Oil consumption prediction method, device and system and storage medium

Abstract

The invention discloses an oil consumption prediction method, device and system and a storage medium. The method comprises the steps that a joint simulation model composed of a whole vehicle transmission system model, a thermal management thermal model, an engine physical model and an engine control model is constructed; simplifying the engine physical model from a detailed physical model to a parameterized mean value model through neural network training based on rack normal temperature data; generating a correction coefficient for the parameterized mean value model according to the whole vehicle transmission system model, the thermal management thermal model and the engine control model; correcting the parameterized mean value model according to the correction coefficient; the instantaneous specific fuel consumption of the engine is output based on the corrected parameterized mean value model; and calculating the dynamic fuel consumption of the whole vehicle according to the instantaneous specific fuel consumption of the engine in combination with the road spectrum integral of the transmission system of the whole vehicle. By adopting the scheme, the economical efficiency and the calculation efficiency of oil consumption prediction can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automobile engineering, and in particular relates to an oil consumption prediction method, device, system and storage medium. BACKGROUND

[0002] Engine instantaneous fuel consumption, as an important indicator of engine fuel efficiency, is extremely sensitive to changes in driving behavior and vehicle operating conditions. Therefore, how to quickly and accurately predict engine instantaneous fuel consumption has become an important challenge in optimizing vehicle performance and improving energy efficiency. Traditional engine fuel consumption estimation methods, such as Map lookup table method and one-dimensional detailed model analysis of steady state conditions, have many limitations: the former relies on a large amount of bench test data, resulting in high testing costs; the latter has slow calculation speed, and the differences between steady state and transient vehicle boundaries are significant, resulting in large differences in fuel consumption prediction.

[0003] Therefore, how to provide an oil consumption prediction method to improve the economy and calculation efficiency of oil consumption prediction has become a technical problem to be solved. SUMMARY

[0004] The present application provides an oil consumption prediction method, device, system and storage medium to improve the economy and calculation efficiency of oil consumption prediction.

[0005] The present application provides an oil consumption prediction method, comprising: constructing a co-simulation model composed of a vehicle driveline model, a thermal management thermal model, an engine physical model and an engine control model; based on bench normal temperature data, simplifying the engine physical model from a detailed physical model to a parameterized mean model through neural network training; generating correction coefficients for the parameterized mean model according to the vehicle driveline model, the thermal management thermal model and the engine control model; correcting the parameterized mean model according to the correction coefficients; outputting engine instantaneous specific fuel consumption based on the corrected parameterized mean model; calculating vehicle dynamic fuel consumption according to the engine instantaneous specific fuel consumption and integrating the vehicle driveline spectrum.

[0006] The application has the beneficial effects that: a co-simulation model composed of a whole vehicle drive train model, a thermal management thermal model, an engine physical model and an engine control model is constructed; the engine physical model is simplified from a detailed physical model to a parameterized mean model through neural network training based on bench normal temperature data; a correction coefficient of the parameterized mean model is generated according to the whole vehicle drive train model, the thermal management thermal model and the engine control model; the parameterized mean model is corrected according to the correction coefficient; engine instantaneous specific fuel consumption is output based on the corrected parameterized mean model; and whole vehicle dynamic fuel consumption is calculated by integrating the engine instantaneous specific fuel consumption and a whole vehicle drive train spectrum. Since the neural network is trained by using the bench normal temperature data to replace the multi-temperature bench test and the full-temperature performance is verified by the co-simulation model, the real-time fuel consumption prediction under the full-temperature condition can be realized through the bench normal temperature test, the accuracy of the fuel consumption prediction is ensured and the economy of the fuel consumption prediction is improved; and since the engine physical model is simplified from the detailed physical model to the parameterized mean model, the equation solving with high calculation complexity is avoided and the calculation efficiency is improved by relying on the lightweight neural network inference.

[0007] In one embodiment, the parameterized mean model is simplified from the detailed physical model of the engine through neural network training based on the bench normal temperature data, comprising: The steady state operating condition data of the engine under the bench normal temperature are taken as a training set, the speed, the intake manifold temperature / pressure, the exhaust manifold temperature / pressure and the EGR rate are taken as input features, the volumetric efficiency, the indicated effective pressure and the exhaust temperature are taken as output labels, the network weight is optimized through a back propagation algorithm, and a parameterized mapping relationship is generated.

[0008] In one embodiment, the correction coefficient of the parameterized mean model is generated according to the whole vehicle drive train model, the thermal management thermal model and the engine control model, comprising: The engine operating speed and torque are determined according to the whole vehicle drive train model and the thermal management model; The temperature signal corresponding to the engine operating speed and torque is determined through the engine physical model and the thermal management model; The temperature signal is input into the engine control model to generate the correction coefficient of the parameterized mean model.

[0009] In one embodiment, the engine operating speed and torque are determined according to the whole vehicle drive train model, comprising: The drive train energy consumption is calculated according to the whole vehicle drive train model; The accessory power consumption is calculated according to the thermal management model; The whole vehicle demand power is determined according to the drive train energy consumption and the accessory power consumption; The engine operating speed and torque are determined according to the whole vehicle demand power.

