Testing Methodology for Smart Grid Systems: A Multi-Stage Approach Applied to Architectures Utilizing Smart Metering Infrastructure

The energy system as a whole is undergoing a significant transformation process. The driving motivation for this change is the significant reduction of carbon dioxide and other greenhouse gases to limit the extent and the impact of climate change. The foundation for this new energy system will be a generation portfolio primarily based on wind and solar power plants. The future energy system will be, in almost all aspects, a cyber-physical system. It will be characterised by being a system-of-systems, combining software and physical devices and having a high degree of interconnections. These systems are commonly referred to as Smart Grid systems. The needed effort to complete this transformation process is not linear over time or decarbonisation but will be more challenging as it progresses. In the same aspect, the efficient solution for today will presumably be less efficient in 2050. The finding in this projection of the future is that creating and changing technical systems are inevitable factors. The required change rate and the commonly used development and implementation operations need to fit together. The improvement of these processes is the motivation for this contribution. A relevant part of every development process in the energy domain is a laboratory in which to conduct experiments. The experiments test the assumptions made in the conceptual and development phases against the implemented elements. Due to the nature of the energy domain, laboratories typically represent only a sub-system to test a specific aspect. The adjacent systems and their interfaces are modelled in the experiment. Depending on the experimental design, a specific member of the ’X-in-the-Loop’ family of experimental setups is chosen to study the System under Test. The most relevant variants are Power-Hardware-in-the-Loop (PHiL), Controller-Hardware-in-the-Loop(CHiL) and Software-in-the-Loop (SiL). Depending on the application, such a System un-der Test can be a sensor (such as a Smart Meter), an actuator (such as a PV inverter), a control algorithm or a complete control schema incorporating all of the sub-systems as mentioned earlier. This contribution focuses on structuring the previously mentioned methods and experimental setups. The formulated hypothesis guides alongside a step-wise process from a simple lab test to a real-world field trial. These steps are considered a stage, and a particular testing technology is associated with each stage. The underlying research question asks for processes to streamline testing processes for repeated tests based on a mathematical metric. The developed methodology is centred around the value a tested system offers to the rest of the systems and how this value can be measured. The starting point for this analysis are the set of use cases, which describe how the System under Test (SuT) should be used. From this value, relevant observations, which can be made in an experimental setup, such as the measurement of the active power of a PV inverter, are identified. These observations are processed to be condensed into a metric. With each metric, a set of acceptance criteria is derived from the use cases to have a binary result - passedor failed - for each test run. The other relevant use of the metrics is the multi-stage assessment of a complete series of tests. The mathematical tools from variance analysis are used to identify differences in the results of an experiment carried out in different stages and thus identify an under-complex representation of specific effects in earlier stages. The methodology is applied to different systems development in a variety of research and development projects. A selection of these test campaigns is presented in the result chapter of this contribution. This section depicts all steps as described in the method-ology. The relevant findings are: Variance analysis of metrics gathered is suitable to examine multi-stage test campaigns. The results are condensed into a set of recommendations and formulated in the form of commandments to express them clearly. The extension of the methodology and implementation test campaigns with the following concepts have been identified as relevant for further research: Use of Hybrid Stage Testing, implementation of Multi-Lab-Testing and including in Regulatory Sandbox Testing. Another crucial area of further research is the operation and maintenance of complex SmartGrid systems with different requirements profiles compared to research and development projects.