This invention relates to a method for estimation of combustion chamber pressure of an internal combustion engine and to a device to be used for this purpose.
Combustion chamber pressure is often used as a decisive quantity for describing processes in a combustion chamber of an internal combustion engine. Knowledge of combustion may be used for engine control in order to optimize the combustion process. The parameters of the combustion process such as time of ignition and valve control may accordingly be set by engine control unit.
Combustion chamber pressure may be determined by means of a pressure sensor. Sensors such as this are not cost-effective either in manufacture or installation or in maintenance because of the decidedly high pressures to be measured. This disadvantage is even greater in internal combustion engines with a large number of cylinders.
The object of this invention accordingly is to acquire data on the combustion process in the individual combustion chambers of an internal combustion engine.
It is claimed for the invention that this object is attained by means of a method for estimating the combustion chamber pressure of an internal combustion engine by constructing a model of the internal combustion engine with several model parameters in one model including provision of a combustion chamber pressure value and a model alternating torque, determination of actual alternating torque, adjustment of the model alternating torque to the actual alternating torque accompanied by modification of the model parameters, and determination of an estimated value of the combustion chamber pressure in relation to the model on the basis of the modified model parameters.
It is also claimed for the invention that a corresponding device is provided for estimating combustion chamber pressure of an internal combustion engine with a computer system for modeling the internal combustion engine with several model parameters in a model by establishing a combustion chamber pressure value and a model alternating torque, a data acquisition system connected to the computer system for acquiring an actual alternating torque, the computer unit ensuring that the model alternating torque may be adjusted to the actual alternating torque by modifying the model parameters and that an estimated value of the combustion chamber pressure in relation to the model may be determined on the basis of the modified model parameters.
The model claimed for the invention makes it possible to obtain statements regarding energy conversion in each cylinder. It is an advantage that a characteristic diagram with a plurality of parameters need not be plotted for each cylinder in advance in order to obtain data concerning the combustion process for a current run. The model rather makes it possible to obtain realistic parameters for the cycle and thus to effect pollutant or fuel minimization, for example.
By preference a cycle model for description of combustion in a combustion chamber is obtained in the model. Suitable cycle models have long been known and permit simulation of virtually any combustion process with a plurality of parameters.
In addition, the model may comprise a mechanical model for description of a spring-mass system of the internal combustion engine. This makes it possible to take into account the individual mechanism of an internal combustion engine for generation of torque.
Band limitation may be provided for obtaining model alternating torque. Such band limitation makes it possible both to filter out the constant portion and minimize any disturbances in the high-frequency range.
By preference adjustment of the model alternating torque and the actual alternating torque by error calculation and reduction of the error below a prescribed limit value is effected by means of the model parameters through a control circuit. Automatic model validation is effected by means of this control circuit. However, it is also possible to determine optimized model parameters from the difference between the model alternating torque and the actual alternating torque by means of a single computer step, which is also termed a one-step method.
The actual alternating torque may be an estimated value that has been determined by means of an instantaneous estimation model. The actual alternating torque may also be determined metrologically, as was indicated in the introduction.
The present invention will be described in detail below with reference to the attached drawing, which is in the form of a block diagram of connections of the model claimed for the invention for estimating cylinder pressure. The exemplary diagrams described below represent preferred embodiments of the present invention.
The basis of cylinder pressure estimation is represented by comparison of an actually measured or estimated actual alternating torque IW to a model alternating torque MW, which is determined by a suitable model. In the illustration the model is presented as a control loop on the right side. The model is made up essentially of a cycle model 1 and a mechanical model 2. As is indicated by the arrow pointing downward in the illustration, initial values such as those for engine temperature, ignition timing, and the like are first adopted as approximate reference values for current operating values of the engine from engine control. On the basis of these input parameters the cycle model 1 calculates a pressure pattern in the individual internal combustion chambers of the various cylinders.
The mechanical model 2 employs the pressure patterns as determined in the individual cylinders in order to generate a moment pattern of the crankshaft from them. For this purpose the spring-mass system of the internal combustion engine is taken into consideration. In particular torque is computed with a constant portion and an alternating portion. The alternating portion contains torsion moments such as those of the crankshaft and inertia moments of rotating or oscillating masses such as crankshaft, connecting rod, and the like.
The moment pattern obtained from the mechanical model 2 is subjected in block 3 to band limitation. This serves the purpose in particular of achieving freedom from a mean value, that is, freeing of the moment pattern from the constant moment. In addition, the band limitation also eliminates higher residual frequencies, so that the signal-to-noise ratio of the useful remaining signal increases. The output signal of block 3 accordingly is a disturbance-reduced model alternating torque MW.
In block 4 this model alternating torque MW is compared to an actual alternating torque and a corresponding error is determined and prepared as output signal. By preference the root mean square error is employed as the error.
An attempt is made in block 5 to minimize this error. The error is for this purpose compared to an assigned limit value. If the error is larger than the limit value, one or more of the model parameters is/are modified for the cycle model 1. If the root mean square error is smaller than the prescribed limit value, the optimum desired has been reached and the model parameters of the cycle model 1 may be regarded as realistic for the current combustion process.
The optimal model parameters are here found iteratively in a control loop. However, a one-step process involving more extensive use of computational means may also be applied in this instance for the purpose.
The manner in which the actual alternating torque IW is determined is illustrated in the left-hand portion of the drawing. This is effected in this instance by means of a moment estimating process. The model used for this purpose is indicated symbolically by block 6. An engine speed signal obtained by periodic continuous measurement 61 first undergoes sensor wheel error compensation or sensor wheel compensation 62. The sensor wheel error need be memorized only once in advance for each engine and then stored. Subsequent processing by digital filtering and inertial force compensation 63 results in the desired actual alternating torque Estimation of the actual alternating torque may also be replaced by direct measurement of this quantity. However, out of consideration of costs a sensor system generally is not built into mass-produced vehicles.
It may be said in summary, then, that evaluation of the torque signal for estimating cylinder pressure analyzed on the basis of the crank angle may be employed for estimating cylinder pressure. The cylinder pressure estimation made in this manner smoothes the way to cylinder-selective engine management based on engine speed without costly cylinder pressure sensors. Cylinder misfire recognition may be cited as a typical application. The engine data acquired may also be employed for motor vehicle safety planning purposes.
Number | Date | Country | Kind |
---|---|---|---|
102 56 107.9 | Nov 2002 | US | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/EP03/12316 | 11/5/2003 | WO | 3/8/2006 |