Rotational Magnetization

During the last ten years much international effort was taken to study the phenomenon of rotational magnetization (RM) which means that the induction vector B rotates periodically in the xy-plane of laminated soft magnetic material. So-called "rotational single sheet testers" (RSSTs) were developed to produce RM in samples of SiFe sheets or amorphous ribbons. Experimental studies showed that RM may yield increases of energy losses up to several hundred percent. Further, the magnetostriction - a main source of audible noise of machine cores - prove to be increased by up to a full order of magnitude. Most energy loss measurements by means of RSST have been performed for ideal circular or elliptical RM of non- oriented SiFe or amorphous ribbons. Several studies of practical local magnetization patterns B(x,y,t) in cores of transformers or rotating machines indicate that these patterns are not representative for practice in many cases, especially modern grain oriented materials showing very specific patterns. However, even in cases where relevant data exists, the application of the latter is complicated due to strong scatter as a result of specific errors of different RSST designs. With respect to magnetostriction, almost not any data are available. Further, it is a drawback of existing experimental studies that they were performed only under "ideal conditions". Industrial practice is characterized by mechanical stress resulting from both thermal effects and core consolidation through clamping. Pre-studies to this project indicate that RM and its consequences are strongly affected by stress. With the above, it is undisputed that RM is a practically relevant phenomenon. RM contributes to losses and noises of machine cores, but to a non-clarified extent. It would be desirable to modify core designs to attain more compact and more silent machines. However, effective considerations of RM are impeded by the fact that most existing RM studies are of rather academic nature and that data banks do not exist. The aim of the present project is to offer a quantitative clarification of the actual practical relevance of RM for the most important types of materials, including also effects of mechanical stress. It will be tried to attain this aim through the following four-step procedure: (A) Study of the occurrence and extent of the phenomena in machine cores - Data on local magnetization patterns and mechanical stress as being typical for industrial practice of both transformers and rotating machines will be collected from publications, technical experience of two industrial project partners and specific experimental investigation of model cores provided by the partners. Local conditions will be analyzed by a hand-held sensor and a scanning chamber developed in earlier FWF projects. (B) Simulation of the phenomena in the lab - RM-patterns which were found to be practically relevant in Step A will be simulated in an RSST developed in an earlier FWF project. Energy losses and multidirectional magnetostriction will be investigated on samples of different types of material (e.g., nonoriented up to domain refined SiFe, Fe-based amorphous ribbons) in stress free and in stressed state. All data will be related to the case of stress-free sinusoidal magnetization in easy direction. (C) Quantitative estimation of percentage contributions to total losses and vibrations of representative cores - It will be performed by allocating the results of simulation (Step B) to the local pattern distributions (Step A), and evaluating the sum loss for the whole core mass (total building factor) and the sum displacements for main core axes, respectively. The results will be discussed with the industrial partners to establish strategies for modifications of the design or assembling of cores. (D) Establishment of a data bank - All loss and magnetostriction data collected for practically relevant conditions for the different types of material (Step B) will be documented and published on a freely accessibe internet website. Other laboratories will be invited to contribute their own corresponding data. According to the schedule, the project will be performed in a three years work by two PhD students funded through the present application.

During the last ten years, much international effort had been taken to study the phenomenon of rotational magnetization (RM) which means that the induction vector B rotates periodically in the plane of laminated soft magnetic material. So-called "rotational single sheet testers" (RSSTs) - developed to produce RM in samples of material - had shown that RM may yield considerably strong increases of energy losses. Further, some rare studies had indicated even higher ones for magnetostriction - a main source of audible noise of machine cores. However, a controversial situation existed with respect to the practical relevance of these phenomena due to several reasons: (i) The existing loss data showed a strong scatter due to applications of very different RSST designs in rather non- systematic ways, not considering practical magnetization conditions. (ii) Systematic magnetostriction studies had been lacking in a complete way. The aim of the present project was to fill these gaps and to establish data sets which can be applied for practical industrial assessments in a direct way. Both textured materials with respect to transformer cores and isotropic materials with respect to rotating machine cores should be investigated. After a 42% cut of budget, it was decided to study the latter field with reduced emphasis, also considering that the textured case would be more challenging. Apart from this restriction, the main targets of planning could be attained in a very satisfying way. In addition, the project yielded novel theoretical cognitions concerning planar eddy current losses. In more detail, the tasks of project yielded the following results and main conclusions: In order to collect data on practically relevant local magnetization conditions of laminated cores, novel experimental techniques were developed comprising a significantly improved thermal sensor system for automatic scanning of loss distributions on core surfaces. It yielded up to more than 100% excess loss in T-joint regions of transformer model cores. For the study of conditions within the core, an extra-thin sheet-like sensor was designed which detects two- dimensional flux density patterns B(t) based on the needle method. It revealed unexpectedly pronounced planar eddy currents as a further source of losses. For transformer cores, the studies indicated the practical relevance of rhombic patterns of low axis ratio - in contrast to most earlier work. For rotating machine cores, a study of both literature and industrial experience confirm the regional existence even of circular patterns. For the simulation of the phenomena in the lab, an a priori existing 3-phase excited RSST was complemented by a control system which allows for exact definition of magnetization patterns B(t) also with respect to the angular velocity of B. As documented in the also established internet data bank, loss measurements on textured core material yielded relative loss increases up to 500% (compared to alternating magnetization) for circular magnetization. For practically relevant patterns as arising in T-joint regions, the increase was up to 100 %. As well, values up to 100% resulted for non-oriented materials. However, the regional RM concentration means that the industrial relevance of RM is lower than expected, rough assessments being offered by means of the here assembled data. Magnetostriction measurements yielded relative strain increases up to 10.000% for textured materials, the practically relevant values being up to 2.000% with slight further increases for mechanical stress. Non-oriented materials showed increases up to about 100%. For the tested transformer core yokes, the over-all strain increase was close to 20% indicating that strong local phenomena are in contrast with rather low global ones. As a significant side result of the project, a thorough study of the so-called needle method for induction measurement showed that its results are strongly affected by planar eddy currents - a finding which is relevant for studies of both laminated cores and RSST samples. On the other hand, the application of double needle pairs promises the effective assessment of planar eddy currents - and the corresponding losses - as a basis of further work.