Electric vehicle charging station intelligent recommendation method and system based on power quality space-time optimization

Abstract

The invention discloses an electric vehicle charging station intelligent recommendation method and system based on electric energy quality space-time optimization. The recommendation method comprises the steps that power grid operation data, charging station and electric vehicle states, road network operation data, traffic state data and personalized data of all transformer areas are collected in real time; extracting electric energy quality state parameters of each transformer area, and respectively calculating real-time electric energy quality factors; inputting the real-time electric energy quality factor into the electric energy quality prediction model to obtain an electric energy quality factor prediction value of each transformer area; searching a local candidate sub-graph according to the current position and the driving path of the user, and adjusting the edge weight according to the predicted value of the power quality factor in the searching process; and constructing a fusion graph structure, and selecting an optimal charging station as a recommendation result according to the electric energy quality factor prediction value, the traffic cost and the user preference data. The method can improve the power quality, gives consideration to the real-time traffic state, reduces the passing cost of a user, and improves the user experience.

Description

Technical Field

[0001] This invention relates to the field of power system control technology, specifically to an intelligent recommendation method and system for electric vehicle charging stations based on spatiotemporal optimization of power quality. Background Technology

[0002] With the continuous increase in the number of electric vehicles, the corresponding charging demand is also increasing dramatically, which places higher demands on the stability of the power grid and the quality of power. Intelligent charging station recommendation determines the optimal charging station based on user needs and the status of existing charging stations, providing users with personalized recommendations. This improves user experience, optimizes grid load, and enhances the operational efficiency of charging stations.

[0003] Current technologies for intelligent charging station recommendations typically rely on a simple approach based on user travel needs and the availability of charging stations. This involves selecting the nearest available charging station along the user's route. However, this method fails to adequately consider the impact of charging load on power grid quality. A large number of electric vehicles charging simultaneously can lead to harmonic distortion, voltage fluctuations, and grid load imbalances. Therefore, traditional charging station recommendations based solely on user travel needs and station availability can negatively impact the lifespan of grid equipment and electric vehicle batteries. Furthermore, road network traffic conditions are constantly changing. Some road sections may experience severe congestion or temporary traffic control, causing inconvenience, while others may have high traffic efficiency. Simply selecting the nearest charging station based on the user's route can result in traffic inconvenience, higher travel costs, and longer waiting times. Conversely, relying solely on traffic conditions for charging station recommendations fails to address power quality issues. In addition, different electric vehicle users have different preferences. For example, some users may prefer charging speed over charging station distance, while others may prefer charging station distance over charging speed. Traditional intelligent charging station recommendation solutions cannot take into account the impact on power grid quality, real-time road network traffic status, and users' personalized charging needs.

[0004] Therefore, there is an urgent need for an intelligent recommendation method for electric vehicle charging stations that can take into account power quality optimization, real-time road network traffic conditions, and users' personalized charging preferences, in order to improve power quality and user experience while reducing users' travel costs. Summary of the Invention

[0005] The purpose of this invention is to address the technical problems existing in the prior art by proposing an intelligent recommendation method and system for electric vehicle charging stations based on spatiotemporal optimization of power quality. This method predicts the power quality factor of a transformer substation in real time for future moments, performs a local candidate subgraph search within the road network module based on the predicted power quality factor, and then combines power quality factor, travel costs, and users' personalized charging needs data to select the optimal charging station. This approach balances power quality optimization with real-time road network traffic conditions and users' personalized charging preferences, thereby improving power quality, reducing travel costs for users, enhancing user experience, and ensuring the stable operation of the power grid and charging network.

[0006] To achieve this objective, this invention discloses an intelligent recommendation method for electric vehicle charging stations based on spatiotemporal optimization of power quality, the steps of which are as follows: Real-time collection of power grid operation data, charging station status information, electric vehicle status information, road network operation data, traffic status data, and personalized charging demand data of electric vehicle users in a designated area; Based on the real-time collected power grid operation data, the power quality status parameters of each distribution area are extracted. The power quality status parameters include harmonic distortion rate parameters, voltage flicker parameters, voltage fluctuation parameters, and voltage transient event frequency parameters. The real-time power quality factor of each distribution area is calculated based on the power quality status parameters of each distribution area. The real-time power quality factor of each transformer area is input into a pre-trained power quality prediction model to obtain the predicted power quality factor of each transformer area. The power quality prediction model is established by training an ARMA model (Auto-Regression and Moving Average Model) using historical power quality factor data of each transformer area. Based on the current location and driving path of the electric vehicle user, a local candidate subgraph is searched in the road network graph model. The road network graph model is constructed based on the road network operation data of each substation area, with all road intersection locations, charging station locations, and the start and end points of designated important road segments as nodes and road segments as edges. The weight of each edge is configured according to the length of the road segment and the current traffic status data. During the search for local candidate subgraphs, the weight of each edge is adjusted according to the predicted power quality factor value of the substation where the charging station is located. A fusion graph structure is constructed based on the road network model and the feature vectors of each node. The feature vectors of the nodes include the predicted power quality factor, charging load rate, toll cost, and historical user preference charging demand data. Based on the predicted power quality factor of the area where the charging station is located, the toll cost from the current location of the electric vehicle to the charging station, and the personalized preference charging demand data of the current electric vehicle user, the optimal charging station is selected from the local candidate subgraphs as the recommendation result output based on the fusion graph structure.

[0007] As a further improvement of the present invention, the power grid operation data of the distribution area includes the distribution area voltage. Current in the transformer area The charging station status information includes the idle status of the charging piles. The load level of the charging pile The electric vehicle status information includes vehicle type, real-time battery level, real-time location coordinates, driving path, and charging power and charging strategy for different electric vehicles. The personalized charging demand data includes preferences for charging speed, acceptable range for charging costs, and preferred distance to charging stations. The road network operation data includes the location information of all road intersections within a specified area, the location information of charging stations, and information on passable roads. The traffic status data includes traffic flow, road speed, road congestion level, traffic light scheduling information, and traffic event information affecting road capacity. The traffic light scheduling information includes the cycle and remaining time of traffic lights at each intersection. The traffic event information includes information on traffic accidents, construction closures, and temporary traffic control events.

[0008] As a further improvement of the present invention, the harmonic distortion rate parameter is based on the total voltage harmonic distortion rate. and total harmonic distortion of current Calculated total harmonic distortion factor The total harmonic distortion factor is used to characterize the combined distortion state of voltage and current harmonics. The calculation expression is: , , , in, Indicates the harmonic order. Indicates the first The voltage amplitude of the second harmonic Indicates the amplitude of the fundamental voltage. The adjustment index is set according to the harmonic order. Indicates the amplitude of the fundamental current; The voltage flicker parameter is based on long-term flicker values. and short-time flicker value Calculated flicker factor The flicker factor is used to characterize the degree of flicker in a voltage signal. The calculation expression is: , , , , in, Different time constants The flicker visual perception weighting coefficient is below. Indicates the number of sampling time points; Indicates the first The voltage fluctuation sensing value at each sampling time. It is based on the time constant The coefficients are set to be used for further adjustment of different time constants. The weighted effect under the following conditions Represents time constant Quantity, It is a statistical short-time flicker value. The number of Indicates the first A short-time flash value , This represents the maximum tolerable threshold for flicker. Indicates the value of long-term flickering; The voltage fluctuation parameter is a voltage fluctuation factor calculated based on the voltage fluctuation amplitude. The voltage fluctuation factor The calculation expression is:

[0009]

[0010]

[0011]

