A mixed traffic flow dynamic energy consumption measurement method and device and a readable storage medium

Abstract

The application discloses a mixed traffic flow dynamic energy consumption measuring method and device and a readable storage medium, and comprises the following steps: acquiring a first vehicle structure proportion, a second vehicle structure proportion and a mixed vehicle structure proportion of a target region passing in a target period; correcting a reference energy consumption factor according to an environmental temperature, a traffic state parameter and a specific driving distance parameter of a vehicle on a target road section to obtain a corrected energy consumption factor of the target road section; for a first type of road section, determining a traffic flow of the first type of road section by a statistical inference model based on time dimension prior information, space dimension prior information and physical conservation constraints; and calculating the energy consumption of the target region based on the traffic flow of the first type of road section, the mixed vehicle structure proportion and the corrected energy consumption factor. According to the scheme, the hierarchical fusion of the vehicle structure, the coupling correction of the energy consumption factor and the inversion inference of the traffic flow are utilized, and the unobserved road section flow is calculated by using local traffic observation, historical traffic data and road network topology constraints.

Description

Technical Field

[0001] This application belongs to the field of highway traffic energy consumption measurement technology, specifically relating to a method, device and readable storage medium for dynamic energy consumption measurement of mixed traffic flow. Background Technology

[0002] Against the backdrop of green transformation of highway transportation and the planning of integrated transportation and energy infrastructure, accurate dynamic energy consumption calculations for mixed traffic flows including fuel-powered and pure electric passenger vehicles by road segment are fundamental to identifying energy hotspots in the road network and optimizing the layout of energy replenishment facilities. Highway ring roads converge local commuter traffic and cross-regional transit traffic, exhibiting significant tidal characteristics, especially during holidays. Accurate understanding of their energy consumption distribution directly determines the rationality and economic viability of planning distributed photovoltaic and energy storage facilities.

[0003] In existing technologies, energy consumption calculation for mixed traffic flow is typically broken down into three separate processes: vehicle type prediction, unit energy consumption acquisition, and road segment flow acquisition. However, existing technologies have the following shortcomings when applied to tidal traffic flow scenarios during winter holidays: First, in terms of vehicle structure prediction, since a single statistical model can only depict a group of vehicles with a long-term and complete evolutionary pattern, the large influx of heterogeneous vehicles during holidays exceeds the applicability of existing models in terms of both quantity and composition. This results in a systematic deviation between the predicted vehicle structure and the actual situation, which is directly transmitted down to the energy consumption calculation stage.

[0004] Second, at the level of unit energy consumption acquisition, energy consumption correction factors for different operating conditions are usually superimposed in a multiplicative manner, which implicitly assumes that the effects of each operating condition are independent of each other. However, in actual physical processes, under certain combinations of operating conditions, there is a coupling amplification effect caused by one operating condition prolonging the duration of the high-energy-consumption phase of another operating condition. This effect has not been identified and quantified in the current processing method.

[0005] Third, at the level of traffic flow acquisition, in a data environment with only discrete point-based acquisition conditions, there is a spatial gap between the point-based acquisition data and the linear and area-based traffic flow information required for continuous road segments in the road network. Existing technologies are unable to systematically infer and fill this spatial gap by utilizing the topology of the road network itself and the multi-source spatiotemporal patterns of traffic flow.

[0006] Fourth, at the level of the measurement system architecture, the technical means for predicting long-term macro trends and for measuring peak values ​​in specific periods are independent of each other in terms of underlying models, data input and output formats, and lack a unified computing kernel, which makes the system implementation complex and results inconsistent. Summary of the Invention

[0007] The purpose of this application is to provide a method, apparatus and readable storage medium for calculating dynamic energy consumption of mixed traffic flow, which can solve the above-mentioned core technical problems.

[0008] This application provides a method for calculating the dynamic energy consumption of mixed traffic flow, the method comprising: Obtain the proportion of the first type of vehicle type and the proportion of the second type of vehicle type in the target area during the target time period; Using the traffic flow of the first type of vehicles and the second type of vehicles as weights, the proportion of the first vehicle type structure and the proportion of the second vehicle type structure are integrated to obtain the proportion of mixed vehicle type structure in the target area during the target time period; The baseline energy consumption factor is corrected based on ambient temperature, traffic state parameters, and specific travel distance parameters of the vehicle on the target road segment to obtain the corrected energy consumption factor of the target road segment; wherein, the correction includes independent correction based on the ambient temperature and the traffic state parameters, and coupled correction based on the coupling relationship among the ambient temperature, the traffic state parameters, and the specific travel distance parameters. Based on the road network topology, identify the first type of road segments in the target area, where the first type of road segments do not have direct traffic observation data; For the first type of road segment, the traffic flow of the first type of road segment is determined by a statistical inference model based on prior information in the time dimension, prior information in the spatial dimension, and physical conservation constraints. The energy consumption of the target area is calculated based on the traffic flow of the first type of road segment, the proportion of the mixed vehicle type structure, and the corrected energy consumption factor.

[0009] In one optional embodiment of this application, obtaining the proportion of the second vehicle type structure for the second type of vehicles includes: Extract the passage data of the second type of vehicles from historical passage records for multiple historical target time periods. Each passage data includes at least the vehicle's place of origin, the checkpoint location when entering the target road network, the time of passing through the current checkpoint, and the direction of travel. Based on the stated place of origin, determine the vehicle's origin classification; Based on the mileage difference between the checkpoint location where the vehicle was initially captured when it entered the target road network and the current checkpoint location, the distance the vehicle has traveled is estimated, and the total travel distance is calculated based on the distance traveled and the direction of travel, and the total travel distance is divided into a preset travel distance range. Based on the time of passing through the current checkpoint, determine the type of passage period to which the current passage belongs; Multiple cross-classification categories are constructed by using the origin classification, travel distance range, travel time type, and travel direction as dimensions for cross-classification; For each cross-category, the total number of vehicles belonging to that category and the number of vehicles using a specific energy type are counted. The ratio of the number of vehicles using a specific energy type to the total number of vehicles is taken as the proportion of that energy type in that category. The proportion of the second vehicle model structure is determined based on the proportion of the energy type corresponding to one or more cross-classification categories matched during the target time period.

[0010] In one optional embodiment of this application, the baseline energy consumption factor is corrected to obtain the corrected energy consumption factor for the target road segment, including: The first unit mileage energy consumption and the second unit mileage energy consumption are obtained when the vehicle is driven under a baseline traffic condition at a first preset ambient temperature and a second preset ambient temperature; the first preset ambient temperature is normal temperature and the second preset ambient temperature is low temperature. Calculate the ratio of the energy consumption per unit mileage to the energy consumption per unit mileage, and use it as a correction factor for low-temperature single operating conditions. The third and fourth unit mileage energy consumption of the vehicle are obtained when it is driving under the first preset ambient temperature in the first preset traffic state and the second preset traffic state; the traffic efficiency of the second preset traffic state is lower than that of the first preset traffic state. The ratio of the energy consumption per fourth unit mileage to the energy consumption per third unit mileage is calculated and used as a correction factor for a single congestion operating condition. Acquire multiple sets of measured energy consumption data of the vehicle under the second preset ambient temperature and the second preset traffic conditions at different travel lengths; Divide each set of measured energy consumption data by the product of the baseline energy consumption factor, the low temperature single-condition correction factor, and the congestion single-condition correction factor, and determine the portion of the resulting ratio that exceeds 1 as the coupling effect contribution value under that travel length. Establish a functional relationship between the contribution value of the coupling effect and ambient temperature, traffic state parameters and travel length, wherein the contribution value of the coupling effect decreases as the travel length increases; The ambient temperature, traffic condition parameters, and travel length of the current target road segment are obtained, and the current coupling effect contribution value is determined using the aforementioned functional relationship. The coupling amplification factor is obtained by adding 1 to the current coupling effect contribution value. The corrected energy consumption factor for the target road segment is obtained by multiplying the baseline energy consumption factor, the low-temperature single-condition correction factor, the congestion single-condition correction factor, and the coupling amplification factor.

[0011] In one optional embodiment of this application, for fuel-powered vehicles, the first preset ambient temperature is normal temperature, and the second preset ambient temperature is low temperature; the reference traffic state is high-speed constant speed driving state. For electric vehicles, the first preset ambient temperature is normal temperature, and the second preset ambient temperature is low temperature; the reference traffic state is high-speed constant speed driving state.

[0012] In one optional embodiment of this application, for the first type of road segment, the traffic flow of the first type of road segment is determined by a statistical inference model based on time-dimensional prior information, spatial-dimensional prior information, and physical conservation constraints, including: Select historical dates and time periods that have the same traffic characteristics as the target date and the target time period, wherein the same traffic characteristics include at least the same or similar holiday types and weekday attributes; The traffic flow of the first type of road segment under each of the selected historical dates and historical time periods is obtained to form a historical traffic flow sample set; Calculate the weighted average of each sample in the historical traffic sample set as the mean of the time-dimensional prior information, wherein the sample whose time is closer to the target date is assigned a greater weight; Calculate the weighted variance of each sample in the historical traffic sample set as the variance of the prior information in the time dimension; Identify the target upstream road segment with direct traffic observation data that is adjacent to the first type of road segment in the road network topology; Obtain the measured traffic flow of the upstream road segment of the target during the current target time period; The mean of the spatial dimension prior information is calculated based on the pre-established flow transmission relationship from the upstream road segment of the target to the first type of road segment, which is based on historical data. The historical fitting error of the flow transmission relationship is obtained as the variance of the spatial dimension prior information; Obtain the number of samples in the historical traffic sample set, and determine the fusion weight based on the number of samples, wherein the fusion weight increases monotonically with the increase of the number of samples, and the value ranges from 0 to 1; Multiply the fusion weight by the mean of the time-dimensional prior information to obtain the first intermediate value; Multiply the difference between 1 and the fusion weight by the mean of the spatial dimension prior information to obtain the second intermediate value; Add the first intermediate value to the second intermediate value to obtain the mean of the joint prior distribution; The third intermediate value is obtained by multiplying the square of the fusion weight by the variance of the time-dimensional prior information. The fourth intermediate value is obtained by multiplying the square of the difference between 1 and the fusion weight by the variance of the spatial dimension prior information. Add the third intermediate value to the fourth intermediate value to obtain the variance of the joint prior distribution; The vehicle physical conservation relationship of the road network topology between the first type of road segment and adjacent road segments is used as the likelihood constraint of the statistical inference model. Based on the mean and variance of the joint prior distribution and the likelihood constraint, the posterior distribution of the traffic flow of the first type of road segment is determined by a Bayesian inference model, and the mean of the posterior distribution is determined as the traffic flow of the first type of road segment.

[0013] In one optional embodiment of this application, the method further includes: Based on the road network topology, identify a second type of road segment in the target area, wherein the topological hop count of the second type of road segment from the road segment with direct traffic observation data is greater than that of the first type of road segment; For the second type of road segment, the upper quantile of the posterior distribution of its traffic flow is calculated as a conservative estimate of the traffic flow of the second type of road segment based on the posterior distribution of its traffic flow. The conservatively estimated traffic flow is used as the traffic flow for the second type of road segment, and a risk label is generated for the second type of road segment.

[0014] In one optional embodiment of this application, the energy consumption of the target area is calculated based on the traffic flow of the first type of road segment, the proportion of the mixed vehicle type structure, and the corrected energy consumption factor, including: Receive calculation instructions for long-cycle prediction scenarios or short-cycle calculation scenarios. A first set of input parameters is configured for the long-term prediction scenario. The ambient temperature in the first set of input parameters is taken as the historical statistical average value, the traffic state parameter is taken as the parameter value representing the smooth flow state, and the vehicle structure ratio is taken as the first vehicle structure ratio. A second set of input parameters is configured for the short-cycle calculation scenario. The ambient temperature in the second set of input parameters is taken as the historical statistical extreme value, the traffic state parameter is taken as the parameter value representing the congestion state, and the vehicle structure ratio is taken as the ratio of the mixed vehicle structure. The unified underlying computing kernel is invoked to calculate the first energy consumption result of the target area in the long-term prediction scenario and the second energy consumption result in the short-term calculation scenario, based on the input traffic flow of the first type of road segment, the correction energy consumption factor, and the vehicle type structure ratio of the corresponding scenario. Output the first energy consumption result in an annual trend format, and output the second energy consumption result in an hourly distribution format.

[0015] Secondly, this application provides a device for calculating the dynamic energy consumption of mixed traffic flow, the device comprising: The vehicle type structure fusion module is used to obtain the first vehicle type structure ratio of the first type of vehicles and the second vehicle type structure ratio of the second type of vehicles passing through the target area during the target time period; and to fuse the first vehicle type structure ratio and the second vehicle type structure ratio with the traffic flow of the first type of vehicles and the second type of vehicles as weights to obtain the mixed vehicle type structure ratio of the target area during the target time period. An energy consumption factor correction module is used to correct a baseline energy consumption factor based on ambient temperature, traffic state parameters, and specific travel distance parameters of a vehicle on a target road segment, thereby obtaining a corrected energy consumption factor for the target road segment. The correction includes independent correction based on the ambient temperature and traffic state parameters, as well as coupled correction based on the coupling relationship between the ambient temperature, traffic state parameters, and specific travel distance parameters. The traffic flow inversion module is used to identify the first type of road segments in the target area based on the road network topology, where the first type of road segments do not have direct traffic observation data; and for the first type of road segments, the traffic flow is determined by a statistical inference model based on time-dimensional prior information, spatial-dimensional prior information, and physical conservation constraints. An energy consumption calculation module is used to calculate the energy consumption of the target area based on the traffic flow of the first type of road segment, the proportion of the mixed vehicle type structure, and the corrected energy consumption factor.

