Charging station load optimization method and device based on ant colony algorithm and medium

Abstract

The invention relates to the technical field of charging station load optimization, in particular to a charging station load optimization method and device based on an ant colony algorithm and a medium. Comprising the steps that firstly, an optimized multi-objective function is provided, the function comprises the minimum total power cost, the maximum charging pile utilization rate, the balance load and the minimum charging pile loss, and it is ensured that the optimal effect of an optimization result is achieved under the condition that the optimization result is not separated from an actual scene; secondly, a load optimization experience learning model is provided, and the model obtains optimal load optimization experience by correcting historical data; and finally, proposing a constraint of a load optimization model on a load optimization result, fusing an optimal load optimization experience as an optimization suggestion, and solving a charging station load optimization strategy. According to the method, the load optimization effect of the charging station can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of charging station load optimization, and in particular to a charging station load optimization method, device and medium based on an ant colony algorithm. Background Art

[0002] As electric vehicles become more popular, the demand for charging stations has increased significantly. This trend brings many challenges, including the layout of charging infrastructure, the number and distribution of charging piles, and the management of charging load.

[0003] Effective management of charging loads is conducive to solving problems such as power network overload and long charging waiting time, and can improve the operational efficiency of charging stations. Ant colony algorithm is an optimization algorithm that simulates the foraging behavior of ants and is widely used to solve complex optimization problems. Ant colony algorithm performs well in a variety of optimization tasks. Applying ant colony algorithm to charging station load optimization can effectively achieve global optimal solution, adaptive load adjustment, parallel computing processing, and improve the adaptability to dynamic load changes.

[0004] There are still some defects in the application of existing technologies. For example, most methods predict the load of charging piles and make adjustments based on the predicted load. These methods may cause large errors in the optimization of charging station loads in actual applications, which is not conducive to the effective distribution of charging station loads.

[0005] To this end, the present invention proposes a charging station load optimization method, device and medium based on ant colony algorithm. Summary of the invention

[0006] The purpose of the present invention is to provide a method, device and medium for load optimization of charging stations based on ant colony algorithm, mainly starting from the following points; first, the present invention defines a load optimization multi-objective function for load optimization of charging stations, which includes: minimizing total power cost, maximizing charging pile utilization, balancing load and minimizing charging pile loss, thereby improving the economic benefits, service efficiency and equipment life of charging stations; secondly, the historical charging pile usage data of the charging station is collected for preprocessing, and the load optimization experience learning model is used to perform load optimization experience learning on the preprocessed data to obtain optimal load optimization experience; finally, a load optimization model improved based on ant colony algorithm is used to optimize the load of the charging piles currently in use. The model sets load constraints to control the optimization parameters so as not to deviate from the actual situation, and integrates the optimal load optimization experience as a guide into the ant colony algorithm to achieve accurate optimization of the charging pile load.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention proposes a charging station load optimization method based on an ant colony algorithm, comprising:

[0009] Define a multi-objective function for load optimization of the charging station; wherein the multi-objective function for load optimization includes: minimizing total power cost, maximizing charging pile utilization, balancing load and minimizing charging pile loss; wherein, minimizing total power cost includes: defining total power cost; collecting power cost related data of the charging station including: power price data, load demand data and charging status data; performing the preprocessing on the power cost related data to obtain standard power cost related data; establishing a total power cost optimization model; solving the minimization of total power cost using the total power cost optimization model;

[0010] The maximizing charging pile utilization rate comprises: defining the charging pile utilization rate; collecting charging pile data including: charging pile usage data, charging pile available time and charging pile status data; performing the preprocessing on the charging pile data to obtain standard charging pile data; establishing a charging pile usage optimization model; and solving the maximizing charging pile utilization rate using the charging pile usage optimization model;

[0011] The load balancing includes: defining a load balancing index; obtaining relevant load data including load data and resource data; performing the preprocessing on the relevant load data to obtain standard relevant load data; establishing a load balancing optimization model; and solving the load balancing using the load balancing optimization model.

[0012] The minimization of charging pile loss includes: defining a charging pile life loss index; collecting relevant equipment life loss data including: charging pile life loss rate, failure frequency and maintenance cost; preprocessing the relevant equipment life loss data to obtain standard relevant equipment life loss data; establishing a charging pile loss optimization model; and solving the minimization of charging pile loss using the charging pile loss optimization model;

[0013] Furthermore, the load optimization multi-objective function is generated according to the minimization of the total power cost, the maximization of the charging pile utilization, the balancing load and the minimization of the charging pile loss, and is expressed as:

[0014] minimize F(x1,x2,x3,x4)=(w1*f1(x1),-w2*f2(x2),w3*f3(x3),w4*f4(x4));

[0015] Wherein, minimize is represented as a minimization function; F() is represented as the load optimization multi-objective function; x1 is represented as a decision variable for minimizing electricity cost; x2 is represented as a decision variable for maximizing the utilization rate of charging piles; x3 is represented as a decision variable for balancing loads; x4 is represented as a decision variable for minimizing the loss of charging piles; f1() is represented as the total electricity cost optimization model; f2() is represented as a charging pile usage optimization model; f3() is represented as a balanced load optimization model; f4() is represented as a charging pile loss optimization model; w1, w2, w3 and w4 are represented as weight coefficients;

[0016] Further, historical charging pile usage data related to the charging station is obtained; wherein the historical charging pile usage data includes: historical charging pile charging records, historical user behavior data and historical charging pile status data;

[0017] Further, the historical charging pile usage data is preprocessed to obtain standard historical charging pile usage data; wherein the preprocessing includes: data cleaning of the historical charging pile usage data to obtain first historical charging pile usage data; data conversion of the first historical charging pile usage data to obtain second historical charging pile usage data; standardization of the second historical charging pile data to obtain third historical charging pile data; data integration of the third historical charging pile data to obtain the standard historical charging pile usage data;

[0018] Furthermore, the standard historical charging pile usage data is input into a load optimization experience learning model to obtain optimal load optimization experience;

