KR20130043288A - Control system and control method for engine exhaust gas recirculation - Google Patents

Abstract

PURPOSE: An exhaust gas recirculation control system for an engine and a control method thereof are provided to rapidly and precisely obtain air amount controlling performance and to reduce time and human resources for development of a system. CONSTITUTION: An exhaust gas recirculation control system comprises a diesel engine(110), an EGR(Exhaust Gas Recirculation) valve, a MAF(Mass Air Flow) sensor, an identifier(140), and a controller(150). The EGR valve controls the amount of exhaust gas and is installed at a pipe line connecting an exhaust manifold and an intake manifold. The MAF sensor is installed at the pipe line extended from the intake manifold and measures air volume flowing into the diesel engine. The identifier estimates the air volume flowing into the diesel engine. The controller calculates an optimal opening degree of the EGR valve to follow targeted air volume. [Reference numerals] (110) Diesel engine; (140) Identifier; (150) Controller; (AA) Input/output sensitivity; (BB) Estimated air volume; (CC) Estimated air volume error; (DD) Engine rotation speed, fuel injection quantity; (EE) Target air volume; (FF) EGR value opening degree; (GG) Air volume; (HH) Air volume control error;

Description

CONTROL SYSTEM AND CONTROL METHOD FOR ENGINE EXHAUST GAS RECIRCULATION}

The present invention relates to an engine exhaust gas recirculation control system and a control method, and more particularly, to a structure and a method for controlling an optimum EGR valve opening amount for tracking a target air amount in an exhaust gas recirculation control system. .

BACKGROUND Exhaust Gas Recirculation (EGR) systems are commonly used to reduce emissions of nitrogen oxides harmful to humans and the environment in automotive engines, particularly diesel engines. Precise control is essential because the amount of exhaust gas recycled in such an exhaust gas recirculation control system is directly related to the generation of harmful exhaust gas and engine performance.

In the conventional exhaust gas recirculation control system, a linear controller based on a lookup table is applied to control the opening amount of the EGR valve that determines the amount of exhaust gas recycled to the engine, thereby controlling the amount of air flowing into the engine. I'm in control.

However, in the conventional exhaust gas recirculation control system, since the lookup table is optimized based on the engine characteristics obtained under the steady-state operating conditions, it is difficult to control the air volume quickly and precisely in the transient operating conditions. In addition, there is a problem in that it takes a lot of time with a skilled expert because it has to derive an optimized lookup table through a number of engine experiments.

An object of the present invention is to ensure fast and precise air volume control performance under normal operating conditions as well as transient operating conditions, and to recirculate engine exhaust gas, which can save the labor and time of experts for system development and construction. It is to provide a control system and a control method.

The object is, according to the present invention, an engine exhaust gas recirculation control system for introducing a portion of the exhaust gas from an exhaust manifold of an engine into the engine through an intake manifold of the engine, wherein the exhaust manifold and the intake manifold are An EGR valve installed on a conduit connecting the folds to adjust an amount of the exhaust gas flowing into the engine; MAF sensor for measuring the amount of air flowing into the engine from the outside; An identifier for estimating the air amount using a first artificial neural network in which the current EGR valve opening amount is set as an input variable and the amount of air flowing into the engine is set as an output variable; And a controller for calculating an optimum EGR valve opening amount for following the target air amount using a second artificial neural network in which a target air amount is set as an input variable and the EGR valve opening amount is set as an output variable. Engine exhaust gas recirculation control system.

The identifier may compare the estimated value of the air amount with the measured value of the air amount by the MAF sensor and adjust the weight of the first neural network in a direction in which the error is reduced.

The controller may adjust the weight of the second artificial neural network in a direction in which the error is reduced by comparing the value of the target air amount with the measured value of the air amount by the MAF sensor.

The identifier calculates an input / output sensitivity in the process of estimating the air amount, and the controller is configured to optimize the EGR valve opening amount for following the target air amount based on the information on the input / output sensitivity calculated by the identifier. Can be calculated.

The input variable of the first artificial neural network in the identifier may further include an air amount measured by the MAF sensor.

The input variable of the second artificial neural network in the controller may further include the current EGR valve opening amount.