[0016] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement any of the above-described methods for calculating the dynamic energy consumption of mixed traffic flow.

[0017] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for calculating the dynamic energy consumption of mixed traffic flows.

[0018] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements any of the above-described methods for calculating the dynamic energy consumption of mixed traffic flow.

[0019] The proposed solution categorizes vehicle types into two groups with different data generation mechanisms: local steady-state and external tidal traffic. After obtaining the respective proportions, these groups are fused using traffic flow as the weight. This effectively solves the problem of the traditional vehicle transfer matrix's closed state space, which cannot accommodate the instantaneous injection of external vehicles, thus improving the accuracy of vehicle type prediction in holiday tidal traffic scenarios. In the energy consumption correction stage, by introducing specific travel parameters of vehicles on the target road segment, the coupling effect of low temperature and congestion is explicitly modeled as a coupled correction term of the ambient temperature, traffic state parameters, and specific travel parameters. This breaks the assumption of independence between the two in the traditional product correction framework, accurately characterizing the physical mechanism by which congestion prolongs the duration of the low-temperature, high-energy-consumption phase, thereby improving the accuracy of energy consumption calculation under combined operating conditions. In the traffic flow acquisition stage, for the first type of road segments without direct traffic observation data, the system simultaneously utilizes historical periodic information in the time dimension, road network topology transfer relationships in the spatial dimension, and strict physical conservation constraints. A statistical inference model is used to perform optimal estimation through multi-source information fusion, thereby achieving the inversion of traffic flow from sparse checkpoint observations to the entire road network without relying on real-time roadside monitoring equipment. Through hierarchical fusion of vehicle type structure, coupled correction of energy consumption factors, and inversion inference of traffic flow, it eliminates the need for continuous deployment of traffic monitoring equipment on all road segments. Traffic flow on unobserved road segments is inferred using local traffic observations, historical traffic data, and road network topology constraints. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 A flowchart illustrating a method for calculating dynamic energy consumption of mixed traffic flow provided by the present invention; Figure 2 A structural block diagram of a hybrid traffic flow dynamic energy consumption calculation device provided by the present invention; Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0023] The technical solutions disclosed in the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0024] Figure 1 A flowchart illustrating a method for calculating the dynamic energy consumption of mixed traffic flow provided in this application embodiment is shown below. Figure 1 As shown, the method may include: Step S101: Obtain the first vehicle type structure ratio of the first type of vehicles and the second vehicle type structure ratio of the second type of vehicles passing through the target area during the target time period; using the traffic flow of the first type of vehicles and the second type of vehicles as weights, fuse the first vehicle type structure ratio and the second vehicle type structure ratio to obtain the mixed vehicle type structure ratio of the target area during the target time period.

[0025] The target area refers to the entire set of road segments of the expressway ring network covered by this calculation, such as the entire ring expressway of a city and related sections of its radial expressways. The target time period refers to the specific time period of focus in this calculation, such as the peak hours from 9:00 to 11:00 AM during the Spring Festival holiday in 2024 (the third day of the Lunar New Year).

[0026] The first category of vehicles refers to the set of vehicles with long-term local passage records, namely, vehicles registered within the administrative region of the target city and with a consistently stable presence in the city's checkpoint data. The evolution of the vehicle type structure of this type of vehicle is influenced by local vehicle ownership patterns and new energy vehicle promotion policies, exhibiting a slow-changing and predictable data generation mechanism. The second category of vehicles refers to the set of vehicles without long-term local passage records, namely, temporary transit vehicles whose license plates are not registered in the city and do not have a consistently stable presence in the city's checkpoint data. These vehicles surge into the target road network during holidays, and their vehicle type structure is not constrained by local vehicle ownership evolution patterns, exhibiting a fast-changing and pulse-like data generation mechanism.

[0027] The first vehicle type structure ratio is defined as the proportion of vehicles of a certain energy type among the first type of vehicles traveling on a certain road segment within a certain hour, under the assumption that only first type vehicles are considered. The second vehicle type structure ratio is defined as the proportion of vehicles of a certain energy type among external vehicles, under the assumption that only second type vehicles are considered. The mixed vehicle type structure ratio is defined as the equivalent vehicle type ratio obtained by comprehensively considering both types of traffic flow.

[0028] Specifically, for the proportion of the first vehicle type, the existing vehicle transfer matrix is ​​invoked. This matrix is ​​a macro-predictive model that uses vehicle age as a state variable and characterizes the evolution of vehicle ownership over time through transfer probabilities. After inputting the target calculation year and the selected electrification policy scenario, the vehicle transfer matrix outputs the total proportion of each energy type of vehicle at the regional level based on the local vehicle ownership evolution law. Since this proportion is at the regional total level, it needs to be further allocated to each road segment and time period using the annual average daily traffic volume (AADT) as spatial weight and the diurnal variation coefficient of each hour as time weight, ultimately obtaining the proportion of the first vehicle type by road segment and time period.

[0029] The proportion of the second vehicle type is determined based on the proportion of the energy type corresponding to one or more cross-classification categories matched during the target time period, as will be explained in detail later. Traffic flow estimates for the two types of traffic flows are as follows: the first type of traffic flow is obtained by combining the vehicle transfer matrix with historical daily variation coefficients, while the second type of traffic flow is obtained by statistical analysis of checkpoints during similar holidays and time periods in history. Finally, the proportions of the two types of vehicle types are weighted and fused to obtain the mixed vehicle type structure proportion for this road segment and time period. This fusion is performed at the proportion level rather than the quantity level, ensuring that the output result is always a valid probability value between 0 and 1.

[0030] Step S102: Based on ambient temperature, traffic state parameters, and specific travel distance parameters of the vehicle on the target road segment, the baseline energy consumption factor is corrected to obtain the corrected energy consumption factor of the target road segment; wherein, the correction includes independent correction based on the ambient temperature and the traffic state parameters, and coupled correction based on the coupling relationship between the ambient temperature, the traffic state parameters, and the specific travel distance parameters.

[0031] Wherein, ambient temperature T refers to the external atmospheric temperature of the road segment in which the vehicle travels. In this embodiment, the traffic state parameter can be taken as the average vehicle speed v on the target road segment. In this preferred embodiment, the specific travel distance parameter can be taken as the travel length L of the vehicle on the target road segment, in kilometers.

[0032] Baseline energy consumption factor This refers to the energy consumption per unit mileage of a vehicle under standard operating conditions (e.g., ambient temperature 20°C, constant speed 80 km / h, vehicle already in thermal equilibrium). Corrected energy consumption factor. This refers to the dynamic energy consumption per unit mileage after being jointly corrected for temperature, vehicle speed, and coupling effects.

[0033] Specifically, this step involves two levels of correction to the baseline energy consumption factor. The first level is independent correction, which involves separately calibrating the correction factor for a single low-temperature operating condition. Congestion Single Condition Correction Factor . The energy consumption was obtained by testing and normalizing the energy consumption at a constant free-flow vehicle speed of 80 km / h under different temperature levels, reflecting the independent effect of temperature on energy consumption. The energy consumption was measured and normalized at a constant temperature of 20℃ and different average vehicle speeds, reflecting the independent impact of vehicle speed on energy consumption.

[0034] The second level is coupling correction. For gasoline vehicles, low temperatures require a warm-up phase after a cold start, during which fuel consumption is significantly higher. Under normal, smooth traffic conditions, the warm-up process ends after driving 5 to 10 kilometers. However, in congested traffic, vehicle speeds are extremely low, and the actual mileage traveled in the same warm-up time is greatly shortened. Congestion effectively prolongs the proportion of the total mileage in the low-temperature, high-energy-consumption phase. For electric vehicles, low temperatures increase the internal resistance of the battery and require the operation of positive temperature coefficient ceramic (PTC) heaters to heat the battery and cabin. In congested traffic, the drive power decreases due to low vehicle speed, while the PTC heating power remains relatively constant, resulting in a significant amplification of the heating energy consumption proportion.

[0035] Therefore, this step can quantitatively characterize this coupling effect by introducing the travel length L. The shorter the travel, the larger the proportion of the high-energy-consumption phase in the total travel, and the stronger the coupling effect; the longer the travel, the weaker the coupling effect. During the execution phase, the T, v, and L values ​​of the current target road segment are obtained, and finally, the baseline energy consumption factor, low-temperature correction factor, congestion correction factor, and coupling amplification factor are multiplied to obtain the corrected energy consumption factor. The above process is applicable to both fuel vehicles and electric vehicles; only the corresponding calibration data table needs to be switched according to the energy type when calling parameters.

[0036] Additionally, it should be noted that specific travel parameters can also be parameters that characterize the vehicle's thermal state, such as cumulative driving time after startup, cumulative driving distance after startup, parking time before startup, thermal state when entering the target road segment, or the actual overlap time or overlap distance between the high energy consumption phase and the target road segment.

[0037] Specifically, the cumulative driving time and distance after startup can be read directly from the vehicle's CAN bus / on-board diagnostic interface: After the vehicle engine starts (or the electric drive system is powered on), the ECU / VCU continuously records the cumulative running seconds and cumulative mileage of this driving cycle. This data can be obtained in real time through the OBD-II interface, the vehicle networking terminal (T-Box), or the pre-installed data acquisition module, and reported to the cloud or roadside computing unit.

[0038] The parking time before startup can be obtained based on the vehicle's continuous trajectory record: For vehicles with complete trajectory records (such as ride-hailing vehicles and logistics vehicles), the parking time is the time difference between the last time the engine was turned off and the current startup time. It can be directly reported by the onboard terminal.

[0039] The thermal state upon entering the target road segment refers to the degree of warm-up or heating of the engine / battery / cabin when the vehicle enters the target road segment. This can be obtained based on a mapping between cumulative driving parameters and warm-up curves: Warm-up / heating curves are pre-established for different vehicle types (gasoline vehicles, electric vehicles) and ambient temperatures, describing the changes in engine coolant temperature or battery temperature with cumulative driving time (or cumulative driving distance). For example: a warm-up curve for a gasoline vehicle at -10℃, showing the engine coolant temperature rising from -10℃ to 85℃ over driving time; a heating curve for an electric vehicle at 0℃, showing the battery self-heating power changing with driving time. Upon entering the target road segment, based on the vehicle's current cumulative driving time / distance, the current engine / battery temperature is obtained by looking up a table on the warm-up / heating curve at the corresponding ambient temperature, determining whether warm-up is complete (e.g., coolant temperature > 80℃ or battery temperature > 15℃).

[0040] Obtaining the actual overlap time or overlap distance between the high-energy-consumption phase and the target road segment: Obtaining the overlap distance: First, determine the theoretical driving distance corresponding to the high-energy consumption phase of the vehicle under the current environmental conditions. This distance can be obtained from pre-calibrated data based on the vehicle model and ambient temperature. For example, for a certain fuel vehicle, the high-energy consumption phase at -10 degrees Celsius covers the first few kilometers after startup. Simultaneously, obtain the mileage already accumulated by the vehicle when entering the current road segment. This mileage can be obtained from the cumulative mileage reported in real time by the onboard terminal or calculated from historical trajectory data. Compare the vehicle's cumulative mileage with the theoretical distance of the high-energy consumption phase: If the cumulative mileage is greater than or equal to the theoretical distance, it means the vehicle has completed warm-up or heating upon entering the road segment, and the high-energy consumption phase no longer overlaps with the current road segment; the actual overlap distance is zero. If the cumulative mileage is less than the theoretical distance, it means the vehicle is still in the high-energy consumption phase upon entering the road segment. In this case, it is necessary to further compare the remaining high-energy consumption distance with the length of the current road segment. The remaining high-energy consumption distance is obtained by subtracting the cumulative mileage from the theoretical distance. The remaining high-energy-consumption distance is compared with the current road segment length, and the smaller of the two values ​​is taken as the actual overlap distance between the high-energy-consumption stage and the target road segment.

[0041] Obtaining the overlap time: Similar to the distance acquisition approach, first determine the theoretical duration of the high-energy consumption phase of the vehicle under current environmental conditions, such as the number of minutes required to reach warm-up status from startup. Simultaneously, obtain the cumulative driving time when the vehicle enters the target road segment. Compare the cumulative driving time with the theoretical duration: if the cumulative driving time exceeds the theoretical duration, the overlap time is zero. If it does not exceed the theoretical duration, calculate the remaining high-energy consumption time, which is the theoretical duration minus the cumulative driving time. The estimated driving time of the vehicle within the target road segment can be obtained by dividing the segment length by the average speed of the segment. Compare the remaining high-energy consumption time with the estimated driving time on that segment, and take the smaller of the two values; this is the actual overlap time between the high-energy consumption phase and the target road segment.

