An electric heavy truck leasing risk assessment method, device and equipment

Abstract

The application provides an electric heavy truck rental risk assessment method, device and equipment, and relates to the technical field of data processing. The method comprises the following steps: obtaining real-time vehicle operation data, real-time battery state data, energy supplement behavior data, battery replacement behavior data and external application data of a target electric heavy truck, wherein the external application data at least comprises regional peak-valley electricity price table, weather data, maintenance data and freight business data; performing feature extraction on the real-time vehicle operation data, real-time battery state data, energy supplement behavior data and battery replacement behavior data to obtain multi-class feature data of the target electric heavy truck; and performing rental risk assessment on the target electric heavy truck according to the multi-class feature data to obtain a target risk level of the target electric heavy truck. The application guarantees comprehensive data dimensions, makes the assessment result more reliable, and thus improves the early warning capability and reduces asset loss.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to an electric heavy truck leasing risk assessment method, device and equipment. BACKGROUND

[0002] With the promotion of new energy commercial vehicles, electric heavy trucks have been widely used in trunk line logistics, port transportation and mine short dumping scenes. Due to the high cost of vehicle purchase, financing leasing has become the mainstream sales mode, the financing party provides the vehicle body or battery assets, and the lessee pays the rent in installments and is responsible for operation. Under this mode, the financing party faces multiple risks such as overdue rent, improper use of vehicles, accelerated performance degradation of key assets, etc.

[0003] At present, the risk management of electric heavy truck financing leasing still mainly relies on basic credit scoring model for analysis, and the prediction model based on historical repayment data and simple driving mileage is used to assess the default possibility, or after the lessee is overdue, the vehicle terminal is remotely locked or legal collection is started.

[0004] However, the existing model does not fully incorporate the key operating characteristics of electric heavy trucks, resulting in inaccurate assessment, making risk control generally lagging behind, lacking early warning capability, and being difficult to identify high-risk customers before the overdue, leading to difficult prevention of asset losses. SUMMARY

[0005] The purpose of the present application is to overcome the deficiencies in the prior art, provide an electric heavy truck leasing risk assessment method, device and equipment, ensure comprehensive data dimensions, make the assessment results more reliable, and thus improve the early warning capability and reduce asset losses.

[0006] To achieve the above purpose, the technical solutions adopted by the embodiments of the present application are as follows:

[0007] In a first aspect, an embodiment of the present application provides an electric heavy truck leasing risk assessment method, which comprises:

[0008] Obtaining real-time vehicle operation data, real-time battery state data, energy supplement behavior data, battery replacement behavior data and external application data of a target electric heavy truck vehicle, wherein the external application data at least includes regional peak-valley electricity price table, weather data, maintenance data and freight business data;

[0009] Performing feature extraction on the real-time vehicle operation data, the real-time battery state data, the energy supplement behavior data and the battery replacement behavior data to obtain multi-class feature data of the target electric heavy truck vehicle;

[0010] According to the multi-class feature data, performing leasing risk assessment on the target electric heavy truck vehicle to obtain a target risk level of the target electric heavy truck vehicle.

[0011] Optionally, the target electric heavy truck vehicle is subjected to a lease risk assessment according to the multi-type feature data, and a target risk level of the target electric heavy truck vehicle is obtained.

[0012] A plurality of vehicle evaluation input features, a plurality of battery evaluation input features, and a plurality of battery swap evaluation input features are obtained from the multi-type feature data.

[0013] A vehicle risk score is obtained by performing a vehicle risk assessment on the target electric heavy truck vehicle according to the plurality of vehicle evaluation input features.

[0014] A battery risk score is obtained by performing a battery risk assessment on the target electric heavy truck vehicle according to the plurality of battery evaluation input features.

[0015] A battery swap risk score is obtained by performing a battery swap risk assessment on the target electric heavy truck vehicle according to the plurality of battery swap evaluation input features.

[0016] An overdue risk probability is obtained by performing overdue probability prediction using a preset overdue prediction model according to the multi-type feature data.

[0017] The target risk level of the target electric heavy truck vehicle is obtained according to the vehicle risk score, the battery risk score, the battery swap risk score, and the overdue risk probability.

[0018] Optionally, the multi-type feature data includes vehicle operation features, battery behavior features, cost control features, battery swap behavior features, residual value correlation features, and business correlation features.

[0019] The plurality of vehicle evaluation input features, the plurality of battery evaluation input features, and the plurality of battery swap evaluation input features are obtained from the multi-type feature data.

[0020] The plurality of vehicle evaluation input features are obtained from the vehicle operation features and the business correlation features.

[0021] The plurality of battery evaluation input features are obtained according to the battery behavior features, the cost control features, and the residual value correlation features.

[0022] The plurality of battery swap evaluation input features are obtained from the battery swap behavior features.

[0023] Optionally, the vehicle operation features include empty running rate, number of bad driving, working condition intensity, maintenance frequency, and positioning deviation frequency; and the business correlation features include freight order volume year-on-year drop.

[0024] The obtaining the plurality of vehicle evaluation input features from the vehicle operation features and the business correlation features comprises:

[0025] The empty running rate, the number of bad driving, the working condition intensity, the maintenance frequency, and the positioning deviation frequency are obtained from the vehicle operation features.

[0026] The freight order volume month-on-month decrease is obtained from the business correlation features.

[0027] The empty running rate, the number of bad driving, the working condition intensity, the maintenance frequency, the positioning deviation frequency, and the freight order volume month-on-month decrease are determined as the plurality of vehicle evaluation input features.

[0028] Optionally, the obtaining the plurality of battery evaluation input features according to the battery behavior features, the cost control features, and the residual value correlation features comprises:

[0029] The SOH (State of Health) attenuation rate, the fast charging proportion, the high-temperature charging frequency, and the overcharging frequency are obtained from the battery behavior features.

[0030] The valley electricity charging proportion is obtained from the cost control features.

[0031] The residual value decrease amplitude is predicted according to the residual value correlation features, to obtain a predicted residual value decrease amplitude.

[0032] The SOH attenuation rate, the fast charging proportion, the high-temperature charging frequency, the overcharging frequency, the valley electricity charging proportion, and the predicted residual value decrease amplitude are determined as the plurality of battery evaluation input features.

[0033] Optionally, the obtaining the plurality of battery swap evaluation input features from the battery swap behavior features comprises:

[0034] The cumulative battery swap frequency, the SOH parameter before and after battery swap, the battery swap cost proportion, and the battery swap complaint rate are obtained from the battery swap behavior features as the plurality of battery swap evaluation input features.

[0035] Optionally, the obtaining the target risk level of the target electric heavy-duty truck vehicle according to the vehicle risk score, the battery risk score, the battery swap risk score, and the overdue risk probability comprises:

[0036] If any one of the vehicle risk score, the battery risk score, and the battery swap risk score is less than a first preset score value, or the overdue risk probability is greater than a first preset probability, the target risk level is determined as a high risk level.

[0037] If all of the vehicle risk score, the battery risk score and the battery swap risk score are greater than or equal to the first preset score but less than a second preset score, and the overdue risk probability is greater than or equal to a second preset probability but less than or equal to the first preset probability, it is determined that the target risk level is a medium risk level; wherein the second preset score is greater than the first preset score, and the second preset probability is less than the first preset probability.

[0038] If all of the vehicle risk score, the battery risk score and the battery swap risk score are greater than or equal to the second preset score, it is determined that the target risk level is a low risk level.

[0039] Optionally, the method further comprises:

[0040] According to the target risk level, a corresponding intervention measure is generated and output.

[0041] In a second aspect, another embodiment of the present application provides an electric heavy truck rental risk assessment device, the device comprising:

[0042] An acquisition module is configured to acquire real-time vehicle operation data, real-time battery state data, energy supplement behavior data, battery swap behavior data and external application data of a target electric heavy truck, wherein the external application data at least includes regional peak-valley electricity price table, weather data, maintenance data and freight business data.

[0043] An extraction module is configured to perform feature extraction on the real-time vehicle operation data, the real-time battery state data, the energy supplement behavior data and the battery swap behavior data, to obtain multi-class feature data of the target electric heavy truck.

[0044] An evaluation module is configured to perform rental risk assessment on the target electric heavy truck according to the multi-class feature data, to obtain a target risk level of the target electric heavy truck.

[0045] In a third aspect, another embodiment of the present application provides an electronic device, comprising a processor, a memory and a bus, the memory stores machine readable instructions executable by the processor, when the electronic device is running, the processor and the memory communicate through the bus, the processor executes the machine readable instructions to perform the steps of the electric heavy truck rental risk assessment method of any one of the above first aspect.

[0046] In a fourth aspect, another embodiment of the present application provides a storage medium, the storage medium stores a computer program, when the computer program is run by a processor, the steps of the electric heavy truck rental risk assessment method of any one of the above first aspect are executed.

