Daghrah, M, Zhang, X ORCID: https://orcid.org/0000-0002-8790-0313, Wang, Z, Liu, Q, Jarman, P and Walker, D (2019) Flow and temperature distributions in a disc type winding-part I: Forced and directed cooling modes. Applied Thermal Engineering, 165. ISSN 1359-4311
|
Published Version
Available under License Creative Commons Attribution. Download (2MB) | Preview |
Abstract
© 2019 The Authors The thermal design of a power transformer determines its lifetime and loading capability, so correct modelling of liquid flow and temperature distributions in the winding is of vital importance. Existing standards and widely used simple calculation models can seriously under/overestimate winding hotspot temperatures under certain circumstances. In this paper, liquid flow and temperature distributions in a physical model representing a disc-type winding in a liquid forced and directed cooling mode are investigated experimentally using Particle Image Velocimetry (PIV) and a temperature measurement system and numerically using Computational Fluid Dynamics (CFD). Dimensional analysis is used to generalise the results into a form useful for design review. The operating conditions investigated include different liquid inlet velocities, inlet temperatures, power losses in individual disc segments, and the effect of alternative liquids. It is shown that hotspot temperature and position within the winding are a non-linear function of liquid inlet velocity, with stagnation and reverse flow demonstrated in both experiments and CFD models. Comparisons of liquid flow and temperature distributions between measurements and CFD simulation results show that 2D CFD results are representative when there are no reverse flows and 3D CFD simulations are needed when reverse flows occur. The results are presented first in dimensional forms to show the effect of each parameter, and then in non-dimensional forms to provide a generalised insight into transformer thermal design.
Impact and Reach
Statistics
Additional statistics for this dataset are available via IRStats2.