FR2901840A1 - Exhaust gas post-processing device e.g. particle filter, control system for e.g. passenger car`s diesel engine, has management module modifying device operating procedure and coupled with estimator providing estimated gas temperature signal - Google Patents

Abstract

The system has a management module (20) that modifies an exhaust gas post-processing device operating procedure e.g. diagnosis procedure. The module is coupled with a neural network estimator (15) that provides an estimated temperature signal (S) of exhaust gas of an internal combustion engine e.g. diesel engine, of a motor vehicle e.g. passenger car. The estimator has a pre-processing module, a neural network and a post-processing module. The estimator comprises a feedback loop that feeds back output variables of the network, to an input of the network. An independent claim is also included for a method for controlling an exhaust gas post-processing device.

Description

The field of the invention is that of the depollution of a motor with

  internal combustion, in particular through the control of the after-treatment devices mounted downstream of the exhaust manifold of the engine. The invention more specifically relates to a method and a system for controlling an exhaust gas after-treatment device of an engine, as well as an internal combustion engine and a vehicle equipped with such a system. . In order to meet the emission of gaseous pollutants from motor vehicles, gas after-treatment devices are generally disposed in the exhaust line of the engines. These devices are designed to reduce emissions of carbon monoxide, unburnt hydrocarbons, particulate matter and nitrogen oxides. These devices operate discontinuously or alternately in that they alternate pollutant storage phases in traps and conversion phases of pollutants stored in non-polluting substances by regeneration traps. In order to be regenerated, the traps require specific modes of combustion guaranteeing adequate levels of heat and / or richness. Optimizing the treatment of all pollutants thus requires better control of the storage and regeneration phases. In particular, it is necessary to estimate over time the trapped masses (ie the particles in the case of the particle filter, the nitrogen oxides in the case of the nitrogen oxide trap). Similarly, it is necessary to know the evolution over time of the masses converted during the regeneration phases. However, the evolution of these masses during the storage and regeneration phases depends directly on the temperature of the support of these traps and the gases passing through them. We therefore seek to know, if not to control, the temperature of the gases that enter these traps.

  Moreover, the increase in the complexity of the engines and their modes of operation requires more and more sophisticated electronic management means and, consequently, more and more measurement or estimation means.

  Currently, the various control procedures of a post-processing device are controlled by means of a management module that uses exhaust gas temperature estimates determined either with the aid of physical sensors or using a model for estimating this temperature implemented by an onboard electronic calculator. But several disadvantages are related to the use of a sensor placed in the exhaust line. The accuracy of a sensor is indeed inversely proportional to the field of use. That is, the more it is desired to measure the temperature over a wide range of use, the lower the measurement accuracy. This accuracy can also derive with the thermal aging or fouling of the sensor. In addition, the cost of using a sensor can be important. It is actually necessary to associate with the intrinsic cost of the sensor that of the 20 connectors, the input port in the computer and the software driver. Moreover, it is also necessary to have suitable means to diagnose the operating state of the sensor. The use of models also has disadvantages. If these models are generally true in steady state gas, they are indeed rather mediocre transient. In addition, many parameters (physical parameters, state variables) necessary for these models are difficult to identify or to measure on the motor. The efficiency of the control procedures of a post-treatment device is then reduced in the dynamic thermal phases.

  It is therefore desirable to improve the control procedures of the post-processing devices by having a control of these procedures which does not have these drawbacks of the state of the art.

  The invention aims to meet this need, and proposes for this purpose, according to a first aspect, a control system of an exhaust gas after-treatment device of an engine, comprising means for modifying at least one control procedure of said device, characterized in that said means are coupled with a neural network estimator adapted to provide an exhaust gas temperature estimate. Some preferred but non-limiting aspects of the system according to the invention are as follows: the neural network is a feedback loop recurrent network; It is arranged in such a way that the estimator has, at the input, information relating to the temperature of the gases upstream of the post-treatment device disposed on the exhaust line of the engine, and so that the estimator provides in outputting the estimate of the exhaust gas temperature upstream, in or downstream of said post-treatment device; the estimator comprises a post-processing module of the temperature estimate made by the neural network; the estimator comprises a pre-processing module of one or more of the input variables; Said means comprise at least one unit capable of executing a control procedure of the post-processing device, as well as means of control and coordination of the unit or units receiving as input said estimated temperature; the control and coordination means are constituted by a state machine adapted to check the current state of the operation of the engine and to switch or not as a consequence of the current state to a