[0047] The beneficial effects of the present application are:

[0048] The application provides an electric heavy truck leasing risk assessment method, device and equipment, real-time vehicle operation data, real-time battery state data, energy supplement behavior data, battery replacement behavior data and external application data of a target electric heavy truck are acquired, multi-class feature data of the target electric heavy truck is obtained by performing feature extraction on the real-time vehicle operation data, the real-time battery state data, the energy supplement behavior data and the battery replacement behavior data, and the target electric heavy truck is subjected to leasing risk assessment according to the multi-class feature data, and a target risk level of the target electric heavy truck is obtained. The application realizes risk assessment of the electric heavy truck through the vehicle operation data, the real-time battery state data, the energy supplement behavior data, the battery replacement behavior data and the external application data, improves the prediction accuracy, and performs fusion feature assessment on multi-dimensional data, guarantees comprehensive data dimensions, and makes the assessment result more reliable. Therefore, early warning is more accurate, and passive defense is changed to active prevention, and the asset safety and business sustainability of the financing leasing company are effectively guaranteed. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor.

[0050] Figure 1 A flowchart of an electric heavy truck leasing risk assessment method provided by the embodiments of the present application is shown in the figure.

[0051] Figure 2 A flowchart of determining a target risk level in an electric heavy truck leasing risk assessment method provided by the embodiments of the present application is shown in the figure.

[0052] Figure 3 A flowchart of determining input features in an electric heavy truck leasing risk assessment method provided by the embodiments of the present application is shown in the figure.

[0053] Figure 4 A flowchart of determining vehicle evaluation input features in an electric heavy truck leasing risk assessment method provided by the embodiments of the present application is shown in the figure.

[0054] Figure 5 A flowchart of determining battery evaluation input features in an electric heavy truck leasing risk assessment method provided by the embodiments of the present application is shown in the figure.

[0055] Figure 6Another electric heavy truck rental risk assessment method provided by the embodiment of the present application provides a flowchart for determining the target risk level.

[0056] Figure 7 A structural diagram of an electric heavy truck rental risk assessment device provided by the embodiment of the present application is shown in the figure.

[0057] Figure 8 A structural diagram of an electronic device provided by the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0058] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in detail with reference to the accompanying drawings in the embodiments of the present application. It should be understood that the accompanying drawings in the present application are only intended to illustrate and describe the present application, and are not intended to limit the protection scope of the present application. In addition, it should be understood that the accompanying drawings are not drawn according to the actual proportions. The flowchart shows the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowchart can not be implemented in sequence, and the steps without logical context relationship can be reversed in sequence or implemented simultaneously. In addition, one or more other operations can be added to the flowchart or one or more operations can be removed from the flowchart under the guidance of the content of the present application.

[0059] In addition, the described embodiments are only some of the embodiments of the present application, not all the embodiments. The components of the embodiments of the present application described and shown in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0060] It should be noted that the term “comprising” will be used in the embodiments of the present application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0061] Currently, the sales of electric heavy-duty commercial vehicles are mainly in the mode of financial leasing. If the lessee fails to repay the loan on time after the delivery of the vehicle, it will directly affect the asset security and fund recovery of the financial leasing company. Therefore, it is of great significance to establish an effective financial leasing asset risk prediction mechanism for improving asset operation efficiency and risk control capability. Due to the commercial mode of separating the vehicle and the battery of the electric heavy-duty truck, the battery can be independently detached and circulated. The traditional method of locking the engine to control the vehicle cannot effectively freeze the core asset (battery), resulting in a decrease in asset control. The vehicle body and the battery of the electric heavy-duty truck usually belong to different investors. In addition to focusing on the overall operation of the vehicle, the vehicle body investor mainly focuses on the vehicle body asset, and the battery investor mainly focuses on the battery asset. Therefore, both parties can take asset control measures: the vehicle body party can lock the vehicle by locking the engine speed, and the battery party can lock the power by the battery management system. For the battery swap vehicle, the power supply adopts the battery swap mode, and the battery and the vehicle have no strong relationship. The battery swap holder is held by the battery swap station, the vehicle fleet or other parties. The existing method does not involve evaluating the leasing risk of the vehicle in the case of separating the vehicle and the battery. Therefore, the present application provides an electric heavy-duty truck leasing risk evaluation method, which obtains real-time vehicle running data, real-time battery state data, power supply behavior data, battery swap behavior data and external application data of a target electric heavy-duty truck vehicle, extracts features from the real-time vehicle running data, real-time battery state data, power supply behavior data and battery swap behavior data to obtain multi-class feature data of the target electric heavy-duty truck vehicle, and evaluates the leasing risk of the target electric heavy-duty truck vehicle according to the multi-class feature data to obtain the target risk level of the target electric heavy-duty truck vehicle. The risk level of the target electric heavy-duty truck vehicle is accurately evaluated by the method of the present application, so as to improve the early warning of the expected risk and reduce the property loss.

[0062] To clearly describe the electric heavy-duty truck leasing risk evaluation method provided by the embodiments of the present application, the method is described as follows in combination with multiple drawings, Figure 1 The flowchart of the electric heavy-duty truck leasing risk evaluation method provided by the embodiments of the present application is shown in Figure 1 As shown in the figure, the method comprises:

[0063] Step 101, obtaining real-time vehicle running data, real-time battery state data, power supply behavior data, battery swap behavior data and external application data of a target electric heavy-duty truck vehicle.

[0064] The external application data at least includes: regional peak-valley electricity price table, weather data, maintenance data and freight service data. The vehicle operation data is used to indicate a set of data related to vehicle state, driving behavior, driver operation, external environment and vehicle system performance, which is recorded and transmitted in the vehicle driving or powered state. The vehicle operation data includes: vehicle identification number (VIN), speed, GPS track, empty running rate, engine running time, and load estimation value. The real-time battery state data is used to indicate a set of dynamic change parameters of the electrical, thermal and health state of the battery at any time. The real-time battery state data includes: state of charge (SOC), state of health (SOH), battery temperature, battery cycle number, whether the battery is fast charging, SOH daily attenuation rate and high-temperature charging cumulative time length. The energy supplement behavior data is a set of data recorded about behavior preference, process parameter, space-time information and economic cost around the electric vehicle energy supplement process. The energy supplement behavior data includes: charging time, start and end SOC, charging pile type, whether in valley electricity time period, valley electricity charging time period offset. The battery replacement behavior data is a set of data recorded about behavior preference, service process, battery state and business transaction in the process of battery replacement by the user at the battery replacement station. The battery replacement behavior data includes: battery identity document (ID), battery replacement time, SOH difference before and after battery replacement, battery replacement time length, single battery replacement fee and battery replacement complaint record.

[0065] Optionally, the running data of the vehicle is obtained through a vehicle terminal of the electric heavy truck, and the vehicle terminal can be a telematics box (T-Box). The battery state data of the electric heavy truck is obtained through a battery management system (BMS) of a battery of the electric heavy truck. The power supplement behavior data of the electric heavy truck is obtained through the BMS of the electric heavy truck and a data interface of an external charging pile. The battery replacement behavior data of the electric heavy truck is obtained through an interface of a battery replacement pipe system corresponding to the electric heavy truck and a vehicle fleet feedback interface. The regional peak-valley electricity price table is obtained through a power grid interface, the weather data is obtained through a weather interface, the maintenance data is obtained through a maintenance platform interface, and the freight business data is obtained through a freight platform interface. All the data is uploaded in real time in a lightweight data exchange format (JSON) through a lightweight message transmission protocol (MQTT) and a transmission control protocol (TCP) communication mode in a publish / subscribe mode.

[0066] Optionally, in the process of obtaining the real-time vehicle running data, the real-time battery state data, the power supplement behavior data, the battery replacement behavior data, and the external application data of the target electric heavy truck, the battery replacement behavior data is preferentially connected to the mainstream battery replacement station standardized interface, and the battery replacement triggering signal is directly collected by the vehicle T-Box and the battery replacement station equipment for the rare battery replacement station. Whether it is a valley power time period in the power supplement behavior data is obtained from a power grid official website or a third-party interface through a timing task and stored in a data warehouse. The timing task can be performed every morning, and the data storage resources can include a region, a date, a peak / flat / valley period, and an electricity price. In the charging behavior analysis, whether the charging is valley power charging is dynamically associated and marked. The weather data in the external application data is obtained by calling a meteorological interface to obtain the regional temperature and precipitation, and stored in a data warehouse for analyzing the influence of low temperature on the battery performance. The maintenance data in the external application data is obtained from a maintenance platform by calling an application programming interface (API), and stored in a data warehouse. The maintenance data can include a maintenance category, a cost, and a maintenance time.

