A novel charging station energy prediction method and system based on a timing large model

Abstract

The present application relates to the technical field of charging station prediction, in particular to a new charging station energy prediction method and system based on time sequence large model, comprising: collecting the energy information of the charging station; combining the energy information in the form of time sequence to construct an energy time sequence model related to the energy consumption of the charging station; using the energy time sequence model to obtain the actual demand time of the charging station in each power consumption period, and determining the representation relationship between charging demand and charging load according to the actual demand time; verifying the power consumption scenario of the energy time sequence model, and setting the scenario state index related to the power consumption scenario; verifying the distribution position of the power consumption equipment and the idle equipment under different power consumption periods under the scenario state index, and setting the power consumption distribution preference index; checking the difference amount of adjacent time power consumption periods in the energy time sequence model based on the power consumption distribution preference index, and screening to obtain the energy prediction strategy; the accuracy of energy prediction and the efficiency of resource allocation are realized.

Description

Technical Field

[0001] This invention relates to the field of charging station prediction technology, specifically a novel method and system for predicting energy consumption at charging stations based on a time-series large model. Background Technology

[0002] With the increasing prevalence of electric vehicles (EVs), energy management at charging stations has become increasingly important. Traditional energy consumption forecasting methods, typically based on historical data and simple statistical models, struggle to cope with complex dynamic changes and varied user behaviors. Modern large-scale time series models (such as LSTM and Transformer) can capture long-term dependencies and nonlinear patterns, enabling more accurate energy consumption forecasting.

[0003] For example, Chinese Patent Publication No. CN117856257A discloses a method, device, equipment, and medium for predicting the electricity load of charging stations, which can be applied to the field of electric vehicle and power grid interaction technology. The method includes: obtaining questionnaire survey results on the charging decision intentions of multiple target objects in a target area; based on a hybrid selection model, performing data fitting on the questionnaire survey results under multiple predetermined charging scenarios to obtain the correlation between attribute information and charging preference information; responding to the received charging demand of objects arriving at the target charging station within a target time period, obtaining the attribute information of the objects to be charged; based on the correlation, and according to the attribute information of the objects to be charged and the scenario information of the target charging station, obtaining the charging probability of the objects to be charged within the target charging station; and based on the charging probabilities and charging demands of multiple objects to be charged, obtaining the electricity load prediction information of the target charging station during the target time period.

[0004] Existing technologies tend to focus on discovering the charging information of objects to be charged, but neglect the distribution of these objects and how they are distributed in the charging decision-making scenario. This results in the identification of charging associations only emphasizing whether charging has occurred, without describing the differences in the distribution and location of charging, leading to significant deviations in the final charging prediction. Summary of the Invention

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a novel energy consumption prediction method for charging stations based on a time-series large model, including: S1, collecting energy consumption information of charging stations, including power consumption value of charging stations, power consumption period, power consumption scenario, power consumption equipment and idle equipment under each power consumption period.

[0006] S2 combines energy consumption information in a time series format to construct an energy consumption time series model related to the energy consumption of charging stations.

[0007] S3 uses an energy consumption time series model to obtain the actual demand time of charging stations in each power consumption period, and determines the characterization relationship between charging demand and charging load based on the actual demand time.

[0008] S4. Verify the electricity consumption scenarios in the energy consumption time series model, analyze the representation relationship between electricity consumption scenarios and charging demand, and set scenario status indicators related to electricity consumption scenarios.

[0009] S5, verify the distribution of power-consuming and idle devices under the scenario status indicators, and set power distribution preference indicators according to the distribution location of power-consuming and idle devices under different power consumption periods.

[0010] S6. Based on the electricity consumption distribution preference index, the difference in electricity consumption between adjacent time periods in the energy consumption time series model is checked, and the energy consumption prediction strategy is selected.

[0011] A novel energy consumption prediction system for charging stations based on a time-series large model includes: a data acquisition module for collecting energy consumption information from charging stations, including power consumption values, power consumption periods, power consumption scenarios, and the electrical equipment and idle equipment during each power consumption period.

[0012] The energy consumption characterization relationship building module is used to build characterization relationships between energy consumption information and form an energy consumption time series model.

[0013] The charging demand and load analysis module is used to analyze the relationship between charging demand and charging load to determine the actual demand time.

[0014] The electricity consumption scenario analysis module is used to analyze the relationship between electricity consumption scenarios and charging demand, and to set scenario status indicators.

[0015] The electricity distribution preference calculation module is used to calculate electricity distribution preference indicators, reflecting users' regional and time-of-day preferences.

