The Working Process of the Internal Combustion Engine
Sustainable Mobility, Transport and Power Generation
26th - 27th September 2019, Graz, Austria
The main objective of the forum is the international networking of research and industry in the field of sustainable shipping technologies. Innovative solutions will be presented and discussed. The program can be found at the website of the Sustainable Shipping Technologies Forum.
In LEC’s research and development processes, systematic measurements are indispensable for evaluating technological advances. To conduct measurement campaigns as efficiently as possible, it is necessary to check the quality of the data immediately after it is generated directly at the engine test bed. In the event of a measurement fault, the prompt detection of the fault helps to ensure that corrective action is taken – e.g., the replacement of a faulty sensor. There are multiple ways to detect measurement faults, from simple limit checks to model-based methods that consider the physical relationships between the variables. Since physical modeling of highly complex systems is limited, however, fault detection methods are required that are based on historical data instead of explicitly physical knowledge.
With our first LEC Data Challenge, we would like to encourage data-enthusiastic students and practitioners from a variety disciplines to develop data-based fault detection methods that are best suited for LEC’s large engine applications. To this end, a training data set is provided, a second test data set contains multiple faults. The goal is to find as many as possible. The data challenge ranking will be based on the number of correctly classified points (faulty or fault-free). Thus, it is important not to identify fault-free records as faulty or vice versa. Participation in the Challenge is possible from May 1 (earliest data release) to August 19, 2019. Further information can be found at the website of the LEC Data Challenge.