Charging pile operation control method based on AC / DC switching backup

Abstract

The present invention discloses a charging pile operation control method based on AC / DC switching as a backup, which relates to the field of charging pile operation control technology, and includes the following steps: S1, analyzing historical charging records, S2, training behavior prediction models, S3, determining the best charging pile, S4, constructing conditional functions, S5, outputting the optimal charging time, and S6, switching the power transmission interface. The present invention calculates the total load of the power grid in different time periods through a load analysis algorithm, and determines the specific time period corresponding to the load imbalance. At the same time, the method is provided with a data particle group, and the local optimal charging start time and the global optimal charging start time are continuously iterated to ensure that the final output scheme is conducive to the safe operation of the power grid, and an AC pile and a DC pile are provided in the charging pile. If any one of the two charging devices has an abnormal situation, the charging device without abnormality will directly take over the power supply of the abnormal device to ensure that the charging process of the electric vehicle can be smoothly executed.

Description

Technical Field

[0001] The present invention relates to the technical field of charging pile operation control, and in particular to a charging pile operation control method based on AC / DC switching for mutual backup. Background Art

[0002] The charging pile operation control method monitors the operating parameters of the charging pile in real time through sensors. Once abnormal parameters are found, an early warning will be issued immediately. The method will analyze historical data and make optimization suggestions. This not only improves charging efficiency and safety, but also reduces maintenance costs. The invention patent with application number 202311278938.5 discloses "electric vehicle charging pile operation control method, system and device, involving the field of electric vehicle charging. The invention includes a charging pile, a power supply port, a payment camera and a camera. A processor is provided inside the charging pile, and the output end of the processor is electrically connected to an electric push rod, a forward and reverse motor, a human-computer interaction page, an identification module, an information storage module and a positioning module. The invention simplifies the charging pile payment steps through the coordination of structures such as an identification module, a control module, a power monitoring module, and a calculation module to meet the needs of member car owners for quick payment and non-member car owners for convenient use. The image captured by the camera enables the charging pile to be operated in different ways for different car owners, thereby simplifying the operation process and avoiding economic losses caused by manual misoperation."

[0003] The above-mentioned existing technology solves the problem that the payment steps of charging piles are too cumbersome and cannot meet the needs of car owners for quick payment. However, during execution, due to the lack of analysis of the historical charging records of each electric vehicle, it is impossible to obtain the upper and lower limits of the adjustable time period corresponding to the user, the scheduling priority and the preference interval in a short time, resulting in the method being unable to recommend suitable charging piles to users, and unable to adjust the charging time of each vehicle, resulting in an unbalanced load on the power grid and affecting the safe use of the charging piles. Summary of the invention

[0004] The object of the present invention is to provide a charging pile operation control method based on AC / DC switching as a backup to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a charging pile operation control method based on AC / DC switching as a backup, comprising the following steps:

[0006] S1. Analyze historical charging records: extract all historical charging records of each electric vehicle, determine the first characteristic data and the second characteristic data, classify them using a clustering model, store similar data in the same set, perform normal fitting on each set corresponding to the first characteristic data, analyze the fitted data using an adjustable time analysis algorithm, obtain the upper and lower limits of the adjustable time period of the electric vehicle, count all charging interval values ​​of each electric vehicle, determine the relevant attributes of each electric vehicle based on the charging interval value and the second characteristic data, determine the scheduling priority of the electric vehicle based on the relevant attributes, count the number of times the charging pile corresponding to each electric vehicle is used, and determine the number of the charging pile within the preferred interval of each electric vehicle based on the number of times it is used;

[0007] S2. Training behavior prediction model: After receiving the upper and lower limits of the adjustable time period, the scheduling priority and the preference interval of all electric vehicles, the historical charging records corresponding to these vehicles are transmitted to the user behavior prediction model in sequence, and the model is used to extract features from these historical charging records, determine the upper and lower limits of the adjustable time period, the scheduling priority and the preference interval corresponding to each electric vehicle, use the optimization function to analyze the deviation between the predicted results and the actual results, and iteratively update the model parameters until the model training is completed;

[0008] S3. Determine the best charging pile: after receiving the charging request information from the electric vehicle, the vehicle is regarded as the vehicle to be charged, all historical charging records of the vehicle to be charged are extracted, and the user behavior prediction model is used to analyze them, and the upper and lower limits of the adjustable time period corresponding to the electric vehicle, the scheduling priority and the preference interval are determined, and the preference interval corresponding to each vehicle to be charged is analyzed by deep learning technology, so as to obtain the best charging pile number for each vehicle to be charged;

[0009] S4. Constructing conditional functions: extracting basic data and behavior data of the vehicles to be charged, calculating the corresponding default charging period according to the connection start time and charging duration in the behavior data of each electric vehicle, and determining the number of charging piles that are not idle at different times, obtaining the total load of the power grid in each period based on these data, and drawing the corresponding load curve, extracting the corresponding electricity value at different times, and constructing the corresponding comprehensive conditional function through the power optimization algorithm;

