DE102008043127A1 - Method for operating fuel supply system of internal combustion engine, involves supplying fuel through fuel line from electric fuel pump, where pressure in fuel line or flow rate of electric fuel pump is controlled to target-value - Google Patents

Abstract

The method involves supplying fuel through a fuel line (16) from an electric fuel pump (14). The pressure (p-Kraftstoff) in the fuel line or the flow rate of the electric fuel pump is controlled to a target-value. The injected fuel quantity is pre-controlled by controlling the electric fuel pump by applying the pressure control depending on the temporal course of the target value. Independent claims are included for the following: (1) a computer program for executing a method for operating a fuel supply system of an internal combustion engine; (2) a controlling unit or regulating unit for operating an internal-combustion engine; and (3) a diesel-cycle or an otto-cycle, particularly the homogeneous load compression ignition method or the control light-emitting diode car ignition method for an internal-combustion engine.

Description

State of the art

From the DE 101 58 950 C2 a fuel supply system of an internal combustion engine is known in which the fuel pressure on the delivery side of an electric fuel pump is determined by a sensor. Depending on the output signals of the pressure sensor, the delivery rate of the electric fuel pump is controlled.

As With all regulations, so can these from the state of the art known regulation structurally only on already occurred Deviations of the pressure from a nominal value react. This is special disadvantageous in highly dynamic driving situations.

Around to address these disadvantages, it is known to use feedforward controls, on stationary transmission factors, characteristic curves or maps are based. These pre-controllers use information about the dynamic behavior of the fuel supply system is not or inadequate and as a result deliver a suboptimal one Dynamic quality of pilot-operated pressure. To this deficit At least partially compensating is among the off the technology known Vorsteuerungen a high application cost to adapt to different driving situations and operating conditions required.

Disclosure of the invention

Of the Invention is based on the object, the quality of the pressure control on the delivery side of the electric fuel pump one Fuel supply system, especially in highly dynamic operating state changes, to improve driveability and operating behavior of a with the fuel supply system according to the invention equipped and after the invention Procedure to improve working internal combustion engine. Furthermore the application effort should be reduced.

These The object is achieved by a method according to the invention for operating a fuel injection system of an internal combustion engine solved with the features of claim 1.

By the feedforward control according to the invention can be the dynamic Behavior of the fuel supply system in the pilot control considered the desired operating point so that the control quality is significantly improved.

The inventive dynamic pilot control is advantageously combined in a so-called 2-degree of freedom structure with a control. In this case, the pilot control for the leadership behavior, that is, for the time course of the target pressure, used and the control takes over the compensation of resulting from model errors and disturbances deviations of the actual pressure p _actual from the predetermined target pressure p _target . Due to the 2-degree of freedom structure according to the invention, a very good leadership behavior can be realized. Nevertheless, the physical modeling of the fuel supply system can be kept relatively simple, since the remaining model errors are compensated by the control.

In further advantageous embodiment of the invention Method is provided, from an inverse hydraulic model of the fuel supply system of the internal combustion engine and off planned nominal curves of fuel pressure and injection quantity to form a target speed for the electric fuel pump.

outgoing from this target speed and that determined by the pilot control Target pressure can control, in particular the drive voltage, the electric fuel pump can be determined. That's it possible, the electric fuel pump without delay to drive, resulting in the high control quality of the invention Contributes method.

In further advantageous embodiment of the invention The method is the pressure prevailing in the fuel line with Help of a model-based estimation algorithm, in particular a state observer or Kalman filter. Thereby a pressure sensor in the fuel line can be omitted without the control quality of the pressure control decreases.

In a further advantageous embodiment of the invention, it is provided that the pressure prevailing in the fuel line in dependence on the power consumption and / or the power consumption and / or the rotation number of electric fuel pump is determined.

