KR100384282B1 - Multimode driving control method for parallel type hybrid vehicle - Google Patents

Abstract

The present invention relates to a multi-mode driving control method of a parallel hybrid vehicle, the method for improving the high fuel economy low pollution characteristics in a particular mode, such as the Federal Test Procedure-75 (FTP-75) mode of the United States was mainly used, but the slope In the case of hybrid vehicles aimed at high fuel efficiency and low pollution, the sensitivity of vehicle performance to driving mode and driver's habits is higher than that of the existing internal combustion engine. If the control method is not high enough, the fuel economy and exhaust gas results may not have the advantage over the internal combustion engine. Accordingly, the present invention has been made to solve the problems of the prior art, the first step of selecting a predetermined driving pattern and learning the driving patterns in the Hamming Network using the driving pattern characteristic parameters Wow; A second step of optimizing the running control parameters related to fuel economy and exhaust gas in each driving pattern through a predetermined virtual experiment when the first step is completed; A third step of storing driving patterns obtained through the first and second steps and driving control parameters associated with each driving pattern; A fourth step of measuring and storing predetermined data (velocity, acceleration, tilt signal, etc.) at a predetermined interval through a sensor during actual driving; Statistically processing the data stored in the fourth step at predetermined time intervals, extracting the driving pattern characteristic parameter, and inputting it to a neural network to select a driving pattern suitable for the current driving among the driving pantones learned in the first step. A fifth step; By providing a method comprising a sixth step of controlling the driving of the vehicle by selecting the driving control parameters according to the driving pattern selected in the fifth step, to ensure the performance of the actual driving conditions including the slope and the various driving modes In addition, as a result of simulating fuel economy and exhaust gas with a pattern made by arbitrarily combining three urban driving patterns among representative driving patterns, a hybrid vehicle using multi-mode driving control is compared with a hybrid vehicle using single-mode driving control. The fuel efficiency and emissions assessment index (J _ICV ) has been improved by 32%, compared with the existing internal combustion engine vehicles by 43%.

Description

MULTIMODE DRIVING CONTROL METHOD FOR PARALLEL TYPE HYBRID VEHICLE}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multi-mode driving control method for a parallel hybrid vehicle. In particular, a neural network using a driving pattern recognition algorithm for maximizing high fuel efficiency / low pollution characteristics of a parallel hybrid vehicle using a dual power supply system is provided. The present invention relates to a multi-mode driving control method of a parallel hybrid vehicle that recognizes a driving pattern in real time and converts the driving control method to be suitable for the driving pattern according to the driving pattern.

A major advantage of a dual power supply system, such as a hybrid vehicle in the past, is that it allows flexibility in power distribution between two power sources, which gives the flexibility to optimize the characteristics of the target vehicle's power train. To make it possible.

1 is an exemplary view showing the structure of a conventional hybrid hybrid vehicle equipped with an automatic transmission, and as shown in FIG. 1, an existing engine and a battery-powered electric motor are simultaneously used as power sources, and an automatic transmission using pneumatic transmission is shown. Was applied.

The Hybrid Electric Control Unit (HECU), which is connected to each element of the hybrid power train, uses a high-performance digital signal processor (DSP) chip to control pneumatic solenoid valves mounted on an automatic transmission in addition to inverter control. And directly control the engine load controller. Therefore, the HECU ultimately controls the vehicle as a whole for optimal performance and efficiency with respect to driving conditions and driver commands.

The fuel economy suggested by most automakers differs greatly from the actual driving performance, because vehicle performance, including fuel consumption and exhaust emissions, varies greatly depending on driving mode and driver's driving habits.

Therefore, it is essential to establish a control method suitable for each driving mode. In order to improve high fuel economy and low pollution in certain modes, such as the US-based Federal Test Procedure-75 (FTP-75) mode, which is commonly used by automobile companies, the performance is guaranteed under actual driving conditions including slopes and various driving modes. In the case of a hybrid vehicle for high fuel efficiency and low pollution, the vehicle performance is more sensitive to the driving mode and the driver's habits than the existing internal combustion engine, and the fuel consumption and exhaust gas results are low when the control method is inappropriate. Problems may arise that do not have advantages over organs.

