Search

As the complexity and scope of VLSI designs continue to grow, fault detection processes in the pre-silicon stage have become crucial to guaranteeing reliability in IC design. Most fault detection algorithms can be solved by transforming them into a satisfiability (SAT) problem decipherable by SAT solvers. However, SAT solvers consume significant computational time, as a result of the search space explosion problem. This ever- increasing amount of data can be handled via machine learning techniques known as deep learning algorithms. In this paper, we propose a new approach utilizing deep learning for fault detection (FD) of combinational and sequential circuits in a type of stuck-at-faults. The goal of the proposed semi-supervised FD model is to avoid the search space explosion problem by taking advantage of unsupervised and supervised learning processes. First, the unsupervised learning process attempts to extract underlying concepts of data using Deep sparse autoencoder. Then, the supervised process tends to describe rules of classification that are applied to the reduced features for detecting different stuck-at faults within circuits. The FD model proposes good performance in terms of running time about 187 × compared to other FD algorithm based on SAT solvers. In addition, it is compared to common classical machine learning models such as Decision Tree (DT), Random Forest (RF) and Gradient Boosting (GB) classifiers, in terms of validation accuracy. The results show a maximum validation accuracy of the feature extraction process at 99.93%, using Deep sparse autoencoder for combinational circuits. For sequential circuits, stacked sparse autoencoder presents 99.95% as average validation accuracy. The fault detection process delivers around 99.6% maximum validation accuracy for combinational circuits from ISCAS'85 and 99.8% for sequential circuits from ISCAS'89 benchmarks. Moreover, the proposed FD model has achieved a running time of about 1.7x, compared to DT classifier and around 1.6x, compared to RF classifier and GB machine learning classifiers, in terms of validation accuracy in detecting faults occurred in eight different digital circuits. Furthermore, the proposed model outperforms other FD models, based on Radial Basis Function Network (RBFN), achieving 97.8% maximum validation accuracy. [ABSTRACT FROM AUTHOR]

Wind energy has proven to be the highest reliable source of renewables due to the maturity of the technology. Wind turbine (WT) systems are complex systems with undergoing development process that requires innovative methods of design and control. In this paper, WTs are studied through hybrid systems framework. Hybrid models of WT are extracted with representable dynamics from nonlinear complex design code. The WT model is formulated into a mixed-logical dynamical model (MLD) and a piecewise affine (PWA) model. Then, a receding horizon control strategy is applied to WT hybrid model resulting in a hybrid model predictive control (HMPC) with mixed-integer programming (MIP) problem. The performance of the proposed controller is compared against the baseline controller within a simulation environment for the National Renewable Energy Laboratory (NREL) 5MW benchmark WT as a case study. The analysis and investigation of HMPC highlight its capability as a potential tool for exploiting the control objectives of WT systems. [ABSTRACT FROM AUTHOR]

Helicopter systems are considered a complex and challenging control problem due to strong couplings and high non-linearities. In this paper, simulated annealing (SA), as one of the leading methods in search and optimization, is applied to tune a multivariable controller of a lab-scale helicopter system. The lab-scale helicopter system is a multivariable experimental aerodynamic test rig that resembles the behaviour of a real helicopter. The control objectives are quickly to reach a desired position or track a trajectory. A centralized cross-coupled PID controller is used to achieve these objectives. First, SA optimizations are carried out with 24 different initial configurations. Then, the best results of these SA configurations are compared with other controllers obtained with evolutionary algorithms (EAs) of genetic algorithms (GAs), modified particle swarm optimization (MPSO) and differential evolution (DE). The comparisons are based on statistical measures of 20 independent trials, non-linear computer simulations of different input signals and real-time measurements for various commands of positions or trajectories. Results show that SA obtained the best performance index and acceptable time-domain performance on reaching hovering point, following a step command and tracking a sine trajectory compared with the investigated EAs. [ABSTRACT FROM AUTHOR]

The increasing heterogeneity and scalability of the Internet of everything, especially, the Energy Internet (EI), is a prompt for novel engineering paradigms. The current infrastructure of hardware technologies besides the available software stack is becoming obsolete for handling the advanced requirements of the smart energy ecosystem. Accordingly, software definitions of the Internet of Things (IoT) stack functions are gaining interest to provide more flexible and scalable implementations. This paper investigates a sophisticated softwarization explicit hybrid model predictive control strategy for energy storage facility systems through the accessor design pattern. This strategy is presumed to support and facilitate the integration of higher-level analytic and control functionalities within the IoT stack at edges and nodes. Simulation is conducted with the Ptolemy II software to evaluate the execution semantics of the proposed strategy. In addition, real-time experiments on a lab-scale hydraulic tank process with both abrupt and gradual commands test scenarios showed the efficiency and reliability of the accessor-based controllers for achieving the control objectives for the benchmark storage facility with an abstract implementation. The analysis shows that this process can serve as a benchmark to mimic the charge and discharge cycles of storage elements in EI ecosystems. • A new paradigm for hydrogen energy storage interfacing within energy Internet ecosystems is proposed and investigated. • An actor-oriented approach is applied for implementing real-time control systems of hydrogen storage. • Software-defined model predictive control is implemented within a Node.js accessor host. • A benchmark process is developed as an emulator for hydrogen energy storage to evaluate the accessors design pattern. [ABSTRACT FROM AUTHOR]