It has further proven to be advantageous when the estimation algorithm a simulator part and a correction term. The simulator part includes a dynamic physical model of the fuel supply system and is used as a data-driven online simulator. to Compensation of model uncertainties and model inaccuracies This is a correction term in analogy to the return of a Control switched to the estimated by the simulator Correct sizes so that they are against the appropriate ones physical values converge. By this separation of the estimation algorithm 'is it is possible to simulate the in the fuel supply system to simplify ongoing physical processes, because the errors due to the simplifications are caused by the Correction term be compensated immediately.

The inventive method can be further improved when the feedforward control with a parameter estimation combined to the parameters contained in the algorithms Models depending on the current working point and Operational state to adapt online and thus the respective accuracy to improve.

Further Advantages and advantageous embodiments of the invention are the subsequent drawing, the description and the claims removable. All in the drawing, the description and the claims disclosed features can be used both individually and in Any combination with each other essential to the invention.

drawing

It demonstrate:

1 a block diagram of an operating according to the invention internal combustion engine,

2 a block diagram of a pressure control according to the invention with dynamic feedforward control,

3 a further embodiment of the method according to the invention with pressure sensor,

4 an embodiment of the method according to the invention with observers and

5 . 6 : Block Diagrams of the Observer.

In 1 is an internal combustion engine 10 shown greatly simplified. The fuel supply system of the internal combustion engine 10 includes a fuel tank 12 , an electric fuel pump 14 , a fuel line 16 and injectors 20 ,

The electric fuel pump 14 represents a pressure p _fuel in the fuel line 16 ready and promotes via the fuel line 16 Fuel to the injectors 20 , The injectors 20 inject fuel into a suction pipe 22 the internal combustion engine 10 one. The in the suction pipe 22 indicated throttle valve is shown without reference numerals.

At the in 1 illustrated internal combustion engine 10 It is an operating according to the Otto process internal combustion engine with intake manifold injection. The one from the electric fuel pump 14 provided fuel and the associated pressure p _fuel in the fuel line 16 correspond in this embodiment, the injection pressure of the injectors 20 ,

If it is an internal combustion engine 10 is with gasoline direct injection, is between the electric fuel pump 14 and the injection valves still a high-pressure pump available. The method according to the invention then relates to the regulation of the pressure in the fuel line 16 on the suction side of the high pressure pump, not shown, or the delivery side of the electric fuel pump 14 ,

The same applies to internal combustion engines that use the diesel process work, which also prefeed pump and a high-pressure pump exhibit.

The inventive method is also applicable in fuel supply systems with alternative fuels, such as ethanol, liquefied natural gas (CNG) and other fuels or fuel mixtures.

A control and regulating device 18 the internal combustion engine 10 is via signal lines (without reference numeral) with the injectors 20 connected and controls them so that a fuel quantity q _POI corresponding to the torque request of the driver is injected. The electric fuel pump 14 is also via electrical lines (without reference numeral) with the control and regulating device 18 connected. The control and regulating device receives over these lines 18 Information about the current consumption or the power consumption of the electric fuel pump 14 and the speed n _is the electric fuel pump 14 , On the other hand, the inventive control of the electric fuel pump via these lines 14 ,

In 2 is a block diagram of the dynamic pilot control according to the invention shown.

The control and regulating device 18 gives time-dependent reference variables q _{POI, w} (t) for the fuel quantity to be injected and p _w (t) for the fuel pressure.

The time-dependent reference variables q _{POI, w} (t) and p _w (t) are determined on the basis of the current driving situation and are input quantities for a trajectory generation 34 , In trajectory generation 34 corresponding time curves of the setpoint values q _POI (t) and p _setpoint (t) are calculated. The control signal is determined taking into account the system dynamics by using an inverse model ^{-1 of} the controlled system.

In general, the trajectories are determined with the aid of state variable filters, since these additionally provide the temporal derivatives of the reference variables q _{POI, w} and p _w (see also: Wolfram, A. and Vogt, M .: Time Discrete Filtering Algorithms for the Generation of Time Derivatives. Automation Technology 50, 2002 ). The time derivatives of the reference variables q _{POI, w} and p _w are again input variables for a dynamic precontrol 36 needed.

