Tibetan Multi-Dialect Speech and Dialect Identity Recognition

Tibetan language has very limited resource for conventional automatic speech recognition so far. It lacks of enough data, sub-word unit, lexicons and word inventories for some dialects. And speech content recognition and dialect classification have been treated as two independent tasks and modeled respectively in most prior works. But the two tasks are highly correlated. In this paper, we present a multi-task WaveNet model to perform simultaneous Tibetan multi-dialect speech recognition and dialect identification. It avoids processing the pronunciation dictionary and word segmentation for new dialects, while, in the meantime, allows training speech recognition and dialect identification in a single model. The experimental results show our method can simultaneously recognize speech content for different Tibetan dialects and identify the dialect with high accuracy using a unified model. The dialect information used in output for training can improve multi-dialect speech recognition accuracy, and the low-resource dialects got higher speech content recognition rate and dialect classification accuracy by multi-dialect and multi-task recognition model than task-specific models.     

金沙江水电工程智能建造技术体系研究与实践

运用"全面感知、真实分析、实时控制"的智能闭环控制理论,提出了以工程建设资源要素动态精准管理、业务流程数字化管理、工艺过程智能控制、实物成本精确分析、结构安全与进度耦合分析及联动调控的数字化技术和智能化技术为核心,以水电工程数据模型如大坝全景信息模型DIM为基础,以智能建造管理平台iDam为主体的智能建造技术体系,并对工程数据结构分解与编码体系、工程数据感知传输共享技术、iDam平台的系统架构与业务架构、智能建造工程绩效等关键技术进行了分析。通过统一规范的工程数据结构与编码体系,iDam平台可集成工程建设全过程的基础数据、环境数据、过程数据和监测数据,为各业务模块服务;建立的基于工程技术和管理大数据的机理分析功能,可为工程建设的业主、施工、监理、设计、科研及运行等单位服务。智能建造初步实践表明,本文构建的技术体系是科学可行的,对基础设施智能建造技术与管理体系的研究与实践具有参考价值。     

Challenges in supply chain redesign for the Circular Economy: a literature review and a multiple case study

Despite companies face several challenges when redesigning their supply chain for the Circular Economy, the literature lacks a systematisation of such challenges and of the ways to overcome them. Through a systematic literature review, this paper identifies and systematises 24 challenges that may hamper a supply chain redesign for the Circular Economy. Sixteen among these challenges are well known from research in related topics. On the contrary, the remaining eight are relatively new or take a different relevance within the Circular Economy context. A multiple case study in the household appliance supply chain is carried out, to explore how these challenges appear in practice and how companies may tackle them. The cases analysed involve actors at different supply chain levels, and findings suggest that a great degree of vertical integration by one actor in the supply chain is not a necessary condition for Circular Economy implementation. The empirical study, in conjunction with the literature analysis, leads to the development of a framework linking the challenges to specific levers that companies may pursue to overcome them. The framework can be seen as a reference for managers undertaking the path towards Circular Economy.     

A new experimental simulation method for the development of non-renewable hydrogen energy—Oil and gas resources

At present, the proportion of tight oil in non-renewable hydrogen energy is increasing. According to an initial exploration and attemptable practice on the exploration of tight oil, it is found that the cost can be controlled effectively and positive effects are achieved. But this technique cannot make sure the proppants filled uniformly in the long fracture. Several researches on the proppants migration experiment devices and factors influencing on proppant setting are reviewed and a new set of experimental device to simulate the laws of proppants setting in long fracture is developed. This device can simulate the main factors influencing proppants setting performance. It analyzes several factors such as wall filtration, construction displacement, sand concentration, proppant size and density, viscosity of fracturing fluid is used to rank the influencing degree of every factor. Considering the effects of mutual interference between proppants, width of fracture, rough fracture surface and fracture surface filtration during the proppants setting progress, the mathematical model of proppant setting is modified by adding sand concentration correction factor, wall effect correction factor and filtration correction factor. The experimental data verify the accuracy of the settlement model is established using the data getting from experiment.     