[0010] S5. Output the optimal charging time: after receiving the scheduling priority, total number, comprehensive condition function, state of charge constraint function and charging time constraint function of the vehicles to be charged, determine the dimension of the search space according to the total number of vehicles to be charged, set the size of the data particle group and related parameters, determine the initial speed vector and initial position vector of different data particles in each dimensional space according to the scheduling priority of each vehicle, calculate the corresponding condition function value according to the speed vector and position vector of each data particle, and update the local optimal charging start time and the global optimal charging start time, analyze the new speed vector and position vector of each data particle in each dimensional space according to the constraint function, repeat the operation until the number of cycles reaches the set number, and output the optimal charging start time for each electric vehicle;

[0011] S6. Switch the power transmission interface: Each charging pile is equipped with an AC pile and a DC pile, and the AC pile is set with a direct output interface, and the DC pile is set with a DC output interface. After extracting the real-time monitoring data of the AC pile and the DC pile, transmit it to the fault prediction model for analysis. If the model determines that there is an abnormal component in the AC pile, it switches to the DC output interface. If the model determines that there is an abnormal component in the DC pile, it switches to the direct output interface.

[0012] Preferably, the step S1 specifically includes the following steps:

[0013] S101, extracting all historical charging records of each electric vehicle, wherein the historical charging records include user number, connection start time, charging start time, connection end time, charging end time and charging pile number, and analyzing the charging time and connection time corresponding to each historical charging record, combining the connection start time and connection time into the first feature data, and then combining the charging time and connection time into the second feature data;

[0014] S102, using a clustering model to classify the first characteristic data and the second characteristic data of each electric vehicle, and storing the same type of data in the same set, performing normal fitting on each set corresponding to the first characteristic data, and statistically analyzing the data corresponding to the highest point in each set after fitting, wherein the clustering model is specifically a GMM clustering model;

[0015] S103, calculate the corresponding weight according to the amount of data in each set, use the adjustable time analysis algorithm to analyze the data corresponding to the highest point of each set and the weight, and obtain the upper limit and lower limit of the current adjustable time period of the electric vehicle. The adjustable time analysis algorithm is specifically:

[0016]

[0017]

[0018] In the formula, Indicates The upper limit of the adjustable time period of an electric vehicle, Indicates The total number of categories included in electric vehicles, Indicates After clustering, the electric vehicles The weight of the class data. Indicates After clustering, the electric vehicles The highest point of the connection start time in the normal distribution of class data, Indicates After clustering, the electric vehicles The highest point of the connection time of the normal distribution of class data, Indicates The lower limit of the adjustable time period of an electric vehicle, and Indicates a parameter.

[0019] Preferably, the step S1 specifically further comprises the following steps:

[0020] S104, analyzing multiple charging interval values ​​from all historical charging records of each electric vehicle, and counting the proportion of charging interval values ​​within 36 hours. If the proportion is higher than a first threshold, the electric vehicle is determined to be frequently scheduled, otherwise it is determined to be infrequently scheduled. Determine the proportion of the difference between the connection time and the charging time in the second characteristic data that is not zero. If the proportion is higher than a second threshold, the electric vehicle is determined to be strongly responsive. Otherwise, it is determined to be weakly responsive.

[0021] S105, determining the relevant attributes of each electric vehicle, if the attribute is frequent scheduling and strong response scheduling, then determining the scheduling priority of the electric vehicle to be level one, if the attribute is infrequent scheduling and strong response scheduling, then determining the scheduling priority of the electric vehicle to be level two, if the attribute is frequent scheduling and weak response scheduling, then determining the scheduling priority of the electric vehicle to be level three, if the attribute is infrequent scheduling and weak response scheduling, then determining the scheduling priority of the electric vehicle to be level four;

[0022] S106: Count the usage times of the charging pile corresponding to each electric vehicle, and store the numbers of the charging piles whose usage times are greater than a third threshold into the preference interval of the corresponding vehicle.

[0023] Preferably, the step S3 specifically includes the following steps:

[0024] S301, after receiving the charging request information sent by the electric vehicle, the corresponding user number is stored in the set of vehicles to be charged, and the database is queried in sequence according to the numbers contained in the set. If the corresponding historical charging record is not found, the upper and lower limits of the adjustable time period are directly determined from the vehicle user, and the scheduling priority is set to level 4, and the preference interval is empty. Otherwise, all corresponding historical charging records are extracted;

[0025] S302, using a user behavior prediction model to analyze historical charging records of different electric vehicles in turn, and determine the upper and lower limits of the adjustable time period, the scheduling priority, and the preferred interval corresponding to each electric vehicle, wherein the user behavior prediction model is specifically a BiGRU residual prediction model;

[0026] S303: Analyze the preference interval corresponding to each car to be charged by deep learning technology, so as to obtain the best charging pile number and charging pile location information of each car to be charged, and return it to the car user interface.