In trajectory generation 34 (please refer 2 ) Is determined on the basis of the model of the hydraulic from the planned target profiles of the reference variables q and p _{POI set} a target rotational speed n _set or target angular speed _to the electric fuel pump 14 calculated.

With this setpoint speed n _setpoint and the setpoint pressure p _setpoint, the dynamic precontrol is determined 36 on the basis of an inverse model ^{-1 of} the controlled system and taking into account the dynamics of the fuel supply system, a control u for the electric motor of the electric fuel supply pump 14 ,

For this purpose, in a first step, based on an inverse model ^{-1 of} the hydraulic system of the fuel supply system, the target angular speed ought ^{to be} the electric fuel pump 14 certainly.

With:

With:
Q soll / mot: set value of the fuel quantity to be injected into the internal combustion engine.

In a second step, the setpoint voltages U _{soll are} determined based on an inverse electrical and mechanical model of the fuel supply system. It is assumed in the embodiment described below that the electric drive is designed as a permanent-magnet synchronous machine and the setpoint voltages U _{should be written} in the known from the literature rotor-fixed dq coordinates. The variables I _d and I _q respectively describe the currents and the variables U _d and U _{q describe} the drive voltages in the d or q direction. The angular velocity is calculated on the basis of a moment balance with the electrical torque M _mot and the applied load torque M _Last .

With:

By controlling the electric fuel pump 14 is the pressure p _Kraftsoff in the fuel line 16 as a controlled variable of the fuel supply system 38 pilot.

Significant influence on the quality of the dynamic precontrol 36 has the modeling of the fuel supply system. In the following, a simple model is shown as an example, wherein the electric fuel pump is driven by a synchronous electric motor. Alternatively, other engine types, such as model for electric fuel pump

Model for hydraulics

• p. = B V (Q EKP (p, w) - Q JP (p) - Q mot ) For example, a DC motor conceivable.

The pressure p in the rail or in the fuel line 16 obtained by evaluating a corresponding differential equation, wherein Q _EKP the funded by the fuel delivery pump fuel, Q _JP , a design-related loss term and Q _mot represents the amount of fuel injected into the engine.

In summary, it can be stated that on the basis of such a model design the dynamic precontrol 36 is made by an inverse model ^{-1 of} the controlled system. From the inverse hydraulic model, the target angular velocity _to the fuel pump 14 determined. As already explained, the control u for controlling the electric fuel pump is determined from the inverse electrical and mechanical model.

The model inversion and calculation of the precontrol can be done by taking advantage of the flatness property (see: Zeitz, M: Flatness-based design of linear and nonlinear feedforward controls. Automation Technology 52, 2004 ) respectively. This flatness property implies that all model states as well as the input signal can be calculated as functions of so-called "flat" outputs as well as a finite number of time derivatives of these quantities. Such a flat output y for the example system is given, for example, by y = (I _d , p, Q _mot ) ^T. In this way, the input signals U _d and U _q , ie the drive voltages of the electric motor of the fuel pump, can be calculated by the variables I _d , p, Q _mot and corresponding time derivatives. The time derivatives, in turn, are calculated, as already mentioned, with the aid of a state variable filter. In addition, reference should be made to the relevant literature, in particular Rothfuß, R. et al: Flatness: a new approach to the control and regulation of nonlinear systems. automation sierungstechnik, 45, 1997 ; Hagenmeyer, V. and Zeitz, M: Flatness-based design of linear and non-linear pilot controls. Automation Technology 52, 2004

A block diagram of an advantageous 2-degree-of-freedom structure is shown in FIG 3 shown. For the sake of clarity is in the 3 and 4 the trajectory generation 34 not shown.

It serves a regulation 40 to compensate for physical effects not contained in the inverse model ^-1 . These effects, if they have not been found in the modeling input, also not in the feedforward 36 be considered and would lead to a reduced quality of control without a suitable compensation.

At the in 3 Block diagram shown is a pressure sensor in the fuel line 16 present, whose output signal p _actual in an adder with the setpoint p _Soll merged and the controller 40 is supplied as input.