[0012] in, This represents the average voltage fluctuation value. Indicates the width of the time window. , Representing voltage signals respectively , The effective value, To normalize the voltage fluctuation amplitude, To determine the normalized voltage fluctuation amplitude The mean value calculated from a sample selected from historical average voltage fluctuation data. This represents the maximum amplitude of the voltage fluctuation. , These are the mean and standard deviation of historical average voltage fluctuation data, respectively. The period of the voltage signal; The voltage transient event frequency parameter is a sag / surge factor calculated based on the statistical frequency of voltage sag events and voltage surge events. The frequency and depth of voltage transient events are used to characterize these events. The frequency of voltage sag events and voltage swell events are obtained using adaptive threshold statistics, where the voltage amplitude is below a dynamic threshold. If the duration exceeds a specified time window, it is determined to be a voltage sag event. When the voltage amplitude is higher than the dynamic threshold... If the duration exceeds a specified time window, it is determined to be a voltage spurt event. , The preset rated voltage amplitude threshold, Harmonic distortion rate The weighted average of the voltage amplitude influence factor and the load level of the transformer area, i.e. , This represents the current load rate of the transformer area. , , These are the weighting coefficients, This serves as a reference value for voltage fluctuations. The transient decrease / increase factor The calculation expression is: , in, To temporarily reduce the frequency of incidents, To temporarily increase the frequency of incidents, , These are the depth weights for voltage sag events and voltage swell events, respectively. express The maximum value.

[0013] As a further improvement of the present invention, the calculation expression for the real-time power quality factor of each transformer area is as follows: , in, For the first Real-time power quality factor of each transformer substation area The weights for the harmonic distortion rate parameter, The weights for voltage flicker parameters, The weights for voltage fluctuation parameters, The weights for the frequency parameter of voltage transient events. For the first Harmonic distortion rate parameters for each transformer area For the first Voltage flicker parameters for each transformer area For the first Voltage fluctuation parameters for each distribution area For the first Frequency parameters of voltage transient events in each transformer substation.

[0014] As a further improvement of the present invention, the power quality prediction model is established by training an ARMA model using historical power quality factor data of each transformer area, including: The historical power quality factor data of each transformer area is normalized and then arranged in chronological order to form a power quality factor time series. A sliding window is used to divide the power quality factor time series of each transformer area into multiple subsequences; The ARMA model is used as the prediction model, and an independent prediction model is established for each transformer area. Model, in which Let the order be the autoregressive order. This represents the order of the moving average. Model order. The determination is achieved using ACF (Autocorrelation Function) and PACF (Partial Autocorrelation Function) analysis. Preliminary identification is performed by calculating the ACF and PACF plots of the series: ACF measures the linear correlation between the time series and its corresponding lagged version, to preliminarily determine the order of the moving average component. PACF is used to assess the partial correlation between a sequence and a specific lag term after controlling for the effects of intermediate lags, and it helps to initially identify the order of the autoregressive component. .

[0015] Input each subsequence of each transformer area into... The model is trained, and the error between the predicted and actual values ​​is calculated during the training process. Adjustments are made based on the error results. Model order Alternatively, a new model can be selected until the model reaches the preset accuracy requirements, thus completing the model training and obtaining the trained power quality prediction model.

[0016] As a further improvement of the present invention, the step of searching for local candidate subgraphs in the road network graph model based on the current location and driving path of the electric vehicle user includes: The weight of each edge in the road network model is set according to the following formula:

[0017] in, Represents a node With nodes The corresponding edge of the road segment between them Weight of time, Represents a node With nodes The length of the road segment between them For nodes With nodes The section between Average traffic speed at any given time For nodes With nodes The section between Road congestion index at any given time. Represents a node With nodes The end point of the road section is at Traffic light waiting time at any time For nodes With nodes Unexpected event factors in the road sections between them , , To adjust the weights; Starting from the current location of the electric vehicle user, search the road network model for all nodes and their corresponding paths within the passable range of the target direction and within the maximum acceptable travel time threshold to form a local candidate subgraph. During the search for local candidate subgraphs, the weights of the corresponding edges of the current road segment are adjusted according to the predicted power quality factor of the transformer area using the following formula:

[0018] in, Represents a node With nodes The corresponding edge of the road segment between them The weights are adjusted over time. express Time Node With nodes The section between the two roads belongs to the same district. The predicted value of the power quality factor. This represents the preset linkage coefficient between the power grid and the road network.

[0019] As a further improvement of the present invention, the step of constructing the fusion graph structure based on the road network map model and the feature vectors of each node includes: The fusion graph structure is defined based on the road network map model. , This represents a node in a road network diagram model. Represents the edges in the road network diagram model. This represents the edge weight matrix, which is formed by the weights of each edge in the road network graph model. express Node feature vectors of all nodes at time 1 The resulting node feature matrix, where each node in the fused graph structure represents a charging station, and the corresponding node feature vector for each node... Includes multiple fusion features:

[0020] in Indicates the first The candidate charging stations belong to the following areas: Predicted power quality factor at time [time]. Indicates the first One candidate charging station Charging load rate at any given time This indicates the current electric vehicle user's position from the current location to the [number]th [location]. The toll cost of each candidate charging station. , This indicates the electric vehicle user's location from the current position to the next... The path to each candidate charging station Represents a node With nodes The corresponding edge of the road segment between them Weight of time, Indicates the first The candidate charging stations are located in the following areas: Electricity price at any time For the first One candidate charging station The recommended score at time point is calculated using the following expression: in, Represents a node With nodes The weights of the corresponding edges between road segments. This represents the maximum value of the passage cost. This indicates that the data obtained based on the personalized charging needs of current electric vehicle users is used to analyze the first... Preference ratings for each candidate charging station These are the weighting coefficients.

[0021] As a further improvement of the present invention, the step of constructing a fusion graph structure based on the road network graph model and the feature vectors of each node further includes: Node feature matrix As the initial input matrix, i.e. And based on the edge weight matrix Construct the time-normalized adjacency matrix: ; The first fusion graph structure The layer information transmission process is defined as follows: In the formula, normalized adjacency matrix , For the first Layer node embedding representation, For the current number The trainable weight matrix of the layer, For activation functions; A graph attention mechanism is used to assign differentiated weights to different neighbors when propagating information on edges, where the attention weights are... Indicates the first Layer nodes Receive from neighboring nodes The relative importance of information is defined as:

[0022] In the formula, This is a learnable attention vector; This represents a vector concatenation operation; For nodes The set of neighboring nodes, , Represents a node , The input feature vector, Representing neighboring nodes Its characteristics.

[0023] As a further improvement of the present invention, the step of selecting the optimal charging station from the local candidate subgraph based on the fusion graph structure as the recommended result output includes: A Markov decision process is established based on the fusion graph structure, where the current state... Based on the user's current location and path direction The set of embedding vectors of candidate charging station nodes, and the action space. Let represent the set of selectable charging stations, and let the policy function be . , indicating the state Select action The probability of The parameters of the policy network are set with the objective function of maximizing the expected cumulative reward from the current moment into the future. :

[0024] in, A discount factor representing future rewards; For the system in the first The immediate reward function obtained after a recommendation decision. In the policy function Cumulative reward for all possible trajectories Expected value; The instant reward function is defined as follows: in, These are the weighting coefficients. for Moment Action The corresponding predicted power quality factor value, for Moment Action The corresponding charging load rate, for Moment Action The corresponding toll cost value, for Moment Action The corresponding preference rating.

[0025] The present invention also discloses an intelligent recommendation system for electric vehicle charging stations based on spatiotemporal optimization of power quality, including a processor and a memory. The memory is used to store a computer program, and the processor is used to execute the computer program to perform the method described above.