[0042] When it is not possible to obtain precise mileage or time data for each vehicle, the thermal state distribution of a vehicle entering a road segment can be estimated based on road segment-level statistical information, such as the average start frequency and single-trip characteristics of similar vehicle types and ambient temperatures, and thus the overlap can be estimated.

[0043] Step S103: Based on the road network topology, identify the first type of road segment in the target area. The first type of road segment does not have direct traffic observation data. For the first type of road segment, determine the traffic flow of the first type of road segment through a statistical inference model based on time-dimensional prior information, spatial-dimensional prior information, and physical conservation constraints.

[0044] The first category of road segments refers to road segments in the road network where neither the segment itself nor its origin and destination have direct checkpoints, making it impossible to directly obtain traffic flow data. Temporal prior information refers to the prior flow distribution obtained based on historical flow data of the road segment during similar holidays and time periods, stemming from the statistical periodicity of traffic flow with daily, weekly, and yearly cycles. Spatial prior information refers to the prior flow distribution predicted based on the current measured flow of upstream checkpoint-connected road segments with topological relationships to the road segment, through a pre-established spatial flow transmission relationship based on historical data, stemming from the determined physical laws of vehicle transmission between upstream and downstream road networks. The physical conservation constraint states that between consecutive adjacent road segments without ramp merging or merging, the incoming flow should approximately equal the outgoing flow. At the hourly time granularity, the change in vehicle storage within a single road segment is much smaller than the through flow; therefore, this approximation is physically highly valid. The statistical inference model adopted in this preferred embodiment is a Bayesian inference model.

[0045] It should be noted that in this application, the first type of road segment refers to a road segment for which there is no direct traffic flow observation data within the target time period. However, this does not mean that there is no observation data for this road segment in all time periods. In fact, for the vast majority of urban roads, verifiable traffic flow information has been accumulated through various historical means. This historical traffic flow data can serve as a sample source for constructing the prior distribution of traffic flow in the time dimension of the first type of road segment. Specifically, historical traffic flow data includes, but is not limited to, the following: Firstly, there is historical data from fixed monitoring equipment. The target area may have previously been equipped with coil detectors, microwave detectors, geomagnetic detectors, or checkpoint devices on or near Class I road sections. The collected traffic flow data, once archived, can be used as historical samples. Even if these devices cease operation during the target period due to maintenance, removal, or relocation, their historical records still have reference value.

[0046] Secondly, historical temporary traffic survey data. Traffic flow surveys conducted regularly or irregularly by traffic management departments or research institutions, including manual counting surveys and data collection by temporary portable detection equipment, cover some road sections that do not have fixed observations during the target time period.

[0047] Third, path reconstruction data based on ETC transaction records. By utilizing the transaction sequences of vehicles passing through ETC gantries, combined with road network topology and shortest path or most probable path algorithms, the driving trajectory of vehicles on each road segment between adjacent gantries can be reconstructed, thereby indirectly estimating the historical cross-sectional traffic flow of the first type of road segment. This method has wide coverage and a large sample size, and is particularly suitable for highways and urban expressways equipped with ETC gantries.

[0048] Fourth, historical traffic data based on video recognition. By playing back or analyzing historical surveillance videos near the target road segment, and using vehicle detection and tracking algorithms, traffic flow and vehicle type information for historical time periods can be extracted as a verified source of historical traffic data.

[0049] Fifth, historical estimation based on floating car trajectory data. By utilizing floating car data with GPS trajectory records from ride-hailing vehicles, taxis, and logistics vehicles, historical traffic flow of road segments can be inferred through methods such as trajectory matching and penetration inflation. Since floating car data typically has a long time span, it can provide multi-year historical traffic flow references for most urban roads.

[0050] In certain cases, if a Class I road segment truly lacks any direct or indirect historical observation records—such as newly opened roads or remote road segments without any long-term monitoring coverage—then historical traffic flow for that segment can be obtained using road network topology reconstruction and interpolation methods. Specifically, a reference road segment is selected that is at the same or similar functional level in the road network topology as the Class I road segment, has similar traffic demand characteristics, and has complete historical observation data. Based on the historical traffic flow time series of the reference road segment, and combined with factors such as the ratio of road segment length, lane number ratio, and similarity of land use along the route between the target road segment and the reference road segment, a traffic flow transfer factor is constructed, mapping the historical traffic flow of the reference road segment to the estimated historical traffic flow of the target road segment. This estimated value needs to be validated before being used for time-dimensional prior construction. The verification method can be as follows: Select one or more road segments with historically available observation data but similar topological characteristics to the target road segment as the verification set. Use the same flow transfer factor for backtesting and compare the deviation between the estimated results and the actual observations. If the mean absolute percentage error is within an acceptable range (e.g., no more than 15%), the reconstruction method is considered effective, and the reconstructed flow sequence can be used as a historical flow sample for the first type of road segment. Simultaneously, to address the additional uncertainties introduced by the reconstruction, the variance of the prior distribution can be appropriately increased when constructing the time-dimensional prior to reflect differences in data reliability.

[0051] By utilizing the aforementioned diverse sources of historical traffic data and their supplementary reconstruction methods, a sufficient and verifiable sample basis can be provided for the temporal prior distribution of the first type of road segment in this application, thereby supporting the effective implementation of subsequent Bayesian inference.

[0052] Specifically, the traffic inversion module first performs a classification of road segment checkpoint coverage levels across the entire road network. Based on the number of road segments (topology hops) required to reach the nearest checkpoint-equipped road segment from the current road segment, the entire road network segments are divided into Class I coverage (traffic can be directly read from checkpoint records), Class II coverage, and Class III coverage (all belonging to Class I road segments, requiring inference to obtain traffic). Then, a dual-source prior is constructed for each Class I road segment. The temporal prior is based on a weighted statistical analysis of historical traffic samples from similar days to obtain the mean and variance, while the spatial prior is based on predictions of the current traffic of known upstream road segments combined with historical regression relationships to obtain the mean and variance. The dual-source priors are adaptively fused into a joint prior based on the effective historical sample size.

[0053] Finally, the physical conservation law is quantified as a likelihood constraint that the flow difference between adjacent road segments follows a normal distribution with zero mean and minimal variance. Bayesian inference is used to fuse the joint prior and the conserved likelihood, yielding the posterior mean and posterior variance. The posterior mean is used as the optimal flow estimate. For road segments extremely far from checkpoints (where the number of topological hops exceeds a threshold), due to significant inference uncertainty, the 95th percentile of the posterior distribution is used as a conservative estimate, with an added risk label.

[0054] Step S104: Calculate the energy consumption of the target area based on the traffic flow of the first type of road segment, the proportion of the mixed vehicle type structure, and the corrected energy consumption factor.

[0055] Specifically, the energy consumption calculation module obtains the complete outputs of the aforementioned steps from the vehicle type structure fusion module, energy consumption factor correction module, and traffic flow inversion module, respectively. The total traffic flow of each road segment is multiplied by the proportion of mixed vehicle types to obtain the traffic flow by energy type. This is then multiplied by the corresponding corrected energy consumption factor and the physical length of the road segment to obtain the energy consumption by type for each road segment at each time period. Finally, the energy consumption distribution of the target area is obtained by summing the data across the entire road network at all times. Furthermore, this method supports simultaneously outputting dual-scenario results for long-term prospective prediction and short-term pre-holiday peak calculation on the same underlying calculation logic by switching the parameter configuration layer. The detailed implementation of the parameter configuration and mid-layer switching architecture will be explained later.

[0056] The proposed solution categorizes vehicle types into two groups with different data generation mechanisms: local steady-state and external tidal traffic. After obtaining the respective proportions, these groups are fused using traffic flow as the weight. This effectively solves the problem of the traditional vehicle transfer matrix's closed state space, which cannot accommodate the instantaneous injection of external vehicles, thus improving the accuracy of vehicle type prediction in holiday tidal traffic scenarios. In the energy consumption correction stage, by introducing specific travel parameters of vehicles on the target road segment, the coupling effect of low temperature and congestion is explicitly modeled as a coupled correction term of the ambient temperature, traffic state parameters, and specific travel parameters. This breaks the assumption of independence between the two in the traditional product correction framework, accurately characterizing the physical mechanism by which congestion prolongs the duration of the low-temperature, high-energy-consumption phase, thereby improving the accuracy of energy consumption calculation under combined operating conditions. In the traffic flow acquisition stage, for the first type of road segments without direct traffic observation data, the system simultaneously utilizes historical periodic information in the time dimension, road network topology transfer relationships in the spatial dimension, and strict physical conservation constraints. A statistical inference model is used to perform optimal estimation through multi-source information fusion, thereby achieving the inversion of traffic flow from sparse checkpoint observations to the entire road network without relying on real-time roadside monitoring equipment. Through hierarchical fusion of vehicle type structure, coupled correction of energy consumption factors, and inversion inference of traffic flow, it eliminates the need for continuous deployment of traffic monitoring equipment on all road segments. Traffic flow on unobserved road segments is inferred using local traffic observations, historical traffic data, and road network topology constraints.

[0057] In one optional embodiment of this application, obtaining the proportion of the second vehicle type structure for the second type of vehicles includes: Extract the passage data of the second type of vehicles from historical passage records for multiple historical target time periods. Each passage data includes at least the vehicle's place of origin, the checkpoint location when entering the target road network, the time of passing through the current checkpoint, and the direction of travel. Based on the stated place of origin, determine the vehicle's origin classification; Based on the mileage difference between the checkpoint location where the vehicle was initially captured when it entered the target road network and the current checkpoint location, the distance the vehicle has traveled is estimated, and the total travel distance is calculated based on the distance traveled and the direction of travel, and the total travel distance is divided into a preset travel distance range. Based on the time of passing through the current checkpoint, determine the type of passage period to which the current passage belongs; Multiple cross-classification categories are constructed by using the origin classification, travel distance range, travel time type, and travel direction as dimensions for cross-classification; For each cross-category, the total number of vehicles belonging to that category and the number of vehicles using a specific energy type are counted. The ratio of the number of vehicles using a specific energy type to the total number of vehicles is taken as the proportion of that energy type in that category. The proportion of the second vehicle model structure is determined based on the proportion of the energy type corresponding to one or more cross-classification categories matched during the target time period.

[0058] Specifically, firstly, out-of-town vehicles are identified. Information on each vehicle passing through a checkpoint is extracted from years of offline checkpoint time-series identification data during holidays. The criterion for determining whether a vehicle is an out-of-town vehicle is based on a joint judgment of two conditions. Condition 1: The license plate's registration location is not within the city. This is determined by the prefix characters of the license plate number or the location code directly output by the license plate recognition system. This condition means that the vehicle's registration location is not within the local administrative jurisdiction, its purchase, scrapping, and renewal are not directly affected by local policies, and it is not within the state space of the local vehicle transfer matrix. Condition 2: There is no long-term stable record of this license plate appearing in the city's checkpoint data. The cumulative number of days this license plate appears in all checkpoint records within the city over the past year (or multiple recent non-holiday periods) is counted. If the cumulative number of days is lower than a preset threshold (e.g., 5 days / year), the vehicle is determined to lack the characteristics of regular local travel and is a temporary transit vehicle. This condition means that vehicles with out-of-town license plates but actually used locally for extended periods are filtered out. Although these vehicles are registered outside the city, their actual usage is no different from that of local vehicles, and they should be classified as Category 1 vehicles. Vehicles that meet both of the above conditions are marked as out-of-town vehicles and included in Category 2 traffic samples.

[0059] Secondly, for records marked as external vehicles, a four-dimensional cross-classification statistical model based on "origin location × travel distance × travel time type × travel direction" is constructed. Specifically: Dimension one is the place of origin. Vehicles are categorized by the municipal-level administrative division where their license plates are registered (e.g., Langfang, Baoding, Zhangjiakou, Tianjin, etc.). Different places of origin have varying levels of new energy vehicle promotion and charging infrastructure development, which directly affects the proportion of electric vehicles among vehicles from that region. For example, the penetration rate of new energy passenger vehicles in Beijing is approximately 35%, in Tianjin approximately 30%, in Langfang approximately 20%, and in Zhangjiakou approximately 15%. The higher the new energy vehicle penetration rate in the place of origin, the higher the proportion of electric vehicles among vehicles originating from that region.

[0060] Dimension Two is travel distance. The estimated travel distance from the vehicle's origin to the current checkpoint is divided into three ranges: short-distance (less than 100 kilometers), medium-distance (100 to 300 kilometers), and long-distance (more than 300 kilometers). The estimation method is as follows: the distance difference between the vehicle's initial checkpoint location (usually the provincial border toll station on the highway entering the city or the first checkpoint on a radial highway) and the current checkpoint is taken as the vehicle's travel distance. Then, the total travel distance is calculated based on the vehicle's direction of travel. The implication of this dimension is that electric vehicles are subject to range constraints. In low-temperature winter conditions, the actual range of mainstream pure electric passenger vehicles is typically between 200 and 350 kilometers. Long-distance travel requires charging en route, while short-distance travel can be completed with a single full charge. Therefore, electric vehicles account for a higher proportion of vehicles entering the city in the short-distance range and a significantly lower proportion in the long-distance range.