[0027] Preferably, the step S4 specifically includes the following steps:

[0028] S401, after obtaining the total number of charging piles, charging power, charging efficiency and basic load data in the current charging station, extract the basic data and behavior data of the vehicles to be charged, and calculate the corresponding default charging period according to the connection start time and charging duration in the behavior data of each electric vehicle, the basic data includes the vehicle type, power consumption per 100 kilometers and battery capacity, and the behavior data includes the travel mileage, connection start time and charging duration;

[0029] S402, according to the default charging period of each electric vehicle, determine the number of charging piles in a non-idle working state at different times, and analyze the working state, charging power and basic load of the power grid corresponding to each charging pile at different times by a load analysis algorithm, so as to obtain the total load of the power grid in different time periods, and store it in a database. The load analysis algorithm is specifically as follows:

[0030]

[0031] In the formula, Indicates that the current power grid is The base load at the time, Indicates the charging power of the charging pile. Indicates Charging stations at the time working status, When Charging stations at the time The working status is non-idle, When Charging stations at the time The working status is idle. Indicates the total number of charging piles. and Represents parameters;

[0032] S403, repeating the operation until all electric vehicle data are extracted, drawing a corresponding load curve using the total load of the power grid in each time period, and determining the maximum and minimum values ​​corresponding to the curve;

[0033] S404, after extracting the electricity values ​​corresponding to different times, multiple conditional functions are constructed based on these data through the electricity optimization algorithm, a corresponding weight coefficient is set for each conditional function, and the conditional function and the weight coefficient are combined to generate a comprehensive conditional function;

[0034] S405: Determine the connection end time of each electric vehicle according to the upper limit and lower limit of the adjustable time period of the vehicle to be charged and the connection start time.

[0035] Preferably, the step S5 specifically includes the following steps:

[0036] S501, determining the initial state of charge and the minimum state of charge of each vehicle to be charged, calculating the actual state of charge using the initial state of charge, charging time, battery capacity, charging power and efficiency of the vehicle, constructing a state of charge constraint function based on these data, and constructing a charging time constraint function using the connection start time and connection end time of each vehicle;

[0037] S502, receiving the scheduling priority, total number, comprehensive condition function, state of charge constraint function and charging time constraint function of the vehicles to be charged, determining the dimension of the search space according to the total number of vehicles to be charged, setting the data particle group size, number of cycles, maximum moving speed, initial weight value, initial local factor and initial global factor, and determining the initial velocity vector and initial position vector of different data particles in each dimensional space according to the scheduling priority of each vehicle;

[0038] S503, calculating the corresponding conditional function value according to the velocity vector and position vector of each data particle, if the value is greater than the original conditional function value, updating the local optimal charging start time and the global optimal charging start time, otherwise the update operation is not performed;

[0039] S504, using a parameter update algorithm to calculate new weight values, local factors and global factors according to the current number of cycles, and analyzing new velocity vectors and position vectors of each data particle in each dimensional space according to the constraint function;

[0040] S505, determining whether the speed values ​​of each data particle in different dimensional spaces are all less than the maximum moving speed, if they are all less than the maximum moving speed, no operation is performed, otherwise they are adjusted according to the maximum moving speed;

[0041] S506, repeat the operation until the number of cycles reaches the set number, output the optimal charging start time for each electric vehicle, perform charging operations according to these data and update the corresponding basic load data in different time periods.

[0042] Preferably, the step S6 specifically includes the following steps:

[0043] S601. An AC pile and a DC pile are provided in each charging pile, and two output interfaces are provided in the charging pile, namely a direct output interface of the AC pile and a DC output interface for converting AC power into DC power;

[0044] S602. Use multiple temperature and humidity sensors to collect temperature and humidity data at various locations of the AC pile and the DC pile, use a device detector to analyze the AC pile and the DC pile respectively, so as to obtain the voltage value and current value of each point inside these devices, and transmit these data to the database. The temperature and humidity sensor is specifically a DHT11 temperature and humidity sensor, and the device detector is specifically a CPT200 charging pile inspection instrument.

[0045] Preferably, the step S6 specifically further comprises the following steps:

[0046] S603, after extracting the real-time monitoring data of the AC pile and the DC pile, transmit it to the fault prediction model for analysis. If the model determines that there is an abnormal component in the AC pile, the power supply of the AC pile direct output interface will be immediately stopped, switched to the DC output interface, and a fault warning message will be sent to the control center. If the model determines that there is an abnormal component in the DC pile, the power supply of the DC output interface will be immediately stopped, switched to the AC pile direct output interface, and a fault warning message will be sent to the control center. The fault prediction model is specifically an LSTM model;

[0047] S604. If the model determines that the temperature inside the AC pile is too high, the power supply of the AC pile direct output interface will be stopped immediately and switched to the DC output interface. If the model determines that the temperature inside the DC pile is too high, the power supply of the DC output interface will be stopped immediately and switched to the AC pile direct output interface.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] The present invention extracts historical charging records of different electric vehicles from a database, analyzes these historical charging records one by one, determines the upper and lower limits of the adjustable time period, the scheduling priority and the preferred interval corresponding to each electric vehicle, uses these data to complete the training of the user behavior prediction model, and continuously optimizes the model parameters to make the model output value more in line with reality. When receiving the request information sent by the electric vehicle, the trained model is used to analyze the historical charging records of the vehicle that has sent the request, and the upper and lower limits of the adjustable time period, the scheduling priority and the preferred interval corresponding to the vehicle can be determined in a short time, so that the method can recommend more suitable charging piles to users, and is also convenient for subsequent adjustment of the charging time of the electric vehicle;

