Private charging pile real-time load prediction method based on edge calculation

Abstract

The invention discloses a private charging pile real-time load prediction method based on edge calculation, and relates to the technical field of load prediction, and the method comprises the following steps: obtaining historical charging efficiency trajectory data and environmental influence data of a private charging pile; constructing a private charging pile real-time load prediction model based on a bp neural network according to the charging efficiency trajectory data and the environmental influence data; acquiring real-time data of the private charging pile, inputting the real-time data into the private charging pile real-time load prediction model, and performing real-time load prediction on the edge computing equipment to obtain a load prediction value of the private charging pile in the future t time; and the prediction result is sent to the user terminal equipment through the communication module, and a charging plan is reasonably arranged according to the prediction result. According to the invention, the powerful processing capability of the edge computing device is fully utilized, and rapid acquisition, preprocessing and load prediction of the real-time data of the private charging pile are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of load forecasting, and more specifically, to a real-time load forecasting method for a private charging pile based on edge computing. Background Art

[0002] As the number of electric vehicles continues to rise, the number of private charging stations is also growing rapidly. While this development trend provides convenient charging services for electric vehicle users, it also brings many new challenges to the stable operation of the power system.

[0003] In terms of power load, the charging behavior of private charging piles has significant randomness and uncertainty. The daily travel habits of different users vary greatly, resulting in extremely scattered and unpredictable charging time distribution. For example, some users may charge in the evening after get off work, while others charge in their spare time during the day. This disordered charging behavior further exacerbates the peak-to-valley difference in the power system load, increasing the difficulty and cost of power allocation.

[0004] Traditional power load forecasting methods focus on the analysis of large-scale industrial electricity consumption or the overall electricity consumption pattern of residents, and it is difficult to accurately adapt to small-scale, highly decentralized and highly random electricity consumption units such as private charging piles. Edge computing, as an emerging computing model, is a distributed open platform that integrates network, computing, storage, and application core capabilities at the edge of the network close to the source of objects or data. Unlike the traditional centralized cloud computing model, edge computing emphasizes storing computing and data close to the data source or user end to reduce data transmission delays, reduce network bandwidth pressure, and better protect data privacy and security. In the private charging pile scenario, edge computing can directly process charging-related data on edge devices locally or near the charging pile, avoiding the various disadvantages of transmitting a large amount of scattered charging pile data to the cloud for processing over long distances. In view of this, there is an urgent need for a real-time load forecasting method specifically for private charging piles, which can make full use of the advantages of edge computing, quickly process and analyze charging pile data at the edge end close to the data source, and predict its real-time load in a timely and accurate manner. This will not only help users plan charging time reasonably and improve charging efficiency, but also provide strong support for the stable operation of the power system, achieve optimal allocation and efficient utilization of power resources, and thus promote the coordinated and sustainable development of the electric vehicle industry and the power system.

[0005] The above disclosed technical solutions have at least the following technical problems: the real-time load of private charging piles is affected by multiple factors, such as user behavior, weather, traffic, electricity prices, etc. Most of the current prediction models are difficult to fully consider the interaction and coupling relationship between these factors at the same time, resulting in inaccurate and incomplete prediction results. For example, when considering the impact of electricity price factors on load, the changes in user sensitivity to electricity prices under different weather conditions may be ignored. In response to the above problems, the present invention proposes a solution. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a real-time load prediction method for private charging piles based on edge computing. By constructing a real-time load prediction model for private charging piles, the interaction and coupling relationship between these factors are fully considered to solve the problems raised in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions: A real-time load prediction method for private charging piles based on edge computing includes the following steps: obtaining historical charging efficiency trajectory data and environmental impact data of private charging piles; constructing a real-time load prediction model for private charging piles based on a BP neural network according to the charging efficiency trajectory data and the environmental impact data; collecting real-time data of private charging piles and inputting them into the real-time load prediction model for private charging piles, performing real-time load prediction on edge computing devices, and obtaining load prediction values ​​of private charging piles within a future time of t; sending the prediction results to user terminal devices through a communication module, and reasonably arranging a charging plan according to the prediction results.