[0019] The load optimization experience learning model includes: using feature engineering technology to perform feature processing on the standard historical charging pile usage data to obtain historical charging pile usage feature data; integrating the historical charging pile usage feature data to obtain a historical charging pile usage feature set; inputting the historical charging pile usage feature set into the load optimization multi-objective function to perform target value fitting, and calculating the fitting error; using the fitting error to perform data correction on the data in the historical charging pile usage feature set, and calculating the correction loss; when the correction loss and the fitting error converge, the optimal load optimization experience is obtained;

[0020] Further, obtaining current charging pile usage data of the charging station;

[0021] Further, the current charging pile usage data is preprocessed and input into the load optimization model together with the optimal load optimization experience, and the current charging station load optimization strategy is output; wherein, the load optimization model includes: setting the load constraint conditions of the charging station includes: Among them, CULi Represents the load of the charging pile i in use; MaxL i represents the maximum load of the charging pile i in use; M represents the number of charging piles in use; ToL represents the total load of the charging station; CUL i (t) represents the load of the charging pile i at time t; MaxD represents the maximum load change between time t and time t-1;

[0022] The number of ants, pheromone volatility coefficient and heuristic function weight are set; the pheromone concentration is initialized according to the optimal load optimization experience; the probability of each ant selecting the charging pile load distribution path is calculated according to the pheromone concentration and the heuristic function; wherein, the parameters of the preprocessed current charging pile usage data are adjusted according to the load constraint conditions on the path; the load optimization multi-objective function value of the path selected by each ant is calculated using the fitness function; the pheromone concentration is updated according to the path selected by each ant; the information concentration of the global optimal path is updated; when the load optimization multi-objective function value of the global optimal path reaches the optimal value, the update is stopped to obtain the current charging station load optimization strategy.

[0023] Furthermore, the current charging station load optimization strategy is evaluated.

[0024] In a second aspect, the present invention proposes a charging station load optimization device based on an ant colony algorithm, comprising: a data acquisition module, a data storage module, a data preprocessing module, a load optimization experience learning module, a load optimization module, an evaluation module, an execution module and a feedback module;

[0025] Wherein, the data collection module is used to collect the usage data of the charging piles in the charging station;

[0026] The data storage module is used to store historical charging pile usage data in the charging station;

[0027] The data preprocessing module is used to preprocess the usage data of the charging pile;

[0028] The load optimization experience learning module is used to learn the load optimization experience of the charging station; the load optimization experience learning module adopts a load optimization experience learning model, specifically including: using feature engineering technology to perform feature processing on standard historical charging pile usage data to obtain historical charging pile usage feature data;

[0029] Further, the historical charging pile usage feature data is integrated to obtain a historical charging pile usage feature set;

[0030] Furthermore, the historical charging pile usage feature set is input into the load optimization multi-objective function for target value fitting, and the fitting error is calculated; wherein the load optimization multi-objective function includes: minimizing the total power cost, maximizing the charging pile utilization, balancing the load and minimizing the charging pile loss; wherein, minimizing the total power cost includes: defining the total power cost; collecting the power cost related data of the charging station including: power price data, load demand data and charging status data; performing the preprocessing on the power cost related data to obtain the standard power cost related data; establishing a total power cost optimization model; and solving the minimized total power cost using the total power cost optimization model;

[0031] The maximizing charging pile utilization rate comprises: defining the charging pile utilization rate; collecting charging pile data including: charging pile usage data, charging pile available time and charging pile status data; performing the preprocessing on the charging pile data to obtain standard charging pile data; establishing a charging pile usage optimization model; and solving the maximizing charging pile utilization rate using the charging pile usage optimization model;

[0032] The load balancing includes: defining a load balancing index; obtaining relevant load data including load data and resource data; performing the preprocessing on the relevant load data to obtain standard relevant load data; establishing a load balancing optimization model; and solving the load balancing using the load balancing optimization model.

[0033] The minimization of charging pile loss includes: defining a charging pile life loss index; collecting relevant equipment life loss data including: charging pile life loss rate, failure frequency and maintenance cost; preprocessing the relevant equipment life loss data to obtain standard relevant equipment life loss data; establishing a charging pile loss optimization model; and solving the minimization of charging pile loss using the charging pile loss optimization model;

[0034] The load optimization multi-objective function is generated according to the minimization of the total power cost, the maximization of the charging pile utilization, the balancing load and the minimization of the charging pile loss, and is expressed as:

[0035] minimize F(x1,x2,x3,x4)=(w1*f1(x1),-w2*f2(x2),w3*f3(x3),w4*f4(x4));

[0036] Wherein, minimize is represented as a minimization function; F() is represented as the load optimization multi-objective function; x1 is represented as a decision variable for minimizing electricity cost; x2 is represented as a decision variable for maximizing the utilization rate of charging piles; x3 is represented as a decision variable for balancing loads; x4 is represented as a decision variable for minimizing the loss of charging piles; f1() is represented as the total electricity cost optimization model; f2() is represented as a charging pile usage optimization model; f3() is represented as a balanced load optimization model; f4() is represented as a charging pile loss optimization model; w1, w2, w3 and w4 are represented as weight coefficients;

[0037] Further, the fitting error is used to correct the data in the historical charging pile usage feature set, and the correction loss is calculated;

[0038] Further, when the correction loss and the fitting error reach convergence, an optimal load optimization experience is obtained;

[0039] The load optimization module is used to optimize the load of the charging pile in use; wherein the load optimization module adopts a load optimization model, specifically including: setting the load constraint conditions of the charging station includes: Among them, CUL i Represents the load of the charging pile i in use; MaxL i represents the maximum load of the charging pile i in use; M represents the number of charging piles in use; ToL represents the total load of the charging station; CUL i (t) represents the load of the charging pile i at time t; MaxD represents the maximum load change between time t and time t-1;

[0040] Set the number of ants, pheromone volatility coefficient and heuristic function weight; initialize pheromone concentration according to the optimal load optimization experience; calculate the probability of each ant selecting the charging pile load distribution path according to the pheromone concentration and heuristic function; wherein, in the path, adjust the parameters of the pre-processed current charging pile usage data according to the load constraint condition; use the fitness function to calculate the load optimization multi-objective function value of the path selected by each ant; update the pheromone concentration according to the path selected by each ant; update the information concentration of the global optimal path; stop updating when the load optimization multi-objective function value of the global optimal path reaches the optimal value, and obtain the current charging station load optimization strategy;

[0041] The evaluation module is used to evaluate the optimization result of the load optimization module;

[0042] The execution module is used to execute the charging pile load optimization strategy; the feedback module is used to feed back the optimization results to relevant personnel.

