Thermal Lifetime Evaluation of Electrical Machines Using Neural Network

Abstract

This paper proposes a surrogate approach which utilises an supervised neural network to significantly shorten the time required for thermal qualification of electrical machines' insulation. The proposed approach is based on a feedforward neural network trained with Bayesian Regularization Back-Propagation (BRP) algorithm. The network predicts the winding's insulation resistance trend with respect to its thermal aging time. The predicted insulation resistance is evaluated against experimental measurements and an excellent match is found. Its trend is used for estimating the sample's time to failure under thermal stress at various temperatures. The temperature index of the insulating material, predicted by the neural network, matches with an error of just 0.4% margin against the experimental findings.

Lifetime Estimation of Enameled Wires Under Accelerated Thermal Aging Using Curve Fitting Methods

“…Curve Fitting models estimate the points anywhere along the curve, and can be utilized for estimating the lifetime of an insulation system by predicting the IR, followed by the time-to-failure associated with each sample. This section describes three different CF models which have been developed for predicting the lifetime of low voltage enameled wire at thermal exposures of 250 • C, 270 • C and 290 • C, which is then extrapolated under normal stress levels through least mean square method [6], [7]. The details of CF models used to predict the lifetime are described below.…”

“…The Logarithmic function fits a curve through a given dataset in the form of (7). For this case, if the coefficient m is negative, the function represents logarithmic decline whereas, if the coefficient m is positive, the function represents a logarithmic growth or increase [14].…”

“…But, the major concern with the standard method [3], [4] of thermal lifetime evaluation is the amount of time required to estimate the lifetime of electrical machines which can take up to thousands of testing hours, lasting several months [5]. Therefore, fast and accurate tools are required which are capable of estimating motor insulation lifetime quickly and precisely, thus, enabling faster manufacturing of newly designed electrical machines [6], [7].…”

“…The IR of electrical equipment was predicted for early detection of potential failures, while variations in ambient humidity and air temperature were taken into consideration. In [6], [7], authors propose an alternative approach relying on a supervised NN that can significantly reduce the time required for lifetime evaluation under thermal aging tests. The supervised NN is based on a feedforward NN trained with Bayesian Regularisation Backpropagation (BRP) algorithm.…”

“…This is mainly due to the intrinsic characteristics of the NN technique, such as its 1) ability to automatically learn the relationship between inputs and outputs; 2) size independency and complexity of the problem; 3) generalisation capacity; and 4) integration facility with other computational tools. Therefore, this paper proposes a surrogate approach that estimates the lifetime of electrical insulation systems using CF models which is then compared with NN approach proposed in [6], [7]. Three CF models (i.e.…”