Abstract

CONTRERAS., Wilmer; ARICHAVALA, Mauricio  and  JEREZ, Cristian. Determination of the maximum compression pressure of an engine provoked ignition based on a recurrent artificial neural network. Ingenius [online]. 2018, n.19, pp.9-18. ISSN 1390-860X.  https://doi.org/10.17163/ings.n19.2018.01.

In the present research the explanation of the applied methodology for the determination of the maximum pressure of compression of an engine of internal combustion alternative of ignited provoked (MEP), that is based on a study that starts from the characterization of the curves of the amperage rating of the starter motor. A protocol for data acquisition and subsequent statistical analysis is applied. The statistical values of the signal as energy, average, standard deviation, variance, kurtosis, asymmetry, maximum, minimum and crest factor are selected in function of the greater contribution of information for the characterization of the experiment; these values generate databases that are applied for the creation and training of a recurrent artificial neural network (RNAR) in which an absolute error of less than 2% is obtained. In the first instance, the test methodology is applied in an engine assembled in a didactic bank and then the application of the method is applied in vehicles.

Keywords : diagnosis; compression pressure; RNAR; MEP; Elman Network; Layer Appellant.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )