Search

Combining level set image segmentation with the prior shape information, we proposed an improved prior shape model in face recognition. Firstly, we introduced a local tensile invariant to X and Y direction and a shear invariant based on the shape statistics. Then, shape energy term with rotation, scale, shear and translation invariance was reconstructed in level set C-V model. The new model considers global and local image change and makes face contour evolving stably. Experimental results demonstrate that our model can segment obscured face images effectively.

To research the effect of electromagnetic stirring process parameters on the parts’ quality in semi-solid forming of aluminum alloy, the material of ZL017 was heated and melted, and electromagnetic stirred under different frequency and power, then material microstructure gained by different process parameters was observed. The elementary alteration rule of fluid field was analyzed with the alteration of stirring frequency, current density, stirring time，stirring temperature and fluid viscosity. Suitable process parameters for electromagnetic stirring of ZLl07 was concluded.

To research the problem of part’s dimensional accuracy error easily occurred in hot press forming, a B-Pillar on one domestic car was set as example, the dimensional of the actual part and its model diagram were compared by means of ATOS optical scanner. By recording the the process parameters used in practical production, the effect trend of three key process parameters on part’s dimensional accuracy was summarized, process parameters suitable for practical production were gained.

To solve the problem of part’s section thinning easily occurred in hot press forming, the numerical simulation model of hot press forming process of a B-Pillar on one domestic car was established, the forming process was simulated by PAM-STAMP 2G software, the area thinning on the part were forecasted. Then through modifying the corner radius and the tiny shape of the part, the thinning level lightened greatly, the effect was certificated by new simulation results and practical production.

Two interstitial-free steel grades, one based on combined niobium and titanium additions and the other based on titanium only addition, were investigated. Annealing experiments were used to study the effect of niobium on the mechanical property. And tensile-shear fatigue tests were used to study the effect of niobium on the fatigue property of spot welding. The experimental results show that niobium addition not only can significantly improve the mechanical properties of the basic materials by way of grain refinement, but also has beneficial effects to the fatigue strength of spot welding.

This paper presents a novel approach to detect the fault categories for rolling element bearings based on the continuous wavelet transform and hidden Markov models. With the Morlet wavelet transform, the effective fault information is extracted from the time-frequency domain of the vibration signals. Then, the wavelet coefficients are divided into multi-segments, and the infinity-norms of each segment is applied as the features to construct the observation vector of the hidden Markov models. Finally, the experimental results on bearing faults identification and isolation are illustrated.