[0010] In one embodiment, the determination of the temperature signal corresponding to the engine operating speed and torque by the engine physical model and the thermal management model comprises: The engine physical model calculates the pre-intercooler intake air temperature, intake air flow and water heat transfer amount by the engine operating speed and torque; The thermal management model outputs the temperature signal by the pre-intercooler intake air temperature, intake air flow and water heat transfer amount in combination with the vehicle thermal boundary condition, wherein the temperature signal comprises the post-intercooler intake air temperature and engine cooling water temperature.

[0011] In one embodiment, the input of the temperature signal into the engine control model to generate the correction coefficient of the parameterized mean model comprises: The engine control model calculates the EGR rate correction and ignition angle correction under different temperature boundaries according to the temperature signal in combination with the ambient temperature signal; The correction coefficient of the parameterized mean model is determined according to the theoretical influence of the EGR rate correction and ignition angle correction on fuel consumption.

[0012] In one embodiment, the determination of the correction coefficient of the parameterized mean model according to the theoretical influence of the EGR rate correction and ignition angle correction on fuel consumption comprises: A mathematical mapping relationship between the correction value and the fuel consumption change is established by analyzing the influence of EGR rate change on in-cylinder gas composition and combustion efficiency and the influence of ignition angle change on combustion phase and indicated thermal efficiency; The EGR rate correction coefficient and the ignition angle correction coefficient are coupled to generate a correction coefficient applicable to the current working condition.

[0013] The application also provides an oil consumption prediction device, comprising: A construction module is configured to construct a joint simulation model composed of a vehicle driveline model, a thermal management thermal model, an engine physical model and an engine control model; A simplification module is configured to simplify the engine physical model from a detailed physical model to a parameterized mean model by neural network training based on bench normal temperature data; A generation module is configured to generate a correction coefficient of the parameterized mean model according to the vehicle driveline model, the thermal management thermal model and the engine control model; A correction module is configured to correct the parameterized mean model according to the correction coefficient; An output module is configured to output engine instantaneous specific fuel consumption based on the corrected parameterized mean model; A calculation module is configured to calculate vehicle dynamic fuel consumption by integrating the vehicle driveline spectrum according to the engine instantaneous specific fuel consumption.

[0014] In one embodiment, the simplification module is further configured to: The steady-state operating condition data of the engine at the bench ambient temperature are taken as a training set, the rotating speed, intake manifold temperature / pressure, exhaust manifold temperature / pressure and EGR rate are taken as input features, the volumetric efficiency, indicated effective pressure and exhaust temperature are taken as output labels, the network weight is optimized through a back propagation algorithm, and a parameterized mapping relationship is generated.

[0015] In one embodiment, the generation module comprises: The first determination submodule is configured to determine the engine operating rotating speed and torque according to the vehicle driveline model and the thermal management model. The second determination submodule is configured to determine the temperature signal corresponding to the engine operating rotating speed and torque through the engine physical model and the thermal management model. The generation submodule is configured to input the temperature signal into the engine control model to generate a correction coefficient of the parameterized mean model.

[0016] In one embodiment, the first determination submodule is further configured to: The driveline energy consumption is calculated according to the vehicle driveline model. The accessory power consumption is calculated according to the thermal management model. The vehicle demand power is determined according to the driveline energy consumption and the accessory power consumption. The engine operating rotating speed and torque are determined according to the vehicle demand power.

[0017] In one embodiment, the second determination submodule is further configured to: The engine physical model calculates the pre-intercooler intake air temperature, intake air flow and water heat transfer amount through the engine operating rotating speed and torque. The thermal management model outputs the temperature signal including the post-intercooler intake air temperature and engine cooling water temperature through the pre-intercooler intake air temperature, intake air flow and water heat transfer amount in combination with the vehicle thermal boundary condition.

[0018] In one embodiment, the generation submodule is further configured to: The engine control model calculates the EGR rate correction and ignition angle correction under different temperature boundaries according to the temperature signal in combination with the ambient temperature signal. The correction coefficient of the parameterized mean model is determined according to the theoretical influence of the EGR rate correction and the ignition angle correction on fuel consumption.

[0019] In one embodiment, the correction coefficient of the parameterized mean model is determined according to the theoretical influence of the EGR rate correction and the ignition angle correction on fuel consumption, comprising: By analyzing the influence of the EGR rate change on the in-cylinder gas composition and combustion efficiency, and the influence of the ignition angle change on the combustion phase and indicated thermal efficiency, a mathematical mapping relationship between the correction value and the fuel consumption change is established. The EGR rate correction coefficient and the ignition angle correction coefficient are coupled to generate a correction coefficient suitable for the current working condition.

[0020] The application also provides an oil consumption prediction system, comprising: at least one processor; and, a memory in communication connection with the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to implement the oil consumption prediction method described in any of the above embodiments.

[0021] The application also provides a computer-readable storage medium, when the instructions in the storage medium are executed by the processor corresponding to the oil consumption prediction system, the oil consumption prediction system can implement the oil consumption prediction method described in any of the above embodiments.

