Performance Assessment of Residential PV Battery Systems

Declining costs for high-performance batteries are leading to worldwide increased use of storage systems in residential building. Especially in conjunction with reduced photovoltaic (PV) feed-in tariffs, a large market has developed in Germany for PV battery systems to increase self-sufficiency. Presently, more than 120,000 PV battery systems are operating in German households. They differ significantly in the type of coupling between PV and battery, the nominal capacities of their components, and their degree of integration. A high system performance is particularly important to achieve profitability for the operator. This thesis develops and evaluates methods for uniform determination of PV battery system performance. They may be used by the public to compare available systems and serve as an indicator in the purchase decision. For this purpose, influences on system performance are analysed and used to derive requirements regarding the test procedures and key performance indicators (KPIs) as final metrics of the system evaluation. Already this requirement analysis reveals that a performance comparison of PV battery systems must cover the efficiency and effectiveness during system operation. As the assessment of effectiveness based on application-independent investigations is impossible, an application-oriented evaluation proves to be necessary. Based on these findings, this thesis proposes a method that combines an application test of system operation during multiple days with additional application-independent investigations. A methodology to define suitable test profiles for application tests is developed and used to systematically select representative days from long-term measurements and for their subsequent combination. By using the results of the application test, two KPIs are calculated, namely: (i) the efficiency as a classic performance figure and (ii) the effectiveness to evaluate control and energy management of the system. The full methodology for performance assessment is applied to five different devices under test (DUT) and proves good applicability with all of them. Its advantages and drawbacks are identified and discussed by comparison to other approaches, namely (a) aggregation of component efficiencies and (b) utilization of generic performance models. Here, a particular focus is on compliance of the different assessment methods with the initially formulated requirements. The analysis reveals that the proposed method does not only sufficiently address all specifications but is also beneficial for the end-user in terms of result comprehensibility and advantageous concerning KPI validity.