Search

This study simulated the evolution of Typhoon Hato (2017) with the Weather Research and Forecasting model using three bulk schemes and one bin scheme. It was found that the track of the typhoon was insensitive to the microphysics scheme, whereas the degree of correspondence between the simulated precipitation and cloud structure of the typhoon was closest to the observations when using the bin scheme. The different microphysical structure of the bin and three bulk schemes was reflected mainly in the cloud water and snow content. The three bulk schemes were found to produce more cloud water because the application of saturation adjustment condensed all the water vapor at the end of each time step. The production of more snow by the bin scheme could be attributed to several causes: (1) the calculations of cloud condensation nuclei size distributions and supersaturation at every grid point that cause small droplets to form at high levels, (2) different fall velocities of different sizes of particles that mean small particles remain at a significant height, (3) sufficient water vapor at high levels, and (4) smaller amounts of cloud water that reduce the rates of riming and conversion of snow to graupel. The distribution of hydrometeors affects the thermal and dynamical structure of the typhoon. The saturation adjustment hypothesis in the bulk schemes overestimates the condensate mass. Thus, the additional latent heat makes the typhoon structure warmer, which increases vertical velocity and enhances convective precipitation in the eyewall region.

A new type of heat pipe assisted thermoelectric generator (HP-TEG) for automobile exhaust waste heat recovery is proposed in this paper. To confirm the optimized configurations of the thermoelectric device, an experimental prototype was constructed to obtain the best heat pipes insertion depth (Dhp), and the optimum angle between the heat pipe row and the gas flow direction (θhp). A HP-TEG device was designed, constructed after the optimization of heat pipes depth and angle. The operating parameters were further investigated to enhance the performance of thermoelectric generation (TEG). The experimental results show that the power output of HP-TEG preferred high exhaust temperature, cold water flow rate and mass flow rate. The maximum open circuit voltage of 36 thermoelectric modules (TEMs) is measured as 81.09 V. The corresponding power output and pressure drop are 13.08 W and 1657 Pa, respectively, with an optimized thermoelectric power generation efficiency of the HP-TEG as 2.58%.

Waste heat recovery through thermoelectric generators is a promising way to improve energy conversion efficiency. This paper proposes a type of heat pipe assisted thermoelectric generator (HP-TEG) system. The expandable evaporator and condenser surface of the heat pipe facilitates the intensive assembly of thermoelectric (TE) modules to compose a compact device. Compared with a conventional layer structure thermoelectric generator, this system is feasible for the installment of more TE couples, thus increasing power output. To investigate the performance of the HP-TEG and the optimal number of TE couples, a theoretical model was presented and verified by experiment results. Further theoretical analysis results showed the performance of the HP-TEG could be further improved by optimizing the parameters, including the inlet air temperature, the thermal resistance of the heating section, and thermal resistance of the cooling structure. Moreover, applying a proper number of TE couples is important to acquire the best power output performance.

The Weather Research and Forecasting model version 3.2.1 with the Lin microphysics scheme was used herein to simulate super typhoon Usagi, which occurred in 2013. To investigate the effect of the concentration of cloud condensation nuclei (CCN) on the development of typhoon Usagi, a control simulation was performed with a CCN concentration of 100 cm−3, together with two sensitivity tests: C10 and C1000, having CCN concentrations of 10 cm−3 and 1000 cm−3, respectively. The path, intensity, precipitation, microphysical processes, and the release of latent heat resulting from the typhoon in all three simulations were analyzed to show that an increase in CCN concentration leads to decreases in intensity and precipitation, an increase of the cloudless area in the eye of the typhoon, a more disordered cloud system, and less latent heat released through microphysical processes, especially the automatic conversion of cloud water into rainwater. Overall, an increase in CCN concentration reduces the total latent heat released during the typhoon suggesting that typhoon modification by aerosol injection may be optimized using numerical simulations to ensure the strongest release of latent heat within the typhoon.

