Robert Bosch LLC is a subsidiary of the Bosch Group, a leading global supplier of technology and services, active in the fields of automotive technology, energy and building technology, industrial technology, and consumer goods. According to preliminary figures, more than 306,000 associates generated sales of 52.3 billion euros ($67.2 billion) in 2012. The Bosch Group comprises Robert Bosch GmbH and its more than 350 subsidiaries and regional companies in some 60 countries. This worldwide development, manufacturing, and sales network is the foundation for further growth. Bosch spent some 4.5 billion euros ($5.8 billion) for research and development in 2012, and applied for over 4,700 patents worldwide. In the U.S., Canada and Mexico, Bosch employs over 22,500 associates in more than 100 locations, with sales of $9.8 billion in fiscal year 2011.
Improving Simulations of Advanced Combustion Engines
The automotive industry is facing the challenge of improving fuel economy and reducing emissions while delivering the performance that consumers demand. Despite the advent of the advanced battery electric vehicle and developments in electric powertrains using hybrid systems, liquid-fueled internal combustion engines are expected to remain an integral technology in future vehicle designs for some time. Consequently, optimizing engine performance is pivotal in reducing national energy consumption, both in the near term and far into the future.
One way that next-generation engines will deliver improved performance is through more advanced combustion concepts, such as multimode combustion operation. In multimode combustion, traditional spark ignition is augmented with homogeneous charge compression ignition to reduce emissions and maximize thermal efficiency over all driving conditions—including acceleration and cruising. However, successfully employing this engine technology requires advanced control algorithms to switch the engine between combustion modes according to operating conditions.
This project teams researchers at Robert Bosch LLC in Palo Alto and at Lawrence Livermore National Laboratory to run simulations for understanding the transitions from spark ignition (SI) to homogeneous charge compression ignition (HCCI) in automobile engines so that an engine controller can be developed. The transition from SI to HCCI is a complex problem that has occupied the engine research community for some time. Accurate simulations of the transient characteristics requires the use of advanced simulation techniques such as Large Eddy Simulation (LES). Running an engine under the homogeneous charge compression ignition regime allows for more efficient fuel usage in propelling the automobile under steady engine loads. Under the SI regime, which most cars on US roads use, the automobile has sufficient power to accelerate – useful in merging onto freeways and starting from stops. Pairing the desirable qualities of spark ignition with those of homogeneous charge compression ignition allows consumers to choose cars that deliver good performance and fuel efficiency. By simulating 10 engine cycles (two under the SI regime followed by eight under the HCCI regime) the team will test different operating strategies for their effect on smooth transitions.
In this collaboration, the team has been able to complete multiple test cycles of engine operations to validate the multi-mode combustion model.
Models for liquid spray injection and spark and auto-ignition modes have been tested. With the collaboration and use of LLNL’s HPC Innovation Center, there is a 70% reduction in calculation time for each engine cycle, which makes the simulation of an entire transition process feasible.
Prior to the hpc4energy incubator
These simulations were first tackled by researchers at Robert Bosch LLC using commercial Reynolds-Averaged Navier-Stokes (RANS) codes that are unable to resolve the unsteady cycle-to-cycle dynamics of the engine. Large Eddy Simulation (LES) can be used to investigate the unsteady nature of the SI to HCCI transition; however the additional computational cost made the simulation of multiple cycles infeasible with the previously available resources.
With the resources available through the hpc4energy incubator, the computation group at Robert Bosch LLC was able to use the Stanford University LES software CIAO, a structured, three-dimensional explicit compressible flow code for these calculations. The simulation of one engine cycle requires two weeks on 1,000 processors to calculate the chemistry and flow. The results of the 10 engine cycle simulation allows researchers to understand the effect of an operating strategy on the transition, which under this method translates to twenty weeks of complete and uninterrupted computer calculations.