Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

The Clean Development Mechanism (CDM) is a project-based carbon trade mechanism that subsidizes the users of climate-friendly technologies and encourages technology transfer. The CDM has provided financial support for a large share of Chinese wind projects since 2002. Using pooled cross-sectional data of 486 registered CDM wind projects in China from 2002 to 2009, we examine the determinants of technological change in wind power from a learning perspective. We estimate the effects of different channels of learning—learning through R&D in wind turbine manufacturing, learning from previous experience of installation, and learning through the network interaction between project developer and turbine manufacturer—on technological change, measured as reductions in projected costs or as increased capacity factor across CDM wind projects. While we find that a manufacturer's R&D and previous installation experience matter, interactions between wind turbine manufacturers and wind project developer lead to the largest cost reductions. Whereas existing literature suggests that wind power firms can learn from the experience of other wind farm developers, our results indicate that wind power firms mainly learn from their own experience and that knowledge spillovers mostly occur within certain partnerships between wind project developer and foreign turbine manufacturers in China's wind power industry.

Zonder onderwerpscode: economie

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

The scope of this book is to map China's city clusters and their individual directions for the national-level strategies in line with the 2060 carbon neutrality plan. Since China announced the carbon neutrality plan in autumn 2020, no study has looked at the role of city clusters in achieving this long-term plan. Hence, this study is believed to be the first attempt to explore this important topic from the city cluster perspective. It explores the challenges, opportunities, and directions of all 19 city clusters, allowing readers to have a clear picture of China's historical and ongoing progress, as well as the challenges and opportunities that lie ahead. In a short time, China's city clusters have helped boost regional economic development, infrastructure development, trade and business, and better urban-rural integration. With enhanced coordination of connection and transport networks in and between the city clusters, we see a growing number of initiatives beyond just the initial economic strategies. The dual approach of top-down policies and infrastructure systems and bottom-up governance and investments has helped China consider urban-rural development strategies and regional sustainable development. These factors are essential to be explored from the city cluster perspective and in line with China's sustainable development and carbon neutrality directions. Hence, the book covers these points holistically, ensuring that regional planning and development are favored in the face of uneven urbanization trends. We anticipate this book to be a valuable resource for local governments and authorities, urban planners and practitioners, developers, and urban researchers. While the focus is on China's city clusters, we believe there are similar examples elsewhere. Hence, lessons learnt from this book could apply to other countries, regions, and subregions. Lastly, the book aims to put regional sustainable development at the heart of longer-term strategies and plans, such as the case of China's carbon neutrality plan.

Political philosophy. Social philosophy --- Environmental protection. Environmental technology --- milieubeleid

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

As the world population grows and urbanizes, our society is producing an ever-increasing amount of waste, estimated to reach approximately 2.6 billion tons a year in 2030. This has been creating huge impacts on the underlying recycling chain for the Material Recycling Facilities (MRFs) where our waste and recyclables are processed. Not only do they require more capacity, but the composition of the PMD is also changing drastically, with lots of new packaging and materials making the sorting process far more complex. How to ensure that we recycle more and recycle right is an important challenge and requires efficient recycling strategies and technologies to be developed. State-of-art computer vision methods, especially Convolutional Neural Network (CNN) models, are considered to play a key role in improving waste sorting efficiencies. However, much academic research on developing CNN-based waste classification models has relied on public waste datasets. Those publicly available waste datasets mainly contain relatively clean and intact waste objects at home or from common public places, whereas in MRFs the wastes are often severely deformed and compacted. This thesis work has collected a wide range of real-life waste objects from MRFs and constructed a waste dataset of 3136 unique objects. To simulate the varying light condition, which is a frequent issue when deploying CNN-based waste classification models across different locations, 9 pairs of camera exposure and gain values have been configured to capture the same object under various light conditions. A series of models (each for a light condition) were trained using pre-trained Resnet18 and finetuning techniques. Then they were applied to predict new waste items in not only their trained light condition but also other conditions. The results have shown that models trained in very dark or very bright light condition drastically lose their good performance when testing on objects captured under light level far from the training situation. Whereas, models based on more neutral light level could still maintain moderate accuracies when testing on extreme light levels. Moreover, different waste objects demonstrated different sensitivity patterns to the light influence, mostly related to their surface reflectance properties like how shiny or transparent they are. Several strategies have been attempted to reduce the light influence and improve the models' performances. It was found that preprocessing test images so that their brightness values are close to the training samples had very positive effects. Meanwhile, retraining the models with brightness augmentation or instance normalization could also improve models' accuracies on varying light conditions. Overall, the main contributions of this study work are two-folds: we constructed a representative waste dataset and formed a first work studying the influence of light conditions variability on the performance of CNN trained to classify waste packages.