[0067] Optionally, after obtaining the real-time vehicle running data, real-time battery state data, energy supplement behavior data, battery swap behavior data, and external application data of the target electric heavy truck vehicle, the data is cleaned and verified, and then feature extraction is performed. Specifically, for continuous signals such as speed, SOC, and temperature, sliding window mean filtering is used to eliminate instantaneous jumps, where the sliding window can be 10 seconds. For GPS trajectory, simple Kalman filtering is used for trajectory smoothing, and electronic fence and map matching algorithms are used to identify invalid drift points. For charging identification events, charging events are identified based on SOC state, for example, 0x01 for parking charging, 0x02 for energy recovery while the vehicle is running, 0x03 for non-charging state, 0x04 for charging completion, 0xFE for abnormality, and 0xFF for invalid charging. For battery swap event identification, based on three conditions of battery ID change, SOH mutation, and battery swap station location matching, valid battery swap is identified, and abnormal states such as incomplete plugging are excluded. For SOC reasonableness verification, if the single charging increment is less than or equal to the rated capacity of the battery x 1.05, the verification is successful.

[0068] Step 102, feature extraction is performed on the real-time vehicle running data, real-time battery state data, energy supplement behavior data, and battery swap behavior data to obtain multi-class feature data of the target electric heavy truck vehicle.

[0069] Optionally, feature extraction is performed on the real-time vehicle running data, real-time battery state data, energy supplement behavior data, and battery swap behavior data to determine the association information between multi-class data, and based on the association information of multi-class data, multi-class feature data of the target electric heavy truck vehicle is obtained.

[0070] Step 103, based on the multi-class feature data, the target electric heavy truck vehicle is subjected to rental risk assessment to obtain a target risk level of the target electric heavy truck vehicle.

[0071] Optionally, based on the multi-class feature data, the target electric heavy truck vehicle is subjected to rental risk assessment to determine the influence of different feature data on the rental risk of the heavy truck vehicle, and based on the influence of different feature data on the rental risk of the heavy truck vehicle, a target risk level of the target electric heavy truck vehicle is obtained.

[0072] The application obtains real-time vehicle running data, real-time battery state data, energy supplement behavior data, battery replacement behavior data and external application data of a target electric heavy truck vehicle, extracts features from the real-time vehicle running data, the real-time battery state data, the energy supplement behavior data and the battery replacement behavior data, and obtains multi-class feature data of the target electric heavy truck vehicle; and performs lease risk assessment on the target electric heavy truck vehicle according to the multi-class feature data, and obtains a target risk level of the target electric heavy truck vehicle. The application realizes risk assessment of the electric heavy truck through vehicle running data, real-time battery state data, energy supplement behavior data, battery replacement behavior data and external application data, improves the prediction accuracy, and performs fusion feature assessment on multi-dimensional data to ensure comprehensive data dimensions and make the assessment result more reliable. Therefore, the early warning is more accurate, and the passive defense is changed to active prevention, effectively ensuring the asset safety and business sustainability of the financial leasing company.

[0073] On the basis of the above-mentioned embodiments, the application further provides a process for determining a target risk level in an electric heavy truck lease risk assessment method, Figure 2 A process for determining a target risk level in an electric heavy truck lease risk assessment method provided by the embodiments of the application is shown in Figure 2 According to the multi-class feature data, the target electric heavy truck vehicle is subjected to lease risk assessment in step 103, and a target risk level of the target electric heavy truck vehicle is obtained, which includes:

[0074] Step 201: Obtain a plurality of vehicle evaluation input features, a plurality of battery evaluation input features and a plurality of battery replacement evaluation input features from the multi-class feature data.

[0075] The vehicle evaluation input features are used to indicate the overall condition, configuration and market attributes of the vehicle. The battery evaluation input features are used to indicate the battery health status and performance-related data of the vehicle. The battery replacement evaluation input features are used to indicate the battery liquidity, battery replacement behavior and hardware loss information of the vehicle.

[0076] Step 202: According to the plurality of vehicle evaluation input features, the target electric heavy truck vehicle is subjected to vehicle risk assessment, and a vehicle risk score is obtained.

[0077] The vehicle risk score is used to evaluate the operation ability and operation compliance of the lessee. A low vehicle risk score indicates poor operation ability, low repayment willingness and repayment ability, and poor operation compliance, which represents potential fraud or asset loss risk.

[0078] Optionally, according to the plurality of vehicle evaluation input features, the target electric heavy truck vehicle is subjected to vehicle risk assessment, and a vehicle risk score is obtained. According to the vehicle risk score, the operation ability and operation compliance of the lessee are determined.

[0079] Step 203, performing battery risk assessment on the target electric heavy truck vehicle according to a plurality of battery assessment input features to obtain a battery risk score.

[0080] The battery risk score is used to assess the health status and use risk of the power battery as the core asset. The power battery is the most valuable part of the financing lease asset, and rapid degradation of its health or improper use will directly result in a significant loss of asset value. Improper use can include overcharging and high-temperature charging.

[0081] Optionally, the battery risk score is obtained by performing battery risk assessment on the target electric heavy truck vehicle according to a plurality of battery assessment input features, thereby obtaining the asset value of the battery.

[0082] Step 204, performing battery swap risk assessment on the target electric heavy truck vehicle according to a plurality of battery swap assessment input features to obtain a battery swap risk score.

[0083] The battery swap risk score is used to assess the lessee's performance and abnormal patterns during the use of the battery swap service. Battery swap is a necessary condition for maintaining vehicle operation. Abnormal behavior here is a strong signal of the lessee's financial strain or default tendency. Abnormal behavior can include overdue service fees and illegal cross-station.

[0084] Optionally, the battery swap risk score is obtained by performing battery swap risk assessment on the target electric heavy truck vehicle according to a plurality of battery swap assessment input features, thereby determining the lessee's performance and abnormal patterns during the use of the battery swap service.

[0085] Step 205, performing overdue probability prediction using a preset overdue prediction model according to a plurality of feature data to obtain an overdue risk probability.

[0086] The preset overdue prediction model is a preset large language model or neural network model. The overdue risk probability is the likelihood of the lessee's repayment overdue event.

[0087] Optionally, the overdue risk probability is obtained by performing overdue probability prediction on the likelihood of the lessee's repayment overdue event using a preset overdue prediction model according to a plurality of feature data.

[0088] Step 206, obtaining a target risk level of the target electric heavy truck vehicle according to the vehicle risk score, the battery risk score, the battery swap risk score, and the overdue risk probability.

[0089] Optionally, the vehicle risk score, the battery risk score, the battery replacement risk score, and the overdue risk probability are weighted and summed to obtain a target risk level of the target electric heavy truck. Specifically, the weights corresponding to the vehicle risk score, the battery risk score, the battery replacement risk score, and the overdue risk probability can be determined according to historical default data, and the embodiments of the present application do not limit this.

[0090] Optionally, the vehicle risk level, the battery risk level, the battery replacement risk level, and the overdue risk level are respectively determined according to the vehicle risk score, the battery risk score, the battery replacement risk score, and the overdue risk probability, so as to determine the target risk level according to the vehicle risk level, the battery risk level, the battery replacement risk level, and the overdue risk level.

[0091] In the embodiments of the present application, a plurality of vehicle evaluation input features, a plurality of battery evaluation input features, and a plurality of battery replacement evaluation input features are obtained from the plurality of feature data; the target electric heavy truck is subjected to vehicle risk evaluation according to the plurality of vehicle evaluation input features to obtain a vehicle risk score; the target electric heavy truck is subjected to battery risk evaluation according to the plurality of battery evaluation input features to obtain a battery risk score; the target electric heavy truck is subjected to battery replacement risk evaluation according to the plurality of battery replacement evaluation input features to obtain a battery replacement risk score; an overdue risk probability is predicted by using a preset overdue prediction model according to the plurality of feature data; and a target risk level of the target electric heavy truck is obtained according to the vehicle risk score, the battery risk score, the battery replacement risk score, and the overdue risk probability. The vehicle risk score, the battery risk score, the battery replacement risk score, and the overdue risk probability are obtained by using a plurality of features, a multi-level and complementary risk view is constructed, even if a certain method deviates, the results of other dimensions can provide cross-validation, so that the final comprehensive judgment is more comprehensive and stable, and the misjudgment rate is reduced. The plurality of risk scores are more in line with the actual business scene, can effectively distinguish local risks and overall crisis, and thus precise intervention strategies can be executed.

[0092] On the basis of the above-mentioned embodiments, the plurality of feature data comprises vehicle operation features, battery behavior features, cost control features, battery replacement behavior features, residual value association features, and business association features. The present application further provides a process for determining input features in an electric heavy truck rental risk assessment method, Figure 3 A process for determining input features in an electric heavy truck rental risk assessment method provided in the embodiments of the present application is shown in Figure 3 As shown in the above-mentioned step 201, a plurality of vehicle evaluation input features, a plurality of battery evaluation input features, and a plurality of battery replacement evaluation input features are obtained from the plurality of feature data, which comprises:

[0093] Step 301: A plurality of vehicle evaluation input features are obtained from the vehicle operation features and the business association features.