[0050] The present invention calculates the total load of the power grid in different time periods through a load analysis algorithm and draws corresponding curve graphs. Such a design can, on the one hand, more intuitively observe the impact of excessively concentrated charging time on the power grid, and on the other hand, determine the specific time period corresponding to the load imbalance, which is convenient for subsequent optimization measures. At the same time, the method is provided with a data particle group, and the local optimal charging start time and the global optimal charging start time are continuously iterated to ensure that the final output solution is not only conducive to the safe operation of the power grid, but also reduces the cost required for most users' personal charging. In addition, an AC pile and a DC pile are provided in the charging pile. If any one of the two charging devices has an abnormal situation, the charging device without abnormality will directly take over the power supply of the abnormal device, thereby ensuring that the charging process of the electric vehicle can be smoothly executed. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 An overall method flow chart is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0052] 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.

[0053] See also Figure 1 The present invention provides a technical solution: a charging pile operation control method based on AC / DC switching as a backup, comprising the following steps:

[0054] S1. Analyze historical charging records: extract all historical charging records of each electric vehicle, determine the first characteristic data and the second characteristic data, classify them using a clustering model, store similar data in the same set, perform normal fitting on each set corresponding to the first characteristic data, analyze the fitted data using an adjustable time analysis algorithm, obtain the upper and lower limits of the adjustable time period of the electric vehicle, count all charging interval values ​​of each electric vehicle, determine the relevant attributes of each electric vehicle based on the charging interval value and the second characteristic data, determine the scheduling priority of the electric vehicle based on the relevant attributes, count the number of times the charging pile corresponding to each electric vehicle is used, and determine the number of the charging pile within the preferred interval of each electric vehicle based on the number of times it is used;

[0055] S2. Training behavior prediction model: After receiving the upper and lower limits of the adjustable time period, the scheduling priority and the preference interval of all electric vehicles, the historical charging records corresponding to these vehicles are transmitted to the user behavior prediction model in sequence, and the model is used to extract features from these historical charging records, determine the upper and lower limits of the adjustable time period, the scheduling priority and the preference interval corresponding to each electric vehicle, use the optimization function to analyze the deviation between the predicted results and the actual results, and iteratively update the model parameters until the model training is completed;

[0056] S3. Determine the best charging pile: after receiving the charging request information from the electric vehicle, the vehicle is regarded as the vehicle to be charged, all historical charging records of the vehicle to be charged are extracted, and the user behavior prediction model is used to analyze them, and the upper and lower limits of the adjustable time period corresponding to the electric vehicle, the scheduling priority and the preference interval are determined, and the preference interval corresponding to each vehicle to be charged is analyzed by deep learning technology, so as to obtain the best charging pile number for each vehicle to be charged;

[0057] S4. Constructing conditional functions: extracting basic data and behavior data of the vehicles to be charged, calculating the corresponding default charging period according to the connection start time and charging duration in the behavior data of each electric vehicle, and determining the number of charging piles that are not idle at different times, obtaining the total load of the power grid in each period based on these data, and drawing the corresponding load curve, extracting the corresponding electricity value at different times, and constructing the corresponding comprehensive conditional function through the power optimization algorithm;

[0058] S5. Output the optimal charging time: after receiving the scheduling priority, total number, comprehensive condition function, state of charge constraint function and charging time constraint function of the vehicles to be charged, determine the dimension of the search space according to the total number of vehicles to be charged, set the size of the data particle group and related parameters, determine the initial speed vector and initial position vector of different data particles in each dimensional space according to the scheduling priority of each vehicle, calculate the corresponding condition function value according to the speed vector and position vector of each data particle, and update the local optimal charging start time and the global optimal charging start time, analyze the new speed vector and position vector of each data particle in each dimensional space according to the constraint function, repeat the operation until the number of cycles reaches the set number, and output the optimal charging start time for each electric vehicle;

[0059] S6. Switch the power transmission interface: Each charging pile is equipped with an AC pile and a DC pile, and the AC pile is set with a direct output interface, and the DC pile is set with a DC output interface. After extracting the real-time monitoring data of the AC pile and the DC pile, transmit it to the fault prediction model for analysis. If the model determines that there is an abnormal component in the AC pile, it switches to the DC output interface. If the model determines that there is an abnormal component in the DC pile, it switches to the direct output interface.