[0043] In a fourth aspect, the present invention proposes a charging station load optimization medium based on an ant colony algorithm, wherein the medium may be a computer program executed by an electronic device, and when the program runs on the electronic device, the electronic device executes the steps of the method described in any one of claims 1 to 5.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] 1. The present invention proposes a load optimization multi-objective function for explaining the load optimization target of the charging station; the function takes the electricity cost, charging pile utilization, balanced load and charging pile loss as the final optimization target; in the process of setting the four targets, four model functions are used to quantify the results respectively, so as to maximize the economic benefits, improve the equipment utilization, enhance the system stability and extend the equipment life; thereby realizing the effective distribution of the charging station load.

[0046] 2. The present invention proposes a charging station load optimization experience learning method; the method collects historical charging pile usage data of the charging station and characterizes it; uses the load optimization multi-objective function to fit the target value of the characterized data, calculates the error value in the fitting process, and corrects the characteristic data according to the error value; obtains the optimal load optimization experience by judging when the correction loss is minimized; the charging station load optimization experience learning method can improve the accuracy of load optimization, reduce errors, and achieve more efficient power use and load distribution.

[0047] 3. The present invention proposes a load optimization model based on ant colony algorithm for load optimization of the charging piles currently in use; the model sets load constraints to adjust parameters of each charging pile in use to achieve the expected value of the multi-objective function value of load optimization; in the optimization process, the model uses the optimal load optimization experience as the adjustment variable of the pheromone concentration of ants in the path selection process based on the ant colony algorithm, thereby achieving the optimal configuration of the charging pile load. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A flowchart of a charging station load optimization method based on an ant colony algorithm provided in an embodiment of the present invention;

[0049] Figure 2 A structural diagram of a charging station load optimization device based on an ant colony algorithm provided in an embodiment of the present invention;

[0050] Figure 3 A schematic diagram of an iterative process of load optimization experience learning provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] 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.

[0052] Charging stations are facilities that provide charging services for electric vehicles. They usually include multiple charging piles, which can quickly charge electric vehicles, support different charging standards, and improve the convenience of using electric vehicles. Charging stations can be set up in public parking lots, commercial areas, residential areas, etc.

[0053] At present, the optimization of charging stations mainly focuses on the optimization of charging loads; a good charge optimization method can achieve efficient use of power resources and grid stability. In this research area, the prior art has achieved good results by applying the ant colony algorithm to the problem of charging station load optimization. The ant colony algorithm can significantly improve the load optimization efficiency of charging stations, ensure the efficient use of charging piles and the stable operation of the power grid; however, there are still some defects in the application of the prior art, such as most methods predict the load of the charging piles and adjust them according to the predicted load; these methods may cause large errors in the actual application of charging station load optimization. Therefore, the present invention proposes a charging station load optimization method, device and medium based on the ant colony algorithm; it will be described in detail from the following embodiments.

[0054] Embodiment 1:

[0055] In the embodiment of the present application, the load of the charging station in area A is optimized by combining the method and device proposed in the present invention; the charging station in area A has 20 charging piles and is a modern electric vehicle charging facility designed to provide convenient charging services for electric vehicles in the area. Generally, the charging station supports different charging standards and can meet the needs of various electric vehicles. In the process of load optimization of the charging station in area A, the following are mainly used: Figure 1 and Figure 2 To achieve together; among them, Figure 1 The steps in Figure 2 The middle part modules correspond; Figure 1 The steps include: S10. defining the optimization target of the charging station; S20. obtaining the historical charging pile usage data; S30. preprocessing the historical charging pile data; S40. learning the load optimization experience; S50. collecting the current charging pile usage data; S60. optimizing the load of the currently used charging pile; S70. evaluating the load optimization results. Figure 2The structure of the device includes: data acquisition module, data storage module, data preprocessing module, load optimization experience learning module, load optimization module, evaluation module, execution module and feedback module; according to Figure 1 and Figure 2 The contents are described as follows;

[0056] according to Figure 1 Step S10 is to set an optimization target for load optimization of the charging station in zone A; wherein, in the embodiment of the application, the optimization target is generated by a load optimization multi-objective function including: minimizing electricity cost, maximizing charging pile utilization, balancing load, and minimizing loss during charging;

[0057] The specific process of generating the load multi-objective function includes:

[0058] The minimization of the total power cost includes: defining the total power cost; collecting the power cost related data of the charging station including: power price data including power prices in different time periods (such as peak power prices, valley power prices, etc.), load demand data including power demand in each time period of the charging station and charging status data including the usage status and charging progress of each charging pile; performing the preprocessing on the power cost related data to obtain standard power cost related data; establishing a total power cost optimization model; the model function of the total power cost optimization model is expressed as: Wherein, minimize represents the minimization function; f1() represents the total power cost optimization model; x1 represents the decision variable for minimizing the power cost; P i,j It is expressed as the functional requirement of charging pile i in time period j; Q i,j It is represented by the charging status of charging pile i in time period j; C j It is represented by the electricity price in time period j; T is represented by the total number of time periods; N is represented by the total number of charging piles;

[0059] The total electricity cost optimization model sets electricity constraints;

[0060] (1) Power demand constraints: Among them, L j It is expressed as the total load demand in time period j;

[0061] (2) Charging pile status constraints: Among them, 0 represents the uncharged state; 1 represents the charged state;

[0062] (3) Power limit constraints: Among them, P i,max It is expressed as the maximum power of charging pile i;

[0063] The total electricity cost optimization model is used to solve the minimized total electricity cost.

