Simulation test method, device, equipment and product based on vehicle model

Abstract

The invention provides a simulation test method and device based on a vehicle model, equipment and a product. The method comprises the steps that a power distribution strategy of a hybrid power vehicle is acquired; performing a simulation test on the power distribution strategy through a first engine model and a first motor model in the physical model of the hybrid vehicle to obtain first predicted fuel consumption and first predicted motor electric energy; based on a second engine model and a second motor model, simulation testing is conducted on the power distribution strategy, second predicted oil consumption and second predicted motor electric energy are obtained, and the second engine model and the second motor model are data driving models obtained based on historical data training; fusing the first predicted fuel consumption and the second predicted fuel consumption to obtain target predicted fuel consumption; and fusing the first predicted motor electric energy and the second predicted motor electric energy to obtain target predicted motor electric energy. Therefore, in combination with the physical model and the data driving model of the vehicle, the accuracy of vehicle energy consumption prediction is improved.

Description

Technical Field

[0001] This application relates to the field of vehicle technology, and in particular to a simulation testing method, apparatus, equipment and product based on a vehicle model. Background Technology

[0002] With the escalating global energy crisis and environmental pollution, intelligent energy management of vehicles has gradually become a hot research topic in the automotive industry. Intelligent energy management can not only optimize vehicle energy efficiency, improve vehicle range and power performance, but also reduce vehicle exhaust emissions.

[0003] Vehicle modeling, as one of the core technologies of intelligent energy management, provides theoretical support for vehicle energy consumption calculation, power distribution, and energy-saving optimization. Commonly used vehicle modeling methods include physical modeling and empirical modeling. Physical modeling is based on physical principles, analyzing the vehicle's dynamics, thermodynamics, and electrical characteristics to establish a physical model. This method relies on a large number of physical parameters and experimental data, making the modeling process complex, difficult to perform real-time calculations, and exhibiting poor model accuracy in complex and changing driving environments. Empirical modeling is based on the experience of vehicle developers and vehicle experimental data, using statistical analysis or machine learning methods to establish an empirical model of the vehicle. This method is limited by the experience of vehicle developers, the quality of vehicle experimental data, and the coverage of vehicle experimental data, and also suffers from insufficient model accuracy.

[0004] In summary, all of the above vehicle models suffer from insufficient accuracy in energy consumption prediction. Summary of the Invention

[0005] Based on the above-mentioned technological status, this application provides a simulation testing method, device, equipment and product based on a vehicle model, which can improve the accuracy of vehicle energy consumption prediction by establishing a hybrid vehicle model that combines the advantages of a physical model of vehicle energy consumption and the advantages of a data-driven model of vehicle energy consumption.

[0006] To achieve the above-mentioned technical objectives, this application proposes the following technical solution:

[0007] According to a first aspect of this application, a simulation testing method based on a vehicle model is provided, comprising: acquiring a power distribution strategy of a hybrid vehicle; performing simulation testing on the power distribution strategy using a first engine model and a first motor model to obtain a first predicted fuel consumption and a first predicted motor energy, wherein the first engine model and the first motor model are respectively the engine model and the motor model in the physical model of the hybrid vehicle; performing simulation testing on the power distribution strategy based on a second engine model and a second motor model to obtain a second predicted fuel consumption and a second predicted motor energy, wherein the second engine model and the second motor model are data-driven models trained based on historical data; fusing the first predicted fuel consumption and the second predicted fuel consumption to obtain a target predicted fuel consumption under the power distribution strategy; and fusing the first predicted motor energy and the second predicted motor energy to obtain a target predicted motor energy under the power distribution strategy.

[0008] In some implementations, fusing the first predicted fuel consumption and the second predicted fuel consumption to obtain the target predicted fuel consumption under the power distribution strategy includes: weighting the first predicted fuel consumption and the second predicted fuel consumption according to the weight parameters corresponding to the first engine model and the second engine model to obtain the target predicted fuel consumption; and / or, fusing the first predicted motor energy and the second predicted motor energy to obtain the target predicted motor energy under the power distribution strategy includes: weighting the first predicted motor energy and the second predicted motor energy according to the weight parameters corresponding to the first motor model and the second motor model to obtain the target predicted motor energy.

[0009] In some implementations, the weight parameters corresponding to the first engine model and the second engine model support dynamic updates. The dynamic update process for the weight parameters corresponding to the first engine model and the second engine model includes: acquiring first actual driving data of the hybrid vehicle; determining a fuel consumption prediction error based on the first actual driving data, the first engine model, and the second engine model; updating the weight parameters corresponding to the first engine model and the second engine model based on the fuel consumption prediction error; and / or, the weight parameters corresponding to the first motor model and the second motor model support dynamic updates. The dynamic update process for the weight parameters corresponding to the first engine model and the second engine model includes: acquiring second actual driving data of the hybrid vehicle; determining an energy prediction error based on the second actual driving data, the first motor model, and the second motor model; updating the weight parameters corresponding to the first motor model and the second motor model based on the energy prediction error.

[0010] In some implementations, the fuel consumption prediction error includes the instantaneous fuel consumption prediction error of the first engine model and the instantaneous fuel consumption prediction error of the second engine model. Updating the weight parameters corresponding to the first engine model and the second engine model based on the fuel consumption prediction error includes: determining updated weight parameters corresponding to the first engine model and the second engine model based on the instantaneous fuel consumption prediction errors of the first and second engine models; determining the updated weight parameters corresponding to the first engine model as the weight parameters corresponding to the first engine model, and determining the updated weight parameters corresponding to the second engine model as the weight parameters corresponding to the second engine model.

[0011] In some implementations, determining the updated weight parameter corresponding to the first engine model as the weight parameter corresponding to the first engine model, and determining the updated weight parameter corresponding to the second engine model as the weight parameter corresponding to the second engine model, includes: determining the cumulative fuel consumption error obtained by combining the first engine model and the second engine model for fuel consumption prediction based on the instantaneous fuel consumption prediction error of the first engine model, the instantaneous fuel consumption prediction error of the second engine model, and the actual fuel consumption in the reference period in the first actual driving data; if the cumulative fuel consumption error is less than the fuel consumption error threshold, determining the updated weight parameter corresponding to the first engine model as the weight parameter corresponding to the first engine model, and determining the updated weight parameter corresponding to the second engine model as the weight parameter corresponding to the second engine model.

[0012] In some implementations, after determining the cumulative fuel consumption error obtained by combining the first engine model and the second engine model for fuel consumption prediction, the method further includes: if the cumulative fuel consumption error is greater than or equal to the fuel consumption error threshold, re-determining the update weight parameters corresponding to the first engine model and the update weight parameters corresponding to the second engine model, and / or retraining the second engine model.

[0013] In some implementations, the power distribution strategy includes allocated engine speed and allocated engine output torque. The power distribution strategy is simulated and tested using the first engine model to obtain the first predicted fuel consumption. This includes: determining the predicted engine specific fuel consumption based on the allocated engine speed and allocated engine output torque, where the predicted engine specific fuel consumption is the engine specific fuel consumption when the hybrid vehicle's engine operates at the allocated engine speed and allocated engine output torque; and inputting the allocated engine speed, allocated engine output torque, and predicted engine specific fuel consumption into the first engine model to obtain the first predicted fuel consumption.

[0014] In some implementations, determining the predicted engine specific fuel consumption based on the allocated engine speed and the allocated engine output torque includes: interpolating the engine specific fuel consumption from engine test data based on the allocated engine speed and the allocated engine output torque; or, inputting the allocated engine speed and the allocated engine output torque into a specific fuel consumption mapping model to obtain the predicted engine specific fuel consumption; wherein the engine test data includes multi-dimensional data consisting of engine speed, engine output torque, and engine specific fuel consumption, and the specific fuel consumption mapping model is a model of the mapping relationship between engine speed, engine output torque, and engine specific fuel consumption, and the specific fuel consumption mapping model is obtained by fitting based on the engine test data.

[0015] In some implementations, the power distribution strategy includes allocated motor speed and allocated motor output torque. The power distribution strategy is simulated and tested using the first motor model to obtain the first predicted motor energy. This includes: determining the predicted motor charging efficiency and predicted motor discharging efficiency based on the allocated motor speed and allocated motor output torque. The predicted motor charging efficiency and predicted motor discharging efficiency are the charging efficiency and discharging efficiency of the hybrid vehicle's motor when operating at the allocated motor speed and allocated motor output torque; inputting the allocated motor speed, allocated motor output torque, and predicted motor charging efficiency into the motor charging model of the first motor model to obtain the predicted charging energy in the first predicted motor energy; and inputting the allocated motor speed, allocated motor output torque, and predicted motor discharging efficiency into the motor discharging model of the first motor model to obtain the predicted discharging energy in the first predicted motor energy.

[0016] In some implementations, determining the predicted motor charging efficiency and predicted motor discharging efficiency based on the allocated motor speed and the allocated motor output torque includes: interpolating the predicted motor charging efficiency and predicted motor discharging efficiency from motor test data based on the allocated motor speed and the allocated motor output torque; or, inputting the allocated motor speed and the allocated motor output torque into an efficiency mapping model to obtain the predicted motor charging efficiency and predicted motor discharging efficiency; wherein the motor test data includes multi-dimensional data consisting of motor speed, motor output torque, motor charging efficiency, and motor discharging efficiency, and the efficiency mapping model is a model of the mapping relationship between motor speed, motor output torque, motor charging efficiency, and motor discharging efficiency, and the efficiency mapping model is obtained by fitting based on the motor test data.

[0017] In some implementations, the physical model further includes a battery model. After obtaining the power distribution strategy of the hybrid vehicle, the method further includes: performing simulation tests on the power distribution strategy using the battery model to obtain a predicted battery SOC. In some implementations, the model parameters of the battery model include a rated capacitance, which supports dynamic updates. The updating process of the rated capacitance includes: obtaining third actual driving data of the hybrid vehicle; determining the actual capacitance based on the third actual driving data; comparing the actual capacitance with the rated capacitance to obtain a comparison result; and updating the rated capacitance based on the comparison result.

[0018] In some implementations, the physical model further includes a vehicle driving model, and the step of obtaining the power distribution strategy for the hybrid vehicle includes: obtaining vehicle information of the hybrid vehicle and the driver's required torque of the hybrid vehicle, wherein the vehicle information includes first input data required by the vehicle driving model and second input data required by the strategy model; obtaining the predicted driving speed of the hybrid vehicle through the vehicle driving model based on the first input data; and generating the power distribution strategy through the strategy model based on the predicted driving speed, the driver's required torque, and the second input data.

[0019] In some implementations, the drag coefficient in the vehicle driving model supports dynamic updating. The dynamic updating process of the drag coefficient includes: acquiring fourth actual driving data of the hybrid vehicle; determining the vehicle driving prediction error of the vehicle driving model for the hybrid vehicle based on the fourth actual driving data and the vehicle driving model; if it is determined that the drag coefficient needs to be updated based on the vehicle driving prediction error, determining the updated value of the drag coefficient using a parameter identification method; and updating the drag coefficient based on the updated value.

[0020] In some implementations, after fusing the first predicted motor energy and the second predicted motor energy to obtain the target predicted motor energy under the power distribution strategy, the method further includes: training the strategy model based at least on the target predicted fuel consumption and the target predicted motor energy to obtain the strategy model after the current training.

[0021] According to a second aspect of this application, a vehicle model-based simulation testing device is provided, comprising: a strategy acquisition unit for acquiring a power distribution strategy for a hybrid vehicle; a first testing unit for simulating and testing the power distribution strategy using a first engine model and a first motor model to obtain a first predicted fuel consumption and a first predicted motor energy, wherein the first engine model and the first motor model are physical models; a second testing unit for simulating and testing the power distribution strategy based on a second engine model and a second motor model to obtain a second predicted fuel consumption and a second predicted motor energy, wherein the second engine model and the second motor model are data-driven models trained based on historical data; a fuel consumption fusion unit for fusing the first predicted fuel consumption and the second predicted fuel consumption to obtain a target predicted fuel consumption under the power distribution strategy; and an energy fusion unit for fusing the first predicted motor energy and the second predicted motor energy to obtain a target predicted motor energy under the power distribution strategy.

