結合智慧感測標籤與支持向量機快速判定混凝土裂縫位置;Combining smart sensing tags and support vector machine to determine the location of concrete cracks

NCUIR > Engineering College > Graduate Institute of Civil Engineering > Electronic Thesis & Dissertation > Item 987654321/88998

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/88998

Title:	結合智慧感測標籤與支持向量機快速判定混凝土裂縫位置;Combining smart sensing tags and support vector machine to determine the location of concrete cracks
Authors:	楊謦豪;Yang, Qing-Hao
Contributors:	土木工程學系
Keywords:	RFID 智慧感測標籤;3D列印;SVM支持向量機;AI人工智慧;RFID smart tag;3D printing;SVM support vector machine;AI artificial intelligence model
Date:	2022-08-17
Issue Date:	2022-10-04 10:46:57 (UTC+8)
Publisher:	國立中央大學
Abstract:	台灣之社會與經濟發展至今，國土上的土木建築物逐漸增多，在土地資源有限的情況下已經無法再隨意建造大型建築物了。中華民國內政部建築研究所亦提出高樓建築不能在僅僅是提供空間作居住或辦公使用，還需要考慮使用時的安全、通訊、防災、節能等問題。將高資訊科技化與人性化融入建築設計，建造智慧建築，並且以此間接提升國家競爭力。本研究結合RFID智慧感測標籤與3D列印等技術用作感測混凝土裂縫位置，並加上SVM支持向量機尋找資料特徵，以加速判斷裂縫存在與否。結合以上技術，本研究以智慧化安全防災為目標，開發一套非破壞性檢測系統，用於快速、方便且價格低廉地檢測混凝土裂縫之位置。使用RFID智慧感測標籤可以降低檢測系統中感測器的價格花費。當混凝土有裂縫產生時，該區域的混凝土滲水量將會上升，而埋置於混凝土內部的RFID智慧感測標籤就可以感測環境中的水分變化、回傳相應的數值。使用3D列印技術可以為RFID智慧感測標籤提供一定的保護，避免芯片遭受混凝土侵蝕。經由讀取器收集到的數據可由SVM支持向量機做分類，可以獲得快速且準確的結果。經過本研究的嘗試，當混凝土結構物是處於潮濕的環境中，經由SVM支持向量基分類，正確找到混凝土裂縫是否有經過的機率為98.82%。且該結果已經透過其他數項AI人工智慧模型的效果指標驗證其精度。透過正確分類數據，可以推斷產生該數據的感測區域是否有裂縫通過。;Integrate high information technology and humanization into architectural design, build smart buildings, and indirectly enhance national competitiveness. Taiwan′s social and economic development has gradually increased the number of civil buildings on the land. Given the limited land resources, building large buildings at will is no longer possible. The Architectural Research Institute of the Ministry of Internal Affairs of the Republic of China also pointed out that high-rise buildings should not only provide space for living or office use but also consider issues such as safety, communication, disaster prevention, and energy saving during use. In this study, technologies such as RFID smart sensing tags and 3D printing are used to sense the location of concrete cracks, and SVM support vector machine is used to find data features to speed up the judgment of the existence of cracks. Combining the above technologies, this research aims at intelligent safety and disaster prevention and develops a non-destructive detection system to detect the location of concrete cracks quickly, conveniently, and inexpensively. Using RFID smart sensor tags can reduce the cost of sensors in the detection system. When there are cracks in the concrete, the amount of water seepage in the concrete will increase, and the RFID smart sensor tag embedded in the concrete can sense the moisture change in the environment and return the corresponding value. 3D printing technology can provide specific protection for the RFID smart sensor tag to prevent the chip from being corroded by concrete. The data collected via the reader can be sorted by the SVM support vector machine, which can obtain fast and accurate results. After the attempts of this study, when the concrete structure is in a wet environment, the probability of correctly finding whether the concrete crack has passed through the SVM support vector basis classification is 98.82%. Moreover, the result has been verified by the performance indicators of several other AI artificial intelligence models. By correctly classifying the data, it is possible to infer whether a crack has passed through the sensing area where the data was generated.
Appears in Collections:	[Graduate Institute of Civil Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	35	View/Open

社群 sharing

Loading...