Fast Prototyping of a Racing Diesel Engine Control System 2008-01-2942
This paper shows how Rapid Control Prototyping (RCP) and Computational Fluid Dynamics (CFD) techniques have been applied to design and implement an engine control system for a common rail diesel engine. The project aim is to setup a high performance engine in order to participate to the Italian Tractor Pulling Championship (Prostock category).
The original engine is a John Deere 6081 Tier2 model, already equipped with a common rail system. Engine performance is substantially determined by the control system, which is in charge of limiting engine speed, boost pressure and Air to Fuel Ratio (AFR). Given that typically the information and equipment needed to change control parameters are not accessible to customers, the first step of the project has been to replace the original control system, while maintaining injectors and pumps. This solution can guarantee the best performance, but it requires time to design the new control system, both in terms of hardware and software. The challenge of setting up a new control system in a three months time is even more demanding if the control parameters cannot be tuned on the test bench. As it is well known, a common rail system requires the calibration of many parameters (rail pressure, injection timing and splitting, boost pressure, minimum AFR, etc.). This demands for intensive testing, which is expensive and time-consuming; moreover, tests must be carried out in racing conditions (i.e., high engine speed, high boost pressure, low AFR), affecting engine durability. Cost reduction is a key factor in this type of competition, therefore Computational Fluid Dynamics (CFD) simulations were used to assess a range of control parameters setting that could be used as a decent starting point for a fine optimization on the test bench.
The time constraints oriented the hardware choice toward an RCP (Rapid Control Prototyping) system, programmable by means of object oriented software. Computational power, harsh environment compatibility, availability of third party products (such as power modules for the injectors) were taken into account in order to chose the hardware-software package. The ECU has been finally implemented on a National Instruments cRIO 9074 equipped with a 400 MHz controller, 2 Mgate FPGA and Analog-digital I/O modules.
The control system was designed to guarantee performance, driveability and easy access: it was equipped with an easy-to-use interface, allowing the user to change crucial control parameters, and to manage data logging, that can be used to monitor the engine performance after the race.