  another state, then whether or not to modify one and / or the other control procedures of the post-processing device by acting respectively on one and / or the other of the units; the control procedure of the post-treatment device is a procedure selected from the group consisting of a procedure for regulating a temperature, a diagnostic procedure, a procedure for resetting the drifts, a procedure for Regeneration management and load estimation procedure. According to another aspect, the invention relates to a method for controlling a device for aftertreatment of the exhaust gases of an engine, characterized in that an exhaust gas temperature is estimated at using a neural network, and in that by means of said estimated temperature, at least one control procedure of the post-processing device is modified. According to still other aspects, the invention relates to an internal combustion engine and to a motor vehicle equipped with a system according to the first aspect of the invention. Other aspects, objects and advantages of the invention will appear on reading the following detailed description of preferred embodiments thereof, given by way of nonlimiting example and with reference to the appended drawings in which: - Figure 1 is a diagram of an internal combustion engine according to one aspect of the invention; FIG. 2 is a diagram of a system for estimating an exhaust gas temperature used in the context of the invention; FIG. 3 is a diagram illustrating the coupling according to the invention of the estimation system of FIG. 2 with a module for managing the operation of a post-processing device. FIG. 1 schematically represents an internal combustion engine 1 intended to equip a vehicle such as a

  automobile. The engine 1 is for example a diesel engine turbocharged four-cylinder in-line and direct fuel injection. The engine 1 comprises an intake circuit 2 providing its air supply, an engine control computer 3, a pressurized fuel circuit 4, and an exhaust line 5 of the gases. The injection of the fuel into the cylinders is ensured by injectors, not shown, opening into the combustion chambers and driven by the computer 3 from the circuit 4.

  At the outlet of the engine 1, the exhaust gas discharged into the exhaust line 5 passes through one or more after-treatment devices 6 (for example a particulate filter, a nitrogen oxide filter). A turbocharger 7 comprises a compressor disposed on the intake circuit 2 and a turbine disposed on the exhaust line 5. Between the compressor and the engine 1, the intake circuit 2 comprises a heat exchanger 8 for cooling the engine. compressed air at the outlet of the compressor and thus to increase its density, an air intake flap 9 controlled by the computer 3, and a pressure sensor 10 connected to the computer 3. At the output of the engine 1, upstream of the turbine, the exhaust line 5 further comprises means 30 adapted to provide information relating to the temperature of the exhaust gas upstream of the turbine. Said means 30 are for example constituted by a temperature measuring probe or by an estimator provided to provide an estimate of said gas temperature upstream of the turbine. The engine 1 also comprises an exhaust gas recycling circuit 11 equipped with a valve 12 whose opening is controlled by the computer 3; it is thus possible to reintroduce exhaust gases into the intake circuit 2. An air flow meter 13 is mounted in the

  intake circuit 2 upstream of the compressor to provide the computer 3 information relating to the flow of the intake air supplying the engine. Sensors 14, for example pressure or temperature, may also be provided.

  The computer 3 is adapted to process the signals coming from the different sensors, to deduce the states of the engine and to generate the appropriate control signals for the various controlled actuators such as the injectors. The computer 3 thus controls the fuel pressure in the circuit 4 and the opening of the injectors, and this, from the information supplied by the various sensors and in particular the air mass admitted, the engine speed, and stored calibrations to achieve the desired levels of consumption and performance. In the context of the invention, the motor further comprises a neural network estimator 15 provided for making an estimate of the temperature of the exhaust gases upstream of the after-treatment system or systems 6. It will be noted that for reasons of simplicity and sharing of a certain number of resources, in particular computation and memory, it is particularly advantageous to have (as is shown in FIG. 1) the estimator 15 within the calculator 3. management module 20 (detailed further below) adapted to modify at least one operating process of the post-processing device 6 is coupled with the estimator 15 so as to receive as input the temperature information provided by the estimator 15. As for the estimator 15, this module 20 is also advantageously available (as shown in FIG. 1) within the computer. As illustrated in FIG. 2, the estimator 15 comprises a preprocessing module 16, a neural network 17 and a