Each of the input variables of the first artificial neural network and the second artificial neural network may further include at least one of a rotation speed of the engine, a fuel injection amount, and an intake manifold pressure.

The object is, in an engine exhaust gas recirculation system for introducing a portion of the exhaust gas from an exhaust manifold of an engine into the engine via an intake manifold of the engine, according to the present invention. A method for controlling an EGR valve for adjusting an amount, the method comprising: (a) measuring an amount of air flowing into the engine using a MAF sensor; (b) estimating the air amount using a first artificial neural network in which the current EGR valve opening amount is set as an input variable and the amount of air flowing into the engine is set as an output variable; And (c) calculating an optimum EGR valve opening amount for following the target air amount using a second artificial neural network in which a target air amount is set as an input variable and the EGR valve opening amount is set as an output variable. It can be achieved by the engine exhaust gas recirculation control method characterized in that.

The step (b) may include comparing the estimated value of the air amount with the measured value of the air amount by the MAF sensor and adjusting the weight of the first artificial neural network in a direction in which the error is reduced.

The step (c) may include adjusting a weight of the second artificial neural network in a direction in which an error thereof is reduced by comparing the value of the target air amount with the measured value of the air amount by the MAF sensor.

The step (b) includes calculating an input / output sensitivity in the process of estimating the air amount, and the step (c) includes the target based on the information on the input / output sensitivity calculated in the step (b). And calculating the optimal opening amount of the EGR valve for following the air amount.

The above object can be achieved by a computer-readable recording medium having recorded thereon a program for performing the above-described engine exhaust gas recirculation control method according to the present invention.

The present invention, in the engine exhaust gas recirculation control system calculates the optimal EGR valve opening amount for tracking the target air volume by using an identifier and an artificial neural network to estimate the amount of air flowing into the engine using the artificial neural network By controlling the amount of air flowing into the engine, including a controller for controlling the EGR valve based on this, a conventional lookup table based on the optimized lookup table based on the engine characteristics obtained in the steady-state operating conditions Unlike the system, it is possible to obtain fast and precise air flow control performance even under excessive operating conditions, and it is not necessary to derive an optimized lookup table through numerous engine experiments, thereby reducing the effort and time required for the professional development and construction of the system. You can save.

1 is a schematic configuration diagram of an engine exhaust gas recirculation control system according to an embodiment of the present invention.
FIG. 2 is a control block diagram of the engine exhaust gas recirculation control system of FIG. 1.
FIG. 3 is a schematic diagram for describing input and output variables of an identifier artificial neural network in the engine exhaust gas recirculation control system of FIG. 1.
4 is a schematic diagram illustrating input and output variables of a controller artificial neural network in the engine exhaust gas recirculation control system of FIG. 1.
5 is a schematic flowchart illustrating a method for controlling engine exhaust gas recirculation according to an embodiment of the present invention.

In order to fully understand the present invention, operational advantages of the present invention, and objects achieved by the practice of the present invention, reference should be made to the accompanying drawings and the accompanying drawings which illustrate preferred embodiments of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, well-known functions or constructions are not described in order to avoid unnecessary obscuration of the present invention.

1 is a schematic configuration diagram of an engine exhaust gas recirculation control system according to an embodiment of the present invention, and FIG. 2 is a control configuration diagram of the engine exhaust gas recirculation control system of FIG. 1.

1 and 2, the exhaust gas recirculation control system (EGR System) according to the present embodiment is a diesel engine 110 in order to reduce the emission of nitrogen oxides harmful to humans and the environment. An exhaust gas recirculation control system for introducing some of the exhaust gas from the exhaust manifold 113 into the diesel engine 110 through the intake manifold 111 of the diesel engine 110. The diesel engine 110 and the EGR valve 120, MAF sensor 130, identifier 140, and controller 150.

On the other hand, the present invention will be mainly applied to the diesel engine 110 is a relatively large amount of nitrogen oxides, but is not limited to this can be sufficiently applied to the exhaust gas recirculation control system of other internal combustion engine (engine), including gasoline engine. . For reference, the diesel engine 110 is a compression ignition internal combustion engine used as a power source of an automobile. The fuel supplied from the fuel tank 171 after compressing the air sucked through the intake manifold 111 to a high pressure in a cylinder is used. Sprayed to the compressed air to cause an explosion and is used as a driving force for driving the axle 175 of the vehicle.