[0064] Refer to Table 1, which provides an explanation of the electricity prices and power load requirements of the 20 charging piles in the charging station in area A at different time periods;

[0065] Table 1 Power description of charging stations in Zone A The maximization of charging pile utilization includes: defining charging pile utilization; collecting charging pile data including: charging pile usage data including the actual usage time of each charging pile including the total available time of the charging pile, such as daily working hours and charging amount, charging pile available time, and charging pile status data including whether the charging pile is in use, fault status, and maintenance time, etc.;

[0066] The charging pile data is preprocessed to obtain standard charging pile data; a charging pile usage optimization model is established; the charging pile usage optimization model is expressed as: Where maximize is the maximization function; f2() is the charging pile usage optimization model; x2 is the decision variable for maximizing the utilization rate of the charging pile; T used,i It is represented by the actual usage time of charging pile i; T available,i is the total available time of charging pile i;

[0067] Set utilization constraints for the charging pile using the optimization model;

[0068] (1) Charging pile status constraints:

[0069] (2) Time allocation constraints: Among them, TAT represents the total available time of the charging pile;

[0070] The maximum charging pile utilization rate is solved by using the charging pile utilization optimization model. Referring to Table 2, Table 2 provides instructions for the use of 20 charging piles in the charging station in Area A;

[0071] Table 2 Instructions for using the charging piles at the charging station in Area A

[0072]

[0073] The load balancing includes: defining a load balancing index; obtaining relevant load data including: load data including historical load data, real-time load data, load forecast data, etc. The data may be load values ​​recorded hourly, daily or monthly and resource data including the capacity of available power supply, backup power supply, power generation cost, operating status, etc.; performing the preprocessing on the relevant load data to obtain standard relevant load data; establishing a load balancing optimization model; the load balancing optimization model is expressed as: Among them, f3() represents the load balancing optimization model; x3 represents the decision variable for load balancing; L j It is expressed as the load in time period j; Expressed as the average value of the load;

[0074] Setting load constraints for the balanced load optimization model;

[0075] (1) Load balancing constraints: Among them, S i,j It is represented by the supply power of charging pile i in time period j;

[0076] (2) Resource capacity constraints: Among them, MaxR i It is represented by the maximum resource supply capacity of charging pile i in any time period;

[0077] (3) Load demand constraints: Among them, L min Expressed as minimum load demand;

[0078] The balanced load is solved using the balanced load optimization model.

[0079] Refer to Table 3, which gives the load description of the 20 charging piles in the charging station in area A;

[0080] Table 3 Description of charging pile loads at charging stations in area A

[0081] Charging station type Single pile load (kW) Total load (kW) Load balance index (kW) Fast charging pile 60 480 5 Slow charging pile 20 200 7 Super charging pile 100 200 3

[0082] The minimization of charging pile loss includes: defining a charging pile life loss index; collecting relevant equipment life loss data including: a charging pile life loss rate describing the proportion of the expected life lost by the charging pile due to use, a failure frequency describing the frequency of failures of the charging pile per unit time, and a maintenance cost describing the cost of maintaining and repairing the charging pile;

[0083] Refer to Table 4 to collect relevant equipment life loss data;

[0084] Table 4 Life loss data of relevant equipment in charging stations in Area A The relevant equipment life loss data is preprocessed to obtain standard relevant equipment life loss data; a charging pile loss optimization model is established; the charging pile loss optimization model is expressed as: Among them, f4() represents the charging pile loss optimization model; x4 represents the decision variable for minimizing the charging pile loss; LR i Expressed as the life loss rate of charging pile i; FF i It is represented by the fault frequency of charging pile i; MC iIt is expressed as the maintenance cost of charging pile i; w L Expressed as the weight of life loss; w F Expressed as the weight of the fault frequency; w M Expressed as the weight of maintenance cost;

[0085] Set negative loss constraints for the electric pile loss optimization model;

[0086] (1) Lifespan limit: Among them, MaxLR represents the maximum life loss rate;

[0087] (2) Restrictions on Use: Among them, U i It represents the usage time of charging pile i; MaxU represents the actual maximum usage time;

[0088] (3) Fault: Among them, MaxFF represents the maximum failure frequency;

[0089] (4) Maintenance cost: Among them, MaxMC represents the maximum maintenance cost of the charging pile;

[0090] The charging pile loss optimization model is used to solve the minimization of charging pile loss.

[0091] Refer to Table 5, which gives the loss description of 20 charging piles in the charging station in area A;

[0092] Table 5 Charging pile power consumption description of charging station in area A

[0093] Charging station type Failure frequency (times / kWh) Life loss rate Standardized maintenance costs Fast charging pile 2.5 0.12 0.75 Slow charging pile 1.5 0.10 0.50 Super charging pile 1.0 0.08 0.25

[0094] The load optimization multi-objective function is generated according to the minimization of the total power cost, the maximization of the charging pile utilization, the balancing load and the minimization of the charging pile loss, and is expressed as:

[0095] minimize F(x1,x2,x3,x4)=(w1*f1(x1),-w2*f2(x2),w3*f3(x3),w4*f4(x4));

[0096] Wherein, F() represents the load optimization multi-objective function; w1, w2, w3 and w4 represent weight coefficients.

[0097] In the embodiment of the present application, the load objective function is set to set the standard for load optimization of the charging station in area A, including: minimizing the total electricity cost, maximizing the utilization rate of charging piles, balancing the load and minimizing the loss of charging piles; this can achieve optimization among reducing costs, improving equipment utilization, maintaining load balance and reducing losses, thereby improving overall operational efficiency.

