Assessment of a hydraulic network model for zig–zag cooled power transformer windings

Abstract

The prediction of mass flow distribution is a first, though crucial, step in the thermal design of zig-zag cooled power transformer windings. Typically this prediction is based on thermo-hydraulic network models, which critically depend on the applied correlations. In this paper new correlations for flow through elbows and for dividing/merging flow in T-junctions are numerically extracted from dedicated CFD studies and applied to a case study. It is shown that the new correlations are superior for the prediction of the mass flow distribution for this test case in comparison to correlations thus far available in literature.

FluSHELL – A Tool for Thermal Modelling and Simulation of Windings for Large Shell-Type Power Transformers

“…However, in the FluSHELL tool these temperature profiles are approximated by lumped coefficients: the friction coefficient, f , identified in Equation (4.9), represents the fluid velocity profile and the heat transfer coefficient, Therefore, FluSHELL accuracy is dependent on these coefficients and hence an adequate methodology to obtain such coefficients is paramount. This has been the main objective of recent works that used CFD to obtain improved friction and heat transfer correlations for the windings of typical core-type transformers (Campelo et al, 2012;Coddé et al, 2015;. In two of these works the authors conducted sets of CFD simulations in a theoretical T-Junction configuration in order to obtain improved correlations for the frictional losses and heat transfer coefficients .…”

“…CFD simulations usually involve a much finer spatial discretization and enable the computation of the governing differential equations without 118 simplifications. However, its direct integration in the design-cycles is not envisaged in the mid-term (Cigre, 2016) and CFD is currently regarded as a virtual R&D laboratory used to extract relevant information to validate or to include in simpler modelling approaches (Campelo et al, 2012;Coddé et al, 2015;.…”

Thermal degradation assessment of Kraft paper in power transformers insulated with natural esters

“…So far, there has been a lot of research about cooling improvement in power transformers using mineral oil [21][22][23][24]. The reason is that hottest spot temperature inside windings of the oilimmersed power transformers is one of the main manifestations of the thermal stress which leads to aging and decomposition of both liquid and solid insulation material (oil and cellulose) more rapidly as compared with less heated spots [25][26][27].…”

Prediction of Pressure Drop and Flow Distribution in Disc-Type Transformer Windings in an OD Cooling Mode

“…It is probably worth mentioning that the correlation equations proposed in [9] were partially verified by measuring the pressure drop over one pass of a test rig; however flow distributions obtained from the calibrated network model were significantly different from those obtained from 2D CFD simulations, especially for the cases of high oil flow rates [11]. The correlation equations for minor pressure losses of flow over elbows, in dividing, and merging T-junctions have been greatly improved in [10]. However, the discrepancy between flow distributions in a winding pass from a CFD simulation and a network model with the newly developed minor loss correlation equations is still not negligible (figure 8 in [10]).…”

“…They are usually obtained from literature pertaining to circular pipe flow and 2D duct heat transfer processes. The applicability of these traditional correlation equations proved to be unsatisfactory when compared with the performance of newly established correlation equations derived from computational fluid dynamics (CFD) models [8][9][10]. It is probably worth mentioning that the correlation equations proposed in [9] were partially verified by measuring the pressure drop over one pass of a test rig; however flow distributions obtained from the calibrated network model were significantly different from those obtained from 2D CFD simulations, especially for the cases of high oil flow rates [11].…”