[0026] Compared with existing technologies, the advantages of this invention are as follows: This invention calculates the power quality status parameters of each transformer substation in real time based on power grid operation data, and then calculates the power quality factor of each substation to assess its power quality status. Furthermore, it uses a power quality prediction model based on the ARMA model to predict the power quality factor of each substation at future times, thus predicting the changing trend of the power quality status of the substations. When it is necessary to recommend charging stations for current electric vehicle users, it first searches for local candidate subgraphs in the real-time road network map model based on the user's current location and driving path. During the search process, it adjusts the power quality factors of each substation in conjunction with the power quality factors of each road segment. The weighting of the edges not only improves recommendation efficiency and reduces search time, but also prioritizes charging stations in areas with better power quality. Finally, by combining the predicted power quality factor, the travel cost from the current location of the electric vehicle to the candidate charging station, and the personalized charging needs of the electric vehicle user, the optimal charging station is selected. This approach can balance power quality optimization, user travel costs, real-time road network traffic conditions, and personalized user charging needs. It can balance the grid load, take into account real-time traffic conditions, reduce user travel costs, improve user experience, and ensure the stable operation of the power grid and charging network. Attached Figure Description

[0027] The present invention will be described in more detail below based on embodiments and with reference to the accompanying drawings, wherein: Figure 1 This is a schematic diagram illustrating the implementation process of the intelligent recommendation method for electric vehicle charging stations based on spatiotemporal optimization of power quality in an embodiment of the present invention. Detailed Implementation

[0028] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments, but this does not limit the scope of protection of the present invention.

[0029] This invention collects real-time data on power grid operation, electric vehicle status, charging station status, road network and traffic conditions, and user preferences. First, it calculates the power quality status parameters for each transformer substation and obtains the power quality factor to assess and predict the power quality level of each substation. A prediction model based on the ARMA model can predict future power quality trends. When recommending charging stations to electric vehicle users, the system searches for local candidate subgraphs in the real-time road network map based on the user's current location and travel path. During the search, the system dynamically adjusts the road segment weights based on the power quality factor of each substation, thereby improving recommendation efficiency, shortening search time, and prioritizing charging stations in areas with better power quality. Finally, by comprehensively considering the predicted power quality, the user's travel cost to the candidate station, and personalized charging needs, the optimal charging station is selected from the candidate subgraphs for recommendation. This approach balances power quality optimization, user travel costs, real-time road network traffic conditions, and personalized charging needs, balancing power grid load, considering real-time traffic conditions, reducing user travel costs, improving user experience, and ensuring the stable operation of the power grid and charging network.

[0030] On the one hand, this invention introduces harmonic distortion, voltage fluctuation, flicker, and voltage transient event depth into the intelligent charging station recommendation process to predict power quality status. It can utilize the spatiotemporal characteristics of power quality to assess the power quality status of a distribution area in real time. During the search for candidate charging stations, the weights of opposite edges in the road network model are adjusted based on the power quality status. This allows for a comprehensive adjustment of the charging station search path based on real-time power quality, improving search efficiency and enabling users to avoid high-load or low-power-quality areas in real time, prioritizing charging stations in areas with better power quality. This avoids power quality problems such as harmonic distortion, voltage fluctuation, and grid load imbalance caused by electric vehicle charging behavior, thereby improving the power quality of the grid, reducing the impact on grid equipment and the lifespan of electric vehicle batteries, and ensuring grid operational stability.

[0031] On the other hand, based on the road network map model search combined with the predicted power quality factor, this invention further selects the optimal charging station by combining the predicted power quality factor, toll costs, and charging demand data. This enables the construction of a deeply integrated transportation-energy control method, which fully utilizes the intelligent deep integration of real-time traffic conditions and energy systems to dynamically determine the recommendation results. It comprehensively considers real-time power quality status, road network traffic conditions, and users' personalized charging needs to achieve a three-in-one collaborative recommendation optimization of "path selection - power distribution quality - intelligent guidance," thereby improving the resilience and stability of the power distribution network and enhancing the charging efficiency and user experience of electric vehicle users.

[0032] Figure 1This paper illustrates a detailed flowchart of an intelligent electric vehicle charging station recommendation method based on spatiotemporal optimization of power quality, according to an embodiment of the present invention. The steps include: Step S01. Collect in real time the power grid operation data of each transformer substation in the designated area, the charging station status information of each transformer substation, the electric vehicle status information, the road network operation data and traffic status data, as well as the personalized charging demand data of current electric vehicle users.

[0033] Specifically, the power grid operation data of the distribution area includes the distribution area voltage. Current in the transformer area The charging station status information includes the idle status of the charging piles. The load level of the charging pile Electric vehicle status information includes vehicle type, real-time battery level, real-time location coordinates, driving route, and charging power and charging control strategies for different electric vehicles. User personalized charging demand data includes preferences for charging speed, acceptable range of charging costs, and preferred distance from charging stations.

[0034] For example, regarding charging station status information, the real-time operating status (idle / busy rate) of each charging pile in the charging station can be obtained. Indicates free time. (Indicates busy) and the current charging power, then calculate the overall load level of the charging station according to the following formula. : (1) in, Indicates the number of charging stations The output power of the charging pile The total design capacity of the charging station, This refers to the number of charging piles within the charging station.

[0035] Regarding electric vehicle status information, during the electric vehicle user's driving process, the system acquires the electric vehicle's current location coordinates, driving path, real-time battery level, etc., and records the charging power of different types of electric vehicles and the charging control strategies of various brands, storing them in the user database. When intelligent recommendations are needed, the system can retrieve the real-time status information of the user's electric vehicle from the user database.

[0036] Based on users' personalized charging needs data, a charging app can provide users with questionnaires or preferences. Users can select their preferences for charging speed (e.g., fast charging preferred, normal charging acceptable), their acceptable range for charging costs (e.g., low price preferred, charging speed prioritized regardless of price), and their preferences for charging station distance (e.g., prioritizing nearby charging stations or having no distance requirement). The submitted selections are stored in a user database. When intelligent recommendations are needed, the user's preferred charging needs data can be retrieved from this database.

[0037] Road network operation data and traffic status data can be obtained through traffic open platforms, navigation system APIs, urban IoT terminals, and intelligent traffic control systems. Specifically, road network operation data includes the location information of all road intersections within a specified area, the location information of charging stations, and information on passable roads. Traffic status data includes traffic flow, road speed, road congestion level, traffic light scheduling information, and traffic event information affecting road capacity. Among these, road speed is the average speed of vehicles on a certain road segment per unit time, used to reflect vehicle traffic efficiency; road congestion level can be calculated using the ratio of actual traffic density to road design capacity, with a value close to 1 during congestion, corresponding to a decrease in traffic capacity; traffic light scheduling information includes the cycle and remaining time of traffic lights at each intersection, used to assess vehicle waiting time at intersections; traffic event information includes abnormal events that affect road traffic efficiency, such as traffic accidents, construction closures, and temporary traffic control.

[0038] Step S02. Extract the power quality status parameters of each transformer area based on the real-time collected power grid operation data. The power quality status parameters include harmonic distortion rate parameters, voltage flicker parameters, voltage fluctuation parameters, and voltage transient event frequency parameters. Calculate the power quality factor of each transformer area based on the power quality status parameters of each transformer area.

[0039] This embodiment extracts power quality status parameters such as harmonic distortion rate, voltage flicker, voltage fluctuation, and voltage transient event frequency parameters based on real-time collected power grid operation data, which can be used to assess the power quality status of each distribution area in real time.

[0040] In this embodiment, the harmonic distortion rate parameter is based on the total voltage harmonic distortion rate. and total harmonic distortion of current Calculated total harmonic distortion factor This is used to characterize the combined distortion state of voltage harmonics and current harmonics.

