Rail pressure estimation for fault diagnosis in high pressure fuel supply and injection system

Engine roughness (ER) is a complex issue in GDI engines and a frequent customer complaint in workshops. Still it is hard to isolate the root cause of the vibration which is received by the driver. The present paper aims to identify the type and extent of faults causing ER that have their origin in the fuel supply and injection system. The method is presented using the injectors as an example, since they have a great impact on engine related vibrations. The injectors affect ER through mixture formation. For example, a deviating amount of injected fuel mass on one cylinder leads to a deviating delivery of work during the combustion cycle and thus causes vibrations. For the investigation additional pressure sensors were installed in the high pressure fuel system to observe the transfer behavior in the hydraulic system. Tests were executed in different reference and fault states, where a fault state is represented by deviating mass flows of an injector. The generated data is used to develop a parameter estimation model, describing the pressure in the fuel rail of the investigated engine. Firstly, a set of reference parameters is generated by a parameter optimization algorithm for each operating point under reference conditions. Then, these sets of parameters are used for initial calibration of the model for the following injector diagnosis. Observing the adaptation of a separate set of diagnostic parameters, allows for a precise pinpointing to a defective injector. It also delivers information about the type of fault and its size. Finally, the results are reconfirmed by executing the diagnosis on data of a healthy system to preclude mis-detections of faults.