[0094] The vehicle operation feature is used to reflect various dynamic or static technical and behavioral parameters generated by the vehicle in actual operation, and is used to describe how the vehicle is used, its running state and operation habits. The business association feature is used to indicate structural information in the upper business logic, service type, business model or organizational relationship related to the vehicle, and embodies the role and connection relationship of the vehicle in a specific business system.

[0095] Optionally, a feature corresponding to the vehicle evaluation is determined from a plurality of features in the vehicle operation feature and the business association feature, and the obtained feature is taken as the plurality of vehicle evaluation input features.

[0096] Step 302, obtaining a plurality of battery evaluation input features according to the battery behavior feature, the cost control feature and the residual value association feature.

[0097] The battery behavior feature is used to record the working state, aging trend and abnormal behavior of the battery in the charging and discharging process, and reflects the performance degradation and safety boundary. The cost control feature is used to quantify the performance of the vehicle in the entire operation cycle, including comprehensive costs such as energy, maintenance, manpower and time. The residual value association feature is various historical data related to the future second-hand market value of the vehicle, which is used to predict the depreciation curve and residual value.

[0098] Optionally, a feature corresponding to the vehicle evaluation is determined from a plurality of features in the battery behavior feature, the cost control feature and the residual value association feature, and the obtained feature is taken as the plurality of battery evaluation input features.

[0099] Step 303, obtaining a plurality of battery replacement evaluation input features from the battery replacement behavior feature.

[0100] The battery replacement behavior feature is used to record the battery replacement station access rule and battery replacement decision logic of the electric vehicle type supporting battery replacement.

[0101] Optionally, a feature corresponding to the vehicle evaluation is determined from a plurality of features in the battery replacement behavior feature, and the obtained feature is taken as the plurality of battery replacement evaluation input features.

[0102] In the embodiments of the present application, the plurality of vehicle evaluation input features are obtained from the vehicle operation feature and the business association feature; the plurality of battery evaluation input features are obtained according to the battery behavior feature, the cost control feature and the residual value association feature; and the plurality of battery replacement evaluation input features are obtained from the battery replacement behavior feature. The present application logically decouples the complex whole vehicle running data according to the function domain, respectively constructs independent but fusable evaluation sub-models, avoids the noise interference caused by the feature mixing in the traditional single model, realizes multi-dimensional decoupling and precise modeling, and improves the evaluation accuracy.

[0103] On the basis of the above embodiment, the vehicle operation features include: empty running rate, bad driving times, working condition intensity, maintenance frequency, and positioning deviation frequency; the business association features include: freight order volume month-on-month drop. The application also provides a process for determining vehicle evaluation input features in an electric heavy truck rental risk assessment method, Figure 4 A process for determining vehicle evaluation input features in an electric heavy truck rental risk assessment method provided by the embodiment of the application is shown in FIG. Figure 4 As shown in step 301, a plurality of vehicle evaluation input features are obtained from the vehicle operation features and the business association features, including:

[0104] Step 401, obtaining the empty running rate, bad driving times, working condition intensity, maintenance frequency, and positioning deviation frequency from the vehicle operation features.

[0105] The empty running rate is the daily empty running mileage ratio of the electric heavy truck during driving. A high empty running rate indicates low revenue. The bad driving times include daily average sudden braking and daily average sudden acceleration. A sudden braking of >5 m / s² is considered as one sudden braking, and a sudden acceleration of >3 m / s² is considered as one sudden acceleration. The working condition intensity is the ratio of daily mountain driving mileage and heavy load driving mileage to total driving mileage. A high ratio of daily mountain driving mileage and heavy load driving mileage to total driving mileage indicates that the vehicle is quickly worn out. The maintenance frequency is the monthly maintenance frequency. The positioning deviation frequency is the deviation frequency of the daily GPS track of the vehicle from the preset driving track.

[0106] Step 402, obtaining the freight order volume month-on-month drop from the business association features.

[0107] The freight order volume is the ratio of the order volume difference in the last 30 days to the order volume of the last month.

[0108] Step 403, determining the empty running rate, bad driving times, working condition intensity, maintenance frequency, positioning deviation frequency, and freight order volume month-on-month drop as the plurality of vehicle evaluation input features.

[0109] Optionally, the empty running rate, bad driving times, working condition intensity, maintenance frequency, positioning deviation frequency, and freight order volume month-on-month drop are taken as the plurality of vehicle evaluation input features.

[0110] Optionally, the vehicle risk score is determined according to the empty running rate, bad driving times, working condition intensity, maintenance frequency, positioning deviation frequency, and freight order volume month-on-month drop.

[0111]

[0112] The vehicle risk score is determined according to the empty running rate, bad driving times, working condition intensity, maintenance frequency, positioning deviation frequency, and freight order volume month-on-month drop. Vehicle risk score, the value range is 0~100 points, the higher the score represents the lower the risk of vehicle operation related to the financial leasing, the lower the score the higher the risk. 100 refers to the basic score of vehicle risk score, which is the preset initial full score value, representing the ideal state of the vehicle without any operation risk. Idle rate deduction item, the specific calculation rules are based on historical data calibration of the vehicle fleet, and the idle rate is the ratio of daily idle mileage to total mileage. When the idle rate is greater than 50%, deduct 20 points; when the idle rate is greater than 40% and less than or equal to 50%, deduct 10 points; when the idle rate is less than or equal to 40%, no deduction. Daily average number of hard braking deduction item, hard braking is defined as vehicle deceleration greater than 5 m / s². When the daily average number of hard braking is greater than 8 times, deduct 15 points; when the daily average number of hard braking is greater than 5 times and less than or equal to 8 times, deduct 8 points; when the daily average number of hard braking is less than or equal to 5 times, no deduction. Work condition intensity proportion deduction item, wherein heavy load is defined as vehicle load greater than 80% of rated load, and work condition intensity proportion is the proportion of heavy load operation time in total operation time on mountainous road sections. When the work condition intensity proportion is greater than 80%, deduct 20 points; when the work condition intensity proportion is greater than 70% and less than or equal to 80%, deduct 12 points; when the work condition intensity proportion is less than or equal to 70%, no deduction. Near 30-day maintenance frequency deduction item, when the number of vehicle maintenance in the near 30 days is greater than 3 times, deduct 25 points; when the number of maintenance is greater than 2 times and less than or equal to 3 times, deduct 15 points; when the number of maintenance is less than or equal to 2 times, no deduction. GPS deviation frequency deduction item, deviation is defined as the deviation between the actual driving track of the vehicle and the planned route greater than 1 km. When the daily average number of GPS deviation is greater than 5 times, deduct 18 points; when the daily average number of deviation is greater than 3 times and less than or equal to 5 times, deduct 10 points; when the daily average number of deviation is less than or equal to 3 times, no deduction. Order volume year-on-year drop deduction item, order volume year-on-year drop is the ratio of the difference between last month's order volume and this month's order volume to last month's order volume. When the drop is greater than 40%, deduct 25 points; when the drop is greater than 30% and less than or equal to 40%, deduct 15 points; when the drop is less than or equal to 30%, no deduction.

[0113] For example, it is known that the operation data of a certain electric heavy truck includes: idle rate 45%, daily average hard braking 6 times, mountainous area and heavy load proportion 75%, near 30-day maintenance 2 times, daily average GPS deviation 4 times, order volume year-on-year drop 35%. The calculation process is as follows: Deduct 10 points, Deduct 8 points, Deduct 12 points, Deduct 0 points, Deduct 10 points, Deduct 15 points. The vehicle risk score is Part.

[0114] In the embodiment of the present application, the empty running rate, the number of bad driving, the working condition intensity, the maintenance frequency and the positioning deviation frequency are obtained from the vehicle operation characteristics; the freight order volume month-on-month decrease is obtained from the business association characteristics; the empty running rate, the number of bad driving, the working condition intensity, the maintenance frequency, the positioning deviation frequency and the freight order volume month-on-month decrease are determined as the input characteristics for evaluating multiple vehicles. The present application integrates the empty running rate, the number of bad driving, the working condition intensity, the maintenance frequency, the positioning deviation frequency and the freight order volume month-on-month decrease to obtain the operation efficiency, the driving behavior, the use intensity, the asset health status, the compliance and the income stability of the lessee, thereby improving the accuracy and the forward-looking of the risk assessment, effectively identifying the decrease in the repayment ability and the potential default risk, and supporting the differentiated risk control decision and the active intervention.

[0115] On the basis of the above-mentioned embodiment, the present application further provides a process for determining the battery evaluation input characteristics in the electric heavy truck leasing risk assessment method, Figure 5 A process schematic diagram for determining the battery evaluation input characteristics in the electric heavy truck leasing risk assessment method provided in the embodiment of the present application is shown in FIG. 5. Figure 5 As shown in the step 302, the multiple battery evaluation input characteristics are obtained according to the battery behavior characteristics, the cost control characteristics and the residual value association characteristics, including:

[0116] In step 501, the battery health degree SOH decay rate, the fast charging proportion, the high-temperature charging frequency and the overcharging frequency are obtained from the battery behavior characteristics.

