Avoiding Spiraling Effects on Helical Windings with Appropriate Dynamic Axial Short-Circuit-Force Calculation and Evaluation Methods

The classical axial winding force calculation technique (static approach) determines winding end thrust and axial forces by accumulating the electromagnetic Lorentz-forces [2] neglecting the mechanical properties of the winding block (e.g., conductor mass, spacerbord and paper insulation stiffness) as well as the external winding clamping forces. Friction forces calculated using this pure static approach for evaluating the spiraling withstand capability do not correlate by this design approach with data and results from passed and failed short-circuit laboratory tests on full scale power transformers.

Also in the IEC standard 60076-5 Annex A [3], the spiraling force (or so-called thrust force Tf on a winding lead exit) is only addressed through an absolute comparison with a spiraling force of a winding from a reference transformer that has already been successfully tested. With this comparative approach the spiraling force is calculated only statically without considering friction forces generated by the external clamping force and the dynamic axial short-circuit forces, winding resonances and vibrations as well as any mechanically lead-exit support measures. Especially in winding assemblies with axial gaps between disks and layers, a pure static approach will not cover all the axial winding and friction forces and may differ from design to design. Therefore, any comparison with short-circuit tested reference transformers have extremely limited evaluation value for the strength of the winding against spiraling.

To improve this situation, this paper will provide novel evaluation methods for predicting spiraling effects in layer and helical windings of core-type oil-immersed medium- and large power transformers (MPT and LPT).

A mechanical dynamic vibration model was developed and integrated into the manufacturer’s design process to better determine the friction between winding turns and predict and avoid winding spiraling effects.

Considering the computation speed and model usability in an everyday short-circuit design process, it is vital to keep the simulation model as simple as feasible, while not ignoring essential mechanical effects. For this purpose, a simple spring-mass model (dynamic approach) which considers the axial component of the exciting electromagnetic Lorentz-forces is sufficient [4]. The simulation model includes the most relevant mass and spring elements of the transformer’s active part from a mechanical vibration perspective, such as all windings of one phase, the clamping ring or segment-chain (on the top of the winding block), the mounting plate and winding table (on the bottom of the winding block), the clamping plates, the flitch plates or tie rods, and the stacked magnetic core yokes.

The nonlinear behavior ( [5], [6], [7]) of the winding insulation material (insulation paper, pressboard rings, and distance spacers made of transformer-board) having viscoelastic material properties [8]), represented by nonlinear springs in the model, complicates solving the dynamic vibration model.

The stiffness of the materials varies with clamping pressure, making it difficult to determine their viscoelastic behavior. System damping is another challenging parameter to define within the dynamic vibration model.

During a short-circuit incident, strong resonant conductor vibrations are dampened by a variety of mechanisms. These effects include conductor movement in insulation fluids, friction between insulation materials and radial supports. During the first short-circuit cycles, oil in impregnated spacers is squeezed out of the material, altering its stiffness. Furthermore, it is difficult to identify the damping effects of the surrounding fluid flow. For these reasons, a simplified logarithmic damping technique [9] was successfully implemented.