[0060] Step S1 specifically includes the following steps:

[0061] S101, extracting all historical charging records of each electric vehicle, wherein the historical charging records include user number, connection start time, charging start time, connection end time, charging end time and charging pile number, and analyzing the charging time and connection time corresponding to each historical charging record, combining the connection start time and connection time into the first feature data, and then combining the charging time and connection time into the second feature data;

[0062] S102, using a clustering model to classify the first characteristic data and the second characteristic data of each electric vehicle, and storing the same type of data in the same set, performing normal fitting on each set corresponding to the first characteristic data, and statistically analyzing the data corresponding to the highest point in each set after the fitting, wherein the clustering model is specifically a GMM clustering model;

[0063] S103, calculate the corresponding weight according to the amount of data in each set, use the adjustable time analysis algorithm to analyze the data corresponding to the highest point of each set and the weight, and obtain the upper limit and lower limit of the current adjustable time period of the electric vehicle. The adjustable time analysis algorithm is specifically as follows:

[0064]

[0065]

[0066] In the formula, Indicates The upper limit of the adjustable time period of an electric vehicle, Indicates The total number of categories included in electric vehicles, Indicates After clustering, the electric vehicles The weight of the class data, Indicates After clustering, the electric vehicles The highest point of the connection start time in the normal distribution of class data, Indicates After clustering, the electric vehicles The highest point of the connection time of the normal distribution of class data, Indicates The lower limit of the adjustable time period of an electric vehicle, and Represents parameters;

[0067] Step S1 specifically also includes the following steps:

[0068] S104, analyzing multiple charging interval values ​​from all historical charging records of each electric vehicle, and counting the proportion of charging interval values ​​within 36 hours. If the proportion is higher than a first threshold, the electric vehicle is determined to be frequently scheduled, otherwise it is determined to be infrequently scheduled. Determine the proportion of the difference between the connection time and the charging time in the second characteristic data that is not zero. If the proportion is higher than a second threshold, the electric vehicle is determined to be strongly responsive. Otherwise, it is determined to be weakly responsive.

[0069] S105, determining the relevant attributes of each electric vehicle, if the attribute is frequent scheduling and strong response scheduling, then determining the scheduling priority of the electric vehicle to be level one, if the attribute is infrequent scheduling and strong response scheduling, then determining the scheduling priority of the electric vehicle to be level two, if the attribute is frequent scheduling and weak response scheduling, then determining the scheduling priority of the electric vehicle to be level three, if the attribute is infrequent scheduling and weak response scheduling, then determining the scheduling priority of the electric vehicle to be level four;

[0070] S106, counting the number of times the charging pile corresponding to each electric vehicle is used, and storing the number of the charging pile whose number of uses is greater than a third threshold into the preference interval of the corresponding vehicle;

[0071] Step S3 specifically includes the following steps:

[0072] S301, after receiving the charging request information sent by the electric vehicle, the corresponding user number is stored in the set of vehicles to be charged, and the database is queried in sequence according to the numbers contained in the set. If the corresponding historical charging record is not found, the upper and lower limits of the adjustable time period are directly determined from the vehicle user, and the scheduling priority is set to level 4, and the preference interval is empty. Otherwise, all corresponding historical charging records are extracted;

[0073] S302, using a user behavior prediction model to analyze historical charging records of different electric vehicles in turn, and determine the upper and lower limits of the adjustable time period, the scheduling priority, and the preference interval corresponding to each electric vehicle, wherein the user behavior prediction model is specifically a BiGRU residual prediction model;

[0074] S303, analyzing the preference interval corresponding to each vehicle to be charged by deep learning technology, thereby obtaining the best charging pile number and charging pile location information for each vehicle to be charged, and returning the information to the vehicle user interface;

[0075] Step S4 specifically includes the following steps:

[0076] S401. After obtaining the total number of charging piles, charging power, charging efficiency and basic load data in the current charging station, extract the basic data and behavior data of the vehicles to be charged, and calculate the corresponding default charging period according to the connection start time and charging duration in the behavior data of each electric vehicle. The basic data includes the vehicle type, power consumption per 100 kilometers and battery capacity, and the behavior data includes the travel mileage, connection start time and charging duration;

[0077] S402, according to the default charging period of each electric vehicle, determine the number of charging piles in non-idle working state at different times, and analyze the working state, charging power and basic load of the power grid corresponding to each charging pile at different times by load analysis algorithm, so as to obtain the total load of the power grid in different time periods, and store it in the database. The load analysis algorithm is specifically as follows:

[0078]

[0079] In the formula, Indicates that the current power grid is The base load at the time, Indicates the charging power of the charging pile. Indicates Charging stations at the time working status, When Charging stations at the time The working status is non-idle, When Charging stations at the time The working status is idle. Indicates the total number of charging piles. and Represents parameters;

[0080] S403, repeating the operation until all electric vehicle data are extracted, drawing a corresponding load curve using the total load of the power grid in each time period, and determining the maximum and minimum values ​​corresponding to the curve;

[0081] S404, after extracting the electricity values ​​corresponding to different times, multiple conditional functions are constructed based on these data through the electricity optimization algorithm, and a corresponding weight coefficient is set for each conditional function. The conditional function and the weight coefficient are combined to generate a comprehensive conditional function. The electricity optimization algorithm is specifically as follows:

[0082]

[0083] In the formula, represents the peak-to-valley difference condition function, Indicates the maximum total load of the power grid throughout the day. Indicates the minimum total load of the power grid throughout the day. represents the load variance conditional function, express The total grid charge at time, Indicates the total number of time periods throughout the day. represents the electricity price condition function, Indicates the total number of charging piles. Indicates Charging stations at the time working status, Indicates the charging power of the charging pile. Indicates the time interval. express The value of electricity at a given moment, represents the weight coefficient, represents the comprehensive conditional function, and Represents parameters;

