Gearbox fault diagnosis based on deep random forest fusion of acoustic and vibratory signals

Vibration signals of gearbox are sensitive to the existence of the fault. Based on vibration signals, this paper presents an implementation of deep learning algorithm convolutional neural network (CNN) used for fault identification and classification in gearboxes. Different combinations of condition patterns based on some basic fault conditions are considered. 20 test cases with different combinations of condition patterns are used, where each test case includes 12 combinations of different basic condition patterns. Vibration signals are preprocessed using statistical measures from the time domain signal such as standard deviation, skewness, and kurtosis. In the frequency domain, the spectrum obtained with FFT is divided into multiple bands, and the root mean square (RMS) value is calculated for each one so the energy maintains its shape at the spectrum peaks. The achieved accuracy indicates that the proposed approach is highly reliable and applicable in fault diagnosis of industrial reciprocating machinery. Comparing with peer algorithms, the present method exhibits the best performance in the gearbox fault diagnosis.

/pdf/gearbox-fault-identification-and-classification-with-1snl9otyt1.pdf

Gearbox Fault Identification and Classification with Convolutional Neural Networks

Multimodal deep support vector classification with homologous features and its application to gearbox fault diagnosis

Intelligent machine health monitoring and fault diagnosis are becoming increasingly important for modern manufacturing industries. Current fault diagnosis approaches mostly depend on expert-designed features for building prediction models. In this paper, we proposed IDSCNN, a novel bearing fault diagnosis algorithm based on ensemble deep convolutional neural networks and an improved Dempster-Shafer theory based evidence fusion. The convolutional neural networks take the root mean square (RMS) maps from the FFT (Fast Fourier Transformation) features of the vibration signals from two sensors as inputs. The improved D-S evidence theory is implemented via distance matrix from evidences and modified Gini Index. Extensive evaluations of the IDSCNN on the Case Western Reserve Dataset showed that our IDSCNN algorithm can achieve better fault diagnosis performance than existing machine learning methods by fusing complementary or conflicting evidences from different models and sensors and adapting to different load conditions.

/pdf/an-ensemble-deep-convolutional-neural-network-model-with-153in8bjpu.pdf

An Ensemble Deep Convolutional Neural Network Model with Improved D-S Evidence Fusion for Bearing Fault Diagnosis.

An oscillatory neural network model of multiple object tracking is described. The model works with a set of identical visual objects moving around the screen. At the initial stage, the model selects into the focus of attention a subset of objects initially marked as targets. Other objects are used as distractors. The model aims to preserve the initial separation between targets and distractors while objects are moving. This is achieved by a proper interplay of synchronizing and desynchronizing interactions in a multilayer network, where each layer is responsible for tracking a single target. The results of the model simulation are presented and compared with experimental data. In agreement with experimental evidence, simulations with a larger number of targets have shown higher error rates. Also, the functioning of the model in the case of temporarily overlapping objects is presented.

http://www.impb.ru/pdf/Kazanovich_Borisyuk_NeralComp06.pdf

An oscillatory neural model of multiple object tracking

We investigate storage capacity and retrieval property for a synchronous fully connected neural network with a non-monotonic transfer function which retrieves sequences of patterns, by an analytic method and also by numerical simulations. Because of asymmetry of interactions and non-monotonicity of the transfer function, it is difficult to use conventional methods of the equilibrium statistical mechanics in order to investigate the network. We then use a generating-function method of path-integral representation, and obtain equations for dynamical order parameters in the stationary state. We clarify that the network with the non-monotonic transfer function retrieves more sequences of patterns than that with a monotonic transfer function at zero temperature when non-monotonicity of the transfer function is selected optimally. It is also clarified that some chaotic behavior appears in solutions for the equations of the dynamical order parameters when non-monotonicity of the transfer function increases. The ...

Sequence Processing Neural Network with a Non-Monotonic Transfer Function

We propose a mathematical model of selective visual attention using a two-layered neural network with neurons described by the Hodgkin---Huxley equation in order to investigate part of the assumption proposed by Desimone and Duncan. The neural network consists of a layer of hippocampal formation and of visual cortex. A frequency of firing and a firing time for each neuron and also a correlation of the firing times between neurons are calculated numerically to clarify an attention state, a nonattention state, and an attention shift. We find that synchronous phenomena occur not only for the frequency but also for the firing time between the neurons in the hippocampal formation and those in a part of the visual cortex in our model. It also turns out that the attention shift is performed quickly in our model.

Neural network model of selective visual attention using Hodgkin---Huxley equation

We investigate storage capacity and generalization ability for two types of fully connected layered neural networks with non-monotonic transfer functions; random patterns are embedded into the networks by a Hebbian learning rule. One of them is a layered network in which a non-monotonic transfer function of even layers is different from that of odd layers. The other is a layered network with intra-layer connections, in which the non-monotonic transfer function of inter-layer is different from that of intra-layer, and inter-layered neurons and intra-layered neurons are updated alternately. We derive recursion relations for order parameters for those layered networks by the signal-to-noise ratio method. We clarify that the storage capacity and the generalization ability for those layered networks are enhanced in comparison with those with a conventional monotonic transfer function when non-monotonicity of the transfer functions is selected optimally. We also point out that some chaotic behavior appears in the order parameters for the layered networks when non-monotonicity of the transfer functions increases.

Layered neural networks with non-monotonic transfer functions

We propose a model for a system with middle temporal neurons and medial superior temporal (MST) neurons by using a three-layered autoencoder. Noise effect is taken into account by using the framework of statistical physics. We define a cost function of the autoencoder, from which a learning rule is derived by a gradient descent method, within a mean-field approximation. We find a pair of values of two noise levels at which a minimum value of the cost function is attained. We investigate response properties of the MST neurons to optical flows for various types of motion at the pair of optimal values of two noise levels. We obtain that the response properties of the MST neurons are similar to those obtained from neurophysiological experiments.

Model of MT and MST areas using an autoencoder

We investigate storage capacity of a fully connected layered neural network with Q(⩾2)-states clock neurons, including Q=∞ (corresponding to oscillatory neurons) and with intra-layer connections, where random Q-values patterns are embedded into the network by the Hebbian learning rule. We assume that both of inter-layered neurons and intra-layered neurons are updated simultaneously. We are able to treat a recurrent network, a feed-forward network and a layered network with the intra-layer connections within the same framework by adjusting a parameter. We analyze an energy function of the network by using the replica method. We clarify that the storage capacity of the layered network with the intra-layer connections is enhanced in comparison with that of the recurrent network and that of the feed-forward network for Q=2,3,4 and 5 at zero temperature; this is due to the competition between the inter-layer connections and the intra-layer connections. For ⩾6, including Q=∞, the storage capacity at zero temperature is maximum for the feed-forward network. We obtain phase diagrams in a load-parameter versus temperature plane for the layered network with the intra-layer connections.

Katsuki Katayama

Papers

Sequence Processing Neural Network with a Non-Monotonic Transfer Function

Neural network model of selective visual attention using Hodgkin---Huxley equation

Layered neural networks with non-monotonic transfer functions

Model of MT and MST areas using an autoencoder

Layered neural network with intra-layer connections using Q-states clock neurons