[0022] According to a third aspect of this application, an electronic device is provided, including a memory and a processor; the memory is connected to the processor and is used to store a program; the processor is used to implement the vehicle model-based simulation testing method as described in the first aspect or any implementation thereof by running the program in the memory.

[0023] According to a fourth aspect of this application, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the vehicle model-based simulation testing method as described in the first aspect or any implementation thereof.

[0024] According to a fifth aspect of this application, a storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the vehicle model-based simulation testing method as described in the first aspect or any implementation thereof.

[0025] This application provides a simulation testing method, apparatus, equipment, and product based on a vehicle model. Using a first engine model, a first motor model, a second engine model, and a second motor model, the power distribution strategy of a hybrid vehicle is simulated and tested to obtain a first predicted fuel consumption, a first predicted motor energy consumption, a second predicted fuel consumption, and a second predicted motor energy consumption. The first and second predicted fuel consumptions are fused to obtain a target predicted fuel consumption, and the first and second predicted motor energy consumptions are fused to obtain a target predicted motor energy consumption. The first engine model and the first motor model are the engine and motor models in the physical model of the hybrid vehicle, while the second engine model and the second motor model are data-driven models trained based on historical data. The first engine model and the first motor model, as physical models, and the second engine model and the second motor model, as data-driven models, leverage their respective advantages and achieve model complementarity in predicting energy consumption for the power distribution strategy, thereby improving the accuracy of vehicle energy consumption prediction. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0027] Figure 1 A flowchart illustrating a simulation testing method based on a vehicle model, provided for an embodiment of this application;

[0028] Figure 2 A flowchart illustrating the dynamic update process of the weight parameters corresponding to the first engine model and the weight parameters corresponding to the second engine model in a vehicle model-based simulation testing method provided in this application embodiment;

[0029] Figure 3 This is a flowchart illustrating the dynamic update process of the weight parameters corresponding to the physical model and the weight parameters corresponding to the data-driven model.

[0030] Figure 4 Example diagram of model training for the second engine model;

[0031] Figure 5 Example diagram of model training for the second motor model;

[0032] Figure 6 This is a flowchart illustrating the dynamic update process of the rated capacitance in a battery model.

[0033] Figure 7An example diagram illustrating energy flow in hybrid vehicles;

[0034] Figure 8 Example diagrams showing the physical model, its input data, and its output data;

[0035] Figure 9 A flowchart illustrating the dynamic updating of model parameters in a vehicle driving model;

[0036] Figure 10 An example diagram of an intelligent energy management system for hybrid vehicles;

[0037] Figure 11 This is a schematic diagram of the structure of a simulation testing device based on a vehicle model provided in an embodiment of this application;

[0038] Figure 12 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0039] The technical solution proposed in this application is applicable to the energy consumption prediction scenario of hybrid vehicles, aiming to improve the accuracy of energy consumption prediction for hybrid vehicles through vehicle models. By employing the technical solution described in this application, the physical model and data-driven model of the hybrid vehicle can be combined to simulate and test the power distribution strategy of the hybrid vehicle, achieving dual-model fusion for vehicle energy consumption prediction and improving the accuracy of vehicle energy consumption prediction.

[0040] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0041] In related technologies, vehicle models are typically limited to either physical or empirical models, which are used for driving simulation. However, these models suffer from flaws in their modeling process, resulting in poor generalization and low accuracy in practical applications.

[0042] For example, based on fixed conditions, a physical or empirical model formula is designed, and then interpolation is performed using traditional interpolation modeling methods and experimental data to determine the parameters of the model formula. In this modeling approach, the physical or empirical model has poor generalization ability and is difficult to adapt to complex and ever-changing real-world driving environments. Furthermore, the interpolation modeling method relies on discrete points in the experimental data, making it difficult to capture the continuous changes in the vehicle's behavior in actual driving scenarios, resulting in insufficient model accuracy.

[0043] For example, by using machine learning and deep learning technologies, effective features can be extracted and learned from massive amounts of vehicle operation data to build a data-driven model of the vehicle. Compared to a physical model, a data-driven model built in this way may achieve higher accuracy. However, data-driven model building requires the support of a large amount of high-quality labeled data, and has high requirements for the quality and quantity of training data. Without sufficient high-quality training data, the accuracy of a data-driven model may be lower than that of a physical model. Data-driven models may perform well under specific operating conditions, but perform poorly when faced with new operating conditions or actual changes, and their generalization ability is limited. Data-driven models are usually trained offline, making it difficult to update and dynamically adjust them online.

[0044] To address the insufficient accuracy of a single model, this application combines a first engine model from the physical model of a hybrid vehicle and a second engine model from the data-driven model to predict engine fuel consumption under the power distribution strategy of the hybrid vehicle, obtaining a target predicted fuel consumption. Similarly, it combines a first motor model from the physical model and a second motor model from the data-driven model to predict motor energy under the power distribution strategy, obtaining a target predicted energy. Compared to the data-driven model, the accuracy of the physical model depends on precise physical modeling and does not require a large amount of high-quality training data. Conversely, the data-driven model relies less on modeling but more on a large amount of high-quality training data. Furthermore, the data-driven model can capture the continuous variation patterns of the vehicle's power components (engine and motor) under actual operating conditions. Therefore, combining the physical model and the data-driven model for vehicle energy consumption prediction allows for the complementary advantages of both models, improving the accuracy of vehicle energy consumption prediction based on the power distribution strategy of hybrid vehicles.

[0045] Hybrid vehicles refer to vehicles whose powertrain (or drive system) includes at least two power components (or drive components) capable of operating simultaneously. In this embodiment, taking a hybrid vehicle's power components including an engine and at least one electric motor as an example, when the power components include multiple electric motors: the physical model of the hybrid vehicle may include motor models corresponding to each of the multiple motors, with identical model formulas and model parameters determined based on the actual fuel consumption of the motors under operating conditions; the model parameters may be the same or different. The data-driven model of the hybrid vehicle may include a shared motor model among the multiple motor models or individual motor models corresponding to each of the multiple motors, with identical model structures and model parameters trained based on historical operating data of the motors; the model parameters may be the same or different.

[0046] Exemplary methods

[0047] Figure 1 This is a flowchart illustrating a simulation testing method based on a vehicle model, provided as an embodiment of this application. Figure 1 As shown, the simulation testing method based on a vehicle model provided in this embodiment includes the following steps S101 to S105:

[0048] S101, obtain the power distribution strategy for hybrid vehicles.

[0049] The power components of a hybrid vehicle include an engine and an electric motor. The power distribution strategy of a hybrid vehicle includes allocating the engine's output power and the electric motor's output power. In the case of a hybrid vehicle with multiple electric motors, the power distribution strategy may include allocating the output power of multiple motors. Through an optimized power distribution strategy, the output power of the engine and the electric motor can be rationally allocated, achieving better fuel efficiency and power performance in the hybrid vehicle. Therefore, it is necessary to simulate the power distribution strategy using a vehicle model to obtain the engine's fuel consumption and the electric motor's energy consumption under the optimized power distribution strategy.

[0050] In this embodiment, the power distribution strategy generated by the strategy model can be obtained to predict the engine fuel consumption and motor power consumption under the power distribution strategy generated by the strategy model through subsequent simulation testing. The strategy model, also known as the controller model, can adapt to the vehicle's power demand to generate the corresponding power distribution strategy. Of course, the power distribution strategy can also be obtained from a database or other devices, or it can be obtained from the power distribution strategy input by the user.

[0051] S102, through the first engine model and the first motor model, the power distribution strategy is simulated and tested to obtain the first predicted fuel consumption and the first predicted motor power. The first engine model and the first motor model are the engine model and the motor model in the physical model of the hybrid vehicle, respectively.

[0052] In the physical model of the hybrid vehicle, the engine model is a mathematical description based on the fuel consumption of the engine under different operating conditions (including different output power), and the motor model is a mathematical description based on the electrical energy of the motor under different operating conditions (including different output power). For simplicity, in this embodiment, the engine model in the physical model of the hybrid vehicle is referred to as the first engine model, and the motor model in the physical model of the hybrid vehicle is referred to as the first motor model.

[0053] The motor energy may include motor charging energy and / or motor discharging energy, and the first predicted motor energy may include first motor charging energy and / or first motor discharging energy.

[0054] In this embodiment, a first engine model and a first motor model can be pre-constructed. The output power allocated to the engine and the output power allocated to the motor are obtained from the power distribution strategy. The fuel consumption under the allocated engine output power can be predicted using the first engine model to obtain a first predicted fuel consumption, thus achieving simulation testing of engine fuel consumption under the power distribution strategy. Similarly, the motor energy under the allocated motor output power can be predicted using the first motor model to obtain a first predicted motor energy, thus achieving simulation testing of motor energy under the power distribution strategy.

[0055] S103, based on the second engine model and the second motor model, simulates and tests the power distribution strategy to obtain the second predicted fuel consumption and the second predicted motor energy. The second engine model and the second motor model are data-driven models trained based on historical data.

[0056] The historical data includes engine operating data and motor operating data from the past. The engine operating data includes the engine's operating conditions and the engine's actual fuel consumption under those operating conditions. The motor operating data includes the motor's operating conditions and the motor's actual electrical energy (motor charging energy and / or motor discharging energy) under those operating conditions.

[0057] Historical data can come from different vehicles. During the training of the initially constructed data-driven model, historical data from different hybrid vehicles can be used to train the model, resulting in a pre-trained data-driven model. To improve the accuracy of the data-driven model, these different hybrid vehicles can be of the same model. After deploying the pre-trained data-driven model on the hybrid vehicle, to further improve its accuracy and provide a personalized data-driven model for the hybrid vehicle, the model can be retrained based on the actual driving data of that hybrid vehicle.

[0058] In the data-driven model of the hybrid vehicle, the engine model is a mapping model between engine operating conditions and engine fuel consumption, obtained by training the model based on past engine operating data of the hybrid vehicle; the motor model is a mapping model between motor operating conditions and motor electrical energy, obtained by training the model based on past motor operating data of the hybrid vehicle. For simplicity, in this embodiment, the engine model in the data-driven model of the hybrid vehicle is referred to as the second engine model, and the motor model in the data-driven model of the hybrid vehicle is referred to as the second motor model.

[0059] In this embodiment, a second engine model and a second motor model can be pre-trained based on historical data. The output power allocated to the engine and the output power allocated to the motor are obtained from the power distribution strategy. The fuel consumption under the allocated engine output power can be predicted using the second engine model to obtain a second predicted fuel consumption, thus achieving simulation testing of engine fuel consumption under the power distribution strategy. Similarly, the motor energy under the allocated motor output power can be predicted using the second motor model to obtain a second predicted motor energy, thus achieving simulation testing of motor energy under the power distribution strategy. The second predicted motor energy may include the charging energy of the second motor and / or the discharging energy of the second motor.

[0060] S104, the first predicted fuel consumption and the second predicted fuel consumption are fused to obtain the target predicted fuel consumption under the power distribution strategy.

[0061] In this embodiment, a first fusion method can be used to fuse the first predicted fuel consumption and the second predicted fuel consumption to obtain the target predicted fuel consumption under the power distribution strategy. For example, the first predicted fuel consumption and the second predicted fuel consumption can be added together and the average value can be calculated to obtain the target predicted fuel consumption; or, the difference between the first predicted fuel consumption and the second predicted fuel consumption can be compared, and the first predicted fuel consumption or the second predicted fuel consumption can be adjusted according to the difference to obtain the target predicted fuel consumption value.

[0062] The first predicted fuel consumption can be the instantaneous fuel consumption predicted by the first engine model, the second predicted fuel consumption can be the instantaneous fuel consumption predicted by the second engine model, and the target predicted fuel consumption can also be the instantaneous fuel consumption. After fusing the first and second predicted fuel consumptions to obtain the target predicted fuel consumption, the target predicted fuel consumption is integrated over time to obtain the predicted cumulative fuel consumption within a unit time period.

[0063] S105, the first predicted motor energy and the second predicted motor energy are fused to obtain the target predicted motor energy under the power distribution strategy.

