발파로 인한 환경문제 및 암석파괴물성 예측에 ANN 알고리즘의 적용 = Application of ANN algorithms for the prediction of blasting-induced environmental problems and rock fracture properties
저자
발행사항
전주 : 전북대학교 일반대학원, 2023
학위논문사항
학위논문(박사)-- 전북대학교 일반대학원 : 자원.에너지공학과 Rock Mechanics 2023. 8
발행연도
2023
작성언어
영어
주제어
발행국(도시)
전북특별자치도
형태사항
xxii, 375 p. ; 26 cm
일반주기명
지도교수: 조상호
UCI식별코드
I804:45011-000000057255
소장기관
The exponential and continuous growth of the human population has led to increasing demand for mineral resources, the utilization of underground spaces, and improved amenities worldwide. These demands have led to increasing pressure and challenges for geomechanical, rock, and geotechnical engineers and policymakers to devise practical solutions to the prevailing sustainability and environmental problems resulting from the extraction of minerals and the development of various geo-structures for human use. Rock blasting and the characterization of rock geomechanical properties are the two main means of achieving practical engineering solutions. Blasting is the most economical and flexible method for rock excavation and has been applied in various geomechanics fields, including surface and underground mining, civil engineering construction, and tunneling. Nevertheless, the confined explosive energy used in rock blasting generates not only fragments of the rock but also many unfriendly and deleterious environmental problems, such as ground vibration (PPV), air blast overpressure (AOp), and fly rock. These deleterious environmental impacts of blasting often result in higher cost overruns for mining and construction companies, and severe safety concerns for nearby human settlements. The characterization of rock geomechanical properties is a crucial task in geomechanics because the parameters of rock properties are required for the successful and efficient design of important geomechanical structures. However, the accurate characterization of rock geomechanical properties is challenging owing to the variability and inhomogeneous characteristics of rock materials and the nonuniformity of testing conditions, leading to measurement errors. An inaccurate estimation of these properties may lead to unexpected failures of geomechanical structures, safety issues, and high-budget overruns. Therefore, there is a need to develop robust prediction and characterization models to address these problems.
Several studies have been undertaken with the aim of proposing prediction models for estimating PPV and AOp and characterizing the geomechanical properties of rocks. These studies utilized different techniques, including empirical equations and analytical, theoretical, and numerical methods, with promising results. However, the techniques employed by previous researchers for the prediction of PPV and AOp were developed for mine production blasts and a single site and may not be suitable for construction excavation and tunneling projects proceeded from multiple sites and used fewer explosives but shorter distances to human settlements. In addition, most geomechanical rock property prediction models used in engineering designs and analyses consider static rock behavior but neglect the dynamic nature of most geomechanical structures and events like rock bursts. This study aims to address these challenges by developing an ANN-based approach for blast-induced PPV and AOp prediction in construction and tunnel excavations, as well as rock geomechanical property characterization under static and dynamic conditions.
An ANN-based approach was first developed for the prediction of blast-induced PPV and AOp from ten different construction excavation sites and geological terrains in South Korea. The ANN-based framework utilized 115 blast event records and geological conditions of all ten study sites as input parameters. The selected input parameters included burden (B), spacing (S), hole diameter (D), hole depth (L), stemming length (T), monitoring distance (DIS), charge per delay (Q), powder factor (PF), and rock mass rating (RMR), while both blast-induced PPV and AOp were the targeted outputs. For the blast-induced PPV, the best-performing models were the grasshopper algorithm and slime mold algorithm-optimized ANNs, ANN-GOA, and ANN-SMA with R2 = 0.9955 and 0.9820, respectively, for the validation datasets. For the blast-induced AOp, the best-performing models were the slime mold algorithm and multiverse optimized ANNs, ANN-SMA, and ANN-MVO, with R2 = 0.8670 and 0.8300, respectively, for the testing datasets. The PPV model was tested with 20 new blasting event records from two construction excavation sites not used in the validation and showed high testing results of R2 = 0.985 and 0.975 for ANN-GOA and ANN-SMA, respectively. The proposed ANN-based framework explicitly addresses the black-box, site-specific, and non-large-scale deployment of existing blast-induced PPV and AOp models by transforming the models into engineer-friendly closed-form equations for the easy estimation of PPV and AOp from rock excavation. The RMR was found to be the most sensitive parameter for the generation of PPV and AOp, and this showed the site-specific problem of the previous models. The PF and Q values of 0.25, 0.30, and 0.35 kg/m3, and between 0.25 and 4 kg, respectively, were suggested to be optimum blast design parameters to reduce PPV (mm/s) and AOp (dB) in bench blasting at the shortest distance of 10–40 m to the blast location.
A robust prediction model was developed to predict blast-induced PPV during tunnel excavation at five different tunnel sites with different cross-sectional areas and geological terrain in South Korea. 221 field databases comprising seven effective parameters, including charge per delay (Q), number of holes (n), monitoring distance (DIS), hole depth (L), rock mass rating (RMR), total charge (QT), and tunnel cross-sectional area, were used for the model development. The ANN-GOA 7-17-1 model was optimal for the developed models with R2 = 0.99062, MSE = 0.00161, and VAF = 99.0544% for the test datasets. The variable importance analysis of the input parameters showed that the total charge, rock mass rating, and tunnel cross-sectional area are important parameters for the prediction of PPV in tunnel blasting.