[0016] The strategy selection module is used to identify the differences in electricity consumption between adjacent time periods in the energy consumption time series model and select the optimal energy consumption prediction strategy.

[0017] The beneficial effects of this invention are as follows: First, by constructing an energy consumption time series model and analyzing the relationship between charging demand and charging load, this invention can more accurately predict the future energy consumption of charging stations, which helps charging stations to better plan and manage resources and improve operational efficiency.

[0018] Second, this invention, through in-depth analysis of electricity consumption scenarios and user distribution preferences, enables a more accurate understanding of user behavior and needs. This helps charging stations provide more personalized services and improve the user experience.

[0019] Third, this invention, through real-time monitoring and analysis of energy consumption, can promptly identify and resolve issues such as uneven resource allocation. This helps charging stations utilize resources more efficiently and reduce operating costs.

[0020] Fourth, by providing comprehensive energy consumption forecasting and strategy selection functions, this invention can provide strong support for decision-making at charging stations. This helps managers formulate operational strategies and development plans more scientifically and rationally. Attached Figure Description

[0021] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0022] Figure 1 This is a flowchart illustrating a novel energy consumption prediction method for charging stations based on a large time-series model.

[0023] Figure 2 This is a flowchart illustrating step S2 of a novel energy consumption prediction method for charging stations based on a large time-series model.

[0024] Figure 3 This is a flowchart illustrating step S22 of a novel energy consumption prediction method for charging stations based on a time-series large model.

[0025] Figure 4 This is a flowchart illustrating step S3 of a novel energy consumption prediction method for charging stations based on a time-series large model.

[0026] Figure 5 This is a flowchart illustrating step S6 of a novel energy consumption prediction method for charging stations based on a time-series large model.

[0027] Figure 6 This is a system framework diagram of a novel energy consumption prediction system for charging stations based on a large time-series model. Detailed Implementation

[0028] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. Where specific techniques or conditions are not specified in the embodiments, they shall be performed in accordance with the techniques or conditions described in the literature in the art or in accordance with the product manual.

[0029] See Figure 1 A novel energy consumption prediction method for charging stations based on a time-series large model includes: S1, collecting energy consumption information of charging stations, including power consumption value, power consumption period, power consumption scenario, power consumption equipment and idle equipment in each power consumption period.

[0030] S2 combines energy consumption information in a time series format to construct an energy consumption time series model related to the energy consumption of charging stations.

[0031] S3 uses an energy consumption time series model to obtain the actual demand time of charging stations in each power consumption period, and determines the characterization relationship between charging demand and charging load based on the actual demand time.

[0032] S4. Verify the electricity consumption scenarios in the energy consumption time series model, analyze the representation relationship between electricity consumption scenarios and charging demand, and set scenario status indicators related to electricity consumption scenarios.

[0033] S5, verify the distribution of power-consuming and idle devices under the scenario status indicators, and set power distribution preference indicators according to the distribution location of power-consuming and idle devices under different power consumption periods.

[0034] S6. Based on the electricity consumption distribution preference index, the difference in electricity consumption between adjacent time periods in the energy consumption time series model is checked, and the energy consumption prediction strategy is selected.

[0035] Electricity consumption refers to the total electricity used by a charging station within a specific time period, usually measured in kilowatt-hours (kWh). Understanding electricity consumption can help assess the overall energy consumption level of a charging station, identify high-energy-consumption periods, and formulate energy-saving strategies or adjust service fee policies accordingly.

[0036] Electricity consumption periods refer to the distribution of electricity consumption within different time intervals, which can be broken down to the hour or even shorter time intervals. Analyzing electricity consumption periods helps to identify peak and off-peak periods, thereby enabling better planning of electricity procurement, resource allocation, and maintenance work, while also providing users with more flexible charging options.

[0037] Electricity usage scenarios describe the operational status of charging stations under specific conditions, such as routine operation or special events, and how weather conditions affect user behavior. By understanding different electricity usage scenarios, it is possible to predict demand changes under specific conditions, prepare in advance, and ensure that service quality is not affected.

[0038] Electrical equipment refers to the number and type of charging piles or other facilities that are currently operating and consuming electricity. Idle equipment refers to the number of charging piles that are available but not currently connected to for charging. Monitoring the status of electrical equipment and idle equipment allows for real-time understanding of the utilization rate of charging stations, timely response to demand fluctuations, and optimization of resource allocation. It also helps identify potential problems, such as a large number of devices being idle during a certain period, which may be due to inconvenient location or insufficient publicity.

