•Damage detection in structures by applying using artificial neural network (ANN) and cuckoo search (CS) algorithm.•Structural damage localization and quantification in bridges and beam-like ...structures.•Improve ANN training parameters using CS.•ANN combined with CS (ANN-CS) is accurate and cost effective.
This paper presents a new approach for damage detection in structures by applying a flexible combination based on an artificial neural network (ANN) and cuckoo search (CS) algorithm. ANN has become one of the most powerful tools employing computational intelligence techniques to tackle complex problems in numerous fields. However, due to the application of backpropagation algorithms based on gradient descent, a major drawback of ANN is the common problem of local minima that acts as a great hindrance to the search for the best solution. To overcome this disadvantage, we propose to combine ANN with evolutionary algorithms based on global search techniques. This paper employs CS to improve ANN training parameters (weight and bias) by minimizing the difference between real and desired outputs and then using these parameters to generate the network. Two numerical models, comprising a steel beam calibrated using experimental measurements and a large-scale truss bridge, are used to assess the robustness of the proposed approach. The results demonstrate that ANN combined with CS (ANN-CS) is accurate and requires a lower computational time than ANN, and evolutionary algorithm (EA) alone in terms of structural damage localization and quantification.
The antenna array synthesis problem has long been known as a tough issue, which attracts considerable interests to explore high-performance low-complexity optimization techniques. In this paper, an ...efficient artificial neural network (ANN) for time-modulated arrays (TMA) synthesis is proposed. By defining the equivalent excitation properly, TMA synthesis can be first transformed to a generalized phased array optimization. Next, a two-stage ANN framework composed of two encoders and a universal decoder is established to optimize the equivalent excitation, and then a single-input single-output (SISO) sinc -1 -ANN is proposed to solve inverse of sinc(·) efficiently. To achieve fast and accurate pattern prediction, the decoder is pre-trained to be a real-time array analyzer, while the encoder is designed as an array synthesizer to develop online training. By minimizing the loss function related to radiation pattern and equivalent excitation, the desired pattern can be achieved. Then, with the help of the pre-trained SISO sinc -1 -ANN, the static excitation coefficient, switch-on duration and starting time are acquired. Simulation results of different types of desired TMA patterns are provided to verify the superiority, effectiveness and efficiency of the proposed approach.
There has been an increasing interest in using model predictive control (MPC) for power electronic applications. However, the exponential increase in computational complexity and demand of computing ...resources hinders the practical adoption of this highly promising control technique. In this article, a new MPC approach using an artificial neural network (termed ANN-MPC) is proposed to overcome these barriers. A power converter with a virtual MPC controller is first designed and operated under a circuit simulation or power hardware-in-the-loop simulation environment. An artificial neural network (ANN) is then trained offline with the input and output data of the virtual MPC controller. Next, an actual FPGA-based MPC controller is designed using the trained ANN instead of relying on heavy-duty mathematical computation to control the actual operation of the power converter in real time. The ANN-MPC approach can significantly reduce the computing need and allow the use of more accurate high-order system models due to the simple mathematical expression of ANN. Furthermore, the ANN-MPC approach can retain the robustness for system parameter uncertainties by flexibly setting the input elements. The basic concept, ANN structure, offline training method, and online operation of ANN-MPC are described in detail. The computing resource requirement of the ANN-MPC and conventional MPC are analyzed and compared. The ANN-MPC concept is validated by both simulation and experimental results on two kW-class flying capacitor multilevel converters. It is demonstrated that the FPGA-based ANN-MPC controller can significantly reduce the FPGA resource requirement (e.g., 2.11 times fewer slice LUTs and 2.06 times fewer DSPs) while offering a control performance same as the conventional MPC.
Wind speed forecasting is critical for wind energy conversion systems since it greatly influences the issues such as the scheduling of a power system, and the dynamic control of the wind turbine. In ...this paper, we present a comprehensive comparison study on the application of different artificial neural networks in 1-h-ahead wind speed forecasting. Three types of typical neural networks, namely, adaptive linear element, back propagation, and radial basis function, are investigated. The wind data used are the hourly mean wind speed collected at two observation sites in North Dakota. The performance is evaluated based on three metrics, namely, mean absolute error, root mean square error, and mean absolute percentage error. The results show that even for the same wind dataset, no single neural network model outperforms others universally in terms of all evaluation metrics. Moreover, the selection of the type of neural networks for best performance is also dependent upon the data sources. Among the optimal models obtained, the relative difference in terms of one particular evaluation metric can be as much as 20%. This indicates the need of generating a single robust and reliable forecast by applying a post-processing method.