[0064] In this embodiment, a second fusion method can be used to fuse the first predicted motor energy and the second predicted motor energy to obtain the target predicted motor energy under the power distribution strategy. The second fusion method can be the same as or different from the first fusion method. For example, the target predicted motor energy can be obtained by adding the first predicted motor energy and the second predicted motor energy and then averaging them; or, the target predicted motor energy can be obtained by comparing the difference between the first predicted motor energy and the second predicted motor energy and adjusting either the first predicted motor energy or the second predicted motor energy based on the difference.

[0065] The target predicted motor energy may include target predicted motor charging energy and target predicted motor discharging energy. When the first predicted motor energy includes first predicted motor charging energy and the second predicted motor energy includes second predicted motor charging energy, the first predicted motor charging energy and the second predicted motor charging energy are fused to obtain the target predicted motor charging energy under the power distribution strategy; when the first predicted motor energy includes first predicted motor discharging energy and the second predicted motor energy includes second predicted motor discharging energy, the first predicted motor discharging energy and the second predicted motor discharging energy are fused to obtain the target predicted motor discharging energy under the power distribution strategy.

[0066] The first predicted motor energy can be the instantaneous motor energy predicted by the first motor model, the second predicted motor energy can be the instantaneous motor energy predicted by the second motor model, and the target predicted motor energy can also be the instantaneous motor energy. After fusing the first and second predicted motor energy to obtain the target predicted motor energy, the target predicted motor energy is integrated over time to obtain the cumulative predicted motor energy within a unit time period.

[0067] In this embodiment, by combining the first engine model in the physical model of the hybrid vehicle and the second engine model in the data-driven model of the hybrid vehicle, the engine fuel consumption under the power distribution strategy of the hybrid vehicle is predicted to obtain the target predicted fuel consumption. Similarly, by combining the first motor model in the physical model of the hybrid vehicle and the second motor model in the data-driven model of the hybrid vehicle, the motor energy under the power distribution strategy of the hybrid vehicle is predicted to obtain the target predicted energy. By combining the physical model and the data-driven model for vehicle energy consumption prediction, the advantages of both models are combined, improving the accuracy of vehicle energy consumption prediction based on the power distribution strategy of the hybrid vehicle.

[0068] Since dual-model fusion is one of the key aspects of this application's embodiments, the following provides an embodiment of dual-model fusion (i.e., fusing the first predicted fuel consumption and the second predicted fuel consumption, and fusing the first predicted motor energy and the second predicted motor energy).

[0069] In some embodiments, fusing the first predicted fuel consumption and the second predicted fuel consumption to obtain the target predicted fuel consumption under the power distribution strategy may include: weighting the first predicted fuel consumption and the second predicted fuel consumption according to the weight parameters corresponding to the first engine model and the second engine model to obtain the target predicted fuel consumption. And / or, fusing the first predicted motor energy and the second predicted motor energy to obtain the target predicted motor energy under the power distribution strategy may include: weighting the first predicted motor energy and the second predicted motor energy according to the weight parameters corresponding to the first motor model and the second motor model to obtain the target predicted motor energy.

[0070] In this embodiment, the target predicted fuel consumption can be obtained by adding the product of the first predicted fuel consumption and the weight parameter corresponding to the first engine model, and the product of the second predicted fuel consumption and the weight parameter corresponding to the second engine model; and / or, the target predicted motor energy can be obtained by adding the product of the first predicted motor energy and the weight parameter corresponding to the first motor model, and the product of the second predicted motor energy and the weight parameter corresponding to the second motor model. When the first predicted motor energy includes both the first predicted motor charging energy and the second predicted motor charging energy, the target predicted motor charging energy can be obtained by adding the product of the first predicted motor charging energy and the weight parameter corresponding to the first motor model, and the product of the second predicted motor charging energy and the weight parameter corresponding to the second motor model. When the first predicted motor energy includes both the first predicted motor discharging energy and the second predicted motor energy includes both the second predicted motor discharging energy, the target predicted motor discharging energy can be obtained by adding the product of the first predicted motor charging / discharging energy and the weight parameter corresponding to the first motor model, and the product of the second predicted motor discharging energy and the weight parameter corresponding to the second motor model.

[0071] Therefore, by using a weighted approach, the fusion of the first and second predicted fuel consumption, as well as the fusion of the first and second predicted motor energy, is achieved. The accuracy of the target predicted fuel consumption obtained by fusing the first and second predicted fuel consumption can be improved by setting appropriate weight parameters for the first and second engine models respectively. For example, when the training data for the second engine model is insufficient, a larger weight parameter is set for the first engine model compared to the second engine model; conversely, when the training data for the second engine model is sufficient, a larger weight parameter is set for the second engine model compared to the first engine model. Similarly, the accuracy of the target predicted motor energy obtained by fusing the first and second predicted motor energy can be improved by setting appropriate weight parameters for the first and second motor models respectively. For example, when the training data for the second motor model is insufficient, a larger weight parameter is set for the first motor model compared to the second motor model; conversely, when the training data for the second motor model is sufficient, a larger weight parameter is set for the second motor model compared to the first motor model.

[0072] In some embodiments, the weight parameters corresponding to the first engine model and the second engine model support dynamic updates. And / or, the weight parameters corresponding to the first motor model and the second motor model support dynamic updates. The foregoing description of related technologies mentions that vehicle models have limited generalization ability and cannot be updated online or dynamically adjusted. To address this issue, this embodiment proposes a scheme that supports dynamic updates of the weighted parameters used in fusing the vehicle's physical model and data-driven model, providing a high-precision, dynamically updatable dual-model fusion scheme. This scheme improves the accuracy of vehicle energy consumption prediction.

[0073] Specifically, the weight parameters corresponding to the first engine model and the second engine model can be dynamically updated based on the fuel consumption prediction error of the models (first engine model and second engine model) to improve the accuracy of the dynamic update of the weight parameters corresponding to the first engine model and the second engine model. And / or, the weight parameters corresponding to the first motor model and the second motor model can be dynamically updated based on the energy prediction error of the models (first motor model and second motor model) to improve the accuracy of the dynamic update of the weight parameters corresponding to the first motor model and the second motor model.

[0074] Figure 2This is a flowchart illustrating the dynamic update process of the weight parameters corresponding to the first engine model and the weight parameters corresponding to the second engine model in a vehicle model-based simulation testing method provided in this application embodiment. Figure 2 As shown, the dynamic update process of the weight parameters corresponding to the first engine model and the weight parameters corresponding to the second engine model includes the following steps S201 to S203:

[0075] S201, acquire the first actual driving data of the hybrid vehicle.

[0076] In this embodiment, with the authorization or permission of the hybrid vehicle owner, the first actual driving data of the hybrid vehicle can be stored on a cloud server. Therefore, the first actual driving data can be obtained from the cloud server. The first actual driving data includes the engine operating data of the hybrid vehicle during actual driving, and the engine operating data includes the engine's operating conditions and the engine's actual fuel consumption under those operating conditions.

[0077] S202, based on the first actual driving data, the first engine model, and the second engine model, determine the fuel consumption prediction error.

[0078] In this embodiment, based on the engine's operating conditions in the first actual driving data, the engine fuel consumption under the operating conditions can be predicted using a first engine model and a second engine model, respectively, to obtain the predicted fuel consumption output by the first engine model and the predicted fuel consumption output by the second engine. Based on the predicted fuel consumption output by the first engine, the predicted fuel consumption output by the second engine, and the actual fuel consumption of the engine under the operating conditions in the first actual driving data, the fuel consumption prediction error is determined.

[0079] S203, based on the fuel consumption prediction error, update the weight parameters corresponding to the first engine model and the weight parameters corresponding to the second engine model.

[0080] In this embodiment, the fuel consumption prediction error reflects the prediction accuracy of the first engine model and the second engine model. To reduce the fuel consumption prediction error, the weight parameters corresponding to the first engine model and the second engine model can be updated to improve the rationality of the weight parameters corresponding to the first engine model and the second engine model respectively, thereby improving the accuracy of energy consumption prediction by combining the first engine model and the second engine model.

[0081] Optionally, the fuel consumption prediction error includes the instantaneous fuel consumption prediction error of the first engine model and the instantaneous fuel consumption prediction error of the second engine model. S202 may include: predicting the instantaneous fuel consumption of the engine under the operating conditions in the first actual driving data using the first engine model to obtain a first predicted instantaneous fuel consumption; predicting the instantaneous fuel consumption of the engine under the operating conditions in the first actual driving data using the second engine model to obtain a second predicted instantaneous fuel consumption; comparing the first predicted instantaneous fuel consumption with the actual instantaneous fuel consumption of the engine in the first actual driving data to obtain the instantaneous fuel consumption prediction error of the first engine model; comparing the second predicted instantaneous fuel consumption with the actual instantaneous fuel consumption of the engine in the first actual driving data to obtain the instantaneous fuel consumption prediction error of the second engine model. S203 may include: determining the updated weight parameters corresponding to the first engine model and the second engine model based on the instantaneous fuel consumption prediction error of the first engine model and the instantaneous fuel consumption prediction error of the second engine model; determining the updated weight parameters corresponding to the first engine model as the weight parameters corresponding to the first engine model, and determining the updated weight parameters corresponding to the second engine model as the weight parameters corresponding to the second engine model.

[0082] In this optional approach, during the process of determining the updated weight parameters corresponding to the first engine model and the second engine model based on the instantaneous fuel consumption prediction errors of the first and second engine models, a larger updated weight parameter can be determined for the second engine model if the instantaneous fuel consumption prediction error of the first engine model is large, and vice versa. Thus, by referencing the instantaneous fuel consumption prediction errors of the first and second engine models, the rationality of the updated weight parameters corresponding to the first and second engine models is improved.

[0083] Furthermore, based on the instantaneous fuel consumption prediction errors of the first engine model and the second engine model, determining the updated weight parameters corresponding to the first and second engine models can include: adding the instantaneous fuel consumption prediction errors of the first and second engine models to obtain a sum of instantaneous fuel consumption prediction errors; determining the ratio of the instantaneous fuel consumption prediction error of the first engine model to the sum of the instantaneous fuel consumption prediction errors as the updated weight parameter corresponding to the second engine model; and determining the ratio of the instantaneous fuel consumption prediction error of the second engine model to the sum of the instantaneous fuel consumption prediction errors as the updated weight parameter corresponding to the first engine model. It can be seen that through this process of calculating the updated weight parameters, the following conditions are met: when the instantaneous fuel consumption prediction error of the first engine model is greater than that of the second engine model, the updated weight parameter corresponding to the first engine model is less than that of the second engine model; when the instantaneous fuel consumption prediction error of the first engine model is less than that of the second engine model, the updated weight parameter corresponding to the first engine model is greater than that of the second engine model.

[0084] Furthermore, based on the instantaneous fuel consumption prediction errors of the first engine model, the second engine model, and the actual fuel consumption during the reference period in the first actual driving data, the cumulative fuel consumption error obtained by combining the first and second engine models for fuel consumption prediction can be determined. If the cumulative fuel consumption error is less than the fuel consumption error threshold, the updated weight parameters corresponding to the first engine model are determined as the weight parameters corresponding to the first engine model, and the updated weight parameters corresponding to the second engine model are determined as the weight parameters corresponding to the second engine model. Thus, using the cumulative fuel consumption error as an indicator of whether weight parameters need to be updated improves the rationality of updating the weight parameters corresponding to the first and second engine models, avoids unnecessary weight updates, and reduces the consumption of computational resources.

[0085] In this further approach, the reference time period can be a unit time period or other time periods. The instantaneous fuel consumption prediction error of the first engine model and the instantaneous fuel consumption prediction error of the second engine model can be added together and then integrated in the time dimension to obtain the cumulative fuel consumption error.

[0086] Alternatively, the updated weight parameters corresponding to the first engine model and the updated weight parameters corresponding to the second engine model can be used to weight the first predicted instantaneous fuel consumption and the second predicted instantaneous fuel consumption to obtain the predicted instantaneous fuel consumption obtained by combining the first engine model and the second engine model; the predicted instantaneous fuel consumption is integrated over time based on the reference period to obtain the predicted fuel consumption for the reference period; the predicted fuel consumption for the reference period is compared with the actual fuel consumption for the reference period to obtain the cumulative fuel consumption error.