Next, the suitability of the developed ANN-based prediction framework was assessed for the prediction of the high strain-rate loading-dependent UCS of rocks. An ANN-based model was developed using a database of 94 direct laboratory measurements comprising six effective input parameters: rock disc sample diameter (D), rock disc sample thickness (L), strain rate (Ɛ), rock bulk density (ρ), static UCS (σc), and P-wave velocity (PWV). The salp swarm algorithm hybrid ANN (ANN-SSA 6-10-1) model was determined to be the optimum of the developed models, with the lowest error metrics and highest coefficient of correlation of R = 0.99270, RRMSE = 0.07384, VAF = 98.49%, and a20-index = 1.0000 for the testing dataset. To ensure an easy and large-scale implementation of the optimum ANN-SSA 6-10-1 and other developed ANN-based models, the models were transformed into an intuitive closed-form equation. The static UCS, strain rate, and rock disc sample diameter were found to be the most sensitive parameters to the dynamic UCS.
The ANN-based framework was extended to study the fracture behavior of rock. An ANN-based prediction model framework was developed for the ISRM-suggested Semi-Circular Bend (SCB) specimen Mode-I fracture toughness (KIc) using 121 experimental data points obtained from the fracture toughness tests on SCB rock specimens. Four effective parameters affecting KIc, namely the Brazilian tensile strength (σt), disc specimen radius (R), thickness (T), and crack length (a), were selected as the input parameters. The ANN-GOA 4-9-1 model was determined to be the best-performing model based on the error metrics, with R = 0.98498, MSE = 0.0036, VAF = 97.02%, and a20-index = 0.96694 for the overall datasets. To ensure easy implementation of the optimum ANN-GOA 4-9-1, the model was transformed into a tractable closed-form explicit equation. The Brazilian tensile strength was found to be the most sensitive parameter for KIc. The ANN-based framework and closed-form equations proposed in this thesis provide an engineer-friendly, non-time-consuming, and reliable method for predicting blast-induced PPV and AOp in rock excavation and tunneling and for the accurate characterization of rock properties that can be integrated into geomechanical design and analysis frameworks.
서지정보 내보내기(Export)
닫기소장기관 정보
닫기권호소장정보
닫기오류접수
닫기오류 접수 확인
닫기음성서비스 신청
닫기음성서비스 신청 확인
닫기이용약관
닫기학술연구정보서비스 이용약관 (2017년 1월 1일 ~ 현재 적용)
학술연구정보서비스(이하 RISS)는 정보주체의 자유와 권리 보호를 위해 「개인정보 보호법」 및 관계 법령이 정한 바를 준수하여, 적법하게 개인정보를 처리하고 안전하게 관리하고 있습니다. 이에 「개인정보 보호법」 제30조에 따라 정보주체에게 개인정보 처리에 관한 절차 및 기준을 안내하고, 이와 관련한 고충을 신속하고 원활하게 처리할 수 있도록 하기 위하여 다음과 같이 개인정보 처리방침을 수립·공개합니다.
주요 개인정보 처리 표시(라벨링)
목 차
3년
또는 회원탈퇴시까지5년
(「전자상거래 등에서의 소비자보호에 관한3년
(「전자상거래 등에서의 소비자보호에 관한2년
이상(개인정보보호위원회 : 개인정보의 안전성 확보조치 기준)개인정보파일의 명칭 | 운영근거 / 처리목적 | 개인정보파일에 기록되는 개인정보의 항목 | 보유기간 | |
---|---|---|---|---|
학술연구정보서비스 이용자 가입정보 파일 | 한국교육학술정보원법 | 필수 | ID, 비밀번호, 성명, 생년월일, 신분(직업구분), 이메일, 소속분야, 웹진메일 수신동의 여부 | 3년 또는 탈퇴시 |
선택 | 소속기관명, 소속도서관명, 학과/부서명, 학번/직원번호, 휴대전화, 주소 |
구분 | 담당자 | 연락처 |
---|---|---|
KERIS 개인정보 보호책임자 | 정보보호본부 김태우 | - 이메일 : lsy@keris.or.kr - 전화번호 : 053-714-0439 - 팩스번호 : 053-714-0195 |
KERIS 개인정보 보호담당자 | 개인정보보호부 이상엽 | |
RISS 개인정보 보호책임자 | 대학학술본부 장금연 | - 이메일 : giltizen@keris.or.kr - 전화번호 : 053-714-0149 - 팩스번호 : 053-714-0194 |
RISS 개인정보 보호담당자 | 학술진흥부 길원진 |
자동로그아웃 안내
닫기인증오류 안내
닫기귀하께서는 휴면계정 전환 후 1년동안 회원정보 수집 및 이용에 대한
재동의를 하지 않으신 관계로 개인정보가 삭제되었습니다.
(참조 : RISS 이용약관 및 개인정보처리방침)
신규회원으로 가입하여 이용 부탁 드리며, 추가 문의는 고객센터로 연락 바랍니다.
- 기존 아이디 재사용 불가
휴면계정 안내
RISS는 [표준개인정보 보호지침]에 따라 2년을 주기로 개인정보 수집·이용에 관하여 (재)동의를 받고 있으며, (재)동의를 하지 않을 경우, 휴면계정으로 전환됩니다.
(※ 휴면계정은 원문이용 및 복사/대출 서비스를 이용할 수 없습니다.)
휴면계정으로 전환된 후 1년간 회원정보 수집·이용에 대한 재동의를 하지 않을 경우, RISS에서 자동탈퇴 및 개인정보가 삭제처리 됩니다.
고객센터 1599-3122
ARS번호+1번(회원가입 및 정보수정)