[0039] In one embodiment of the present invention, step S2 mainly involves combining energy consumption information in the form of a time series to construct an energy consumption time series model. When constructing the energy consumption time series model, in addition to preprocessing the energy consumption information and converting it into dimensionless vector data, it is also necessary to determine the correlation information in the energy consumption information and set the energy consumption time series model according to the correlation information, thereby achieving the effects of multi-energy complementarity and fixed-mobile fusion.

[0040] For example, energy consumption information can be divided into multiple types, and each type can be set based on the correlation information of energy consumption information at different time scales. Then, the collaborative work between the devices corresponding to the energy consumption information and the specific device configuration can be judged to verify whether the current charging station has resource feedback and load balancing. Finally, the content corresponding to these data will be used to build an energy consumption time series model to improve the prediction and allocation effect of energy consumption at the charging station.

[0041] The associated information includes correlations on a time scale, correlations between devices working together, correlations between user charging needs, and correlations with the external environment.

[0042] The correlation on a time scale can be represented as the changing trend of power consumption values ​​at different time periods, dates, and seasons; the correlation of collaborative work between devices can be represented as the specific changes in power consumption values ​​of powered and idle devices during different peak periods or emergency situations, at which point it is necessary to identify how powered and idle devices charge and discharge; the correlation of user charging needs can be represented as the correlation between user charging habits, locations, and charging times and power consumption periods; the correlation of the external environment can be represented as how powered and idle devices in the power consumption information are set up under different weather conditions.

[0043] like Figure 2 As shown, step S2 may also include the following implementation: S21, preprocessing the energy consumption information and dividing the preprocessed energy consumption information into multiple types.

[0044] S22, Obtain the correlation information of each type of energy consumption information at different time scales, and use the correlation information to construct the representation relationship between energy consumption information. The representation relationship between energy consumption information will include all the features of the correlation information to represent the specific situation of the current energy consumption information.

[0045] S23, in response to the representational relationship between energy consumption information, respectively obtains the temporal scale correlation, equipment collaborative correlation, user charging demand correlation and external environment correlation of energy consumption information.

[0046] S24. Based on the temporal scale correlation, equipment coordination correlation, user charging demand correlation, and external environment correlation of energy consumption information, the allocation probability of user charging within the charging station is obtained. An energy consumption time series model is then constructed based on the values ​​of this allocation probability at different time scales. This allocation probability represents the probability of user allocation within the charging station under the corresponding values ​​of the temporal scale correlation, equipment coordination correlation, user charging demand correlation, and external environment correlation of energy consumption information. It describes how users are allocated to charge within the charging station under the corresponding data. This allocation probability is combined with the feature information corresponding to the temporal scale correlation, equipment coordination correlation, user charging demand correlation, and external environment correlation to obtain values ​​at different time scales to form the energy consumption time series model.

[0047] like Figure 3 As shown, the implementation method of the correlation information of energy consumption information at different time scales in step S22 also includes: S221, extracting the trend value of electricity consumption at different electricity consumption periods from each type of energy consumption information.

[0048] S222, based on the trend value of power consumption during different power consumption periods, obtain the location of power-consuming equipment and idle equipment under the corresponding trend value.

[0049] S223, analyze the probability distribution of user charging and the trend values ​​of power consumption equipment and idle equipment under different power consumption periods to obtain the correlation information of energy consumption information at different time scales.

[0050] This explains how energy consumption information can be processed from data on electricity consumption values, electricity consumption periods, electricity consumption scenarios, and the electrical equipment and idle equipment during each electricity consumption period. This processing will generate associated information according to the specific circumstances of the energy consumption information. At the same time, when constructing the representational relationship between energy consumption information using the associated information, the representational relationship is mainly the content after extracting features from the data in the associated information. These data are represented as multiple strings, and these strings will be associated with the specific content in the associated information.

[0051] In step S23, when obtaining the time-scale correlation, device collaboration correlation, user charging demand correlation, and external environment correlation of energy consumption information, these correlations will represent the correlation values ​​between the corresponding data. For example, the Pearson correlation coefficient and Spearman rank correlation coefficient, which are algorithms for correlation analysis, are used to describe these correlations.

[0052] If the time-scale correlation is represented as the trend of electricity consumption over time, then the time points corresponding to the electricity consumption period are input, and the Pearson correlation coefficient and Spearman rank correlation coefficient are used to obtain the corresponding time-scale correlation value. The average value calculated by the Pearson correlation coefficient and the Spearman rank correlation coefficient is taken as the time-scale correlation at this time.