[0041] Specifically, the total harmonic distortion factor It can be calculated according to the following process: Real-time collected transformer area voltage Current in the transformer area Perform Fast Fourier Transform (FFT) on each signal to convert the time-domain signal to the frequency-domain signal. Assume the transformer substation voltage signal... The frequency domain representation is obtained after FFT. Current signal The frequency domain representation is obtained after FFT. Based on the frequency domain signal analysis, the content of each harmonic can be determined. The total voltage harmonic distortion rate can be calculated using the following formula: (2) in, Indicates the harmonic order. Indicates the first The voltage amplitude of the second harmonic Indicates the amplitude of the fundamental voltage. This is the adjustment index related to the harmonic order.

[0042] Calculate the total harmonic distortion of current. The expression can be represented as: (3) in, Indicates the harmonic order. Indicates the first The voltage amplitude of the second harmonic Indicates the amplitude of the fundamental current. This is the adjustment index related to the harmonic order.

[0043] Total harmonic distortion of voltage and total harmonic distortion of current Obtain the total harmonic distortion factor : (4) As shown in equation (4) above, the total harmonic distortion factor Capable of integrating total voltage harmonic distortion and total harmonic distortion of current The influence of these factors is used to comprehensively reflect the quality status of the total harmonic distortion of voltage and current in the power grid data.

[0044] In this embodiment, the voltage flicker parameter is based on the long-term flicker value. and short-time flicker value Calculated flicker factor This is used to characterize the flicker level of a voltage signal. Specifically, the flicker factor... The following calculation process can be used to obtain it: Real-time collected transformer substation voltage Envelope detection of the signal yields the rate of change of voltage fluctuation amplitude. Input it into the following sensing function to calculate the voltage fluctuation sensing value. : (5) in, It serves as a reference value for voltage fluctuations, used to standardize the perceived intensity.

[0045] right Using the sliding window integration method, different time constants were statistically analyzed. The following is a weighted coefficient for the flicker visual perception. : (6) In the formula: Indicates the number of sampling time points; Indicates the first Voltage fluctuation sensing value at each sampling time point.

[0046] Finally, calculate the short-time flicker value using the following formula. : (7) in, Different time constants The flicker visual sensitivity weighting coefficient is used to reflect the sensitivity of the human eye to light flicker caused by voltage fluctuations at different time scales. For example, since the human eye is more sensitive to rapid flicker, the flicker corresponding to rapidly changing voltage fluctuations in a short period of time is... Larger. It is related to the time constant The relevant coefficients are used to further adjust the weighting effect under different time constants. Represents time constant The quantity.

[0047] In this embodiment, the short-time flicker value is calculated according to equations (5) to (7). Introducing coefficients The weighting coefficients for the visual perception of flicker can be adjusted at different time scales. Flash value in a short time The contribution of the value, thereby increasing the short-time flicker value. The accuracy and reliability of the calculation.

[0048] Further based on multiple short-time flicker values Calculate long-term flicker value Long-term flicker value The calculation formula is: (8) in, It is a statistical short-time flicker value. The number of items, that is, the number calculated at certain time intervals within a specified time period. The number of values, Indicates the first A short-time flash value .

[0049] Finally, the short-time flicker value was considered. Long-term flicker value Obtain the flicker factor : (9) in, This is the maximum tolerable threshold for flicker, for example, it can be 4.0.

[0050] The flicker factor can be obtained through the above formula (9). The value is mapped to the range (0,1]. If the short-term flicker value of a certain station area is... Long-term flicker value The sum of these values ​​reaches or exceeds the maximum tolerable flicker threshold. The worst power quality indicates a severe voltage flicker problem in the area, significantly impacting lighting quality and user comfort; when the short-term flicker value... Long-term flicker value sum At that time, flicker factor This corresponds to the optimal power quality, thereby utilizing the flicker factor. It can effectively assess the flicker quality status of electrical energy.

[0051] In this embodiment, the voltage fluctuation parameter is the voltage fluctuation factor calculated based on the voltage fluctuation amplitude. Voltage fluctuation factor Specifically, the following process can be used to calculate it: The effective voltage value refers to the DC voltage value that generates the same amount of heat as the AC voltage across the same resistor within one cycle. This applies to the real-time acquired transformer substation voltage. Signal Calculate its effective value for: (10) in, Let be the period of the voltage signal. In practical calculations, the effective value of the voltage within the window can be approximated by applying a sliding window processing technique to the voltage time series data. .

[0052] Then calculate the average voltage fluctuation value. : (11) Among them, the adjustment coefficient The width of the time window. , Representing voltage signals respectively , The effective value. By setting a reasonable time window, the voltage variation within a certain period can be accurately reflected. The average voltage fluctuation value is calculated according to the formula above. It can reflect the relativity of voltage fluctuations.

[0053] Then normalize the voltage fluctuation amplitude according to the following formula: (12) in, , The mean and standard deviation of historical average voltage fluctuation data are used to standardize the fluctuation amplitude.

[0054] This embodiment, based on normalization, considers the normalized voltage fluctuation amplitude. The mean value is obtained by selecting samples from historical average voltage fluctuation data, removing outliers, and recalculating. To further improve calculation accuracy. For example, it is possible to eliminate those that satisfy the following conditions. The extreme fluctuation values ​​were sampled, and only samples within the normal range were retained for recalculating the mean. .

[0055] Finally, the voltage fluctuation factor was calculated. : (13) in, This represents the maximum amplitude of the voltage fluctuation. According to equation (13), it can be... The value is mapped to the range (0,1].

[0056] In this embodiment, the voltage fluctuation factor is calculated according to the above formulas (10) to (13). First, calculate the average voltage fluctuation value. This can provide a physical quantification of the degree of voltage disturbance, and then utilize the normalized voltage fluctuation amplitude. Outliers are removed, and normalization eliminates scale differences, ensuring uniform comparability of fluctuation amplitudes across different transformer areas or time periods, thus improving the final voltage fluctuation factor. It can stably reflect the power quality status of the distribution area.

[0057] In this embodiment, the voltage transient event frequency parameter is the voltage sag / surge factor calculated based on the statistical frequency of voltage sag events and voltage surge events. The frequency and depth of voltage transient events are used to characterize these events. The frequency of voltage sag events and voltage swell events are obtained using adaptive threshold statistics, where the voltage amplitude is below a dynamic threshold. If the duration exceeds a specified time window, it is determined to be a voltage sag event. At this time, the sag depth is recorded. When the voltage amplitude is higher than the dynamic threshold... If the duration exceeds the specified time window, it is determined to be a voltage spurt event, and the spurt magnitude is recorded.

[0058] In specific application embodiments, the dynamic threshold can be calculated according to the following formula. : (14) in, It is a weighted average of harmonic distortion rate, voltage amplitude influence factor, and transformer area load level to comprehensively reflect the power quality disturbance risk level. This is the preset rated voltage amplitude threshold. The calculation formula is as follows: (15) In the formula: This represents the current load rate of the transformer area. , , These are the weighting coefficients. By setting appropriate weighting coefficients, the identification threshold for voltage transient events can be dynamically adjusted under different operating scenarios, making the system more sensitive to high-risk areas and improving the accuracy and intelligence of event detection.

[0059] This embodiment combines the operating status of the transformer area, such as load level and harmonic distortion rate, and introduces an adaptive threshold model to dynamically adjust the judgment conditions for voltage sags and swells. This can fully consider the dynamic harmonic distortion and load level of the transformer area, thereby improving the recognition accuracy of voltage sag and swell events.

[0060] In a specific application embodiment, the transient decrease / increase factor The calculation expression can be represented as: (16) in, To temporarily reduce the frequency of incidents, To temporarily increase the frequency of incidents, , For the depth weights of voltage sag events and voltage swell events, express The maximum value. For example, suppose The typical range is 0 to 1000, so we take 1000. According to formula (16), we can... The value is mapped to the range (0,1].

