Online Incremental Learning for High Voltage Bushing Condition Monitoring

The problem of fault diagnosis of machine has been an ongoing research in various industries. Many machine learning tools have been applied to this problem using static machine learning structures such as neural network and support vector machine that are unable to accommodate new information as it becomes available into their existing models. This paper introduces the incremental learning approach to the problem of condition monitoring. The paper starts by giving a brief definition of incremental learning. Two incremental learning techniques are applied to the problem of condition monitoring. The first method uses the incremental learning ability of Fuzzy ARTMAP (FAM) and explores whether ensemble approach can improve the performance of the FAM. The second technique uses Learn++ that uses an ensemble of MLP classifiers.