Long short-term memory neural network based fault detection and isolation for electro-mechanical actuators

The prediction of tropospheric ozone concentrations is vital due to ozone’s passive impacts on atmosphere, people’s health, flora and fauna. However, ozone prediction is a complex process and the wide range of traditional models is incapable to obtain an accurate prediction. “Artificial intelligence”, “machine learning” and “ozone prediction model” search terms in the title, abstract or keywords are involved. Inclusion criteria include subject area (engineering, computer science), English language and being published from 2015. This criterion obtained 156 articles, which were categorized into 4 areas of interest based on the machine learning technique applied. Recently as a result of the rapid development in the technology and the increase in the number of measured data, artificial intelligence techniques have been intensively used in predicting ozone concentration as an alternative to the traditional models. Therefore, the main objective of this study is to investigate the most developed techniques that have been used in predicting ozone concentrations as well as theoretic approaches such as information set approaches, fuzzy set approach and probabilistic set approaches. It is clearly stated that the standalone algorithms such as decision tree (DT) and support vector machine (SVM) outperformed multilayer perceptron (MLP); however, the latter is massively implemented by many researchers in the prediction of ozone concentrations. This review paper investigated artificial intelligence techniques integrated with optimization approaches. It can be concluded that hybrid algorithms have significantly improved the prediction accuracy. However, the majority of the proposed hybrid models have limitations; thus, there is a need to develop better hybrid algorithm that is able to tackle all the drawbacks of the improved algorithms and capable to capture the ozone concentration changes with a high level of accuracy.

Ozone Concentration Forecasting Based on Artificial Intelligence Techniques: A Systematic Review

https://repositorio.ufjf.br/jspui/bitstream/ufjf/6034/1/renanpiazzarolifinottiamaral.pdf

Type-1 and singleton fuzzy logic system trained by a fast scaled conjugate gradient methods for dealing with binary classification problems

We carried out a simultaneous study of ground-based magnetic field and lower ionospheric anomalies during major earthquakes occurring around Japan in 2010 and 2012. Ultra Low Frequency (ULF) geomagnetic field waveforms of Esashi station and Very Low Frequency (VLF) Japanese transmitter (JJY) electric signal amplitude received in Moshiri station Hokkaido during nighttime (22:00-02:00 LT) were used to minimize the local interference. Twenty earthquakes having magnitude greater than 5.5 were considered for the data analysis for two years. Nighttime amplitude fluctuations and polarization from the received VLF transmitter signal amplitude and ULF magnetic field respectively were calculated to identify anomalous signatures in relation to every earthquake. We found most earthquakes analyzed indicating VLF amplitude anomalies simultaneously occurred with ULF magnetic field anomalies within a week prior to the earthquakes. Stronger anomalies were observed for larger magnitude and shallower earthquakes. Focal mechanism of earthquakes was also examined to identify the effectiveness of generating anomaly. Both VLF and ULF anomalies were observed for reverse fault type earthquakes occurring under the strong pressure in the crust. Obtained results may indicate the common anomaly source both for VLF and ULF in the lithosphere and are consistent with currently proposed Lithosphere-Atmosphere-Ionosphere (LAI) coupling scenarios during the earthquake preparation period.

Simultaneous Study of VLF/ULF Anomalies Associated with Earthquakes in Japan

The electromagnetic field amplitude of the subionospheric Very Low Frequency (VLF) propagation is sensitive to the lower ionospheric conditions. Accordingly, VLF waves have been proposed to study and monitor the lower ionosphere (D/E region). In this paper, the NARXNN (Nonlinear Autoregressive with Exogenous Input Neural Network) is used as a method for predicting the daily nighttime mean amplitude of VLF transmitter signals indicating the ionospheric perturbation around the transmitter-receiver path. The NARXNN has a good accuracy in predicting time series data and thus are more suitable for dynamic modeling. The NARX constructed model, which was built based on daily input variables of various physical parameters with the time interval from 1 January 2011 to 4 February 2013 such as stratospheric and mesospheric temperatures, cosmic rays, total column ozone, F10.7, Kp, AE, and Dst indices. We used the constructed model to predict high-(NLK-CHF), middle-(NPM-CHF) and low-latitude (NWC-CHF) paths. As a result, the constructed models are capable of performing reasonably good 5-day ahead predictions of the daily nighttime of VLF electric field amplitude for NPM-CHF path with the Pearson correlation coefficient (r) of 0.84 and with Root Mean Square Error (RMSE) of 3.12 dB, NLK-CHF (r = 0.80, RMSE = 3.57 dB) and NWC-CHF (r = 0.79, RMSE = 2.60 dB). We conclude that the constructed NARX NN model is capable of predicting the VLF electric field amplitude variation for different latitude paths.