The artificial neural network (ANN)-based compact modeling methodology is evaluated in the context of advanced field-effect transistor (FET) modeling for Design-Technology-Cooptimization (DTCO) and ...pathfinding activities. An ANN model architecture for FETs is introduced, and the results clearly show that by carefully choosing the conversion functions (i.e., from ANN outputs to device terminal currents or charges) and the loss functions for ANN training, ANN models can reproduce the current-voltage and charge-voltage characteristics of advanced FETs with excellent accuracy. A few key techniques are introduced in this work to enhance the capabilities of ANN models (e.g., model retargeting, variability modeling) and to improve ANN training efficiency and SPICE simulation turn-around-time (TAT). A systematical study on the impact of the ANN size on ANN model accuracy and SPICE simulation TAT is conducted, and an automated flow for generating optimum ANN models is proposed. The findings in this work suggest that the ANN-based methodology can be a promising compact modeling solution for advanced DTCO and pathfinding activities.
Approximation Attacks on Strong PUFs Shi, Junye; Lu, Yang; Zhang, Jiliang
IEEE transactions on computer-aided design of integrated circuits and systems,
10/2020, Letnik:
39, Številka:
10
Journal Article
Recenzirano
Physical unclonable function (PUF) is a promising lightweight hardware security primitive for resource-constrained systems. It can generate a large number of challenge-response pairs (CRPs) for ...device authentication based on process variations. However, attackers can collect the CRPs to build a machine learning (ML) model with high prediction accuracy for the PUF. Recently, a lot of ML-resistant PUF structures have been proposed, e.g., a multiplexer-based PUF (MPUF) was introduced to resist ML attacks and its two variants (rMPUF and cMPUF) were further proposed to resist reliability-based and cryptanalysis modeling attacks, respectively. In this article, we propose a general framework for ML attacks on strong PUFs, then based on the framework, we present two novel modeling attacks, named logical approximation and global approximation, that use artificial neural network (ANN) to characterize the nonlinear structure of MPUF, rMPUF, cMPUF, and XOR Arbiter PUF. The logical approximation method uses linear functions to approximate logical operations and builds a precise soft model based on the combination of logical gates in the PUF. The global approximation method uses the function sinc with filtering characteristics to fit the mapping relationship between the challenge and response. The experimental results show that the proposed two approximation attacks can successfully model the (<inline-formula> <tex-math notation="LaTeX">n </tex-math></inline-formula>, <inline-formula> <tex-math notation="LaTeX">k </tex-math></inline-formula>)-MPUF (<inline-formula> <tex-math notation="LaTeX">k= 3, 4 </tex-math></inline-formula>), (<inline-formula> <tex-math notation="LaTeX">n </tex-math></inline-formula>, <inline-formula> <tex-math notation="LaTeX">k </tex-math></inline-formula>)-rMPUF (<inline-formula> <tex-math notation="LaTeX">k = 2, 3 </tex-math></inline-formula>), cMPUF (<inline-formula> <tex-math notation="LaTeX">k = 4, 5 </tex-math></inline-formula>), and <inline-formula> <tex-math notation="LaTeX">l </tex-math></inline-formula>-XOR Arbiter PUF (<inline-formula> <tex-math notation="LaTeX">l= 3, 4, 5 </tex-math></inline-formula>) (<inline-formula> <tex-math notation="LaTeX">n = 32, 64 </tex-math></inline-formula>) with the average accuracies of 96.85%, 95.33%, 94.52%, and 96.26%, respectively.
Optoelectronic synaptic devices that mimic biological synapses are critical building blocks of artificial neural networks (ANN) based on optoelectronic integration. Here it is shown that an ...optoelectronic synaptic device based on the hybrid structure of silicon nanocrystals (Si NCs) and poly(3‐hexylthiophene) (P3HT) can work with dual modes, exhibiting versatile synaptic plasticity. In the three‐terminal mode, the device is a synaptic transistor, which has wavelength‐selective synaptic plasticity due to potential wells enabled by the Si NCs/P3HT hybrid structure. In the two‐terminal mode, it is a synaptic metal‐oxide‐semiconductor (MOS) device, which is capable of mimicking spike‐rate‐dependent plasticity (SRDP) and metaplasticity with optical stimulation. Based on the wavelength‐selective synaptic plasticity a light‐stimulated ANN is proposed to recognize handwritten digits with an accuracy of around 90.4%. In addition, the SRDP and metaplasticity may be well used for the simulation of edge detection of images, facilitating real‐time image processing.