[0117] The battery health degree SOH decay rate is the daily battery SOH difference average value. The fast charging proportion is the daily fast charging frequency proportion. The high-temperature charging frequency is the number of times of charging when the battery temperature is greater than 45 degrees Celsius per week, and the overcharging frequency is the number of times of battery SOC > 98% and > 1 hour per month. The fast charging proportion is high or the fast charging frequency is high, which indicates that the battery decays fast.

[0118] In step 502, the valley electricity charging proportion is obtained from the cost control characteristics.

[0119] The cost control characteristics include the valley electricity charging proportion, the unit electricity cost and the valley electricity time period offset. The valley electricity charging proportion is the daily valley electricity charging amount proportion, the unit electricity cost is the ratio of the daily charging bill and the charging amount, and the valley electricity time period offset is the difference between the actual charging and the valley electricity time period. The low valley electricity charging proportion indicates that the vehicle use cost is high, and the corresponding repayment pressure is large.

[0120] In step 503, the residual value decrease amplitude is predicted according to the residual value association characteristics.

[0121] The residual value-related features include: battery cycle number, high-temperature charging cumulative duration, SOH-cycle number fitting value. The battery cycle number is the daily cycle number of the battery, the high-temperature charging cumulative duration is the monthly high-temperature charging cumulative duration, and the SOH-cycle number fitting value is data obtained by fitting according to historical SOH.

[0122] In step 504, the SOH decay rate, fast charging ratio, high-temperature charging number, overcharging number, valley electricity charging ratio, and predicted residual value drop are determined as a plurality of battery evaluation input features.

[0123] Optionally, the SOH decay rate, fast charging ratio, high-temperature charging number, overcharging number, valley electricity charging ratio, and predicted residual value drop are used as a plurality of battery evaluation input features.

[0124] Optionally, the SOH decay rate, fast charging ratio, high-temperature charging number, overcharging number, valley electricity charging ratio, and predicted residual value drop are used for battery risk assessment on the target electric heavy truck vehicle to obtain a battery risk score.

[0125]

[0126] Wherein: The battery risk score is a value ranging from 0 to 100 points, and the higher the score represents the lower the financing and leasing risk related to the battery state, and the lower the score represents the higher the risk. SOH value deduction item, SOH is the ratio of actual capacity to rated capacity, and is predicted based on a long-short term memory model. When SOH is less than 75%, deduct 30 points; when SOH is greater than or equal to 75% and less than 80%, deduct 20 points; when SOH is greater than or equal to 80% and less than 85%, deduct 10 points; when SOH is greater than or equal to 85%, no deduction. Fast charging ratio deduction item, fast charging ratio is the ratio of fast charging number to total charging number. When the fast charging ratio is greater than 60%, deduct 25 points; when the fast charging ratio is greater than 50% and less than or equal to 60%, deduct 15 points; when the fast charging ratio is less than or equal to 50%, no deduction. Weekly high-temperature charging number deduction item, high-temperature charging is defined as charging when the battery temperature is greater than 45℃. When the weekly high-temperature charging number is greater than 5 times, deduct 20 points; when the weekly high-temperature charging number is greater than 3 times and less than or equal to 5 times, deduct 12 points; when the weekly high-temperature charging number is less than or equal to 3 times, no deduction. Overcharging number deduction item, overcharging is defined as battery SOC greater than 98% and duration greater than 1 hour. When the monthly overcharging number is greater than 3 times, the deduction is: (8x overcharging number) + 10 points; when the monthly overcharging number is less than or equal to 3 times, the deduction is 8x overcharging number points. The valley electricity charging proportion deduction item refers to the ratio of the charging amount in the valley electricity period to the total charging amount. When the valley electricity proportion is less than 30%, 25 points are deducted; when the valley electricity proportion is greater than or equal to 30% and less than 40%, 15 points are deducted; and when the valley electricity proportion is greater than or equal to 40%, no points are deducted. The residual value risk deduction item refers to the residual value reduction based on the SOH attenuation and the two-stage model of cycle times. When the predicted 1-year residual value reduction is greater than 30%, 25 points are deducted; when the predicted 1-year residual value reduction is greater than 20% and less than or equal to 30%, 15 points are deducted; and when the predicted 1-year residual value reduction is less than or equal to 20%, no points are deducted.

[0127] For example, it is known that the battery operation data is: SOH = 82%, fast charging proportion is 55%, high-temperature charging is 4 times a week, overcharging is 3 times a month, valley electricity charging proportion is 35%, and predicted 1-year residual value reduction is 25%. The calculation process is as follows: deduct 10 points, deduct 15 points, deduct 12 points, deduct 24 points, deduct 15 points, deduct 15 points. The battery risk score is points.

[0128] In the embodiments of the present application, the SOH attenuation rate, the fast charging proportion, the high-temperature charging times, and the overcharging times are obtained from the battery behavior characteristics; the valley electricity charging proportion is obtained from the cost control characteristics; the predicted residual value reduction is obtained according to the residual value correlation characteristics; and the SOH attenuation rate, the fast charging proportion, the high-temperature charging times, the overcharging times, the valley electricity charging proportion, and the predicted residual value reduction are determined as multiple battery evaluation input characteristics. The present application reflects the health status, use habit and economy of the battery by fusing the SOH attenuation rate, the fast charging proportion, the high-temperature charging times, the overcharging times, the valley electricity charging proportion, and the predicted residual value reduction, accurately captures the core factors affecting the battery life and asset value, quantifies the influence of the user's cost control ability on the repayment stability, and effectively improves the risk identification accuracy and management initiative of the battery asset in the electric heavy truck financing lease.

[0129] On the basis of the above-mentioned embodiments, the present application further provides a process for determining the battery replacement evaluation input characteristics in the electric heavy truck lease risk assessment method, and the multiple battery replacement evaluation input characteristics are obtained from the battery replacement behavior characteristics in the above-mentioned step 303, including:

[0130] The cumulative battery replacement times, the SOH parameters before and after battery replacement, the battery replacement cost proportion, and the battery replacement complaint rate are obtained from the battery replacement behavior characteristics as the multiple battery replacement evaluation input characteristics.

[0131] Wherein, the cumulative battery replacement frequency is the battery replacement frequency in 30 days, the SOH parameter before and after battery replacement is the average value of the SOH parameter difference before and after battery replacement in a month, the battery replacement cost ratio is the ratio of the total monthly battery replacement cost to the monthly operation cost, and the battery replacement complaint rate is the ratio of the quarterly complaint frequency to the battery replacement frequency.

[0132] Optionally, the target electric heavy truck is subjected to battery replacement risk assessment according to the cumulative battery replacement frequency, the SOH parameter before and after battery replacement, the battery replacement cost ratio, and the battery replacement complaint rate, to obtain a battery replacement risk score.

[0133]

[0134] Wherein: The battery replacement risk score is a value in the range of 0-100, and a higher score represents a lower financing leasing risk related to the battery replacement scene, and a lower score represents a higher risk. 100 is the base score of the battery replacement risk score, which is a preset initial full score value representing an ideal state of no risk and stable operation in the battery replacement process. The 30-day battery replacement frequency deduction item is calculated based on the operation data of the battery swap station. When the 30-day battery replacement frequency is greater than or equal to 8 times, 20 points are deducted; when the 30-day battery replacement frequency is greater than or equal to 5 times and less than 8 times, 12 points are deducted; and when the 30-day battery replacement frequency is less than 5 times, no points are deducted. The SOH difference after battery replacement deduction item is the SOH difference after battery replacement, which is the SOH of the battery after battery replacement minus the SOH of the battery before battery replacement. When the difference is less than -10%, i.e., the SOH of the new battery is lower than that of the original battery, 15 points are deducted; when the difference is greater than or equal to -10% and less than 0, 5 points are deducted; and when the difference is greater than or equal to 0, no points are deducted. The battery replacement cost ratio deduction item is the ratio of the total monthly battery replacement cost to the total monthly operation cost. The calculation rule is: when the ratio is greater than 30%, 18 points are deducted; when the ratio is greater than 25% and less than or equal to 30%, 8 points are deducted; and when the ratio is less than or equal to 25%, no points are deducted. The battery replacement complaint frequency deduction item is the number of complaints related to battery replacement, such as insufficient power and reduced range. When the quarterly battery replacement complaint frequency is greater than or equal to 5 times, 20 points are deducted; when the quarterly battery replacement complaint frequency is greater than or equal to 3 times and less than 5 times, 10 points are deducted; and when the quarterly battery replacement complaint frequency is less than 3 times, no points are deducted.

