Structure Principles and Measurement Methods of Various Models Based on the engine power simulation models ADVISOR, PSAT and EVSIM are fuel consumption models that are widely used in the automotive engineering field. These models use a driving cycle to simulate the operating status of the selected vehicle and the corresponding power flow. Complex vehicle, engine, emission control, and driving cycle parameters are required to calculate torque, power, vehicle traction, and fuel consumption. Take ADVISOR as an example: the model provides the driving trajectory that the car should meet by the driving cycle module, and requests the required speed from the vehicle module. The vehicle module calculates the wheel speed and force required to meet this speed request using the vehicle driving equation. Requires a request to the wheel and axle module to request the relay to pass through the backward path to the modules such as the main reducer, transmission, clutch, and mechanical load of the upper module, until the engine and the fuel conversion module calculate the actual power required by the engine and fuel.
The above model can calculate the fuel consumption of a particular vehicle under a specified driving cycle with very high accuracy, but the disadvantages are also very obvious, requiring a lot of input parameters and calculation processes, and can only be aimed at a specific vehicle. In order to evaluate the fuel consumption of a vehicle technology rather than a specific vehicle, Simpson designed the PAMVEC model and used three parameters (average speed, speed ratio, acceleration characteristic value) to characterize the vehicle's driving mode and predict its fuel consumption.
Models based on driving pattern decomposition This type of model classifies typical driving conditions that a motor vehicle can encounter under different conditions and measures the corresponding fuel consumption emissions according to each driving pattern. Akcelik et al. divided the three driving modes of idling, uniform speed, and acceleration and deceleration, and later developed four modes and applied it to the SIDRA model. Such models assume that the fuel consumption emissions of each model are independent of each other, and the total fuel consumption emissions are equal to the sum of fuel consumption emissions of each model. Because this type of model has a clear concept and a simple structure, and it is easy to establish an interface with the traffic model, the model is widely used. The basic formula of the model can be expressed by formula (1): F=f1Xsf2dsf1h(1) where F is the total fuel consumption or emissions of the driving zone, g; Xs is the distance of the driving zone, km; ds is the average delay of each vehicle s;h is the average number of stops per vehicle; f1 is the fuel consumption or emission factor when the vehicle is running at a constant speed, g/km; f2 is the fuel consumption or emission rate when the vehicle is idle, g/s; f3 is the number of times the vehicle is parked The resulting additional fuel consumption or emissions, g. In order to calculate more accurately, the model suggests that the user provide the initial speed and final speed of each acceleration and deceleration. If it is not available, the model will be replaced by the time and distance of the observation or by a full stop. With respect to parking, the model needs to define a minimum uniform speed and acceleration to distinguish between two different types of parking: parking during constant speed driving and parking during queuing.
This type of model is simple and easy to use, but the disadvantage is that the acceleration and deceleration of the model under the average driving state gives the fuel consumption emission parameters of the model, but it cannot distinguish the acceleration and deceleration due to different drivers and different traffic conditions. Subsequent tests have proved that different accelerating effects on fuel consumption emissions are very large, and the emissions and fuel consumption of short-distance travel are more affected by acceleration and deceleration, so the model has a poor prediction effect at short distances. Post et al.'s research shows that 4km is the minimum travel distance for this type of model application, and this model can explain the 90% fuel consumption change of 4km and above.
The most intuitive way to describe the driving state of a motor vehicle based on the speed-acceleration statistical model is to establish a velocity-acceleration matrix. Therefore, the most convenient micro-modeling method is to build a query table corresponding to the velocity-acceleration matrix, that is, to give the average fuel consumption and emission rate for each velocity-acceleration unit under the matrix according to experimental data. Andre et al. and Joumard et al. used the product of velocity and acceleration in place of the acceleration variable. The MODEM model is a representative model of this method. The model classifies the fuel consumption emission data according to “speed†and “velocity acceleration product†to predict the instantaneous Fuel consumption is based on "speed" and "speed acceleration product"
Combine for selection and calculation of values.
Cernuschi divided the data into five types of acceleration modes, representing high deceleration, low deceleration, uniform velocity, high acceleration, and low acceleration modes. He applied a regression method for each type of model to best fit the emission-velocity curve. Differentiate idle fuel consumption.
This method is actually a special case of the velocity-acceleration matrix method. The dimension of the acceleration vector is defined as 5, which reduces the data requirements for modeling and also reduces the amount of computation during model prediction. Ahn et al. did not directly use the velocity and acceleration variables to form a matrix look-up table, but did perform product combinations of different powers for speed and acceleration.
In the VT-Micro model established by Ahn et al., the test vehicles are first divided into several categories using the method of categorical regression trees, and then each type of vehicle averages the discharge data according to the speed and acceleration to form representative vehicle emission data. For each type of vehicle, the best fit is determined by using the product combination of speed and acceleration with different powers for each emission. The basic form of VT-Micro is shown in formula (2): (2) where MOEe is the instantaneous fuel consumption rate or emission rate, mg/s; Lei,j is the acceleration speed, and the power is “iâ€, and the acceleration power is The model regression coefficient at "j"; Mei, j is the model regression coefficient when the speed power is "i" at deceleration and the acceleration power is at "j"; s is the instantaneous speed of the vehicle, km/h; a is the vehicle Instantaneous acceleration, m/s2.
The problem based on the velocity-acceleration statistical model is above all the speed-acceleration resolution. In theory, the higher the resolution of the matrix, the higher the prediction accuracy. However, high matrix resolution also places very high demands on the basic data collection and model calculation efficiencies of various models.
Kenworthy et al. pointed out that it is unrealistic to construct a speed car energy-saving and environmental protection-acceleration matrix with a resolution of 0.1 km/h. Secondly, although this statistical model based purely on velocity and acceleration may have a good fitting effect on specific vehicle data, it does not consider the principle of fuel consumption and emissions of vehicles. Therefore, the versatility of the model needs to be further verified.
Power model based on power demand In order to overcome the shortcomings of statistical methods that cannot explain the principle of vehicle emissions, many studies have established models from the perspective of vehicle power requirements. This method draws heavily on the model method based on the engine power of automobiles, but it is different from the research on automotive engineering in the microcosmic degree of parameter selection. The concept of "work" demand in the earliest fuel consumption model appears in Watson et al.'s study in the 1980s. They added a variable to the kinetic energy of the vehicle based on the average fuel consumption model to reflect the acceleration of the vehicle. The demand for work is shown in equations (3) and (4).
Conclusion Based on the design needs of automobiles in terms of energy consumption and emissions, the micro fuel consumption and emissions model first appeared in the field of automotive engineering. Although the model parameters and the traffic parameters are far apart, these models are very useful for later model development in the transportation sector. Great reference value. In the field of transportation, research on fuel consumption emission models has been limited for a long time to statistical analysis of speeds and accelerations and the establishment of matrix look-up tables. The calculation efficiency is low and the physical relationship between fuel consumption emissions and traffic behavior cannot be explained. Represented by the CMEM and MOVES models, the model development method based on the vehicle power demand variables is currently the mainstream method. In particular, the vehicle specific power (VSP) variable, which is independent of the vehicle weight, not only retains the advantages of the physical model in the automotive engineering field, but also reduces the complexity of the model's in-depth parameters, and is suitable for linking with traffic parameters. It is a fuel consumption model. Research and application direction.
Nantong Tobel Intelligent Controlling Equipment CO;LTD , https://www.tobel-tech.com