[0061] Dimension three is based on travel time type. Holidays are divided into Spring Festival (Lunar New Year's Eve to the sixth day of the first lunar month), National Day (October 1st to October 7th), May Day (May 1st to May 5th), and other short holidays (Qingming Festival, Dragon Boat Festival, Mid-Autumn Festival, etc., 3-day holidays). Different holiday types correspond to different travel behavior patterns: Spring Festival is mainly for returning home to visit relatives (high proportion of long-distance inter-provincial travel, high passenger load per vehicle, and a high concentration of departure and destination points in specific cities); National Day is mainly for tourism (dispersed destinations, relatively high proportion of short and medium-distance trips). These differences in travel behavior directly affect the vehicle type structure of vehicles from outside the region.

[0062] Dimension four is the direction of travel. This distinguishes between inbound (from outside the city boundary to the city center) and outbound (from the city center to outside the city boundary). For inbound vehicles, most are long-distance arrivals (originating from their place of origin and destined for the city), and electric vehicles are less frequently chosen for long-distance travel due to range anxiety. For outbound vehicles, a large number are transit vehicles passing through the city to other areas, as well as short-distance trips to surrounding areas (of which electric vehicles account for a higher proportion because range anxiety is not an issue for short distances). Therefore, the proportion of electric vehicles among inbound vehicles is systematically lower than that among outbound vehicles.

[0063] For each cross-combination of the above four dimensions, the total number of vehicles from outside the area and the number of vehicles using pure electric power are counted in that category. The ratio of electric vehicles to the total number of vehicles is then calculated as the proportion of electric vehicles from outside the area in that category. .

[0064] Finally, for the current target measurement period, the proportion of external traffic disturbance is calculated segment by segment and time period by time. A weighted average method with sample size as the weight is adopted:

[0065] in, This is the set of cross-classification categories that match road segment i in time period t (i.e., the summary of multiple historical statistical classifications). Weights for each category (using the historical sample size of that category as the weight to reflect statistical reliability). This represents the percentage of imported electric vehicles within this category. The percentage of gasoline-powered vehicles in the second vehicle type structure is obtained through normalization. Weighted averages use sample size as the weight because categories with larger sample sizes tend to have more stable and reliable statistical proportions; this is an adaptive weighting strategy based on statistical reliability. Indicates road segment , Indicates time period , Indicates electric vehicle, Indicates gasoline-powered cars, This is the traversal symbol.

[0066] In obtaining the first type structure proportion of the first category of vehicles At the same time, obtain the proportion of the second type of vehicle structure in the second category. Similarly, the estimated value of the first type of traffic flow was obtained. The estimated value of the second type of traffic flow is obtained by combining the vehicle transfer matrix with historical diurnal variation coefficients. This data was extrapolated from historical statistics of checkpoints during similar holidays and time periods. Substituting these statistics into a fusion formula weighted by traffic flow, the final proportion of mixed vehicle types was obtained. :

[0067] Compared with existing technologies, the technical improvements of this solution are reflected in the following aspects: First, it identifies two different types of traffic flow and customizes the matching of different modeling methods, using a vehicle transfer matrix to handle local traffic flow and multi-dimensional cross-classification to handle incoming traffic flow. Second, it introduces a refined classification based on four dimensions: origin, travel distance, time period type, and direction of travel, capturing the deep behavioral patterns that determine the energy type selection of incoming vehicles. Third, the weighted average fusion uses historical sample size as the weight, ensuring automatic weight reduction in sparse sample categories and giving greater influence to statistically stable categories, thus improving the robustness and reliability of the proportion estimation.

[0068] In one optional embodiment of this application, the baseline energy consumption factor is corrected to obtain the corrected energy consumption factor for the target road segment, including: The first unit mileage energy consumption and the second unit mileage energy consumption are obtained when the vehicle is driven under a baseline traffic condition at a first preset ambient temperature and a second preset ambient temperature; the first preset ambient temperature is normal temperature and the second preset ambient temperature is low temperature. Calculate the ratio of the energy consumption per unit mileage to the energy consumption per unit mileage, and use it as a correction factor for low-temperature single operating conditions. The third and fourth unit mileage energy consumption of the vehicle are obtained when it is driving under the first preset ambient temperature in the first preset traffic state and the second preset traffic state; the traffic efficiency of the second preset traffic state is lower than that of the first preset traffic state. The ratio of the energy consumption per fourth unit mileage to the energy consumption per third unit mileage is calculated and used as a correction factor for a single congestion operating condition. Acquire multiple sets of measured energy consumption data of the vehicle under the second preset ambient temperature and the second preset traffic conditions at different travel lengths; Divide each set of measured energy consumption data by the product of the baseline energy consumption factor, the low temperature single-condition correction factor, and the congestion single-condition correction factor, and determine the portion of the resulting ratio that exceeds 1 as the coupling effect contribution value under that travel length. Establish a functional relationship between the contribution value of the coupling effect and ambient temperature, traffic state parameters and travel length, wherein the contribution value of the coupling effect decreases as the travel length increases; The ambient temperature, traffic condition parameters, and travel length of the current target road segment are obtained, and the current coupling effect contribution value is determined using the aforementioned functional relationship. The coupling amplification factor is obtained by adding 1 to the current coupling effect contribution value. The corrected energy consumption factor for the target road segment is obtained by multiplying the baseline energy consumption factor, the low-temperature single-condition correction factor, the congestion single-condition correction factor, and the coupling amplification factor.

[0069] Furthermore, for fuel-powered vehicles, the first preset ambient temperature is normal temperature, and the second preset ambient temperature is low temperature; the reference traffic state is high-speed constant speed driving state. For electric vehicles, the first preset ambient temperature is normal temperature, and the second preset ambient temperature is low temperature; the reference traffic state is high-speed constant speed driving state.

[0070] Specifically, the process of correcting the baseline energy consumption factor is as follows: 1. Purity determination of independent correction factors Before performing coupling correction, the independent effects of temperature and vehicle speed must first be calibrated under pure conditions.

[0071] Low Temperature Single Operating Condition Correction Factor The calibration method is as follows: Based on bench test data, the vehicle speed is controlled at a constant 80km / h, and only the ambient temperature is changed from multiple levels from -20℃ to 40℃. Alternatively, from the OBD measured data, travel segments with "free-flow speed (greater than 70km / h), travel length greater than 20km (negligible during warm-up), and different temperatures" are selected, and the energy consumption per unit mileage is extracted. Normalization was performed using 20℃ as the baseline, resulting in... The method obtains a continuous curve. By fixing the conditions of "free-flow speed and long stroke", it eliminates the interference of congestion and warm-up effect, and can purify the independent effect of temperature on energy consumption.

[0072] Congestion Single Condition Correction Factor The calibration method is as follows: Based on bench test data, the temperature is controlled at 20℃, and only the average vehicle speed of the test cycle is changed from multiple levels from 20km / h to 120km / h, or the travel segments with "normal temperature (15℃ to 25℃) and travel length greater than 20km" are selected from the OBD data, and the energy consumption per unit mileage at different average vehicle speeds is extracted. Normalization is performed using 80km / h as the baseline to obtain... The method yields a continuous curve. It eliminates the interference of low temperature and warm-up effects, and establishes the independent impact of pure calibration speed on energy consumption.

[0073] 2. Mathematical and physical modeling with coupling correction First, let's analyze the physical processes of gasoline-powered vehicles and electric vehicles. For gasoline-powered vehicles, after a cold start, there is a warm-up process before the engine coolant temperature reaches the normal operating range (approximately 85°C to 95°C). At low temperatures, gasoline atomization is poor and lubricating oil viscosity is high; during this stage, fuel consumption is higher than that under normal warm-up conditions. Times (typical value 1.3 to 1.8). Warm-up process duration. Influenced by two key factors: the lower the ambient temperature (T), the more heat needs to be absorbed, and the longer the warm-up time; the lower the vehicle speed (v), the lower the engine output power, the less heat is generated per unit, and the longer the warm-up time. Therefore, when both T and v decrease, The impact is amplified in two ways. For electric vehicles, the internal resistance of the battery can be 2 to 4 times that at room temperature at low temperatures, requiring the activation of a PTC electric heater (typically 3 to 6 kW) to heat the battery pack and passenger compartment. At -10°C, the PTC heating power and drive power (approximately 10 to 15 kW at a constant speed of 60 km / h) are already quite significant. In congested conditions, the drive power decreases due to reduced vehicle speed, but the PTC heating power remains relatively constant, further amplifying the proportion of heating energy consumption in the total energy consumption.

[0074] Based on the above physical analysis, a unified coupling correction model is designed for this step, which is applicable to both gasoline-powered and electric vehicles. The core derivation process of the modeling is as follows.

[0075] Consider a car traveling at an average speed v on a road segment of length L, with a travel time of L / v. The initial travel time after the vehicle starts... During this period, it is in a high-energy-consumption state (for gasoline vehicles, this is the warm-up phase). Electric vehicles are in the heating stage. The energy consumption rate at this stage is Afterwards, it enters a normal thermal equilibrium state, with an energy consumption rate of... Let the ratio of energy consumption rate under high energy consumption conditions to that under normal energy consumption conditions be _____. The average energy consumption over the entire journey can be expressed as a segmented weighted average with the travel distance as the weighting factor:

[0076] in, The distance traveled during the high-energy-consumption phase. This represents the distance traveled under normal conditions. Substitute into the above formula and extract the common factor. The results are as follows:

[0077] in, This is the independently adjusted expected energy consumption value. The portion outside the brackets (1) represents the additional energy consumption amplification ratio due to the coupling effect, from which the coupling amplification function is defined. for:

[0078] The physical meaning of each parameter in the expression is broken down. It reflects the proportion of energy consumption exceeding that under high-energy-consumption conditions compared to normal conditions, and is calibrated through bench tests. The distance traveled under high energy consumption conditions and in the aforementioned ambient temperature and traffic conditions. It is affected by both T and v; the lower T is, the lower v is. The longer. / L This represents the proportion of high-energy-consuming travel distance to the total travel distance. This proportion is inversely proportional to the travel length L. The shorter L is, the larger the proportion of high-energy-consuming travel, and the more significant the coupling effect. Conversely, the longer L is, the higher the energy-consuming travel distance is diluted by the normal travel distance, and the coupling effect tends to disappear. (The rest of the text appears to be incomplete and requires further context.) The physical meaning of this term is: the amplification factor of additional energy consumption caused by the extension of the low-temperature, high-energy-consumption stage by low speed.

[0079] For gasoline vehicles, the high-energy-consumption state specifically refers to the warm-up phase from engine cold start to when the coolant temperature reaches normal operating temperature (approximately 85°C to 95°C), and its coupling amplification function is:

[0080] in, The ratio of energy consumption during the warm-up phase to that during normal operation is determined through bench cold start tests, with a typical range of 1.3 to 1.8. The length of the warm-up time window needs to be calibrated separately through bench tests under different combinations of temperature and vehicle speed.

[0081] For electric vehicles, the high-energy-consumption state specifically refers to the heating phase from low-temperature start-up to when the battery temperature reaches a suitable operating temperature (approximately 15°C to 35°C) and the cabin temperature reaches a comfortable temperature (approximately 18°C ​​to 22°C). Let the PTC heating power be... The driving power is The ratio of energy consumption rate during the heating phase to energy consumption rate under normal conditions is... for:

[0082] in, The typical value for PTC heating power in an economy electric vehicle without a heat pump is 3 to 6 kW. The driving power, which varies with vehicle speed v, can be estimated using a vehicle dynamics model. Substituting the appropriate values... The expression yields the electric vehicle coupling amplification function as follows:

[0083] In this expression, The ratio of heating power to driving power reflects the ratio of vehicle speed to driving power. The lower the vehicle speed and the lower the driving power, the larger this ratio is, and the more prominent the proportion of heating energy consumption. This reflects the percentage of the total distance traveled during the heating phase. The product of these two parts represents the additional energy consumption of the electric vehicle due to heating. For electric vehicles equipped with heat pumps, the heat pump's coefficient of performance (COP) is typically 2 to 3, equivalent to... The coupling effect is significantly smaller than that of PTC models, thus the effect is mitigated in higher-spec models.

[0084] 3. Calibration of parameters related to coupling terms and Calibration: Based on bench tests, under different ambient temperatures and different test cycles (representing different average vehicle speeds), record the time from start-up to coolant temperature reaching 85°C (fuel vehicle) or the time from battery minimum temperature reaching 15°C and cabin temperature reaching 20°C (electric vehicle).

[0085] Complete calibration of the coupling strength coefficient: The above-mentioned coupling correction model is a theoretical derivation. To scale it to the actual scale, a coupling strength coefficient λ is introduced. The calibration method is as follows: real-world travel segments with "low temperature (T<0℃) + congestion (v<40km / h) + different travel lengths" are selected from the on-board diagnostic (OBD) measured data. The measured energy consumption of each segment is calculated. Compared with the expected energy consumption after independent correction only The ratio of the two values, with the excess representing the measured coupling effect of that segment, is then compared with the theoretical value. The values ​​are fitted using regression to obtain the calibration value of λ. The physical meaning of λ is: the collective correction factor for the coupling effect of unmodeled factors other than the travel length L (such as differences in thermal management strategies during warm-up, differences in warm-up behavior among different drivers, etc.) on the coupling effect. Ideally, λ should be close to 1.0, indicating that the theoretical model is in high agreement with reality; if λ deviates significantly from 1.0, it means that the scaling of the coupling term in the formula needs additional adjustment.