Exhaust Gas Recirculation Valve (EGR) valve 120 is a conventional means for controlling the amount of exhaust gas flowing into or recirculating the diesel engine 110 as shown in FIG. 1. It may be installed on a conduit connecting the fold 113 and the intake manifold of the diesel engine 110. When the EGR valve 120 is opened, some of the exhaust gas discharged from the exhaust manifold 113 of the diesel engine passes through the EGR cooler 125, as shown in FIG. 1, and then the intake manifold 111. It may be introduced into the diesel engine 110 through. At this time, the amount of exhaust gas recycled to the diesel engine 110 is determined according to the opening degree of the EGR valve 120, that is, the opening amount of the EGR valve. In addition, in the diesel engine 110 to which the exhaust gas recirculation control system is applied, the amount of air required for the current operating conditions (the amount of air introduced into the diesel engine 110 from the outside) is controlled by adjusting the opening amount of the EGR valve. Meanwhile, in the EGR cooler 125, the exhaust gas discharged from the exhaust manifold 113 at a very high temperature may be cooled to a predetermined temperature by the engine coolant.

The MAF sensor 130 (mass air flow sensor) is a conventional means for measuring the amount of air (air amount) flowing into the diesel engine 110 from the outside, as shown in Figure 1, the intake of the diesel engine 110 It may be installed on a conduit extending from the manifold 111. On the other hand, in this embodiment, in order to increase the output of the diesel engine 110, the air passing through the MAF sensor 130 is compressed by the compressor (163, Compressor) to the diesel engine 110 through the intake manifold (111). In this case, the rotary shaft 162 of the compressor 163 may be connected to the turbine 161 using the energy of the exhaust gas. That is, in this embodiment, a turbo charger system for increasing the output of the diesel engine 110 may be applied. Specifically, some of the exhaust gas discharged from the exhaust manifold 113 of the diesel engine 110 is recycled to the diesel engine 110 through the intake manifold 111 and the remaining exhaust gas drives the compressor 163. After passing through the turbine 161 is discharged to the outside. On the other hand, since the air compressed by the compressor 163 rises due to adiabatic compression, it is cooled through the inter cooler 165 (Inter Cooler) as shown in FIG. 1 and then flows into the diesel engine 110. desirable.

FIG. 3 is a schematic diagram for describing input and output variables of an identifier artificial neural network in the engine exhaust gas recirculation control system of FIG. 1, and FIG. 4 is a schematic diagram for describing input and output parameters of a controller artificial neural network in the engine exhaust gas recycling control system of FIG. 1. to be.

1 to 4, the identifier 140 estimates the amount of air flowing into the diesel engine 110 (air amount), and the controller 150 determines an optimal EGR for tracking the target air amount. To calculate the valve opening amount, both the identifier 140 and the controller 150 are based on an artificial neural network (ANN). In other words, the exhaust gas recirculation control system according to the present invention employs an indirect adaptive control structure based on an artificial neural network, and identifies the system 140 and identification information of the system that largely simulates the dynamic characteristics of the system. The controller 150 may calculate an optimal EGR valve opening amount based on the controller 150. To this end, the identifier 140 calculates the input / output sensitivity for the control system in estimating the air volume, and the controller 150 follows the target air volume based on the information on the input / output sensitivity calculated by the identifier 140. The optimum opening amount of the EGR valve can be calculated. The identifier 140 and the controller 150 may be a microprocessor, a microcontroller, or the like, and may be implemented in the form of software or firmware including an artificial neural network algorithm. . Hereinafter, for convenience of explanation, the artificial neural network of the identifier 140 is referred to as a 'first artificial neural network', and the artificial neural network of the controller 150 is referred to as a 'second artificial neural network'.

