The project, carried out from 1 January 2020 to 31 December 2023 and extended by seven months to 31 July 2023 because of the COVID‑19 pandemic, aimed to develop a surface‑plasmon resonance (SPR) sensor capable of monitoring the ageing of insulating oils used in power transformers. The sensor was designed to detect changes in antioxidant concentration, polar degradation products and other ageing markers that influence the dielectric strength of the oil. The research was funded by the German Federal Ministry of Economic Affairs and Energy (BMWi) through the µSPIN programme, which provided an existing SPR‑imaging laboratory that could be optimised for the project’s specific needs.
In the first work package the team defined the technical specifications and modular interfaces for the sensor, the fluidic system, the light source and the software. The sensor’s central element was fabricated by the University of Oldenburg (OTH) and integrated with fluidic components developed by the University of Rostock (UR), the Technical University of Dresden (THD) and the Institute of Electrical Engineering (SOL). Light‑source specifications were agreed upon by SOL, TOP and OTH, while the software interface and the planned incorporation of artificial‑intelligence modules were coordinated by the German Aerospace Center (GS) and OTH. The project was organised into nine work packages, each focusing on a distinct aspect of the sensor’s development, from component analysis to prototype testing.
The core scientific work involved analysing a large pool of over 100 oil samples, including both conventional mineral oils and ester‑based oils. The samples were selected according to defined parameters and stored separately. The University of Rostock used these samples to test the sensor’s ability to distinguish oil types. Initial results showed that the SPR sensor could differentiate between oil types, but the influence of the oil’s intrinsic properties made it difficult to assess ageing differences. Consequently, the team shifted focus to “trend samples” that varied in only one parameter, enabling clearer detection of ageing effects.
A key objective was to monitor the concentration of the antioxidant 2,6‑di‑tert‑butyl‑p‑hydroxy‑cinnamic acid (DBPC), which is added to fresh oils to protect against oxidative degradation. The concentration of DBPC was measured in mineral oils using DIN EN 60666 and in ester oils using ASTM D6971, the latter being the standard method for turbine oils. Calibration standards were prepared by OEL with known DBPC concentrations in a blank oil, allowing quantitative determination in both oil types. Comparison of the two analytical methods revealed only limited correlation because degradation products interfered with the ASTM D6971 measurement. Nevertheless, the resulting data set provided a robust basis for training the SPR sensor to recognise antioxidant levels.
To create realistic test fluids, OEL produced defined mixtures of oils with controlled refractive indices and varying degrees of polar degradation products or oxidation. These control fluids were characterised for absolute and relative water content, density, breakdown voltage and surface tension, eliminating disturbances in subsequent tests at OTH. The sensor’s surface was engineered using graphene materials derived from flake graphite. The team optimised the Hummers synthesis to produce graphene flakes of suitable size, then reduced the graphene oxide with elemental potassium under inert gas at low temperatures to achieve variable oxidation degrees. The resulting graphene layers adhered strongly to the gold film, maintained good wettability, and could be functionalised to adjust affinity for ageing markers, thereby preventing irreversible binding that would compromise continuous monitoring.
Throughout the project, the consortium maintained close collaboration. The partners—OEL, UR, OTH, SGB, THD, SOL, TOP, GS, Fuchs GmbH and Maschinenfabrik Rheinhauen—contributed complementary expertise in oil chemistry, sensor fabrication, fluidics, optics and software. The project’s outcomes were integrated into five master’s theses and two doctoral dissertations, demonstrating its impact on training the next generation of researchers. The final deliverables include a fully characterised SPR sensor prototype, a database of antioxidant concentrations, a set of calibrated control fluids, and a software platform incorporating AI modules for real‑time ageing assessment.