[0022] Other features and advantages of the application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the application. The objects and other advantages of the application can be realized and obtained by the structure particularly pointed out in the written description, claims, and drawings.

[0023] The technical solutions of the application will be further described in detail below with the help of the drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0024] The accompanying drawings are used to provide a further understanding of the application, and constitute a part of the specification, together with the embodiments of the application, to explain the application, and do not constitute a limitation of the application. In the drawings: Figure 1 A flowchart of an oil consumption prediction method in an embodiment of the application; Figure 2 A structural schematic diagram of an oil consumption prediction device in an embodiment of the application; Figure 3 A hardware structural schematic diagram of an oil consumption prediction system in an embodiment of the application. DETAILED DESCRIPTION

[0025] The preferred embodiments of the application will be described below in conjunction with the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to explain and illustrate the application, and do not limit the application.

[0026] Figure 1This is a flowchart of a fuel consumption prediction method in one embodiment of this application, as shown below. Figure 1 As shown, the method can be implemented as follows: S101-S106: In step S101, a joint simulation model is constructed, consisting of a vehicle powertrain model, a thermal management model, an engine physical model, and an engine control model. In step S102, based on bench ambient temperature data, the engine physical model is simplified from a detailed physical model to a parameterized mean model through neural network training. In step S103, correction coefficients for the parameterized mean model are generated based on the vehicle powertrain model, thermal management thermal model, and engine control model. In step S104, the parameterized mean model is corrected according to the correction coefficient; In step S105, the instantaneous specific fuel consumption of the engine is output based on the modified parameterized mean model; In step S106, the dynamic fuel consumption of the vehicle is calculated based on the instantaneous specific fuel consumption of the engine and the road spectrum integral of the vehicle's transmission system.

[0027] A joint simulation model was constructed, consisting of a vehicle powertrain model, a thermal management model, an engine physical model, and an engine control model. Constructing this joint simulation model is the core technological foundation for achieving dynamic fuel consumption prediction across the entire temperature range. This model integrates four key sub-models to form a complete vehicle energy management system simulation platform.

[0028] The vehicle drivetrain model calculates drivetrain energy consumption based on road spectrum data (vehicle speed, gradient) and outputs the target engine speed / torque. The thermal management model simulates the heat exchange process between the engine cooling system and the intake system. By receiving heat transfer data from the drivetrain, engine water circuit, and the ambient thermal boundary, it outputs the intercooled intake temperature, coolant temperature, and accessory power consumption.

[0029] Engine physical models are typically built based on calibration data during bench testing, and their core function is to calculate key parameters such as combustion efficiency and emission characteristics. Traditional methods employ detailed physical equations for modeling, using complex combustion, fluid dynamics, and heat transfer equations to calculate core parameters like volumetric efficiency, indicated effective pressure, and exhaust temperature. However, such detailed physical models require solving high-dimensional nonlinear equations, resulting in high computational complexity and long execution times, making them unsuitable for real-time simulation and rapid evaluation.

[0030] To address this issue, this application proposes an optimization scheme: using a neural network model to replace or partially replace the computationally complex parameter calculation process in the detailed physical model. Specifically, this can be achieved in the following two ways: The first way is to completely replace the entire engine physical model with a parameterized mean model based on a neural network. The neural network model is trained by collecting a large amount of steady-state operating condition data in a constant temperature bench test, and learns the complex nonlinear mapping relationship between the input parameters (such as speed, intake pressure, exhaust temperature, EGR rate, etc.) and the output parameters (volume efficiency, indicated effective pressure, exhaust temperature, etc.).

[0031] The second way is to replace the neural network only for the calculation of specific parameters in the detailed physical model that are particularly complex. For example, the physical model calculation of other parameters is retained, and the calculation process of the volume efficiency, indicated effective pressure, and exhaust temperature parameters is replaced by a neural network model.

[0032] Taking the second way as an example, the specific replacement process is as follows: Collect steady-state operating condition data at constant temperature on the bench, and use speed, intake manifold temperature / pressure, exhaust manifold temperature / pressure, and EGR rate as input feature vectors, and volume efficiency, indicated effective pressure, and exhaust temperature as output label vectors. A feedforward neural network is constructed, the number of input layer nodes is equal to the input feature dimension (4), the number of output layer nodes is equal to the output label dimension (3), and the hidden layer can be set to 1-3 layers according to the data complexity. Using the back propagation algorithm, the network weight and bias parameters are iteratively optimized by minimizing the mean square error between the predicted output and the true label. An independent validation set is used to evaluate the model performance to ensure that the trained model has good generalization ability. The trained neural network model is deployed in the simulation system to replace the original complex physical model, achieving fast and accurate engine parameter prediction.

[0033] It can be understood that to realize the first way, i.e., to completely replace the entire engine physical model with a parameterized mean model based on a neural network, other parameters can also be replaced with a parameterized mean model based on a neural network according to the corresponding example of the second way, so as to completely replace the entire engine physical model with a parameterized mean model based on a neural network.