In general, an artificial neural network is composed of an input layer, a hidden layer, and an output layer, as shown in FIGS. 3 and 4, and each layer includes a plurality of neurons. At this time, the neurons of the input layer and the neurons of the hidden layer are interconnected, the neurons of the hidden layer and the neurons of the output layer are interconnected, so that the output of each layer is transferred to the input of the next layer. In this case, the hidden layer is located between the input layer and the output layer and receives the output of the input layer and delivers it to the input of the output layer. The hidden layer may be composed of one or a plurality of layers. Correlated and a predetermined correlation is formed between the input of the input layer and the output of the output layer by weight. That is, the correlation between the input of the input layer and the output of the output layer is adjusted by modifying the weight. This weight is initially set to an arbitrary value, and is modified in a direction in which an error between the output value of the artificial neural network and the predetermined target value decreases through an iterative learning process. In other words, the 'learning process of artificial neural network' refers to a process of finding weights optimized for a system to which an artificial neural network is applied. In this case, the weight may include a first weight that determines a correlation between the hidden layer and the output layer, and a second weight that determines a correlation between the input layer and the hidden layer. On the other hand, the learning process of the artificial neural network, that is, the process of adjusting the weight suitable for exhibiting the maximum control performance in the system is already well known in the art, a detailed description thereof will be omitted herein.

Specifically, the identifier 140 includes a first artificial neural network in which the current EGR valve opening amount is set as an input variable and the amount of air flowing into the diesel engine 110 is set as an output variable as shown in FIGS. 2 and 3. Using this, the amount of air can be estimated. That is, the input variable of the first artificial neural network in the identifier 140 may include the current EGR valve opening amount, and the output variable of the first artificial neural network may include the amount of air flowing into the diesel engine 110. At this time, the identifier 140 compares the estimated value of the air amount with the measured value of the air amount by the MAF sensor 130 through the artificial neural network learning process, that is, the first artificial neural network in a direction in which the error of the identifier 140 decreases. It can be optimized by adjusting the weight of.

Similarly, the controller 150 includes a second artificial neural network in which the target air amount is set as an input variable and the EGR valve opening amount is set as an output variable as shown in FIGS. 2 and 4, and is optimal for following the target air amount using the same. The EGR valve opening amount can be calculated. That is, in the controller 150, the input variable of the second artificial neural network may include a target air amount, and the output variable of the second artificial neural network may include an EGR valve opening amount. At this time, the controller 150 compares the value of the target air amount with the measured value of the air amount by the MAF sensor 130 through the artificial neural network learning process, and thus, the second artificial mantle in a direction in which the control error of the controller 150 is reduced. It can be optimized by adjusting the weight of the neural network. For reference, the target air volume is the air volume required for the best performance under the current driving conditions and is automatically set by analyzing various factors related to the engine in a typical vehicle.

In general, since the input / output variables of the neural network have a great influence on the control performance, it is very important to select an appropriate input variable of the neural network in order to obtain excellent control performance. In this aspect, in the present embodiment, the static and dynamic characteristics of the exhaust gas recirculation control system or the air control system of the diesel engine 110 have been analyzed through many experiments and simulations. Similarly, input variables suitable for the first artificial neural network of the identifier 130 and the second artificial neural network of the controller 150 were derived.

According to this, the input variable of the first artificial neural network in the identifier 140 is based on the current EGR valve opening amount directly affecting the air amount and the air amount measured by the current MAF sensor 130 as the main input variables. It is preferable to use the rotational speed of the engine, the fuel injection amount, and the intake manifold pressure which indirectly affect the dynamic characteristics of the air flow path. Similarly, in the controller 150, the input variables of the second artificial neural network are the main input variables of the target air amount and the current EGR valve opening amount, and in addition, the rotational speed and fuel of the engine indirectly affecting the dynamic characteristics of the air flow path. It is preferable to use the injection amount and the intake manifold pressure.

In other words, the input variable of the first artificial neural network in the identifier 140 preferably further includes the amount of air measured by the MAF sensor 130 as well as the current EGR valve opening amount as shown in FIGS. 2 and 3. More preferably engine rotational speed, fuel injection amount and intake manifold pressure. Similarly, the input variable of the second artificial neural network in the controller 150 preferably further includes not only the target air amount but also the current EGR valve opening amount, as shown in FIGS. 2 and 4, and more preferably the engine rotation. Speed, fuel injection amount and intake manifold pressure.