In this study, the Weather Research and Forecasting model coupled with chemistry was used to simulate cloud microphysics processes of a mesoscale urban heavy precipitation event over the Pearl River Delta, China. Two intensities of anthropogenic emissions were considered: the E2010 simulation adopted the level and distribution of present-day pollution, while the E1960 simulation adopted the aerosol condition before urbanization. The modeling reproduced the precipitation process and the results showed a more centralized distribution of precipitation and a rain rate that was 20% higher at the mature stage in E2010. A larger area of convective-type precipitation occurred in E2010 (51.2%) than in E1960 (42.3%), but less stratiform- and mixed-type precipitation occurred in E2010 (7.6% and 7.3%, respectively) than in E1960 (12.1% and 9.1%, respectively). In comparison with E1960, E2010 produced larger quantities of cloud droplets and ice-phase particles. Additional release of condensation/deposition latent heating promoted vertical upward motion and convection in clouds, which further promoted the riming process between hail and raindrops/cloud droplets, as well as the colliding, merging, and collecting processes between raindrops and droplets that ultimately generated raindrops with larger radii (600–700 μm) but smaller number concentrations. Furthermore, hydrometeors convert to precipitation more effectively in polluted conditions. In E2010, the centralized distribution and greater mass magnitude of hail and melting hail resulted in a focused distribution of rainfall. In E1960, the decentralized distributions and smaller mass magnitude of hail and melting hail led to a dispersed distribution of rainfall.

The physical and chemical characteristics of aerosols have a significant effect on scattering coefficients (σsp). In this research, a single particle aerosol mass spectrometer (SPAMS) and a scanning mobility particle sizer spectrometer (SMPS) were used to measure the chemical composition and size distribution of aerosol number concentration in Nanjing, China. During the observation time, the average σsp varied from 11.64 to 485.84 Mm−1, with an average of 114.16 ± 77.87 Mm−1. For relative humidity greater than 75.0%, the optical hygroscopic growth factor (f(RH)) increased rapidly, with an approximate distribution of 3.41 ± 0.78. During the night of July 15th, 2015, nitrate and sulfate increased by 5.93 and 1.32 times, respectively, compared with daytime values. The peak of the average size distribution of the aerosol number concentration during the observation period (341.0 cm−3) was located in the Aitken mode of 58.3 nm. After particle size conversion from the vacuum aerodynamic diameter to the mobility diameter, the particle size range of the aerosol mass spectrometer ranged from 138.2–723.4 nm, all of which are distributed in accumulation mode. Source apportionment revealed that secondary, biomass, and carbonaceous aerosols are the primary facilitators in 327.8–406.8 nm, accounting for 21.3%, 25.7%, and 33.3%, respectively. The fixed scattering coefficients were calculated using the Mie model, according to the source apportionment of aerosol number concentration for modifying the aerosol refractive index. From July 6th–11th, the calculation and observation results of σsp exhibit excellent consistency and a Pearson coefficient of 0.66. Therefore, the accurate size distribution of the chemical composition and source apportionment of aerosols are of great significance for the simulation of atmospheric optical scattering.

As the information technology develops rapidly, Web servers are easily to be attacked because of their high value. Therefore, Web security has aroused great concern in both academia and industry. Anomaly detection plays a significant role in the field of Web security, and log messages recording detailed system runtime information has become an important data analysis object accordingly. Traditional log anomaly detection relies on programmers to manually inspect by keyword search and regular expression match. Although the programmers can use intrusion detection system to reduce their workload, yet the log system data is huge, attack types are various, and hacking skills are improving, which make the traditional detection not efficient enough. To improve the traditional detection technology, many of anomaly detection mechanisms have been proposed in recent years, especially the machine learning method. In this paper, an anomaly detection system for web log files has been proposed, which adopts a two-level machine learning algorithm. The decision tree model classifies normal and anomalous data sets. The normal data set is manually checked for the establishment of multiple HMMs. The experimental data comes from the real industrial environment where log files have been collected, which contain many true intrusion messages. After comparing with three types of machine learning algorithms used in anomaly detection, the experimental results on this data set suggest that this system achieves higher detection accuracy and can detect unknown anomaly data.

International audience; Studying the spatial distribution characters of rubber yield and soil nutrients and the rule of spatial variability are important for suitable fertilization strategy in rubber plantation. This paper selected Hongquan Branch, Guangba Branch and Gongai Branch of Guangba Farm in Hainan province as study area and total of 327 samples were selected in the rubber plantation. The spatial distribution characters of rubber yield and five soil nutrients, including organic matter (OM), total nitrogen (TN), available phosphorus (AP), available potassium (AK), exchangeable calcium (Ga), were studied using traditional analysis method and geo-statistics analysis method. The results showed that: (1) The average value of rubber yield was 3.55 kg/hm2 with moderate spatial variability and the average values of OM, N, P, K and Ga were 11.65 g/kg, 0.07%, 16.23 mg/kg, 49.65 mg/kg and 84.44 mg/kg, respectively. Soil OM, TN, AK and Ga had moderate spatial variability but AP had strong spatial variability. (2) Rubber yield and soil total nitrogen (N) nutrient had strong spatial dependence; soil OM, AP, AK and Ga had moderate spatial dependence. (3) Based on the previous reports of normal range of soil nutrients, soil OM and TN nutrient content were very low in the studied rubber plantation of Guangba Farm. Therefore, more nitrogen fertilizer should be applied in the rubber plantation in future.

