New energy charging pile operation state monitoring and optimal scheduling method based on cloud computing

Abstract

The invention provides a new energy charging pile operation state monitoring and optimal scheduling method based on cloud computing, and relates to the technical field of new energy, and the method comprises the steps: collecting charging pile data, and uploading the data to a cloud server; constructing a charging pile health assessment model to carry out health state comprehensive assessment and dynamic classification on the charging piles; a load prediction model is constructed based on user reservation data and historical data, a charging load balance coefficient is calculated, and a hierarchical scheduling strategy is implemented in combination with peak, flat peak and valley periods. Therefore, the problems of insufficient running state monitoring capability, inaccurate health state evaluation, low load prediction precision and single scheduling strategy of the charging pile in the prior art can be solved, equipment overload and user waiting time are reduced, and meanwhile, the quality and reliability of charging service are improved.

Description

Technical Field

[0001] The present invention relates to the field of new energy technology, and more specifically, to a new energy charging pile operation status monitoring and optimization scheduling method based on cloud computing. Background Art

[0002] With the transformation of the global energy structure and the rapid development of the low-carbon economy, the application of new energy technologies in the field of transportation has become increasingly widespread, and the penetration rate of new energy vehicles has increased significantly. In line with this, new energy charging piles, as the core infrastructure for energy supply of new energy vehicles, have been rapidly deployed and promoted. At present, the construction of charging piles not only needs to meet the growing demand for charging, but also needs to improve the intelligence and efficiency of charging services to cope with problems such as power load fluctuations, diversified user needs, and complex equipment operation and maintenance management. Traditional charging pile operation and management methods mainly rely on fixed rules or manual intervention, lack of real-time monitoring and analysis of the operating status of charging piles, and it is difficult to achieve intelligent load prediction and optimal scheduling. In addition, the existing technology still has deficiencies in processing the health status assessment of charging pile equipment. It is usually based on a single static indicator and fails to fully combine dynamic operation data for comprehensive assessment, resulting in equipment operation risks that are difficult to predict and control in a timely manner.

[0003] The shortcomings of the existing technology are mainly reflected in the following aspects: First, the operating status monitoring capability of the charging pile equipment is limited, and it can usually only achieve basic fault alarms or simple data collection, and fails to deeply explore the potential value of the operating data, so it is impossible to effectively identify the changing trend of the equipment health status; secondly, in terms of charging load management, most of the existing methods are based on simple empirical data or linear prediction models, lack of multi-dimensional comprehensive analysis of user reservation behavior, historical data and real-time load, and it is difficult to deal with the problem of peak power load imbalance; the current scheduling strategy usually fails to fully consider the differences between the status of charging pile equipment and the use period, lacks refined hierarchical management and optimized scheduling mechanism, and therefore has a large room for improvement in charging service efficiency and resource utilization. Therefore, how to achieve real-time monitoring and optimized scheduling of the operating status of charging piles through cloud computing technology has become a key problem that needs to be solved in the current technical field. Summary of the invention

[0004] In order to solve the above technical problems, the present invention is proposed. The present invention provides a cloud computing-based new energy charging pile operation status monitoring and optimization scheduling method, which can solve the problem that due to inaccurate load prediction and single scheduling strategy, it is difficult to achieve peak load balance distribution, which leads to overload operation of some charging piles, waste of resources or long waiting time for users.

[0005] According to one aspect of the present invention, a method for monitoring and optimizing the operation status of a new energy charging pile based on cloud computing is provided, which includes: Collect the operation data, historical data and user reservation data of the charging pile, upload them to the cloud server, and mark the charging pile type; Based on the operating data and the historical data, a charging pile health assessment model is constructed, and the charging piles are dynamically graded based on the health assessment model; Based on the user reservation data and the historical data, a load prediction model is constructed to calculate the charging load balance coefficient, and a hierarchical scheduling strategy is implemented based on the charging load balance coefficient in combination with peak, flat and valley periods.

[0006] Furthermore, the operating data includes charging power, charging voltage, charging current, charging pile temperature, charging interface impedance and power regulation margin; The historical data includes charging time, queuing time and number of charging interruptions; The user reservation data includes a reservation period, an expected charging amount and an expected charging time.

[0007] Furthermore, the charging pile types include fast charging type, slow charging type and mixed type.