[0053] The processing methods for device collaboration correlation, user charging demand correlation, and external environment correlation are the same as those for time-scale correlation. The only difference is that device collaboration correlation uses the power consumption value of the device and the corresponding coordinates of the idle device and the device. Pearson correlation coefficient and Spearman rank correlation coefficient are used with the coordinate probability and the power consumption value of the device. In this case, the Spearman rank correlation coefficient is calculated according to the coordinate probability of the device and the power consumption value, respectively. The average value of the two Spearman rank correlation coefficients is then calculated with the Pearson correlation coefficient calculated with the coordinate probability and power consumption value of the device to set the device collaboration correlation.

[0054] The correlation between user charging demand and electricity consumption is calculated using the distribution probability of user charging power consumption and electricity usage time periods. The average value of the Pearson correlation coefficient and the Spearman rank correlation coefficient is used to set the correlation between user charging demand and electricity consumption.

[0055] External environmental correlation is calculated by statistically analyzing the distribution probabilities of powered and idle equipment under weather conditions. The correlation coefficient between these two distribution probabilities is calculated using the Pearson correlation coefficient and the Spearman rank correlation coefficient, ultimately yielding the external environmental correlation.

[0056] The allocation probability of user charging in subsequent calculations is the sum of the product of the values ​​of time scale correlation, device coordination correlation, user charging demand correlation and external environment correlation at the corresponding time scale and the user charging probability.

[0057] In one embodiment of the present invention, the characterization relationship between charging demand and charging load in step S3 is set by identifying the trend of power consumption value under actual demand time and the distribution probability of actual demand time.

[0058] The actual demand time can be determined by the increase in charging demand and charging load within a given time period. This increase is considered to be a surge in the number of users and a surge in charging demand during that time period, and this time is easily considered to be the time when users need to charge.

[0059] like Figure 4As shown, the implementation method in step S3 includes: S31, obtaining the trend value of power consumption value and the allocation probability of user charging under different power consumption periods of the energy consumption time series model, and calculating the actual charging demand time distribution of power consumption value in each power consumption period.

[0060] S32, compare the actual charging demand time distribution of the power consumption value in each power consumption period with the allocation probability of user charging, identify the probability distribution value under the actual charging demand time distribution, and use the probability distribution value under the actual charging demand time distribution as a representation of the relationship between charging demand and charging load.

[0061] When obtaining the actual charging demand time distribution, the trend value of power consumption is used as the priority value. The settings are based on the frequency of occurrence of the trend value of power consumption and the probability of user charging in each power consumption period. For example, the trend value of power consumption is used as a condition to identify the probability of user charging in each power consumption period. These corresponding probabilities and the trend value of power consumption are clustered to obtain multiple actual charging demand time distributions. The clustering method can use the k-means clustering algorithm, which clusters according to the different distribution probabilities and the trend value of power consumption in each power consumption period.

[0062] Then, the probability of the common occurrence of the power consumption value and the user charging allocation probability in the actual charging demand time distribution within each power consumption period is determined, and this common occurrence probability is used as the probability distribution value of the power consumption value in the actual charging demand time distribution within each power consumption period.

[0063] In one embodiment of the present invention, step S4 mainly verifies the power consumption scenario of the energy consumption time series model, identifies the representational relationship between the power consumption scenario and the charging demand according to the different power consumption scenarios, and finally sets the scenario state index.

[0064] Scenario state indicators are used to indicate whether the scenarios included in the current energy consumption time series model are related to charging demand, and these connections are represented using scenario state indicators. The representation of the relationship between electricity consumption scenarios and charging demand will focus on the user's charging situation under the electricity consumption scenario, such as using scenario matching degree, demand satisfaction rate, and supply-demand balance index to represent the representation of the relationship between electricity consumption scenarios and charging demand at this time; the representation of the relationship between electricity consumption scenarios and charging demand is represented at this time as the conditional probability values ​​of the corresponding values ​​of scenario matching degree, demand satisfaction rate, and supply-demand balance index.

[0065] Scenario matching degree can be represented as the ratio of the minimum value of power consumption and charging demand to the maximum value of power consumption and charging demand at a certain time; demand satisfaction rate is the proportion of the number of completed charging to the number corresponding to the charging demand; supply and demand balance index represents the corresponding situation of the difference between charging demand and power consumption.

[0066] Step S4 also includes: obtaining the scenario matching degree, demand satisfaction rate and supply-demand balance index corresponding to user charging and power consumption in the power consumption scenario, and calculating the scenario status index.

[0067] Scenario matching degree, demand satisfaction rate and supply-demand balance index will be statistically analyzed according to the same time dimension to obtain scenario status indicators.