[0086] 4. Real-time correction of calculations and final output During the execution phase, the ambient temperature T, traffic state parameter v, and travel length L of the current target road segment are obtained. The current contribution value of the coupling effect is determined using the calibrated functional relationship. Mix the value 1 with... The coupling amplification factor is obtained by adding them together (or by introducing λ and setting it as λ). Multiplying the baseline energy consumption factor, the low-temperature single-condition correction factor, the congestion single-condition correction factor, and the coupling amplification factor together, we obtain the complete expression for the corrected energy consumption factor:

[0087] in, Energy consumption factor under baseline operating conditions; This is a correction factor for a single low-temperature operating condition. λ is the correction factor for a single congestion condition; λ is the coupling strength coefficient, which is obtained through regression calibration of measured data.

[0088] The mathematical form of the above-mentioned coupling correction model is exactly the same for both gasoline-powered and electric vehicles; the only difference lies in the source of the input parameter values. still The specific value of λ and The calibration method. At the software implementation level, the same function body can be used to control different parameter tables and sub-functions through the "energy type" parameter, so as to achieve a true one-code adaptation for two types of vehicles.

[0089] After incorporating this scheme, the overall method achieves a leap in accuracy at the energy consumption correction level, moving from "semi-blind fixed product" to "dynamic adaptation based on road conditions." (Coupled amplification function) The coupling amplification effect of low temperature and congestion was explicitly modeled. The travel length L parameter was used to link the ratio of the duration of the high energy consumption stage to the total travel distance. Short road sections were automatically subject to strong coupling correction, while long road sections were automatically reduced to the level of independent correction.

[0090] In one optional embodiment of this application, for the first type of road segment, the traffic flow of the first type of road segment is determined by a statistical inference model based on time-dimensional prior information, spatial-dimensional prior information, and physical conservation constraints, including: Select historical dates and time periods that have the same traffic characteristics as the target date and the target time period, wherein the same traffic characteristics include at least the same or similar holiday types and weekday attributes; The traffic flow of the first type of road segment under each of the selected historical dates and historical time periods is obtained to form a historical traffic flow sample set; Calculate the weighted average of each sample in the historical traffic sample set as the mean of the time-dimensional prior information, wherein the sample whose time is closer to the target date is assigned a greater weight; Calculate the weighted variance of each sample in the historical traffic sample set as the variance of the prior information in the time dimension; Identify the target upstream road segment with direct traffic observation data that is adjacent to the first type of road segment in the road network topology; Obtain the measured traffic flow of the upstream road segment of the target during the current target time period; The mean of the spatial dimension prior information is calculated based on the pre-established flow transmission relationship from the upstream road segment of the target to the first type of road segment, which is based on historical data. The historical fitting error of the flow transmission relationship is obtained as the variance of the spatial dimension prior information; Obtain the number of samples in the historical traffic sample set, and determine the fusion weight based on the number of samples, wherein the fusion weight increases monotonically with the increase of the number of samples, and the value ranges from 0 to 1; Multiply the fusion weight by the mean of the time-dimensional prior information to obtain the first intermediate value; Multiply the difference between 1 and the fusion weight by the mean of the spatial dimension prior information to obtain the second intermediate value; Add the first intermediate value to the second intermediate value to obtain the mean of the joint prior distribution; The third intermediate value is obtained by multiplying the square of the fusion weight by the variance of the time-dimensional prior information. The fourth intermediate value is obtained by multiplying the square of the difference between 1 and the fusion weight by the variance of the spatial dimension prior information. Add the third intermediate value to the fourth intermediate value to obtain the variance of the joint prior distribution; The vehicle physical conservation relationship of the road network topology between the first type of road segment and adjacent road segments is used as the likelihood constraint of the statistical inference model. Based on the mean and variance of the joint prior distribution and the likelihood constraint, the posterior distribution of the traffic flow of the first type of road segment is determined by a Bayesian inference model, and the mean of the posterior distribution is determined as the traffic flow of the first type of road segment.

[0091] Furthermore, the method also includes: Based on the road network topology, identify a second type of road segment in the target area, wherein the topological hop count of the second type of road segment from the road segment with direct traffic observation data is greater than that of the first type of road segment; For the second type of road segment, the upper quantile of the posterior distribution of its traffic flow is calculated as a conservative estimate of the traffic flow of the second type of road segment based on the posterior distribution of its traffic flow. The conservatively estimated traffic flow is used as the traffic flow for the second type of road segment, and a risk label is generated for the second type of road segment.

[0092] Specifically, the following provides a detailed explanation of each step in Bayesian inference: 1. Classification of checkpoint coverage levels along road sections Using GIS road network segments as the smallest calculation unit, the road network topology distance from each segment to the nearest checkpoint is calculated. The so-called road network topology distance refers to the number of road segments required to reach the nearest road segment with a checkpoint by following the road network topology (without skipping ramps or crossing intersections) from the current road segment, and is expressed in "jumps".

[0093] The entire road network is divided into three coverage levels based on topological distance. Level 1 coverage (0 hops): Checkpoints are directly located on the road segment itself or at its origin and destination. Traffic flow can be directly obtained from the checkpoints through records, requiring no inversion, and has the highest confidence. Level 2 coverage (1 to 2 hops): Checkpoints are nearby but not directly deployed. Bayesian inversion can be used, with moderate confidence. Level 3 coverage (more than 2 hops): The road segment is far from the nearest checkpoint, prior information sources are sparse, and inversion uncertainty is high. A conservative estimation strategy should be adopted. The physical basis for this classification is that checkpoints provide "precise but sparse" observations. The closer to the checkpoint, the stronger the constraint of precise observation and the smaller the uncertainty of traffic estimation; the farther from the checkpoint, the weaker the constraint and the greater the uncertainty.

[0094] The "Category 1 road segments" processed in this step include Class II coverage road segments (nearby road segments without checkpoints with 1 to 2 hops), while the "Category 2 road segments," namely Class III coverage road segments (extremely distant road segments without checkpoints with more than 2 hops), need to be processed using the conservative estimation strategy described later. Traffic flow in Class I coverage road segments is obtained directly from actual checkpoint measurements and is not included in the inference scope.

[0095] 2. Construction of temporal priors For each second-class coverage road segment and each target time period, construct a Gaussian prior distribution of the traffic flow Q for that road segment during that time period. The construction scheme for the time-dimensional prior is as follows.

[0096] First, a set of historically similar dates is selected. The definition of similar days is based on the following rules: if the target date is a holiday type (e.g., the third day of the Lunar New Year), then the same day of that holiday type and one day before and after it are selected from each of the past three years; if the target date is a regular weekday, then weekdays of the same week within the past four weeks are selected; if the target date is a regular weekend, then weekends of the same week within the past four weeks are selected. If the available sample size of similar days is insufficient (e.g., 5 days), the similarity criteria are gradually relaxed, first expanding to the same period of the same month, then to the same period of the same season, ensuring at least a minimum sample size of historical data to support the date.

[0097] For the selected set of similar days, extract the traffic flow of the road segment during the corresponding time periods on those historical days to form a historical traffic flow sample set. Then, categorize each element in the sample set according to the number of calendar days between its current date and the target date. Weighting is applied, with the weights using an exponential time decay method:

[0098] in, Let τ be the number of calendar days (in days) between the k-th similar day and the target day. τ is the decay time constant, which physically represents the characteristic scale of the decay of historical data influence over time. The smaller the value of τ, the faster the decay of distant data. A value of 365 days means that the weight of data from one year ago decays to the power of e (approximately 0.37). The physical meaning of this weighting method is that historical data closer to the target day in time better reflects the current actual traffic status of the road network, because the infrastructure, population distribution, and travel habits of the road network slowly evolve over time. The exponential decay form ensures the "nearer, larger; farther, smaller" weight allocation principle while avoiding the complete discarding of more distant historical data.

[0099] The weighted average of each historical sample is calculated using its normalized weights and used as the mean of the time-dimensional prior. Calculate the weighted variance as the variance of the time dimension prior. The larger the variance, the greater the interannual fluctuation of traffic flow in that road segment during that period, and the higher the prior uncertainty.

[0100] 3. Construction of Spatial Dimensional Priors For target road segment i, find its nearest upstream first-class coverage road segment j with direct checkpoint coverage (i.e., the traffic flow of road segment j can be obtained from actual checkpoint measurements). In the historical database, extract traffic flow pairing data for road segment j and road segment i over hundreds of historical time periods. Using least squares linear regression, establish the traffic flow transfer relationship from road segment j to road segment i:

[0101] in, The transmission coefficient, in its physical sense, represents the proportion of vehicles that pass through road segment j that will continue through road segment i. If there is no ramp diversion, α should be close to 1.0; if there is ramp diversion, then α is less than 1.0. This is a constant term, and its physical meaning is the basic flow that newly enters from the ramp and has no linear dependence on the flow of road segment j. Let be the random error term, which is assumed to follow a normal distribution with zero mean. , The traffic flows for road segment i and road segment j are respectively.

[0102] During the current target time period, read the actual measured traffic flow at the checkpoint of road segment j. Substituting this measured value into the previously defined transitive relation, we obtain the mean prediction value of the spatial prior. The variance of the spatial prior is taken as the variance of the historical fit residuals of the regression model. .

[0103] 4. Dual-source adaptive fusion prior The temporal and spatial priors are fused into a joint prior distribution using an adaptive weighting method. The fusion weight γ is based on the effective number of historical similar day samples. Dynamically determined, a preferred calculation formula is:

[0104] in, The number of valid historical similar day samples actually available when constructing the time-dimensional prior (e.g., the number of valid historical days selected when constructing the time-dimensional prior). This is a preset half-saturation constant; the example value is 5 days, indicating that when... At that time, the time dimension and the space dimension each account for approximately 50% of the weight. Much larger hour, Approaching 1, the joint prior primarily trusts the periodic information of the time dimension; when much smaller hour, Approaching 0, the joint prior primarily trusts the spatial topology of the space to transmit information; when At that time, the weights of the two are equal. The value range naturally falls between 0 and 1, depending on... Monotonically increasing.

[0105] Based on the fusion weight γ, the mean and variance of the joint prior are obtained:

[0106]

[0107] This fusion strategy ensures that periodic patterns are fully utilized to improve inference accuracy in road sections with abundant historical data, while spatial physical constraints are used as reliable anchor points in road sections with scarce historical data, and it has universal robustness for road sections with different data abundance.

[0108] 5. Physically Conserved Likelihood Constraints and Bayesian Posterior Derivation For target road segment i and its upstream adjacent road segment i-1, if there are no ramps between them (i.e., no vehicles merging in or out), then according to the principle of vehicle conservation, the traffic flow of the two road segments should be approximately equal. This physical fact can be expressed as a probability constraint in probabilistic form:

[0109] in, This is a preset minimum variance value; the smaller the variance, the higher the reliability of the conservation law. In the example... Taking 1% of the square of the average flow estimate of adjacent road segments reflects a high degree of confidence in the conservation law. This constraint forces the flow estimates of adjacent road segments to converge during Bayesian inference unless there is sufficiently strong data evidence to support their divergence.

[0110] If there is a ramp merging or merging between road segment i and i-1, the above conservation constraint needs to be modified as follows:

[0111] in, and These are estimated values ​​for the inflow and outflow of traffic from the ramps, which can be calculated based on the historical ratio of ramp traffic to mainline traffic and the traffic flow of upstream sections.

[0112] Thus, for each second-class coverage segment in each target time period, the constructed joint prior distribution (Gaussian distribution) is integrated with the aforementioned conserved likelihood constraint (also Gaussian distribution). Since both the prior and likelihood follow normal distributions, based on the favorable property of Gaussian-Gaussian conjugate, the posterior distribution is also Gaussian, and its posterior mean and posterior variance have the following analytical solutions, requiring no numerical optimization iteration:

[0113]

[0114] In the above formula, For the joint prior mean, The joint prior variance (obtained from the dual-source fusion above). The likelihood mean is taken as the upstream traffic flow when there are no ramps. Take when there is a ramp . To conserve the likelihood variance, we take the magnitude of the square of 1% of the average flow of adjacent road segments. The mean is the Bayesian posterior mean, which serves as the optimal estimate of the traffic flow for this road segment during this time period. The posterior variance serves as a measure of the uncertainty of this estimate.

[0115] The posterior mean is a weighted average of the precision (reciprocal of the variance) of the prior guess and the conserved observations—the higher the precision of the information source, the greater its influence on the posterior estimate. Because The value is deliberately set to a minimum, resulting in extremely high accuracy. Therefore, the posterior estimate is strongly guided by conservation constraints, forcing the traffic flow of adjacent road segments to align, thus maintaining physical consistency in the link topology. The posterior variance is necessarily less than the variance of any single source, indicating that the uncertainty of the estimate is reduced by incorporating more information.