[0008] Furthermore, the health assessment model includes static indicators and dynamic indicators; The static indicators reflect the current working status and basic performance of the charging pile, and are evaluated through power output capacity indicators, interface performance indicators and temperature safety indicators; The dynamic indicators reflect the operating stability and performance change trend of the charging pile, including charging interruption assessment, power fluctuation assessment, temperature change trend assessment and performance attenuation assessment.

[0009] Furthermore, the dynamic classification is to calculate a comprehensive health status index based on the health assessment model, and to perform three-level dynamic classification of charging piles based on the comprehensive health status index.

[0010] Furthermore, the calculation of the comprehensive health status index adopts a fuzzy comprehensive evaluation method to establish a two-level evaluation matrix. The first level evaluation calculates the basic health score based on the static indicators, and the second level evaluation corrects the basic score based on the dynamic indicators to obtain the comprehensive health status index.

[0011] Furthermore, the load forecasting model performs time series decomposition on historical charging data, extracts trend, season, cycle and random fluctuation characteristics, combines user reservation data, and establishes an ARIMA-based short-term load forecasting model.

[0012] Furthermore, a load prediction value at a future time is calculated based on the load prediction model, and is associated with a maximum carrying capacity of the charging station, and a load balancing coefficient is constructed with a ratio of the predicted load value to the maximum capacity as a core parameter.

[0013] Furthermore, a hierarchical scheduling strategy is implemented based on the charging load balancing coefficient in combination with peak, off-peak and valley periods.

[0014] According to another aspect of the present invention, a cloud computing-based new energy charging pile operation status monitoring and optimization scheduling system is provided, which includes: The collection module is used to collect the operation data, historical data and user reservation data of the charging pile, upload them to the cloud server, and mark the charging pile type; A grading module, which constructs a charging pile health assessment model based on the operating data and the historical data, and dynamically grades the charging piles based on the health assessment model; The scheduling module constructs a load prediction model based on the user reservation data and the historical data, calculates the charging load balance coefficient, and implements a hierarchical scheduling strategy based on the charging load balance coefficient in combination with peak, flat and valley periods.

[0015] Compared with the prior art, the present invention realizes accurate classification and intelligent scheduling of charging piles by establishing a complete data collection and preprocessing system, combining health status assessment and load forecast analysis. A dynamic and adaptive hierarchical scheduling mechanism is established, and differentiated scheduling strategies are implemented based on load characteristics and charging needs in different time periods, thereby optimizing the allocation efficiency of charging resources. The operating efficiency, equipment utilization and service quality of charging stations are significantly improved, and equipment failure rate and maintenance costs are reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings: Figure 1 A flowchart of a method for monitoring and optimizing the operation status of a new energy charging pile based on cloud computing according to an embodiment of the present invention; Figure 2 The flowchart of constructing a load prediction model according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0017] Below, the exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention, and it should be understood that the present invention is not limited to the exemplary embodiments described here.

[0018] As mentioned in the above background technology, there are three prominent problems in the prior art: First, the operation status monitoring capability of the charging pile equipment is limited, and it can only realize basic fault alarm and simple data collection, and it is impossible to effectively identify the changing trend of the equipment health status through deep data mining; second, the charging load management relies too much on simple empirical data or linear prediction models, lacks multi-dimensional analysis of user behavior, historical data and real-time load, and is difficult to effectively deal with the problem of peak power load imbalance; third, the existing scheduling strategy fails to fully consider the differences between the charging pile equipment status and the use period, lacks refined hierarchical management and optimization scheduling mechanism, resulting in insufficient charging service efficiency and resource utilization. In response to these technical pain points, our invention proposes a new energy charging pile operation status monitoring and optimization scheduling method based on cloud computing. By establishing a complete data collection and preprocessing system, combined with the dual evaluation mechanism of the health status comprehensive index and the load balance coefficient, the accurate classification and intelligent scheduling of charging piles are realized. At the same time, a dynamic adaptive hierarchical scheduling strategy is constructed based on the load characteristics of different time periods, and a complete emergency response mechanism is designed, thereby effectively solving the problems of insufficient equipment status monitoring, inaccurate load prediction and low scheduling efficiency in the prior art.