[0008] In a preferred embodiment, the charging efficiency trajectory data includes a user charging portrait coefficient and a charging status coefficient; the environmental impact data includes a weather impact coefficient; the specific method for obtaining the user charging portrait coefficient is as follows: obtain the user's historical usage data on the charging pile APP, and remove duplicate data, abnormal data and incomplete data; calculate the average power consumption per charge based on all charging records of each user; obtain the user's power demand based on the proportion of the average power consumption per charge to the vehicle battery capacity in combination with the vehicle battery capacity; count the user's charging time and charging frequency based on the historical usage data, and calculate the standard deviation of the charging time; calculate the user's charging portrait coefficient based on the user's power demand combined with the charging time and charging frequency based on a preset user charging portrait coefficient calculation formula.

[0009] In a preferred embodiment, the specific method for obtaining the charging state coefficient is as follows: obtain the current and voltage of the charging pile during the charging process, as well as the working state information of the charging pile; obtain the charging efficiency of the charging pile according to the ratio of the actual output power and the rated power of the charging pile when it is working; calculate the operating index of the charging pile according to the ratio of the theoretical charging time to the actual consumed time; and obtain the charging state coefficient based on the charging efficiency and operating index of the charging pile combined with the battery health coefficient.

[0010] In a preferred embodiment, the specific method for obtaining the weather impact coefficient is as follows: obtain the historical weather data of the private charging pile during time B; analyze the correlation between various weather factors such as temperature, humidity, wind speed, rainfall, and snowfall and the charging pile load by drawing a scatter plot; based on the correlation analysis results, select weather characteristics that have a significant impact on the charging pile load as the key factors of the weather impact coefficient; use a linear regression algorithm to establish a regression model between the charging load and key weather factors; and use the trained model to calculate the corresponding weather impact coefficient based on the input characteristics under different weather conditions.

[0011] In a preferred embodiment, the real-time data of the private charging pile is collected and input into the real-time load prediction model of the private charging pile, and the real-time load prediction is performed on the edge computing device to obtain the load prediction value of the private charging pile in the future t time. Specifically, by connecting with the built-in sensor and communication module of the private charging pile, the current charging power, charging voltage, and charging current data are collected; the user's usage feature data is collected, and the usage feature data includes the remaining battery power of the vehicle, the vehicle type, and the user's charging habits; the MQTT communication protocol is used to transmit the collected various real-time data to the edge computing device safely, stably and efficiently; during the transmission process, the data is encrypted and a verification mechanism is added, and if a data error occurs, it is retransmitted; at the edge computing device end, the received data is first preprocessed, including supplementing missing values, eliminating outliers, and normalizing and standardizing the data; the load prediction features are extracted from the preprocessed data, and the key features are screened out by the feature selection algorithm to reduce the data dimension; the preset real-time load prediction model of the private charging pile is deployed in the edge computing device, and the extracted and selected feature data is input into the model to calculate the load prediction value of the private charging pile at each time node in the future t time.

[0012] The technical effects and advantages of the real-time load prediction method of a private charging pile based on edge computing of the present invention are as follows: 1. The present invention makes full use of the powerful processing capabilities of edge computing devices to achieve rapid collection, preprocessing and load forecasting of real-time data of private charging piles. By deploying the load forecasting model at the edge, key data such as user charging portrait coefficient, charging status coefficient and weather impact coefficient can be analyzed in real time, so as to quickly calculate the load forecast value within the future t time. This localized data processing method greatly reduces data transmission delays and ensures the real-time and accuracy of the forecast results. In addition, edge computing devices also have powerful data preprocessing capabilities, which can effectively deal with problems such as missing data and outliers, and further improve the prediction accuracy.

[0013] 2. The present invention not only provides efficient real-time load forecasting, but also can provide users with personalized charging plans and services based on the forecast results. By deeply analyzing the user's charging portrait coefficient, the system can accurately grasp the user's charging habits and needs, and thus push customized charging strategy recommendations for them, such as choosing to charge during low-load periods to save electricity costs. At the same time, the system also supports users to make reservations for the use of charging piles through terminal devices to achieve reasonable scheduling and allocation of charging resources. This personalized service not only improves the user's charging experience and satisfaction, but also helps to optimize the utilization rate of charging piles and avoid waste of resources. In addition, by real-time monitoring of the load conditions and charging progress of the charging piles, the system can also promptly detect and handle abnormal conditions to ensure the safety and reliability of the charging process. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a structural schematic diagram of a real-time load prediction method for private charging piles based on edge computing in the present invention. DETAILED DESCRIPTION

