User:Jalayer masoud/sandbox1

Signal processing based FDI

Machine learning techniques for signal processing based FDI

See also: Machine learning

In signal processing based FDI, mathematical classification models which in fact belong to supervised learning methods are trained on the training set of a labeled dataset to accurately identify the redundancies, faults and anomalous samples. There are numerous different classification and preprocessing models developed and proposed in this research area. k-nearest-neighbors algorithm(kNN) is one of the oldest techniques which have been used to solve FDI problems. Despite the simple logic it has, there are some problems with large dimensionality and processing time in large datasets. Since kNN is not able to automatically extract the features to overcome curse of dimensionality, so often some data preprocessing techniques like Principal component analysis(PCA), Linear discriminant analysis(LDA) or Canonical correlation analysis(CCA) accompany it to reach a better performance. In many industrial cases it is shown that kNN is outperformed by more complex classification models such as Support Vector Machines(SVMs), which is widely used in this field. However SVMs do not have automatic feature extraction themselves and just like kNN, are often coupled with a data pre-processing technique. Another drawback of SVMs is that their performance is highly sensitive to the initial parameters, particularly to the kernel methods, so in each signal dataset a parameter tuning process is required to be conducted firstly.

Artificial Neural Networks(ANNs) are among the most mature and widely used mathematical classification algorithms in signal processing based FDI. ANNs are well-known for their efficient self-learning capabilities of the complex relations (which generally exists inherently in signal processing based FDI problems) and are easy to operate. Another advantage of ANNs is that they perform automatic feature extraction by allocating negligible weights to the irrelevant features, helping the system to avoid dealing with another feature extractor. However, ANNs tend to over-fits the training set, which will have consequences of having poor validation accuracy on validation set. Hence, often, some regularization terms and prior knowledge are added to the ANN model to avoid over-fiting and achieve higher performance. In general, different SVMs and ANNs models (i.e. Back-Propagation Neural Networks and Multi-Layer Perceptron) have shown successful performances in the signal processing based FDI in industries such as gearbox, machinery parts (i.e. mechanical bearings) and steam turbines fault detection.

Deep learning techniques for signal processing based FDI

See also: Deep learning

Recently, by research advances in ANNs and advent of deep learning algorithms, using deep and complex layers, novel classification models are developed to cope with signal processing problems in FDI.

Time domain waveform(top) and CWTS(bottom) of a normal signal

Most of the shallow learning models extract a few feature values from signals, causing a dimensionality reduction from the original signal. While, by using a Convolutional neural network, the continuous wavelet transform scalogram can be directly classified to normal and faulty classes. Such a technique avoids omitting any important fault message and results in a better performance of FDI. Also by transforming signals to image constructions a 2D Convolutional neural network can be implemented for FDI.

Image Processing based FDI

In the past few