In the embodiment of a method according to the invention 4 there is no feedback of the measured by a pressure sensor pressure p _actual , but it is an inventive observer 42 used, from the control of the electric fuel pump 14 , in particular the control u or of the current I, and the rotational speed n of the fuel pump 14 a pressure p _is estimated. This makes it possible to access a pressure sensor in the fuel line 16 to renounce. As a result, the hardware costs and the assembly costs of the fuel supply system according to the invention are reduced 38 ,

In 5 is the observer 42 shown in more detail. It becomes clear that the observer 42 a simulator part 44 and a correction term 46 includes. The Observer 42 is designed in two steps. First, a physical / dynamic model Σ ^ of the fuel supply system is created. This model Σ ^ forms the simulator part 44 the observer 42 , The model or the simulator part 44 on the one hand must reproduce the physics accurately enough, but in terms of implementation in the control unit 18 and to make the associated computation time as simple as possible. Subsequently, a correction term or controller is created 46 educated.

For the application of the invention has for the design of the observer 42 comprising the simulator part 44 and the correction term 46 , a so-called "extended Luenberger observer" approach proved to be suitable.

The general algorithm of both a state observer and the Kalman filter are known in principle from the literature (see, for example: Zeitz, M., The extending Luenberger observer for nonlinear systems, Systems & Control Letters 9, pp. 149-156, 1987 ). Also, the general algorithm of the state observer and the Kalman filter are known from the literature, see above references. An alternative variant for deriving a suitable estimation algorithm for the fuel pressure p _actual is z. Eg the "Advanced Kalman Filter". For a detailed illustration see for example Föllinger, O., Control Engineering Introduction to Methods and Their Application, Hüthig-Verlag, 1994; Grewal, MS and Andrews, AP, Kalman Filtering: Theory and Practice, Wiley Publishing, 2001 ,

The exemplary nonlinear model of the fuel supply system can be summarized with the state vector x = (i _d , I _q , u, p) ^T and the vector of the input signals u = (U _d , U _q , Q _mot ) ^T in the following basic structure:
Σ: = x. = f (x, u), t> 0, x (0) = x ₀
y = h (x), t ≥ 0

there the vector f () describes the right-hand sides of the above Differential equations of the model of the fuel supply system and the variable y the output quantities respectively Metrics.

The measuring model h of the sensor system is described by h (x). In the present case, the rotational speed n and / or the current of the electric fuel pump 14 measured. As an alternative to the specified model, other model variants of the hydraulics and / or of the electric motor are likewise covered by the method according to the invention. This also applies to observers / filters based on simplified models, which neglect the electric model equations for the currents and which only model the mechanical part of the electric motor and the fuel supply system. Furthermore, the erfin includes Methods according to the invention are observers and / or filters which have been designed with the aid of models which have been extended by a disturbance value approach of the form dz / dt = 0. In this case, the variable z represents an unknown disturbance variable, such as a deviation of the injected fuel quantity from the desired injection quantity (cf. 3 and 4 ).

The observer is then obtained according to the procedure described above 42 to estimate the pressure in the following form

The Model is used as a copy in the observer with a correction term G (t) · (y - ŷ) is switched on. Of the Correction term consists of a target-actual comparison between measurement y and estimate ŷ. The difference between these two Signals are multiplied by the gain G (t). Of the Amplification G (t) design is done with the above mentioned Method. The alternatives given are equivalent Procedures based on identical or very similar structures of the algorithm described in the invention.

In 6 what has been said is shown in the form of a block diagram.

Basically, with the sign Π the physical fuel system,
with the sign Σ ^ of the state estimator 42 and
with the symbol Π ^-1 denotes the inverse model of the physical fuel system.

QUOTES INCLUDE IN THE DESCRIPTION

This list The documents listed by the applicant have been automated generated and is solely for better information recorded by the reader. The list is not part of the German Patent or utility model application. The DPMA takes over no liability for any errors or omissions.