[0084] S405, determining the connection end time of each electric vehicle according to the upper limit and lower limit of the adjustable time period of the vehicle to be charged and the connection start time;

[0085] Step S5 specifically includes the following steps:

[0086] S501, determining the initial state of charge and the minimum state of charge of each vehicle to be charged, calculating the actual state of charge using the initial state of charge, charging time, battery capacity, charging power and efficiency of the vehicle, constructing a state of charge constraint function based on these data, and constructing a charging time constraint function using the connection start time and connection end time of each vehicle;

[0087] S502, receiving the scheduling priority, total number, comprehensive condition function, state of charge constraint function and charging time constraint function of the vehicles to be charged, determining the dimension of the search space according to the total number of vehicles to be charged, setting the data particle group size, number of cycles, maximum moving speed, initial weight value, initial local factor and initial global factor, and determining the initial velocity vector and initial position vector of different data particles in each dimensional space according to the scheduling priority of each vehicle;

[0088] S503, calculating the corresponding conditional function value according to the velocity vector and position vector of each data particle, if the value is greater than the original conditional function value, updating the local optimal charging start time and the global optimal charging start time, otherwise the update operation is not performed;

[0089] S504, using the parameter update algorithm to calculate the new weight value, local factor and global factor according to the current number of cycles, and analyzing the new velocity vector and position vector of each data particle in each dimensional space according to the constraint function. The parameter update algorithm is specifically as follows:

[0090]

[0091] In the formula, Indicates The weight value in the next cycle, represents the minimum weight value, represents the maximum weight value, represents a constant, Indicates the parameters, Indicates the total number of cycles, Indicates The local factor in the sub-cycle, represents the minimum local factor, represents the maximum local factor, Indicates The global factor in the sub-cycle, represents the minimum global factor, represents the maximum global factor, represents a natural constant;

[0092] S505, determining whether the speed values ​​of each data particle in different dimensional spaces are all less than the maximum moving speed, if they are all less than the maximum moving speed, no operation is performed, otherwise they are adjusted according to the maximum moving speed;

[0093] S506, repeat the operation until the number of cycles reaches the set number, output the optimal charging start time for each electric vehicle, perform charging operations according to these data and update the corresponding basic load data in different time periods;

[0094] Step S6 specifically includes the following steps:

[0095] S601. An AC pile and a DC pile are provided in each charging pile, and two output interfaces are provided in the charging pile, namely a direct output interface of the AC pile and a DC output interface for converting AC power into DC power;

[0096] S602, using multiple temperature and humidity sensors to collect temperature and humidity data at various locations of the AC pile and the DC pile, using a device detector to analyze the AC pile and the DC pile respectively, so as to obtain the voltage and current values ​​of each point inside these devices, and transmit these data to the database, the temperature and humidity sensor is specifically a DHT11 temperature and humidity sensor, and the device detector is specifically a CPT200 charging pile inspection instrument;

[0097] Step S6 specifically also includes the following steps:

[0098] S603, after extracting the real-time monitoring data of the AC pile and the DC pile, transmit it to the fault prediction model for analysis. If the model determines that there is a component abnormality in the AC pile, the power supply of the AC pile direct output interface will be immediately stopped, switched to the DC output interface, and a fault warning message will be sent to the control center. If the model determines that there is a component abnormality in the DC pile, the power supply of the DC output interface will be immediately stopped, switched to the AC pile direct output interface, and a fault warning message will be sent to the control center. The fault prediction model is specifically an LSTM model;

[0099] S604. If the model determines that the temperature inside the AC pile is too high, the power supply of the AC pile direct output interface will be stopped immediately and switched to the DC output interface. If the model determines that the temperature inside the DC pile is too high, the power supply of the DC output interface will be stopped immediately and switched to the AC pile direct output interface.