[0135] For example, it is known that a vehicle has the following battery replacement data: 30-day battery replacement frequency of 6 times, SOH difference after battery replacement of -8%, battery replacement cost ratio of 28%, and quarterly battery replacement complaint frequency of 2 times. The calculation process is as follows: (5 times ≤ 6 times < 8 times) deduct 12 points, (-10% ≤ -8% < 0) deduct 5 points, (25% < 28% ≤ 30%) deduct 8 points, (2 complaints) 0 points deducted. Battery swapping risk score: point.

[0136] In this embodiment, the cumulative number of battery swaps, the State of Health (SOH) parameters before and after swaps, the percentage of swap costs, and the battery swap complaint rate are obtained from battery swapping behavior characteristics as multiple input features for battery swapping evaluation. By using the cumulative number of battery swaps, the SOH difference before and after swaps, the percentage of swap costs, and the battery swap complaint rate as input features for battery swapping evaluation, this application comprehensively reflects the lessee's usage frequency, battery replacement quality, operating cost pressure, and service quality feedback under the battery swapping mode. It effectively identifies asset instability, inefficient energy replenishment, increased economic burden, or potential default tendency caused by frequent battery swapping, thereby improving the accuracy of perception and early warning capability of financial leasing risks in the battery swapping scenario.

[0137] Based on the above embodiments, this application also provides a process for determining the target risk level in another method for risk assessment of electric heavy truck leasing. Figure 6 A flowchart illustrating the determination of the target risk level in another electric heavy-duty truck leasing risk assessment method provided in this application embodiment is shown below. Figure 6 As shown, in step 206 above, the target risk level of the target electric heavy-duty truck is obtained based on the vehicle risk score, battery risk score, battery swapping risk score, and overdue risk probability, including:

[0138] Step 601: If any of the vehicle risk score, battery risk score, and battery swapping risk score is less than the first preset score, or the probability of overdue risk is greater than the first preset probability, then the target risk level is determined to be a high risk level.

[0139] The first preset score is 60 points, and the first preset probability is 80%.

[0140] Optionally, if any of the vehicle risk score, battery risk score, or battery swapping risk score is lower than a first preset score, or if the probability of delinquency risk is greater than a first preset probability, then the target risk level is determined to be high-risk. A corresponding high-risk level label is then assigned to the electric heavy-duty truck, and the label includes a note indicating the risk item whose score is lower than the first preset score. For example, if the vehicle risk score is lower than the first preset score, and the vehicle has an empty-running rate, SOH < 80%, and frequent battery swapping, then the corresponding information is added to the high-risk level label for the target vehicle. Alternatively, if the probability of delinquency risk is greater than the first preset probability, then the delinquency risk value for the target vehicle is added to the high-risk level label.

[0141] Step 602: If all scores in the vehicle risk score, battery risk score, and battery swapping risk score are greater than or equal to the first preset score but less than the second preset score, and the overdue risk probability is greater than or equal to the second preset probability but less than or equal to the first preset probability, then the target risk level is determined to be a medium risk level.

[0142] The second preset score is greater than the first preset score, and the second preset probability is less than the first preset probability. The second preset score is 60-70 points. The second preset probability is 50%-80%.

[0143] Optionally, if all scores in the vehicle risk score, battery risk score, and battery swapping risk score are greater than or equal to a first preset score but less than a second preset score, and the probability of delinquency risk is greater than or equal to the second preset probability but less than or equal to the first preset probability, then the target risk level is determined to be a medium risk level. A corresponding medium risk level label is then set for the electric heavy-duty truck, and the label notes include the risk items corresponding to scores greater than or equal to the first preset score but less than the second preset score. For example, if the battery swapping risk score is greater than or equal to the first preset score but less than the second preset score, and there is insufficient utilization of off-peak electricity and battery swapping costs account for 28%, then the corresponding content is added to the medium risk level label for the target vehicle. Alternatively, if the probability of delinquency risk is greater than or equal to the second preset probability but less than or equal to the first preset probability, then the delinquency risk value of the target vehicle is added to the medium risk level label.

[0144] Step 603: If all scores in the vehicle risk score, battery risk score, and battery swapping risk score are greater than or equal to the second preset score, then the target risk level is determined to be a low risk level.

[0145] Optionally, if all scores in the vehicle risk score, battery risk score, and battery swapping risk score are greater than or equal to the second preset score, the target risk level is determined to be low risk level, indicating that the target vehicle's operational stability cost control is good.

[0146] In the embodiments of the present application, if any of the vehicle risk score, the battery risk score and the battery swap risk score is less than the first preset score, or the overdue risk probability is greater than the first preset probability, the target risk level is determined as a high risk level; if all of the vehicle risk score, the battery risk score and the battery swap risk score are greater than or equal to the first preset score but less than the second preset score, and the overdue risk probability is greater than or equal to the second preset probability but less than or equal to the first preset probability, the target risk level is determined as a medium risk level; if all of the vehicle risk score, the battery risk score and the battery swap risk score are greater than or equal to the second preset score, the target risk level is determined as a low risk level. The present application sets a compound judgment logic of multi-dimensional risk score and overdue risk probability, and constructs a hierarchical, cooperative and dynamically adjustable risk level determination mechanism for determining the target risk level, thereby improving the accuracy and robustness of risk assessment.

[0147] On the basis of the above-mentioned embodiments, the present application further provides a process for determining intervention measures in the electric heavy truck rental risk assessment method, which comprises the following steps on the basis of steps 101-103:

[0148] According to the target risk level, the corresponding intervention measures are generated and output.

[0149] Optionally, when the vehicle and the battery are of the same capital party, if the risk level is a low risk level, an operation optimization suggestion is pushed to the corresponding electric heavy truck, for example, the proportion of valley electricity charging is improved to 50%. If the risk level is a medium risk level, a risk message is sent to the lessee, the entertainment system of the vehicle is frozen, and if there is no improvement for 7 days, the speed is limited to 60 km / h. If there is no improvement, the risk level is not improved. If the risk level is a high risk level, the BMS is immediately locked, and the vehicle is locked and the on-site collection is performed if there is no repayment for 24 hours. The locking operation can control the engine speed below 800 rpm. When the vehicle and the battery are of different capital parties, if the risk level is a low risk level, an operation optimization suggestion is pushed to the vehicle body of the corresponding electric heavy truck, and the risk level information of the vehicle is sent to the battery party and the battery swap capital party. If the risk level is a medium risk level, the vehicle body, the battery and the battery swap party are confirmed to intervene within 48 hours, and the vehicle body speed limiting and the battery current limiting are simultaneously performed. The battery current limiting can reduce the charging power to 30%. If the risk level is a high risk level, the battery is forcibly locked without the confirmation of the capital party, and the vehicle body capital party needs to lock the vehicle within 24 hours and the battery swap station needs to suspend the battery swap permission of the vehicle within 12 hours after the intervention.

[0150] Optionally, a unified intervention control platform is established, which will serve as the only dispatch entrance for all intervention actions, ensuring that all actions are executed through the platform, whether it is locking the car, locking the power, limiting the speed, or suspending the battery replacement permission, all through a standardized interface. The platform will provide a RESTful API and a message queue dual channel, supporting synchronous and asynchronous operations, ensuring that each intervention instruction is executed only once, avoiding repeated execution. Through a unified entrance, the dispersion of execution is eliminated, ensuring the accuracy of each intervention action.

[0151] Optionally, when the vehicle and the battery belong to the same capital party, single-party confirmation can execute the intervention action; when the vehicle and the battery belong to different capital parties, multi-party confirmation is required, each intervention request is accompanied by a risk score and a trigger reason, and enters a multi-signature pool for capital party confirmation. If confirmation is not completed within 48 hours, the system automatically degrades to a reminder intervention. For extremely high-risk scenarios, single-party execution is allowed, and within 12 hours, a traceable record is completed. Among them, the extremely high-risk scenario can be a 95% overdue probability.

[0152] Optionally, when the car locking operation is responsible for the vehicle capital party, the power locking is responsible for the battery capital party, and the battery replacement is responsible for the battery replacement station. Each intervention action will be accompanied by responsibility subject information when executed, ensuring that it can be traced back afterwards, clearly defining responsibility division, and binding related capital parties in the intervention action to avoid disputes caused by unclear responsibilities.

[0153] Optionally, the battery replacement station, BMS and other devices may use different protocols, and the platform will convert these different protocol standards into a unified standard interface through a protocol adapter, so that all parties can be compatible, shielding interface differences. For example, OpenAPI3.0 standard is used to define the interface, and an adaptation layer is provided to quickly connect third-party manufacturers, ensuring that the platform and external devices can be seamlessly connected, reducing the execution problems caused by interface heterogeneity.