[0116] It should be noted that in the Bayesian inference framework of this embodiment, the use of each data source is strictly functionally layered to avoid the problem of the same observation information being repeatedly included in the prior and likelihood estimates, which could lead to underestimation of posterior variance and overestimation of confidence levels. Specifically: The construction of the time-dimensional prior distribution relies solely on traffic flow observation samples of the first type of road segment during historical periods. These historical samples are derived from traffic flow data obtained and verified by the road segment in the past through methods such as fixed detectors, temporary surveys, ETC route reconstruction, video recognition, or floating car penetration expansion, reflecting the periodic patterns and long-term statistical characteristics of traffic flow in the road segment.

[0117] The spatial dimension observation equation is constructed using the current measured traffic flow of road segments with direct observation data during the target time period (i.e., the road segments where checkpoints are located). These measured traffic flows reflect the actual traffic demand during the target time period and are correlated with the estimated traffic flow of the first type of road segments through spatial transfer relationships. The specific form of the spatial transfer relationship can be established based on the traffic flow conversion relationship between adjacent road segments in the road network topology. Its calibration process only uses road segment pairs with complete observations from historical data for fitting, without introducing the observation data of the target time period into the input side of the calibration set.

[0118] Physical constraints are independently constructed based on the traffic flow conservation relationships at road network nodes. For example, the sum of the flow rates of all road segments flowing into an intersection equals the sum of the flow rates of all road segments flowing out of the intersection (a correction term can be added when considering ramp merging and merging). Conservation constraints exist as deterministic or probabilistic equations independent of the aforementioned prior and observation equations. They do not depend on the current observations of a specific road segment and only reflect the physical topological properties of the road network itself.

[0119] Through the aforementioned layering, the Bayesian model in this embodiment ensures that each piece of observation data is used only at one level throughout the entire inference process: historical data only affects the temporal prior, current checkpoint data only affects the spatial observation equation, and the physical conservation relationships of the road segment itself constitute independent constraints. These three elements constitute a joint posterior, which is computationally fused using the Bayesian formula, preventing the problem of double-weighting of evidence caused by the same upstream measured flow being used both to define the spatial prior and to construct the likelihood equation.

[0120] In certain cases, if data sparsity necessitates the reuse of some current observations when constructing spatial observation equations—for example, if checkpoint data serves as both upstream input to the target road segment and nodal verification for the conservation equation—a corresponding observation error correlation structure should be established. This involves explicitly introducing correlation parameters between these current observations into the covariance matrix of the observation equations. These correlation parameters can be pre-calibrated using road segments with repeated observation conditions from historical data. This allows the model to reasonably reflect the internal dependencies of these observations, preventing them from being mistakenly treated as independent evidence.

[0121] 6. Conservative estimates for Category II road sections and tiered outputs for Category I road sections For third-class coverage road sections (i.e., second-class road sections with a topological hop count greater than 2), the variance of the spatial prior is naturally larger due to the extended error propagation chain, and the effective historical sample size of the temporal prior is usually insufficient. This results in a large joint prior variance and an excessively wide confidence interval for the posterior estimate. Directly using the posterior mean as a point estimate carries the risk of unreliability.

[0122] A conservative estimation strategy is adopted for this type of road segment. The posterior mean is not output. Instead, it outputs the 95th quantile of the posterior distribution as a conservative estimate of the flow:

[0123] Wherein, 1.645 is the 95th percentile Z-value of the standard normal distribution, which physically means that the actual flow rate is lower than... The probability is approximately 95%. For the posterior standard deviation, take the posterior variance. The positive square root. Simultaneously, a prominent risk label is added to the output for this road segment, distinguishing it from the regular labels of road segments with higher confidence levels.

[0124] For Class II coverage sections (the portion of Class I sections closest to the checkpoint), directly output the posterior mean. As a point estimate of flow, and output As a numerical reference for the confidence level.

[0125] 7. Integration of vehicle-type traffic flow breakdown with energy consumption calculation The total traffic flow of each road segment and time period obtained from the above inversion Multiplied by the final output of the hybrid vehicle structure ratio The vehicle flow rate by energy type is obtained: Traffic diversion for electric vehicles:

[0126] Traffic diversion for fuel-powered vehicles:

[0127] Among them, for the second-class coverage sections Take the posterior mean For third-class coverage road sections Take a conservative estimate The results of this breakdown can be directly input into the energy consumption calculation module.

[0128] Understandably, after determining the corrected energy consumption factor and traffic flow for each road segment, this embodiment performs a segment-by-segment summary calculation of the total energy consumption of the target area during the target time period. Unlike methods that only target a single type of road segment, this embodiment's summary calculation covers all target road segments within the target area and processes them according to the source type of traffic flow data for each road segment: For road segments within the target area with direct traffic flow observation data, the measured traffic flow of that road segment during the target time period is directly used as the calculation input. These road segments are typically equipped with checkpoint devices, loop detectors, or other real-time traffic monitoring devices, and their flow data can accurately reflect the number of vehicles actually passing through that section during the target time period.

[0129] For the first type of road segment, which is a road segment that does not have direct observations during the target time period but has verified traffic flow observation data during historical time periods, the posterior traffic flow estimate obtained through Bayesian inference is used as the calculation input.

[0130] For the second type of road segment, namely the long-distance unobserved road segment, the traffic flow calculated based on the estimated traffic flow of the first type of road segment and the upstream and downstream transmission relationship of the road network topology is used as the calculation input.

[0131] After obtaining the traffic flow data for each road segment within the target area, the energy consumption value is calculated segment by segment, taking into account the vehicle type composition, energy consumption correction factor, and road segment length. The vehicle type composition is used to decompose the total traffic flow of a road segment into fuel-powered vehicle traffic and electric vehicle traffic. The energy consumption correction factor reflects the energy consumption per unit mileage of the road segment under the current ambient temperature, traffic conditions, and vehicle thermal conditions. The road segment length is used to convert the unit energy consumption into the total energy consumption of the road segment. The sum of the energy consumption of all road segments is the total energy consumption of the target area during the target time period.

[0132] By summarizing the data segment by segment, category by category, and with full coverage as described above, this embodiment can fully cover the energy consumption contribution of all target road segments within the target area, avoiding systematic underestimation or omission of regional energy consumption due to using only the calculation results of a single type of road segment.

[0133] After incorporating this approach, the overall method achieves a triple improvement in traffic acquisition capabilities. First, information utilization is significantly enhanced. Simultaneously, by utilizing constraints of three different properties—time-dimensional periodicity, spatial-dimensional topological transitivity, and physical conservation laws—and fusing them through a Bayesian framework, it improves both accuracy and robustness compared to methods relying solely on a single information source. Second, the adaptive fusion weight design (automatically adjusting γ based on the effective historical sample size K) allows the model to fully trust periodic patterns in data-rich road segments and regress the hard constraints of spatial topology in data-scarce road segments, providing universal adaptability to road segments with varying data abundance across the entire road network. Third, uncertainty is transformed from a "hidden risk" into a "visible and manageable planning parameter," with improved posterior variance. The t quantifies the confidence level of the inference results for each road segment. For the second type of road segment, the 95th percentile is used as a conservative estimate and a risk label is added. This allows planners to configure the safety capacity margin of energy replenishment facilities differently according to the confidence level of each road segment. This avoids the dilemma of investment waste and omission of weak links caused by treating all road segments the same and blindly reserving the same safety factor in the traditional scheme.

[0134] In one optional embodiment of this application, the energy consumption of the target area is calculated based on the traffic flow of the first type of road segment, the proportion of the mixed vehicle type structure, and the corrected energy consumption factor, including: Receive calculation instructions for long-cycle prediction scenarios or short-cycle calculation scenarios. A first set of input parameters is configured for the long-term prediction scenario. The ambient temperature in the first set of input parameters is taken as the historical statistical average value, the traffic state parameter is taken as the parameter value representing the smooth flow state, and the vehicle structure ratio is taken as the first vehicle structure ratio. A second set of input parameters is configured for the short-cycle calculation scenario. The ambient temperature in the second set of input parameters is taken as the historical statistical extreme value, the traffic state parameter is taken as the parameter value representing the congestion state, and the vehicle structure ratio is taken as the ratio of the mixed vehicle structure. The unified underlying computing kernel is invoked to calculate the first energy consumption result of the target area in the long-term prediction scenario and the second energy consumption result in the short-term calculation scenario, based on the input traffic flow of the first type of road segment, the correction energy consumption factor, and the vehicle type structure ratio of the corresponding scenario. Output the first energy consumption result in an annual trend format, and output the second energy consumption result in an hourly distribution format.

[0135] Specifically, the final step of calculating the energy consumption of the target area is concretized into a parallel simulation scheme with long and short cycles based on a three-layer separation architecture. The so-called three-layer separation means that the entire simulation system is divided into: the bottom layer - a unified energy consumption accumulation calculation kernel; the middle layer - a separate parameter configuration layer, responsible for preparing differentiated input parameter sets for different scenarios; and the top layer - a separate output layer, responsible for rendering and distributing results according to the scenario format.

[0136] 1. Underlying Layer – Unified Computing Kernel The underlying computational kernel is a pure energy consumption accumulation function, whose input is a complete set of parameters, including road segment number i, time period t, traffic flow Q, and correction energy consumption factor. and the proportion of hybrid vehicle models The function outputs the energy consumption by energy type for that road segment during that time period. It does not contain any scenario judgment or parameter selection logic; it only performs deterministic calculations based on given parameters and returns the results. Its core processing flow is as follows: First, the total traffic flow is multiplied by the proportion of electric vehicles and fuel vehicles in the mixed vehicle structure to obtain the traffic flow by energy type; then, each traffic flow is multiplied by its corresponding corrected energy consumption factor and the physical length of each road segment. The system obtains the energy consumption value of the road segment during the specified time period; finally, it returns the calculation results for the upper-level module to accumulate and format.

[0137] 2. Middle layer – Separate parameter configuration layer The middle-level parameter configuration layer is responsible for preparing input parameters for the underlying kernel. The parameter configuration logic of long-cycle links is completely different from that of short-cycle links, but the output parameter format is exactly the same so that the kernel can call them uniformly.

[0138] (1) Parameter configuration logic of long-period links The ambient temperature T is taken as the historical average temperature of the winter (December to February) of the target year. The physical basis is that long-term planning focuses on the average energy consumption level over many years, rather than the peak value of individual extreme days. On an annual scale, energy storage facilities and grid layout are driven by average energy density, and short-term peaks can be absorbed by energy storage buffers. Therefore, long-term parameters should reflect the central trend of statistics. The traffic state parameter v is taken as the free-flow vehicle speed, typically 0.85 to 0.90 times the design speed of the road segment. The physical basis is that, on an annual average basis, highways are mostly uncongested, with congestion occurring only during a few peak hours, thus having a limited weight in contributing to the total annual energy consumption. The travel length L is taken as the physical length of the road segment. The physical basis is that the total travel distance of vehicles passing through this road segment is usually much greater than the length of a single road segment; approximating the travel length with the road segment length is reasonable on an annual average basis. The vehicle type structure ratio only takes the ratio of the first vehicle type. Tidal disturbances are excluded because they only occur during a few holidays throughout the year and their contribution to total annual energy consumption is usually less than 3%, which can be ignored from a long-term planning perspective. Scenario variables are adjusted for the proportion of electric vehicles in future years according to electrification policy scenarios (such as aggressive, neutral, or conservative), with different scenarios corresponding to different policy intensities.

[0139] (2) Parameter configuration logic for short-cycle links The ambient temperature T is taken as the historical low temperature extreme value for the same period during the target holiday, such as the lowest 10th percentile temperature of the Spring Festival holiday in the past five years. The physical basis is that the goal of short-cycle links is to measure "peak energy consumption," the maximum load that the energy replenishment facilities must withstand under extreme conditions. Peak energy consumption occurs under the most severe combination of conditions (extremely low temperature + congestion + peak traffic flow), therefore extreme parameters should be used rather than average parameters. The traffic state parameter v is taken as the historical measured low vehicle speed during the peak period of the target holiday, such as the measured average vehicle speed during the hour of maximum traffic flow, also to capture peak conditions. The travel length L is estimated based on the license plate trajectory reconstruction technology at checkpoints to estimate the average travel length of vehicles on the road segment. The physical basis is that under extreme conditions, a high proportion of short-travel vehicles will exacerbate the coupling effect, and the travel length directly affects the strength of the coupling correction, so it should be accurately captured. The vehicle type structure ratio is taken as the mixed vehicle type structure ratio superimposed with tidal disturbances. This is because short-period scenarios are precisely when tidal disturbances are strongest. The coupling correction parameters use the extreme condition parameter set from Module 2 (λ takes the upper limit value, ...). (Taking the longest value, etc.) to ensure that the peak value calculation is not overly optimistic.

[0140] 3. Upper layer – Separate output layer The long-cycle output layer annually extrapolates the average daily energy consumption density of each road segment under three electrification scenarios (such as aggressive, neutral, and conservative policy scenarios), generating trend curves. It also identifies key time points, such as "the year when the energy consumption of electric vehicles on a certain road segment first exceeds 50%," indicating that large-scale deployment of charging facilities should be completed on that road segment before that year. The short-cycle output layer outputs the instantaneous energy consumption (kWh / h) of each road segment hourly for target holidays, generating 24 frames of energy consumption heatmaps using a geographic information system, with color temperature indicating energy consumption levels. The heatmaps are overlaid with the calculation reliability level of each road segment (distinguished by different line types or transparency according to the aforementioned first to third level coverage), and the second-class road segments (third-level coverage segments) are marked with specific color borders or symbols as "calculation risk hotspots."