[0019] Figure 1 FIG. 1 is a flow chart of a method for monitoring and optimizing the operation status of a new energy charging pile based on cloud computing according to an embodiment of the present invention. Figure 1 As shown, in the cloud computing-based new energy charging pile operation status monitoring and optimization scheduling method, it includes: S1: Collect the operation data, historical data and user reservation data of the charging pile, upload them to the cloud server, and mark the charging pile type.

[0020] Collect the operation data, historical data and user reservation data of charging piles.

[0021] The operation data includes: Charging power, used to reflect the actual output power status of the charging pile, and record the power value and power factor; Charging voltage and charging current are used to monitor the stability of the charging process, including voltage fluctuation rate, current ripple factor and harmonic content; Charging pile temperature, used to evaluate whether the operating temperature of the charging pile is within a safe range, and monitor the charging gun head temperature, transformer temperature, and power module temperature; Charging interface impedance, used to detect the connection quality of the charging interface, including contact resistance and insulation resistance; The power regulation margin is used to characterize the power regulation capability of the charging pile. The power regulation margin is equal to the difference between the rated maximum power and the current actual power, and records the response time and regulation curve characteristics during the power regulation process.

[0022] The frequency of collecting operating data is once every 1 minute. When abnormal fluctuations are detected, it is automatically increased to once every 10 seconds to ensure that changes in the working status of the charging pile can be reflected in a timely manner.

[0023] Historical data is based on charging data within the past 30 days, including: The charging amount in each period is used to analyze the time distribution of charging demand and record the charging amount and charging power curve per hour; Charging time, used to evaluate the charging characteristics of different types of vehicles, including charging start time, end time, and pause time during charging; Queuing time, which is used to calculate the efficiency of charging facilities and record the time when users arrive at the waiting area and the actual time when charging starts; The number of charging interruptions is used to evaluate the reliability of the charging process and record the interruption cause classification and interruption duration.

[0024] The historical data is statistically summarized every hour, and time series data sets with four granularities of 15 minutes, 30 minutes, 1 hour and 24 hours are established.

[0025] User reservation data includes: The scheduled time period is the time when the user expects charging to start, which supports time selection with an accuracy of 30 minutes; Expected charge capacity, used to estimate the required charging time, calculated based on the charging characteristic curve in the vehicle model database; The expected charging time is used to reasonably arrange the charging sequence, including the shortest charging time requirement and the maximum waiting time limit.

[0026] It supports users to make charging reservations 24 hours in advance, with the minimum unit of the reservation period being 30 minutes, and provides intelligent recommended time periods based on historical data.

[0027] The collected data is uploaded to the cloud server in real time through the communication network. The communication network uses 5G network or wired network, and adopts dual-channel redundant design to ensure the real-time and reliability of data transmission. A distributed database is established in the cloud server to classify, store and pre-process different types of data, including data cleaning, outlier processing and data standardization. Data cleaning rules include: over-limit value filtering, mutation value detection, data continuity verification and timestamp calibration.

[0028] At the same time, the charging piles are marked and divided into three types: fast charging type, which supports DC charging with a maximum power of ≥60kW and has overvoltage protection, overcurrent protection and short circuit protection functions; slow charging type, which provides AC charging of ≤7kW and is equipped with leakage protection and grounding protection devices; hybrid type, which has dynamic power adjustment capabilities and can adjust the output power according to demand within the range of 7kW-60kW, and is equipped with an intelligent power management system. The power adjustment response time of the hybrid charging pile is no more than 30 seconds, and the adjustment accuracy is no less than 5% of the rated power. It supports two modes: stepless adjustment and preset gear adjustment.

[0029] After data collection and labeling, a structured data set is formed. The adaptability and prediction accuracy of the model are guaranteed through the continuous update of real-time data streams. During data transmission and storage, the AES-256 encryption algorithm is used to ensure data security, and a disaster recovery backup mechanism with real-time synchronization of master and slave databases is set up to prevent data loss. The data retention period is 30 days for real-time data, 1 year for historical statistical data, and permanent preservation of important event records.

[0030] S2: Based on the operation data and historical data, a health assessment model for charging piles is constructed. The health assessment model includes static indicators and dynamic indicators. The charging piles are dynamically graded based on the health assessment model.