[0061] After extracting the aforementioned power quality status parameters, this embodiment further calculates the power quality factor for different transformer substations by integrating these parameters. It can comprehensively consider the impact of factors such as harmonics, flicker, voltage sags and swells on power quality, and accurately reflect the power quality status of the distribution area.

[0062] In specific application examples, the power quality factor of each transformer area can be calculated using the following formula: (17) in, The weights for the harmonic distortion rate parameter, The weights for voltage flicker parameters, The weights for voltage fluctuation parameters, The weights for the frequency parameter of voltage transient events. For the first Harmonic distortion rate parameters for each transformer area For the first Voltage flicker parameters for each transformer area For the first Voltage fluctuation parameters for each distribution area For the first Frequency parameters of voltage transient events in each transformer substation.

[0063] Step S03. Input the power quality factor of each transformer area into the pre-trained power quality prediction model to predict the power quality factor of each transformer area at future times. The power quality prediction model is established by training the ARMA model using the historical power quality factor data of each transformer area.

[0064] This embodiment pre-trains a power quality prediction model based on the ARMA model, and the steps include: Step S301. Normalize the historical power quality factor data of each transformer area and arrange them in chronological order to form a power quality factor time series.

[0065] Specifically, the voltage, current, harmonics, flicker, and other data collected in step S02 can be sorted by transformer substation number. Classify and generate independent time series datasets: (18) in The total number of stations. This represents the total number of time points.

[0066] Power quality factor for each transformer area Perform independent preprocessing: remove missing values ​​and outliers (such as...) Map the data to the [0,1] interval: (19) Step S302. Use a sliding window to divide the power quality factor time series of each transformer area into multiple subsequences.

[0067] Specifically, the normalized data for each transformer substation are arranged in chronological order to form an independent time series, where each transformer substation... The corresponding power quality factor sequence is The sliding window technique is used to divide the time series of each transformer area into multiple subsequences.

[0068] For example, by window length (e.g., 24-hour data points), step size Divide the time into segments (e.g., 1 hour), with each window containing the following data: To be used to predict the power quality factor at the next moment. .

[0069] Step S303. Use the ARMA (Autoregressive Moving Average) model as the prediction model, and determine the order of the ARMA model based on the autocorrelation function ACF and the partial autocorrelation function PACF. ,in Let the order be the autoregressive order. To determine the moving average order, an independent moving average is established for each substation. The model is used to predict the power quality factor of the transformer area at future times.

[0070] Specifically, model order The determination of the moving average component is achieved using ACF (autocorrelation function) and PACF (partial autocorrelation function) analysis. Preliminary identification is performed by calculating the ACF and PACF plots of the series. The ACF measures the linear correlation between the time series and its corresponding lagged version, thus providing an initial assessment of the order of the moving average component. PACF is used to assess the partial correlation between a sequence and a specific lag term after controlling for the effects of intermediate lags, and it helps to initially identify the order of the autoregressive component. .

[0071] Specifically, establish an independent [system / mechanism] for each [station / area]. The model can be represented as: (20) in For the first Each district The power quality factor at any given time. is a constant term representing the intercept of the model. The autoregressive coefficient represents the past... Each time point The effect of the value on the current value. The moving average coefficient represents the past... The impact of the error term at each time point on the current value. For time points The error term (white noise), assuming .

[0072] Step S304. Input each sub-sequence of each transformer area into the ARMA model for training. During the training process, calculate the error between the predicted value and the actual value, and adjust the model according to the error results. Model order Alternatively, a new model can be selected until the model reaches the preset accuracy requirements, thus completing the model training and obtaining a well-trained power quality prediction model.

[0073] Specifically, the mean squared error (MSE) can be used to calculate the error between the predicted and actual values. The calculation expression is as follows: (twenty one) in For the first Each district Predicted power quality factor at time [time]. This represents the number of samples.

[0074] Then, based on the error analysis results, the order of the ARMA model is adjusted. Alternatively, a different model can be selected to improve prediction accuracy.

[0075] After the model is trained, the trained ARMA model is used to predict the future of each transformer area. Power quality factor at time: (twenty two) in, For the first Time point of each station area The predicted value of the power quality factor. The autoregressive coefficient represents the past... Each time point The effect of the value on the current value. The moving average coefficient represents the past... The impact of the error term at each time point on the current value. For time points The error term follows the order of distributed.

[0076] This embodiment predicts the power quality factor of each transformer area at future times. Based on the predicted power quality factor It can analyze the trend of power quality changes in the future and determine whether power quality will decline or rise and the magnitude of the change.

[0077] Step S04. Search for local candidate subgraphs in the road network graph model based on the current location and driving route of the electric vehicle user. The road network graph model is constructed based on the road network operation data of each substation area, with all road intersection locations, charging station locations, and the start and end points of designated important road segments as nodes and road segments as connecting nodes as edges. The weight of each edge is configured according to the length of the road segment and the current traffic status data. During the search for local candidate subgraphs, the weight of each edge is adjusted according to the real-time power quality factor prediction value of the substation where the charging station is located.

[0078] This embodiment first constructs a road network map model based on electronic map data of the target area. The graphical model uses all road intersections, charging station locations, and the start and end points of important road sections as nodes in the graph. Each node corresponds to a unique geographical coordinate spatial location, reflecting the set of accessible spatial points. Road segments are abstracted as a set of directed edges. Each edge This indicates that vehicles can originate from nodes. Proceed along the actual road to the node The directionality of edges follows road traffic rules, such as one-way street settings, traffic light guidance sequence, and branch road priority. Each edge also has a dynamic attribute weight. This is used to characterize the current travel cost of that road segment, i.e., the path cost. The above road network diagram model... It can be represented as .

[0079] In this embodiment, the real-time operating status of roads within the target area is continuously collected and updated, and the road network map model is updated synchronously. For example, real-time road network operation data and traffic status data are collected, standardized, and synchronized with time series to form a traffic status matrix. This matrix and the road network diagram model Binding allows the attributes of each roadside to be perceived in real time and dynamically adjusted over time, giving the road network map structure dynamic perceptibility. By searching for charging stations based on the road network map model constructed above, dynamic perception of traffic factors and route guidance can be achieved.

[0080] In this embodiment, the road network graph model supports continuous weighting of edge attributes, and the weights of each edge in the graph are... The road network map model is calculated based on a combination of factors, including road segment length and traffic condition data (real-time vehicle speed, congestion level, traffic light delay, construction closure status, etc.). It can comprehensively describe the traffic accessibility and impedance characteristics of the road network in the current region at the current moment. This road network map model is used... It can estimate the shortest path and travel time between electric vehicle users and charging stations, and serve as the foundation layer for constructing a fusion graph of "road network-distribution network-charging network" to enable subsequent adjacency propagation calculation and recommendation strategy determination in graph neural networks.

[0081] In a specific application embodiment, a road network map model is completed. After constructing and collecting real-time traffic status data, the weight of each edge is set according to the following formula. To accurately describe the time, resistance, or energy consumption required for a vehicle to travel from one node to another: (twenty three) In the formula, Represents a node With nodes The corresponding edge of the road segment between them Weight of time, Represents a node With nodes The length of the road segment between them For nodes With nodes The section between The average speed at any given time, and the ratio of the two. Characterizes the ideal travel time; For nodes With nodes The section between The road congestion index at any given time is used to quantify the impact of traffic flow density on traffic efficiency. Represents a node With nodes The end point of the road section is at The traffic light waiting time at any given moment is used to reflect the delay effect caused by traffic light control. For nodes With nodes Unexpected event factors on the road section, such as traffic accidents, road closures, and construction, should be considered. The weight of an item is significantly increased (e.g., by a specified percentage) to dynamically adjust path risk; , , Adjustment weights are assigned to corresponding factors to support flexible adjustment of the influence of different factors on edge weights as needed.