[0087] Furthermore, the difference between the predicted fuel consumption and the actual fuel consumption during the reference period can be calculated, and this difference can be divided by the actual fuel consumption during the reference period to obtain the cumulative fuel consumption error.

[0088] Optionally, if the cumulative fuel consumption error is greater than or equal to the fuel consumption error threshold, the updated weight parameters corresponding to the first engine model and the second engine model are redefined, and / or the second engine model is retrained. Thus, by redefining the weight parameters and retraining the engine models, the accuracy of predicting fuel consumption under the power distribution strategy by combining the first and second engine models is effectively improved.

[0089] In this optional approach, if the cumulative fuel consumption error is greater than or equal to the fuel consumption error threshold, it indicates that the updated weight parameters corresponding to the first engine model and the second engine model are inappropriate, or that the accuracy of the second engine model is low. The updated weight parameters corresponding to the first engine model and the second engine model can be re-determined according to the aforementioned process, and / or the second engine model can be retrained based on the actual driving data of the hybrid vehicle (such as the first actual driving data).

[0090] Furthermore, if the cumulative fuel consumption error is greater than or equal to the fuel consumption error threshold, the updated weight parameters for the first engine model and the second engine model are redefined. Based on these redefined parameters, the cumulative fuel consumption error is recalculated. If this recalculated error is greater than or equal to the threshold, the second engine model is retrained. Thus, by first redefined the updated weight parameters for both models, and then retraining the second engine model if the redefined parameters still fail to resolve the issue of the cumulative fuel consumption error exceeding the threshold, the prediction accuracy of the second engine model is improved.

[0091] In some embodiments, the dynamic update process of the weight parameters corresponding to the first motor model and the second motor model includes: acquiring second actual driving data of the hybrid vehicle; determining the energy prediction error based on the second actual driving data, the first motor model, and the second motor model; and updating the weight parameters corresponding to the first motor model and the second motor model based on the energy prediction error. The second actual driving data includes motor operating data of the hybrid vehicle during actual driving, and the motor operating data includes the motor's operating conditions and the actual energy of the motor under those operating conditions.

[0092] Optionally, the energy prediction error includes the instantaneous energy prediction error of the first motor model and the instantaneous energy prediction error of the second motor model. The first motor model can be used to predict the instantaneous energy of the motor under the operating conditions of the second actual driving data, resulting in a first predicted instantaneous energy. The second motor model can be used to predict the instantaneous energy of the motor under the operating conditions of the second actual driving data, resulting in a second predicted instantaneous energy. The instantaneous energy prediction error of the first motor model is obtained by comparing the first predicted instantaneous energy with the actual instantaneous energy of the engine in the second actual driving data. The instantaneous energy prediction error of the second motor model is obtained by comparing the second predicted instantaneous energy with the actual instantaneous energy of the motor in the second actual driving data. Based on the instantaneous energy prediction errors of the first and second motor models, updated weight parameters corresponding to the first and second motor models can be determined. The updated weight parameters corresponding to the first motor model are then defined as the weight parameters corresponding to the first motor model, and the updated weight parameters corresponding to the second motor model are defined as the weight parameters corresponding to the second motor model.

[0093] Furthermore, the instantaneous energy prediction error of the first motor model can be added to the instantaneous energy prediction error of the second motor model to obtain the sum of instantaneous energy prediction errors; the ratio of the instantaneous energy prediction error of the first motor model to the sum of instantaneous energy prediction errors is determined as the update weight parameter corresponding to the second motor model; the ratio of the instantaneous energy prediction error of the second motor model to the sum of instantaneous energy prediction errors is determined as the update weight parameter corresponding to the first motor model.

[0094] Furthermore, based on the instantaneous energy prediction error of the first motor model, the instantaneous energy prediction error of the second motor model, and the actual fuel consumption during the reference period in the second actual driving data, the cumulative energy error obtained by combining the first motor model and the second motor model for energy prediction can be determined. If the cumulative energy error is less than the energy error threshold, the updated weight parameter corresponding to the first motor model is determined as the weight parameter corresponding to the first motor model, and the updated weight parameter corresponding to the second motor model is determined as the weight parameter corresponding to the second motor model.

[0095] Optionally, if the cumulative power error is greater than or equal to the power error threshold, the updated weight parameters corresponding to the first motor model and the updated weight parameters corresponding to the second motor model are re-determined, and / or the second motor model is retrained.

[0096] Furthermore, if the cumulative energy error is greater than or equal to the energy error threshold, the updated weight parameters corresponding to the first motor model and the second motor model are redefined. Based on the redefined updated weight parameters corresponding to the first motor model and the second motor model, the cumulative energy error is recalculated. If the recalculated cumulative energy error is greater than or equal to the energy error threshold, the second motor model is retrained.

[0097] The dynamic update process of the weight parameters corresponding to the first motor model and the weight parameters corresponding to the second motor model can be referred to the dynamic update process of the weight parameters corresponding to the first engine model and the weight parameters corresponding to the second engine model, and will not be repeated here.

[0098] As an example, Figure 3 This is a flowchart illustrating the dynamic update process of the weight parameters corresponding to the physical model and the weight parameters corresponding to the data-driven model.

[0099] exist Figure 3 In this model, the physical model is the first engine model and the data-driven model is the second engine model, or the physical model is the first motor model and the data-driven model is the second motor model.

[0100] like Figure 3As shown, the dynamic update process of the weight parameters may include: First, predicting vehicle energy consumption based on actual driving data and a physical model to obtain the predicted instantaneous energy consumption (e.g., predicted instantaneous fuel consumption or predicted instantaneous electrical energy) y_pre_A of the physical model; and predicting vehicle energy consumption based on actual driving data and a data-driven model to obtain the predicted instantaneous energy consumption y_pre_B of the data-driven model. Second, comparing y_pre_A with the actual instantaneous energy consumption in the actual driving data to obtain the instantaneous energy consumption prediction error (e.g., instantaneous fuel consumption prediction error or instantaneous electrical energy prediction error) mse_A of the physical model; and comparing y_pre_B with the actual instantaneous energy consumption to obtain the instantaneous energy consumption prediction error mse_B of the data-driven model. Then, summing mse_A and mse_B to obtain mse_A + mse_B, and determining the updated weight parameter weight_A of the physical model as mse_B / (ms The update weight parameter weight_B of the data-driven model is determined as mse_A / (mse_A+mse_B). Based on weight_A and weight_B, y_pre_A and y_pre_B are weighted to obtain the predicted instantaneous energy consumption y_pre_C obtained by combining the physical model and the data-driven model for energy consumption prediction. y_pre_C is then integrated over time based on a reference period to obtain the predicted energy consumption E_pre_C for the reference period. E_pre_C is subtracted from the actual energy consumption E_test in the actual driving data for the reference period and then divided by E_test to obtain the cumulative energy consumption error mse_C. If the cumulative energy consumption error is less than the energy consumption error threshold of 5%, the weight parameter of the physical model is determined as weight_A, and the weight parameter of the data-driven model is determined as weight_B; otherwise, the weight parameters are recalculated. If mse_C is still greater than or equal to the energy consumption error threshold after reweighting, the data-driven model is trained.

[0101] The above describes the implementation examples of the dual-model fusion process and the parameter update process of dual-model fusion. The following provides corresponding implementation examples for the simulation test process based on dual models.

[0102] In some embodiments, the power distribution strategy includes allocated engine speed and allocated engine output torque (i.e., the allocated engine output power in the aforementioned embodiments includes allocated engine speed and allocated engine output torque). The power distribution strategy is simulated and tested using a first engine model to obtain a first predicted fuel consumption. This includes: determining the predicted engine specific fuel consumption based on the allocated engine speed and allocated engine output torque; the predicted engine specific fuel consumption is the engine specific fuel consumption when the hybrid vehicle's engine operates at the allocated engine speed and allocated engine output torque; and inputting the allocated engine speed, allocated engine output torque, and predicted engine specific fuel consumption into the first engine model to obtain the first predicted fuel consumption. This process improves the accuracy of predicting generator fuel consumption using the first generator model.

[0103] Among them, based on the working characteristics of the engine in the hybrid vehicle, the engine specific fuel consumption when the engine works according to the allocated engine speed and allocated engine output torque can be predicted to obtain the predicted engine specific fuel consumption.

[0104] The first engine model can be a mapping model of engine speed, engine output torque, engine specific fuel consumption, and engine fuel consumption. Therefore, using this mapping model as the first engine model improves the accuracy of predicting engine fuel consumption using this model.

[0105] Optionally, the first engine model is represented as:

[0106]

[0107] In the above formula for the first engine model, N eng T represents the allocated engine speed. q This indicates the allocated engine output torque, M represents the unit conversion factor (e.g., M = 9550), bsfc represents the predicted engine specific fuel consumption, and Mass... fuel This indicates the first predicted fuel consumption.

[0108] Optionally, the predicted engine specific fuel consumption is determined based on the allocated engine speed and allocated engine output torque. This includes: interpolating the engine specific fuel consumption from engine test data based on the allocated engine speed and allocated engine output torque; or inputting the allocated engine speed and allocated engine output torque into a specific fuel consumption mapping model to obtain the predicted engine specific fuel consumption. The engine test data includes multidimensional data consisting of engine speed, engine output torque, and engine specific fuel consumption (e.g., tabular data containing a multidimensional sequence of engine speed, engine output torque, and engine specific fuel consumption). The specific fuel consumption mapping model is a model of the mapping relationship between engine speed, engine output torque, and engine specific fuel consumption, obtained by fitting the engine test data. The engine test data can be obtained through engine bench testing, and interpolation methods include linear interpolation, spline interpolation, etc. Therefore, by using interpolation or fitting based on engine test data, the accuracy of predicting engine specific fuel consumption from discrete test data is improved.

[0109] In some embodiments, the power distribution strategy includes allocated motor speed and allocated motor output torque (i.e., the allocated motor output power in the aforementioned embodiments includes allocated motor speed and allocated motor output torque). The power distribution strategy is simulated and tested using a first motor model to obtain a first predicted motor energy. This includes: determining the predicted motor charging efficiency and predicted motor discharging efficiency based on the allocated motor speed and allocated motor output torque. The predicted motor charging efficiency and predicted motor discharging efficiency are the charging efficiency and discharging efficiency of the hybrid vehicle's motor when operating at the allocated motor speed and allocated motor output torque; inputting the allocated motor speed, allocated motor output torque, and predicted motor charging efficiency into the motor charging model of the first motor model to obtain the predicted charging energy in the first predicted motor energy; and inputting the allocated motor speed, allocated motor output torque, and predicted motor discharging efficiency into the motor discharging model of the first motor model to obtain the predicted discharging energy in the first predicted motor energy. This process improves the accuracy of predicting motor energy using the first motor model.

[0110] Among them, based on the working characteristics of the motor in the hybrid vehicle, the motor charging efficiency and motor discharging efficiency when the motor works according to the allocated motor speed and allocated motor output torque can be predicted to obtain the predicted motor charging efficiency and predicted motor discharging efficiency.

[0111] The motor charging model in the first motor model can be a mapping model of motor speed, motor output torque, motor charging efficiency, and motor charging energy. Similarly, the motor discharging model in the first motor model can be a mapping model of motor speed, motor output torque, motor discharging efficiency, and motor discharging energy. Therefore, by using the corresponding mapping model as the first motor model, the accuracy of predicting engine fuel consumption using the first motor model is improved.

[0112] Optionally, the motor discharge model can be represented as:

[0113] E_generator = n motor ×Tq motor ÷K×Efficency1

[0114] The motor charging model is represented as:

[0115] E_driven = n motor ×Tq motor ÷K÷Efficency2

[0116] In the above formulas for the motor discharge model and the motor charging model, n motor Tq represents the allocated motor speed. motor This indicates the allocated motor output torque, K represents the unit conversion factor (e.g., K = 9550), Efficency1 represents the predicted motor discharge efficiency, Efficency2 represents the predicted motor charging efficiency, E_generator represents the predicted discharge energy, and E_driven represents the predicted charging energy.