[0100] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0101] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A charging pile operation control method based on AC / DC switching backup, characterized in that: The method comprises the following steps: S1. Analyze historical charging records: extract all historical charging records of each electric vehicle, determine the first characteristic data and the second characteristic data, classify them using a clustering model, store similar data in the same set, perform normal fitting on each set corresponding to the first characteristic data, analyze the fitted data using an adjustable time analysis algorithm, obtain the upper and lower limits of the adjustable time period of the electric vehicle, count all charging interval values ​​of each electric vehicle, determine the relevant attributes of each electric vehicle based on the charging interval value and the second characteristic data, determine the scheduling priority of the electric vehicle based on the relevant attributes, count the number of times the charging pile corresponding to each electric vehicle is used, and determine the number of the charging pile within the preferred interval of each electric vehicle based on the number of times it is used; S2. Training behavior prediction model: After receiving the upper and lower limits of the adjustable time period, the scheduling priority and the preference interval of all electric vehicles, the historical charging records corresponding to all electric vehicles are sequentially transmitted to the user behavior prediction model, and the user behavior prediction model is used to extract features of these historical charging records, and the upper and lower limits of the adjustable time period, the scheduling priority and the preference interval corresponding to each electric vehicle are determined. The deviation value between the predicted result and the actual result is analyzed using the optimization function, and the model parameters are iteratively updated until the model training is completed; S3. Determine the best charging pile: after receiving the charging request information from the electric vehicle, the electric vehicle that issued the charging request is regarded as the vehicle to be charged, all historical charging records of the vehicle to be charged are extracted, and the user behavior prediction model is used to analyze them, and the upper and lower limits of the adjustable time period corresponding to the electric vehicle, the scheduling priority and the preference interval are determined, and the preference interval corresponding to each vehicle to be charged is analyzed by deep learning technology, so as to obtain the best charging pile number for each vehicle to be charged; S4. Constructing conditional functions: extracting basic data and behavior data of the vehicles to be charged, calculating the corresponding default charging period according to the connection start time and charging duration in the behavior data of each electric vehicle, and determining the number of charging piles in non-idle working state at different times, obtaining the total load of the power grid in each time period according to the working state, charging power of each charging pile at different times and the basic load corresponding to the power grid in different time periods, and drawing the corresponding load curve diagram, extracting the corresponding electricity value at different times, and constructing the corresponding comprehensive conditional function through the power optimization algorithm; S5. Output the optimal charging time: after receiving the scheduling priority, total number, comprehensive condition function, state of charge constraint function and charging time constraint function of the vehicles to be charged, determine the dimension of the search space according to the total number of vehicles to be charged, set the data particle group size and the number of cycles, the maximum moving speed, the initial weight value, the initial local factor and the initial global factor, determine the initial velocity vector and the initial position vector of different data particles in each dimensional space according to the scheduling priority of each vehicle, calculate the corresponding condition function value according to the velocity vector and position vector of each data particle, and update the local optimal charging start time and the global optimal charging start time, analyze the new velocity vector and position vector of each data particle in each dimensional space according to the constraint function, repeat the operation until the number of cycles reaches the set number, and output the optimal charging start time for each electric vehicle; S6. Switch the power transmission interface: Each charging pile is equipped with an AC pile and a DC pile, and the AC pile is set with a direct output interface, and the DC pile is set with a DC output interface. After extracting the real-time monitoring data of the AC pile and the DC pile, transmit it to the fault prediction model for analysis. If the fault prediction model determines that there is a component abnormality in the AC pile, switch to the DC output interface. If the fault prediction model determines that there is a component abnormality in the DC pile, switch to the direct output interface.

2. The charging pile operation control method based on AC / DC switching backup according to claim 1 is characterized in that: The step S1 specifically includes the following steps: S101, extracting all historical charging records of each electric vehicle, wherein the historical charging records include user number, connection start time, charging start time, connection end time, charging end time and charging pile number, and analyzing the charging time and connection time corresponding to each historical charging record, combining the connection start time and connection time into the first feature data, and then combining the charging time and connection time into the second feature data; S102, using a clustering model to classify the first characteristic data and the second characteristic data of each electric vehicle, and storing the same type of data in the same set, performing normal fitting on each set corresponding to the first characteristic data, and statistically analyzing the data corresponding to the highest point in each set after the fitting; S103, calculating the corresponding weight according to the amount of data in each set, analyzing the data corresponding to the highest point of each set and the weight according to the weight using the adjustable time analysis algorithm, and obtaining the upper limit and lower limit of the current adjustable time period of the electric vehicle.

3. The charging pile operation control method based on AC / DC switching backup according to claim 2 is characterized in that: The step S1 specifically further comprises the following steps: S104, analyzing multiple charging interval values ​​from all historical charging records of each electric vehicle, and counting the proportion of charging interval values ​​within 36 hours. If the proportion is higher than a first threshold, the electric vehicle is determined to be frequently scheduled, otherwise it is determined to be infrequently scheduled. Determine the proportion of the difference between the connection time and the charging time in the second characteristic data that is not zero. If the proportion is higher than a second threshold, the electric vehicle is determined to be strongly responsive. Otherwise, it is determined to be weakly responsive. S105, determining the relevant attributes of each electric vehicle, if the attribute is frequent scheduling and strong response scheduling, then determining the scheduling priority of the electric vehicle to be level one, if the attribute is infrequent scheduling and strong response scheduling, then determining the scheduling priority of the electric vehicle to be level two, if the attribute is frequent scheduling and weak response scheduling, then determining the scheduling priority of the electric vehicle to be level three, if the attribute is infrequent scheduling and weak response scheduling, then determining the scheduling priority of the electric vehicle to be level four; S106: Count the usage times of the charging pile corresponding to each electric vehicle, and store the numbers of the charging piles whose usage times are greater than a third threshold into the preference interval of the corresponding vehicle.

4. The charging pile operation control method based on AC / DC switching backup according to claim 1 is characterized in that: The step S3 specifically comprises the following steps: S301, after receiving the charging request information sent by the electric vehicle, the corresponding user number is stored in the set of vehicles to be charged, and the database is queried in sequence according to the numbers contained in the set. If the corresponding historical charging record is not found, the upper and lower limits of the adjustable time period are directly determined from the vehicle user, and the scheduling priority is set to level 4, and the preference interval is empty. Otherwise, all corresponding historical charging records are extracted; S302, using the user behavior prediction model to analyze the historical charging records of different electric vehicles in turn, and determine the upper and lower limits of the adjustable time period, the scheduling priority, and the preferred interval corresponding to each electric vehicle; S303: Analyze the preference interval corresponding to each car to be charged by deep learning technology, so as to obtain the best charging pile number and charging pile location information of each car to be charged, and return it to the car user interface.

