Smart Data model generation for big data models in a solar thermal power plant

Abstract

FIWARE is a collection of open source components that can be used to realize a platform for smart solutions. The basic idea is to collect data from heterogenous sources using different types of agents. The agents unify data according to NGSI-LD standard and FIWARE smart data models and finally forward it to a central IoT broker where it can be further processed. To allow integration of new data sources configuration files for the smart data model and the mapping of the agent have to be created which describe all entities, attributes, relations and map them to original property names of the data source. The effort is manageable for the integration of data sources with only a few attributes, but results in a lot of manual work in scenarios with big data models. As an example, this presentation demonstrates the integration of a process control system (PCS) of a solar thermal power plant into FIWARE platform. The PCS has about 55 entities and 90000 available cryptic variables that result in huge integration effort and error-prone manual work. Using the example of PCS integration, it is shown that name systems can help to automate FIWARE resource generation for the integration of data sources with big data models.