[0015] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0016] Embodiment 1, Figure 1 The present invention provides a real-time load prediction method for a private charging pile based on edge computing, comprising the following steps: S1, obtain the historical charging efficiency trajectory data and environmental impact data of private charging piles; The charging efficiency trajectory data includes a user charging profile coefficient and a charging state coefficient; the environmental impact data includes a weather impact coefficient; The user charging portrait coefficient is a key quantitative indicator for accurately understanding the charging behavior and habits of private charging pile users. It is like a fine ruler that deeply analyzes the user's charging characteristics from multiple dimensions, laying a solid foundation for charging pile operation management and load forecasting. The user charging portrait coefficient can not only accurately locate each user's charging preferences, but also provide charging pile operators with valuable decision-making basis. With the help of this coefficient, operators can push time-of-use electricity price reminders and charging reservation services in a targeted manner; when forecasting load, fully consider the concentrated charging periods of different types of users, prepare for the allocation of power resources and equipment operation and maintenance in advance, and comprehensively improve the operational efficiency and service level of private charging piles.

[0017] The specific method for obtaining the user charging portrait coefficient is as follows: Obtain the user's historical usage data on the charging pile APP and remove duplicate data, abnormal data, and incomplete data; Based on all charging records of each user, calculate the average amount of electricity per charge; The user's power demand is obtained based on the ratio of the average power consumed per charge to the vehicle's battery capacity in combination with the vehicle's battery capacity; Statistics on users’ charging time and charging frequency based on historical usage data, and calculate the standard deviation of charging time; The user charging profile coefficient is calculated based on the user's power demand combined with the charging time and charging frequency based on the preset user charging profile coefficient calculation formula.

[0018] The specific calculation formula for the user's power demand is:

[0019] The specific calculation formula of the user charging portrait coefficient is:

[0020] in, is the user's power demand, n is the number of charging times, is the amount of electricity charged each time, R is the vehicle battery capacity, The time for each charge.

[0021] The user charging profile coefficient has the following advantages for real-time load prediction of private charging piles: Improve load forecasting accuracy: The user charging portrait coefficient integrates multi-dimensional information such as time preference, power demand, charging frequency, and electricity price response, and comprehensively outlines the user's unique charging habits. Based on this, forecasters can accurately locate the charging pattern of each user and know when and under what power demand different users start charging. For example, the portrait shows that some time-sensitive users always charge between 19:00 and 21:00 after get off work, and will recharge in advance if they have special travel arrangements; after understanding these details, the load increase during a specific period can be accurately estimated, avoiding forecast deviations caused by general estimates, and making the real-time load forecast value closer to the actual situation.

[0022] Smooth load fluctuations: With the help of the portrait coefficient, the user groups that are sensitive to electricity prices and have flexible charging times can be accurately identified, so that the operator or power department can push time-of-use electricity price discounts, peak-shifting charging rewards and other strategies to them. When a large number of flexible demand users respond to the call and actively avoid peak electricity consumption and turn to charging during low electricity price periods, the overall load curve of private charging piles will become smoother. In this way, the pressure on the power grid during peak hours can be reduced and overload tripping can be avoided; the redundant power during the valley period can be fully utilized to improve the efficiency of power resource utilization, so that the real-time load can be maintained in a controllable and predictable stable range.

[0023] Optimize equipment operation and maintenance and resource allocation: The portrait coefficient clearly shows the user's power demand and charging frequency, and the operation and maintenance team can use this to understand the busyness and wear conditions of different charging piles. Charging piles with frequent high power demand wear and age quickly and are more prone to failures, so regular inspections and parts replacements need to be arranged in advance; for equipment with high idle rates, the operation and maintenance investment can be appropriately reduced, and resources can be allocated to key areas. From the perspective of power resources, by understanding the load changes in each time period, the power supply company can accurately dispatch power generation resources, ensure sufficient power supply, reduce power generation and transmission costs, maintain the stability of the power system, and allow real-time load forecasting results to directly serve efficient resource matching.

[0024] Empowering personalized services and enhancing user stickiness: By understanding the user's charging profile, the charging pile operator can tailor charging plans for users and provide personalized services such as appointment charging, power warning, exclusive discount packages, etc. When users truly enjoy the intimate services that meet their needs, their satisfaction and loyalty will increase, and stickiness will increase; a stable user group is conducive to the accumulation of long-term and complete charging data, which feeds back to the load forecasting model, continuously optimizes the forecast accuracy, and forms a virtuous data-driven closed loop.