Multi step ahead prediction of nighttime VLF amplitude signal for low-, mid-and high-latitude paths

The electric field amplitude of very low frequency (VLF) transmitter from Hawaii (NPM) has been continuously recorded at Chofu (CHF), Tokyo, Japan. The VLF amplitude variability indicates lower ionospheric perturbation in the D region (60–90 km altitude range) around the NPM-CHF propagation path. We carried out the prediction of daily nighttime mean VLF amplitude by using Nonlinear Autoregressive with Exogenous Input Neural Network (NARX NN). The NARX NN model, which was built based on the daily input variables of various physical parameters such as stratospheric temperature, total column ozone, cosmic rays, Dst, and Kp indices possess good accuracy during the model building. The fitted model was constructed within the training period from 1 January 2011 to 4 February 2013 by using three algorithms, namely, Bayesian Neural Network (BRANN), Levenberg Marquardt Neural Network (LMANN), and Scaled Conjugate Gradient (SCG). The LMANN has the largest Pearson correlation coefficient (r) of 0.94 and smallest root-mean-square error (RMSE) of 1.19 dB. The constructed models by using LMANN were applied to predict the VLF amplitude from 5 February 2013 to 31 December 2013. As a result the one step (1 day) ahead predicted nighttime VLF amplitude has the r of 0.93 and RMSE of 2.25 dB. We conclude that the model built according to the proposed methodology provides good predictions of the electric field amplitude of VLF waves for NPM-CHF (midlatitude) propagation path.

https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1002/2016RS006022

One day prediction of nighttime VLF amplitudes using nonlinear autoregression and neural network modeling

Very low frequency (VLF) waves have been used as a powerful tool to monitor and study the lower ionosphere (D/E region). In this paper, nonlinear physical processes of VLF signals propagation can be well represented by nonlinear autoregressive with exogenous input neural network (NARXNN) model. Further, a study of NARXNN model to predict the daily nighttime mean amplitude of VLF propagation wave to recognize the ionospheric perturbation along the great circle path. The NARXNN model is powerful in predicting time series data and suitable representations of a variation of nonlinear models. The daily input variables of various physical parameters with the time interval from 15 March 2014 to 26 May 2016 were used to build prediction model. The results of the built models are performing reasonably good for one-step ahead (OSA) predictions of the daily nighttime of VLF electric field amplitude. The NARXNN model has good performance for predicting the VLF amplitude variation for different latitude paths.

Prediction of VLF Sub-Ionospheric Wave Propagation Using Nonlinear System Identification

The modeling and numerical analysis of seismo-electromagnetics perturbation are approached using finite-difference time-domain (FDTD) method. The research related seismo-electromagnetics perturbation recently has been implemented for earthquakes (EQs) precursor analysis by utilizing very low frequency (VLF) wave propagation. Here, the proposed method in two-dimensional (2D) form is applied to compute electromagnetics fields of the observed VLF wave and to investigate the phenomena of seismo-electromagnetics perturbation by modeling the geographical maps in two dimensions. To truncate the region of computation, the boundary condition of split-field perfectly matched layer (PML) is implemented. The parameters of investigation including the change of amplitude, frequency and time-shifting in the VLF wave are interpreted in term of changing the ionosphere layer which is confirmed as seismo-electromagnetics perturbation. From numerical result, it shows that the approaching technique using FDTD method is successfully to model the correlation between VLF wave propagation and perturbation wave associated with earthquake for different location. Moreover, some results related to time-shifting of VLF wave and frequency-influence of perturbation wave are also presented.

http://www.ijeei.org/docs-71660999653cde5d0328e9.pdf

Hendy Santosa

Papers

One day prediction of nighttime VLF amplitudes using nonlinear autoregression and neural network modeling

Prediction of VLF Sub-Ionospheric Wave Propagation Using Nonlinear System Identification

Modeling and Numerical Analysis of Seismo-Electromagnetics Perturbation Using FDTD Method