  post-processing module 18. The estimator 15 comprises a feedback loop provided for returning to the input of the neural network one or more variables at the output of said network. The estimator 15 furthermore comprises an entry relating to the information relating to the temperature of the exhaust gases upstream of the post-processing device 6, as provided for example by the means 30. The estimator may also comprise a or several other entries relating to physical quantities representative of the state of the engine, such as, for example, the flow rate of the gases; the gas pressure upstream of the turbine; backpressure of the gases downstream of the turbine; the position of the turbocharger fins; the speed of the vehicle; the ambient air temperature; the position of the exhaust gas recirculation valve; the supercharging pressure measured by the pressure sensor disposed downstream of the compressor, the heat exchanger and the air intake flap; the temperature of the air in the intake circuit. The estimator can also receive engine speed, fuel flow and airflow as input. The neural network 17 is made up of a number of neurons defined by their parameters (weight, bias), by learning and by their activation functions. The output s of a neuron is related to the inputs (el, e2, ..., en) of the neuron by s = F (ei * w1 + e2 * w2 + ... + in * wn + b), where F is the activation function of the neuron, wI, w2, ... wn the weights and b the bias. The number of neurons in the network, the values of the weights and the biases of the different neurons are calibration parameters which can in particular be determined during a learning phase which will be described further later. The preprocessing module 16 makes it possible to process one or more input variables of the estimator by performing calculations based on

  known physical relations. The preprocessing module 16 makes it possible in particular to reduce the number of inputs of the neural network 17 by calculating one or more variables al, a2 each representative of one or more physical quantity (s), measurable (s). ) or not, absent at the input of the computer 3, from several input variables. The module 16 also makes it possible to filter certain inputs that may take outliers. The neural network 17 has an output channel relating to the desired temperature estimation, as well as possibly one or more other output channels relating to non-measurable state quantities intended to be looped back, directly or indirectly. to the input of the network 17. The post-processing module 18 makes it possible to process the available temperature estimate at the output of the neuron network 17 to emit a signal S of estimated temperature at the output of the estimator, available in particular the management module 20 of the different control procedures of the post-processing device 6. The processing carried out by the post-processing module 18 may be for example a filtering making the estimator robust by limiting the return to the input of an outlier. As mentioned above, one or more of the output variables of the neural network follow a feedback loop and are thus returned to the input of the neural network. Some of the output variables of the network may be processed (eg, filtered, delayed, or associated with other entries in the preprocessing module) before being returned to the input of the network. With reference to FIG. 2, the case is thus illustrated in which state magnitudes 161, 1 (32, / 33 (not measurable and known only during learning) available at the output of the network 17 are looped back to the earth. input of said network 17.

  We have also illustrated the case where the output S of the estimator (that is to say the estimate of the temperature downstream of the turbine disposed on the gas exhaust line) available at the output of the post module treatment 18 (and whose value is here measurable during learning) is on the one hand looped directly to the input of the network 17, and secondly looped back after being delayed (see self-timer represented by the symbol z- ') to the input' of the network 1. By implementing such feedback of one or more of the output variables of the neural network, the network 17 uses as input of its estimate function one or several of the values predicted at the preceding calculation steps by this same estimation function, and whether these values were or were not measured during the learning phase. Such feedback loopback has the advantage of allowing the taking into account of highly dynamic or non-linear phenomena. The neural network 17 can thus be described as recurrent or dynamic, and differs in this respect from a static solution, without a feedback loop, such as that described, for example, in the publication FR 2 864 155. The estimator 15 allows finally, to obtain an estimate of the temperature downstream of the turbine and upstream of a post-treatment device. The estimator can thus be seen as implementing a transfer function taking input temperature information upstream of the turbine, as well as possibly one or more other information relating in particular to physical quantities, to provide in particular output the desired temperature estimate. With reference to FIG. 3, there is shown the management module 20 able to modify at least one control procedure of the post-processing device 6. The module 20 is coupled with the estimator 15 so as to receive the input exhaust gas temperature information S provided by the estimator 15. i0

  The module comprises at least one unit 41, 42, 43, 44 adapted to perform a control procedure of the post-processing device 6, as well as means 30 adapted to control and coordinate the different units 41, 42, 43, 44.

  As examples of control procedures of the operating device 6, each procedure being capable of being performed by a corresponding unit 41, 42, 43, 44, there may be mentioned in particular: a procedure for regulating a temperature of the device 6 (for example the temperature at the inlet, internally or at the output of said system 6) - a regeneration management procedure (in the case of particle filters), purges (in the case of Nox traps), exotherm ( Ox catalyst), etc., adapted to the type of post-processing device 6 used; A diagnostic procedure of the device 6; a procedure for resetting the drifts of the device 6; a charge estimation procedure specific to the type of device 6 (soot for the particulate filter, Nox and Sox for the Nox Trap ...); 20 - etc. The means 30 take for example, as schematically represented in FIG. 3, the form of a state machine adapted to check the current state of the operation of the engine and to switch or not as a consequence of the current state to another state (a state being represented by a circle in FIG. 3), which then modifies or not the one and / or the other of the control procedures of the post-processing device 6 by acting respectively on one and / or the other units 41-44. In the context of the present invention, the state machine 30 is fed in particular by the temperature estimation S provided by the recurrent (or dynamic) neural network estimator. 2901840 He