[0117] Optionally, the predicted motor charging efficiency and predicted motor discharging efficiency are determined based on the assigned motor speed and assigned motor output torque. This includes: interpolating the assigned motor speed and assigned motor output torque into the motor test data to obtain the predicted motor charging efficiency and predicted motor discharging efficiency; or, inputting the assigned motor speed and assigned motor output torque into an efficiency mapping model to obtain the predicted motor charging efficiency and predicted motor discharging efficiency. The motor test data includes multidimensional data consisting of motor speed, motor output torque, motor charging efficiency, and motor discharging efficiency (e.g., tabular data containing a multidimensional sequence of motor speed, motor output torque, motor charging efficiency, and motor discharging efficiency). The efficiency mapping model is a model of the mapping relationship between motor speed, motor output torque, motor charging efficiency, and motor discharging efficiency, and is obtained by fitting the motor test data. The motor test data can be obtained through motor bench tests, and interpolation methods include linear interpolation, spline interpolation, etc. Therefore, by using interpolation or fitting based on motor test data, the accuracy of obtaining the predicted motor charging efficiency and predicted motor discharging efficiency from discrete test data is improved.

[0118] In some embodiments, in the data-driven model, the second engine model is a mapping model between engine operating conditions and engine fuel consumption trained based on historical data. The second engine model can be a neural network or other regression prediction model (such as a linear regression model or a support vector machine regression model) to improve the fuel consumption prediction accuracy of the second engine model. Among them, the neural network is, for example, a long short-term memory network (LSTM).

[0119] Optionally, the input data for the second engine model includes engine speed, engine output torque, engine intake air temperature, and engine coolant temperature. Based on the second engine model, simulation testing of the power distribution strategy is performed to obtain a second predicted fuel consumption. This may include: acquiring vehicle sensor data, including engine intake air temperature and engine coolant temperature; inputting the engine speed, engine output torque, engine intake air temperature, and engine coolant temperature allocated in the power distribution strategy into the second engine model; and performing fuel consumption prediction in the second engine model to obtain the second predicted fuel consumption. This provides the second engine model with engine characteristics related to engine fuel consumption, improving the accuracy of the second engine model's fuel consumption prediction.

[0120] Figure 4 This is an example diagram showing the model training for the second engine model. Figure 4As shown, the input data for the second engine model includes engine speed, engine output torque, engine intake air temperature, and engine coolant temperature. The output data includes engine fuel consumption, such as the amount of fuel injected. During training, the neural network can be trained based on training data containing the above input data and their corresponding labels. After multiple training iterations, the second engine model is obtained.

[0121] In some embodiments, in the data-driven model, the second motor model is a mapping model between motor operating conditions and motor energy obtained by training based on historical data. The second motor model can be a neural network (e.g., LSTM) or other regression prediction models (e.g., linear regression model, support vector machine regression model) to improve the accuracy of the second motor model in predicting electrical energy.

[0122] Optionally, the input data for the second motor model includes motor speed, motor output torque, and motor body temperature. Based on the second motor model, simulation testing of the power distribution strategy is performed to obtain the second predicted motor energy. This may include: acquiring vehicle sensor data, which includes the motor body temperature; inputting the motor speed, motor output torque, and motor body temperature allocated in the power distribution strategy into the second motor model; and performing energy prediction within the second motor model to obtain the second predicted energy. This provides the second motor model with motor characteristics related to motor energy, improving the accuracy of the second motor model's energy prediction.

[0123] Figure 5 This is an example diagram showing the model training for the second motor model. (Example:) Figure 5 As shown, the input data for the second motor model includes motor speed, motor output torque, and motor body temperature, while the output data is motor electrical energy. During training, the neural network can be trained based on training data containing the above input data and corresponding labels. After multiple training iterations, the second motor model is obtained.

[0124] In some embodiments, the physical model further includes a battery model. After obtaining the power distribution strategy of the hybrid vehicle, the model further includes: performing simulation tests on the power distribution strategy using the battery model to obtain a predicted battery state of charge (SOC). Therefore, it can provide not only generator fuel consumption and motor energy under the power distribution strategy, but also predicted battery SOC, facilitating a multi-faceted evaluation of the energy consumption of the power distribution strategy.

[0125] In this embodiment, the loads on the hybrid vehicle (such as the vehicle air conditioner and vehicle audio system) consume battery power. The load energy consumption and the current battery SOC can be obtained. The load energy consumption is input into the battery model, and the battery SOC is predicted by the battery model to obtain the predicted battery SOC.

[0126] Optionally, considering that the load on a hybrid vehicle is affected by actual usage scenarios and is not easily accurately simulated through models, obtaining load energy consumption includes: obtaining the input load energy consumption. For example, an application that calculates load energy consumption can be deployed on the vehicle's infotainment system, and the load energy consumption input to this application can be obtained.

[0127] Optionally, the battery model can be represented as:

[0128]

[0129] Where E represents load energy consumption, c BATTRY Indicates the rated capacitance, SOC start1 This indicates the battery's current state of charge (SOC). end1 This represents the predicted battery SOC, where Start1 represents the current time and end1 represents a future time. Therefore, the battery SOC at future times can be predicted using this battery model.

[0130] Optionally, the battery model's parameters include the rated capacitance, which supports dynamic updates. The updating process for the rated capacitance includes: acquiring third-party real-world driving data of the hybrid vehicle; determining the actual capacitance based on the third-party real-world driving data; comparing the actual capacitance with the rated capacitance to obtain a comparison result; and updating the rated capacitance based on the comparison result. Therefore, considering that the capacitance of the battery in a hybrid vehicle will degrade to a certain extent with increasing usage time, the dynamic updating of the battery model's rated capacitance can simulate capacitance degradation and improve the accuracy of the battery model.

[0131] The third type of actual driving data includes parameters related to the hybrid vehicle's battery. These parameters can be input into the battery capacitance calculation formula to obtain the actual capacitance. If the battery's rated capacitance has degraded based on a comparison between the actual capacitance and the rated capacitance, the rated capacitance in the battery model is updated to the actual capacitance; otherwise, the rated capacitance remains unchanged.

[0132] Optionally, the third actual driving data includes a battery data set, which includes battery current, battery voltage, battery SOC corresponding to the start time of the actual driving period, and battery SOC corresponding to the end time of the actual driving period. The battery current, battery voltage, battery SOC corresponding to the start time of the actual driving period, and battery SOC corresponding to the end time of the actual driving period can be input into the battery capacitance calculation formula to obtain the actual capacitance.

[0133] Alternatively, the formula for calculating the actual capacitance can be expressed as:

[0134]

[0135] Where U represents battery voltage, I represents battery current, Start2 represents the start time of the actual driving period, end2 represents the end time of the actual driving period, and SOC... start2 This indicates the battery SOC corresponding to the start time. end2 C represents the battery SOC corresponding to the end time. 实际 This represents the actual capacitance.

[0136] Optionally, the third set of actual driving data includes multiple sets of battery data. Multiple actual capacitances can be calculated using the above process. A target capacitance (e.g., the median) can be selected from these actual capacitances, and compared with the rated capacitance to obtain a comparison result. If the comparison result is within a set range, the rated capacitance is considered to have not degraded, and the rated capacitance remains unchanged. Otherwise, the rated capacitance in the battery model can be updated to the target capacitance to further improve the accuracy of the battery model. For example, if the ratio of the median to the rated capacitance is within a set range, the rated capacitance remains unchanged; otherwise, the rated capacitance can be updated to the median.

[0137] As an example, Figure 6 This is a flowchart illustrating the dynamic update process of the rated capacitance in a battery model. Figure 6 As shown, firstly, actual driving data is obtained from the cloud server. This data includes battery current, battery voltage, and SOC. Based on this data and the calculation formula for actual capacitance (see the above embodiment), multiple sets of actual capacitances are obtained. Secondly, the median of these multiple sets of actual capacitances is obtained, and the ratio of the median to the rated capacitance in the battery model is calculated. Then, if this ratio is less than a first threshold and greater than a second threshold, the rated capacitance of the battery is considered to have not degraded, and the rated capacitance in the battery model remains unchanged. Otherwise, the rated capacitance in the battery model is updated to the median. The first threshold is greater than the second threshold; for example, the first threshold is 1.05, and the second threshold is 0.95.

[0138] As an example, Figure 7 This is an example diagram illustrating energy flow in a hybrid vehicle. Figure 7As shown, in a hybrid vehicle, components related to vehicle energy consumption include an engine, motors, a battery, and a load. There can be multiple motors, such as motor P1, motor P2, and motor P3. The engine and motor P1 are directly connected (e.g., a direct rigid connection), and the engine can charge motor P1 when it is operating. There is no direct connection between the engine and motors P2 and P3, so the engine cannot charge motors P2 and P3. Energy can flow bidirectionally between motors and between the battery and motors. Because the flow of energy within a vehicle is complex in real-world scenarios, a center point is used to represent the energy flow path, without showing the specific energy flow direction.

[0139] In some embodiments, the physical model further includes a vehicle driving model to obtain a power distribution strategy for the hybrid vehicle, including: obtaining vehicle information of the hybrid vehicle and the driver's required torque of the hybrid vehicle, wherein the vehicle information includes first input data required by the vehicle driving model and second input data required by the strategy model; obtaining the predicted driving speed of the hybrid vehicle through the vehicle driving model based on the first input data; and generating a power distribution strategy through the strategy model based on the predicted driving speed, the driver's required torque, and the second input data.

[0140] In this embodiment, the first input data is input into the vehicle driving model, and the predicted driving input of the hybrid vehicle under the driver's required torque is calculated through the vehicle driving model. The predicted driving speed, the driver's required torque, and the second input data are input into the strategy model. In the strategy model, a power distribution strategy is generated based on these input data. Thus, the vehicle driving model provides an accurate predicted driving speed for the strategy model, improving the accuracy of the power distribution strategy generated by the strategy model.

[0141] The vehicle information for hybrid vehicles may include, at the current time, the accelerator pedal travel, brake pedal travel, battery SOC, battery charging and discharging power, engine speed, engine output torque, engine status (such as engine intake air temperature and engine coolant temperature), motor speed, motor output torque, vehicle gear, vehicle slope angle (an indicator reflecting the vehicle's climbing energy), vehicle weight, vehicle frontal area, wheel radius, final drive ratio, transmission ratio, and transmission system mechanical efficiency.

[0142] The first input data may include engine speed, engine output torque, motor speed, motor output torque, vehicle power mode, vehicle gear, vehicle slope angle, vehicle weight, vehicle frontal area, wheel radius, final drive ratio, and transmission ratio; the second input data may include accelerator pedal travel, brake pedal travel, battery SOC, battery charging and discharging power, engine speed, engine output torque, engine status, motor speed, and motor output torque.

[0143] Optionally, the vehicle information includes an engine start-stop signal. Before conducting simulation tests on the power distribution strategy using the first engine model and the second engine model respectively, the engine start-stop signal is determined to indicate engine start, thereby improving the accuracy of predicting engine fuel consumption.

[0144] In this optional approach, the engine speed changes linearly during actual driving. When the engine start-stop signal indicates that the engine is off, the actual engine fuel consumption is zero, but the engine speed may not be zero. If the engine fuel consumption is predicted based on the engine speed and engine output torque, it may be predicted that the engine fuel consumption is not zero. Therefore, by first determining that the engine start-stop signal indicates that the engine is on, and then simulating and testing the power distribution strategy through the first engine model and the second engine model respectively, the accuracy of predicting engine fuel consumption can be improved.