[0082] This embodiment sets the weights of each edge in the manner described above and assigns these weights to the road network diagram model. By focusing on edge attributes and combining them with the edge update mechanism of the subsequent graph neural network, real-time feedback and adjustment of charging recommendation results based on road condition changes can be achieved. This can effectively improve the charging recommendation system's responsiveness to the actual road traffic environment, making route planning and recommendation results more accurate and reasonable. It is particularly suitable for intelligent navigation and load guidance in dynamic environments such as morning and evening rush hours, special events, or severe weather.

[0083] To further improve the system's computational efficiency and recommendation response speed, this embodiment focuses on constructing a road network map model. Based on the global road network map structure, a user travel path pruning mechanism is further adopted to generate local candidate sub-maps within the user's reachable area in real time. Specifically, based on the user's current location... Starting from this point, and combining the user's input of the target direction or navigation intent, the path planning module uses a graph model... Search for all accessible nodes within its drivable range. Based on the preset maximum acceptable passage time threshold Only charging station nodes that meet the following conditions will be retained: (twenty four) In the formula, This indicates the route from the user's current location to the candidate charging station, taking into account the current traffic conditions. The estimated travel time can be obtained by summing the edge weights of all edges in the path: (25) Finally, all charging station nodes that meet the time threshold requirements, along with their paths, are combined to form a local candidate subgraph for the current user request. Using this graph as the input region for graph neural network processing can significantly reduce computational complexity and avoid global traversal on complex graphs.

[0084] Furthermore, this embodiment searches for local candidate subgraphs. During the process, the corresponding edge weights are dynamically adjusted based on the predicted power quality factor of the transformer area, so that the route recommendation process is not only affected by traffic conditions, but also takes into account the operation quality of the distribution system of the transformer area along the route. This realizes the positive guidance and coordinated control of electric vehicle charging behavior on the power grid operation status, forming a linkage control mechanism between the road network and the distribution network.

[0085] Specifically, during the search process, if the power quality of the area traversed by the path is poor, i.e., the predicted power quality factor... The value is lower than the set safety threshold If so, then the edge weights corresponding to the road segments surrounding that area will be adjusted as a penalty. For example, suppose a certain road The connected area belongs to the transformer substation. The weights of its edges can be adjusted according to the following formula: (26) In the formula, Represents a node With nodes The corresponding edge of the road segment between them The weights are adjusted over time. express Time Node With nodes The section between the two roads belongs to the same district. The predicted value of the power quality factor. This is the preset linkage coefficient between the power grid and the road network. The larger the value, the more sensitive it is to the impact on power quality.

[0086] As shown in equation (26), when a certain transformer area Energy quality factor of k When the value is close to 1 (i.e., very good), the passage cost remains almost unchanged; while when... When the power quality approaches zero (i.e., extremely poor power quality), the cost of the travel path will increase significantly, thus enabling the recommendation system to automatically avoid the area and achieve the goal of "avoiding weak power areas and guiding balanced load distribution".

[0087] This embodiment, through the aforementioned control mechanism, can effectively guide charging behavior to areas with stronger grid capacity and better operational quality without affecting the efficiency of users' basic travel routes. Combined with the control method of deep integration of transportation and energy, it achieves coordinated recommendation optimization of "route selection - power distribution quality - intelligent guidance", effectively improving the resilience and stability of the power distribution network.

[0088] Step S05. Construct a fusion graph structure based on the road network map model and the feature vectors of each node. The feature vectors of the nodes include the predicted power quality factor, charging load rate, toll cost, and historical user preference charging demand data. Based on the predicted power quality factor of the transformer substation where the charging station is located, the toll cost from the electric vehicle's location to the charging station, and the personalized charging demand data of electric vehicle users, select the optimal charging station from the local candidate subgraphs as the recommendation result output.

[0089] In this embodiment, a fusion graph structure is constructed based on the road network graph model and the feature vectors of each node to integrate and express three types of heterogeneous information—distribution network operation status, charging station business load, and road network traffic conditions—in a unified graph structure, thereby constructing a dynamic heterogeneous graph that supports state updates and decision optimization.

[0090] Specifically, the fusion graph structure is defined based on the road network graph model. , This represents a node in a road network diagram model. Represents the edges in the road network diagram model. This represents the edge weight matrix, which is formed by the weights of each edge in the road network graph model. express Node feature vectors of all nodes at time 1 The resulting node feature matrix, where each node in the fused graph structure represents a charging station, and the corresponding node feature vector for each node... Includes multiple fusion features: ,in Indicates the first The candidate charging stations belong to the following areas: The predicted power quality factor at time [0,1] is given. Indicates the first One candidate charging station Charging load rate at any given time Indicates the first The candidate charging stations are located Electricity price at any time This indicates that the data on the personalized charging needs of current electric vehicle users is used to determine the charging situation in the first district. Preference ratings for each candidate charging station For the first One candidate charging station Recommended score at any time This represents a traffic accessibility indicator, specifically the distance a current electric vehicle user can travel from their current location to the next... The toll cost of each candidate charging station can be calculated using the following formula: (27) in, This indicates the distance from the current electric vehicle user's current location to the [number]th [location]. The path to each candidate charging station.

[0091] Further, the node feature matrix As the initial input matrix And according to the edge weights Constructing a time-normalized adjacency matrix: (28) Adjacency Matrix Used in information propagation and graph convolution computation in graph neural networks, enabling the system to identify differences in access impedance, impact range, and service capacity between different sites.

[0092] In this embodiment, a comprehensive service quality scoring function is further defined to calculate the recommendation score. Used to evaluate each charging station The overall recommendation value at any given time is calculated using the following expression: (29) in, Indicates the first One candidate charging station Recommended score at any time Represents a node With nodes The weights of the corresponding edges between road segments. This represents the maximum value of the passage cost. This indicates that the data obtained based on the personalized charging needs of current electric vehicle users is used to analyze the first... Preference ratings for each candidate charging station These are the weighting coefficients, and Let be the weighting coefficient, satisfying .

[0093] In this embodiment, the first step of the fusion graph structure is further defined. The layer information transmission process is as follows: (30) In the formula, normalized adjacency matrix , For the first Layer node embedding representation, The trainable weight matrix of the current layer, This is the activation function.

[0094] To enhance the model's expressive power under heterogeneous node attributes, this embodiment further employs a graph attention mechanism to assign differentiated weights to different neighbors when propagating information along edges, wherein the attention weights... Indicates the first Layer nodes Receive from neighboring nodes The relative importance of information is defined as: (31) In the formula, This is a learnable attention vector; This represents a vector concatenation operation; For nodes The set of neighboring nodes, , Represents a node , The input feature vector, nodes It is a node The neighboring nodes, Representing neighboring nodes Its characteristics.

[0095] In this embodiment, a fused graph structure is further used. Based on the node representations of its graph neural network output, a Markov decision process framework is established. By autonomously learning recommendation strategies through interaction with the environment, the adaptability to optimal recommendation paths under complex and multi-constraint conditions can be further improved.