[0034] This method replaces the traditional complex physical model with a high-efficiency neural network model through data-driven approach. The significance of this replacement method is that on the one hand, the fast forward inference capability of the neural network model significantly improves the calculation efficiency; on the other hand, since the neural network model is trained based on real bench data, it can maintain high calculation accuracy, thereby significantly reducing the calculation complexity while ensuring prediction accuracy.

[0035] Since the parameterized mean model is obtained based on the bench normal temperature data, at present, only the fuel consumption predicted in the normal temperature environment is relatively accurate, therefore, in the present application, a correction coefficient of the parameterized mean model is generated according to the whole vehicle drive train model, the thermal management thermal model and the engine control model. The purpose of the correction coefficient is to dynamically correct the output of the parameterized mean model trained based on the normal temperature data to an accurate value that can reflect the influence of the current actual temperature environment (especially the intake temperature and the cooling water temperature).

[0036] In the generation of the correction coefficient of the parameterized mean model, the engine operating speed and torque are determined according to the whole vehicle drive train model and the thermal management model; the temperature signal corresponding to the engine operating speed and torque is determined through the engine physical model and the thermal management model; and the temperature signal is input into the engine control model to generate the correction coefficient of the parameterized mean model.

[0037] In the determination of the engine operating speed and torque according to the whole vehicle drive train model and the thermal management model, the following steps can be implemented: The drive train energy consumption is calculated according to the whole vehicle drive train model; for example, the driving resistance power of the vehicle is calculated according to the vehicle speed and the road slope in combination with the vehicle parameters, and the power is the power required to drive the vehicle; the driving resistance power of the vehicle is multiplied by the pre-calibrated drive train coefficient to obtain the drive train energy consumption. The accessory power consumption (such as air conditioner compressor, cooling fan and other thermal management accessories) is calculated according to the thermal management model; the whole vehicle demand power is determined according to the drive train energy consumption and the accessory power consumption, specifically, the whole vehicle demand power can be determined according to the sum of the drive train energy consumption and the accessory power consumption. The engine operating speed and torque are determined according to the whole vehicle demand power.

[0038] In the determination of the temperature signal corresponding to the engine operating speed and torque through the engine physical model and the thermal management model, the following steps can be implemented: The engine physical model calculates the intercooler front intake temperature, intake flow and water route heat transfer amount through the engine operating speed and torque.

[0039] GT-SUITE is a professional multi-physical field simulation analysis software package, mainly used for performance analysis and optimization in the fields of vehicles, engines, energy systems, etc. In the present application, when the intercooler front intake temperature, intake flow and water route heat transfer amount are calculated through the engine operating speed and torque, GT-SUITE and other software can be used.

[0040] The heat management model calculates the heat transfer amount of the drive train according to the vehicle drive train model, and outputs a temperature signal in combination with the vehicle thermal boundary condition, wherein the temperature signal comprises the intake air temperature after the intercooler and the engine cooling water temperature. The heat transfer amount of the drive train can be obtained by multiplying the energy consumption of the drive train calculated by the vehicle drive train model by the pre-calibrated drive train efficiency.

[0041] The step of inputting the temperature signal into the engine control model to generate the correction coefficient of the parameterized mean model can be implemented as the following steps: Firstly, the engine control model calculates the EGR rate correction and the ignition angle correction under different temperature boundaries according to the temperature signal and the ambient temperature signal. Then, the correction coefficient of the parameterized mean model is determined according to the theoretical influence of the EGR rate correction and the ignition angle correction on fuel consumption. Specifically, the mathematical mapping relationship between the correction value and the fuel consumption change is established by analyzing the influence of the EGR rate change on the in-cylinder gas composition and the combustion efficiency, and the influence of the ignition angle change on the combustion phase and the indicated thermal efficiency. The EGR rate correction coefficient and the ignition angle correction coefficient are coupled to generate a correction coefficient suitable for the current working condition.

[0042] The parameterized mean model is corrected according to the correction coefficient. Specifically, if the current actual temperature is consistent with the normal temperature data, no correction is needed, and if the current actual temperature is inconsistent with the normal temperature data, correction is needed. In this way, not only can the fuel consumption of the engine be accurately predicted in a normal temperature environment, but also accurate fuel consumption prediction results can be obtained in a high temperature environment or a low temperature environment. When training the parameterized mean model, only the bench normal temperature data is used, and the final mean model can realize "full temperature range" prediction under the action of the correction coefficient, thereby reducing the dependence on multi-temperature bench tests.

