Internal Combustion Engine Calibration considering the Embedded Systems Sensor Accuracy

Sammanfattning: In recent years, the increasing economic, environmental and legal demands on internal combustion engines to reduce emissions and fuel consumption are a major challenge. A lot of effort has been put into achieve reliable methods of diagnosis, fault detection and monitoring to increase safety and efficiency in vehicles. A sensor that is not accurate can affect the overall performance of the system. Therefore, a method able to guarantee that the sensors are providing nominal measurements is required. In the process of engine calibration in the test cell intervene two types of sensors, the laboratory sensors that are the ones that belong to the test cell and provide accurate measurements, and the production sensors that are the ones mounted in the engines during normal operation of the vehicles and do not provide as much accurate measurements as the laboratory sensors. The aim of the project is to develop a method able to ensure that the production sensors are providing the correct measurements according to the physical input property. For this purpose, the laboratory sensors were used as reference measurements. The method was divided into automated calibration and manual calibration. The main difference was that the manual calibration requires more intervention of the engineer in charge of the test cell to place extra sensors. In truck industry, the control of emissions (CO, NOx, HC and PM) has been done mainly through SCR catalysts, DOCs and DPFs. The measurements provided by the sensors on which the correct functioning of these systems depends are therefore of vital importance. That is why this Thesis was focused on the performance of the back-pressure sensor located in the exhaust manifold upstream the turbine; the High Temperature Sensor with its three probes positioned in different points of the aftertreatment system and the NOx sensor located at the aftertreatment too. They were compared with their corresponding laboratory sensor. In the proposed method, the gain and the offset are calculated, and comparing them with the supplier specifications, it was possible to determine if the sensor is in the need for calibration or not. Then, modifications would have to be conducted in their parameters stored in the ECUCombustion Engine Calibration considering the Embedded Systems Sensor Accuracyin order to get nominal values. In order to develop the method, previous tests were studied performed in the company, and the implementation of the method was done with actual tests. The final test performed with the parameters modifications has revealed that all the gain and offset values for the sensors analyzed improved their performance, providing values closer to their ideal behavior. However, the final results still lie outside the acceptance threshold. Theoretically, if this method is repeated with further tests, the parameters would continue improving, staying inside the threshold. Therefore, it is in the need of further implementation and verification to be considered completely validated in order to achieve that the sensors remain within specifications during several tests.