Electric heavy truck charging load prediction method, system, device and medium

Abstract

The present application relates to the technical field of power grid load forecasting, and more particularly to an electric heavy truck charging load forecasting method, system, device and medium, which comprises: obtaining operation characteristic data of electric heavy trucks; classifying the operation characteristic data by scene to generate a plurality of typical scenes of electric heavy truck application; calculating the scene-by-scene retention data of each typical scene by a scene-based structure distribution method; constructing a charging behavior probability model according to the electric heavy truck retention data of each typical scene; generating the charging load curve corresponding to each typical scene based on each charging behavior probability model, and performing time sequence fusion processing on each charging load curve to obtain a regional total charging load forecasting result. In this way, the technical problem that the prior art cannot balance the electric heavy truck charging load forecasting accuracy and scene adaptability is solved, reliable data support is provided for power grid planning and charging facility layout, and the forecasting accuracy of the electric heavy truck charging load is improved.

Description

Technical Field

[0001] This invention relates to the field of power grid load forecasting technology, and in particular to a method, system, device and medium for forecasting the charging load of electric heavy trucks. Background Technology

[0002] The electrification transformation of the transportation sector is accelerating. As a major contributor to road carbon emissions, the electrification process of heavy-duty trucks (referred to as "heavy trucks") poses new challenges to power grid planning and safe operation. The large-scale, high-power charging load of electric heavy trucks, especially at key nodes such as highway service areas and logistics parks, is prone to forming concentrated peak loads. Accurately predicting their charging demand has become a crucial link in supporting power grid upgrading and energy dispatch.

[0003] However, the inventors found that existing charging load prediction technologies often overlook the scenario-dependent nature of electric heavy-duty truck operation modes, leading to a disconnect between the prediction model and actual charging behavior. Specifically, prior art, such as prior art document 1 (application number 202510668468.6), discloses a charging station planning method based on vehicle charging demand, which predicts charging demand through multi-source data collection and fusion processing. However, this method has significant limitations when applied to electric heavy-duty trucks: it does not fully consider the essential differences between electric heavy-duty trucks and passenger cars in terms of battery capacity, charging power, and operating patterns, and it fails to distinguish the specificity of charging behavior under different transportation scenarios.

[0004] Specifically, the method in Comparison Document 1 tends to treat all types of electric vehicles as a homogeneous group for aggregate prediction, or to use a single model to uniformly describe the charging behavior of all heavy-duty trucks. This approach has the following drawbacks: (1) When attempting to improve the prediction range through macro statistics, the model cannot capture the unique peak charging time and intensity of each scenario because the scenario is not distinguished, resulting in the peak load being averaged and the prediction accuracy being insufficient.

[0005] (2) Existing methods lack the ability to dynamically depict scenario-based charging behavior and are difficult to adapt to the differences in charging patterns under different scenarios such as trunk transportation and port operations, resulting in large deviations in prediction results in local areas.

[0006] Therefore, due to the homogenization of models, existing technologies cannot simultaneously ensure the accuracy of electric heavy-duty truck charging load prediction and scenario adaptability. They are unable to accurately reconstruct load curves in complex and ever-changing operating environments and cannot meet the actual needs of lean power grid planning and precise investment in charging facilities. Summary of the Invention

[0007] In view of the above-mentioned shortcomings or disadvantages, the present invention provides a method, system, device and medium for predicting the charging load of electric heavy trucks, which can solve the technical problem that the existing technology cannot simultaneously take into account the accuracy of electric heavy truck charging load prediction and the adaptability of scenarios.

[0008] This invention provides a method for predicting the charging load of electric heavy-duty trucks, comprising: Obtain operational characteristic data of electric heavy-duty trucks.

[0009] Based on preset scenario segmentation rules, the operational characteristic data is classified into scenarios to generate multiple typical scenarios for electric heavy truck applications.

[0010] Based on the current regional heavy-duty truck market size and electrification progress forecasts, the vehicle ownership data for each typical scenario is calculated through a scenario-based structural allocation method.

[0011] A charging behavior probability model is constructed based on the electric heavy truck ownership data for each typical scenario. The charging behavior probability model is configured to describe the statistical regularity of charging behavior under the corresponding typical scenario.

[0012] Based on the probability models of various charging behaviors, charging load curves corresponding to each typical scenario are generated through random simulation. The charging load curves are then fused over time to obtain the prediction results of the total regional charging load.

[0013] According to a second aspect, the present invention provides an electric heavy-duty truck charging load prediction system, comprising: The operational characteristic data acquisition module is used to acquire operational characteristic data of electric heavy trucks.

[0014] The typical scenario generation module is used to classify operational feature data into scenarios based on preset scenario division rules, and will generate multiple typical scenarios for electric heavy truck applications.

[0015] The vehicle ownership data calculation module is used to calculate the vehicle ownership data for each typical scenario based on the current regional heavy truck market size and the predicted results of electrification progress through a scenario-based structure allocation method.

[0016] The charging probability model building module is used to build a charging behavior probability model based on the electric heavy truck ownership data of each typical scenario. The charging behavior probability model is configured to describe the statistical regularity of charging behavior under the corresponding typical scenario.

[0017] The charging load prediction module is used to generate charging load curves corresponding to each typical scenario based on the probability model of each charging behavior through random simulation, and to perform time-series fusion processing on each charging load curve to obtain the prediction result of the total regional charging load.

[0018] According to a third aspect, the present invention provides an electronic device comprising: At least one processor; and The memory that is communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform any of the electric heavy-duty truck charging load prediction methods in the embodiments of the present invention.

[0019] According to another aspect of the present invention, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause a computer to execute any of the electric heavy-duty truck charging load prediction methods in the embodiments of the present invention.

[0020] The present invention provides a method for predicting the charging load of electric heavy-duty trucks across multiple scenarios. This method is achieved through a collaborative process of acquiring operational characteristic data, scenario-based classification, vehicle inventory prediction, behavioral modeling, and load fusion. Scenario-based classification identifies vehicle groups under different operating modes, while the charging behavior probability model describes the statistical patterns of charging in specific scenarios. The method includes: first, acquiring operational characteristic data of electric heavy-duty trucks and classifying them according to preset scenario division rules to generate multiple independent typical scenarios; then, based on the predicted results of the regional heavy-duty truck market size and electrification process, calculating the vehicle inventory data for each typical scenario through a scenario-based structure allocation method; next, constructing a charging behavior probability model based on the vehicle inventory data of each typical scenario; and finally, generating charging load curves for each typical scenario through random simulation based on each charging behavior probability model, and performing time-series fusion processing on the curves to obtain the predicted result of the total regional charging load.