[0098] Furthermore, the data storage module of the device obtains the historical charging pile usage data of the charging station in zone A for nearly one year, corresponding to the above-mentioned step S20; wherein the historical charging pile usage data includes: historical charging pile charging records, historical user behavior data and historical charging pile status data.

[0099] In this embodiment of the application, historical data of the charging stations in area A is obtained, including: historical charging records of charging piles, historical user behavior data and historical charging pile status data; this helps to conduct load optimization experience learning based on comprehensive historical information, thereby improving the accuracy and effectiveness of optimization decisions.

[0100] See Table 6, which shows the historical charging pile usage data according to detailed attribute descriptions.

[0101] Table 6 Description of historical charging pile usage data of charging stations in Area A Furthermore, the data preprocessing module of the device preprocesses the historical charging pile usage data, corresponding to the above-mentioned step S30; the specific process includes: performing data cleaning on the historical charging pile usage data to obtain first historical charging pile usage data; performing data conversion on the first historical charging pile usage data to obtain second historical charging pile usage data; performing standardization on the second historical charging pile data to obtain third historical charging pile data; and performing data integration on the third historical charging pile data to obtain the standard historical charging pile usage data.

[0102] In the embodiment of the present application, data preprocessing is performed on the historical charging station usage data collected by the charging stations in zone A, which can improve the accuracy and effect of load optimization and optimize the experience learning based on historical data.

[0103] Further, the load optimization experience learning module of the device is used to learn the load optimization experience from the standard historical charging pile usage data, corresponding to the above-mentioned step S40; wherein, the load optimization experience learning model is used for learning the load optimization experience, and the specific process includes: using feature engineering technology to perform feature processing on the standard historical charging pile usage data to obtain historical charging pile usage feature data; the feature extraction using feature engineering technology mainly includes the following aspects;

[0104] 1. Historical charging records of charging piles: number of charges, charging time, charging amount and charging time period

[0105] (1) Statistical characteristics: average charging time, total charge capacity, and rate of change of charging times;

[0106] (2) Time characteristics: number of charging times during peak hours and number of charging times during off-peak hours;

[0107] (3) Distribution characteristics: charge capacity distribution (mean, variance, maximum and minimum value of charge capacity) and charging time distribution (mean, variance, maximum and minimum value of charging time);

[0108] 2. Historical user behavior data: user charging frequency, charging time preference, charging capacity requirements and charging mode

[0109] (1) Behavioral pattern characteristics: user activity and time preference;

[0110] (2) Demand characteristics: average charging demand and charging mode distribution;

[0111] (3) Behavioral sequence characteristics: changes in charging frequency;

[0112] 3. Historical charging pile status data: number of failures, maintenance records, working status and operating time

[0113] (1) Fault characteristics: fault frequency and fault type;

[0114] (2) Maintenance characteristics: maintenance interval and maintenance duration;

[0115] (3) Operation status characteristics: normal operation time and downtime time;

[0116] As shown in Table 7, historical charging pile usage characteristic data corresponding to standard historical charging pile usage data in a certain period of time is given in Table 7;

[0117] Table 7 Historical charging pile usage characteristics data for a certain period of time

[0118] Feature data items Eigenvalue Average charging time (hours) 1.5 Total charging capacity (kWh) 800 Peak charging times 50 Average charge capacity (kWh) 6.67 Average charging time (hours) 1.55 Charge capacity variance (kWh2) 10.5 Charging time variance (hours2) 0.2 Failure frequency (times / day) 0.1 Maintenance interval (days) 30 Normal operation time (hours) 280 Downtime (hours) 10

[0119] Further, the historical charging pile usage feature data is integrated to obtain a historical charging pile usage feature set;

[0120] Furthermore, the historical charging pile usage feature set is input into the load optimization multi-objective function to perform target value fitting, and the fitting error is calculated; the fitting error is calculated using a mean square error method, and the formula is: Among them, FE represents the fitting error value; AF k Represented as the kth historical charging pile usage feature; FFD k It is represented as the k-th fitting charging pile usage feature; K is represented as the total number of feature data items;

[0121] Further, the fitting error is used to correct the data in the historical charging pile usage feature set, and the correction loss is calculated; comprising the following steps:

[0122] Step 1: Identify the data features that need to be corrected based on the calculated fitting error;

[0123] Step 2: Use regression correction method to correct the data features that need to be corrected;

[0124] Step 3: Input the corrected data characteristics into the load optimization multi-objective function and recalculate the target value;

[0125] Step 4: Calculate the characteristic data adjustment error during the correction process as the correction loss;

[0126] Further, when the correction loss and the fitting error reach convergence, the optimal load optimization experience is obtained;

[0127] See also Figure 3 ,exist Figure 3 The modified loss and the change of the modified loss in the process of multiple learning of load optimization experience are given in the figure. When the iterative learning reaches 80 times, the values ​​of the modified loss and the fitting error reach convergence.

[0128] Further, the data collection module of the device is used to collect the data of the charging piles currently in use at the charging station in area A, corresponding to the above step S50; see Table 8;

[0129] Table 8 Data description of charging piles currently in use at charging stations in area A Further, the data in Table 8 is processed by the data preprocessing module and input into the load optimization module of the equipment together with the optimal load optimization experience, corresponding to the above step S60;

[0130] The load optimization module uses a load optimization model to perform load optimization, and the specific process includes:

[0131] The load constraint conditions for setting the charging station include: Among them, CUL i Represents the load of the charging pile i in use; MaxL i represents the maximum load of the charging pile i in use; M represents the number of charging piles in use; ToL represents the total load of the charging station; CUL i (t) represents the load of the charging pile i at time t; MaxD represents the maximum load change between time t and time t-1;

[0132] Set the number of ants, pheromone volatility coefficient and heuristic function weight;

[0133] Initializing pheromone concentration according to the optimal load optimization experience;

[0134] The probability of each ant selecting a charging pile load distribution path is calculated according to the pheromone concentration and the heuristic function; wherein, in the path, the parameters of the pre-processed current charging pile usage data are adjusted according to the load constraint condition;

