Hot-Spot Temperature Forecasting of the Instrument Transformer Using an Artificial Neural Network

Abstract

Cast resin medium voltage instrument transformer are highly used because of several benefits over other type of transformers. Nevertheless, the high operating temperatures affects their performance and durability. It is important to forecast the hot spots in the transformer. The aim of this study is to develop a model based on Artificial Neural Networks (ANN) theory to be able to forecast the temperature in seven points, taking into account twenty-six input data of transformer design features. 792 simulations were carried out in COMSOL Multiphysics®to emulate the heat transfer in the transformer. The data obtained were used to train 1110 ANN with different number of neurons and hidden layers. The ANN with the best performance (R = 1, MSE=0.003455) has three hidden layers with 10, 9 and 9 neurons respectively. The ANN predictions were validated with finite element simulations and laboratory thermal tests which present similar patterns. With this accuracy in the prediction of hot-spot temperature, this ANN can be used to optimize the design of instrument transformers. INDEX TERMS Artificial neural networks, resin-cast instrument transformer, epoxy resins, finite element analysis, hot-spot temperature

Inversion Detection of Transformer Transient Hot Spot Temperature

“…ɸ is the activation function of hidden layer, and the sigmoid activation function is used. Ψ is the linear activation function of output layer, and y1 is the final output value [29], [30].…”

Forecasting thermal parameters for ultra‐high voltage transformers using long‐ and short‐term time‐series network with conditional mutual information

“…In order to demonstrate the superiority of the proposed method more clearly in short‐term prediction, the CMI method is applied to LSTNet and typical existing methods CNN‐LSTM [36], PSO‐SVR [21], LSTM [45] and ANN [23] models, respectively, in comparative experiments. The one‐ and multi‐step prediction of OT west in one to five steps is carried out, that is, to predict change tendency in the next 1–5 days.…”

“…At present, several methods have been so far proposed to predict winding and top oil temperature in ordinary transformers [20][21][22]. The authors in [23] simulated the heat transfer process inside the transformer through 792 groups of experiments and established the prediction model of the hot spot temperature based on the artificial neural network (ANN) theory. Qi et al [24] introduced the top oil temperature of the Susa thermal circuit model by the nuclear extreme learning machine (ELM) to reduce the forecast error.…”

Data Driven Transformer Thermal Model for Condition Monitoring

“…A power transformer thermal model that was developed using artificial neural network (ANN) was also more accurate compared with the traditional approach [11]. In addition, the ANN was used to forecast the hot-spot temperature for an instrument transformer [12]. Performance of the prediction made by binary regression tree, generalized linear model, gaussian process regression and support vector machine over three units lightly loaded was investigated [13].…”