[0145] Optionally, the vehicle driving model is represented as follows:

[0146] F t =F f +F w +F i +F j

[0147] F t =(T tq *i0*i k *n t ) / r D

[0148] F f =m*g*f*cosα

[0149] F w =(C d *A*u a ) / ρ

[0150] F i =m*g*sinα

[0151] F j =m*δ*(d u / d t )

[0152] Among them, F t F represents the driving force. f F represents rolling resistance. w F represents air resistance. i F represents the slope resistance. j Indicates acceleration resistance; T tq i represents the engine output torque; i0 represents the final drive ratio; i kIndicates the gear ratio of the transmission in K gear; n t The mechanical efficiency of the transmission system is represented by m; the total mass of the vehicle is represented by g; the acceleration due to gravity is represented by f; the rolling resistance coefficient is represented by α; and the slope angle is represented by C. d Indicates the air drag coefficient; A represents the vehicle's frontal area; u a The vehicle speed is represented by δ; the rotational mass conversion factor is represented by d. u / d t Indicates acceleration; r D This represents the wheel radius, and ρ represents a constant, for example, ρ = 21.15.

[0153] Optionally, in addition to the predicted driving speed of the hybrid vehicle, the output data of the vehicle driving model may also include the predicted driving acceleration and / or predicted driving range of the hybrid vehicle.

[0154] As an example, Figure 8 Example diagrams showing the physical model, its input data, and its output data. (Example...) Figure 8 As shown, the physical model includes a vehicle driving model, an engine model (i.e., the first engine model), a motor model (i.e., the first motor model), and a battery model. The input data for the vehicle driving model includes the engine output torque, engine start-stop signal, vehicle gear, and vehicle slope angle at the current time. The output data for the vehicle driving model includes predicted driving speed, predicted driving acceleration, and predicted driving range. The input data for the engine model includes the engine speed allocated in the power distribution strategy, the engine output torque allocated in the power distribution strategy, and the engine start-stop signal (which is not directly input to the engine model but serves as a judgment signal for whether to predict fuel consumption through the engine). The output data for the vehicle driving model includes predicted fuel consumption (i.e., the first predicted fuel consumption). The input data for the motor model includes the motor speed and motor output torque allocated in the power distribution strategy. The output data for the motor model includes predicted motor energy (i.e., the first predicted motor energy). The input data for the battery model includes load energy consumption, and the output data for the motor model includes predicted battery SOC.

[0155] Optionally, the drag coefficient in the vehicle driving model can be dynamically updated. While the drag coefficient in the vehicle driving model is a fixed value in simulation tests, in real-world scenarios, the drag coefficient changes with road conditions and is not a fixed value. Furthermore, the drag coefficient is difficult to obtain experimentally. To improve the accuracy of the vehicle driving model, the drag system can be dynamically updated.

[0156] Furthermore, the drag coefficients that support dynamic updates may include one or more of the following: rolling drag coefficient, air drag coefficient, and acceleration drag coefficient (i.e., rotational mass conversion coefficient), in order to improve the accuracy of the rolling drag, air drag, and / or acceleration drag calculated in the vehicle driving model by dynamically updating the rolling drag coefficient, air drag coefficient, and / or acceleration drag coefficient.

[0157] Optionally, the dynamic update process of the drag coefficient in the vehicle driving model includes: acquiring the fourth actual driving data of the hybrid vehicle; determining the vehicle driving prediction error of the vehicle driving model for the hybrid vehicle based on the fourth actual driving data and the vehicle driving model; determining the updated value of the drag coefficient by using parameter identification when the drag coefficient needs to be updated based on the vehicle driving prediction error; and updating the drag coefficient based on the updated value.

[0158] In this optional approach, the fourth actual driving data includes the input data required by the vehicle driving model and the actual vehicle speed. The input data from the fourth actual driving data can be input into the vehicle driving model, and the model can predict vehicle movement to obtain the predicted speed corresponding to the fourth actual driving data. Based on the predicted speed and the actual speed in the fourth actual driving data, it can be determined whether the drag coefficient needs to be updated. If it is determined that the drag coefficient needs to be updated, a parameter identification method is used to determine the updated value of the drag coefficient, and the drag coefficient is updated to this value. Therefore, by using actual driving data, the prediction accuracy of the vehicle driving model can be measured, accurately determining whether the drag coefficient needs to be updated, and the parameter identification method can improve the accuracy of the updated drag coefficient value.

[0159] Furthermore, the correlation coefficient between the predicted driving speed corresponding to the fourth actual driving data and the actual driving speed in the fourth actual driving data can be calculated. If the correlation coefficient is greater than the correlation threshold, the drag coefficient is kept unchanged; otherwise, the drag coefficient is updated. Thus, by analyzing the correlation between the predicted and actual driving speeds, the accuracy of determining whether to update the drag coefficient is improved, avoiding unnecessary update operations.

[0160] Furthermore, the correlation coefficient between predicted driving speed and actual driving speed can be the Pearson correlation coefficient between the predicted driving speed and actual driving speed.

[0161] As an example, Figure 9 This is a flowchart illustrating the dynamic updating of model parameters in a vehicle driving model. Figure 9As shown, firstly, actual driving data (i.e., the fourth type of actual driving data) is obtained from the cloud server. Input data required for the vehicle driving model, such as actual vehicle control data, is then obtained from the actual driving data. The actual vehicle control data can be input into the vehicle driving model to calculate the predicted driving speed. The actual driving speed is then obtained from the actual driving data. The Pearson correlation coefficient between the actual driving speed and the predicted driving speed is calculated. If the Pearson correlation coefficient is greater than the correlation threshold, the drag coefficient is not updated; otherwise, the drag coefficient is updated using a parameter identification method.

[0162] Optionally, a parameter identification method is used to determine the updated value of the drag coefficient, including: determining the actual values ​​of the input variables and output variables in the vehicle driving model based on the fourth set of actual driving data; obtaining the output data of the regression model by performing multiple regression on the corresponding regression model based on the actual values ​​of the input and output variables in the vehicle driving model; and obtaining the updated value of the drag coefficient in the vehicle driving model by performing inverse parameter decomposition based on the output data of the regression model. Thus, by utilizing actual driving data and multiple regression, the accuracy of the drag coefficient update is improved.

[0163] Furthermore, the least squares method can be used to perform multiple regression on the regression model corresponding to the vehicle driving model, reducing the computational load of the regression process and improving the efficiency of updating the drag coefficient.

[0164] Furthermore, in the process of constructing the regression model corresponding to the vehicle driving model, the vehicle driving model can be summarized into the following formula:

[0165] F t =β1u 2 +β2cosα+β3(d u / d t )+β4sinα

[0166] Based on this formula, the regression model is obtained as follows:

[0167] Y = F t

[0168] X1=u 2

[0169] X2=cosα

[0170] X3=d u / d t

[0171] X4=sinα

[0172] The solution process for the above regression model includes: based on the actual values ​​of the input variables and the actual values ​​of the output variables in the vehicle driving model, using the least squares method, performing multiple regression on the regression model to obtain the input data for the regression model. Specifically, by inputting Y, X1, X2, X3, and X4, β1, β2, β3, and β4 are obtained. The formula for calculating Y is:

[0173] F t =(T tq *i0*i k *n t ) / r D

[0174] After obtaining β1, β2, β3, and β4, the parameters of β1, β2, and β3 are solved inversely to obtain the updated values ​​of the drag coefficients. The inverse parameter solution is expressed as:

[0175] C d =2*β1 / (ρ*A)

[0176] f = β² / (m*g)

[0177] δ=β3 / m

[0178] Optionally, by performing multiple regression on the regression model, the slope resistance term β4 can be obtained. By verifying the slope resistance term β4, it can be determined whether the updated value of the resistance coefficient is reliable. If the updated value of the resistance coefficient is reliable, the resistance coefficient can be updated to this updated value.

[0179] For example, if β4≈g, then the updated value of the drag coefficient is reliable.

[0180] As can be seen from the above formula, the actual driving speed, actual slope angle, actual driving acceleration, engine output torque, vehicle mass, wheel radius, vehicle frontal area, final drive ratio, transmission ratio in K gear, and transmission system mechanical efficiency can be obtained from the fourth set of actual driving data. Based on the engine output torque, final drive ratio, transmission ratio in K gear, transmission system mechanical efficiency, and wheel radius, the actual driving force (i.e., the actual value of the output variable in the vehicle driving model, the output variable being Y in the aforementioned formula) is determined. Based on the actual driving speed, actual slope angle, and actual driving acceleration, the actual values ​​of the input variables in the vehicle driving model (i.e., X1, X2, X3, and X4 in the aforementioned formula) are determined. The vehicle mass and vehicle frontal area are then used in the subsequent inverse solution.

[0181] The meanings of the variables in the above formulas can be found in the explanations of the foregoing embodiments, and will not be repeated here.

[0182] In some embodiments, after fusing the first predicted motor energy and the second predicted motor energy to obtain the target predicted motor energy under the power distribution strategy, the method further includes: training the strategy model based at least on the target predicted fuel consumption and the target predicted motor energy to obtain the strategy model after the current training. The target predicted fuel consumption and the target predicted motor energy reflect the rationality of the power distribution strategy generated by the strategy model from an energy consumption perspective. Training the strategy model based on the target predicted fuel consumption and the target predicted motor energy can improve the training effect of the strategy model, enabling it to generate a better power distribution strategy that minimizes vehicle energy consumption.

[0183] Optionally, if the physical model also includes a battery model, the strategy model can be trained based on the target predicted energy consumption, the target predicted motor energy, and the predicted battery SOC to obtain the strategy model after the current training, thereby further improving the training effect of the strategy model.

[0184] Optionally, during the training of the policy model, reinforcement learning algorithms and model distillation methods can be used to train the policy model in order to improve the training effect and efficiency of the policy model.

[0185] Optionally, after training the strategy model, the strategy model can be deployed in the vehicle's onboard equipment (e.g., vehicle infotainment system) to generate a reasonable power distribution strategy for the driver's required torque during vehicle operation.

[0186] As an example, Figure 10 An example diagram of an intelligent energy management system for hybrid vehicles, such as... Figure 10 As shown, the intelligent energy management system involves the training of a strategy model, the modeling of a vehicle model, and the prediction of energy consumption from the vehicle model. First, during the training of the strategy model, the required input data is fed into the model to obtain the power distribution strategy generated by the strategy model. Next, this power distribution strategy is sent to the vehicle simulation system. The vehicle simulation system includes a physical model obtained through physics-based driven modeling and a data-driven model obtained through training data driven modeling. The physical model and the data model are fused to predict energy consumption, resulting in the vehicle's energy consumption performance under the power distribution strategy. Then, the vehicle's energy consumption performance under the power distribution strategy is fed back to the strategy model's optimization algorithm (e.g., a reinforcement learning algorithm). Based on this vehicle energy consumption performance, the optimization algorithm trains the strategy model. After the strategy model training is complete, during its application, vehicle control commands can be issued to the real vehicle based on the power distribution strategy generated by the strategy model, thereby improving the rationality of power distribution during real vehicle operation.

[0187] like Figure 10As shown, real-world driving data from real vehicles can be uploaded to a cloud server. This data can be used for dynamic updates of the vehicle model, including parameter identification and optimization of the physical model. For example, it can be used to update the parameters of the vehicle driving model and battery model in the physical model. It can also be used to dynamically optimize the weight parameters involved in the fusion of the physical model and the data-driven model, and it can also be used for model training of the data-driven model.

[0188] Exemplary device

[0189] Corresponding to the above-mentioned vehicle model-based simulation testing method, this application also provides a vehicle model-based simulation testing device. Figure 11 This is a schematic diagram of the structure of a simulation testing device based on a vehicle model provided in an embodiment of this application. Figure 11 As shown in the embodiment of this application, the vehicle model-based simulation testing device 1100 includes: a strategy acquisition unit 1101, a first testing unit 1102, a second testing unit 1103, a fuel consumption fusion unit 1104, and an electric energy fusion unit 1105.

[0190] The strategy acquisition unit 1101 is used to acquire the power distribution strategy of the hybrid vehicle;

[0191] The first test unit 1102 is used to simulate and test the power distribution strategy using the first engine model and the first motor model to obtain the first predicted fuel consumption and the first predicted motor power. The first engine model and the first motor model are physical models.

[0192] The second test unit 1103 is used to simulate and test the power distribution strategy based on the second engine model and the second motor model to obtain the second predicted fuel consumption and the second predicted motor energy. The second engine model and the second motor model are data-driven models trained based on historical data.