[0025] Quickly adapt to dynamic changes and flexibly adjust forecasting strategies: User living habits and travel patterns are not static and are affected by many factors such as seasonal changes, job transfers, and new car purchases. Thanks to the dynamic update mechanism of the portrait coefficient, once the user behavior pattern changes, the coefficient can be quickly captured and adjusted, and the real-time load forecast will immediately adapt to the new situation; avoid forecast failures caused by outdated data and fixed models, and always keep the load forecast in line with actual needs, maintaining high timeliness and flexibility.

[0026] The state of charge coefficient is a quantitative indicator used to comprehensively evaluate the charging state and related characteristics of electric vehicles. It can reflect the actual situation of the vehicle during the charging process from multiple dimensions and plays an important role in many aspects.

[0027] On the one hand, the state of charge coefficient can intuitively reflect the current charging progress. For example, through this coefficient, it can be clearly seen how much percentage the vehicle battery has been charged, so that users can see the charging level at a glance. On the other hand, it can reflect the efficiency of charging. If the coefficient grows faster, it indicates that the charging efficiency is high. Otherwise, there may be charging anomalies or facility problems. At the same time, the state of charge coefficient is also related to the health of the battery. The long-term trend of the coefficient change can be used as one of the bases for judging the degree of battery aging. In addition, for the power system, this coefficient helps to grasp the charging load situation in real time, so as to carry out reasonable power dispatch and distribution and ensure the stable operation of the power grid.

[0028] The specific method for obtaining the charging state coefficient is as follows: Obtain the current and voltage of the charging pile during the charging process, as well as the working status information of the charging pile; The charging efficiency of the charging pile is obtained according to the ratio of the actual output power and the rated power of the charging pile when it is working; Calculate the operation index of the charging pile based on the ratio of theoretical charging time to actual consumed time; The charging state coefficient is obtained based on the charging efficiency and operating indicators of the charging pile combined with the battery health coefficient.

[0029] The specific calculation formula of the charging state coefficient is:

[0030] in, is the charge state coefficient, is the actual power of the charging pile, is the rated power of the charging pile, is the actual time consumed, is the battery health factor, is the battery capacity, is the current battery capacity.

[0031] The charging state coefficient has the following advantages for real-time load prediction of private charging piles: Real-time reflection of charging status: The charging status coefficient can reflect the state changes of electric vehicles in the charging process in real time, such as the dynamic adjustment of charging power, the growth trend of battery power, etc. By incorporating these real-time status information into the load forecasting model, the forecast results can be closer to the changes in the actual charging load, thereby improving the accuracy of the forecast.

[0032] Consider individual differences: Electric vehicles of different brands and models, as well as different user charging habits, will lead to differences in the charging process. The charging state coefficient can be calculated and analyzed individually for each charging process, fully considering the impact of these individual differences on the charging load, and further improving the accuracy of the prediction.

[0033] Adapt to different charging scenarios: Whether in different scenarios such as home charging piles, commercial charging piles or public charging piles, the charging state coefficient can be flexibly adjusted and calculated according to the actual charging situation. It can adapt to changes in various factors such as different charging power, charging time, ambient temperature, etc., thereby providing a more flexible solution for real-time load forecasting in various charging scenarios.

[0034] Dynamically adjust the prediction model: As the charging process progresses, the charging state coefficient will be continuously updated. The prediction model can adjust the prediction strategy and parameters in time according to these real-time changing coefficients, so that the prediction results can better adapt to the dynamic changes of the charging load, enhancing the flexibility and adaptability of the prediction.

[0035] Providing accurate load information: Accurate real-time load forecasting helps the grid dispatching department to grasp the electricity demand of private charging piles in advance, and the charging status coefficient can provide more detailed and accurate information for load forecasting, enabling grid dispatching personnel to better understand the load characteristics and changing trends of each charging pile, so as to arrange the grid operation mode and dispatching plan more reasonably.

[0036] Reduce grid operation risks: By real-time monitoring and analysis of the charging status coefficient, the grid dispatching department can promptly detect potential load peaks or troughs and take corresponding regulatory measures in advance, such as adjusting power generation output, optimizing the charging and discharging of energy storage equipment, etc., effectively avoiding grid voltage fluctuations, frequency instability and other operational risks caused by sudden changes in the load of private charging piles, and improving the safety and reliability of the grid.