For reference, the input variables of the identifier 140 and the controller 150 form a vector by collecting the current measured value and the previously measured N values through a Tapped Delay Line (TDL). It is preferable in view of improving the accuracy of system control after being processed into a structure to be input to the artificial neural network.

As described above, the engine exhaust gas recirculation control system according to the present invention is an optimal EGR for following the target air amount using the artificial neural network and the identifier 140 for estimating the amount of air flowing into the engine using the artificial neural network. By calculating the valve opening amount and controlling the amount of air flowing into the engine, including a controller 150 for controlling the EGR valve 120 based on the valve opening amount, the lookup table optimized based on the engine characteristics obtained under the normal operating conditions ( Unlike the conventional system based on the lookup table, it is possible to ensure fast and precise air volume control performance even during operation in the transient section. That is, the engine exhaust gas recirculation control system according to the present invention can significantly improve the performance of air volume control under transient driving conditions by replacing the conventional lookup table based control algorithm with the artificial neural network based control algorithm.

In addition, since the engine exhaust gas recirculation control system according to the present invention does not need to derive an optimized lookup table through numerous engine experiments, it is possible to save the labor and time of a professional manpower for system development and construction.

In addition, in the engine exhaust gas recirculation control system according to the present invention, since the tracking performance of the target air amount is improved compared to the conventional system, unnecessary movement of the EGR valve 120 is significantly reduced, and thus the EGR valve 120 and It is possible to improve the life of the actuator for driving it.

5 is a schematic flowchart illustrating a method for controlling engine exhaust gas recirculation according to an embodiment of the present invention.

Hereinafter, the engine exhaust gas recirculation control method according to the present invention will be described in detail with reference to FIGS. 1 to 5. However, descriptions overlapping with the above-described engine exhaust gas recirculation control system will be omitted.

1 to 5, in the engine exhaust gas recirculation control method according to the present invention, a part of the exhaust gas emitted from the exhaust manifold 113 of the diesel engine 110 is transferred to the intake manifold of the diesel engine 110 ( In the engine exhaust gas recirculation control system as described above, which flows into the diesel engine 110 through 111, as a method for controlling the EGR valve 120 for adjusting the amount of exhaust gas flowing into the diesel engine 110. The air amount measuring step S110, the air amount estimating step S120, and an optimum EGR valve opening amount calculating step S130 may be included.

In the air amount measuring step S110, the air amount flowing into the diesel engine 110 is measured using the MAF sensor 130. In this case, the MAF sensor 130 may be installed on a conduit extending from the intake manifold 111 of the diesel engine 110.

In the air amount estimating step (S120), the air amount flowing into the diesel engine 110 is estimated using the first artificial neural network. In this case, the input variable of the first artificial neural network may be set to the current EGR valve opening amount, and the output variable of the first artificial neural network may be set to the amount of air flowing into the diesel engine 110. In addition, the input variable of the first artificial neural network may further include an air amount measured by the MAF sensor 130, and may further include an engine rotation speed, a fuel injection amount, an intake manifold pressure, and the like. Meanwhile, the air amount estimating step S120 may include comparing the estimated value of the air amount with the measured value of the air amount by the MAF sensor 130 and adjusting the weight of the first artificial neural network in a direction in which the error is reduced.

In the optimum EGR valve opening amount calculation step (S130), an optimal EGR valve opening amount for following the target air amount is calculated using the second artificial neural network. In this case, the input variable of the second artificial neural network may be set to a target air amount, and the output variable of the second artificial neural network may be set to an EGR valve opening amount. In addition, the input variable of the second artificial neural network may further include the current EGR valve opening amount, and may further include the rotational speed of the engine, the fuel injection amount and the intake manifold pressure. Here, the target air amount is the amount of air necessary to exhibit the best performance under the current driving conditions and is automatically set by analyzing various factors related to the engine in a normal vehicle. On the other hand, the optimum EGR valve opening amount calculation step (S130) is a step of adjusting the weight of the second artificial neural network in a direction in which the error is reduced by comparing the value of the target air amount and the measured value of the air amount by the MAF sensor 130 It may include.