[0021] In this technical solution, the present invention addresses the deficiency described in the background art of treating electric heavy-duty trucks in different operating scenarios as a homogeneous load group for prediction. It employs a scenario-based classification mechanism to categorize vehicles into typical scenarios based on their operational characteristics, thus resolving the issue of load characteristics being averaged due to model homogenization in traditional aggregation prediction methods. Furthermore, to address the problem that existing single models cannot accurately reflect the unique charging behavior patterns of each scenario, the invention constructs an independent probabilistic model of charging behavior for each typical scenario, achieving differentiated characterization of behaviors such as charging start time and power selection under different scenarios, including trunk line transportation, regional short-distance transportation, and closed-loop transport. Finally, to address the difficulty in identifying load peak time shifts and composite peaks, the invention generates load curves for each scenario and performs time-series fusion, ensuring accurate reconstruction of composite peak characteristics after load superposition from multiple scenarios. Therefore, the technical solution of this invention solves the technical problem of existing technologies failing to balance the accuracy and scenario adaptability of electric heavy-duty truck charging load prediction, achieving accurate prediction of electric heavy-duty truck charging load and providing reliable data support for power grid planning and charging facility layout. Attached Figure Description

[0022] Figure 1 This is a flowchart of an embodiment of the electric heavy-duty truck charging load prediction method of the present invention; Figure 2 This is a flowchart illustrating the overall process of predicting the charging load of electric heavy-duty trucks according to other embodiments of the present invention. Figure 3 This is a structural block diagram of an electric heavy-duty truck charging load prediction system according to an embodiment of the present invention; Figure 4 This is a block diagram of an electronic device used to implement embodiments of the present invention. Detailed Implementation

[0023] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0024] During the development of this invention, the inventors, through extensive experiments and data analysis, revealed the fundamental flaws of traditional charging load prediction methods: traditional aggregation prediction methods not only struggle to accurately capture the essential differences in charging characteristics between electric heavy trucks and passenger vehicles, but also severely smooth out load peaks by treating vehicles in different operating scenarios as a homogeneous group. Based on this discovery, the inventors innovatively proposed this technical solution, which utilizes a scenario-based classification mechanism to identify typical scenarios through operational characteristic data, and combines scenario-specific vehicle ownership prediction with behavioral modeling to achieve a precise depiction of differentiated charging patterns across multiple scenarios, embodying the core concept of "classification first, modeling second, and integration third."

[0025] Specifically, through comparative experiments, the invention team discovered that traditional single-model prediction has significant technical bottlenecks: the uniform probability distribution, which does not differentiate between scenarios, cannot simultaneously adapt to the bimodal characteristics of trunk transportation and the shift coupling characteristics of port operations. These technical deficiencies cause the prediction curve to exhibit an averaged shift at peak times, making it impossible to identify compound peaks. However, the scenario-based independent modeling method proposed in this invention can improve the accuracy of restoring the shape characteristics of the load curve and achieve precise superposition of loads from multiple scenarios.

[0026] Therefore, this invention provides a method for predicting the charging load of electric heavy-duty trucks according to the first aspect. This method can be applied to an electric heavy-duty truck charging load prediction and management system (hereinafter referred to as the "system"). The system runs on a server cluster or distributed computing platform through local deployment or cloud computing to complete accurate prediction and data analysis of the charging load of electric heavy-duty trucks in multiple scenarios. The physical equipment deployed in the system includes, but is not limited to, data acquisition servers, model calculation servers, storage arrays, and network switching equipment. These physical devices need to have high-performance parallel data processing and large-scale storage capabilities to support real-time acquisition of multi-source operational characteristic data, efficient training of charging behavior probability models, and parallel simulation calculation of load curves, ensuring that the system can reliably and timely output the regional total charging load prediction results.

[0027] like Figure 1 As shown, the method may include: Step S110: Obtain operational characteristic data of electric heavy trucks.

[0028] Among them, operational characteristic data refers to the set of multi-dimensional attribute data generated by electric heavy trucks during operation, which is used to characterize the vehicle's operating characteristics and charging needs. This includes information such as average daily mileage (unit: kilometers), transportation route characteristics (such as highways, urban roads, etc.), operating time patterns (such as weekday and holiday operating modes), vehicle type distribution data (such as tractor trucks, dump trucks, etc.), battery capacity parameters (unit: kilowatt-hours), and charging power parameters (unit: kilowatts).

[0029] Specifically, the system can acquire operational characteristic data in real time or in batches through multiple data interfaces such as vehicle network platforms, charging pile operator management systems, and traffic management department databases, and perform data verification (such as checking data integrity) and format standardization (such as unifying timestamp format) operations on the acquired data.

[0030] For example, the system obtains operational characteristic data of 500 electric heavy trucks from the vehicle monitoring platform of a regional logistics company. The average daily mileage ranges from 150 km to 1000 km, and the battery capacity ranges from 200 kWh to 500 kWh. The system removes invalid data points (approximately 1% of the total) caused by sensor failures through data verification, and unifies the timestamps to Beijing time format, finally generating a standardized operational characteristic dataset (approximately 100 megabytes in total size).

[0031] Step S120: Classify the operational feature data according to the preset scenario classification rules to generate multiple typical scenarios for electric heavy truck applications.

[0032] Among them, the scenario division rule refers to the classification standard set based on the threshold or combination of operational characteristic data, which is used to divide the electric heavy truck group into subsets with similar operating modes; typical scenarios refer to independent operating mode categories generated through classification, such as long-distance trunk transportation scenarios, regional short-distance transportation scenarios, and port park closed transshipment scenarios.

[0033] Specifically, the system can perform scenario-based classification through a rule engine or clustering algorithm (such as K-means clustering, a distance-based unsupervised machine learning algorithm used to automatically classify data points into K clusters): First, the classification conditions are defined based on the average daily mileage (e.g., more than 500 kilometers are classified as trunk line scenarios), transportation route characteristics (e.g., mainly highways) and operation time patterns (e.g., 24-hour continuous operation). Then, the operational characteristic data is mapped to the corresponding scenarios, and a unique identifier is assigned to each scenario.