An optoelectronic synaptic device based on the hybrid structure of silicon nanocrystals (Si NCs) and poly(3‐hexylthiophene) (P3HT) can work with dual modes. In the three‐terminal mode, the device shows the wavelength‐selective synaptic plasticity, which is applied to a light‐stimulated artificial neural network. In the two‐terminal mode, the device simulates the spike‐rate‐dependent plasticity and metaplasticity, which may be used for edge detection.
The use of artificial neural networks (ANNs) for the selection of weighting factors in cost function of the finite-set model-predictive control (FS-MPC) algorithm can speed up selection without ...imposing additional computational burden to the algorithm and ensure that optimum weights are selected for the specific application. In this article, the ANN-based design process of the weighting factors is used for predictive torque control (PTC) in a motor drive. In the design process, the weighting factors in the cost function and the reference flux value are obtained using different fitness functions. The results show that different operating conditions of the drive will have new optimum parameters of the cost function; therefore, sweeping parameters like load torque or reference speed can optimize the PTC for the whole operating range of the drive. A good match of the performance metrics predicted by the ANN and the simulation model is also observed. The experiments demonstrate that the selected cost function parameters can provide a fast drive start and good performance during different loading conditions and also in reversing of the drive.
•The use of machine learning technology in biodiesel research is thoroughly reviewed.•The artificial neural network approach is the most popular machine learning tool in this field.•The main goal of ...using machine learning technology has been to model biodiesel systems.•The pros and cons of the use of machine learning methods in biodiesel research are outlined.•Machine learning technology has the potential to monitor/control biodiesel systems in real-time.
Biodiesel has the potential to significantly contribute to making transportation fuels more sustainable. Due to the complexity and nonlinearity of processes for biodiesel production and use, fast and accurate modeling tools are required for their design, optimization, monitoring, and control. Data-driven machine learning (ML) techniques have demonstrated superior predictive capability compared to conventional methods for modeling such highly complex processes. Among the available ML techniques, the artificial neural network (ANN) technology is the most widely used approach in biodiesel research. The ANN approach is a computational learning method that mimics the human brain's neurological processing ability to map input-output relationships of ill-defined systems. Given its high generalization capacity, ANN has gained popularity in dealing with complex nonlinear real-world engineering and scientific problems. This paper is devoted to thoroughly reviewing and critically discussing various ML technology applications, with a particular focus on ANN, to solve function approximation, optimization, monitoring, and control problems in biodiesel research. Moreover, the advantages and disadvantages of using ML technology in biodiesel research are highlighted to direct future R&D efforts in this domain. ML technology has generally been used in biodiesel research for modeling (trans)esterification processes, physico-chemical characteristics of biodiesel, and biodiesel-fueled internal combustion engines. The primary purpose of introducing ML technology to the biodiesel industry has been to monitor and control biodiesel systems in real-time; however, these issues have rarely been explored in the literature. Therefore, future studies appear to be directed towards the use of ML techniques for real-time process monitoring and control of biodiesel systems to enhance production efficiency, economic viability, and environmental sustainability.
Display omitted
Summary
In this study, five different water based ZrO2 nanofluids were prepared at volumetric concentrations of 0.0125%, 0.025%, 0.05%, 0.1%, and 0.2%. In the preparation of nanofluids, two‐step ...method was preferred, magnetic stirrer and ultrasonic homogenizer were used. Their thermal conductivity was measured experimentally in the temperature range of 10°C to 65°C. Using the obtained experimental data, a multi‐layer perceptron feed‐forward back‐propagation artificial neural network was developed. In addition, a new correlation was proposed for the calculation of the thermal conductivity values of the ZrO2/Water nanofluid. The results showed that the ZrO2/Water nanofluid had higher thermal conductivity compared to the base fluid and the thermal conductivity increases with the increase in temperature and concentration. While the artificial neural network developed with experimental data predicted the thermal conductivity of ZrO2/Water nanofluid with an average error of −0.41%, the new correlation developed predicted it with an average error of −0.02%. These values were an indication that the results obtained from the developed artificial neural network and the correlation are in perfect agreement with the experimental data.
The thermal conductivity of the water‐based ZrO2 nanofluid, prepared at five different volumetric concentrations have experimentally measured between 10 ‐ 65 °C. Using experimental data, an artificial neural network and a new mathematical correlation have been developed. The outputs obtained from the artificial neural network and mathematical correlation have compared with the experimental results.
PDF