In addition, the air amount estimating step (S120) includes calculating an input / output sensitivity in the process of estimating the air amount, and the optimal EGR valve opening amount calculating step (S130) includes an input / output sensitivity calculated in the air amount estimating step (S120). The optimal opening amount of the EGR valve for tracking the target air amount can be calculated based on the information.

On the other hand, the engine exhaust gas recirculation control method according to the present invention as described above, can be implemented as computer-readable code on a computer-readable recording medium. That is, the present invention can be implemented as a program for performing the above-described engine exhaust gas recirculation control method, which can be recorded on a computer-readable recording medium. Computer-readable recording media can include any type of recording device that stores data that can be read by a computer system. And functional programs, codes and code segments for implementing the present invention can be easily inferred by programmers in the art to which the present invention belongs.

For example, a program for performing the engine exhaust gas recirculation control method according to the present invention and a computer-readable recording medium recording the same may be implemented in the form of one module constituting an electronic control unit (ECU) of a vehicle. .

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit and scope of the invention. Accordingly, such modifications or variations are intended to fall within the scope of the appended claims.

110: diesel engine
111: intake manifold
113: exhaust manifold
120: Exhaust Gas Recirculation Valve
125: Exhaust Gas Recirculation Cooler
130: MAF sensor (mass air flow sensor)
140: identifier
150: controller
161: Turbine
163: Compressor
165: Inter Cooler
171: fuel tank
175 axle

Claims (8)

An engine exhaust gas recirculation control system for introducing a portion of the exhaust gas from the engine exhaust manifold into the engine through the engine intake manifold,
An EGR valve installed on a conduit connecting the exhaust manifold and the intake manifold to adjust the amount of the exhaust gas flowing into the engine;
MAF sensor for measuring the amount of air flowing into the engine from the outside;
An identifier for estimating the air amount using a first artificial neural network in which the current EGR valve opening amount is set as an input variable and the amount of air flowing into the engine is set as an output variable; And
And a controller configured to calculate an optimal amount of the EGR valve opening amount for following the target air amount by using a second artificial neural network in which a target air amount is set as an input variable and the EGR valve opening amount is set as an output variable. Engine exhaust gas recirculation control system.
The method of claim 1, wherein
The identifier,
And comparing the estimated value of the air amount with the measured value of the air amount by the MAF sensor and adjusting the weight of the first artificial neural network in a direction in which the error is reduced.
The method of claim 1, wherein
The controller,
And the weight of the second artificial neural network is adjusted in a direction in which the error is reduced by comparing the target air amount with the measured value of the air amount by the MAF sensor.
The method of claim 1, wherein
The identifier calculates an input / output sensitivity in the process of estimating the amount of air
And the controller calculates an optimum amount of the EGR valve opening degree for following the target air amount based on the information on the input / output sensitivity calculated by the identifier.
5. The method according to any one of claims 1 to 4,
The input variable of the first artificial neural network in the identifier,
Engine exhaust gas recirculation control system further comprises the amount of air measured by the MAF sensor.
The method of claim 5,
The input variable of the second artificial neural network in the controller,
And the present EGR valve opening amount.
The method according to claim 6,
Each of the input variables of the first artificial neural network and the second artificial neural network,
And at least one of a rotational speed of the engine, a fuel injection amount, and an intake manifold pressure.
In an engine exhaust gas recirculation system in which some of the exhaust gas from the engine's exhaust manifold is introduced into the engine through the engine's intake manifold, controlling an EGR valve that controls the amount of the exhaust gas flowing into the engine. For the way,
(a) measuring the amount of air flowing into the engine using a MAF sensor;
(b) estimating the air amount using a first artificial neural network in which the current EGR valve opening amount is set as an input variable and the amount of air flowing into the engine is set as an output variable; And
(c) calculating an optimum EGR valve opening amount for following the target air amount using a second artificial neural network in which a target air amount is set as an input variable and the EGR valve opening amount is set as an output variable; Engine exhaust gas recirculation control method characterized in that.

Images