[0043] For example, the engine control model receives temperature signals (including intercooled intake air temperature and engine coolant water temperature) from the thermal management model and ambient temperature signals. Based on these temperature information, the control model simulates the operating characteristics of the engine under different thermal boundary conditions. For example, when the engine coolant water temperature is too high, the model will reduce the exhaust gas recirculation (EGR) rate and / or delay the ignition timing to prevent knock from occurring, and the specific reduction amount is determined according to the actual temperature of the engine coolant water. The adjustment amount of these control strategies is the EGR rate correction value and the ignition angle correction value. Secondly, based on the above control correction value, the theoretical influence on fuel consumption is analyzed. The increase of EGR rate will dilute the oxygen concentration in the mixture, change the composition of the cylinder gas, affect the combustion rate and combustion efficiency, and thus cause the change of fuel consumption. The delay of ignition angle will change the combustion phase, affect the completeness of combustion and indicated thermal efficiency, and thus affect the fuel consumption. Through theoretical analysis and bench calibration data, a quantitative relationship between the EGR rate correction value, the ignition angle correction value and the fuel consumption change is established. Finally, the EGR rate correction coefficient and the ignition angle correction coefficient are coupled. Considering the mutual influence of the two correction factors in actual working conditions, the two independent correction coefficients are integrated into a unified correction coefficient through mathematical methods (such as multiplication combination). This final correction coefficient can accurately reflect the change proportion of the actual fuel consumption of the engine relative to the predicted value of the normal temperature model under the current temperature condition and control strategy, so as to accurately correct the output of the parameterized mean model.

[0044] Based on the corrected parameterized mean model output engine instantaneous specific fuel consumption, the specific implementation is as follows: First, the engine operating condition parameters (including corrected speed, torque, intake temperature, EGR rate, etc.) after temperature correction are input into the parameterized mean model that has been temperature corrected. The parameterized mean model is based on neural network or other machine learning algorithms and has learned the mapping relationship between engine parameters and specific fuel consumption under normal temperature conditions. By introducing temperature correction factors, the model can map these corrected input parameters to the instantaneous specific fuel consumption value under the current working condition. This process is essentially using the trained model to perform forward calculation on the corrected input features to obtain the corresponding fuel consumption output, thereby realizing the rapid prediction of engine instantaneous fuel consumption.

[0045] According to the engine instantaneous specific fuel consumption, the dynamic fuel consumption of the whole vehicle is calculated by integrating the vehicle drive train spectrum, and the specific implementation is as follows: After obtaining the engine instantaneous specific fuel consumption, it is necessary to combine it with the dynamic road spectrum information of the vehicle driving. The vehicle driveline road spectrum integration process first needs to obtain the vehicle speed, acceleration, slope and other driving state information under a specific driving condition, which constitutes the driving road spectrum. Then, the engine instantaneous specific fuel consumption is matched with the driving state at the corresponding time, and the engine instantaneous fuel consumption at each time is accumulated through time integration, so as to calculate the total fuel consumption of the vehicle under the driving condition. Specifically, by discretizing the driving road spectrum into multiple time steps, the engine instantaneous specific fuel consumption in each time step is integrated and calculated, and finally the dynamic fuel consumption result of the vehicle is obtained. This process considers the dynamic change characteristics of the vehicle driving process, and can accurately reflect the influence of different working condition combinations on the fuel consumption of the vehicle.

[0046] The beneficial effects of the present application are: a co-simulation model composed of a vehicle driveline model, a thermal management thermal model, an engine physical model and an engine control model is constructed; based on bench normal temperature data, the engine physical model is simplified from a detailed physical model to a parameterized mean model through neural network training; correction coefficients of the parameterized mean model are generated according to the vehicle driveline model, the thermal management thermal model and the engine control model; the parameterized mean model is corrected according to the correction coefficients; engine instantaneous specific fuel consumption is output based on the corrected parameterized mean model; and vehicle dynamic fuel consumption is calculated by integrating the vehicle driveline road spectrum according to the engine instantaneous specific fuel consumption. Since the neural network is trained using bench normal temperature data to replace multi-temperature bench tests, and the full-temperature performance is verified through the co-simulation model, real-time fuel consumption prediction under full-temperature conditions can be realized through normal temperature bench tests, ensuring the accuracy of fuel consumption prediction and improving the economy of fuel consumption prediction; and by simplifying the engine physical model from a detailed physical model to a parameterized mean model, high computational complexity equation solving is avoided, and instead, lightweight neural network inference is relied on, improving the computational efficiency.

[0047] In one embodiment, the above step S102 can be implemented as the following steps: The engine steady-state condition data at the bench normal temperature are used as the training set, the speed, intake manifold temperature / pressure, exhaust manifold temperature / pressure and EGR rate are used as the input features, the volumetric efficiency, indicated effective pressure and exhaust temperature are used as the output labels, the network weight is optimized through the back propagation algorithm, and the parameterized mapping relationship is generated.

[0048] In one embodiment, the above step S103 can be implemented as the following steps A1-A3: In step A1, the engine operating speed and torque are determined according to the vehicle driveline model and the thermal management model; In step A2, the engine operating speed and torque corresponding temperature signal is determined by the engine physical model and the thermal management model; In step A3, the temperature signal is input into the engine control model to generate a correction coefficient for the parameterized mean model.

[0049] In one embodiment, the above step A1 can be implemented as steps A11-A14 as follows: In step A11, the powertrain energy consumption is calculated according to the vehicle powertrain model; In step A12, the accessory power consumption is calculated according to the thermal management model; In step A13, the vehicle demand power is determined according to the powertrain energy consumption and the accessory power consumption; In step A14, the engine operating speed and torque are determined according to the vehicle demand power.