[0193] The fuel consumption fusion unit 1104 is used to fuse the first predicted fuel consumption and the second predicted fuel consumption to obtain the target predicted fuel consumption under the power distribution strategy.

[0194] The power fusion unit 1105 is used to fuse the first predicted motor power and the second predicted motor power to obtain the target predicted motor power under the power distribution strategy.

[0195] In some embodiments, the fuel consumption fusion unit 1104 is specifically used to: weight the first predicted fuel consumption and the second predicted fuel consumption according to the weight parameters corresponding to the first engine model and the second engine model to obtain the target predicted fuel consumption; and / or, the electric energy fusion unit is specifically used to: weight the first predicted motor energy and the second predicted motor energy according to the weight parameters corresponding to the first motor model and the second motor model to obtain the target predicted motor energy.

[0196] In some embodiments, the weight parameters corresponding to the first engine model and the second engine model support dynamic updates. The dynamic update process of the weight parameters corresponding to the first engine model and the second engine model includes: acquiring first actual driving data of the hybrid vehicle; determining a fuel consumption prediction error based on the first actual driving data, the first engine model, and the second engine model; updating the weight parameters corresponding to the first engine model and the second engine model based on the fuel consumption prediction error; and / or, the weight parameters corresponding to the first motor model and the second motor model support dynamic updates. The dynamic update process of the weight parameters corresponding to the first engine model and the second engine model includes: acquiring second actual driving data of the hybrid vehicle; determining an energy prediction error based on the second actual driving data, the first motor model, and the second motor model; updating the weight parameters corresponding to the first motor model and the second motor model based on the energy prediction error.

[0197] In some embodiments, the fuel consumption prediction error includes the instantaneous fuel consumption prediction error of the first engine model and the instantaneous fuel consumption prediction error of the second engine model. Based on the fuel consumption prediction error, the weight parameters corresponding to the first engine model and the second engine model are updated, including: determining updated weight parameters corresponding to the first engine model and the second engine model based on the instantaneous fuel consumption prediction errors of the first and second engine models; determining the updated weight parameters corresponding to the first engine model as the weight parameters corresponding to the first engine model, and determining the updated weight parameters corresponding to the second engine model as the weight parameters corresponding to the second engine model.

[0198] In some embodiments, determining the updated weight parameter corresponding to the first engine model as the weight parameter corresponding to the first engine model, and determining the updated weight parameter corresponding to the second engine model as the weight parameter corresponding to the second engine model, includes: determining the cumulative fuel consumption error obtained by combining the first engine model and the second engine model for fuel consumption prediction based on the instantaneous fuel consumption prediction error of the first engine model, the instantaneous fuel consumption prediction error of the second engine model, and the actual fuel consumption in the reference period in the first actual driving data; if the cumulative fuel consumption error is less than the fuel consumption error threshold, determining the updated weight parameter corresponding to the first engine model as the weight parameter corresponding to the first engine model, and determining the updated weight parameter corresponding to the second engine model as the weight parameter corresponding to the second engine model.

[0199] In some embodiments, after determining the cumulative fuel consumption error obtained by combining the first engine model and the second engine model for fuel consumption prediction, the method further includes: if the cumulative fuel consumption error is greater than or equal to the fuel consumption error threshold, re-determining the update weight parameters corresponding to the first engine model and the update weight parameters corresponding to the second engine model, and / or retraining the second engine model.

[0200] In some embodiments, the power distribution strategy includes allocated engine speed and allocated engine output torque. The first test unit 1102 is specifically used to: determine the predicted engine specific fuel consumption based on the allocated engine speed and allocated engine output torque. The predicted engine specific fuel consumption is the engine specific fuel consumption when the engine of the hybrid vehicle operates according to the allocated engine speed and allocated engine output torque; and input the allocated engine speed, allocated engine output torque and predicted engine specific fuel consumption into the first engine model to obtain the first predicted fuel consumption.

[0201] In some embodiments, the first test unit 1102 is specifically used to: interpolate the engine test data to obtain a predicted engine specific fuel consumption based on the allocated engine speed and allocated engine output torque; or, input the allocated engine speed and allocated engine output torque into a specific fuel consumption mapping model to obtain a predicted engine specific fuel consumption; wherein, the engine test data includes multi-dimensional data consisting of engine speed, engine output torque and engine specific fuel consumption, the specific fuel consumption mapping model is a model of the mapping relationship between engine speed, engine output torque and engine specific fuel consumption, and the specific fuel consumption mapping model is obtained by fitting the engine test data.

[0202] In some embodiments, the power distribution strategy includes the allocated motor speed and the allocated motor output torque. The first test unit 1102 is specifically used to: determine the predicted motor charging efficiency and the predicted motor discharging efficiency based on the allocated motor speed and the allocated motor output torque. The predicted motor charging efficiency and the predicted motor discharging efficiency are the charging efficiency and discharging efficiency of the hybrid vehicle's motor when it operates according to the allocated motor speed and the allocated motor output torque; input the allocated motor speed, the allocated motor output torque, and the predicted motor charging efficiency into the motor charging model in the first motor model to obtain the predicted charging energy in the first predicted motor energy; input the allocated motor speed, the allocated motor output torque, and the predicted motor discharging efficiency into the motor discharging model in the first motor model to obtain the predicted discharging energy in the first predicted motor energy.

[0203] In some embodiments, the first test unit 1102 is specifically used to: interpolate the assigned motor speed and assigned motor output torque in the motor test data to obtain the predicted motor charging efficiency and predicted motor discharging efficiency; or, input the assigned motor speed and assigned motor output torque into the efficiency mapping model to obtain the predicted motor charging efficiency and predicted motor discharging efficiency; wherein, the motor test data includes multi-dimensional data consisting of motor speed, motor output torque, motor charging efficiency and motor discharging efficiency, and the efficiency mapping model is a model of the mapping relationship between motor speed, motor output torque, motor charging efficiency and motor discharging efficiency, and the efficiency mapping model is obtained by fitting the motor test data.

[0204] In some embodiments, the physical model further includes a battery model, and the vehicle model-based simulation testing device further includes a third testing unit (not shown in the figure), used to perform simulation testing on the power distribution strategy through the battery model to obtain the predicted battery state of charge (SOC).

[0205] In some implementations, the battery model parameters include the rated capacitance, which supports dynamic updates. The updating process of the rated capacitance includes: acquiring third actual driving data of the hybrid vehicle; determining the actual capacitance based on the third actual driving data; comparing the actual capacitance with the rated capacitance to obtain a comparison result; and updating the rated capacitance based on the comparison result.

[0206] In some embodiments, the physical model further includes a vehicle driving model, and the strategy acquisition unit 1101 is specifically used to: acquire vehicle information of the hybrid vehicle and the driver's required torque of the hybrid vehicle, wherein the vehicle information includes first input data required by the vehicle driving model and second input data required by the strategy model; obtain the predicted driving speed of the hybrid vehicle under the driver's required torque based on the first input data and through the vehicle driving model; and generate a power distribution strategy based on the predicted driving speed, the driver's required torque and the second input data and through the strategy model.

[0207] In some embodiments, the drag coefficient in the vehicle driving model supports dynamic updating. The dynamic updating process of the drag coefficient includes: acquiring fourth actual driving data of the hybrid vehicle; determining the vehicle driving prediction error of the vehicle driving model for the hybrid vehicle based on the fourth actual driving data and the vehicle driving model; determining the updated value of the drag coefficient by using a parameter identification method when it is determined that the drag coefficient needs to be updated based on the vehicle driving prediction error; and updating the drag coefficient based on the updated value.

[0208] In some embodiments, the vehicle model-based simulation testing apparatus further includes: a model training unit (not shown in the figure), used to train the strategy model based at least on the target predicted fuel consumption and the target predicted motor energy, to obtain the strategy model after the current training.

[0209] The vehicle model-based simulation testing device provided in this embodiment belongs to the same concept as the vehicle model-based simulation testing method provided in the above embodiments of this application. It can execute the vehicle model-based simulation testing method provided in any of the above embodiments of this application and has the corresponding functional modules and beneficial effects for executing the vehicle model-based simulation testing method. Technical details not described in detail in this embodiment can be found in the specific processing content of the vehicle model-based simulation testing method provided in the above embodiments of this application, and will not be repeated here.

[0210] The functions implemented by the above-mentioned acquisition unit 1101, first test unit 1102, second test unit 1103, fuel consumption fusion unit 1104 and power fusion unit 1105 can be implemented by the same or different processors, and this application embodiment does not limit them.

[0211] It should be understood that the units in the above device can be implemented by a processor calling software. For example, the device includes a processor connected to a memory containing instructions. The processor calls the instructions stored in the memory to implement any of the above methods or to implement the functions of each unit in the device. The processor can be a general-purpose processor, such as a CPU or microprocessor, and the memory can be internal or external to the device. Alternatively, the units in the device can be implemented as hardware circuits. By designing the hardware circuits, some or all of the unit functions can be implemented. The hardware circuits can be understood as one or more processors. For example, in one implementation, the hardware circuit is an ASIC, and the functions of some or all of the above units are implemented by designing the logical relationships between the components within the circuit. In another implementation, the hardware circuit can be implemented using a PLD, such as an FPGA, which can include a large number of logic gates. The connection relationships between the logic gates are configured through configuration files to implement the functions of some or all of the above units. All units in the above device can be implemented entirely by a processor calling software, entirely by hardware circuits, or partially by a processor calling software with the remaining parts implemented by hardware circuits.

[0212] In this application embodiment, a processor is a circuit with signal processing capabilities. In one implementation, the processor can be a circuit with instruction reading and execution capabilities, such as a CPU, microprocessor, GPU, or DSP. In another implementation, the processor can implement certain functions through the logical relationships of hardware circuits. These logical relationships are fixed or reconfigurable. For example, the processor may be a hardware circuit implemented as an ASIC or PLD, such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document and configuring the hardware circuit can be understood as the processor loading instructions to implement the functions of some or all of the above units. Furthermore, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as an NPU, TPU, or DPU.

[0213] As can be seen, each unit in the above device can be one or more processors (or processing circuits) configured to implement the above methods, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.

[0214] Furthermore, the units in the above devices can be integrated in whole or in part, or they can be implemented independently. In one implementation, these units are integrated together and implemented in the form of a System-on-Chip (SoC). The SoC may include at least one processor for implementing any of the above methods or implementing the functions of the units in the device. The at least one processor may be of different types, such as CPU and FPGA, CPU and artificial intelligence processor, CPU and GPU, etc.

[0215] Exemplary System

[0216] This application provides an electronic device, see [link to relevant documentation] Figure 12 As shown, the electronic device includes a memory 1200 and a processor 1210; wherein the memory 1200 is connected to the processor 1210 and is used to store programs; the processor 1210 is used to implement the vehicle model-based simulation testing method disclosed in any of the above embodiments by running the programs stored in the memory 1200.

[0217] Specifically, the aforementioned electronic device may also include: a bus, a communication interface 1220, an input device 1230, and an output device 1240.

[0218] The processor 1210, memory 1200, communication interface 1220, input device 1230, and output device 1240 are interconnected via a bus. Among them:

[0219] A bus can include a pathway for transmitting information between various components of a computer system.

[0220] The processor 1210 can be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the program of the present invention. It can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0221] The processor 1210 may include a main processor, as well as a baseband chip, modem, etc.

[0222] The memory 1200 stores a program that executes the technical solution of this invention, and may also store an operating system and other key business functions. Specifically, the program may include program code, which includes computer operation instructions. More specifically, the memory 1200 may include read-only memory (ROM), other types of static storage devices capable of storing static information and instructions, random access memory (RAM), other types of dynamic storage devices capable of storing information and instructions, disk storage, flash memory, etc.

[0223] Input device 1230 may include a device for receiving user input data and information, such as a keyboard, mouse, camera, scanner, light pen, voice input device, touch screen, pedometer, or gravity sensor.

[0224] Output device 1240 may include devices that allow information to be output to a user, such as a display screen, printer, speaker, etc.

[0225] The communication interface 1220 may include a device that uses any transceiver to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.

[0226] The processor 1210 executes the program stored in the memory 1200 and calls other devices, and can be used to implement any of the steps of the vehicle model-based simulation test method provided in the above embodiments of this application.