[0037] Reasonable arrangement of charging time: For users, real-time load forecasting based on the charging state coefficient can provide more accurate charging time suggestions. Users can reasonably arrange charging time according to the forecast results, avoid the peak load period of the power grid, enjoy more favorable electricity prices, and also reduce waiting time, improving the user's charging experience.

[0038] Ensure charging efficiency: By analyzing the charging status coefficient, abnormal situations in the charging process can be discovered in time, such as abnormally low charging power, too long charging time, etc., and users can be notified in time or corresponding maintenance measures can be taken to ensure charging efficiency and quality, and further improve user experience.

[0039] The weather impact coefficient is a key indicator in the real-time load forecasting of private charging piles, which is used to quantify the impact of weather factors on the load of charging piles. It comprehensively considers various meteorological conditions such as temperature, humidity, wind speed, rainfall, snowfall, etc., and determines the changing rules of users' demand for charging piles and charging behavior under different weather conditions by analyzing historical data and actual operation conditions. For example, in hot weather, the use of electric vehicle air conditioning systems will increase energy consumption, which may increase charging demand; while in severe rainy and snowy weather, users' willingness to travel will decrease, and charging demand may decrease. The coefficient value is usually between 0 and 1. The closer it is to 1, the greater the impact of weather on the load of charging piles, and the closer it is to 0, the smaller the impact. By incorporating the weather impact coefficient into the real-time load forecasting model, the accuracy and reliability of the forecast can be improved, providing a basis for the reasonable scheduling and optimized operation of charging piles.

[0040] The specific method for obtaining the weather influence coefficient is as follows: Get the historical weather data of the private charging pile within time B; Analyze the correlation between various weather factors such as temperature, humidity, wind speed, rainfall, snowfall and charging pile load by drawing scatter plots; According to the results of correlation analysis, weather characteristics that have a significant impact on charging pile load are selected as key factors for determining weather impact coefficients; Use linear regression algorithm to build a regression model between charging load and key weather factors; Through the trained model, the corresponding weather impact coefficient is calculated according to the input features under different weather conditions.

[0041] The specific calculation formula of the weather influence coefficient is:

[0042] In the formula, is the weather influence coefficient, … The preset impact scores for different weather characteristics. … is the proportion of a single weather factor in all weather influencing factors, is the type of weather feature, and n is the number of weather features.

[0043] The weather impact coefficient has the following advantages for real-time load forecasting of private charging piles: Consider multiple weather factors: Weather has a significant impact on the charging demand and charging behavior of electric vehicles. By introducing the weather impact coefficient, the impact of multiple weather factors such as temperature, humidity, wind speed, rainfall, snowfall, etc. on the real-time load of charging piles can be comprehensively considered, making the prediction results closer to the actual situation.

[0044] Dynamically adjust the forecast value: Weather conditions change in real time, and different weather combinations have different effects on load. The weather impact coefficient can dynamically adjust the predicted charging pile load based on real-time weather data, promptly reflecting the impact of weather changes, thereby improving the accuracy and timeliness of the forecast.

[0045] Reasonable arrangement of charging resources: Understand the impact of weather on the load of charging piles. Operators can adjust the operation strategy of charging piles in advance according to the weather impact coefficient. When weather factors may lead to an increase in charging demand, reasonable allocation of power resources can ensure the stable operation of charging piles and avoid failures or safety hazards due to overload.

[0046] Reduce operating costs: By accurately predicting the real-time load of charging piles in different weather conditions, operators can optimize personnel scheduling and equipment maintenance plans. Equipment inspection and maintenance can be performed during periods of low load, thereby improving equipment utilization and reducing operating costs.

[0047] Meeting user needs: The weather impact factor helps operators better understand users' charging needs in different weather conditions, thereby providing users with more considerate services. For example, in hot weather, users are reminded to pay attention to charging safety and arrange charging time appropriately to avoid charging immediately after the vehicle is exposed to the sun, which affects battery life and charging efficiency.

[0048] Reduce waiting time: Accurate real-time load forecasting can help users better plan their charging trips and avoid long waits in front of charging piles. Users can choose the appropriate time and place to charge based on the forecast results, improve charging efficiency, and enhance user experience.