[0050] In this embodiment, the total vehicle demand power is first calculated. Specifically, the powertrain energy consumption required to overcome the driving resistance is calculated by the vehicle powertrain model according to the road profile (vehicle speed, slope) and vehicle parameters (mass, wind resistance, etc.); at the same time, the thermal management accessory power consumption of the air conditioner compressor, cooling fan, etc. is calculated by the thermal management model; thus, the energy that the current vehicle must obtain from the power system, i.e. the vehicle demand power = powertrain energy consumption + thermal management accessory power consumption, is obtained.

[0051] Then the engine operating speed and torque are determined according to the vehicle demand power. Corresponding to the above total demand power, the part that needs to be borne by the engine is finally determined through the decision of the vehicle energy management strategy. This decision result is converted into the specific control instruction of the engine, i.e. the target operating speed (RPM) and target output torque (N·m) of the engine. This "speed-torque" point is the "load working condition point" that the engine needs to run at present.

[0052] In one embodiment, the above step A2 can be implemented as steps A21-A22 as follows: In step A21, the engine physical model calculates the intercooler pre-intake temperature, intake flow and water route heat transfer amount through the engine operating speed and torque; In step A22, the thermal management model outputs a temperature signal through the intercooler pre-intake temperature, intake flow and water route heat transfer amount, combined with the vehicle thermal boundary conditions, wherein the temperature signal includes the post-intercooler intake temperature and engine cooling water temperature.

[0053] Specifically, when air is drawn into the cylinder and compressed, the temperature of the air will rise due to compression heat generation. The higher the speed or the greater the torque, the more intense the compression, and the more significant the intake air temperature rise. Therefore, the engine physics model estimates the air temperature before entering the intercooler by analyzing the impact of speed and torque on the compression process. Secondly, the speed directly affects the cylinder charging frequency, and the torque reflects the engine's load demand. At higher speeds, the cylinder charging speed increases, and the intake air flow increases; while in high-torque conditions, the engine needs more air to support combustion, and the intake air flow also increases accordingly. The engine physics model calculates the amount of air entering the engine per unit time by considering the effects of speed and torque. At the same time, speed and torque determine the engine's combustion efficiency and friction loss. High speed or high torque usually accompanies stronger combustion and mechanical friction, resulting in more heat being absorbed by the cooling water. The engine physics model estimates the heat that needs to be transferred by the cooling system by analyzing the impact of these heat sources on the cooling water.

[0054] The intercooler reduces the intake air temperature through a heat dissipation medium such as air or water. The larger the intake air flow, the stronger the cooling capacity of the intercooler; while the higher the intake air temperature, the greater the cooling demand. The thermal management model can combine these parameters to simulate the cooling effect of the intercooler on the intake air, outputting the intake air temperature after intercooling.

[0055] The water route heat transfer reflects the heat generated by the engine, while the radiator efficiency determines the cooling system's ability to release heat. If the ambient temperature is high or the radiator efficiency is low, the cooling water temperature will rise; otherwise, it will drop. The thermal management model estimates the final temperature of the cooling water after absorbing the engine heat, i.e., the engine cooling water temperature, by analyzing these factors.

[0056] In one embodiment, the above step A3 can be implemented as steps A31-A32 as follows: In step A31, the engine control model calculates the EGR rate correction and the ignition angle correction under different temperature boundaries according to the temperature signal combined with the ambient temperature signal; In step A32, the correction coefficients of the parameterized mean model are determined according to the theoretical impact of EGR rate correction and ignition angle correction on fuel consumption.

[0057] In one embodiment, the above step A32 can be implemented as steps A321-A322 as follows: In step A321, a mathematical mapping relationship between the correction value and the fuel consumption change is established by analyzing the impact of EGR rate change on the composition of in-cylinder gas and combustion efficiency, and the impact of ignition angle change on combustion phase and indicated thermal efficiency; In step A322, the EGR rate correction coefficient and the ignition angle correction coefficient are coupled to generate a correction coefficient suitable for the current operating condition.

[0058] Figure 2 Fig. 1 is a schematic diagram of an oil consumption prediction device according to an embodiment of the present application. Figure 2 As shown in the figure, the device comprises: a construction module 201 configured to construct a co-simulation model composed of a vehicle driveline model, a thermal management thermal model, an engine physical model and an engine control model; a simplification module 202 configured to simplify the engine physical model from a detailed physical model to a parameterized mean model through neural network training based on bench constant temperature data; a generation module 203 configured to generate correction coefficients for the parameterized mean model according to the vehicle driveline model, the thermal management thermal model and the engine control model; a correction module 204 configured to correct the parameterized mean model according to the correction coefficients; an output module 205 configured to output engine instantaneous specific fuel consumption based on the corrected parameterized mean model; a calculation module 206 configured to calculate vehicle dynamic fuel consumption by integrating the vehicle driveline model and the vehicle driveline spectrum according to the engine instantaneous specific fuel consumption.

[0059] In an embodiment, the simplification module is further configured to: use engine steady state operating condition data at a bench constant temperature as a training set, use speed, intake manifold temperature / pressure, exhaust manifold temperature / pressure and EGR rate as input features, use volumetric efficiency, indicated mean effective pressure and exhaust temperature as output labels, optimize network weights through a back propagation algorithm, and generate a parameterized mapping relationship.

