| 摘要 |
A method for misfire recognition in an internal combustion engine, with the crankshaft of the internal combustion engine having a sensor for determining the rpm and with at least one crankshaft segment being established for each cylinder. By means of the rpm, the segment time (tSEG) is determined and then the differences ( DELTA tSEG) are determined from two successive times tSEG(n), tSEG(n-1). In a first phase, a plurality of processing units are calculated from the collected values for each operating point and each possible error, with each of these processing units having a weighting vector and information (I). These weighting vectors are stored in a storage unit of the internal combustion engine. In a second phase, during operation of the internal combustion engine under operating conditions, sample vectors are calculated from the current values then being picked up, the vectors being compared with the individual weighting vectors of the processing units that have been learned. The processing unit that has the shortest distance from the weighting vectors of the currently calculated sample vectors is selected. Then, the information from the processing unit is output as a diagnosis of the functioning of the internal combustion engine. |
