Research Group 2401
Optimization-Based Multiscale Control of Low-Temperature Combustion Engines
Due to the growing global demand for energy, fossil fuels will remain indispensable in the foreseeable future. Among other things, this affects the transport sector, for which area-wide availability and a high energy density of the applied energy sources are important. Considering the high energy density and good storage possibilities, the use of liquid fuels for mobile drive systems appears to play an essential role for a long time to come, even if their use can be reduced by increasing electrification, for example through hybrid vehicles. In connection with an increasing shortage of resources and rising environmental pollution, economic and ecological aspects of energy supply are increasingly coming into the focus of interest. The combustion of fossil fuels, which mainly consist of hydrocarbons, produces pollutant emissions such as nitrogen oxides (NOx), carbon monoxide (CO), unburned hydrocarbons (uHC) and soot, which contribute significantly to urban and regional air pollution. In addition, the greenhouse gas CO2 is produced, which leads to changes in the global climate. The reduction of pollutant emissions and greenhouse gases is an important social goal.
In the field of engine combustion, innovative combustion processes are currently being researched to achieve this goal. On the one hand, these processes meet the existing requirements for performance and comfort, and on the other hand, they avoid the production of pollutant emissions by means of internal engine technology while maintaining the highest possible efficiency. In particular, combustion processes with low-temperature combustion (LTC) have proven to be very promising. This applies to both gasoline engines (Gasoline Controlled Auto Ignition, GCAI) and diesel engines (Premixed Charge Compression Ignition, PCCI). Those combustion processes are characterized by a high degree of homogenization of the fuel-air mixture and combustion initiated by self-ignition, which takes place without significant flame front propagation. The homogenization with simultaneously lower peak temperatures leads to a significant reduction of emissions at a very high degree of efficiency. While the concept of LTC offers significant advantages with respect to efficiency and emissions, the challenges faced have not yet been sufficiently solved for a technical realization.
Structure of the research group
The thematic and methodological linking of the subprojects and the associated interdisciplinary cooperation is crucial for the realization of the highly complex project. For example, the planned optimization-based control systems, which need to meet the tough requirements for real time, can only be developed if expertise from the fields of numerical processes, controller development, combustion engineering and combustion chemistry is combined. The close cooperation between the research institutions is supported by the extensive networking of the subprojects.
Workshops on the various topics are held in regular intervals. The focus, as shown here in the photo, is on personal exchange in small groups. As part of the summer campus of the technical faculty of the Albert-Ludwigs-University of Freiburg, the optimization tools CasADi and acados, which are developed at the Institute for Microsystems Technology at the University of Freiburg, were presented to IRT PhD students and the junior professorship for Mechatronic Systems for Combustion Engines at RWTH Aachen University.
Objective
The aim of the research group is to use innovative control concepts to make the promising LTC in combustion engines technically applicable. To accomplish these goals, the research group is using a novel interdisciplinary approach from the fields of chemistry, combustion and control engineering. Currently researched methods for LTC control can be classified as cycle-based control concepts. The established control systems use information from previous cycles to determine cycle-integral surrogate parameters that serve as control variables. On this basis, the current manipulated variables are calculated on a cycle-to-cycle scale. Using a cycle-based approache, process control is feasible in principle, but only in a very restricted range, which is not sufficient for technical application.
The multiscale control approaches, that have to be newly developed within the framework of FOR2401, are to be supplemented by additional inner-cyclic information, which appears to be crucial for robust control of LTC. The addition of inner-cyclic information makes it possible to take into account not only the dynamics occurring from cycle to cycle, but also the decisive effects of LTC combustion processes, which take place on a much smaller time scale than that of a combustion cycle. Altogether, multiscale control consists of two approaches which are combined with each other.
- a) The process is controlled on a finer time scale under consideration of the non-linear process behavior. Thus, considerably more information from the combustion process is used to characterize and control the combustion. The investigations go up to the extreme case that the complete combustion process is (quasi-)continuously controlled (combustion process control). The planned combustion process control allows the utilization of completely new degrees of freedom, which are currently not available. The cycle rate of the combustion process control remains cycle-to-cycle based.
- b) Furthermore, an in-cycle control shall be established, which uses measurements of the current cycle in order to generate control interventions within the same cycle. This means that even if a disturbance occurs, which can cause a significant effect due to the sensitive process behavior, it is still possible to intervene in the same cycle to maintain stability.
With currently researched control concepts, it is only possible to react to this disturbance variable in the next cycle and thus a disturbance variable can lead directly to a combustion misfire with drastically increased emissions. Multiscale control thus allows a more direct control of the combustion process, with which the combustion stability and efficiency can be positively effected and the emissions can be reduced. In order to realize multiscale control, new fundamental research questions must first be investigated. These relate to the development of the required basic understanding of combustion technology with regard to the possibilities of influencing the inner-cyclic processes of the combustion. Based on this, tailor-made control engineering methods for the realization of this concept must be developed.
Funding
External Links
- FOR2401 - Website
- Albert-Ludwigs-Universität Freiburg, Professur für Systemtheorie, Regelungstechnik und Optimierung
- RWTH Aachen University, Institut für Regelungstechnik
- ETH Zürich - Institute for Dynamic Systems and Control
- University of Duisburg Essen - Institute for Combustion and Gas Dynamics
- RWTH Aachen University, institute for combustion technology