[0068] For example, the scenario matching degree is calculated by obtaining the power consumption value and charging demand value under the power consumption scenario, and combining them with the corresponding time point in the energy consumption time series model.

[0069] Where IR represents scene matching degree, PD t CD represents the power consumption value at time point t. t Let t represent the charging demand value at time point t, where T represents the number of time points, and t ranges from 1 to T.

[0070] The demand satisfaction rate is the ratio of the number of charges completed at a given time point to the total number of charges, while the supply-demand balance index represents the average value of the difference between charging demand and electricity consumption.

[0071] The final scenario state index is expressed as follows: in, This represents the scenario status indicators, where DFR represents the demand fulfillment rate, SDBI represents the supply-demand balance index, and w IR The weighting coefficient w represents the scene matching degree. DFR w represents the weighting coefficient for the demand satisfaction rate. SDBI This represents the weighting coefficient of the supply and demand balance index; at this time, the weighting coefficients of scenario matching degree, demand satisfaction rate and supply and demand balance index will be set to 0.37, 0.42 and 0.21 respectively.

[0072] In one embodiment of the present invention, step S5 mainly verifies the distribution of power-consuming and idle devices under the scenario status index. At the same time, according to the different distribution of power-consuming and idle devices, the time corresponding to these power-consuming and idle devices is verified, and finally a power distribution preference index is set.

[0073] When setting the electricity consumption distribution preference index, first obtain the proportion of consuming devices and the proportion of idle devices during the corresponding electricity consumption period. Then, set a coefficient related to user charging based on the values ​​of the consuming device proportion and the idle device proportion at the corresponding location within the charging station. For example, if the charging station is considered to be composed of ten blocks, the proportion of consuming devices and the proportion of idle devices in each block are different. After combining the data from these blocks, the electricity consumption distribution can be displayed. Then, by integrating the data according to the time corresponding to the electricity consumption period, the electricity consumption distribution preference index can be obtained.

[0074] Therefore, the electricity distribution preference index is obtained by setting the user's regional preference and time preference for the current charging station based on the number of times the user charges during the electricity application period, comparing the regional preference and time preference with the proportion of power-consuming equipment and the proportion of idle equipment during the electricity application period, and calculating the electricity distribution preference index.

[0075] Wherein, DPI represents the electricity distribution preference index, and ADR... t IDR represents the proportion of electrical equipment at time point t. t LP represents the proportion of idle devices at time point t. t TP represents the index value of regional preference at time point t. t The value of the indicator representing the time period preference at time point t is α, which represents the adjustment coefficient. The adjustment coefficient is a constant, such as 1.667.

[0076] Regional preference is quantified using regional access frequency or the percentage of charging times within a specific region. This metric represents the proportion of times a user charges at charging stations in that region out of their total charging frequency. Based on the value of this metric, regional preference is categorized into several levels, such as high preference, moderate preference, low preference, or no preference.

[0077] High preference: The user charges a high percentage of times in this area, for example, more than 50%, indicating that the user has a strong dependence on or preference for charging stations in this area.

[0078] Moderate preference: Users charge their devices in this area at a moderate rate, for example, 20%-50%, indicating that users sometimes choose charging stations in this area, but it is not their first choice.

[0079] Low preference: Users charge their devices in this area at a low rate, for example, 5%-20%, indicating that users rarely choose charging stations in this area.

[0080] No preference: The percentage of times a user charges in this area is extremely low or zero, indicating that the user does not show a clear preference for charging stations in this area.

[0081] Time-based preferences are quantified using the percentage of charging times or the frequency of access within a specific time period. This metric represents the proportion of times a user charges within a particular time period relative to the total number of charges. Based on the value of this metric, time-based preferences are categorized into several time periods, such as peak-hour preferences, off-peak-hour preferences, low-peak-hour preferences, or no clear time-based preference.

[0082] Peak hour preference: Users charge their devices during peak hours, such as between 6 pm and 10 pm, indicating that they prefer to charge during periods of higher electricity demand.

[0083] Off-peak period preference: Users charge their phones during off-peak periods, such as from 10 a.m. to 4 p.m., which is a moderate percentage of the time, indicating that users charge their phones more frequently during these periods.

[0084] Off-peak period preference: Users charge more frequently during off-peak periods, such as from 0:00 to 6:00 AM, indicating that users may prefer to charge when electricity prices are lower or charging stations are less busy.

[0085] No obvious time-period preference: The proportion of times users charge at different times is relatively even, indicating that users do not have obvious time-period preferences.