5. The charging pile operation control method based on AC / DC switching backup according to claim 1 is characterized in that: The step S4 specifically comprises the following steps: S401, after obtaining the total number of charging piles, charging power, charging efficiency and basic load data in the current charging station, extract the basic data and behavior data of the vehicles to be charged, and calculate the corresponding default charging period according to the connection start time and charging duration in the behavior data of each electric vehicle; S402, determining the number of charging piles in a non-idle working state at different times according to the default charging time period of each electric vehicle, and analyzing the working state of each charging pile at different times, the charging power, and the basic load corresponding to the power grid at different time periods by a load analysis algorithm, thereby obtaining the total load of the power grid at different time periods, and storing it in a database; S403, repeating the operation until all electric vehicle data are extracted, drawing a corresponding load curve using the total load of the power grid in each time period, and determining the maximum and minimum values ​​corresponding to the curve; S404. After extracting the electricity values ​​corresponding to different times, multiple conditional functions are constructed through the electricity optimization algorithm according to the total charge of the power grid at different times, the total number of charging piles, the working status of each charging pile, the charging power, the electricity value at each time, and the maximum and minimum values ​​corresponding to the curve graph, and a corresponding weight coefficient is set for each conditional function. The conditional function and the weight coefficient are combined to generate a comprehensive conditional function; S405: Determine the connection end time of each electric vehicle according to the upper limit and lower limit of the adjustable time period of the vehicle to be charged and the connection start time.

6. The charging pile operation control method based on AC / DC switching backup according to claim 1 is characterized in that: The step S5 specifically comprises the following steps: S501, determining the initial state of charge and the minimum state of charge of each vehicle to be charged, calculating the actual state of charge using the initial state of charge, charging time, battery capacity, charging power and efficiency of the vehicle to be charged, constructing a state of charge constraint function, and then constructing a charging time constraint function using the connection start time and connection end time of each vehicle; S502, receiving the scheduling priority, total number, comprehensive condition function, state of charge constraint function and charging time constraint function of the vehicles to be charged, determining the dimension of the search space according to the total number of vehicles to be charged, setting the data particle group size, number of cycles, maximum moving speed, initial weight value, initial local factor and initial global factor, and determining the initial velocity vector and initial position vector of different data particles in each dimensional space according to the scheduling priority of each vehicle; S503, calculating the corresponding conditional function value according to the velocity vector and position vector of each data particle, if the conditional function value calculated this time is greater than the original conditional function value, then updating the local optimal charging start time and the global optimal charging start time, otherwise, no update operation is performed; S504, using a parameter update algorithm to calculate new weight values, local factors and global factors according to the current number of cycles, and analyzing new velocity vectors and position vectors of each data particle in each dimensional space according to the constraint function; S505, determining whether the speed values ​​of each data particle in different dimensional spaces are all less than the maximum moving speed, if they are all less than the maximum moving speed, no operation is performed, otherwise they are adjusted according to the maximum moving speed; S506, repeat the operation until the number of cycles reaches the set number, output the optimal charging start time for each electric vehicle, perform the charging operation according to the optimal charging start time for each electric vehicle, and update the corresponding basic load data in different time periods.

7. The charging pile operation control method based on AC / DC switching backup according to claim 1 is characterized in that: The step S6 specifically comprises the following steps: S601. An AC pile and a DC pile are provided in each charging pile, and two output interfaces are provided in the charging pile, namely a direct output interface of the AC pile and a DC output interface for converting AC power into DC power; S602. Use multiple temperature and humidity sensors to collect temperature and humidity data at various locations of the AC pile and the DC pile, use equipment detectors to analyze the AC pile and the DC pile respectively, so as to obtain the voltage value and current value of each point inside the AC pile and the DC pile, and transmit the voltage value and current value of each point to the database.

8. The charging pile operation control method based on AC / DC switching backup according to claim 7 is characterized in that: The step S6 specifically further comprises the following steps: S603, after extracting the real-time monitoring data of the AC pile and the DC pile, transmit it to the fault prediction model for analysis. If the fault prediction model determines that there is a component abnormality in the AC pile, the power supply of the AC pile direct output interface will be immediately stopped, switched to the DC output interface, and a fault warning message will be sent to the control center. If the fault prediction model determines that there is a component abnormality in the DC pile, the power supply of the DC output interface will be immediately stopped, switched to the AC pile direct output interface, and a fault warning message will be sent to the control center; S604. If the fault prediction model determines that the temperature inside the AC pile is too high, the power supply of the AC pile direct output interface will be stopped immediately and switched to the DC output interface. If the fault prediction model determines that the temperature inside the DC pile is too high, the power supply of the DC output interface will be stopped immediately and switched to the AC pile direct output interface.

Images