Therefore, the present invention was devised to solve the above problems of the prior art, by selecting a driving pattern (for example, urban driving pattern, rural road driving pattern and highway driving pattern, etc.) having a representative driving environment. It is an object of the present invention to provide a method for recognizing a driving pattern in real time after learning to a neural network and converting the driving control method according to the driving pattern accordingly.

Figure 1 is an exemplary view showing a structure of a conventional hybrid hybrid vehicle equipped with an automatic transmission.

2 is an exemplary view showing an operation flow of the present invention.

3 is an exemplary view showing six representative driving patterns applied to the present invention.

4 is a table showing the definition of a driving pattern characteristic parameter applied to the present invention.

FIG. 5 is an exemplary diagram in which the characteristics of six representative driving patterns applied to the present invention are expressed by relative values of main driving pattern characteristic parameters. FIG.

6 is a table showing six representative driving patterns applied to the present invention as one embodiment of 24 driving pattern characteristic parameters.

7 is a structural diagram of a neural network used to recognize a driving pattern in the present invention.

8 is an exemplary view showing a vector conversion criteria by statistical analysis of six representative driving patterns in the present invention.

9 is an exemplary view showing a SOC control method in the present invention.

10 is a table showing the optimum driving control parameters extracted through the simulation of the six representative driving patterns in the present invention.

11 is an exemplary view showing a pattern made by arbitrarily combining three urban driving patterns of the representative driving pattern applied to the present invention.

12 is a table showing the fuel consumption and exhaust gas simulation results in FIG.

In order to achieve the above object, a multi-mode driving control method of a parallel hybrid vehicle according to the present invention includes selecting a predetermined driving pattern and then training the driving patterns to a neural network using a driving pattern characteristic parameter. Steps; A second step of optimizing the running control parameters related to fuel economy and exhaust gas in each driving pattern through a predetermined virtual experiment when the first step is completed; A third step of storing driving patterns obtained through the first and second steps and driving control parameters associated with each driving pattern; A fourth step of measuring and storing predetermined data (velocity, acceleration, tilt signal, etc.) at a predetermined interval through a sensor during actual driving; Statistically processing the data stored in the fourth step at predetermined time intervals, extracting the driving pattern characteristic parameter, and inputting it to a neural network to select a driving pattern suitable for the current driving among the driving pantones learned in the first step. A fifth step; And a sixth step of controlling driving of the vehicle by selecting a driving control parameter according to the driving pattern selected through the fifth step.

Hereinafter, an operation process of an embodiment according to the present invention will be described in detail with reference to the accompanying drawings.

FIG. 2 is an exemplary view illustrating an operation flow of the present invention. As shown in the drawing, first, a representative driving pattern including a city driving pattern, a local road driving pattern, and a highway driving pattern representative of a general driving environment is selected. As shown in FIG. 3, three urban driving patterns including two inclinations, two rural road driving patterns, and one highway driving pattern were selected and their relative characteristics are shown in FIG. 5. The characteristic parameter values of the six driving patterns as shown in FIG. 6 extracted by using the driving pattern characteristic parameters C1 to C24 defined in FIG. 4 are trained to a hamming network shown in FIG. 7. The parameter corresponding to is initially applied to the neural network by applying actual values such as 30 km / hr, 10 degrees of inclination, and acceleration of 0.5 m / s ^ 2 to the neural network. Since it may be distorted, in order to prevent this, the values of all parameters C1 to C24 are converted into vectors consisting of 1 and -1 based on a constant statistical process, and the constant statistical process is shown in FIG. As described above, for example, in the case of C1, since there are six representative driving patterns, six values are calculated. The average value of these values (m) and the standard deviation (std) are calculated, and then three values (m) are bounded. + α * std, m, m-α * std) {1,1,1}, {1,1, -1}, {1, -1, -1}, {-1, -1, -1} depending on where the actual value falls Grant. In other words, if it is larger than m and smaller than m + α * std, {1,1, -1} is allocated.

Then, the driving control parameters related to fuel economy and exhaust gas are optimized through simulation or dynamo experiment in each driving pattern. The driving control parameters include weights related to power distribution of engine and motor (fuel consumption, exhaust gas (NOx, CO)). , HC), motor efficiency and engine power ), A method of controlling a battery remaining state (hereinafter referred to as 'SOC'), an SOC usage section, an engine oscillation speed and a shifting algorithm, and the like. In the control method, as shown in FIG. 9, after the SOC becomes 0.5 or less, the control1 performs charging using the engine power, and the battery is frequently charged using the remaining engine power after satisfying the vehicle required driving power. There is a way of Control2.

