An Artificial Neural Network-Based Response Surface Method for Reliability Analyses of c-φ Slopes with Spatially Variable Soil

This paper presents an artificial neural network（ANN）-based response surface method that can be used to predict the failure probability of c-φslopes with spatially variable soil.In this method,the Latin hypercube sampling technique is adopted to generate input datasets for establishing an ANN model;the random finite element method is then utilized to calculate the corresponding output datasets considering the spatial variability of soil properties;and finally,an ANN model is trained to construct the response surface of failure probability and obtain an approximate function that incorporates the relevant variables.The results of the illustrated example indicate that the proposed method provides credible and accurate estimations of failure probability.As a result,the obtained approximate function can be used as an alternative to the specific analysis process in c-φslope reliability analyses.

Hybrid dynamic model of polymer electrolyte membrane fuel cell stack using variable neural network

The polymer electrolyte membrane（PEM） fuel cell has been regarded as a potential alternative power source,and a model is necessary for its design,control and power management.A hybrid dynamic model of PEM fuel cell,which combines the advantages of mechanism model and black-box model,is proposed in this paper.To improve the performance,the static neural network and variable neural network are used to build the black-box model.The static neural network can significantly improve the static performance of the hybrid model,and the variable neural network makes the hybrid dynamic model predict the real PEM fuel cell behavior with required accuracy.Finally,the hybrid dynamic model is validated with a 500 W PEM fuel cell.The static and transient experiment results show that the hybrid dynamic model can predict the behavior of the fuel cell stack accurately and therefore can be effectively utilized in practical application.

VARIABLE STRUCTURE CONTROL APPLIED IN ELECTRO-HYDRAULIC SERVO SYSTEM WITH ANN

The variable structure control （VSC） theory is applied to the electro-hydraulic servo system here. The VSC control law is achieved using Lyapunov method and pole placement. To eliminate the chattering phenomena, a saturation function is adopted. The proposed VSC approach is fairly robust to load disturbance and system parameter variation. Since the distortion. including phase lag and amplitude attenuation occurs in the system sinusoid response, the amplitude and phase control （APC） algorithm, based on Adaline neural network and using LMS algorithm, is developed for distortion cancellation. The APC controller is simple and can on-line adjust, thus it gives accurate tracking.

Transition Features from Simplicity-Universality to Complexity-Diversification Under UHNTF

A large unified hybrid network model with a variable speed growth （LUHNM-VSG） is proposed as third model of the unified hybrid network theoretical framework （UHNTF）. A hybrid growth ratio vg of deterministic linking number to random linking number and variable speed growth index a are introduced in it. The main effects of vg and a on topological transition features of the LUHNM-VSC are revealed. For comparison with the other models, we construct a type of the network complexity pyramid with seven levels, in which from the bottom level-1 to the top level-7 of the pyramid simplicity-universality is increasing but complexity-diversity is decreasing. The transition relations between them depend on matching of four hybrid ratios （dr, fd, gr, vg）. Thus the most of network models can be investigated in the unification way via four hybrid ratios （dr, fd, gr, vg）. The LUHNM-VSG as the level-1 of the pyramid is much better and closer to description of real-world networks as well as has potential application.

Contributions of internal climate variability in driving global and ocean temperature variations using multi-layer perceptron neural network

Roles of internal climate variabilities regulating global and ocean temperature changes is a hot but complex issue of scientific concern,infuencing the comprehensive policy-making in response to global and regional warming.In this study,the time series of monthly global and ocean mean surface temperature(GST and OST,respectively)since 1866 is successflly reconstructed via natural and anthropogenic forcing factors and internal climate variability by using a Multi-Layer Perceptron(MLP)neural network technique.The MLP demonstrates prominent monthly GST and OST reconstruction skills on both interannual and annual time scales.Most of the warming in GST and OST since 1866 is found to be attributable to anthropogenic forcing,while the multidecadal and interannual GST and OST variations are considerably dominated by Atlantic Multidecadal Oscillation(AMO).Internal climate variabilities like Interdecadal Pacific Oscillation(IPO)can amplify the GST and OST changes and explain the global warming slowdown since 1998.Southern Oscillation Index(SOI)performs a similar role as IPO but to a lesser extent.Changes in OST caused by solar forcing are more considerable than those in GST.Moreover,the"biased warmth"during the Second World War is successfully reconstructed in MLP.AMO and IPO can explain most annual and even sub-annual temperature variations during this period,offering an explanation for the existence of this abnormal warm period other than that it was entirely caused by instrumental errors.The generally high accuracy of reconstructions on interannual and annual time scales can enhance the ability to monitor the prompt feedback of specific external radiative forcings and internal variabilities to changes in climate.