[0049] Achieve a balance between supply and demand: The large-scale application of private charging piles poses a challenge to the stability of the power grid. The weather impact coefficient can make the load forecast of charging piles more accurate. The power grid dispatching department can adjust the power supply strategy of the power grid in advance according to the forecast results, achieve a dynamic balance between power supply and demand, and reduce the impact on the power grid caused by sudden changes in the load of charging piles.

[0050] Promote the consumption of renewable energy: Combine the weather impact coefficient and renewable energy power generation forecast to optimize the charging strategy of charging piles, guide users to charge more when there is excess renewable energy power generation, improve the consumption capacity of renewable energy, reduce dependence on traditional fossil energy, and promote the optimization of energy structure and sustainable development.

[0051] S2, builds a real-time load prediction model for private charging piles based on BP neural network according to charging efficiency trajectory data and environmental impact data; The user charging portrait coefficient, charging status coefficient and weather impact coefficient are used to build a real-time load prediction model for private charging piles based on the BP neural network.

[0052] The real-time load prediction model of the private charging pile is specifically as follows:

[0053] in, is the load prediction factor, Charging profile coefficient for users, is the charge state coefficient, is the weather influence coefficient, The preset user charging profile coefficient weight factor, is the preset state of charge coefficient weight factor, It is the preset weather influence coefficient weight factor.

[0054] S3, collect the real-time data of the private charging pile and input it into the real-time load prediction model of the private charging pile, perform real-time load prediction on the edge computing device, and obtain the load prediction value of the private charging pile in the future time t; The real-time data of the private charging pile is collected and input into the real-time load prediction model of the private charging pile, and the real-time load prediction is performed on the edge computing device to obtain the load prediction value of the private charging pile in the future t time, specifically: By connecting to the built-in sensors and communication modules of private charging piles, the current charging power, charging voltage, and charging current data are collected; Collect user usage characteristic data, including vehicle battery remaining power, vehicle type, and user charging habits; Using the MQTT communication protocol, all kinds of collected real-time data are safely, stably and efficiently transmitted to edge computing devices; During the transmission process, the data is encrypted and a verification mechanism is added. If data errors occur, it is retransmitted.

[0055] On the edge computing device side, the received data is first preprocessed, including supplementing missing values, removing outliers, and normalizing the data. Extract features valuable for load forecasting from preprocessed data. For example, calculate the weather impact coefficient based on real-time weather data, and construct a feature vector that can reflect the future load change trend of the charging pile by integrating features such as the remaining battery power and vehicle type. At the same time, filter out key features through feature selection algorithms (such as correlation analysis, recursive feature elimination, etc.), reduce data dimensions, and improve forecasting efficiency. The preset real-time load forecasting model of private charging piles is deployed in the edge computing device, and the extracted and selected feature data is input into the model to calculate the load forecast value of the private charging pile at each time node in the future t time.

[0056] S4, sending the prediction results to the user terminal device through the communication module, and reasonably arranging the charging plan according to the prediction results.

[0057] Based on the load forecast results, charging strategy recommendations are provided to users, such as choosing to charge during low-load periods to save electricity costs, avoiding charging during peak hours to reduce waiting time, etc.

[0058] Intelligent algorithms can be used to provide users with personalized charging plans to meet the needs and preferences of different users.

[0059] Support users to reserve the usage time of charging piles through terminal devices to ensure charging during low load periods or at a time convenient for users.

[0060] Charging scheduling and resource allocation are carried out according to the real-time status and reservation status of the charging piles to improve the utilization rate of the charging piles and user satisfaction.

[0061] During the charging process, the terminal device monitors the load status and charging progress of the charging pile in real time.

[0062] Provides abnormal alarm and fault handling functions to ensure the safety and reliability of the charging process.

[0063] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0064] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0065] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0066] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0067] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0068] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A real-time load prediction method for private charging piles based on edge computing, characterized in that: The following steps are involved: Obtain historical charging efficiency trajectory data and environmental impact data of private charging piles; According to the charging efficiency trajectory data and environmental impact data, a real-time load prediction model for private charging piles is built based on the BP neural network; Collect the real-time data of private charging piles and input it into the real-time load prediction model of private charging piles, perform real-time load prediction on the edge computing device, and obtain the load prediction value of the private charging pile in the future t time; The prediction results are sent to the user terminal device through the communication module, and the charging plan is reasonably arranged according to the prediction results.