[0031] Based on the collected operating data and historical data, a two-layer health assessment model is constructed, including two dimensions: static indicator evaluation and dynamic indicator evaluation.

[0032] Static indicators mainly reflect the current working status and basic performance of the charging pile, and are evaluated through power output capacity indicators, interface performance indicators and temperature safety indicators.

[0033] Furthermore, the power output capacity index is calculated by the ratio of real-time charging power to rated power. The closer the ratio is to 1, the better the power output capacity. During the calculation process, the power output data in the last 24 hours is taken, the data during the idle period of the charging pile is excluded, and the average power ratio value during the effective charging process is calculated.

[0034] Furthermore, the interface performance index is evaluated by the deviation rate of the charging interface impedance from the standard impedance, which reflects the connection quality and conductivity of the charging interface. When calculating, first obtain the standard impedance value of the charging interface under no-load state, then calculate the relative deviation between the interface impedance measured in actual operation and the standard value, and take the average deviation rate of the last 100 charging processes.

[0035] Furthermore, the temperature safety index is calculated by the margin between the temperature of the key parts of the charging pile and the safety temperature threshold. This index reflects the thermal management capability of the charging pile. The calculation method is: monitor the temperature of the charging gun head, transformer and power module respectively, calculate the difference between the temperature of each measuring point and its corresponding safety temperature threshold, and take the minimum temperature margin as the index value.

[0036] In the calculation of static indicators, the three parameters are combined by weighted summation. The weight distribution is determined by the hierarchical analysis method. First, a judgment matrix is ​​established. Based on the reliability analysis report of the charging pile equipment and the evaluation opinions of the operation and maintenance experts, the three indicators of power output capacity, interface performance and temperature safety are compared in pairs, and the characteristic vector is calculated to obtain the weight coefficient of each indicator.

[0037] At the same time, a dynamic weight adjustment mechanism is established based on the operating characteristics and historical fault data of different types of charging piles. When a certain type of fault occurs frequently or an indicator fluctuates abnormally, the weight configuration is adjusted accordingly to improve the sensitivity of the relevant indicator in the evaluation model.

[0038] Dynamic indicators mainly reflect the operating stability and performance change trend of the charging pile, including charging interruption assessment, power fluctuation assessment, temperature change trend assessment and performance attenuation assessment.

[0039] Furthermore, the charging interruption assessment calculates the interruption frequency and average interruption duration per unit time by analyzing the charging interruption records in the last 30 days. The interruption causes are classified and counted according to hardware failure, software abnormality and human interruption, and different types of interruptions are assigned different weight coefficients.

[0040] Furthermore, the power fluctuation assessment calculates the power fluctuation rate and stability index by analyzing the power curve characteristics during the charging process. The calculation method is: during each charging process, the ratio of the standard deviation of the power value to the average value is recorded, and the fluctuation rate distribution characteristics of the last 100 charging times are counted.

[0041] Furthermore, the temperature change trend is evaluated by comparing the temperature rise rate changes in continuous working cycles to evaluate the degree of degradation of heat dissipation performance. During the calculation process, the temperature rise curve under the same ambient temperature and load conditions is selected to analyze the change trend of the temperature rise rate.

[0042] Furthermore, the performance decay is evaluated by comparing the changing trends of key performance parameters in historical data over time, including the decay rates of charging efficiency, response time and regulation accuracy.

[0043] In the calculation of dynamic indicators, the historical data is analyzed by time series weighting. The time series weight is allocated by exponential decay function, and the weight coefficient is determined based on the interval between the data time and the current moment.

[0044] The comprehensive health status index is calculated by integrating static indicators and dynamic indicators. The calculation of the comprehensive health status index adopts the fuzzy comprehensive evaluation method to establish a two-level evaluation matrix. The first-level evaluation calculates the basic health score based on static indicators, and the second-level evaluation modifies the basic score based on dynamic indicators to finally obtain the comprehensive health status index.

