German study finds greater operational safety needed with photovoltaic systems

Leading independent German broker Funk has warned that the worrying rise in the number of cases of damage to photovoltaic systems on industrial buildings is being driven by operators of such systems not yet paying enough attention to important aspects of operational safety.

The Funk Foundation-funded project PV-DiStAnS 2.0: Artificial Intelligence and has now published its core findings.

Together with a research consortium consisting of the German Aerospace Center, twingz development GmbH and Doepke Schaltgeräte GmbH, the aim was to develop and test a new type of monitoring system for inverter systems so that typical errors and age-related wear of photovoltaic systems can be identified at an early stage to minimise risks.

“Above all, the clear signs of ageing and wear on first-generation photovoltaic systems appear to be the cause of increasing numbers of accidents. The current protective devices are primarily reactive and only intervene at the final stage of the damage event, so that, among other things, possible consequential damage cannot be prevented,” Funk says in the report.

Over the course of the multi-year project, typical types of errors in photovoltaic systems were classified and evaluated. Field tests and laboratory tests were also carried out to evaluate the extent to which AI-based analysis algorithms can increase the operational reliability of photovoltaic systems.

The tests were carried out using measuring boxes designed for this purpose, which recorded and stored the necessary data. Since these systems are influenced by seasonally changing conditions in the network and operating states, and in order to obtain as much data as possible, the measurements were carried out over a longer period of time.

Funk says the aim was to develop a database with fault current signatures. The results were used to examine the extent to which AI-based analysis algorithms can detect these errors.

In addition, the error scenarios analysed in the first phase of the project were replicated and the results were also included in the database. Based on the findings from the investigations, the AI models can now be continuously improved and further developed, says the Hamburg-based broker.

Holger Behrends, project manager at the German Aerospace Center, explains: “Using artificial intelligence, we have the opportunity to monitor the operating behaviour of photovoltaic systems around the clock and to react to critical conditions at an early stage.”

Dr Alexander Skorna, managing director of Funk Consulting and consultant to the Funk Foundation, adds: “Intelligent residual current monitoring can detect risk-relevant anomalies in the operation of PV systems at an early stage and alert those responsible. In this way, necessary actions can be carried out in a timely manner through planned maintenance measures, and the risk of unplanned downtimes and even fire incidents can be significantly reduced.”

As a result, greater safety can be ensured when operating photovoltaic systems on industrial buildings, and the sustainability and profitability of companies can be secured preventively, adds Funk.

In addition, the analysis system created by the project can contribute to the insurability of photovoltaic risks and, according to Behrends, “can be adapted if necessary in accordance with the requirements of the insurance industry and associated testing and certification bodies”.

The importance of this project was underlined by a follow-up project launched last year, ‘PV-DiStAnS-3’, funded by the Federal Ministry for Economic Affairs and Climate Protection.