  The solution proposed by the invention thus differs from control systems proposed in the state of the art using simplistic temperature models in their approach (mapping, physical model of the first order, etc.), linear or non-recurring. Thanks to the recurrent and non-linear aspect of the neural network, it is possible to take into account highly dynamic or even non-linear phenomena for the management of the post-processing device. Furthermore, the solution according to the invention has the advantage of being a little resource-consuming of the computer 3 (in terms of computing load and memory resource required) and to use a limited number of parameters. It is thus relatively easy to implement. Furthermore, insofar as a database 19 useful for learning the estimation function implemented by the estimator 15 is sufficiently representative of all the conditions of use, the solution according to FIG. The invention makes it possible to accurately estimate (on the order of a few degrees) the temperature of the exhaust gas, and that it is in steady state or transient state. This results in a significant gain in the accuracy of the data provided by the management module 20 to the postprocessing device 6, via the units 41-44 for executing control procedures. The description below relates to the different phases of implementation of the system according to the invention. A first phase concerns the design and calibration of the management module 20 of the post-processing device 6. It is particularly necessary to implement the implementation in the computer 3 of the management module 20. It is decided in particular during this phase. post-processing strategies that will be adopted.

  During this same phase, the different units 41-44 adapted to each execute a control procedure of the post-processing device 6 are also arranged. A second phase concerns the validation of the management module of the post-processing device (validation of the post-treatment strategy). This is for example a vehicle test phase carried out using a physical temperature sensor, which makes it possible to decorrelate the operation of the system according to the invention from the operation of a system of the state of the art. (using for example a physical sensor). A third phase concerns the design and calibration of the estimator 15. It is an iterative process between the collection of the data necessary for the design and calibration of the neural network, and the definition of the inputs allowing to realize at best (precision, robustness) and at a lower cost (number of inputs, number of calculations, number and difficulty of the tests) the desired estimator. The estimator is then validated during a fourth phase by carrying out a certain number of tests, and in particular a set of tests of rolling representative of the normal operation of the vehicle, then comparing the results of these tests with the results obtained by implementation of a system of the state of the art using a physical sensor. Finally, during a fifth phase, the entire control system of the post-processing device is validated by reiterating the second phase (validation of the management module) but this time using the neural temperature estimator .

Claims (11)

  A control system of an engine exhaust gas aftertreatment device (6) (1), comprising means (20) for modifying at least one control procedure of said device (6), characterized in that said means (20) is coupled with a neural network estimator (15) (17) adapted to provide a temperature estimate (S) of the exhaust gas.   2. System according to the preceding claim, characterized in that the neural network (15) is a recurrent feedback loop network.   3. System according to one of the preceding claims, characterized in that it is arranged so that the estimator (15) has at the input of information relating to the temperature of the gases upstream of the post-treatment device ( 6) disposed on the exhaust line (5) of the engine, and so that the estimator outputs the estimation of the temperature of the exhaust gas upstream, in or downstream of said device (6).   4. System according to one of the preceding claims, characterized in that the estimator (15) comprises a post-processing module (18) of the temperature estimation performed by the neural network (17).   5. System according to one of the preceding claims, characterized in that the estimator (15) comprises a pre-processing module of one or more of the input variables.   6. System according to one of the preceding claims, characterized in that the means (20) comprise at least one unit (41-44) able to carry out a control procedure of the post-processing device, as well as means ( 30) for controlling and coordinating the at least one unit (41-44) receiving as input said estimated temperature.   7. System according to the preceding claim, characterized in that the means (30) for control and coordination are constituted by a state machine adapted to check the current state of the engine operation and to switch or not as a result of the current state to another state, whether or not to modify one or both of the control procedures of the post-processing device (6) by acting respectively on one and / or the other of the units (41-44).   8. System according to one of the preceding claims, characterized in that the control procedure of the post-treatment device is a procedure selected from the group consisting of a temperature control procedure, a diagnostic procedure , a drift resetting procedure, a regeneration management procedure and a load estimation procedure.   9. A method of controlling an aftertreatment device (6) of the exhaust gas of an engine, characterized in that it estimates an exhaust gas temperature using a network of neurons (15) and in that by means of said estimated temperature, at least one control procedure of the post-treatment device is modified.   10. Internal combustion engine (1) comprising a control system of an exhaust aftertreatment device according to any one of claims 1 to 8.   Motor vehicle comprising a system according to any one of claims 1 to 8.