[0154] Optionally, each intervention action must have detailed audit logs, including intervention action ID, trigger time, risk level, responsible subject, execution result, etc. All logs will be stored in the local database and the sidechain of the blockchain at the same time, ensuring that the data cannot be tampered with. Regular audits are also conducted to ensure the compliance and legality of all intervention operations. This effectively avoids disputes caused by improper operations or data tampering, while providing data support for subsequent optimization.

[0155] Optionally, an automatic rollback mechanism is executed when the intervention operation fails. If the executed intervention operation fails to be successfully executed due to system failure or data misjudgment, the related restrictions or the vehicle state will be automatically released within 2 hours. Each failed operation must be marked as an exception and recorded in the log for subsequent analysis and optimization. Manual review can be performed to ensure that the vehicle state can be restored in time in an emergency. The execution failure can be an interface timeout or communication failure.

[0156] Optionally, the newly introduced intervention action must go through a gray release process. A small number of vehicle fleets can be used for pilot testing to verify the safety and execution effect of the new operation. During the gray release period, the execution strategy is adjusted for promotion according to feedback. Clear evaluation indicators such as confirmation timeliness, intervention success rate, and responsibility dispute rate are defined to measure the efficiency and quality of collaborative execution. Through regular evaluation of the indicators, the intervention mechanism can be continuously optimized.

[0157] Optionally, to ensure the safety of the intervention action, an intervention action grading system can be established. According to the risk level, the intervention action is divided into three categories: mild, moderate, and severe. Mild actions include risk message reminders and background markers, which do not produce actual control over the vehicle; moderate actions such as speed or flow limits may affect vehicle speed and charging power; and severe actions such as power lock, vehicle lock, and temporary suspension of power swapping rights are the final mandatory interventions. According to real-time scoring and risk judgment, the corresponding intervention action is determined to ensure that each intervention matches the current risk level, thereby avoiding excessive intervention.

[0158] Optionally, to avoid triggering strong actions in inappropriate scenarios, clear safety scenario constraints are defined for each intervention action. For example, when the vehicle is traveling at a speed exceeding 20 km / h, lock power or lock vehicle operations are not allowed; during the transportation of dangerous goods, only reminder-type or background frozen power swapping rights are allowed, and no mandatory intervention is taken; when driving on special road sections such as tunnels and bridges, mandatory intervention must be avoided. These safety constraints will be judged in real time according to the vehicle location, driving state, and task type to ensure that intervention only occurs in safe and controllable scenarios.

[0159] Optionally, when executing the lock power or flow limiting operation, the platform needs to ensure that the vehicle can maintain sufficient emergency power. Specifically, the platform will reserve at least 10% of the power to ensure that the vehicle can drive away from the current location and find the nearest charging station or power supplement point. If the vehicle is in a power supplement scarce area, the reserved power is increased to 20%, and the power supplement scarce area can be a mountainous or desert area. The emergency power guarantee threshold is dynamically adjusted according to the vehicle's current location, nearby charging facility distribution, historical charging records, and other data to ensure that the vehicle always has sufficient endurance when executing intervention.

[0160] Optionally, to avoid misjudgment in emergency situations, the platform will introduce an execution delay before strong action is taken, for example, a 30-minute countdown. During this period, the driver will receive multi-channel reminders, including car interface, SMS, phone, etc. If the driver completes the repayment or contacts the capital party within the countdown, the system will automatically cancel the intervention to avoid unnecessary forced measures. A cooling-off period mechanism is also established. Within 7 days after intervention, if the risk situation improves, the system will automatically revoke the restriction measures.

[0161] Optionally, to avoid security risks caused by misjudgment of a single data source, the platform will implement multi-source data cross-validation. For example, before executing the lock power or lock car, the platform will cross-verify multiple data sources such as BMS, battery management system, vehicle T-Box, charging station, etc. to ensure data consistency. If there is a significant difference in data, it can be reviewed manually by risk control personnel to ensure the accuracy and reasonableness of the intervention operation.

[0162] Optionally, to ensure the safety of the intervention operation, the platform will set key safety indicators, such as the accident rate caused by intervention must be less than 0.1%, and the user complaint rate after intervention must be less than 5%. These indicators will continuously optimize the intervention strategy through real-time data monitoring and historical event analysis. For example, if a certain intervention method leads to a higher accident rate or user complaint rate, the platform will adjust the corresponding trigger conditions and execution process to ensure that the intervention strategy always complies with the principle of safety first.

[0163] In the embodiments of the present application, corresponding intervention measures are generated and output according to the target risk level. The present application dynamically generates and outputs differentiated intervention measures according to the target risk level, realizing closed-loop management from risk identification to precise response, which not only guarantees asset safety in high-risk scenarios, but also avoids over-intervention on low-risk users, balances risk control effectiveness and operational safety, and significantly improves the risk disposal efficiency and intelligent level of the financing lease business.

[0164] Based on the same inventive concept, the embodiments of the present application also provide an electric heavy truck rental risk assessment device corresponding to the electric heavy truck rental risk assessment method. Since the principle of the device in the embodiments of the present application solves the problem similar to the electric heavy truck rental risk assessment method of the above-mentioned embodiments of the present application, the implementation of the device can be referred to the implementation of the method, and the repeated parts will not be described again.

[0165] Figure 7 The structure diagram of an electric heavy truck rental risk assessment device provided in the embodiments of the present application is shown in Figure 7 As shown in the figure, the device includes:

[0166] The acquisition module 701 is configured to acquire real-time vehicle operation data, real-time battery state data, energy supplement behavior data, battery swap behavior data, and external application data of the target electric heavy truck vehicle, wherein the external application data at least includes regional peak-valley electricity price table, weather data, maintenance data, and freight service data.

[0167] The extraction module 702 is configured to perform feature extraction on the real-time vehicle operation data, the real-time battery state data, the energy supplement behavior data, and the battery swap behavior data, to obtain multi-class feature data of the target electric heavy truck vehicle.

[0168] The evaluation module 703 is configured to perform rental risk evaluation on the target electric heavy truck vehicle according to the multi-class feature data, to obtain a target risk level of the target electric heavy truck vehicle.

[0169] In a possible implementation, the evaluation module 703 is specifically configured to: acquire, from the multi-class feature data, a plurality of vehicle evaluation input features, a plurality of battery evaluation input features, and a plurality of battery swap evaluation input features.

[0170] perform vehicle risk evaluation on the target electric heavy truck vehicle according to the plurality of vehicle evaluation input features, to obtain a vehicle risk score;

[0171] perform battery risk evaluation on the target electric heavy truck vehicle according to the plurality of battery evaluation input features, to obtain a battery risk score;

[0172] perform battery swap risk evaluation on the target electric heavy truck vehicle according to the plurality of battery swap evaluation input features, to obtain a battery swap risk score;

[0173] perform overdue probability prediction according to the multi-class feature data by using a preset overdue prediction model, to obtain an overdue risk probability;

[0174] obtain the target risk level of the target electric heavy truck vehicle according to the vehicle risk score, the battery risk score, the battery swap risk score, and the overdue risk probability.

[0175] In a possible implementation, the multi-class feature data includes vehicle operation features, battery behavior features, cost control features, battery swap behavior features, residual value correlation features, and service correlation features; and the evaluation module 703 is specifically configured to: acquire, from the vehicle operation features and the service correlation features, the plurality of vehicle evaluation input features.

[0176] acquire, from the battery behavior features, the cost control features, and the residual value correlation features, the plurality of battery evaluation input features;

[0177] acquire, from the battery swap behavior features, the plurality of battery swap evaluation input features.

[0178] In a possible implementation, the vehicle operation features include: empty running rate, bad driving times, working condition intensity, maintenance frequency, and positioning deviation frequency; the business correlation features include: freight order volume month-on-month decrease; the evaluation module 703 is specifically configured to: obtain the empty running rate, the bad driving times, the working condition intensity, the maintenance frequency, and the positioning deviation frequency from the vehicle operation features;

[0179] obtain the freight order volume month-on-month decrease from the business correlation features;

[0180] determine the empty running rate, the bad driving times, the working condition intensity, the maintenance frequency, the positioning deviation frequency, and the freight order volume month-on-month decrease as the plurality of vehicle evaluation input features.

[0181] In a possible implementation, the evaluation module 703 is specifically configured to: obtain, from the battery behavior features, a battery health degree SOH attenuation rate, a fast charging proportion, a high-temperature charging times, and an overcharging times;

[0182] obtain, from the cost control features, a valley electricity charging proportion;

[0183] perform residual value decrease prediction according to the residual value correlation features to obtain a predicted residual value decrease;

[0184] determine the SOH attenuation rate, the fast charging proportion, the high-temperature charging times, the overcharging times, the valley electricity charging proportion, and the predicted residual value decrease as the plurality of battery evaluation input features.

[0185] In a possible implementation, the evaluation module 703 is specifically configured to: obtain, from the battery swap behavior features, a cumulative battery swap times, a SOH parameter before and after battery swap, a battery swap cost proportion, and a battery swap complaint rate as the plurality of battery swap evaluation input features.