[0086] Based on the proportions represented by these two preferences, the relative distribution of idle and power-consuming devices within a given time period can be calculated, thus revealing the main power-consuming areas and locations within the charging station.

[0087] In one embodiment of the present invention, step S6 involves checking the differences between adjacent time periods in the energy consumption time series model and filtering according to the differences to obtain an energy consumption prediction strategy.

[0088] In step S6, the difference between adjacent electricity consumption periods in the energy consumption time series model represents the difference in trend values ​​generated by the energy consumption time series model in these adjacent electricity consumption periods. That is, under the scenario where the electricity consumption distribution preference index appears, the trend value of the electricity consumption value and the differences between the corresponding multiple indicators are used to determine the energy consumption forecasting strategy that needs to be adopted.

[0089] For example, considering the relationship between charging demand and charging load, and the relationship between electricity consumption scenarios and charging demand, these two relationships are used as priority judgment methods. The differences between the index values ​​of these two relationships in adjacent electricity consumption periods are compared. Then, the differences between scenario status indicators and electricity distribution preference indicators are compared again. The results of these two separate comparisons are combined to select the final energy consumption forecasting strategy. This energy consumption forecasting strategy represents the potential increase in electricity consumption over a future period and how this electricity consumption will increase. The energy consumption forecasting strategy will pre-set multiple strategies based on the current charging station situation. At this time, it uses input data to find the content most likely to occur in the current charging station, thereby achieving the prediction of the charging station's own energy consumption.

[0090] like Figure 5 As shown, step S6 includes: S61, extracting the first index value corresponding to the representation relationship between charging demand and charging load and the second index value corresponding to the representation relationship between electricity consumption scenario and charging demand, calculating the difference between the first index value and the second index value in adjacent electricity consumption periods, and setting the difference judgment index.

[0091] The first indicator value is the probability distribution value under the actual charging demand time distribution described in step S3; the second indicator value is the conditional probability value of the corresponding values ​​of scenario matching degree, demand satisfaction rate and supply-demand balance index described in step S4.

[0092] The difference judgment index set at this time is set by the magnitude of the first index value and the second index value, and adjusted according to the different differences in adjacent electricity consumption periods.

[0093] Where DJI represents the difference judgment index, ΔFID represents the difference in the first indicator value, ΔSID represents the difference in the second indicator value, FID′ represents the standard value of the difference in the first indicator value, and SID′ represents the standard value of the difference in the second indicator value; the standard values ​​of the difference in the first and second indicator values ​​are set by extracting the corresponding average values ​​from historical data.

[0094] S62 identifies the differences between scenario status indicators and electricity distribution preference indicators in adjacent electricity consumption periods and sets status difference indicators.

[0095] The state difference index can be expressed as: SDI stands for State Difference Index. ΔDPI represents the difference in scenario status indicators, while ΔDPI represents the difference in electricity distribution preference indicators.

[0096] S63, compare the state difference index and the difference judgment index with the preset energy consumption strategy, and select the preset energy consumption strategy corresponding to the maximum similarity between the state difference index and the difference judgment index and the preset energy consumption strategy as the output of the energy consumption prediction strategy.

[0097] When comparing the state difference index and difference judgment index with the preset energy consumption strategy, the process mainly involves judging the two indicators based on their values ​​and corresponding data. The energy consumption strategy with the highest similarity to these two indicators is then identified and output, thus completing the implementation of the energy consumption prediction strategy. The preset energy consumption strategy is pre-set in the database. This data includes predictions for electricity consumption values, electricity consumption periods, electricity consumption scenarios, and the number of electrical devices and idle devices during each electricity consumption period. Finally, the overall energy consumption can be adjusted based on these predictions.

[0098] like Figure 6 As shown, this invention also provides a novel energy consumption prediction system for charging stations based on a time-series large model, comprising: a data acquisition module, an energy consumption characterization relationship construction module, a charging demand and load analysis module, an electricity consumption scenario analysis module, an electricity consumption distribution preference calculation module, and a strategy selection module; the output of the data acquisition module is connected to the energy consumption characterization relationship construction module, the output of the energy consumption characterization relationship construction module is connected to the charging demand and load analysis module, the output of the charging demand and load analysis module is connected to the electricity consumption scenario analysis module, the output of the electricity consumption scenario analysis module is connected to the electricity consumption distribution preference calculation module, and the output of the electricity consumption distribution preference calculation module is connected to the strategy selection module.

[0099] The data acquisition module is used to collect energy consumption information from charging stations, including power consumption values, power consumption periods, power consumption scenarios, and the electrical equipment and idle equipment during each power consumption period.