[0045] The first-level evaluation process is mainly based on static indicators. First, establish a set of evaluation factors for static indicators, including three dimensions: power output capability, interface performance, and temperature safety. Establish a comment set for each evaluation factor, and divide the indicator values ​​into five levels: excellent, good, general, poor, and faulty, with transition intervals set between adjacent levels. Based on fuzzy mathematics theory, construct a membership function to calculate the membership of each indicator value to the comment set. When the indicator value is in the optimal operating range, it is assigned to the excellent level with a membership of 1. If the indicator value begins to deviate from the optimal range but does not reach the next level standard, the membership is allocated between the two adjacent levels in a linear proportion. When the indicator value continues to deteriorate, its membership decreases step by step until it reaches the fault level standard.

[0046] The construction process of the evaluation matrix includes: obtaining the real-time monitoring values ​​of each static indicator and eliminating the dimension effect through standardization. The membership function is used to calculate the membership of each indicator to different evaluation levels to form a fuzzy relationship matrix. Based on the previously calculated indicator weight vector, a fuzzy synthesis operation is performed, and the static indicator weight vector and the fuzzy relationship matrix are matrix multiplied. When the calculation is completed, a comprehensive evaluation vector is obtained, and each element in the vector represents the membership of the final evaluation result to the corresponding level. If a single evaluation result is required, the level with the largest membership is selected as the final evaluation level to obtain the comprehensive evaluation result of the static indicator, that is, the basic health score.

[0047] The second-level evaluation process introduces dynamic indicators for correction. The evaluation factor set of dynamic indicators is established, including four dimensions: charging interruption, power fluctuation, temperature change trend and performance degradation. A five-level comment set is also established for each dynamic indicator, and the corresponding membership function is constructed.

[0048] The dynamic correction process adopts a progressive evaluation method. First, the real-time value of each dynamic indicator is compared with its corresponding evaluation standard, the individual score of each indicator is calculated, and the comprehensive score of the dynamic indicator is obtained by the weighted average method; then the correction coefficient is determined based on the comprehensive score of the dynamic indicator. When the comprehensive score is higher than the benchmark value, the correction coefficient is greater than 1; when the comprehensive score is lower than the benchmark value, the correction coefficient is less than 1. The specific value of the correction coefficient is nonlinearly related to the degree of deviation of the comprehensive score. Influence factors are introduced in the correction process, and the correction coefficient is multiplied by the influence factor to obtain the final correction parameter. The correction parameter is multiplied by the basic health score to obtain the corrected health status assessment result.

[0049] Furthermore, the calculation process of the impact factor is determined in a hierarchical and progressive manner, including: Calculate the basic impact factors of the four dynamic indicators separately. For each dynamic indicator, calculate the basic impact factor based on the degree of deviation between its current value and the standard interval. When the indicator value is at the center of the standard interval, the basic impact factor is 1; when the indicator value deviates from the standard interval, the basic impact factor changes linearly with the degree of deviation.

[0050] Perform time series analysis on the historical data of each dynamic indicator to extract the characteristics of the change trend. When the indicator shows a trend of continuous deterioration, the trend impact coefficient is superimposed on the basic impact factor; when the indicator fluctuates periodically, the fluctuation impact coefficient is added. The size of the trend impact coefficient and the fluctuation impact coefficient is proportional to the rate of change and the amplitude of fluctuation.

[0051] The revised impact factors of the four dynamic indicators are weighted according to their importance, and the comprehensive impact factor is calculated by weighted average method. When multiple indicators are abnormal at the same time, a nonlinear weighting method is used to strengthen the impact of the abnormal state.

[0052] Through the above two-level evaluation, a comprehensive health status index between 0 and 1 is obtained. When the index is close to 1, it means that the device is in the best health state; when the index is close to 0, it means that the device is in a fault state.

[0053] According to the comprehensive health status index, the first threshold, the second threshold, and the third threshold are set, and the first threshold> the second threshold> the third threshold, and the charging piles are dynamically classified into three levels: When the comprehensive health status index continues to remain above the first threshold and there is no warning record in the most recent evaluation cycle, the charging pile will be classified as Class A.

[0054] If the index of the Class A charging pile equipment is stable for three consecutive evaluation cycles and the power output fluctuation rate is lower than the threshold, its full power stable operation capability is confirmed. When multiple charging piles meet the dispatch conditions at the same time, the charging pile with the highest health index will be dispatched first.

[0055] If the comprehensive health status index is between the second threshold and the first threshold, or the comprehensive health status index is higher than 0.85, but a dynamic warning has occurred in the most recent evaluation cycle; or the comprehensive health status index shows a downward trend in volatility, the charging pile will be classified as Class B.