[0141] By incorporating this solution, the overall approach achieves the system architecture advantages of "one core for multiple uses, long-term and short-term synergy." First, it breaks the traditional deadlock of two independent systems. Long-term forecasts and short-term calculations share the same underlying computational logic, achieving scenario differentiation only by switching parameter layers. This eliminates the need to maintain two codebases, two databases, and two operation and maintenance teams, reducing system implementation and long-term operation and maintenance costs. Second, it ensures the physical consistency between long-term trends and short-term peak results. Because they share the same kernel and are based on the same physical model (coupling correction and Bayesian inference), the two types of outputs can corroborate each other and maintain consistency in spatial hotspots and temporal trends—something that two completely different models cannot inherently guarantee. Third, it provides a unified data foundation for integrated planning and operation. Long-term forecasts identify the strategic direction with the fastest future energy replenishment demand growth, while short-term calculations provide peak load heat maps and risk road segment lists for the next holiday. Planners can strategically deploy energy replenishment sites from decades in the long term, while precisely configuring the maximum instantaneous backup capacity of each site from days to hours in the short term, achieving true full lifecycle planning and scheduling optimization of photovoltaic energy storage and replenishment facilities.

[0142] Figure 2 This application provides a structural block diagram of a hybrid traffic flow dynamic energy consumption measurement device, as shown below. Figure 2 As shown, the device includes: The vehicle type structure fusion module 201 is used to obtain the first vehicle type structure ratio of the first type of vehicles and the second vehicle type structure ratio of the second type of vehicles passing through the target area during the target time period; and to fuse the first vehicle type structure ratio and the second vehicle type structure ratio with the traffic flow of the first type of vehicles and the second type of vehicles as weights to obtain the mixed vehicle type structure ratio of the target area during the target time period. The energy consumption factor correction module 202 is used to correct the baseline energy consumption factor based on ambient temperature, traffic state parameters and specific travel distance parameters of the vehicle on the target road segment to obtain the corrected energy consumption factor of the target road segment; wherein, the correction includes independent correction based on the ambient temperature and the traffic state parameters, and coupled correction based on the coupling relationship between the ambient temperature, the traffic state parameters and the specific travel distance parameters. The traffic flow inversion module 203 is used to identify a first type of road segment in the target area based on the road network topology, where the first type of road segment does not have direct traffic observation data; and for the first type of road segment, based on time-dimensional prior information, spatial-dimensional prior information and physical conservation constraints, to determine the traffic flow of the first type of road segment through a statistical inference model. The energy consumption calculation module 204 is used to calculate the energy consumption of the target area based on the traffic flow of the first type of road segment, the proportion of the mixed vehicle structure, and the correction energy consumption factor.

[0143] The proposed solution categorizes vehicle types into two groups with different data generation mechanisms: local steady-state and external tidal traffic. After obtaining the respective proportions, these groups are fused using traffic flow as the weight. This effectively solves the problem of the traditional vehicle transfer matrix's closed state space, which cannot accommodate the instantaneous injection of external vehicles, thus improving the accuracy of vehicle type prediction in holiday tidal traffic scenarios. In the energy consumption correction stage, by introducing specific travel parameters of vehicles on the target road segment, the coupling effect of low temperature and congestion is explicitly modeled as a coupled correction term of the ambient temperature, traffic state parameters, and specific travel parameters. This breaks the assumption of independence between the two in the traditional product correction framework, accurately characterizing the physical mechanism by which congestion prolongs the duration of the low-temperature, high-energy-consumption phase, thereby improving the accuracy of energy consumption calculation under combined operating conditions. In the traffic flow acquisition stage, for the first type of road segments without direct traffic observation data, the system simultaneously utilizes historical periodic information in the time dimension, road network topology transfer relationships in the spatial dimension, and strict physical conservation constraints. A statistical inference model is used to perform optimal estimation through multi-source information fusion, thereby achieving the inversion of traffic flow from sparse checkpoint observations to the entire road network without relying on real-time roadside monitoring equipment. Through hierarchical fusion of vehicle type structure, coupled correction of energy consumption factors, and inversion inference of traffic flow, it eliminates the need for continuous deployment of traffic monitoring equipment on all road segments. Traffic flow on unobserved road segments is inferred using local traffic observations, historical traffic data, and road network topology constraints.

[0144] It should be noted that in this embodiment, vehicles in mixed traffic flow are divided into two main categories based on power type: gasoline vehicles and electric vehicles. Hybrid vehicles (including hybrid electric vehicles (HEVs) and plug-in hybrid electric vehicles (PHEVs)) are not considered as a separate third category. In actual implementation, if floating car samples or checkpoint identification data within the target area can provide detailed statistical information for HEVs or PHEVs, their energy consumption contribution can be approximated as follows to obtain more accurate calculation results: the energy consumption factor for HEVs is calculated using the corresponding parameters for gasoline vehicles; for PHEVs, their traffic flow can be proportionally split based on the average electric drive mileage ratio of PHEV models in the regional statistical data, and then classified into electric and gasoline vehicle categories for corresponding calculations. This processing method maintains the existing framework, only requiring the introduction of optional sub-parameters, and has strong compatibility and scalability.

[0145] In one optional embodiment of this application, obtaining the proportion of the second vehicle type structure for the second type of vehicles includes: Extract the passage data of the second type of vehicles from historical passage records for multiple historical target time periods. Each passage data includes at least the vehicle's place of origin, the checkpoint location when entering the target road network, the time of passing through the current checkpoint, and the direction of travel. Based on the stated place of origin, determine the vehicle's origin classification; Based on the mileage difference between the checkpoint location where the vehicle was initially captured when it entered the target road network and the current checkpoint location, the distance the vehicle has traveled is estimated, and the total travel distance is calculated based on the distance traveled and the direction of travel, and the total travel distance is divided into a preset travel distance range. Based on the time of passing through the current checkpoint, determine the type of passage period to which the current passage belongs; Multiple cross-classification categories are constructed by using the origin classification, travel distance range, travel time type, and travel direction as dimensions for cross-classification; For each cross-category, the total number of vehicles belonging to that category and the number of vehicles using a specific energy type are counted. The ratio of the number of vehicles using a specific energy type to the total number of vehicles is taken as the proportion of that energy type in that category. The proportion of the second vehicle model structure is determined based on the proportion of the energy type corresponding to one or more cross-classification categories matched during the target time period.

[0146] In one optional embodiment of this application, the baseline energy consumption factor is corrected to obtain the corrected energy consumption factor for the target road segment, including: The first unit mileage energy consumption and the second unit mileage energy consumption are obtained when the vehicle is driven under a baseline traffic condition at a first preset ambient temperature and a second preset ambient temperature; the first preset ambient temperature is normal temperature and the second preset ambient temperature is low temperature. Calculate the ratio of the energy consumption per unit mileage to the energy consumption per unit mileage, and use it as a correction factor for low-temperature single operating conditions. The third and fourth unit mileage energy consumption of the vehicle are obtained when it is driving under the first preset ambient temperature in the first preset traffic state and the second preset traffic state; the traffic efficiency of the second preset traffic state is lower than that of the first preset traffic state. The ratio of the energy consumption per fourth unit mileage to the energy consumption per third unit mileage is calculated and used as a correction factor for a single congestion operating condition. Acquire multiple sets of measured energy consumption data of the vehicle under the second preset ambient temperature and the second preset traffic conditions at different travel lengths; Divide each set of measured energy consumption data by the product of the baseline energy consumption factor, the low temperature single-condition correction factor, and the congestion single-condition correction factor, and determine the portion of the resulting ratio that exceeds 1 as the coupling effect contribution value under that travel length. Establish a functional relationship between the contribution value of the coupling effect and ambient temperature, traffic state parameters and travel length, wherein the contribution value of the coupling effect decreases as the travel length increases; The ambient temperature, traffic condition parameters, and travel length of the current target road segment are obtained, and the current coupling effect contribution value is determined using the aforementioned functional relationship. The coupling amplification factor is obtained by adding 1 to the current coupling effect contribution value. The corrected energy consumption factor for the target road segment is obtained by multiplying the baseline energy consumption factor, the low-temperature single-condition correction factor, the congestion single-condition correction factor, and the coupling amplification factor.

[0147] In one optional embodiment of this application, for fuel-powered vehicles, the first preset ambient temperature is normal temperature, and the second preset ambient temperature is low temperature; the reference traffic state is high-speed constant speed driving state. For electric vehicles, the first preset ambient temperature is normal temperature, and the second preset ambient temperature is low temperature; the reference traffic state is high-speed constant speed driving state.

[0148] In one optional embodiment of this application, for the first type of road segment, the traffic flow of the first type of road segment is determined by a statistical inference model based on time-dimensional prior information, spatial-dimensional prior information, and physical conservation constraints, including: Select historical dates and time periods that have the same traffic characteristics as the target date and the target time period, wherein the same traffic characteristics include at least the same or similar holiday types and weekday attributes; The traffic flow of the first type of road segment under each of the selected historical dates and historical time periods is obtained to form a historical traffic flow sample set; Calculate the weighted average of each sample in the historical traffic sample set as the mean of the time-dimensional prior information, wherein the sample whose time is closer to the target date is assigned a greater weight; Calculate the weighted variance of each sample in the historical traffic sample set as the variance of the prior information in the time dimension; Identify the target upstream road segment with direct traffic observation data that is adjacent to the first type of road segment in the road network topology; Obtain the measured traffic flow of the upstream road segment of the target during the current target time period; The mean of the spatial dimension prior information is calculated based on the pre-established flow transmission relationship from the upstream road segment of the target to the first type of road segment, which is based on historical data. The historical fitting error of the flow transmission relationship is obtained as the variance of the spatial dimension prior information; Obtain the number of samples in the historical traffic sample set, and determine the fusion weight based on the number of samples, wherein the fusion weight increases monotonically with the increase of the number of samples, and the value ranges from 0 to 1; Multiply the fusion weight by the mean of the time-dimensional prior information to obtain the first intermediate value; Multiply the difference between 1 and the fusion weight by the mean of the spatial dimension prior information to obtain the second intermediate value; Add the first intermediate value to the second intermediate value to obtain the mean of the joint prior distribution; The third intermediate value is obtained by multiplying the square of the fusion weight by the variance of the time-dimensional prior information. The fourth intermediate value is obtained by multiplying the square of the difference between 1 and the fusion weight by the variance of the spatial dimension prior information. Add the third intermediate value to the fourth intermediate value to obtain the variance of the joint prior distribution; The vehicle physical conservation relationship of the road network topology between the first type of road segment and adjacent road segments is used as the likelihood constraint of the statistical inference model. Based on the mean and variance of the joint prior distribution and the likelihood constraint, the posterior distribution of the traffic flow of the first type of road segment is determined by a Bayesian inference model, and the mean of the posterior distribution is determined as the traffic flow of the first type of road segment.

[0149] In one optional embodiment of this application, the method further includes: Based on the road network topology, identify a second type of road segment in the target area, wherein the topological hop count of the second type of road segment from the road segment with direct traffic observation data is greater than that of the first type of road segment; For the second type of road segment, the upper quantile of the posterior distribution of its traffic flow is calculated as a conservative estimate of the traffic flow of the second type of road segment based on the posterior distribution of its traffic flow. The conservatively estimated traffic flow is used as the traffic flow for the second type of road segment, and a risk label is generated for the second type of road segment.

[0150] In one optional embodiment of this application, the energy consumption of the target area is calculated based on the traffic flow of the first type of road segment, the proportion of the mixed vehicle type structure, and the corrected energy consumption factor, including: Receive calculation instructions for long-cycle prediction scenarios or short-cycle calculation scenarios. A first set of input parameters is configured for the long-term prediction scenario. The ambient temperature in the first set of input parameters is taken as the historical statistical average value, the traffic state parameter is taken as the parameter value representing the smooth flow state, and the vehicle structure ratio is taken as the first vehicle structure ratio. A second set of input parameters is configured for the short-cycle calculation scenario. The ambient temperature in the second set of input parameters is taken as the historical statistical extreme value, the traffic state parameter is taken as the parameter value representing the congestion state, and the vehicle structure ratio is taken as the ratio of the mixed vehicle structure. The unified underlying computing kernel is invoked to calculate the first energy consumption result of the target area in the long-term prediction scenario and the second energy consumption result in the short-term calculation scenario, based on the input traffic flow of the first type of road segment, the correction energy consumption factor, and the vehicle type structure ratio of the corresponding scenario. Output the first energy consumption result in an annual trend format, and output the second energy consumption result in an hourly distribution format.