[0096] Specifically, set the current state. Based on the user's current location and path direction The set of embedding vectors of candidate charging station nodes, and the action space. Let represent the set of selectable charging stations, and let the policy function be . , indicating the state Select action The probability of These are the parameters of the policy network. That is, the current state of the system is defined as: (32) In the formula Indicates the current electric vehicle user Location at any given moment Indicates the user's destination direction (such as going home, heading to the office area or business district, etc.); This represents the candidate charging station node embedding vector, which is the final output of the graph neural network. This represents the set of candidate charging station nodes in the current cropped subgraph. The state space integrates user behavioral intent, graph structure encoding, and node feature information, enabling a comprehensive characterization of the environmental awareness state in the current recommendation scenario.

[0097] The action space is defined as the set of all reachable candidate charging station nodes in the current subgraph: (33) Furthermore, the recommendation objective function is to maximize the expected cumulative reward from the current moment into the future. (34) in, This represents a discount factor for future rewards. In the policy function Cumulative reward for all possible trajectories Expected value; For the system in the first The immediate reward function obtained after the recommendation decision can be specifically defined as follows: (35) in, These are the weighting coefficients. for The reward value at any moment, for Moment Action The corresponding predicted power quality factor value, for Moment Action The corresponding charging load rate, for Moment Action The corresponding toll cost value, for Moment Action The corresponding recommendation score. Specifically, it is calculated according to formula (29), that is, the recommended score. On the one hand, it will be embedded into the graph structure as a node feature, participating in the graph attention propagation and information fusion process between adjacent charging stations. On the other hand, it is used to calculate the immediate reward function in the Markov decision process to reflect the immediate comprehensive benefits brought by choosing each charging station, thereby guiding the optimization of reinforcement learning strategies. This can not only improve recommendation accuracy, but also achieve a continuous connection from multi-factor evaluation to graph structure learning and dynamic decision optimization, ensuring the unity of recommendation strategies in terms of data fusion and rationality of behavioral decisions.

[0098] As shown in equation (32) above, the larger the reward function value, the higher the comprehensive value of the recommendation in the current state. The system will continuously improve the probability of selecting high-quality sites through the reinforcement learning strategy optimization process, thereby guiding the strategy network to evolve towards "grid friendly, traffic efficient, and user satisfied", and finally selecting the optimal recommendation scheme that can take into account power quality, real-time traffic status and user charging needs.

[0099] This embodiment is based on the current user state and the optimal policy function after training. The system selects the optimal target station from the set of candidate charging stations in a local candidate subgraph and outputs the recommendation result. The recommendation process is based on the user's current location. Real-time traffic status map Embedded Graph Nodes Together they constitute the current system state Policy network according to Output recommended actions That is, select the target site .

[0100] This embodiment also provides an intelligent recommendation system for electric vehicle charging stations based on spatiotemporal optimization of power quality, including a processor and a memory. The memory is used to store computer programs, and the processor is used to execute the computer programs to perform the methods described above.

[0101] This invention introduces power quality optimization, comprehensively considering real-time power quality status, road network traffic status, and users' personalized charging needs. It achieves a three-in-one collaborative recommendation optimization of "route selection, power distribution quality, and intelligent guidance," which can take into account power quality optimization, user travel costs, real-time road network traffic status, and users' personalized charging needs to achieve intelligent charging station recommendation. This significantly improves the stability of power grid operation and the user charging experience, and can be applied to application scenarios such as smart grids and electric vehicle charging networks that provide intelligent charging station recommendations for electric vehicle users.

[0102] Although the invention has been described with reference to preferred embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, the technical features mentioned in the various embodiments can be combined in any manner as long as there is no structural conflict. The invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for intelligent recommendation of electric vehicle charging stations based on spatiotemporal optimization of power quality, characterized by the following steps: include: Real-time collection of power grid operation data, charging station status information, electric vehicle status information, road network operation data, traffic status data, and personalized charging demand data of electric vehicle users in a designated area; Based on the real-time collected power grid operation data, the power quality status parameters of each distribution area are extracted. The power quality status parameters include harmonic distortion rate parameters, voltage flicker parameters, voltage fluctuation parameters, and voltage transient event frequency parameters. The real-time power quality factor of each distribution area is calculated based on the power quality status parameters of each distribution area. The real-time power quality factor of each transformer area is input into a pre-trained power quality prediction model to obtain the predicted power quality factor of each transformer area. The power quality prediction model is established by training an ARMA model using historical power quality factor data of each transformer area. Based on the current location and driving path of electric vehicle users, local candidate subgraphs are searched in the road network graph model. The road network graph model is constructed based on the road network operation data of each substation area, with all road intersection locations, charging station locations, and the start and end points of designated important road segments as nodes and road segments as connecting nodes as edges. The weight of each edge is configured according to the length of the road segment and the current traffic status data. During the search for local candidate subgraphs, the weight of each edge is adjusted according to the predicted power quality factor value of the substation where the charging station is located. A fusion graph structure is constructed based on the road network map model and the feature vectors of each node. The feature vectors of the nodes include the predicted power quality factor, charging load rate, toll cost, and historical user preference charging demand data. Based on the predicted power quality factor of the area where the charging station is located, the toll cost from the current location of the electric vehicle to the charging station, and the personalized preference charging demand data of the current electric vehicle user, the optimal charging station is selected from the local candidate subgraphs as the recommendation result output based on the fusion graph structure.

2. The intelligent recommendation method for electric vehicle charging stations based on spatiotemporal optimization of power quality according to claim 1, characterized in that, The power grid operation data for the transformer substation includes the substation voltage. Current in the transformer area The charging station status information includes the idle status of the charging piles. The load level of the charging pile The electric vehicle status information includes vehicle type, real-time battery level, real-time location coordinates, driving path, and charging power and charging strategy for different electric vehicles. The personalized charging demand data includes preferences for charging speed, acceptable range for charging costs, and preferred distance to charging stations. The road network operation data includes the location information of all road intersections within a specified area, the location information of charging stations, and information on passable roads. The traffic status data includes traffic flow, road speed, road congestion level, traffic light scheduling information, and traffic event information affecting road capacity. The traffic light scheduling information includes the cycle and remaining time of traffic lights at each intersection. The traffic event information includes information on traffic accidents, construction closures, and temporary traffic control events.

3. The intelligent recommendation method for electric vehicle charging stations based on spatiotemporal optimization of power quality according to claim 1, characterized in that, The harmonic distortion rate parameter is based on the total harmonic distortion rate of the voltage. and total harmonic distortion of current Calculated total harmonic distortion factor The total harmonic distortion factor is used to characterize the combined distortion state of voltage and current harmonics. The calculation expression is: ， ， ， in, Indicates the harmonic order. Indicates the first The voltage amplitude of the second harmonic Indicates the amplitude of the fundamental voltage. The adjustment index is set according to the harmonic order. Indicates the amplitude of the fundamental current; The voltage flicker parameter is based on long-term flicker values. and short-time flicker value Calculated flicker factor The flicker factor is used to characterize the degree of flicker in a voltage signal. The calculation expression is: ， ， ， ， in, Different time constants The flicker visual perception weighting coefficient is below. Indicates the number of sampling time points; Indicates the first The voltage fluctuation sensing value at each sampling time. It is based on the time constant The coefficients are set to be used for further adjustment of different time constants. The weighted effect under the following conditions Represents time constant Quantity, It is a statistical short-time flicker value. The number of Indicates the first A short-time flash value , This represents the maximum tolerable threshold for flicker. Indicates the value of long-term flickering; The voltage fluctuation parameter is a voltage fluctuation factor calculated based on the voltage fluctuation amplitude. The voltage fluctuation factor The calculation expression is: ， ， ， ， in, This represents the average voltage fluctuation value. Indicates the width of the time window. , Representing voltage signals respectively , The effective value, To normalize the voltage fluctuation amplitude, To determine the normalized voltage fluctuation amplitude The mean value calculated from a sample selected from historical average voltage fluctuation data. This represents the maximum amplitude of the voltage fluctuation. , These are the mean and standard deviation of historical average voltage fluctuation data, respectively. The period of the voltage signal; The voltage transient event frequency parameter is a sag / surge factor calculated based on the statistical frequency of voltage sag events and voltage surge events. The frequency and depth of voltage transient events are used to characterize these events. The frequency of voltage sag events and voltage swell events are obtained using adaptive threshold statistics, where the voltage amplitude is below a dynamic threshold. If the duration exceeds a specified time window, it is determined to be a voltage sag event. When the voltage amplitude is higher than the dynamic threshold... If the duration exceeds a specified time window, it is determined to be a voltage spurt event. , The preset rated voltage amplitude threshold, Harmonic distortion rate The weighted average of the voltage amplitude influence factor and the load level of the transformer area, i.e. , This represents the current load rate of the transformer area. , , These are the weighting coefficients, This serves as a reference value for voltage fluctuations. The transient decrease / increase factor The calculation expression is: ， in, To temporarily reduce the frequency of incidents, To temporarily increase the frequency of incidents, , These are the depth weights for voltage sag events and voltage swell events, respectively. express The maximum value.

