A novel VC (voice conversion) method based on hybrid SVR (support vector regression) and GMM (Gaussian mixture model) is presented in the paper, the mapping abilities of SVR and GMM are exploited to map the spectral features of the source speaker to those of target ones. A new strategy of F0 transformation is also presented, the F0s are modeled with spectral features in a joint GMM and predicted from the converted spectral features using the SVR method. Subjective and objective tests are carried out to evaluate the VC performance; experimental results show that the converted speech using the proposed method can obtain a better quality than that using the state-of-the-art GMM method. Meanwhile, a VC method based on non-parallel data is also proposed, the speaker-specific information is investigated using the SVR method and preliminary subjective experiments demonstrate that the proposed method is feasible when a parallel corpus is not available.
In the process of coal extraction, a fractured zone is developed in the overburden above the goaf. If the fractured zone is connected with an aquifer, then water inrush may occur. Hence, research and analysis of the height of overburden fractured zone (HOFZ) are of considerable significance. This study focuses on the HOFZ determination in deep coal mining. First, general deformation failure characteristics of overburden were discussed. Second, a new method, numerical simulation by orthogonal design(NSOD), have been proposed to determinate the HOFZ in deep coal mining. Third, the validity of NSOD is verified in the practical application, compared with empiric al formula in Chinese Regulations and in-situ test. These three methods were applied to determine the HOFZ of working face No. 111303 in No. 5 coal mine. The pre dicted HOFZ of NSOD is found to be similar to the result of the in-situ test (8.9% relative error), whereas the HOFZ calculated by the empirical formula has extremely large error (25.7% relative error). Results show that the NSOD can reliably predict the HOFZ in deep coal mining and reduce time and expenses required for in-situ test.
The magnetic properties of the U-type ferrite synthesized by a sol-gel process had studied by substituting cobalt with manganese or zinc in cobalt-based U-type ferrite. The substituted U-type ferrite showed a dominant crystal structure at a different substitution ratio of manganese and zinc. The change of the starting temperature of U-type ferrite formation according to substitutional elements was confirmed by TG-DTA analysis. In the case of manganese substitution, the starting temperature of U-type ferrite formation lowered, and on the contrary, when zinc was substituted, it became higher. The magnetic properties of the U-type ferrite substituted with manganese showed a tendency that the saturation magnetization was decreased and the coercivity was increased as the manganese ratio increased. The highest saturation magnetization was 57.9 emu/g in the specific composition (Ba4Co0.5Zn1.5Fe36O60) substituted with zinc.
The densification behavior of H13 tool steel powder by dual speed laser scanning strategy have been characterized for selective laser melting process, one of powder bed fusion based metal 3d printing. Under limited given laser power, the laser re-melting increases the relative density and hardness of H13 tool steel with closing pores. The single melt-pool analysis shows that the pores are located on top area of melt pool when the scanning speed is over 400 mm/s while the low scanning speed of 200 mm/s generates pores beneath the melt pool in the form of keyhole mode with the high energy input from the laser. With the second laser scanning, the pores on top area of melt pools are efficiently closed with proper dual combination of scan speed. However pores located beneath the melt pools could not be removed by second laser scanning. When each layer of 3d printing are re-melted, the relative density and hardness are improved for most dual combination of scanning. Among the scan speed combination, the 600 mm/s by 400 mm/s leads to the highest relative density, 99.94 % with hardness of 53.5 HRC. This densification characterization with H13 tool steel laser re-melting can be efficiently applied for tool steel component manufacturing via metal 3d printing.