• KIEAE Journal
• /
• v.17 no.1
• /
• pp.77-82
• /
• 2017

In power supply systems for urban areas, issues such as a progressive tax have escalated recently. In this regard, photovoltaic power generation, which is appraised as an alternative power generation system, is drawing attention increasingly for its high stability and applicability to existing infrastructure. This study assessed the realistic feasibility of photovoltaic power generation and also analyzed the economic benefits expected when it is linked with green roof, which is likely to promote ecological functions in urban areas, based on the Seoul solar map, RPS, and actual monitoring data. The economics analysis of 30kW photovoltaic power generation applied with the monthly average horizontal solar radiation of six grades in the Seoul solar map showed that positive NPV was up to grade 4, while grade 5 or poorer showed negative NPV and indicated that it is difficult to assure appropriate feasibility. Compared with non-afforestation, when green roof was applied, monthly average power improvement efficiency was 7.2% at highest and 3.7% at lowest based on yearly actual monitoring data. The annual average was 5.3%, and the efficiency was high relatively in summer, including September and November. As for the economic benefits expected when 30kw photovoltaic power generation is combined with green roof based on the average horizontal solar radiation of grade 1 in the Seoul solar map, SP has improved 0.2 years to 7.4 years, and EP has improved 0.5 years to 8.3 years.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.21 no.3
• /
• pp.419-432
• /
• 2019

Most of deep learning model training was proceeded by supervised learning, which is to train labeling data composed by inputs and corresponding outputs. Labeling data was directly generated manually, so labeling accuracy of data is relatively high. However, it requires heavy efforts in securing data because of cost and time. Additionally, the main goal of supervised learning is to improve detection performance for 'True Positive' data but not to reduce occurrence of 'False Positive' data. In this paper, the occurrence of unpredictable 'False Positive' appears by trained modes with labeling data and 'True Positive' data in monitoring of deep learning-based CCTV accident detection system, which is under operation at a tunnel monitoring center. Those types of 'False Positive' to 'fire' or 'person' objects were frequently taking place for lights of working vehicle, reflecting sunlight at tunnel entrance, long black feature which occurs to the part of lane or car, etc. To solve this problem, a deep learning model was developed by simultaneously training the 'False Positive' data generated in the field and the labeling data. As a result, in comparison with the model that was trained only by the existing labeling data, the re-inference performance with respect to the labeling data was improved. In addition, re-inference of the 'False Positive' data shows that the number of 'False Positive' for the persons were more reduced in case of training model including many 'False Positive' data. By training of the 'False Positive' data, the capability of field application of the deep learning model was improved automatically.