Magnetic Foil Sensors for Analyses in Transformer Cores

Project aim is a completely novel sensor concept for analyses of the interior physical performance of laminated magnetic machine cores, e.g. of transformers or motors etc. Specific relevance concerns the many millions of distribution transformers that exhibit uninterrupted energy loss and disturbing audible noise, both being enhanced in globalized grid systems that contain alternative energy sources. Domain refinement by laser-treatments yielded core materials as hi-tech products. But optimum use in transformers is impeded by unknown behaviours of the magnetic core that represents a complex 3D-system with 3D flux-distributions. Worldwide attempts for lowered losses and noises are handicapped by ineffective numerical modelling and lacking experimental sensors for the core interior - traditional (laborious) sensors causing artefacts through inter-laminar air gaps. This project aims at extra-thin sensor foils, several sensors being mounted on a ca. 0.03 mm thick plastic-foil as a carrier and positioner. A total thickness round 0.1 mm may result in thinnest sensors what-so-ever, so-called nano-sensors usually being much thicker due to their substrate. A ca. 50 mm wide, and for large machines up to meters long foil will be arranged in the core during its assembling. Being a low-cost product, it remains there for long time core diagnostics. For model core tests, a foil may be re-used at different crucial core regions. Sensor contacts of increased thickness remain outside the core, being mounted at a foil end. For simple, rapid and precise sensor manufacturing, a combined 3D/2D-printer will be developed, with three nozzles. Already established 3D-components will allow for plastic elements, such as insulations or for mounting of magnetic glasses as specific sensor components. But for conductive elements, 2D-components will be used. It is expected that the simple 3D/2D combination represents a novelty. Sensor types concern the most relevant local physical quantities: flux density components, losses, temperature-increases, as well as strains and vibrations as being relevant for noise generation. For sensor testing, a ca. 1 m sized model transformer core will be established that represents the most important features of industrial design. The test results will be proved true by means of so called MACC modelling a novel computer-based method that firstly considers non-linear variations of material characteristics for all three directions of space. Final interpretations should contribute to a deeper understanding of 3D-features of transformer cores as a basis for improved machines. With their modern industrial designs, they represent hi-tech products, a fact that has not found general public awareness so far. Carrying-out of project will be by a lab of Vienna University of Technology that is among the leading ones in the given field in cooperation with ABB Transformers, one of the largest producers of electric machines.

Worldwide grid systems comprise millions of electric transformers, with high masses, up to the order of 500 tons. Ecological and economic problems result from considerably high energy losses that arise in continuous ways, due to the fact that transformers work 24 hours per day. Since 100 years, attempts are made to reduce these energy losses. However, they are concentrated in the interior of the machines soft magnetic core, of silicon iron laminations of about 1/30 mm thickness. Analyses of the physical state of the interior were performed by micro sensors as moved through channels drilled through the whole core, or as arranged in inter-laminar ways, respectively. However, both methods need highest expenditure, apart from causing artifacts on interior physical conditions. The aim of project was the development of a novel family of band sensors with total thickness of about 1/10 mm, i.e. not much more than a hair. For manufacturing, a specific 3D/2D Assembler was developed that combines technologies of 3D printing and 2D printing, for combinations of four different conductive and non-conductive print layers. This yielded several sensor types that partly represent world-wide novelties. For analyses of 3D distributions of magnetic flux within complexely built-up machine cores, three individual sensor principles were established. While one is rather simple, the two others need the integration of very thin foils of so-called magnetic glasses, with print-round electric windings. Though this was a crucial challenge, the results prove to be stable and robust enough to allow for repeated applications in different core regions. However, it was not succeeded to integrate a further counter band for sensoring of vibrations. For the measurement of inner thermal profiles and energy losses, the printed sensors reached resolutions of 1/10000 C. As well, very sensitive strain sensors were attained for resolutions of 1/1000000 strain, with the novelty that the usual agglutination is not needed, due to compression that is given in cores a priori. However, exact sensor calibration was not attained. As a conclusion, the targets of project could be more than fulfilled. In addition, an a priori non- planned portable test box was established that offers synchronized registrations from up to 16 sensors, directly at the site of a machine, for a later off-line evaluation.