Applying Artificial Intelligence to Improve Ultrasonic Pulse Velocity Test in Determining Concrete Compressive Strength
碩士 === 國立高雄應用科技大學 === 土木工程與防災科技研究所 === 104 === In the construction industry, using non-destructive testing (NDT) methods (Ultrasonic Pulse Velocity Test, UPV) to examine the compressive strength of the concrete is quite economical and feasible. Without damaging the structure, it can effectively eval...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2016
|
Online Access: | http://ndltd.ncl.edu.tw/handle/00906193816315945596 |
id |
ndltd-TW-104KUAS0653023 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-104KUAS06530232016-12-22T04:19:08Z http://ndltd.ncl.edu.tw/handle/00906193816315945596 Applying Artificial Intelligence to Improve Ultrasonic Pulse Velocity Test in Determining Concrete Compressive Strength 應用人工智慧方法改善超音波檢測混凝土之抗壓強度 CHEN,YI-MIN 陳宜旻 碩士 國立高雄應用科技大學 土木工程與防災科技研究所 104 In the construction industry, using non-destructive testing (NDT) methods (Ultrasonic Pulse Velocity Test, UPV) to examine the compressive strength of the concrete is quite economical and feasible. Without damaging the structure, it can effectively evaluate the uniformity and relative quality of concrete structures. Ultrasonic Pulse Velocity has some advantages such as the low cost, easier to operate and convenient to carry. But, using UPV test to estimate concrete compressive strength have an average of over 20% mean absolute percentage error (MAPE) when comparing to the actual compressive strength obtained by destructive the tests. As a result, this research collected a total of 312 sample data from a material testing laboratory in Chia-Yi. The collected data are used to train and validate the concrete strength prediction model developed by this research. Firstly, linear and nonlinear regression prediction models are developed and then three artificial intelligence techniques, artificial neural network (ANN), support vector machine (SVM) and adaptive neural fuzzy inference system (ANFIS), are adopted to develop the AI prediction models. The objective is to establish a best prediction model for the UPV test. The results show that both of the MAPEs for the linear and nonlinear regression models are 11.17% and 17.66% respectively. The MAPEs for ANFIS, ANN and SVM models are 9.86%, 10.94% and 10.495% respectively. The research results can provide valuable information when UPV tests are used to estimate concrete compressive strength. WANG,YU-REN 王裕仁 2016 學位論文 ; thesis 105 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立高雄應用科技大學 === 土木工程與防災科技研究所 === 104 === In the construction industry, using non-destructive testing (NDT) methods (Ultrasonic Pulse Velocity Test, UPV) to examine the compressive strength of the concrete is quite economical and feasible. Without damaging the structure, it can effectively evaluate the uniformity and relative quality of concrete structures. Ultrasonic Pulse Velocity has some advantages such as the low cost, easier to operate and convenient to carry. But, using UPV test to estimate concrete compressive strength have an average of over 20% mean absolute percentage error (MAPE) when comparing to the actual compressive strength obtained by destructive the tests.
As a result, this research collected a total of 312 sample data from a material testing laboratory in Chia-Yi. The collected data are used to train and validate the concrete strength prediction model developed by this research. Firstly, linear and nonlinear regression prediction models are developed and then three artificial intelligence techniques, artificial neural network (ANN), support vector machine (SVM) and adaptive neural fuzzy inference system (ANFIS), are adopted to develop the AI prediction models. The objective is to establish a best prediction model for the UPV test. The results show that both of the MAPEs for the linear and nonlinear regression models are 11.17% and 17.66% respectively. The MAPEs for ANFIS, ANN and SVM models are 9.86%, 10.94% and 10.495% respectively. The research results can provide valuable information when UPV tests are used to estimate concrete compressive strength.
|
author2 |
WANG,YU-REN |
author_facet |
WANG,YU-REN CHEN,YI-MIN 陳宜旻 |
author |
CHEN,YI-MIN 陳宜旻 |
spellingShingle |
CHEN,YI-MIN 陳宜旻 Applying Artificial Intelligence to Improve Ultrasonic Pulse Velocity Test in Determining Concrete Compressive Strength |
author_sort |
CHEN,YI-MIN |
title |
Applying Artificial Intelligence to Improve Ultrasonic Pulse Velocity Test in Determining Concrete Compressive Strength |
title_short |
Applying Artificial Intelligence to Improve Ultrasonic Pulse Velocity Test in Determining Concrete Compressive Strength |
title_full |
Applying Artificial Intelligence to Improve Ultrasonic Pulse Velocity Test in Determining Concrete Compressive Strength |
title_fullStr |
Applying Artificial Intelligence to Improve Ultrasonic Pulse Velocity Test in Determining Concrete Compressive Strength |
title_full_unstemmed |
Applying Artificial Intelligence to Improve Ultrasonic Pulse Velocity Test in Determining Concrete Compressive Strength |
title_sort |
applying artificial intelligence to improve ultrasonic pulse velocity test in determining concrete compressive strength |
publishDate |
2016 |
url |
http://ndltd.ncl.edu.tw/handle/00906193816315945596 |
work_keys_str_mv |
AT chenyimin applyingartificialintelligencetoimproveultrasonicpulsevelocitytestindeterminingconcretecompressivestrength AT chényímín applyingartificialintelligencetoimproveultrasonicpulsevelocitytestindeterminingconcretecompressivestrength AT chenyimin yīngyòngréngōngzhìhuìfāngfǎgǎishànchāoyīnbōjiǎncèhùnníngtǔzhīkàngyāqiángdù AT chényímín yīngyòngréngōngzhìhuìfāngfǎgǎishànchāoyīnbōjiǎncèhùnníngtǔzhīkàngyāqiángdù |
_version_ |
1718405414940311552 |