[0056] Class B charging piles are mainly used for peak load operation during peak load periods. When the load rate of Class A charging piles exceeds the preset threshold, Class B charging piles are started to share the load.

[0057] If the comprehensive health status index is between the third threshold and the second threshold, the charging pile is classified as Class C.

[0058] Class C charging piles are only enabled in the following situations: Class A and Class B charging piles fail at the same time or operate at full capacity; an emergency occurs and emergency charging services are required; or performance testing is performed before equipment maintenance.

[0059] When using Class C charging piles, key monitoring is required, and timely inspection and maintenance should be arranged after use.

[0060] Charging piles whose comprehensive health status index is lower than the third threshold will be marked as awaiting maintenance and their use will be suspended.

[0061] S3: Based on the user reservation data and the historical data, a load prediction model is constructed to calculate the charging load balance coefficient, and a hierarchical scheduling strategy is implemented based on the charging load balance coefficient in combination with peak, off-peak and valley periods.

[0062] Through in-depth analysis of user reservation data and historical data, a multi-dimensional load forecasting model is constructed. First, a time series analysis is performed on the historical data to extract the periodic characteristics of the charging load, including the daily load variation pattern, weekly load distribution characteristics, and seasonal load fluctuation patterns. Combined with key elements such as the expected charging time and expected charging amount in the user reservation information, a short-term load forecasting model based on time series is established.

[0063] Specifically, the original charging load data is decomposed into time series. The data is decomposed into trend items, seasonal items, cycle items, and random fluctuation items. The trend item reflects the long-term trend of charging demand, such as the year-on-year increase in charging demand due to the increase in the number of electric vehicles; the seasonal item captures the change pattern of fixed cycles, such as the increase in charging demand due to the increase in air-conditioning power consumption in summer, or the increase in charging frequency due to the decrease in battery efficiency in winter; the cycle item reflects the fluctuation of non-fixed cycles, such as the load difference between weekdays and weekends; the random fluctuation item includes all irregular and unpredictable changes.

[0064] By removing the seasonal effect from the original data, we get a seasonally adjusted data series, and apply the ARIMA model to analyze the autocorrelation and moving average characteristics of historical data to establish the correlation between data points. During the calculation process, we need to ensure the stationarity of the data and perform differential processing when necessary.

[0065] After obtaining the forecast results of the ARIMA model, it is necessary to add the seasonal factors back into the forecast values. Ensure that the final forecast results contain all important time characteristics. If the difference processing is performed during the modeling process, it is also necessary to perform an inverse difference transformation to convert the forecast value back to the original scale.

[0066] In practical applications, for example, when predicting the load of a charging station next month, we first analyze the seasonal patterns in the historical data, including the monthly and quarterly changes. Then we use ARIMA to predict the data after removing the seasonality, and finally add the seasonal factors back to the prediction results. The prediction results take into account both the continuity of historical data and the influence of seasonality, and can better reflect the actual changes in the charging load.

[0067] Furthermore, the prediction results of the ARIMA model considering seasonal factors are combined with user reservation information to establish a short-term load forecasting model based on time series.

[0068] Specifically, for each user who has made an appointment, calculate his historical behavior score as the reliability coefficient and use Laplace smoothing. For example, if a user has 8 appointment records and has actually fulfilled 7 appointments, then his historical behavior score is 7 / 8 plus 1 divided by 8 plus 2, which equals 0.8. Multiply the charging power reserved by the user by the proportion of the user's charging time in a specific time period, and then multiply it by the calculated historical behavior score to obtain the weighted appointment load of the user in this time period. Sum up the weighted appointment loads of all users who have made appointments in this time period to obtain the overall appointment load for this time period.

[0069] The overall reservation load is converted into a correction factor. Specifically, the ratio of the overall reservation load to the predicted value of the ARIMA model is used as the basic correction value, and the time-varying weight function is introduced to adjust the correction value. The time-varying weight function is constructed in the form of exponential decay, where the exponential term contains the product of the decay parameter and the time interval, and the value range of the decay parameter is in the interval [0.1, 0.3]. The relative deviation value is multiplied by the output value of the time-varying adjustment function, and superimposed with the benchmark value to obtain the final correction factor.