[0100] The energy consumption characterization relationship building module is used to build characterization relationships between energy consumption information and form an energy consumption time series model.

[0101] The charging demand and load analysis module is used to analyze the relationship between charging demand and charging load to determine the actual demand time.

[0102] The electricity consumption scenario analysis module is used to analyze the relationship between electricity consumption scenarios and charging demand, and to set scenario status indicators.

[0103] The electricity distribution preference calculation module is used to calculate electricity distribution preference indicators, reflecting users' regional and time-of-day preferences.

[0104] The strategy selection module is used to identify the differences in electricity consumption between adjacent time periods in the energy consumption time series model and select the optimal energy consumption prediction strategy.

[0105] The modules described above can perform the following functions: collect energy consumption data of charging stations in real time or periodically, perform preprocessing operations such as cleaning, filling missing values, and standardization on the collected data; divide the preprocessed data into multiple types, and extract the time scale correlation, equipment collaboration correlation, user charging demand correlation and external environment correlation of different types of data.

[0106] Extract trend values ​​of electricity consumption at different electricity consumption periods, obtain the positions of power-consuming and idle devices under the corresponding trend values, analyze the probability distribution of user charging and power-consuming and idle devices under the trend values ​​at different electricity consumption periods; construct the time scale correlation, device collaboration correlation, user charging demand correlation and external environment correlation among energy consumption information.

[0107] Calculate the actual charging demand time distribution of electricity consumption value in each electricity consumption period, compare the actual charging demand time distribution with the user charging allocation probability, and identify the probability distribution value; use the probability distribution value under the actual charging demand time distribution as a representation of the relationship between charging demand and charging load.

[0108] Verify the scenario matching degree, demand satisfaction rate, and supply-demand balance index corresponding to user charging and power consumption in different electricity consumption scenarios; calculate scenario status indicators to evaluate demand satisfaction and service quality in different electricity consumption scenarios.

[0109] Based on the number of times a user charges during different electricity consumption periods, the user's regional and time-period preferences for the current charging station are set; the regional and time-period preferences are compared with the proportion of powered equipment and the proportion of idle equipment to calculate the electricity distribution preference index.

[0110] Extract the first indicator value corresponding to the representation relationship between charging demand and charging load, and the second indicator value corresponding to the representation relationship between electricity consumption scenario and charging demand; calculate the difference between the first indicator value and the second indicator value in adjacent electricity consumption periods, and set the difference judgment index; identify the difference between the scenario state index and the electricity distribution preference index in adjacent electricity consumption periods, and set the state difference index; compare the state difference index and the difference judgment index with the preset energy consumption strategy, and select the preset energy consumption strategy with the highest similarity as the final output.

[0111] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention, which are still covered within the protection scope of the present invention.

Claims

1. A novel energy consumption prediction method for charging stations based on a large time-series model, characterized in that, include: S1 collects energy consumption information of charging stations, including power consumption value, power consumption period, power consumption scenario, power consumption equipment and idle equipment in each power consumption period; S2, combine energy consumption information in the form of time series to construct an energy consumption time series model related to the energy consumption of charging stations; S3 uses an energy consumption time series model to obtain the actual demand time of charging stations in each power consumption period, and determines the characterization relationship between charging demand and charging load based on the actual demand time. S4. Verify the electricity consumption scenarios in the energy consumption time series model, analyze the representation relationship between electricity consumption scenarios and charging demand, and set scenario status indicators related to electricity consumption scenarios. S5, verify the distribution of power-consuming and idle devices under the scenario status indicators, and set power distribution preference indicators according to the distribution location of power-consuming and idle devices under different power consumption periods; S6. Based on the electricity consumption distribution preference index, check the difference in electricity consumption periods between adjacent time periods in the energy consumption time series model, and select the energy consumption prediction strategy. Step S6 includes: S61, extracting the first index value corresponding to the representation relationship between charging demand and charging load and the second index value corresponding to the representation relationship between electricity consumption scenario and charging demand, calculating the difference between the first index value and the second index value in adjacent electricity consumption periods, and setting a difference judgment index; S62, identifying the difference between scenario state index and electricity consumption distribution preference index in adjacent electricity consumption periods, and setting a state difference index; S63, comparing the state difference index and the difference judgment index with a preset energy consumption strategy, and selecting the preset energy consumption strategy corresponding to the maximum similarity between the state difference index and the difference judgment index and the preset energy consumption strategy as the output of the energy consumption prediction strategy; The difference judgment index is expressed as follows: ； in, Indicators representing differences This indicates the difference in the value of the first indicator. This indicates the difference in the values ​​of the second indicator. The standard value representing the difference in the first indicator value. The standard value representing the difference in the second indicator value; The state difference index is expressed as follows: ； in, Indicators representing state differences This indicates the difference in scenario status indicators. This indicates the difference in electricity consumption distribution preference indicators.