[0034] For example, the system classifies the data of the above 500 electric heavy trucks according to the application scenario rules, identifying 200 vehicles belonging to the long-distance trunk transportation scenario (average daily mileage > 500 km, highway proportion > 80%), 150 vehicles belonging to the regional short-distance transportation scenario (average daily mileage 100~500 km, mainly urban roads), and 100 vehicles belonging to the port park closed transshipment scenario (average daily mileage < 100 km, fixed operating area). The system generates a classification report for each scenario, including scenario name, number of vehicles, and statistics of main characteristics.

[0035] Step S130: Based on the current regional heavy truck market size and the predicted results of electrification process, calculate the sub-scenario ownership data of each typical scenario through scenario-based structural allocation.

[0036] Among them, the regional heavy-duty truck market size refers to the predicted total number of regional heavy-duty trucks during the forecast period (e.g., 2025); the configuration of the electrification process forecast results refers to the electrification penetration rate forecast parameters based on policy objectives, technological maturity (e.g., battery technology improvements), and economic factors (e.g., cost reductions); the scenario-based structural allocation method refers to the mathematical method of allocating the total number of electric heavy-duty trucks to each typical scenario according to the mapping relationship between vehicle type and scenario; and the scenario-based ownership data refers to the predicted number of electric heavy-duty trucks in each typical scenario.

[0037] Specifically, the system can predict the total market size of regional heavy trucks through a multiple regression model, and predict the electrification penetration rate using a growth curve model (such as the Logistic model, a mathematical model that describes the S-shaped curve change law of the initial slow growth, the middle acceleration, and the later saturation) to obtain the total number of electric heavy trucks in operation. Then, based on vehicle distribution data (such as tractor vehicles accounting for 60%) and scenario mapping matrix (such as tractor vehicles mainly corresponding to trunk line scenarios), the system allocates the number of vehicles in operation through weighted calculation.

[0038] For example, the system predicts that the total market size of regional heavy trucks will be 50,000 units in 2025, with an electrification penetration rate of 25%, so the total number of electric heavy trucks will be 12,500 units. Based on the vehicle type-scenario mapping (80% of tractor units are allocated to trunk line scenarios), the number of units in the trunk line scenario is calculated to be 6,000 units (i.e., calculated by 12,500 × 60% × 80%), and other scenarios are calculated in the same way.

[0039] Step S140: Construct a charging behavior probability model based on the electric heavy truck ownership data for each typical scenario. The charging behavior probability model is configured to describe the statistical regularity of charging behavior under the corresponding typical scenario.

[0040] Among them, the charging behavior probability model refers to a mathematical model that quantifies the random characteristics of charging behavior through probability distribution functions, including the charging start time probability distribution (describing the probability of the start of charging), the charging power selection probability distribution (describing the probability of selecting different charging power levels), and the starting charging state probability distribution (describing the probability of the battery's charge state when charging begins).

[0041] Specifically, the system can fit the probability distribution using historical charging data: for each typical scenario, it collects charging start time, charging power and initial charge data, uses kernel density estimation or maximum likelihood method to fit the probability distribution function, and verifies the model fit (such as chi-square test).

[0042] For example, in the case of long-distance transportation on trunk lines, the system fits the probability distribution of charging start time based on historical data and finds that the peak occurs at 12:00 (probability 0.3) and 20:00 (probability 0.4); the probability distribution of charging power selection shows that the probability of 120 kW power being selected is 0.6; the average value of the probability distribution of starting charging state is 30% (standard deviation 10%).

[0043] Step S150: Based on the probability models of each charging behavior, generate charging load curves corresponding to each typical scenario through random simulation, and perform time-series fusion processing on each charging load curve to obtain the prediction result of the total regional charging load.

[0044] Among them, random simulation refers to the method of generating a sequence of charging events that conforms to a probability distribution by random sampling, such as Monte Carlo simulation; time series fusion processing refers to the operation of superimposing and merging multiple time series load curves on a unified time axis.

[0045] Specifically, the system can perform parallel Monte Carlo simulations: for each typical scenario, a virtual vehicle queue is generated based on the scenario-specific vehicle inventory data, and the charging parameters (such as start time and power) of each vehicle are randomly sampled using a charging behavior probability model. The charging load curve of a single vehicle (power × time) is calculated, and then all vehicle curves are aggregated to obtain the scenario load curve. Finally, the curves of each scenario are superimposed according to time points (such as adding them at 15-minute intervals).

[0046] For example, the system simulates 6,000 electric heavy trucks in the trunk line scenario and generates a 24-hour load curve with a peak load of 5,000 kilowatts; the peak load curve for the regional short-distance scenario is 3,000 kilowatts; through time-series fusion, the peak load curve reaches 8,000 kilowatts, and the system outputs a prediction result table (including time points and load values).

[0047] In other embodiments, such as Figure 2 The diagram illustrates the overall process of the electric heavy-duty truck charging load forecasting method. Taking an electric heavy-duty truck load forecasting project in a coastal province as an example, the system first executes step one: based on operational characteristic data such as average daily mileage and transportation route characteristics, the province's 50,000 electric heavy-duty trucks are divided into three typical scenarios—30,000 trucks in a long-distance trunk transportation scenario (average daily mileage > 500 km), 15,000 trucks in a regional short-distance transportation scenario (average daily mileage 200-500 km), and 5,000 trucks in a port closed transshipment scenario (average daily mileage < 200 km). Next, step two is executed: based on the province's GDP growth rate of 8% and new energy vehicle subsidy policies, the total heavy-duty truck market size is predicted to be 80,000 vehicles in 2025, with an electrification penetration rate of 30%, resulting in a total electric heavy-duty truck fleet of 24,000 vehicles; then, the fleet size for each scenario is calculated using a vehicle type-scenario mapping matrix (e.g., 80% of tractor units are allocated to trunk route scenarios). Step three is then executed: For trunk line scenarios, a dedicated probability model is constructed based on historical charging station data. The results show that the charging start time exhibits a bimodal distribution (midday peak probability 0.35, nighttime peak probability 0.45). Finally, step four is executed: 10,000 Monte Carlo simulations are performed on the existing capacity for each scenario. After generating scenario load curves, time-series fusion is performed to obtain a predicted curve for the province's peak load of 125 MW. This result is directly applied to the province's power grid upgrade planning scheme.