Cited patent literature

• - DE 10158950 C2 [0001]

Cited non-patent literature

• - Wolfram, A. and Vogt, M .: Time discrete filter algorithms for generating time derivatives. Automation Technology 50, 2002 [0031]
• - Zeitz, M: Flatness-based design of linear and nonlinear feedforward controls. Automation Technology 52, 2004 [0043]
• - Rothfuß, R. et al: Flatness: a new approach to the control and regulation of nonlinear systems. Automation Technology , 45, 1997 [0043]
• - Hagenmeyer, V. and Zeitz, M: Flatness-based design of linear and non-linear pilot controls. Automation Technology 52, 2004 [0043]
• - Zeitz, M., The extending Luenberger observer for nonlinear systems, Systems & Control Letters 9, pp. 149-156, 1987 [0050]
• - Föllinger, O., Control Engineering Introduction to Methods and Their Application, Hüthig-Verlag, 1994; Grewal, MS and Andrews, AP, Kalman Filtering: Theory and Practice, Wiley-Verlag, 2001 [0050]

Claims (11)

Method for operating a fuel supply system of an internal combustion engine ( 10 ), in which fuel from an electric fuel pump ( 14 ) in a fuel line ( 16 ), with the fuel line ( 16 ) prevailing pressure (p _fuel ) and / or the delivery rate of the electric fuel pump ( 14 ) is controlled and / or regulated to a desired value (p _setpoint ), characterized in that when using a control system in the fuel line ( 16 ) and / or a fuel rail prevailing pressure (p _fuel ) by a dynamic precontrol ( 36 ) as a function of the time profile of the desired value (p _desired ) and the desired time profile of the fuel quantity to be injected (q _POI (t)) by triggering ( 36 ) of the electric fuel pump ( 14 ) is piloted. Method according to claim 1, characterized in that for the precontrol ( 36 ) an inverse model ( ^-1 ) of the electric fuel pump ( 14 ) with an electrical and a mechanical portion and an inverse hydraulic model of the fuel supply system of the internal combustion engine ( 10 ) is formed. A method according to claim 2, characterized in that from the inverse hydraulic model of the fuel supply system of the internal combustion engine ( 10 ) taking into account the planned nominal curves of fuel pressure (p _soll (t)) and injection quantity (q _POI (t)) a setpoint speed (n _setpoint (t)) for the electric fuel pump ( 14 ) is determined. A method according to claim 3, characterized in that from the setpoint speed (n _setpoint ) of the electric fuel pump ( 14 ) and the set pressure (p _setpoint ) the control, in particular the current ( 1 ) and / or the voltage (U), the electric fuel pump ( 14 ) is determined. Method according to one of the preceding claims, characterized in that in the fuel line ( 16 ) prevailing pressure (p _actual ) with the aid of a model-based estimation algorithm, in particular a state observer ( 42 . 46 ) or a Kalman filter. Method according to one of the preceding claims, characterized in that in the fuel line ( 16 ) prevailing pressure (p _actual ) as a function of the current consumption (I) or the power consumption (P _el ) and the rotational speed (n) of the electric fuel pump ( 14 ) is determined. Method according to one of Claims 5 or 6, characterized in that the estimation algorithm comprises a simulator part ( 42 ) and a correction term ( 46 ). Method according to one of the preceding claims, characterized in that a parameter estimate provided is, and that the parameter estimation the parameters of depending on the models contained in the algorithms adapted to the current operating point and operating status online. Computer program, characterized in that it for carrying out the method according to one of the preceding Claims is appropriate when executed on a computer becomes. Control and / or regulating device ( 34 ) for operating an internal combustion engine, characterized in that it comprises a memory on which a computer program according to claim 9 is stored. According to the diesel process, the Otto process and / or an alternative process, in particular the HCCI process (HCCI = Homogeneous Charge Compression Ignition) or the CAI process (CAI = Controlled Auto Ignition) working internal combustion engine ( 10 ), with injection devices ( 20 ), which inject the fuel directly or indirectly into the combustion chambers, and with a fuel supply system comprising at least one electric fuel pump ( 14 ), characterized in that it is a control and / or regulating device ( 18 ) according to claim 10.