[0070] The load forecasting model is obtained by combining the correction factor with the ARIMA forecast value using a multiplicative structure, which is expressed as: ; in, is the final predicted load value at time t, is the prediction result of the ARIMA model, is the total reservation load at time t, is the attenuation parameter, is the predicted target time, t is the current time, is the seasonal factor.

[0071] Furthermore, ; ; ; in, is the order of the autoregressive term, is the order of the moving average term, is the autoregressive parameter (i=1,..., ), is the moving average parameter (j=1,..., ), is the actual load value at the i-th time point before time t, is the prediction error at the jth time point before time t; is the number of historical periods considered when calculating the seasonal factor, is the historical load value for the same period, is the historical trend value for the same period, s is the length of the seasonal cycle, is the reserved charging power of the jth user, is the charging proportion of the jth user in this period, is the historical behavior score of the jth user, and n is the total number of scheduled users.

[0072] Furthermore, the load forecast value for the future is calculated based on the load forecast model, but this forecast value cannot be directly used for charging scheduling decisions. It is necessary to establish an association with the maximum carrying capacity of the charging station to determine the available status of the charging station at a specific moment. Charging scheduling has obvious time characteristics, and the sensitivity of the decision needs to be adjusted according to the distance of the predicted time point. At the same time, the uncertainty brought about by load fluctuations in actual operations requires scheduling decisions to have a certain degree of fault tolerance and room for adjustment. Therefore, a load balancing coefficient is constructed, with the ratio of the predicted load value to the maximum capacity as the core parameter, which intuitively reflects the load level of the charging station at a specific moment, expressed as: ; ; in, is the load balancing coefficient at time t, with a value range of [0,1], is the load value at time t obtained by the prediction model, is the maximum carrying capacity of the charging station, λ is the time sensitivity parameter, Predict the time for the target, is the load fluctuation correction factor, For the recent The standard deviation of the load at each time point, is the maximum standard deviation of historical load fluctuations.

[0073] According to the maximum carrying capacity of the charging station, the warning upper limit, middle limit and lower limit of the charging load balance coefficient are determined.

[0074] When the charging load balance coefficient exceeds the upper limit, it enters the overload warning state, indicating that the charging station load is close to saturation; when it exceeds the middle limit, it indicates that the charging station enters a high load state; when it exceeds the lower limit, it indicates that a scheduling strategy is needed.

[0075] Based on the actual operation characteristics of charging stations, the 24 hours of each day are divided into three main periods: peak period, off-peak period and off-peak period. The peak period is mainly concentrated in the lunch period from 11:30 to 14:00 and the night period from 20:00 to 23:00. The off-peak period covers the two periods of 9:00 to 11:30 and 14:00 to 20:00. The off-peak period includes the night and early morning periods from 23:00 to 9:00 the next day.

[0076] Based on the charging load balance coefficient and the charging pile classification results, a hierarchical progressive scheduling strategy is implemented: when it is at peak time and the charging load balance coefficient exceeds the medium limit, priority is given to fast charging and hybrid charging piles among Class A dispatchable charging piles, of which the proportion of fast charging piles is not less than 70%, and hybrid charging piles operate within the range of 85%-100% of the maximum power through dynamic power adjustment; When the charging load balance coefficient continues to rise and approaches the upper limit, the Class B dispatchable charging pile is started to share the load, and the power of the hybrid charging pile is adjusted to the maximum value; When the charging load balance coefficient exceeds the upper limit, the Class C emergency standby charging pile is activated to build an emergency charging network and start time-slot reservation at the same time; When the charging load balance coefficient is less than the lower limit, the slow charging type of Class A and Class B dispatchable charging piles will be prioritized, accounting for no less than 60%, and the power of the hybrid charging piles will be adjusted to operate within the range of 60%-80% to extend the service life of the equipment; Based on user reservation data, dynamic configuration and power pre-adjustment of charging piles are completed in advance to achieve optimal allocation of charging resources.