[0048] Therefore, according to the above implementation method, the system can achieve its goals through a collaborative process of acquiring operational characteristic data, scenario-based classification, vehicle ownership prediction, behavioral modeling, and load fusion. Scenario-based classification is used to identify vehicle groups under different operating modes, and the charging behavior probability model is used to describe the charging statistical patterns under specific scenarios. The method includes: first, acquiring operational characteristic data of electric heavy-duty trucks and classifying them according to preset scenario division rules to generate multiple independent typical scenarios; then, based on the prediction results of the regional heavy-duty truck market size and electrification process, calculating the vehicle ownership data for each typical scenario through a scenario-based structure allocation method; then, constructing a charging behavior probability model based on the vehicle ownership data of each typical scenario; finally, based on each charging behavior probability model, generating charging load curves for each typical scenario through random simulation, and performing time-series fusion processing on the curves to obtain the regional total charging load prediction result.

[0049] In this technical solution, this embodiment addresses the deficiency described in the background technology of treating electric heavy-duty trucks in different operating scenarios as a homogeneous load group for prediction. It solves the problem of load characteristics being averaged due to model homogenization in traditional aggregation prediction methods by classifying vehicles into typical scenarios based on their operating characteristics through a scenario-based classification mechanism. To address the issue that existing single models cannot accurately reflect the unique charging behavior patterns of each scenario, a charging behavior probability model is independently constructed for each typical scenario, achieving differentiated characterization of behaviors such as charging start time and power selection under different scenarios such as trunk line transportation, regional short-distance transportation, and closed-loop transport. To address the difficulty in identifying load peak time offsets and composite peaks, load curves are generated for each scenario and time-series fusion is performed, ensuring accurate restoration of composite peak characteristics after load superposition from multiple scenarios. Therefore, the technical solution of this embodiment solves the technical problem of existing technologies being unable to balance the accuracy and scenario adaptability of electric heavy-duty truck charging load prediction, achieving accurate prediction of electric heavy-duty truck charging load and providing reliable data support for power grid planning and charging facility layout.

[0050] In some embodiments, operational characteristic data includes average daily mileage, transportation route characteristics, operating time patterns, vehicle type distribution data, battery capacity parameters, and charging power parameters; the operational characteristic data is classified according to preset scenario division rules to generate multiple typical scenarios for electric heavy-duty truck applications, including: Multiple operating modes are defined based on average daily mileage, transportation route characteristics, and operating time patterns.

[0051] Among them, the operating model refers to a standardized operating behavior paradigm formed based on the combination of key operating characteristics, which is used to characterize a group of vehicles with similar operating patterns.

[0052] Specifically, the system defines the operation mode through feature combination analysis: the average daily driving mileage is divided into three ranges: long mileage (>500 km), medium mileage (200~500 km) and short mileage (<200 km); the transportation route characteristics are divided into three types: highways, national and provincial highways and urban roads; and the operation time pattern is divided into single shift (8 hours), double shift (16 hours) and continuous operation (24 hours).

[0053] For example, the system defines a "long-distance highway double shift" operation mode, with the following characteristics: average daily mileage > 500 kilometers, highway proportion > 80%, and operation time of 16 hours.

[0054] Set corresponding scenario classification standards for each operating model.

[0055] The scenario classification standard refers to the set of judgment rules that map the operation mode to typical scenarios, including feature thresholds and logical conditions.

[0056] Specifically, the system sets scenario classification standards through a decision tree rule engine: setting characteristic threshold conditions for each operation mode. For example, for the "long-distance highway double shift system" mode, the system sets a daily average driving mileage threshold of 500 kilometers, a highway mileage ratio threshold of 80%, and an effective working time threshold of 14 hours.

[0057] For example, for the "long-distance highway double shift system" mode, the system sets the scenario classification standard as follows: when a vehicle simultaneously meets the following conditions, such as an average daily mileage of ≥500 kilometers, a highway mileage ratio of ≥80%, and an operating time of ≥14 hours, it is classified as a long-distance trunk line transportation scenario.

[0058] Based on the scenario classification criteria, the application of electric heavy trucks is categorized into several independent typical scenarios.

[0059] The classification operation refers to the process of performing pattern matching and classification labeling on electric heavy truck data based on scenario division criteria.

[0060] Specifically, the system performs classification through a rule matching algorithm: it traverses the vehicle dataset, matches the operational characteristics of each vehicle with the classification criteria for each scenario, and vehicles that meet all threshold conditions are labeled as the corresponding scenario, while unmatched vehicles enter the waiting queue for manual review.

[0061] For example, the system processes data from 1,000 electric heavy-duty trucks, of which 320 trucks simultaneously meet the conditions of an average daily mileage of ≥500 kilometers, a highway ratio of ≥80%, and an operating time of ≥14 hours, and are classified as long-distance trunk transportation scenarios.

[0062] For each typical scenario after classification, determine the typical values ​​of the corresponding vehicle model distribution data, battery capacity parameters, and charging power parameters.

[0063] Typical values ​​refer to parameter values ​​or distribution ranges that can represent the main characteristics of the scenario, obtained through statistical analysis methods.

[0064] Specifically, the system determines typical values ​​through descriptive statistical analysis: calculating the vehicle model distribution ratio, average battery capacity, and mode of charging power for vehicle data in each scenario.

[0065] For example, for the 320 vehicles already classified for long-haul scenarios, the system statistics show that the vehicle type distribution is as follows: tractor trucks account for 85%, dump trucks account for 10%, and other vehicle types account for 5%; the typical battery capacity is 350 kWh (mean ± standard deviation: 350 ± 50 kWh); and the typical charging power is 240 kW (the most frequently occurring power level).

[0066] Typical scenarios include long-distance trunk line transportation, short-distance regional transportation, and closed-loop transshipment in port parks.

[0067] Specifically, the system maintains typical scene definitions through the scene label management module: creating a unique identifier for each scene and storing scene feature templates and classification rules.

[0068] For example, the system database creates three scenario records: Scenario ID001 (long-distance trunk transportation scenario) corresponds to vehicles with an average daily mileage of >500 kilometers and a highway ratio of >80%; Scenario ID002 (short-distance regional transportation scenario) corresponds to vehicles with an average daily mileage of 200~500 kilometers and mainly urban roads; Scenario ID003 (closed-loop transshipment scenario in port parks) corresponds to vehicles with an average daily mileage of <200 kilometers and fixed operating areas.

[0069] Therefore, according to the above implementation method, the system can realize the fine classification of electric heavy truck operation scenarios, providing an accurate scenario basis for subsequent differentiated modeling.