Nitrogen dry deposition to Lake Superior and Lake Michigan

Nitrate (NO₃^?) levels in Lake Superior have increased from historic levels of about 5?μM to its current concentration of about 25?μM. The atmosphere makes a substantial contribution to the nitrogen budgets for Lake Superior and Lake Michigan. This study provides a more well-defined estimate of nitrogen dry deposition rates derived from the measurement of over-water concentrations, and in situ meteorological measurements, which were input into the Resistance Model. We obtained a nitrogen dry deposition rate of [(3.41?±?2.26)?×?10⁷?kg?N/yr; (5.90?±?3.91)?kg?N/ha/yr] over Lake Michigan, and [(1.54?±?1.06)?×?10⁷?kg?N/yr; (1.87?±?1.27)?kg?N/ha/yr] over Lake Superior. Nitric acid (HNO₃), which originates from the combustion of fossil fuels, contributes 84% of the total nitrogen dry deposition to Lake Michigan; and 66% to Lake Superior. Ammonia (NH₃), which originates from agricultural activities and gasoline combustion, is the second highest contributor of nitrogen dry deposition to both lakes: contributing 13% to Lake Michigan and 32% to Lake Superior. The nitrogen dry deposition is approximately 68% of the nitrogen wet deposition over Lake Superior, and approximately 80% of wet deposition over Lake Michigan. The over-water dry deposition velocity of HNO₃ and NH₃ were also evaluated. We obtained morning deposition velocities of 0.099?cm/s for NH₃ and 0.095?cm/s for HNO₃; and afternoon values of 0.137?cm/s for NH₃ and 0.132?cm/s for HNO₃. Another key finding is that the atmospheric concentrations of nitrogen compounds near Lake Michigan and Lake Superior have decreased since 2003.     

Hydrogen production by oxidative steam reforming of methanol over anodic aluminum oxide-supported Cu-Zn catalyst

Oxidative steam reforming of methanol (OSRM), which is a convenient reaction for producing hydrogen, suffers from the hot spot formation problem when conventional particle catalysts are used. Recently, an anodic aluminum oxide (AAO)-supported Cu-Zn catalyst was proposed as an OSRM catalyst for its high thermal conductivity through the aluminum metal body. In this study, OSRM was conducted in a prototype reactor packed with the AAO plate catalyst strips. It was verified that the high thermal conductivity of the catalyst effectively suppresses the hot spot formation and makes the temperature profile smooth along the reactor. The start-up time of the reactor depended on the preheating temperature and was very short (less than 2 min) for preheating over 503 K. The methanol conversion and reactor temperature increased with increasing O₂/CH₃OH mole ratio, indicating that the mole ratio can be used as a control variable to operate the reactor at desired conditions. Further, a reactor model was developed and verified, and the simulation showed that for a given total reactor volume, an optimal reactor configuration could be achieved by shortening the reactor length while widening the cross-sectional area.     

Study on internal stress damage detection in long-distance oil and gas pipelines via weak magnetic method

Weak magnetic stress detection is an important issue in oil–gas pipeline internal detection area. In order to verify the characteristics of weak magnetic stress internal detection signals, we built herein a magneto-mechanics equivalent model having a balanced magnetic field. First, we calculated the relationship between the stress and the weak magnetic signals; consequently, the analysis propagation laws of the weak magnetic signals with non-magnetic saturation were pointed out. Finally, the theoretical model was validated by a systematic experimental research. The analytical results show that a one-to-one linear link between the weak magnetic signals and the stress concentration is clear. Instead, the change of the weak magnetic signals with the liftoff is nonlinear, therefore we are proposing the Boltzmann liftoff correction factor whose degree of adaptability of the equivalent model can reach the value of 94%. It is possible to note that when the liftoff is in the approximate linear stage, the relevance ratio and the recognition rate of the magneto-mechanics curve show a high-quality. This conclusion is important in the engineering field for the set of the liftoff.     

Attitudinal choice models with applications in human decision making

A new family of attitudinal discrete choice models is proposed by considering the attitudinal character and the weight vector, both of which are specific to a decision maker (DM). Given the attribute values of different alternatives, the proposed models give varying choice probabilities, as per the DM's-specific attitudinal character and the weight vector. It is also shown that the conventional discrete choice models are the special cases of the proposed attitudinal models. The proposed choice models are also generalized through an additional parameter to add to their capabilities. An application on real data is included to demonstrate their usefulness in the real world.     

Evaluation of dynamic modulus measurement for C/C-SiC composites at different temperatures

The determination of elastic properties at application temperature is fundamental for the design of fibre reinforced ceramic composite components. An attractive method to characterize the flexural modulus at room and high temperature under specific atmosphere is the nondestructive Resonant Frequency Damping Analysis (RFDA). The objective of this paper was to evaluate and validate the modulus measurement via RFDA for orthotropic C/C-SiC composites at the application temperature. At room temperature flexural moduli of C/C-SiC with 0/90° reinforcement were measured under quasi-static 4-point bending loads and compared with dynamic moduli measured via RFDA longitudinally to fibre direction. The dynamic modulus of C/C-SiC was then measured via RFDA up to 1250°C under flowing inert gas and showed an increase with temperature which fitted with literature values. The measured fundamental frequencies were finally compared to those resulting from numerical modal analyses. Dynamic and quasi-static flexural moduli are comparable and the numerical analyses proved that bending modes are correctly modeled by means of dynamic modulus measured via RFDA. The nondestructive RFDA as well as the numerical modeling approach are suitable for evaluation of C/C-SiC and may be transferred to other fibre reinforced ceramic composite materials.