[0077] In summary, the cloud computing-based new energy charging pile operation status monitoring and optimization scheduling method based on the embodiment of the present invention is explained, which realizes the accurate classification and intelligent scheduling of charging piles by establishing a complete data acquisition and preprocessing system, combining health status assessment and load forecast analysis. Through the dual evaluation system of health status comprehensive index and load balance coefficient, it not only ensures the reliable operation of a single charging pile, but also realizes the load optimization of the entire charging station. A dynamic and adaptive hierarchical scheduling mechanism is established, and differentiated scheduling strategies are implemented based on load characteristics and charging needs in different time periods to optimize the allocation efficiency of charging resources. The operating efficiency, equipment utilization and service quality of the charging station are significantly improved, and the equipment failure rate and maintenance cost are reduced.

Claims

1. A new energy charging pile operation status monitoring and optimization scheduling method based on cloud computing, characterized in that: include: Collect the operation data, historical data and user reservation data of the charging pile, upload them to the cloud server, and mark the charging pile type; Based on the operating data and the historical data, a charging pile health assessment model is constructed, and the charging piles are dynamically graded based on the health assessment model; Based on the user reservation data and the historical data, a load prediction model is constructed to calculate the charging load balance coefficient, and a hierarchical scheduling strategy is implemented based on the charging load balance coefficient in combination with peak, flat and valley periods.

2. The cloud computing-based new energy charging pile operation status monitoring and optimization scheduling method according to claim 1 is characterized in that: The operating data includes charging power, charging voltage, charging current, charging pile temperature, charging interface impedance and power regulation margin; The historical data includes charging time, queuing time and number of charging interruptions; The user reservation data includes a reservation period, an expected charging amount and an expected charging time.

3. The cloud computing-based new energy charging pile operation status monitoring and optimization scheduling method according to claim 2 is characterized in that: The charging pile types include fast charging type, slow charging type and mixed type.

4. The cloud computing-based new energy charging pile operation status monitoring and optimization scheduling method according to claim 3 is characterized in that: The health assessment model includes static indicators and dynamic indicators; The static indicators reflect the current working status and basic performance of the charging pile, and are evaluated through power output capacity indicators, interface performance indicators and temperature safety indicators; The dynamic indicators reflect the operating stability and performance change trend of the charging pile, including charging interruption assessment, power fluctuation assessment, temperature change trend assessment and performance attenuation assessment.

5. The cloud computing-based new energy charging pile operation status monitoring and optimization scheduling method according to claim 2 is characterized in that: The dynamic classification is to calculate a comprehensive health status index based on the health assessment model, and to dynamically classify the charging piles into three levels based on the comprehensive health status index.

6. The cloud computing-based new energy charging pile operation status monitoring and optimization scheduling method according to claim 5 is characterized in that: The calculation of the comprehensive health status index adopts a fuzzy comprehensive evaluation method to establish a two-level evaluation matrix. The first level of evaluation calculates the basic health score based on the static indicators, and the second level of evaluation corrects the basic score based on the dynamic indicators to obtain the comprehensive health status index.

7. The cloud computing-based new energy charging pile operation status monitoring and optimization scheduling method according to claim 6 is characterized in that: The load forecasting model decomposes the historical charging data into time series, extracts the trend, season, cycle and random fluctuation characteristics, combines the user reservation data, and establishes a short-term load forecasting model based on ARIMA.

8. The cloud computing-based new energy charging pile operation status monitoring and optimization scheduling method according to claim 7 is characterized in that: The load prediction value at a future moment is calculated based on the load prediction model, and is associated with the maximum carrying capacity of the charging station, and the load balancing coefficient is constructed with the ratio of the predicted load value to the maximum capacity as the core parameter.

9. The cloud computing-based new energy charging pile operation status monitoring and optimization scheduling method according to claim 8 is characterized in that: A hierarchical scheduling strategy is implemented based on the charging load balancing coefficient in combination with peak, flat and valley periods.

10. The cloud computing-based new energy charging pile operation status monitoring and optimization scheduling system is characterized by: include: The collection module is used to collect the operation data, historical data and user reservation data of the charging pile, upload them to the cloud server, and mark the charging pile type; A grading module, which constructs a charging pile health assessment model based on the operating data and the historical data, and dynamically grades the charging piles based on the health assessment model; The scheduling module constructs a load prediction model based on the user reservation data and the historical data, calculates the charging load balance coefficient, and implements a hierarchical scheduling strategy based on the charging load balance coefficient in combination with peak, flat and valley periods.

Images