Applied Mechanics and Materials Vols. 204-208

Abstract: By the aid of Tangent-Cross-Point feature, a novel improved SIFT algorithm is proposed in this paper. This method is more efficient and reliable in some cases, since it adopts Tangent-Cross-Point feature to improve the efficiency and robustness of SIFT algorithm. With a pair of matching feature points and a pair of curves nearby, it’s very easy to estimate the values of local transformation parameters and validate the correctness of this matching by the aid of Tangent-Cross-Point feature.

Abstract: A virtual reality system was developed to simulate emergency evacuations during fires. The spreading of the flame and smoke in the virtual fire was modeled based on numerical fire simulations, so that the conditions are similar to real life. The paper investigates a number of evacuation strategies in high-rise buildings and the set of human factors affecting high-rise evacuations. It also discusses the challenges that building occupants and fire departments face during a fire emergency.

Abstract: Wireless smart sensor networks (WSSN) have many advances compared with traditional structural health monitoring (SHM) such as wireless process, real-time calculation and low cost. However power consumption is considered as one of the most limitations in this field. The unique features offered by decentralized data aggregation (DDA) technique with the potential to overcome power consumption enable implementation of the dense array of WSSN on large structures. This paper presents a system identification of a simply supported plate based on the random decrement technique (RDT) and natural excitation technique (NExT) in combination with eigensystem realization algorithm (ERA) using Illinois Structural Health Monitoring Project (ISHMP) Services Toolsuite. Finally the system parameters including natural frequency and mode shapes are in accordance with numerical simulation showing efficacy and feasibility of decentralized NExT/ERA and RDT/ERA system identification.

Abstract: Researched how the multi-pipeline processor accelerates the running of thread ,found that when the branch predictor facing the random branch instruction, the hit rate will become very low, so bring out a new method that using the free pipeline to accelerate the running of branch instruction. If the right prediction from branch predictor is less than 70% and there is a free pipeline, then using two pipelines to run the two sides of a branch instruction at the same time. In order to test the new method, the HLA (High Level architecture) architecture-based simulation system is established, the results show that the new method can really reduce the time when processing the random branch instructions.

Abstract: D-S evidence theory is a theory of dealing with uncertain information which was empowered by Dempster and expanded by Shafer. However, as a theory founded in 1970s, D-S theory is invalid when there has conflicts between evidences. This papal reviewed classical models and applications of D-S evidence theory. These models revised the shortcomings of D-S theory and put forward some new methods. The applications of D-S theory involved decision-making, information fusion and many other fields.

Abstract: The statistical sensitivity of training neural networks by B-splines weight functions and its applications for digital modulation signal recognition (DMSR) is discussed in this paper. By extracting some instantaneous parameters using the technology of DMSR, and using the sensitivity formula of B-splines weight function neural networks proposed in this paper, we show that the classifier of B-splines weight function neural networks has optimized architecture and high recognition rate.

Abstract: Based on large amounts of datas, BP neural network model was applied to studying the DO influence on fluidized bed step- feed process. The output values and the actual values reached a high fitting degree. The fitting degree of NH4+-N was 0.979 and TN of 0.965. So BP neural network model could be used as a new type analysis method in water treatment. It could predict operating results under other conditions without experimental measure.

Abstract: The CFD software FLUENT was used as the foundation to develop the numerical wave flume, in which the governing equations are the Reynolds-averaged Navier-Stokes (RANS) equations and the standard k~ε turbulence model. The wave generating and absorbing were introduced into the RANS equations as the source terms using the relaxation approach. A new module of the wave generating and absorbing function, which is suitable for FLUENT based on the volume of fluid method (VOF), was established. Within the numerical wave flume, the reflected waves from the model within the computation domain can be absorbed effectively before second reflection appears due to the wave generating boundary. The computational results of the wave pressures on the bottom of the rectangular slab were validated for the different relative clearance by the experimental data. Good agreements were found.