract the optimal repair time point; Step S55: Make a fault repair diagnosis decision based on the optimal repair time point and build a fault repair diagnosis strategy.

8. The charging pile production test data collection and analysis method according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: dynamically adjusting the production process parameters based on the fault repair diagnosis strategy to obtain multiple parameter adjustment schemes; Step S62: performing production simulation on the charging pile based on multiple parameter adjustment schemes to generate multiple parameter adjustment production response data; Step S63: multiple parameters are adjusted to produce response data to perform production performance evaluation, and obtain the production performance evaluation result of the adjustment scheme; Step S64: fine-tune and optimize the production parameters of the charging pile according to the production performance evaluation results of the adjustment plan, and build an intelligent production parameter control model.

9. A charging pile production test data acquisition and analysis system, characterized in that: The method for collecting and analyzing charging pile production test data according to claim 1 comprises: The parameter processing module is used to perform real-time production monitoring sampling of charging piles based on multiple sensors, obtain real-time status monitoring parameters of multiple parts, perform time sequence segmentation and fitting on the real-time status monitoring parameters of multiple parts, and construct a time series sequence of status monitoring parameters of multiple parts; The state trend evolution module is used to perform parameter change correlation mining on the time series of state monitoring parameters of multiple parts, and conduct multi-time point deep learning to construct a charging pile state trend evolution map; The fault analysis module is used to perform real-time production test and detection on the charging pile status trend evolution map, and identify abnormal faults to obtain the abnormal fault points of the charging pile; The fault location module is used to trace the fault propagation path of the abnormal fault point of the charging pile, and then perform fault attribution analysis to obtain the cause of the fault at the abnormal part; The fault repair decision module is used to predict the future cycle fault evolution of the abnormal fault point of the charging pile according to the fault cause of the abnormal part, make fault repair diagnosis decisions, and build a fault repair diagnosis strategy; The fine-tuning optimization module is used to simulate the adjustment of production process parameters of charging piles based on the fault repair diagnosis strategy, generate multiple parameter adjustment production response data; perform production parameter fine-tuning optimization based on multiple parameter adjustment production response data, and build an intelligent production parameter control model.

Images