The headwater of Yellow River Basin (HYRB) is crucial for the water resources of the whole basin in Northwest China. Based on the semi-distributed hydrological model "Soil and Water Assessment Tool" (SWAT), the spatiotemporal change trends of blue water and green water resources in the HYRB were analyzed quantificationally. By using the Sequential Uncertainty Fitting program (SUFI-2), the model was calibrated at Tangnaihai hydrological station and uncertainty analysis was performed. The results showed that the total water resources decreased by 1.08 billion m 3 over the past five decades in the HYRB. Blue water and green water storage (soil water) presented the downtrend, while green water flow (actual evapotranspiration) increased between 1961 and 2010. The decrease in blue water resources were mainly attributed to the decrease in precipitation in the southwest parts of the study area while the increase in actual evapotranspiration and the decrease in soil water were the results of the uptrend of air temperature. In 1990s, an enormous transition occurred between the blue water (24.86 %) and green water flow (63.46 %). At seasonal scale, the largest down trend of blue water and uptrend of actual evapotranspiration all occurred in autumn. The decrease ratios of them were 88.3 and 83.1 % in inter-annual variation, respectively. The study can provided a scientific basis for integrated water resources management under the background of global climate change and human activity.

No abstract

No abstract

No abstract

Compressive sensing provides a new idea for machinery monitoring, which greatly reduces the burden on data transmission. After that, the compressed signal will be used for fault diagnosis by feature extraction and fault classification. However, traditional fault diagnosis heavily depends on the prior knowledge and requires a signal reconstruction which will cost great time consumption. For this problem, a deep belief network (DBN) is used here for fault detection directly on compressed signal. This is the first time DBN is combined with the compressive sensing. The PCA analysis shows that DBN has successfully separated different features. The DBN method which is tested on compressed gearbox signal, achieves 92.5 % accuracy for 25 % compressed signal. We compare the DBN on both compressed and reconstructed signal, and find that the DBN using compressed signal not only achieves better accuracies, but also costs less time when compression ratio is less than 0.35. Moreover, the results have been compared with other classification methods.