[0186] In a possible implementation, the evaluation module 703 is specifically configured to: if any one of the vehicle risk score, the battery risk score, and the battery swap risk score is less than a first preset score, or the overdue risk probability is greater than a first preset probability, determine that the target risk level is a high risk level;

[0187] if all of the vehicle risk score, the battery risk score, and the battery swap risk score are greater than or equal to the first preset score but less than a second preset score, and the overdue risk probability is greater than or equal to a second preset probability but less than or equal to the first preset probability, determine that the target risk level is a medium risk level; the second preset score is greater than the first preset score, and the second preset probability is less than the first preset probability;

[0188] if all of the vehicle risk score, the battery risk score, and the battery swap risk score are greater than or equal to the second preset score, determine that the target risk level is a low risk level.

[0189] In one possible implementation, the device further includes an intervention module, which is specifically used to generate and output corresponding intervention measures based on the target risk level.

[0190] The processing flow of each module in the device and the interaction flow between each module can be referred to the relevant descriptions in the above method embodiments, and will not be detailed here.

[0191] This application also provides an electronic device. Figure 8 This application provides a schematic diagram of an electronic device structure, such as... Figure 8 As shown, the electronic device includes a processor 801 and a memory 802, and optionally, a bus 803. The memory 802 stores machine-readable instructions executable by the processor 801. When the electronic device is running, the processor 801 and the memory 802 communicate via the bus 803. When the machine-readable instructions are executed by the processor 801, the steps of the above-described electric heavy truck leasing risk assessment method are performed.

[0192] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the above-described electric heavy truck rental risk assessment method.

[0193] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some communication interfaces; the indirect coupling or communication connection of devices or modules can be electrical, mechanical, or other forms.

[0194] In addition, each function unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. When the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the parts of the technical solutions that are essential or contribute to the prior art can be embodied in the form of software products, which are stored in a storage medium and include a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0195] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application.

Claims

1. An electric heavy truck rental risk assessment method, characterized in that, The method includes: The system acquires real-time vehicle operation data, real-time battery status data, energy replenishment behavior data, battery swapping behavior data, and external application data of the target electric heavy-duty truck. The external application data includes at least: regional peak and valley electricity price tables, weather data, maintenance data, and freight business data. Feature extraction is performed on the real-time vehicle operation data, the real-time battery status data, the energy replenishment behavior data, and the battery swapping behavior data to obtain multiple types of feature data of the target electric heavy-duty truck. Multiple vehicle evaluation input features, multiple battery evaluation input features, and multiple battery swapping evaluation input features are obtained from the aforementioned multi-type feature data; Based on the multiple vehicle assessment input features, a vehicle risk assessment is performed on the target electric heavy-duty truck to obtain a vehicle risk score. Based on the multiple battery assessment input features, a battery risk assessment is performed on the target electric heavy-duty truck to obtain a battery risk score. Based on the multiple battery swapping assessment input features, a battery swapping risk assessment is performed on the target electric heavy-duty truck to obtain a battery swapping risk score. Based on the aforementioned multi-type feature data, a preset overdue prediction model is used to predict the probability of overdue payments, thereby obtaining the probability of overdue risk. The preset overdue prediction model is a preset large language model or neural network model, and the probability of overdue risk is the likelihood of the lessee defaulting on repayments. The target risk level of the target electric heavy-duty truck is obtained based on the vehicle risk score, the battery risk score, the battery swapping risk score, and the overdue risk probability.

2. The method of claim 1, wherein, The various types of feature data include: vehicle operation features, battery behavior features, cost control features, battery swapping behavior features, residual value correlation features, and business correlation features; The step of obtaining multiple vehicle evaluation input features, multiple battery swapping evaluation input features, and multiple battery swapping evaluation input features from the multiple types of feature data includes: The multiple vehicle evaluation input features are obtained from the vehicle operation features and the business association features; Based on the battery behavior characteristics, the cost control characteristics, and the residual value correlation characteristics, the multiple battery evaluation input characteristics are obtained; The multiple battery swapping evaluation input features are obtained from the battery swapping behavior features.

3. The method of claim 2, wherein, The vehicle operation characteristics include: empty-run rate, number of poor driving incidents, workload intensity, maintenance frequency, and location deviation frequency; the business-related characteristics include: month-on-month decrease in freight order volume; The step of obtaining the multiple vehicle evaluation input features from the vehicle operation features and the business association features includes: The empty-run rate, number of bad driving incidents, working conditions intensity, maintenance frequency, and positioning deviation frequency are obtained from the vehicle operation characteristics. Obtain the month-on-month decrease in freight order volume from the aforementioned business-related characteristics; The empty-run rate, the number of bad driving incidents, the working conditions intensity, the maintenance frequency, the positioning deviation frequency, and the month-on-month decrease in freight order volume are determined as the multiple vehicle evaluation input features.

4. The method of claim 2, wherein, The step of obtaining the multiple battery evaluation input features based on the battery behavior features, the cost control features, and the residual value correlation features includes: The battery health status, SOH degradation rate, fast charging rate, high-temperature charging cycles, and overcharging cycles are obtained from the battery behavior characteristics. Obtain the off-peak electricity charging ratio from the aforementioned cost control characteristics; Based on the residual value correlation characteristics, the residual value reduction rate is predicted to obtain the predicted residual value reduction rate; The SOH decay rate, the fast charging ratio, the number of high-temperature charging cycles, the number of overcharge cycles, the off-peak charging ratio, and the predicted residual value reduction are determined as the multiple battery evaluation input features.

5. The method of claim 2, wherein, The step of obtaining the multiple battery swapping evaluation input features from the battery swapping behavior features includes: The cumulative number of battery swaps, the state of health (SOH) parameters before and after battery swaps, the percentage of battery swap costs, and the battery swap complaint rate are obtained from the battery swapping behavior characteristics and used as the multiple battery swapping evaluation input features.

6. The method according to claim 1, characterized in that, The step of obtaining the target risk level of the target electric heavy-duty truck based on the vehicle risk score, the battery risk score, the battery swapping risk score, and the overdue risk probability includes: If any of the vehicle risk score, the battery risk score, and the battery swapping risk score is less than a first preset score, or if the overdue risk probability is greater than a first preset probability, then the target risk level is determined to be a high-risk level. If all scores in the vehicle risk score, the battery risk score, and the battery swapping risk score are greater than or equal to the first preset score but less than the second preset score, and the overdue risk probability is greater than or equal to the second preset probability but less than or equal to the first preset probability, then the target risk level is determined to be a medium risk level; wherein the second preset score is greater than the first preset score, and the second preset probability is less than the first preset probability. If all scores in the vehicle risk score, the battery risk score, and the battery swapping risk score are greater than or equal to the second preset score, then the target risk level is determined to be a low risk level.

7. The method according to claim 1, characterized in that, The method further includes: Based on the target risk level, generate and output corresponding intervention measures.

8. A risk assessment device for electric heavy-duty truck leasing, characterized in that, The device includes: The acquisition module is used to acquire real-time vehicle operation data, real-time battery status data, energy replenishment behavior data, battery swapping behavior data, and external application data of the target electric heavy-duty truck. The external application data includes at least: regional peak and valley electricity price table, weather data, maintenance data, and freight business data. The extraction module is used to extract features from the real-time vehicle operation data, the real-time battery status data, the energy replenishment behavior data, and the battery swapping behavior data to obtain multiple types of feature data of the target electric heavy-duty truck. The assessment module is used to assess the leasing risk of the target electric heavy-duty truck based on the multi-type feature data, and obtain the target risk level of the target electric heavy-duty truck. The evaluation module is specifically used to: obtain multiple vehicle evaluation input features, multiple battery evaluation input features, and multiple battery swapping evaluation input features from the multiple types of feature data; Based on the multiple vehicle assessment input features, a vehicle risk assessment is performed on the target electric heavy-duty truck to obtain a vehicle risk score. Based on the multiple battery assessment input features, a battery risk assessment is performed on the target electric heavy-duty truck to obtain a battery risk score. Based on the multiple battery swapping assessment input features, a battery swapping risk assessment is performed on the target electric heavy-duty truck to obtain a battery swapping risk score. Based on the aforementioned multi-type feature data, a preset overdue prediction model is used to predict the probability of overdue payments, thereby obtaining the probability of overdue risk. The preset overdue prediction model is a preset large language model or neural network model, and the probability of overdue risk is the likelihood of the lessee defaulting on repayments. The target risk level of the target electric heavy-duty truck is obtained based on the vehicle risk score, the battery risk score, the battery swapping risk score, and the overdue risk probability.

9. An electronic device, characterized in that, include: The processor and memory, the memory storing machine-readable instructions executable by the processor, which, when the computer device is running, execute the machine-readable instructions to perform the steps of the electric heavy truck leasing risk assessment method as described in any one of claims 1 to 7.

Images