[0135] Use the fitness function to calculate the load optimization multi-objective function value of the path selected by each ant;

[0136] Update pheromone concentrations based on the path chosen by each ant;

[0137] Update the information density of the global optimal path; stop updating when the load optimization multi-objective function value of the global optimal path reaches the optimal value, and obtain the current charging station load optimization strategy;

[0138] Refer to Table 9, which shows the comparison of the load optimization of the charging piles in use at the charging station in area A;

[0139] Table 9 Comparison results of currently used charging pile load after optimization

[0140]

[0141] In the embodiment of the present application, a load optimization model based on an ant colony algorithm is used to optimize the charging piles currently in use; the model sets load optimization conditions during the optimization process to prevent the optimized charging pile parameters from being out of touch with the actual situation; and the learned optimal load optimization experience is used in the adjustment strategy of ants seeking paths in the ant colony algorithm; this effectively improves the system's global optimization capability, dynamic adaptability, and practical feasibility, while accumulating experience for continuous optimization. This method can not only improve system performance and user experience, but also reduce energy waste and ensure the stable and efficient operation of charging piles in practical applications.

[0142] Further, the execution module of the device performs optimization operations according to the current charging station load optimization strategy obtained by the load optimization module;

[0143] Furthermore, the evaluation module of the device evaluates the optimized result, corresponding to the above step S70; wherein the evaluation of the load optimization result is mainly carried out from the following aspects:

[0144] (1) Power distribution: Check whether the optimized charging power of each charging pile meets the expected load distribution strategy to ensure that the total power meets the system load requirements.

[0145] (2) Load balance: Compare the power distribution before and after optimization to evaluate whether the optimization has improved load balance. For example, whether the power reduction of the fast charging pile helps to avoid overload.

[0146] (3) Charging efficiency: Analyze the impact of optimized power on charging efficiency. Confirm whether the charging efficiency is improved or whether the reduction in power affects the charging speed.

[0147] (4) System load: Calculate the total power requirement of the optimized system to ensure that the needs of all charging piles are met without exceeding the system load.

[0148] (5) User experience: Evaluate the impact of charging power adjustment on user experience and ensure that the optimized charging time and speed are still within an acceptable range.

[0149] Refer to Table 10, which gives the evaluation results after load optimization.

[0150] Table 10A District Charging Station Load Optimization Evaluation

[0151]

[0152] Furthermore, the feedback module of the device submits feedback information to relevant staff.

[0153] In the example of the present application, the load optimization of the charging station in area A is realized by the present invention, mainly from the following aspects: first, a load optimization multi-objective function is established according to the actual operation standard of the charging station; the function realizes the optimal load optimization effect by constraining the electricity cost, charging pile utilization, load balance and charging pile loss; wherein, the present invention adopts a load optimization experience learning model to learn the historical charging pile usage data of the charging station in area A to obtain the optimal load optimization strategy; finally, the present invention uses the input load optimization model based on the ant colony algorithm to optimize the load of the charging pile currently in use; in the model, load constraints are set to control the optimization parameters not to deviate from the actual situation, and the optimal load optimization experience is integrated into the ant colony algorithm as a guide to realize the precise optimization of the charging pile load.

[0154] Embodiment 2:

[0155] In Example 1, the method and device proposed in the present invention are used to perform real-time load optimization on the charging station in Area A. In order to further illustrate the effectiveness of the present invention, in Example 2, the load optimization is performed again on the charging station in Area B.

[0156] The specific process of load optimization includes: defining a load optimization multi-objective function for the charging station in zone B; wherein the load optimization multi-objective function includes: minimizing the total power cost, maximizing the utilization rate of the charging pile, balancing the load and minimizing the loss of the charging pile;

[0157] Wherein, the minimization of the total power cost includes: defining the total power cost; collecting power cost related data of the charging station in area B including: power price data, load demand data and charging status data; performing the preprocessing on the power cost related data to obtain standard power cost related data; establishing a total power cost optimization model; solving the minimized total power cost using the total power cost optimization model;

[0158] Refer to Table 11 for power description of charging stations in Area B;

[0159] Table 11B Area Charging Station Power Description The maximizing charging pile utilization rate comprises: defining the charging pile utilization rate; collecting charging pile data including: charging pile usage data, charging pile available time and charging pile status data; performing the preprocessing on the charging pile data to obtain standard charging pile data; establishing a charging pile usage optimization model; and solving the maximizing charging pile utilization rate using the charging pile usage optimization model;

[0160] The load balancing includes: defining a load balancing index; obtaining relevant load data including load data and resource data; performing the preprocessing on the relevant load data to obtain standard relevant load data; establishing a load balancing optimization model; and solving the load balancing using the load balancing optimization model.

[0161] The minimization of charging pile loss includes: defining a charging pile life loss index; collecting relevant equipment life loss data including: charging pile life loss rate, failure frequency and maintenance cost; preprocessing the relevant equipment life loss data to obtain standard relevant equipment life loss data; establishing a charging pile loss optimization model; and solving the minimization of charging pile loss using the charging pile loss optimization model;

[0162] Furthermore, the load optimization multi-objective function is generated according to the minimization of the total power cost, the maximization of the charging pile utilization, the balancing load and the minimization of the charging pile loss, and is expressed as:

[0163] minimize F(x1,x2,x3,x4)=(w1*f1(x1),-w2*f2(x2),w3*f3(x3),w4*f4(x4));

[0164] Wherein, minimize is represented as a minimization function; F() is represented as the load optimization multi-objective function; x1 is represented as a decision variable for minimizing electricity cost; x2 is represented as a decision variable for maximizing the utilization rate of charging piles; x3 is represented as a decision variable for balancing loads; x4 is represented as a decision variable for minimizing the loss of charging piles; f1() is represented as the total electricity cost optimization model; f2() is represented as a charging pile usage optimization model; f3() is represented as a balanced load optimization model; f4() is represented as a charging pile loss optimization model; w1, w2, w3 and w4 are represented as weight coefficients;