[0151] Figure 3 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 3 As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute a method for calculating the dynamic energy consumption of mixed traffic flow. This method includes: obtaining the proportion of a first type of vehicle type and the proportion of a second type of vehicle type passing through the target area during a target time period; using the traffic flow of the first type of vehicle and the second type of vehicle as weights, fusing the proportions of the first and second types of vehicle types to obtain the proportion of mixed vehicle types in the target area during the target time period; and correcting the baseline energy consumption factor based on ambient temperature, traffic state parameters, and specific travel distance parameters of vehicles on the target road segment to obtain the corrected energy consumption factor for the target road segment. The correction includes independent correction based on the ambient temperature and the traffic state parameters, and coupled correction based on the coupling relationship between the ambient temperature, the traffic state parameters, and the specific travel distance parameters. According to the road network topology, a first type of road segment is identified in the target area, where no direct traffic observation data is available. For the first type of road segment, traffic flow is determined using a statistical inference model based on time-dimensional prior information, spatial-dimensional prior information, and physical conservation constraints. Based on the traffic flow of the first type of road segment, the proportion of mixed vehicle types, and the corrected energy consumption factor, the energy consumption of the target area is calculated.

[0152] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0153] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the dynamic energy consumption calculation method for mixed traffic flow provided by the above methods. The method includes: obtaining the first vehicle type structure ratio of a first type of vehicle and the second vehicle type structure ratio of a second type of vehicle passing through a target area during a target time period; using the traffic flow of the first type of vehicle and the second type of vehicle as weights, fusing the first vehicle type structure ratio and the second vehicle type structure ratio to obtain the mixed vehicle type structure ratio of the target area during the target time period; and based on ambient temperature, traffic state parameters, and specific vehicle behavior on the target road segment... The driving distance parameter is used to correct the baseline energy consumption factor to obtain the corrected energy consumption factor for the target road segment. The correction includes independent correction based on the ambient temperature and traffic state parameters, and coupled correction based on the coupling relationship between the ambient temperature, traffic state parameters, and the specific driving distance parameter. According to the road network topology, a first type of road segment in the target area is identified, where no direct traffic observation data is available. For the first type of road segment, traffic flow is determined using a statistical inference model based on time-dimensional prior information, spatial-dimensional prior information, and physical conservation constraints. Based on the traffic flow of the first type of road segment, the proportion of mixed vehicle types, and the corrected energy consumption factor, the energy consumption of the target area is calculated.

[0154] Furthermore, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements a method for calculating the dynamic energy consumption of mixed traffic flow provided by the methods described above. This method includes: obtaining the proportion of a first type of vehicle type and the proportion of a second type of vehicle type passing through a target area during a target time period; using the traffic flow of the first type of vehicle and the second type of vehicle as weights, fusing the proportion of the first type of vehicle type and the proportion of the second type of vehicle type to obtain the proportion of mixed vehicle type in the target area during the target time period; and adjusting the baseline energy consumption factor based on ambient temperature, traffic state parameters, and specific travel distance parameters of vehicles on the target road segment. The correction process involves obtaining a corrected energy consumption factor for the target road segment. This correction includes independent corrections based on the ambient temperature and traffic state parameters, as well as coupled corrections based on the coupling relationship between the ambient temperature, traffic state parameters, and specific travel distance parameters. Based on the road network topology, a first type of road segment is identified within the target area, where no direct traffic observation data is available. For the first type of road segment, traffic flow is determined using a statistical inference model based on time-dimensional prior information, spatial-dimensional prior information, and physical conservation constraints. Finally, based on the traffic flow of the first type of road segment, the proportion of mixed vehicle types, and the corrected energy consumption factor, the energy consumption of the target area is calculated.

[0155] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0156] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0157] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

[0158] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for calculating the dynamic energy consumption of mixed traffic flow, characterized in that, The method includes: Obtain the proportion of the first type of vehicle type and the proportion of the second type of vehicle type in the target area during the target time period; Using the traffic flow of the first type of vehicles and the second type of vehicles as weights, the proportion of the first vehicle type structure and the proportion of the second vehicle type structure are integrated to obtain the proportion of mixed vehicle type structure in the target area during the target time period; The baseline energy consumption factor is corrected based on ambient temperature, traffic state parameters, and specific travel distance parameters of the vehicle on the target road segment to obtain the corrected energy consumption factor of the target road segment; wherein, the correction includes independent correction based on the ambient temperature and the traffic state parameters, and coupled correction based on the coupling relationship among the ambient temperature, the traffic state parameters, and the specific travel distance parameters. Based on the road network topology, identify the first type of road segments in the target area, where the first type of road segments do not have direct traffic observation data; For the first type of road segment, the traffic flow of the first type of road segment is determined by a statistical inference model based on prior information in the time dimension, prior information in the spatial dimension, and physical conservation constraints. The energy consumption of the target area is calculated based on the traffic flow of the first type of road segment, the proportion of the mixed vehicle type structure, and the corrected energy consumption factor.

2. The method according to claim 1, characterized in that, To obtain the proportion of the second type of vehicle structure in the second category, including: Extract the passage data of the second type of vehicles from historical passage records for multiple historical target time periods. Each passage data includes at least the vehicle's place of origin, the checkpoint location when entering the target road network, the time of passing through the current checkpoint, and the direction of travel. Based on the stated place of origin, determine the vehicle's origin classification; Based on the mileage difference between the checkpoint location where the vehicle was initially captured when it entered the target road network and the current checkpoint location, the distance the vehicle has traveled is estimated, and the total travel distance is calculated based on the distance traveled and the direction of travel, and the total travel distance is divided into a preset travel distance range. Based on the time of passing through the current checkpoint, determine the type of passage period to which the current passage belongs; Multiple cross-classification categories are constructed by using the origin classification, travel distance range, travel time type, and travel direction as dimensions for cross-classification; For each cross-category, the total number of vehicles belonging to that category and the number of vehicles using a specific energy type are counted. The ratio of the number of vehicles using a specific energy type to the total number of vehicles is taken as the proportion of that energy type in that category. The proportion of the second vehicle model structure is determined based on the proportion of the energy type corresponding to one or more cross-classification categories matched during the target time period.

3. The method according to claim 1, characterized in that, The corrected energy consumption factor for the target road segment is obtained by correcting the baseline energy consumption factor, including: The first unit mileage energy consumption and the second unit mileage energy consumption are obtained when the vehicle is driven under a baseline traffic condition at a first preset ambient temperature and a second preset ambient temperature; the first preset ambient temperature is normal temperature and the second preset ambient temperature is low temperature. Calculate the ratio of the energy consumption per unit mileage to the energy consumption per unit mileage, and use it as a correction factor for low-temperature single operating conditions. The third and fourth unit mileage energy consumption of the vehicle are obtained when it is driving under the first preset ambient temperature in the first preset traffic state and the second preset traffic state; the traffic efficiency of the second preset traffic state is lower than that of the first preset traffic state. The ratio of the energy consumption per fourth unit mileage to the energy consumption per third unit mileage is calculated and used as a correction factor for a single congestion operating condition. Acquire multiple sets of measured energy consumption data of the vehicle under the second preset ambient temperature and the second preset traffic conditions at different travel lengths; Divide each set of measured energy consumption data by the product of the baseline energy consumption factor, the low temperature single-condition correction factor, and the congestion single-condition correction factor, and determine the portion of the resulting ratio that exceeds 1 as the coupling effect contribution value under that travel length. Establish a functional relationship between the contribution value of the coupling effect and ambient temperature, traffic state parameters and travel length, wherein the contribution value of the coupling effect decreases as the travel length increases; The ambient temperature, traffic condition parameters, and travel length of the current target road segment are obtained, and the current coupling effect contribution value is determined using the aforementioned functional relationship. The coupling amplification factor is obtained by adding 1 to the current coupling effect contribution value. The corrected energy consumption factor for the target road segment is obtained by multiplying the baseline energy consumption factor, the low-temperature single-condition correction factor, the congestion single-condition correction factor, and the coupling amplification factor.

4. The method according to claim 3, characterized in that, For fuel-powered vehicles, the first preset ambient temperature is normal temperature, and the second preset ambient temperature is low temperature; the reference traffic state is high-speed constant speed driving state. For electric vehicles, the first preset ambient temperature is normal temperature, and the second preset ambient temperature is low temperature; the reference traffic state is high-speed constant speed driving state.

5. The method according to claim 1, characterized in that, For the first type of road segment, based on prior information in the time dimension, prior information in the spatial dimension, and physical conservation constraints, the traffic flow of the first type of road segment is determined through a statistical inference model, including: Select historical dates and time periods that have the same traffic characteristics as the target date and the target time period, wherein the same traffic characteristics include at least the same or similar holiday types and weekday attributes; The traffic flow of the first type of road segment under each of the selected historical dates and historical time periods is obtained to form a historical traffic flow sample set; Calculate the weighted average of each sample in the historical traffic sample set as the mean of the time-dimensional prior information, wherein the sample whose time is closer to the target date is assigned a greater weight; Calculate the weighted variance of each sample in the historical traffic sample set as the variance of the prior information in the time dimension; Identify the target upstream road segment with direct traffic observation data that is adjacent to the first type of road segment in the road network topology; Obtain the measured traffic flow of the upstream road segment of the target during the current target time period; The mean of the spatial dimension prior information is calculated based on the pre-established flow transmission relationship from the upstream road segment of the target to the first type of road segment, which is based on historical data. The historical fitting error of the flow transmission relationship is obtained as the variance of the spatial dimension prior information; Obtain the number of samples in the historical traffic sample set, and determine the fusion weight based on the number of samples, wherein the fusion weight increases monotonically with the increase of the number of samples, and the value ranges from 0 to 1; Multiply the fusion weight by the mean of the time-dimensional prior information to obtain the first intermediate value; Multiply the difference between 1 and the fusion weight by the mean of the spatial dimension prior information to obtain the second intermediate value; Add the first intermediate value to the second intermediate value to obtain the mean of the joint prior distribution; The third intermediate value is obtained by multiplying the square of the fusion weight by the variance of the time-dimensional prior information. The fourth intermediate value is obtained by multiplying the square of the difference between 1 and the fusion weight by the variance of the spatial dimension prior information. Add the third intermediate value to the fourth intermediate value to obtain the variance of the joint prior distribution; The vehicle physical conservation relationship of the road network topology between the first type of road segment and adjacent road segments is used as the likelihood constraint of the statistical inference model. Based on the mean and variance of the joint prior distribution and the likelihood constraint, the posterior distribution of the traffic flow of the first type of road segment is determined by a Bayesian inference model, and the mean of the posterior distribution is determined as the traffic flow of the first type of road segment.

6. The method according to claim 1, characterized in that, The method further includes: Based on the road network topology, identify a second type of road segment in the target area, wherein the topological hop count of the second type of road segment from the road segment with direct traffic observation data is greater than that of the first type of road segment; For the second type of road segment, the upper quantile of the posterior distribution of its traffic flow is calculated as a conservative estimate of the traffic flow of the second type of road segment based on the posterior distribution of its traffic flow. The conservatively estimated traffic flow is used as the traffic flow for the second type of road segment, and a risk label is generated for the second type of road segment.

7. The method according to claim 1, characterized in that, Based on the traffic flow of the first type of road segment, the proportion of the mixed vehicle type structure, and the corrected energy consumption factor, the energy consumption of the target area is calculated, including: Receive calculation instructions for long-cycle prediction scenarios or short-cycle calculation scenarios. A first set of input parameters is configured for the long-term prediction scenario. The ambient temperature in the first set of input parameters is taken as the historical statistical average value, the traffic state parameter is taken as the parameter value representing the smooth flow state, and the vehicle structure ratio is taken as the first vehicle structure ratio. A second set of input parameters is configured for the short-cycle calculation scenario. The ambient temperature in the second set of input parameters is taken as the historical statistical extreme value, the traffic state parameter is taken as the parameter value representing the congestion state, and the vehicle structure ratio is taken as the ratio of the mixed vehicle structure. The unified underlying computing kernel is invoked to calculate the first energy consumption result of the target area in the long-term prediction scenario and the second energy consumption result in the short-term calculation scenario, based on the input traffic flow of the first type of road segment, the correction energy consumption factor, and the vehicle type structure ratio of the corresponding scenario. Output the first energy consumption result in an annual trend format, and output the second energy consumption result in an hourly distribution format.

8. A device for measuring the dynamic energy consumption of mixed traffic flow, characterized in that, The device includes: The vehicle type structure fusion module is used to obtain the first vehicle type structure ratio of the first type of vehicles and the second vehicle type structure ratio of the second type of vehicles passing through the target area during the target time period; and to fuse the first vehicle type structure ratio and the second vehicle type structure ratio with the traffic flow of the first type of vehicles and the second type of vehicles as weights to obtain the mixed vehicle type structure ratio of the target area during the target time period. An energy consumption factor correction module is used to correct a baseline energy consumption factor based on ambient temperature, traffic state parameters, and specific travel distance parameters of a vehicle on a target road segment, thereby obtaining a corrected energy consumption factor for the target road segment. The correction includes independent correction based on the ambient temperature and traffic state parameters, as well as coupled correction based on the coupling relationship between the ambient temperature, traffic state parameters, and specific travel distance parameters. The traffic flow inversion module is used to identify the first type of road segments in the target area based on the road network topology, where the first type of road segments do not have direct traffic observation data; and for the first type of road segments, the traffic flow is determined by a statistical inference model based on time-dimensional prior information, spatial-dimensional prior information, and physical conservation constraints. An energy consumption calculation module is used to calculate the energy consumption of the target area based on the traffic flow of the first type of road segment, the proportion of the mixed vehicle type structure, and the corrected energy consumption factor.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images