[0070] In some embodiments, the configuration of the regional heavy-duty truck market size and electrification progress forecast results includes the configuration of macroeconomic indicators and policy technology driving factors; based on the current configuration of the regional heavy-duty truck market size and electrification progress forecast results, the scenario-specific vehicle ownership data for each typical scenario is calculated through a scenario-based structural allocation method, including: The predicted value of the total regional heavy truck market size is calculated based on the correlation between macroeconomic indicators and historical sales data. Historical sales data is used to characterize the historical patterns and trends of the regional heavy truck market.

[0071] Among them, macroeconomic indicators include regional GDP (unit: 100 million yuan), fixed asset investment (unit: 100 million yuan), and highway freight turnover (unit: 100 million ton-kilometers); correlation refers to the mathematical model relationship established through regression analysis.

[0072] Specifically, the system uses a multiple linear regression model to establish predictive relationships: historical sales data is used as the dependent variable, and macroeconomic indicators lagged by one period are used as independent variables. The regression coefficients are estimated using the least squares method.

[0073] For example, a regression equation can be established based on data from a certain province from 2015 to 2022: ; Model fit It reached 0.85.

[0074] By inputting policy and technology driving factors into a preset growth curve model, the electrification penetration rate is predicted through the growth curve model, and the predicted value of the total number of electric heavy trucks is calculated.

[0075] Among them, policy-driven technology factors include purchase subsidies (unit: RMB 10,000 / vehicle), charging infrastructure coverage (unit: %), and battery energy density improvement rate (unit: %); the growth curve model refers to a mathematical model that uses an S-shaped curve to describe the diffusion law of technology.

[0076] Specifically, the system uses a Logistic growth model for forecasting: it sets a maximum market penetration threshold, adjusts the growth rate parameter based on policy and technology driving factors, and calculates the penetration rate for each year through iterative calculation.

[0077] For example, if the maximum penetration rate is set at 40% and the basic growth rate is 0.15, and the growth rate is adjusted upward to 0.18 based on policy support, the electrification penetration rate is predicted to reach 25% in 2025. Combined with a total market size of 50,000 vehicles, the total number of electric heavy trucks in operation is 12,500.

[0078] Based on the mapping relationship between the current vehicle configuration and various typical scenarios, the predicted total number of electric heavy trucks is allocated to each typical scenario to obtain the scenario-specific number of trucks.

[0079] Among them, the mapping relationship refers to the corresponding ratio matrix of vehicle models and scenarios determined through statistical analysis; the allocation calculation refers to the numerical calculation of proportional allocation according to weight coefficients.

[0080] Specifically, the system constructs a vehicle-scenario allocation matrix: statistically analyzes the distribution ratio of each vehicle model in different scenarios in historical data, and weights the total number of vehicles based on the proportion of each vehicle model and the scenario distribution coefficient.

[0081] For example, if tractor units account for 60% of the total fleet, and 80% of those are used in mainline transportation, then: (vehicles); Dump trucks account for 30%, of which 70% are used in short-distance regional scenarios. Therefore: (vehicles).

[0082] Therefore, according to the above implementation method, the system can realize a complete calculation chain from macro forecasting to scenario-based allocation, providing accurate basic data on the number of units in each typical scenario.

[0083] In some embodiments, a probabilistic model of charging behavior is constructed based on the electric heavy-duty truck ownership data for each typical scenario, including: Based on the historical operation records of charging stations and vehicle operation data, historical charging behavior data under various typical scenarios are extracted. The historical charging behavior data is used to characterize the statistical features of charging behavior under different scenarios.

[0084] Historical charging behavior data refers to a time-series data set collected from the charging station monitoring system and the vehicle telematics system that records the charging process of electric heavy trucks. It includes fields such as charging start timestamp, charging power value, initial battery charge (unit: percentage) and charging duration (unit: minutes).

[0085] Specifically, the system can extract data through the Extract-Transform-Load (ETL) process: exporting raw records from the charging station database, mapping fields and unifying units (such as unifying charging power to kilowatts), and filtering data by scenario tags.

[0086] For example, the system obtains data for the entire year of 2023 from a charging network in a certain region, extracting 15,000 charging records for long-distance transportation scenarios on trunk lines. The peak charging start time field shows that the peak occurs at 12:00 and 20:00, and the charging power value is mainly distributed in the range of 120 kW to 240 kW.

[0087] Based on historical charging behavior data, we establish probability distribution models for charging start time, charging power selection, and starting charging state.

[0088] Among them, the charging start time probability distribution model is a mathematical model that describes the distribution of the time when a vehicle starts charging through a probability density function; the charging power selection probability distribution model is a discrete probability model that quantifies the probability of different charging power levels being selected; and the starting charging state probability distribution model is a continuous probability model that characterizes the random characteristics of the battery's charge state when charging begins.

[0089] Specifically, the system can fit a continuous probability distribution (such as the initial charging state) using kernel density estimation (KDE) and a discrete probability distribution (such as charging power selection) using maximum likelihood estimation (MLE).

[0090] For example, for trunk line scenario data, the system uses Gaussian kernel density estimation to fit the probability distribution of the starting charging time, resulting in a bimodal curve with peak probabilities of 0.3 (12:00) and 0.4 (20:00), respectively; the probability distribution of charging power selection is fitted by a multinomial distribution, showing that the probability of 120 kW power being selected is 0.6, and the probability of 240 kW power is 0.3.

[0091] A probability model for charging behavior is constructed by integrating the probability distribution models of charging start time, charging power selection, and initial charging state.

[0092] Integration refers to the operation of combining multiple independent probability models into a unified model framework, which is usually achieved by joint probability distribution or model concatenation.

[0093] Specifically, the system can achieve integration by constructing a multidimensional joint probability distribution table: the output probabilities of the three models are used as joint random variables, and the overall charging behavior probability is calculated through conditional probability.

[0094] For example, the system constructs a three-dimensional probability table with dimensions of time slices (96 15-minute intervals), power levels (5 levels), and power ranges (10 ranges). The probability value of each cell is the product of the probabilities of three independent models, ultimately generating a joint distribution table containing 4800 probability values.

[0095] Among them, the charging start time probability distribution model is used to describe the probability distribution of the time when the vehicle starts charging, the charging power selection probability distribution model is used to describe the probability distribution of the selection of different power levels, and the starting charging state probability distribution model is used to describe the probability distribution of the battery charge state when charging starts.