However, as a double example, the results shown in FIG. 10 are obtained through simulations using weights and SOC control method related to power distribution of the engine and the motor as driving control parameters.

Thereafter, the driving pattern obtained above and driving control parameters related to each driving pattern are stored in a vehicle driving controller (not shown), and then a speed, acceleration, and inclination signal are set at predetermined intervals through a sensor (not shown) during actual driving. For example, 1 second) is measured and stored. In the weight matrix 1 of the neural network shown in FIG. 7 mounted on the driving controller, characteristic parameters C1 to C24 of six representative driving patterns are converted into 1 and -1. It is stored in the form of a vector, and the signal from the sensor is also converted into 1 and -1 to perform the dot product with the weight matrix 1, and the dot product results in the vector with the six elements, the dot product in the recurrent layer. Find this biggest factor through competition. The weight matrix 1 is a 6 * 72 matrix, where 6 represents the number of representative driving patterns, and 72 is a number obtained by multiplying 24 characteristic parameters by 3 elements consisting of 1 and -1.

After extracting the driving pattern characteristic parameter by statistically processing the stored signal at predetermined time intervals (for example, 300 seconds), the driving pattern suitable for the current driving is selected from the driving patterns inputted to the neural network. The operation is repeatedly performed, and the driving control parameter is selected according to the selected driving pattern to control the driving of the vehicle.

In addition, not only the parallel hybrid vehicle with the automatic transmission shown in FIG. 1 but also the parallel hybrid vehicle with the continuously variable transmission, in this case, as a driving control parameter, factors related to battery charge and discharge amount, factors related to transmission ratio control, At the initial start of the vehicle, factors that determine the motor-driven section are used.

As described above, the multi-mode driving control method of the parallel hybrid vehicle according to the present invention can guarantee the actual driving conditions including the inclination and the performance in various driving modes, and as shown in FIG. As a result of performing fuel economy and exhaust gas simulation with a pattern made by arbitrarily combining three urban driving patterns, as shown in FIG. 12, the hybrid vehicle (HEV_MM) using multi-mode driving control uses a single mode driving control as shown in FIG. 12. Compared with (HEV_SM), the fuel efficiency and exhaust gas evaluation index (J _ICV ) value is improved by 32%, and compared with the existing internal combustion engine vehicle (ICV) by 43%.

Claims (4)

Selecting a predetermined representative driving pattern and learning the representative driving patterns in a Hamming Network using a driving pattern characteristic parameter; A second step of optimizing the running control parameters related to fuel economy and exhaust gas in each representative driving pattern through a predetermined virtual experiment when the first step is completed; A third step of storing the representative driving patterns obtained through the first and second steps and a driving control parameter associated with each representative driving pattern; A fourth step of measuring and storing predetermined data (velocity, acceleration, tilt signal, etc.) at a predetermined interval through a sensor during actual driving; Representatives learned in the first step by statistically processing data measured and stored at a predetermined interval at predetermined intervals, extracting a driving pattern characteristic parameter for a predetermined time before statistical processing, and inputting the same into a neural network. Selecting a representative driving pattern most suitable for the extracted driving pattern characteristic parameter among driving patterns; And a sixth step of controlling driving of the vehicle by selecting a driving control parameter according to the representative driving pattern selected through the fifth step. The parallel hybrid of claim 1, wherein the predetermined representative driving pattern of the first step comprises a predetermined urban driving pattern including a slope, a predetermined rural road driving pattern, and a predetermined highway driving pattern. Multi-mode driving control method of the vehicle. According to claim 1, The driving control parameters of the second step is a value of 1 and -1 by the criteria by a constant statistical processing in order to prevent the driving performance pattern recognition is distorted by the value of one large parameter Multi-mode driving control method of a parallel hybrid vehicle, characterized in that converted to a vector consisting of. The driving control parameter of claim 1, wherein the driving control parameter is a factor related to a battery charge amount and discharge amount, a factor related to a shift ratio control, and a section that operates only as a motor at an initial start of the vehicle in the case of a parallel hybrid vehicle having a continuously variable transmission. Multimode driving control method of a parallel hybrid vehicle further comprising.