[0060] In an embodiment, the generation module comprises: a first determination sub-module configured to determine engine operating speed and torque according to the vehicle driveline model and the thermal management model; a second determination sub-module configured to determine temperature signals corresponding to the engine operating speed and torque through the engine physical model and the thermal management model; a generation sub-module configured to input the temperature signals into the engine control model to generate correction coefficients for the parameterized mean model.

[0061] In an embodiment, the first determination sub-module is further configured to: calculate driveline energy consumption according to the vehicle driveline model; calculate accessory power consumption according to the thermal management model; determine vehicle demand power according to the driveline energy consumption and the accessory power consumption; determine engine operating speed and torque according to the vehicle demand power.

[0062] In an embodiment, the second determining sub-module is further configured to: The engine physical model calculates the intercooler front intake air temperature, intake air flow and water heat transfer amount through the engine operating speed and torque; The thermal management model outputs a temperature signal including the post-intercooler intake air temperature and engine cooling water temperature through the intercooler front intake air temperature, intake air flow and water heat transfer amount, in combination with the vehicle thermal boundary condition.

[0063] In an embodiment, the generating sub-module is further configured to: The engine control model calculates the EGR rate correction and ignition angle correction under different temperature boundaries in combination with the ambient temperature signal according to the temperature signal; According to the theoretical influence of the EGR rate correction and ignition angle correction on fuel consumption, a correction coefficient of the parameterized mean model is determined.

[0064] In an embodiment, the correction coefficient of the parameterized mean model is determined according to the theoretical influence of the EGR rate correction and ignition angle correction on fuel consumption, including: By analyzing the influence of EGR rate change on in-cylinder gas composition and combustion efficiency, and the influence of ignition angle change on combustion phase and indicated thermal efficiency, a mathematical mapping relationship between the correction value and the fuel consumption change is established; The EGR rate correction coefficient and the ignition angle correction coefficient are coupled to generate a correction coefficient applicable to the current working condition.

[0065] Figure 3 A hardware structure diagram of an oil consumption prediction system in an embodiment of the present application is shown in FIG. 3. Figure 3 The oil consumption prediction system includes: at least one processor 320; and a memory 304 in communication with the at least one processor 320; wherein The memory 304 stores instructions executable by the at least one processor 320, and the instructions are executed by the at least one processor 320 to implement the oil consumption prediction method described in any of the above embodiments.

[0066] Referring to FIG. 3, Figure 3 The oil consumption prediction system 300 can include one or more of the following components: a processing component 302, a memory 304, a power supply component 306, an input / output (I / O) interface 308, a sensor component 310, and a communication component 312.

[0067] The processing component 302 generally controls the overall operation of the fuel consumption prediction system 300. The processing component 302 can include one or more processors 320 to execute instructions for performing all or a subset of the steps of the methods described above. Furthermore, the processing component 302 can include one or more modules to facilitate the interaction between the processing component 302 and other components. The processor 320 can be a central processing unit (CPU), a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic component, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. The general purpose processor can be a microprocessor but, in the alternative, the processor can be any conventional processor, controller, microcontroller, or state machine. A processor can also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.

[0068] The memory 304 is configured to store various types of data to support the operation of the fuel consumption prediction system 300. Examples of such data include instructions for any application or method operating on the fuel consumption prediction system 300. The memory 304 can be an internal storage unit of the terminal device, such as a hard disk or memory of the terminal device. The memory 304 can also be an external storage device of the terminal device, such as a plug-in hard disk equipped on the terminal device. The memory 304 can be realized by any type of volatile or nonvolatile storage devices, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable programmable read only memory (EPROM), programmable read only memory (PROM), read only memory (ROM), magnetic storage, flash memory, magnetic or optical disk. The memory 304 is used to store programs and data required by the present application. The memory 304 can also be used to temporarily store data that has been output or will be output.

[0069] The power supply component 306 supplies the various components of the fuel consumption prediction system 300 with power. The power supply component 306 can include a power supply management system, one or more power supplies, and other components associated with generating, managing and distributing power for the fuel consumption prediction system 300.

[0070] The I / O interface 308 provides an interface between the processing component 302 and peripheral interface modules, which can be a keyboard, a click wheel, buttons, and the like.

[0071] The sensor assembly 310 includes one or more sensors for providing status assessment of various aspects for the fuel consumption prediction system 300. In addition, the sensor assembly 310 can detect the on / off status of the fuel consumption prediction system 300, the relative positioning of the components, and can also detect the operational status of the fuel consumption prediction system 300 or one of the components of the fuel consumption prediction system 300. In some embodiments, the sensor assembly 310 can include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor, among others.

[0072] The communication assembly 312 is configured to enable the fuel consumption prediction system 300 to provide wired or wireless communication capability with other devices and cloud platforms. The fuel consumption prediction system 300 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an example embodiment, the communication assembly 316 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In an example embodiment, the communication assembly 316 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0073] In an example embodiment, the fuel consumption prediction system 300 can be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, micro-controllers, microprocessors, or other electronic components, for performing the fuel consumption prediction method as described in any of the embodiments above.