[0227] This application also proposes a chip, which includes a processor and a data interface. The processor reads and runs a program stored in a memory through the data interface to execute the vehicle model-based simulation testing method described in any of the above embodiments. For details of the processing and its beneficial effects, please refer to the above embodiments of the vehicle model-based simulation testing method.

[0228] This application also proposes a vehicle equipped with the aforementioned vehicle-mounted equipment, on which a strategy model trained by the foregoing embodiments is deployed.

[0229] Exemplary computer program products and storage media

[0230] In addition to the methods and devices described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the vehicle model-based simulation testing methods according to various embodiments of this application as described in any of the above embodiments of this specification.

[0231] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0232] Furthermore, embodiments of this application may also be storage media storing computer programs, which are executed by a processor to perform the steps of the vehicle model-based simulation testing method according to various embodiments of this application described in any of the above embodiments of this specification, specifically implementing the steps of the vehicle model-based simulation testing method as described above.

[0233] For the foregoing method embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0234] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0235] The steps in the methods of the various embodiments of this application can be adjusted, merged, or deleted in order according to actual needs, and the technical features described in each embodiment can be replaced or combined.

[0236] The units of the apparatus in the various embodiments of this application can be merged, divided, and deleted according to actual needs.

[0237] It should be understood that the disclosed terminals, devices, and methods can be implemented in other ways, given the several embodiments provided in this application. For example, the terminal embodiments described above are merely illustrative. For instance, the division of modules or sub-modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple sub-modules or modules may be combined or integrated into another module, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0238] The modules or submodules described as separate components may or may not be physically separate. The components that constitute a module or submodule may or may not be physical modules or submodules; that is, they may be located in one place or distributed across multiple network modules or submodules. Some or all of the modules or submodules can be selected to achieve the purpose of this embodiment's solution, depending on actual needs.

[0239] Furthermore, the functional modules or sub-modules in the various embodiments of this application can be integrated into one processing module, or each module or sub-module can exist physically separately, or two or more modules or sub-modules can be integrated into one module. The integrated modules or sub-modules described above can be implemented in hardware or in the form of software functional modules or sub-modules.

[0240] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0241] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software unit executed by a processor, or a combination of both. The software unit can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0242] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0243] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A simulation testing method based on a vehicle model, characterized in that, include: Obtain the power distribution strategy for hybrid vehicles; The power distribution strategy is simulated and tested using a first engine model and a first motor model to obtain a first predicted fuel consumption and a first predicted motor energy. The first engine model and the first motor model are the engine model and motor model in the physical model of the hybrid vehicle, respectively. Based on the second engine model and the second motor model, the power distribution strategy is simulated and tested to obtain the second predicted fuel consumption and the second predicted motor energy. The second engine model and the second motor model are data-driven models trained based on historical data. The first predicted fuel consumption and the second predicted fuel consumption are fused to obtain the target predicted fuel consumption under the power distribution strategy; The first predicted motor energy and the second predicted motor energy are fused to obtain the target predicted motor energy under the power distribution strategy.

2. The simulation testing method based on a vehicle model according to claim 1, characterized in that, The step of fusing the first predicted fuel consumption and the second predicted fuel consumption to obtain the target predicted fuel consumption under the power distribution strategy includes: Based on the weight parameters corresponding to the first engine model and the weight parameters corresponding to the second engine model, the first predicted fuel consumption and the second predicted fuel consumption are weighted to obtain the target predicted fuel consumption. And / or, the fusion of the first predicted motor energy and the second predicted motor energy to obtain the target predicted motor energy under the power distribution strategy includes: Based on the weight parameters corresponding to the first motor model and the second motor model, the predicted motor energy and the predicted motor energy are weighted to obtain the target predicted motor energy.

3. The simulation testing method based on a vehicle model according to claim 2, characterized in that, The weight parameters corresponding to the first engine model and the second engine model support dynamic updates. The dynamic update process for the weight parameters corresponding to the first engine model and the second engine model includes: Obtain the first actual driving data of the hybrid vehicle; Based on the first actual driving data, the first engine model, and the second engine model, determine the fuel consumption prediction error; Based on the fuel consumption prediction error, the weight parameters corresponding to the first engine model and the weight parameters corresponding to the second engine model are updated; And / or, the weight parameters corresponding to the first motor model and the weight parameters corresponding to the second motor model support dynamic updates. The dynamic update process of the weight parameters corresponding to the first engine model and the weight parameters corresponding to the second engine model includes: Obtain the second actual driving data of the hybrid vehicle; The power prediction error is determined based on the second actual driving data, the first motor model, and the second motor model. Based on the power prediction error, the weight parameters corresponding to the first motor model and the weight parameters corresponding to the second motor model are updated.

4. The simulation testing method based on a vehicle model according to claim 3, characterized in that, The fuel consumption prediction error includes the instantaneous fuel consumption prediction error of the first engine model and the instantaneous fuel consumption prediction error of the second engine model. The step of updating the weight parameters corresponding to the first engine model and the second engine model based on the fuel consumption prediction error includes: Based on the instantaneous fuel consumption prediction error of the first engine model and the instantaneous fuel consumption prediction error of the second engine model, the update weight parameters corresponding to the first engine model and the update weight parameters corresponding to the second engine model are determined. The updated weight parameter corresponding to the first engine model is determined as the weight parameter corresponding to the first engine model, and the updated weight parameter corresponding to the second engine model is determined as the weight parameter corresponding to the second engine model.

5. The simulation testing method according to claim 4, characterized in that, The step of determining the updated weight parameter corresponding to the first engine model as the weight parameter corresponding to the first engine model, and determining the updated weight parameter corresponding to the second engine model as the weight parameter corresponding to the second engine model, includes: Based on the instantaneous fuel consumption prediction error of the first engine model, the instantaneous fuel consumption prediction error of the second engine model, and the actual fuel consumption during the reference period in the first actual driving data, the cumulative fuel consumption error obtained by combining the first engine model and the second engine model for fuel consumption prediction is determined. When the cumulative fuel consumption error is less than the fuel consumption error threshold, the updated weight parameter corresponding to the first engine model is determined as the weight parameter corresponding to the first engine model, and the updated weight parameter corresponding to the second engine model is determined as the weight parameter corresponding to the second engine model. After determining the cumulative fuel consumption error obtained by combining the first engine model and the second engine model for fuel consumption prediction, the method further includes: If the cumulative fuel consumption error is greater than or equal to the fuel consumption error threshold, the updated weight parameters corresponding to the first engine model and the updated weight parameters corresponding to the second engine model are re-determined, and / or the second engine model is retrained.

6. The simulation testing method based on a vehicle model according to any one of claims 1 to 5, characterized in that, The power distribution strategy includes the allocated engine speed and the allocated engine output torque. Using the first engine model, the power distribution strategy is simulated and tested to obtain the first predicted fuel consumption, including: Based on the allocated engine speed and the allocated engine output torque, a predicted engine specific fuel consumption is determined. The predicted engine specific fuel consumption is the engine specific fuel consumption when the engine of the hybrid vehicle operates according to the allocated engine speed and the allocated engine output torque. The allocated engine speed, the allocated engine output torque, and the predicted engine specific fuel consumption are input into the first engine model to obtain the first predicted fuel consumption. And / or, the power distribution strategy includes the allocated motor speed and the allocated motor output torque. Using the first motor model, the power distribution strategy is simulated and tested to obtain the first predicted motor energy, including: Based on the allocated motor speed and the allocated motor output torque, the predicted motor charging efficiency and the predicted motor discharging efficiency are determined. The predicted motor charging efficiency and the predicted motor discharging efficiency are the charging efficiency and discharging efficiency of the motor of the hybrid vehicle when it operates according to the allocated motor speed and the allocated motor output torque. The allocated motor speed, the allocated motor output torque, and the predicted motor charging efficiency are input into the motor charging model in the first motor model to obtain the predicted charging energy in the first predicted motor energy. The allocated motor speed, the allocated motor output torque, and the predicted motor discharge efficiency are input into the motor discharge model in the first motor model to obtain the predicted discharge energy in the first predicted motor energy.

7. The simulation testing method based on a vehicle model according to claim 6, characterized in that, Determining the predicted engine specific fuel consumption based on the allocated engine speed and the allocated engine output torque includes: Based on the allocated engine speed and the allocated engine output torque, the predicted engine specific fuel consumption is obtained by interpolation in the engine test data. Alternatively, the allocated engine speed and the allocated engine output torque can be input into the specific fuel consumption mapping model to obtain the predicted engine specific fuel consumption; The engine test data includes multi-dimensional data consisting of engine speed, engine output torque, and engine specific fuel consumption. The specific fuel consumption mapping model is a model of the mapping relationship between engine speed, engine output torque, and engine specific fuel consumption. The specific fuel consumption mapping model is obtained by fitting the engine test data. And / or, determining the predicted motor charging efficiency and predicted motor discharging efficiency based on the allocated motor speed and the allocated motor output torque includes: Based on the allocated motor speed and the allocated motor output torque, the predicted motor charging efficiency and the predicted motor discharging efficiency are obtained by interpolation in the motor test data. Alternatively, the predicted motor charging efficiency and the predicted motor discharging efficiency can be obtained by mapping the allocated motor speed and the allocated motor output torque into the input efficiency model. The motor test data includes multi-dimensional data consisting of motor speed, motor output torque, motor charging efficiency, and motor discharging efficiency. The efficiency mapping model is a model of the mapping relationship between motor speed, motor output torque, motor charging efficiency, and motor discharging efficiency. The efficiency mapping model is obtained by fitting the motor test data.

8. The simulation testing method based on a vehicle model according to any one of claims 1 to 5, characterized in that, The physical model also includes a battery model, and after obtaining the power distribution strategy for the hybrid vehicle, it further includes: The power distribution strategy is simulated and tested using the battery model to obtain the predicted battery state of charge (SOC). The battery model's model parameters include a rated capacitance, which supports dynamic updates. The updating process for the rated capacitance includes: Obtain the third actual driving data of the hybrid vehicle; The actual capacitance is determined based on the third actual driving data; The actual capacitance and the rated capacitance are compared to obtain the comparison result; Based on the comparison results, the rated capacitor is updated.

9. The simulation testing method based on a vehicle model according to any one of claims 1 to 5, characterized in that, The physical model also includes a vehicle driving model, and the process of obtaining the power distribution strategy for hybrid vehicles includes: Obtain the vehicle information of the hybrid vehicle and the driver's required torque of the hybrid vehicle. The vehicle information includes the first input data required by the vehicle driving model and the second input data required by the strategy model. Based on the first input data, the predicted driving speed of the hybrid vehicle is obtained through the vehicle driving model; Based on the predicted driving speed, the driver's required torque, and the second input data, the power distribution strategy is generated through the strategy model. The drag coefficient in the vehicle driving model supports dynamic updating, and the dynamic updating process of the drag coefficient includes: Obtain the fourth actual driving data of the hybrid vehicle; Based on the fourth actual driving data and the vehicle driving model, the vehicle driving prediction error of the vehicle driving model for the hybrid vehicle is determined. When the drag coefficient needs to be updated based on the vehicle driving prediction error, the updated value of the drag coefficient is determined by parameter identification. The drag coefficient is updated based on the updated value.

10. A simulation testing device based on a vehicle model, characterized in that, include: The strategy acquisition unit is used to acquire the power distribution strategy of the hybrid vehicle. The first test unit is used to simulate and test the power distribution strategy using a first engine model and a first motor model to obtain a first predicted fuel consumption and a first predicted motor energy. The first engine model and the first motor model are physical models. The second test unit is used to simulate and test the power distribution strategy based on the second engine model and the second motor model to obtain the second predicted fuel consumption and the second predicted motor power. The second engine model and the second motor model are data-driven models trained based on historical data. A fuel consumption fusion unit is used to fuse the first predicted fuel consumption and the second predicted fuel consumption to obtain the target predicted fuel consumption under the power distribution strategy. The power fusion unit is used to fuse the first predicted motor power and the second predicted motor power to obtain the target predicted motor power under the power distribution strategy.

Images