2. The real-time load prediction method for private charging piles based on edge computing according to claim 1 is characterized in that: The charging efficiency trajectory data includes a user charging profile coefficient and a charging state coefficient; the environmental impact data includes a weather impact coefficient; The specific method for obtaining the user charging portrait coefficient is as follows: Obtain the user's historical usage data on the charging pile APP and remove duplicate data, abnormal data, and incomplete data; Based on all charging records of each user, calculate the average amount of electricity per charge; The user's power demand is obtained based on the ratio of the average power consumed per charge to the vehicle's battery capacity in combination with the vehicle's battery capacity; Statistics on users’ charging time and charging frequency based on historical usage data, and calculate the standard deviation of charging time; The user charging profile coefficient is calculated based on the user's power demand combined with the charging time and charging frequency based on the preset user charging profile coefficient calculation formula.

3. The real-time load prediction method for private charging piles based on edge computing according to claim 2 is characterized in that: The specific method for obtaining the charging state coefficient is as follows: Obtain the current and voltage of the charging pile during the charging process, as well as the working status information of the charging pile; The charging efficiency of the charging pile is obtained according to the ratio of the actual output power and the rated power of the charging pile when it is working; Calculate the operation index of the charging pile based on the ratio of theoretical charging time to actual consumed time; The charging state coefficient is obtained based on the charging efficiency and operating indicators of the charging pile combined with the battery health coefficient.

4. The real-time load prediction method for private charging piles based on edge computing according to claim 3 is characterized in that: The specific method for obtaining the weather influence coefficient is as follows: Get the historical weather data of the private charging pile within time B; Analyze the correlation between various weather factors such as temperature, humidity, wind speed, rainfall, snowfall and charging pile load by drawing scatter plots; According to the results of correlation analysis, weather characteristics that have a significant impact on charging pile load are selected as key factors of weather impact coefficient; Use linear regression algorithm to build a regression model between charging load and key weather factors; Through the trained model, the corresponding weather impact coefficient is calculated according to the input features under different weather conditions.

5. The real-time load prediction method for private charging piles based on edge computing according to claim 4 is characterized in that: The real-time data of the private charging pile is collected and input into the real-time load prediction model of the private charging pile, and the real-time load prediction is performed on the edge computing device to obtain the load prediction value of the private charging pile in the future t time, specifically: By connecting to the built-in sensors and communication modules of private charging piles, the current charging power, charging voltage, and charging current data are collected; Collect user usage characteristic data, including vehicle battery remaining power, vehicle type, and user charging habits; Using the MQTT communication protocol, all kinds of collected real-time data are safely, stably and efficiently transmitted to edge computing devices; During the transmission process, the data is encrypted and a verification mechanism is added. If there is a data error, it is retransmitted; On the edge computing device side, the received data is first preprocessed, including supplementing missing values, removing outliers, and normalizing the data. Extract load forecasting features from preprocessed data, select key features through feature selection algorithm, and reduce data dimension; The preset real-time load forecasting model of private charging piles is deployed in the edge computing device, and the extracted and selected feature data is input into the model to calculate the load forecast value of the private charging pile at each time node in the future t time.

6. The real-time load prediction method for private charging piles based on edge computing according to claim 5 is characterized in that: The specific calculation formula for the user's power demand is: The specific calculation formula of the user charging portrait coefficient is: in, is the user's power demand, n is the number of charging times, is the amount of electricity charged each time, R is the vehicle battery capacity, The time for each charge.

7. The real-time load prediction method for private charging piles based on edge computing according to claim 6 is characterized in that: The specific calculation formula of the charging state coefficient is: in, is the charge state coefficient, is the actual power of the charging pile, is the rated power of the charging pile, is the actual time consumed, is the battery health factor, is the battery capacity, is the current battery capacity.

8. The real-time load prediction method for private charging piles based on edge computing according to claim 7 is characterized in that: The specific calculation formula of the weather influence coefficient is: In the formula, is the weather influence coefficient, … The preset impact scores for different weather characteristics. … is the proportion of a single weather factor in all weather influencing factors, is the type of weather feature, and n is the number of weather features.

9. The real-time load prediction method for private charging piles based on edge computing according to claim 8 is characterized in that: The real-time load prediction model of the private charging pile is specifically as follows: in, is the load prediction factor, Charging profile coefficient for users, is the charge state coefficient, is the weather influence coefficient, The preset user charging profile coefficient weight factor, is the preset state of charge coefficient weight factor, It is the preset weather influence coefficient weight factor.

Images