[0096] Specifically, the system calls each sub-model in the model application: the charging start time model outputs the charging probability at each time point within 24 hours, the charging power model outputs the selection probability of each power level, and the starting charging state model outputs the probability of the power range.

[0097] For example, in the trunk line scenario simulation, the system calls the charging start time model and obtains a charging probability of 0.3 at 12:00; calls the charging power model and obtains a probability of 0.6 for selecting 120 kW; calls the initial state model and obtains a probability of 0.5 for the battery level being below 30%. The joint probability is then calculated as follows: .

[0098] Therefore, according to the above implementation method, the system can construct a probabilistic model that accurately reflects the charging behavior patterns of various typical scenarios, providing a differentiated modeling basis for load forecasting.

[0099] In some embodiments, the step of generating charging load curves for each typical scenario based on each charging behavior probability model through stochastic simulation includes: The number of electric heavy trucks to be simulated in each typical scenario is calculated based on the inventory data of each sub-scenario.

[0100] The number of electric heavy trucks to be simulated refers to the total number of virtual vehicles that need to be generated during the Monte Carlo simulation, and its value is equal to the predicted value of the sub-scenario ownership of this typical scenario.

[0101] Specifically, the system directly obtains the values ​​from the scenario-specific inventory data table through the data reading interface and inputs them as the total simulation parameters into the random simulation engine.

[0102] For example, in the long-haul transportation scenario, the system reads the value of 6,000 vehicles from the scenario-specific inventory data table, which means that the charging behavior of 6,000 electric heavy trucks needs to be simulated in this scenario.

[0103] For each electric heavy-duty truck to be simulated, the charging start time, charging power, and initial charging status are generated by random sampling of probability distribution based on the charging behavior probability model of the corresponding typical scenario.

[0104] Among them, probability distribution random sampling refers to the calculation method of generating random samples according to a preset probability distribution, which is usually implemented by inverse transformation sampling or rejection sampling algorithm.

[0105] Specifically, the system calls a random number generator (such as the Mason twisting algorithm) to generate... Uniformly distributed random numbers are input into the inverse cumulative distribution function of each probability distribution model to obtain the sampling results. For the charging start time, the specific time (e.g., 12:30) is sampled from the time probability distribution model; for the charging power, the power level (e.g., 240 kW) is sampled from the power selection probability distribution model; and for the initial charging state, the SOC (State of Charge) value (e.g., 25%) is sampled from the state of charge probability distribution model.

[0106] For example, the system generates a random number for the i-th virtual heavy truck. The initial charging time was calculated to be 13:45 using the inverse cumulative distribution function; random numbers Select 120 kW power; random number The corresponding initial SOC is 32%.

[0107] Based on the charging start time, charging power, and initial charging status of each electric heavy-duty truck, the charging time of each electric heavy-duty truck is calculated, and the single-vehicle charging load curve of each electric heavy-duty truck is generated.

[0108] Charging time refers to the time required to fully charge a battery from its initial charge level, calculated using battery capacity, charging power, and initial charge level. The single-vehicle charging load curve refers to the stepped curve showing the change in power over time for a single vehicle during the charging process.

[0109] Specifically, the system uses the formula: ; The above formula is used to calculate the duration, and then a rectangular load curve is generated with the charging start time as the starting point, the charging duration as the duration, and the charging power as a constant power value.

[0110] For example, if a car has a battery capacity of 350 kWh, an initial SOC of 30%, and a charging power of 120 kW, then: (Hour).

[0111] The system generates a 120 kW constant load curve that lasts for 2.04 hours starting at 13:45.

[0112] The charging load curves of each individual vehicle are superimposed and aggregated to generate charging load curves corresponding to each typical scenario.

[0113] Among them, the overlay aggregation process refers to the operation of accumulating the values ​​of multiple time series curves at the same time point to generate a total load curve.

[0114] Specifically, the system first discretizes the time axis into fixed intervals (such as 15 minutes), then sums the power values ​​of all single-vehicle load curves within each time interval to obtain the total load value of the scenario at that time point, and finally forms a continuous scenario load curve.

[0115] For example, the system divides 24 hours into 96 15-minute intervals, overlays the load curves of 6,000 vehicles, calculates the total load at each time point (e.g., the total load for the period from 12:00 to 12:15 is 5,000 kilowatts), and finally generates a scene load curve containing 96 data points.

[0116] Therefore, according to the above implementation method, the system can accurately generate charging load curves for each typical scenario through efficient random simulation and aggregate calculation.

[0117] In some embodiments, the step of performing time-series fusion processing on the charging load curves to obtain the regional total charging load prediction result includes: Based on the time synchronization relationship between typical scenarios corresponding to each charging load curve, a time series alignment operation is performed to obtain the aligned time series load dataset.

[0118] Among them, time synchronization relationship refers to the correspondence between the charging load curves of each typical scenario and the same starting point and time interval on the time axis; time series alignment operation refers to the data processing process of adjusting multiple time series data to a unified time baseline; time series load dataset refers to the data set formed after alignment processing, which contains load data from multiple scenarios and has the same timestamp.

[0119] Specifically, the system can perform alignment operations through a time resampling algorithm: first, set a uniform time resolution (such as a 15-minute interval), and then use linear interpolation or nearest neighbor interpolation methods to align the timestamps of each scenario load curve to the standard time grid, ensuring that all curves have the same number and order of time points.

[0120] For example, the system aligns the load curves of trunk line scenarios (original timestamps are 10:00, 10:15, and 10:30) with the curves of regional short-distance scenarios (original timestamps are 10:05, 10:20, and 10:35) to the time points of 10:00, 10:15, and 10:30 through linear interpolation, generating a unified dataset containing 96 time points (24 hours × 4), with a data size of approximately 200 kilobytes.

[0121] Based on the aligned time-series load dataset, the total charging load curve of the region is generated by superimposing the load values ​​at each time point.

[0122] Among them, the time-point load value superposition calculation refers to the arithmetic summation of the load values ​​of multiple scenarios at the same time point in the aligned dataset; the regional total charging load curve refers to the continuous curve reflecting the change of the total charging power of all electric heavy trucks in the region over time.

[0123] Specifically, the system can perform superposition calculations using a parallel accumulation algorithm: iterates through each time point, adds up the load values ​​of all scenarios at that point, and connects the results in chronological order to form the overall curve.