4. The intelligent recommendation method for electric vehicle charging stations based on spatiotemporal optimization of power quality according to claim 1, characterized in that, The calculation expression for the real-time power quality factor of each transformer area is as follows: ， in, For the first Real-time power quality factor of each transformer substation area The weights for the harmonic distortion rate parameter, The weights for voltage flicker parameters, The weights for voltage fluctuation parameters, The weights for the frequency parameter of voltage transient events. For the first Harmonic distortion rate parameters for each transformer area For the first Voltage flicker parameters for each transformer area For the first Voltage fluctuation parameters for each distribution area For the first Frequency parameters of voltage transient events in each transformer substation.

5. The intelligent recommendation method for electric vehicle charging stations based on spatiotemporal optimization of power quality according to claim 1, characterized in that, The power quality prediction model is established by training an ARMA model using historical power quality factor data from each transformer substation, and includes: The historical power quality factor data of each transformer area is normalized and then arranged in chronological order to form a power quality factor time series. A sliding window is used to divide the power quality factor time series of each transformer area into multiple subsequences; The ARMA model is used as the prediction model, and an independent prediction model is established for each transformer area. Model, in which Let the order be the autoregressive order. The moving average order and the model order are given. The determination was made using ACF and PACF analysis. Preliminary identification was performed by calculating the ACF and PACF plots of the sequences. The order of the moving average component was initially determined using ACF. n PACF was used to initially identify the order of the autoregressive components. ; Input each subsequence of each transformer area into... The model is trained, and the error between the predicted and actual values ​​is calculated during the training process. Adjustments are made based on the error results. Model order Alternatively, a new model can be selected until the model reaches the preset accuracy requirements, thus completing the model training and obtaining the trained power quality prediction model.

6. The intelligent recommendation method for electric vehicle charging stations based on spatiotemporal optimization of power quality according to any one of claims 1 to 5, characterized in that, The step of searching for local candidate subgraphs in the road network graph model based on the current location and driving path of the electric vehicle user includes: The weight of each edge in the road network model is set according to the following formula: ， in, Represents a node With nodes The corresponding edge of the road segment between them Weight of time, Represents a node With nodes The length of the road segment between them For nodes With nodes The section between Average traffic speed at any given time For nodes With nodes The section between Road congestion index at any given time. Represents a node With nodes The end point of the road section is at Traffic light waiting time at any time For nodes With nodes Unexpected event factors in the road sections between them , , To adjust the weights; Starting from the current location of the electric vehicle user, search the road network model for all nodes and their corresponding paths within the passable range of the target direction and within the maximum acceptable travel time threshold to form a local candidate subgraph. During the search for local candidate subgraphs, the weights of the corresponding edges of the current road segment are adjusted according to the predicted power quality factor of the transformer area using the following formula: ， in, Represents a node With nodes The corresponding edge of the road segment between them The weights are adjusted over time. express Time Node With nodes The section between the two roads belongs to the same district. The predicted value of the power quality factor. This represents the preset linkage coefficient between the power grid and the road network.

7. The intelligent recommendation method for electric vehicle charging stations based on spatiotemporal optimization of power quality according to any one of claims 1 to 5, characterized in that, The step of constructing the fusion graph structure based on the road network graph model and the feature vectors of each node includes: The fusion graph structure is defined based on the road network map model. , This represents a node in a road network diagram model. Represents the edges in the road network diagram model. This represents the edge weight matrix, which is formed by the weights of each edge in the road network graph model. express Node feature vectors of all nodes at time 1 The resulting node feature matrix, where each node in the fused graph structure represents a charging station, and the corresponding node feature vector for each node... Includes multiple fusion features: in Indicates the first The candidate charging stations belong to the following areas: Predicted power quality factor at time [time]. Indicates the first One candidate charging station Charging load rate at any given time This indicates the current electric vehicle user's position from the current location to the [number]th [location]. The toll cost of each candidate charging station. , This indicates the electric vehicle user's location from the current position to the next... The path to each candidate charging station Represents a node With nodes The corresponding edge of the road segment between them Weight of time, Indicates the first The candidate charging stations are located in the following areas: Electricity price at any time This indicates that the data obtained based on the personalized charging needs of current electric vehicle users is used to analyze the first... Preference ratings for each candidate charging station For the first One candidate charging station The recommended score at time point is calculated using the following expression: in, This represents the maximum value of the passage cost. These are the weighting coefficients.

8. The intelligent recommendation method for electric vehicle charging stations based on spatiotemporal optimization of power quality according to claim 7, characterized in that, The step of constructing the fused graph structure based on the road network graph model and the feature vectors of each node also includes: Node feature matrix As the initial input matrix, i.e. And based on the edge weight matrix Construct the time-normalized adjacency matrix: ; The first fusion graph structure The layer information transmission process is defined as follows: In the formula, normalized adjacency matrix , For the first Layer node embedding representation, For the current number The trainable weight matrix of the layer, For activation functions; A graph attention mechanism is used to assign differentiated weights to different neighbors when propagating information on edges, where the attention weights are... Indicates the first Layer nodes Receive from neighboring nodes The relative importance of information is defined as: ， In the formula, This is a learnable attention vector; This represents a vector concatenation operation; For nodes The set of neighboring nodes, , Represents a node , The input feature vector, Representing neighboring nodes Its characteristics.

9. The intelligent recommendation method for electric vehicle charging stations based on spatiotemporal optimization of power quality according to claim 8, characterized in that, The step of selecting the optimal charging station from the local candidate subgraphs based on the fusion graph structure as the recommendation result output includes: A Markov decision process is established based on the fusion graph structure, where the current state... Based on the user's current location and path direction The set of embedding vectors of candidate charging station nodes, and the action space. Let represent the set of selectable charging stations, and let the policy function be . , indicating the state Select action The probability of The parameters of the policy network are set with the objective function of maximizing the expected cumulative reward from the current moment into the future. : ， in, A discount factor representing future rewards; For the system in the first The immediate reward function obtained after a recommendation decision. In the policy function Cumulative reward for all possible trajectories Expected value; The instant reward function is defined as follows: ， in, These are the weighting coefficients. for Moment Action The corresponding predicted power quality factor value, for Moment Action The corresponding charging load rate, for Moment Action The corresponding toll cost value, for Moment Action The corresponding recommended score.

10. An intelligent recommendation system for electric vehicle charging stations based on spatiotemporal optimization of power quality, comprising a processor and a memory, wherein the memory is used to store computer programs, characterized in that, The processor is used to execute the computer program to perform the method as described in any one of claims 1 to 9.

Images