[0165] Further, historical charging pile usage data related to the charging station is obtained; wherein the historical charging pile usage data includes: historical charging pile charging records, historical user behavior data and historical charging pile status data;

[0166] Further, the historical charging pile usage data is preprocessed to obtain standard historical charging pile usage data;

[0167] Furthermore, the standard historical charging pile usage data is input into a load optimization experience learning model to obtain optimal load optimization experience;

[0168] Further, obtaining current charging pile usage data of the charging station;

[0169] Further, the current charging pile usage data is preprocessed and input into the load optimization model together with the optimal load optimization experience, and the current charging station load optimization strategy is output; wherein, the load optimization model includes: setting the load constraint conditions of the charging station includes: Among them, CUL i Represents the load of the charging pile i in use; MaxL i represents the maximum load of the charging pile i in use; M represents the number of charging piles in use; ToL represents the total load of the charging station; CUL i (t) represents the load of the charging pile i at time t; MaxD represents the maximum load change between time t and time t-1;

[0170] Set the number of ants, pheromone volatility coefficient and heuristic function weight;

[0171] Initializing pheromone concentration according to the optimal load optimization experience;

[0172] The probability of each ant selecting a charging pile load distribution path is calculated according to the pheromone concentration and the heuristic function; wherein, in the path, the parameters of the pre-processed current charging pile usage data are adjusted according to the load constraint condition;

[0173] Use the fitness function to calculate the load optimization multi-objective function value of the path selected by each ant;

[0174] Update pheromone concentrations based on the path chosen by each ant;

[0175] The information concentration of the global optimal path is updated; when the load optimization multi-objective function value of the global optimal path reaches the optimal value, the updating is stopped to obtain the current charging station load optimization strategy.

[0176] Furthermore, the current charging station load optimization strategy is evaluated.

[0177] Refer to Table 12, which gives the evaluation results of the optimized charging station load in area B;

[0178] Table 12B Area Charging Station Load Optimization Evaluation Embodiment three:

[0179] A charging station load optimization medium based on an ant colony algorithm, wherein the medium is a computer program executed by an electronic device, and when the program runs on the electronic device, the electronic device implements all the processes disclosed in the first embodiment of the present invention when executing.

[0180] 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 station load optimization method based on ant colony algorithm, characterized in that: include: Defining a multi-objective function for load optimization of a charging station; obtaining historical charging pile usage data related to the charging station; Preprocessing the historical charging pile usage data to obtain standard historical charging pile usage data; Input the standard historical charging pile usage data into the load optimization experience learning model to obtain the optimal load optimization experience; wherein, the load optimization experience learning model includes: using feature engineering technology to perform feature processing on the standard historical charging pile usage data to obtain historical charging pile usage feature data; integrating the historical charging pile usage feature data to obtain a historical charging pile usage feature set; inputting the historical charging pile usage feature set into the load optimization multi-objective function to perform target value fitting, and calculating the fitting error; using the fitting error to perform data correction on the data in the historical charging pile usage feature set, and calculating the correction loss; when the correction loss and the fitting error reach convergence, the optimal load optimization experience is obtained; Acquire current charging pile usage data of the charging station; input the current charging pile usage data into a load optimization model together with the optimal load optimization experience after preprocessing, and output the current charging station load optimization strategy; and evaluate the current charging station load optimization strategy.

2. The charging station load optimization method based on ant colony algorithm according to claim 1 is characterized in that: The load optimization multi-objective function includes: minimizing the total power cost, maximizing the utilization rate of charging piles, balancing the load and minimizing the loss of charging piles; Wherein, the minimization of the total power cost includes: defining the total power cost; collecting the power cost related data of the charging station including: power price data, load demand data and charging status data; performing the preprocessing on the power cost related data to obtain standard power cost related data; establishing a total power cost optimization model; solving the minimized total power cost using the total power cost optimization model; The maximizing charging pile utilization rate comprises: defining the charging pile utilization rate; collecting charging pile data including: charging pile usage data, charging pile available time and charging pile status data; performing the preprocessing on the charging pile data to obtain standard charging pile data; establishing a charging pile usage optimization model; and solving the maximizing charging pile utilization rate using the charging pile usage optimization model; The load balancing includes: defining a load balancing index; obtaining relevant load data including load data and resource data; performing the preprocessing on the relevant load data to obtain standard relevant load data; establishing a load balancing optimization model; and solving the load balancing using the load balancing optimization model. The minimization of charging pile loss includes: defining a charging pile life loss index; collecting relevant equipment life loss data including: charging pile life loss rate, failure frequency and maintenance cost; preprocessing the relevant equipment life loss data to obtain standard relevant equipment life loss data; establishing a charging pile loss optimization model; and solving the minimization of charging pile loss using the charging pile loss optimization model; The load optimization multi-objective function is generated according to the minimization of the total power cost, the maximization of the charging pile utilization, the balancing load and the minimization of the charging pile loss, and is expressed as: minimize F(x1,x2,x3,x4)=(w1*f1(x1),-w2*f2(x2),w3*f3(x3),w4*f4(x4)); Among them, minimize is represented by the minimization function; F() is represented by the load optimization multi-objective function; x1 is represented by the decision variable for minimizing the electricity cost; x2 is represented by the decision variable for maximizing the utilization rate of the charging pile; x3 is represented by the decision variable for balancing the load; x4 is represented by the decision variable for minimizing the loss of the charging pile; f1() is represented by the total electricity cost optimization model; f2() is represented by the charging pile usage optimization model; f3() is represented by the balanced load optimization model; f4() is represented by the charging pile loss optimization model; w1, w2, w3 and w4 are represented by weight coefficients.

3. The charging station load optimization method based on ant colony algorithm according to claim 1, characterized in that: The historical charging pile usage data includes: historical charging pile charging records, historical user behavior data and historical charging pile status data.