[0124] For example, at 10:00 AM, the load on the main line is 5000 kW, the load on the regional short-distance line is 3000 kW, and the load on the port is 1000 kW. The total load is then... kilowatts; the system repeats this operation at 96 time points to generate a total load curve with a peak load of 12,000 kilowatts.

[0125] Based on the regional total charging load curve, the regional total charging load forecast is generated by extracting key load indicators.

[0126] Among them, key load indicators refer to numerical indicators extracted from the load curve to quantify load characteristics, such as peak load (unit: kilowatt), load factor (unit: percentage), and total charging power (unit: kilowatt-hour); the regional total charging load forecast result refers to the structured output containing key indicators and curve data.

[0127] Specifically, the system can extract indicators through numerical analysis algorithms: calculate the maximum value of the curve to obtain the peak load, calculate the ratio of the average value to the peak value to obtain the load factor, and integrate the curve to obtain the total power consumption.

[0128] For example, the system extracts a peak load of 12,000 kW (occurring at 18:00) from the total load curve, with a load factor of 65% (i.e., average load 7,800 kW / peak load 12,000 kW) and a total electricity consumption of 150,000 kWh; finally, it generates a forecast report, including indicator tables and graphs.

[0129] Therefore, according to the above implementation method, the system can achieve accurate fusion of load curves in multiple scenarios and extraction of key indicators, and output regional total charging load prediction results that can support power grid planning.

[0130] In some embodiments, after generating the predicted total charging load for the region, the method further includes: The total regional charging load forecast results are transmitted to the power grid dispatch system.

[0131] Among them, the power grid dispatching system refers to a computer management system used to monitor, control and optimize the operation of the power system, and has the functions of data reception, storage and analysis; transmission operation refers to the process of sending the forecast results from the load forecasting system to the power grid dispatching system through the data communication interface.

[0132] Specifically, the system establishes a data connection through an Application Programming Interface (API) or a File Transfer Protocol (FTP), and encapsulates the prediction results into data packets in a specific format (such as JSON or JavaScript Object Notation, a lightweight, language-independent data exchange format) for transmission.

[0133] For example, every hour, the system automatically transmits a data file (approximately 50KB in size) containing the load forecast results for the next 24 hours to the data receiving port of the provincial power grid dispatch center via HTTPS (HyperText Transfer Protocol Secure), with a transmission success rate of 99.9%.

[0134] A suggested configuration plan for charging facilities is generated based on the forecast results of the total regional charging load.

[0135] Among them, the charging facility configuration recommendation scheme refers to the optimized configuration scheme of parameters such as the construction location, power capacity and pile type of charging stations based on the load forecast results.

[0136] Specifically, the system processes the prediction results through a spatial load density analysis algorithm: it identifies peak load areas (such as areas with a load density greater than 500 kW / km²), combines land use planning data, and uses an integer programming model to calculate the optimal number of charging piles and power allocation.

[0137] For example, the system identified that the peak load around a logistics park reached 8,000 kilowatts at night and suggested adding 10 240-kilowatt DC fast charging piles in the area, with a total investment budget of about 3 million yuan, which is expected to meet the charging needs for the next three years.

[0138] Based on the current actual charging load data and the predicted total charging load for the region, the prediction difference analysis results are calculated, and the configuration parameters of the charging behavior probability model are updated according to the prediction difference analysis results.

[0139] Among them, the prediction difference analysis results refer to the error statistics (such as mean absolute percentage error) obtained by comparing the actual measured values ​​with the predicted values; the configuration parameter update refers to the process of adjusting the distribution parameters (such as mean and variance) in the probability model based on the error analysis.

[0140] Specifically, the system uses a rolling time window approach to calculate the discrepancy: it acquires actual load data daily, calculates the error sequence between the actual load data and the predicted values, and adjusts the parameters of the probabilistic model using a Bayesian update algorithm or gradient descent method.

[0141] For example, the system monitored for 7 consecutive days and found that the nighttime load forecast was 15% higher than normal. By updating the probability distribution model of charging start time from 0.4 to 0.35, the probability of the evening peak was reduced to less than 5%.

[0142] Therefore, according to the above implementation method, the system can realize the effective application of prediction results and continuous optimization of the model, forming a complete prediction-application-feedback closed loop.

[0143] Figure 3This is a structural block diagram of an electric heavy-duty truck charging load prediction system according to an embodiment of the present invention.

[0144] like Figure 3 As shown, the electric heavy-duty truck charging load prediction system includes: The operational characteristic data acquisition module 210 is used to acquire operational characteristic data of electric heavy trucks.

[0145] The typical scenario generation module 220 is used to classify the operational feature data into scenarios according to preset scenario division rules, and generate multiple typical scenarios for electric heavy truck applications.

[0146] The inventory data calculation module 230 is used to calculate the inventory data of each typical scenario based on the current regional heavy truck market size and the prediction results of the electrification process.

[0147] The charging probability model building module 240 is used to build a charging behavior probability model based on the electric heavy truck ownership data of each typical scenario. The charging behavior probability model is configured to describe the statistical regularity of charging behavior under the corresponding typical scenario.

[0148] The charging load prediction module 250 is used to generate charging load curves corresponding to each typical scenario based on each charging behavior probability model through random simulation, and to perform time-series fusion processing on each charging load curve to obtain the regional total charging load prediction result.

[0149] The specific functions and examples of each module and submodule of the device in this embodiment of the invention can be found in the relevant descriptions of the corresponding steps in the above method embodiments, and will not be repeated here.

[0150] According to embodiments of the present invention, the above-described method of the present invention can be applied to an electronic device and a readable storage medium.

[0151] Figure 4 A schematic block diagram of an example electronic device 600 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0152] like Figure 4As shown, the electronic device 600 includes a computing unit 601, which can perform various appropriate actions and processes based on a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. The RAM 603 may also store various programs and data required for the operation of the electronic device 600. The computing unit 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0153] Multiple components in electronic device 600 are connected to I / O interface 605, including: input unit 606, such as keyboard, mouse, etc.; output unit 607, such as various types of displays, speakers, etc.; storage unit 608, such as disk, optical disk, etc.; and communication unit 609, such as network card, modem, wireless transceiver, etc. Communication unit 609 allows electronic device 600 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0154] The computing unit 601 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 performs the various methods and processes described above, such as a method for predicting the charging load of an electric heavy-duty truck. For example, in some embodiments, a method for predicting the charging load of an electric heavy-duty truck can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as storage unit 608. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the method for predicting the charging load of an electric heavy-duty truck described above can be performed. Alternatively, in other embodiments, the computing unit 601 may be configured, by any other suitable means (e.g., by means of firmware), to perform an electric heavy-duty truck charging load prediction method.

