Detección de fallas a través del modelo Fuzzy Logic o decisión por umbrales en sistemas de inyección electrónica de vehículos con comunicación CAN.

Abstract

In order to contribute to automotive diagnostics, it is intended to provide a system for detecting faults in the electronic injection system of vehicles communicated with the CAN protocol, which will help efficiently reducing the risk and complexity it presents, for which; a DAQ data acquisition card dedicated to this protocol was used with LabView 2016 software to take the identification parameters PIDs that the vehicle owns and later identify the operating ranges to establish the limits that were used in the detection of thresholds of each signal in the same way to determine the ranges of the linguistic variables that are in the functions of entrance membership of the Fuzzy Logic model, which will be in charge of detecting faults in the different systems that involve the electronic injection of the vehicle by means of combinations programmed in rules according to the expertise of the automotive technician and a workshop manual of Hyundai, these failures were recorded in static and dynamic tests. The evaluation of faults was made in a KIA Sportage 2013 car, during the aforementioned tests the faults are presented eventually, so it was necessary to record the frequency with which they present the faults and in this way identify the faults that occur most frequently in each vehicle system at the time of the final report. In this way the system is useful for the planned purpose allowing real-time monitoring of the vehicle, allowing subsequently to identify faults that not only occur eventually, for the analysis must be considered or initially determine which are the PID's in the vehicle to have a good performance, it is important to continue the fault detection study with other models of artificial intelligence.