4. The charging station load optimization method based on ant colony algorithm according to claim 1, characterized in that: The preprocessing includes: performing data cleaning on the historical charging pile usage data to obtain first historical charging pile usage data; performing data conversion on the first historical charging pile usage data to obtain second historical charging pile usage data; performing standardization on the second historical charging pile data to obtain third historical charging pile data; and performing data integration on the third historical charging pile data to obtain the standard historical charging pile usage data.

5. The method for optimizing charging station load based on ant colony algorithm according to claim 1, characterized in that: The load optimization model includes: Setting load constraints of the charging station; Set the number of ants, pheromone volatility coefficient and heuristic function weight; Initializing pheromone concentration according to the optimal load optimization experience; The probability of each ant selecting a charging pile load distribution path is calculated according to the pheromone concentration and the heuristic function; wherein, in the path, the parameters of the pre-processed current charging pile usage data are adjusted according to the load constraint condition; Use the fitness function to calculate the load optimization multi-objective function value of the path selected by each ant; Update pheromone concentrations based on the path chosen by each ant; The information concentration of the global optimal path is updated; when the load optimization multi-objective function value of the global optimal path reaches the optimal value, the updating is stopped to obtain the current charging station load optimization strategy.

6. A charging station load optimization device based on ant colony algorithm, characterized in that: The device includes: a data acquisition module, a data storage module, a data preprocessing module, a load optimization experience learning module, a load optimization module, an evaluation module, an execution module and a feedback module; wherein the data acquisition module is used to collect the usage data of the charging piles in the charging station; the data storage module is used to store the historical usage data of the charging piles in the charging station; the data preprocessing module is used to preprocess the usage data of the charging piles; the load optimization experience learning module is used to learn the load optimization experience of the charging station; the load optimization module is used to load optimize the charging piles in use; the evaluation module is used to evaluate the optimization results of the load optimization module; the execution module is used to execute the charging pile load optimization strategy; the feedback module is used to feedback the optimization results to relevant personnel.

7. The charging station load optimization device based on ant colony algorithm according to claim 6, characterized in that: The load optimization experience learning module adopts a load optimization experience learning model, which specifically includes: using feature engineering technology to perform feature processing on standard historical charging pile usage data to obtain historical charging pile usage feature data; integrating the historical charging pile usage feature data to obtain a historical charging pile usage feature set; inputting the historical charging pile usage feature set into the load optimization multi-objective function to fit the target value, and calculating the fitting error; using the fitting error to correct the data in the historical charging pile usage feature set, and calculating the correction loss; when the correction loss and the fitting error converge, the optimal load optimization experience is obtained.

8. The charging station load optimization device based on ant colony algorithm according to claim 7, characterized in that: The load optimization multi-objective function includes: minimizing the total power cost, maximizing the utilization rate of charging piles, balancing the load and minimizing the loss of charging piles; Wherein, the minimization of the total power cost includes: defining the total power cost; collecting the power cost related data of the charging station including: power price data, load demand data and charging status data; performing the preprocessing on the power cost related data to obtain standard power cost related data; establishing a total power cost optimization model; solving the minimized total power cost using the total power cost optimization model; The maximizing charging pile utilization rate comprises: defining the charging pile utilization rate; collecting charging pile data including: charging pile usage data, charging pile available time and charging pile status data; performing the preprocessing on the charging pile data to obtain standard charging pile data; establishing a charging pile usage optimization model; and solving the maximizing charging pile utilization rate using the charging pile usage optimization model; The load balancing includes: defining a load balancing index; obtaining relevant load data including load data and resource data; performing the preprocessing on the relevant load data to obtain standard relevant load data; establishing a load balancing optimization model; and solving the load balancing using the load balancing optimization model. The minimization of charging pile loss includes: defining a charging pile life loss index; collecting relevant equipment life loss data including: charging pile life loss rate, failure frequency and maintenance cost; preprocessing the relevant equipment life loss data to obtain standard relevant equipment life loss data; establishing a charging pile loss optimization model; and solving the minimization of charging pile loss using the charging pile loss optimization model; The load optimization multi-objective function is generated according to the minimization of the total power cost, the maximization of the charging pile utilization, the balancing load and the minimization of the charging pile loss, and is expressed as: minimize F(x1,x2,x3,x4)=(w1*f1(x1),-w2*f2(x2),w3*f3(x3),w4*f4(x4)); Among them, minimize is represented by the minimization function; F() is represented by the load optimization multi-objective function; x1 is represented by the decision variable for minimizing the electricity cost; x2 is represented by the decision variable for maximizing the utilization rate of the charging pile; x3 is represented by the decision variable for balancing the load; x4 is represented by the decision variable for minimizing the loss of the charging pile; f1() is represented by the total electricity cost optimization model; f2() is represented by the charging pile usage optimization model; f3() is represented by the balanced load optimization model; f4() is represented by the charging pile loss optimization model; w1, w2, w3 and w4 are represented by weight coefficients.

9. The charging station load optimization device based on ant colony algorithm according to claim 6, characterized in that: The load optimization module adopts a load optimization model, which specifically includes: setting load constraints of the charging station; Set the number of ants, pheromone volatility coefficient and heuristic function weight; Initialize pheromone concentration based on optimal load optimization experience; The probability of each ant selecting a charging pile load distribution path is calculated according to the pheromone concentration and the heuristic function; wherein, in the path, the parameters of the pre-processed current charging pile usage data are adjusted according to the load constraint condition; The fitness function is used to calculate the load optimization multi-objective function value of the path selected by each ant; Update pheromone concentrations based on the path chosen by each ant; The information concentration of the global optimal path is updated; when the load optimization multi-objective function value of the global optimal path reaches the optimal value, the updating is stopped to obtain the current charging station load optimization strategy.

10. A charging station load optimization medium based on ant colony algorithm, characterized in that: The medium is a computer program executed by an electronic device, and when the program is run on the electronic device, the electronic device executes the steps of the method according to any one of claims 1 to 5.

Images