Thirty-five surface sediment samples collected from three mangroves in Shantou coastal zone, China in 2007 were analyzed for a suite of polycyclic aromatic hydrocarbons (PAHs). Two mathematical models were used to determine the profiles and relative contributions of PAH sources to the mangroves. The two models are principal component analyses (PCA) with multiple linear regression analysis (MLR) and positive matrix factorization (PMF). Both models identified five factors and gave excellent correlation coefficients between predicted and measured levels of 16 PAH compounds, but the results had some differences. The PAH contribution rate attributed to vehicular emission sources identified by PCA-MLR was 37.20%, but the rate identified by PMF was only 12.37%. The main sources identified by PCA-MLR were combination of biomass/coal combustion and vehicular emissions, whereas the main source identified by PMF was only biomass/coal combustion. The PMF analysis was the preferred model for the paper data set.

Recent developments in information technologies provide facility agriculture monitoring system a new perspective of networking and automation. This paper presents a new agriculture monitoring system based on WSNs (Wireless sensor networks) with IP cameras, which can be controlled remotely to have close vision of plants. WSNs can deploy freely, collect sensor data periodically to the control center, process and store historical data, which could facilitate clients and experts in agriculture to monitor the conditions in a large field. Traditional WSNs monitoring system only focused on environmental parameters, such as temperature and humility. We integrate video information and environmental data. Also the sleeping methods based on small-scale time synchronization and alarming mechanism of our system achieve better communication and less power consumption. Through experiments we validated the system in a greenhouse and outdoors fields in Hainan successfully.

Maximum Entropy Markov Models (MEMM) can avoid the assumption of independence in traditional Hidden Markov Models (HMM), and thus take advantage of context information in most text mining tasks. Because the convergence rate of the classic generalized iterative scaling (GIS) algorithm is too low to be tolerated, researchers proposed a lot of improved methods such as IIS, SCGIS and LBFGS for parameters training in MEMM. However these methods sometimes do not satisfy task requirements in efficiency and robustness. This article modifies the traditional Particle Swarm Optimization (PSO) algorithm by using dynamic global mutation probability (DGMP) to solve the local optimum and infinite loops problems and use the modified PSO in MEMM for estimating the parameters. We introduce the MEMM trained by modified PSO into Chinese Part-of-Speech (POS) tagging, analysis the experimental results and find it has higher convergence rate and accuracy than traditional MEMM.

In order to solve problems caused by the individualism of Chinese architecture more and more researchers focus on Hybrid and improved Hidden Markov Model. However, as the foundation of Chinese natural language processing, studies on Chinese words segmentation based on Second-order Hidden Markov Model (HMM2) are not abundant. A words frequency weighted smoothing method and a Threshold-Viterbi algorithm are proposed and combined to build a Improved Viterbi Algorithm-based HHM2(IV-HMM2) model in this article to overcome the sparse problem and improve the accuracy. Experimental rusults demonstrate that the improved model has better performance and lower overhead than traditional HMM2.

AdaBoost is an excellent committee-based tool for classification. However, its effectiveness and efficiency in multiclass categorization face the challenges from methods based on support vector machine (SVM), neural networks (NN), naïve Bayes, andk-nearest neighbor (kNN). This paper uses a novel multi-class AdaBoost algorithm to avoid reducing the multi-class classification problem to multiple two-class classification problems. This novel method is more effective. In addition, it keeps the accuracy advantage of existing AdaBoost. An adaptive group-basedkNN method is proposed in this paper to build more accurate weak classifiers and in this way control the number of basis classifiers in an acceptable range. To further enhance the performance, weak classifiers are combined into a strong classifier through a double iterative weighted way and construct an adaptive group-basedkNN boosting algorithm (AGkNN-AdaBoost). We implement AGkNN-AdaBoost in a Chinese text categorization system. Experimental results showed that the classification algorithm proposed in this paper has better performance both in precision and recall than many other text categorization methods including traditional AdaBoost. In addition, the processing speed is significantly enhanced than original AdaBoost and many other classic categorization algorithms.