[0155] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0156] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0157] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0158] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0159] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0160] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0161] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0162] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for predicting the charging load of electric heavy-duty trucks, characterized in that, include: Obtain operational characteristic data of electric heavy-duty trucks; The operational feature data is classified into scenarios according to preset scenario division rules to generate multiple typical scenarios for electric heavy truck applications. Based on the current regional heavy truck market size and electrification progress forecast results, the scenario-based ownership data of each typical scenario is calculated through a scenario-based structural allocation method. A charging behavior probability model is constructed based on the electric heavy truck ownership data of each typical scenario. The charging behavior probability model is configured to describe the statistical regularity of charging behavior under the corresponding typical scenario. Based on the charging behavior probability models, charging load curves corresponding to the typical scenarios are generated through random simulation, and the charging load curves are fused over time to obtain the regional total charging load prediction results.

2. The method according to claim 1, characterized in that, The operational characteristic data includes average daily mileage, transportation route characteristics, operating time patterns, vehicle type distribution data, battery capacity parameters, and charging power parameters; the operational characteristic data is classified into scenarios according to preset scenario classification rules to generate multiple typical scenarios for electric heavy truck applications, including: Based on the average daily mileage, the characteristics of the transportation routes, and the patterns of operating times, multiple operating modes are defined; Set corresponding scenario classification standards for each of the aforementioned operating models; Based on the aforementioned scenario classification criteria, the application of electric heavy-duty trucks is categorized into several independent typical scenarios; For each of the categorized typical scenarios, determine the corresponding typical values ​​for the vehicle model distribution data, the battery capacity parameter, and the charging power parameter; The typical scenarios include long-distance trunk line transportation, short-distance regional transportation, and closed-loop transshipment in port parks.

3. The method according to claim 1, characterized in that, The configuration of the regional heavy-duty truck market size and electrification progress forecast results includes the configuration of macroeconomic indicators and policy and technology driving factors; the configuration of the current regional heavy-duty truck market size and electrification progress forecast results, through a scenario-based structural allocation method, calculates the sub-scenario vehicle ownership data for each of the typical scenarios, including: The predicted value of the total regional heavy truck market size is calculated based on the correlation between the macroeconomic indicators and historical sales data. The historical sales data is used to characterize the historical patterns and trends of the regional heavy truck market. The policy and technology driving factors are input into a preset growth curve model, and the electrification penetration rate is predicted through the growth curve model to calculate the predicted value of the total number of electric heavy trucks. Based on the mapping relationship between the current vehicle configuration and each of the typical scenarios, the predicted total number of electric heavy trucks is allocated to each of the typical scenarios to obtain the scenario-specific number of trucks.

4. The method according to claim 3, characterized in that, The step of constructing a charging behavior probability model based on the electric heavy-duty truck ownership data for each of the typical scenarios includes: Based on the historical operation records of charging stations and vehicle operation data, historical charging behavior data under each of the aforementioned typical scenarios is extracted. The historical charging behavior data is used to characterize the statistical features of charging behavior under different scenarios. Based on the historical charging behavior data, a probability distribution model of charging start time, a probability distribution model of charging power selection, and a probability distribution model of starting charging state are established. A probability model for the charging behavior is constructed by integrating the probability distribution model of the charging start time, the probability distribution model of the charging power selection, and the probability distribution model of the initial charging state. The charging start time probability distribution model describes the probability distribution of the time when the vehicle starts charging, the charging power selection probability distribution model describes the probability distribution of different power levels, and the starting charging state probability distribution model describes the probability distribution of the battery charge state when charging starts.

5. The method according to claim 4, characterized in that, The step of generating charging load curves corresponding to each typical scenario based on each of the charging behavior probability models through random simulation includes: The number of electric heavy trucks to be simulated in the corresponding typical scenarios is calculated based on the inventory data of each scenario. For each electric heavy truck to be simulated, the charging start time, charging power and starting charging status are generated by random sampling of probability distribution according to the charging behavior probability model of the corresponding typical scenario. Based on the charging start time, charging power and starting charging status of each electric heavy truck, the charging time of each electric heavy truck is calculated, and the single-vehicle charging load curve of each electric heavy truck is generated. The charging load curves of each vehicle are superimposed and aggregated to generate the charging load curves corresponding to each typical scenario.

6. The method according to claim 5, characterized in that, The step of performing time-series fusion processing on the charging load curves to obtain the regional total charging load prediction result includes: Based on the time synchronization relationship between the typical scenarios corresponding to each of the charging load curves, a time series alignment operation is performed to obtain the aligned time series load dataset. Based on the aligned time-series load dataset, the total charging load curve of the region is generated by superimposing the load values ​​at each time point. Based on the total charging load curve of the region, the predicted result of the total charging load of the region is generated by extracting key load indicators.

7. The method according to claim 6, characterized in that, After generating the predicted total charging load for the region, the method further includes: The total charging load forecast results for the region are transmitted to the power grid dispatching system. A suggested configuration plan for charging facilities will be generated based on the predicted total charging load in the region. Based on the current actual charging load data and the predicted total charging load of the region, the prediction difference analysis results are calculated, and the configuration parameters of the charging behavior probability model are updated according to the prediction difference analysis results.

8. A charging load prediction system for electric heavy-duty trucks, characterized in that, include: The operational characteristic data acquisition module is used to acquire operational characteristic data of electric heavy trucks; The typical scenario generation module is used to classify the operational feature data into scenarios according to preset scenario division rules, and generate multiple typical scenarios for electric heavy truck applications. The inventory data calculation module is used to calculate the inventory data of each typical scenario based on the current regional heavy truck market size and the prediction results of the electrification process through a scenario-based structure allocation method. The charging probability model construction module is used to construct a charging behavior probability model based on the electric heavy truck ownership data of each typical scenario. The charging behavior probability model is configured to describe the statistical regularity of charging behavior under the corresponding typical scenario. The charging load prediction module is used to generate charging load curves corresponding to each of the typical scenarios through random simulation based on each of the charging behavior probability models, and to perform time-series fusion processing on each of the charging load curves to obtain the regional total charging load prediction result.

9. An electronic device, characterized in that, include: At least one processor; as well as The memory that is communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, Computer instructions are used to cause a computer to perform the method according to any one of claims 1-7.

Images