2. The novel energy consumption prediction method for charging stations based on a large time-series model according to claim 1, characterized in that, Step S2 also includes: S21, Energy consumption information is preprocessed, and the preprocessed energy consumption information is divided into multiple types; S22, obtain the correlation information of each type of energy consumption information at different time scales, and use the correlation information to construct the representation relationship between energy consumption information; S23, in response to the representational relationship between energy consumption information, respectively obtain the time scale correlation, equipment coordination correlation, user charging demand correlation and external environment correlation of energy consumption information; S24. Based on the time scale correlation of energy consumption information, the equipment coordination correlation, the user charging demand correlation, and the external environment correlation, the allocation probability of user charging in the charging station is obtained, and an energy consumption time series model is formed according to the values ​​of the user charging allocation probability at different time scales.

3. The novel energy consumption prediction method for charging stations based on a large time-series model according to claim 2, characterized in that, Step S22 also includes: S221, extract the trend value of electricity consumption at different electricity consumption periods from each type of energy consumption information; S222, Based on the trend value of power consumption during different power consumption periods, obtain the position of power-consuming equipment and idle equipment under the corresponding trend value; S223, analyze the probability distribution of user charging and the trend values ​​of power consumption equipment and idle equipment under different power consumption periods to obtain the correlation information of energy consumption information at different time scales.

4. The novel energy consumption prediction method for charging stations based on a large time-series model according to claim 1, characterized in that, Step S3 includes: S31, obtain the trend value of power consumption value and the allocation probability of user charging under different power consumption periods of the energy consumption time series model, and calculate the actual charging demand time distribution of power consumption value in each power consumption period; S32, compare the actual charging demand time distribution of the power consumption value in each power consumption period with the allocation probability of user charging, identify the probability distribution value under the actual charging demand time distribution, and use the probability distribution value under the actual charging demand time distribution as a representation of the relationship between charging demand and charging load.

5. A novel energy consumption prediction method for charging stations based on a large time-series model according to claim 1, characterized in that, Step S4 also includes: obtaining the scenario matching degree, demand satisfaction rate and supply-demand balance index corresponding to user charging and power consumption in the power consumption scenario, and calculating the scenario status index; ； in, Indicates the status indicators of the scene. Indicates scene matching degree. Indicates the demand satisfaction rate. Indicates the supply and demand balance index. The weighting coefficients represent the scene matching degree. The weighting coefficients representing the demand satisfaction rate The weighting coefficients represent the supply and demand balance index.

6. The novel energy consumption prediction method for charging stations based on a large time-series model according to claim 1, characterized in that, The electricity distribution preference index is obtained by setting the user's regional and time preferences for the current charging station based on the number of times the user charges during the electricity application period. The regional and time preferences are then compared with the proportion of power-consuming equipment and the proportion of idle equipment during the electricity application period to calculate the electricity distribution preference index. ； in, Indicators representing preferences for electricity consumption distribution This represents the proportion of electrical equipment at time point t. This represents the proportion of idle devices at time point t. This represents the index value indicating regional preference at time point t. This represents the index value of the time period preference at time point t. Indicates the adjustment factor. This indicates the number of time points, with t ranging from 1 to T.

7. A novel energy consumption prediction system for charging stations based on a time-series large model, used to execute the steps in the novel energy consumption prediction method for charging stations based on a time-series large model as described in any one of claims 1-6, characterized in that, include: The data acquisition module is used to collect energy consumption information from charging stations, including power consumption values, power consumption periods, power consumption scenarios, and the electrical equipment and idle equipment during each power consumption period. The energy consumption characterization relationship construction module is used to construct the characterization relationship between energy consumption information and form an energy consumption time series model. The charging demand and load analysis module is used to analyze the relationship between charging demand and charging load to determine the actual demand time. The electricity consumption scenario analysis module is used to analyze the relationship between electricity consumption scenarios and charging demand, and to set scenario status indicators. The electricity consumption distribution preference calculation module is used to calculate electricity consumption distribution preference indicators, reflecting users' regional and time-of-day preferences. The strategy selection module is used to identify the differences in electricity consumption between adjacent time periods in the energy consumption time series model and select the optimal energy consumption prediction strategy.

Images