[0074] The present application also provides a computer readable storage medium, when instructions in the storage medium are executed by a processor corresponding to the fuel consumption prediction system, the fuel consumption prediction system is enabled to implement the fuel consumption prediction method as described in any of the embodiments above.

[0075] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk memory and optical memory, etc.) containing computer-usable program code.

[0076] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 one or more flowcharts and / or blocks

[0077] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 one or more flowcharts and / or blocks

[0078] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 one or more flowcharts and / or blocks

[0079] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A fuel consumption prediction method, characterized in that, Includes the following steps: Construct a joint simulation model consisting of a vehicle powertrain model, a thermal management model, an engine physical model, and an engine control model; Based on bench ambient temperature data, the engine physical model is simplified from a detailed physical model to a parameterized mean model through neural network training. Correction coefficients for the parameterized mean model are generated based on the vehicle drivetrain model, thermal management thermal model, and engine control model. The parameterized mean model is corrected according to the correction coefficient; The engine instantaneous specific fuel consumption is output based on the modified parameterized mean model; The dynamic fuel consumption of the vehicle is calculated based on the instantaneous specific fuel consumption of the engine and the road spectrum integral of the vehicle's transmission system.

2. The method as described in claim 1, characterized in that, The process of simplifying the detailed physical model of the engine into a parameterized mean model through neural network training based on bench ambient temperature data includes: Using engine steady-state operating data under normal temperature conditions on a test bench as the training set, and taking engine speed, intake manifold temperature / pressure, exhaust manifold temperature / pressure and EGR rate as input features, and volumetric efficiency, indicated effective pressure and exhaust temperature as output labels, the network weights are optimized through backpropagation algorithm to generate parameterized mapping relationships.

3. The method as described in claim 1, characterized in that, The step of generating correction coefficients for the parameterized mean model based on the vehicle drivetrain model, thermal management model, and engine control model includes: The engine operating speed and torque are determined based on the vehicle drivetrain model and thermal management model. The temperature signals corresponding to the engine's operating speed and torque are determined using the engine's physical model and thermal management model. The temperature signal is input into the engine control model to generate correction coefficients for the parameterized mean model.

4. The method as described in claim 3, characterized in that, The step of determining the engine operating speed and torque based on the vehicle drivetrain model includes: Calculate the energy consumption of the transmission system based on the vehicle transmission system model; Calculate the power consumption of the accessory based on the thermal management model; The required power of the whole vehicle is determined based on the energy consumption of the transmission system and the power consumption of the accessories. The engine operating speed and torque are determined based on the required power of the entire vehicle.

5. The method as described in claim 3, characterized in that, The step of determining the temperature signals corresponding to the engine's operating speed and torque through the engine physical model and thermal management model includes: The engine physical model calculates the inlet air temperature, inlet air flow and water circuit heat transfer of the intercooler by the engine's operating speed and torque. The thermal management model outputs a temperature signal by combining the inlet air temperature, inlet air flow and heat transfer through the intercooler with the vehicle's thermal boundary conditions. The temperature signal includes the inlet air temperature after the intercooler and the engine coolant temperature.

6. The method as described in claim 3, characterized in that, The step of inputting the temperature signal into the engine control model to generate correction coefficients for the parameterized mean model includes: The engine control model calculates the EGR rate correction and ignition angle correction under different temperature boundaries based on the temperature signal and the ambient temperature signal. Based on the theoretical impact of EGR rate correction and ignition angle correction on fuel consumption, the correction coefficients for the parameterized mean model are determined.

7. The method as described in claim 6, characterized in that, The determination of correction coefficients for the parameterized mean model based on the theoretical impact of EGR rate correction and ignition angle correction on fuel consumption includes: By analyzing the effects of EGR rate changes on in-cylinder gas composition and combustion efficiency, and the effects of ignition angle changes on combustion phase and indicated thermal efficiency, a mathematical mapping relationship between correction values ​​and fuel consumption changes is established. The EGR rate correction factor is coupled with the ignition angle correction factor to generate a correction factor applicable to the current operating conditions.

8. A fuel consumption prediction device, characterized in that, include: The building module is used to construct a joint simulation model consisting of a vehicle powertrain model, a thermal management model, an engine physical model, and an engine control model. A simplification module is used to simplify the engine physical model from a detailed physical model to a parameterized mean model based on bench ambient temperature data and through neural network training. The generation module is used to generate correction coefficients for the parameterized mean model based on the vehicle drivetrain model, thermal management thermal model, and engine control model. A correction module is used to correct the parameterized mean model according to the correction coefficient; The output module is used to output the engine's instantaneous specific fuel consumption based on the corrected parameterized mean model; The calculation module is used to calculate the dynamic fuel consumption of the vehicle based on the instantaneous specific fuel consumption of the engine and the road spectrum integral of the vehicle's transmission system.

9. A fuel consumption prediction system, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to implement the fuel consumption prediction method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor corresponding to the fuel consumption prediction system, the fuel consumption prediction system is able to implement the fuel consumption prediction method as described in any one of claims 1-7.

Images