https://doi.org/10.11112/jksmi.2026.30.3.1

장주영(Juyoung Jang) ; 조민건(Mingeon Cho) ; 강건모(Gunmo Gang) ; 차기춘(Gichun Cha) ; 박승희(Seunghee Park)

Conventional safety grade estimation of infrastructure facilities is conducted based on inspection and testing results, where structural conditions are assessed according to predefined evaluation criteria. However, such assessment processes largely depend on inspectors’ experience and subjective judgment, often leading to inconsistent and less reliable outcomes. In addition, maintenance-related information is commonly stored in unstructured document formats, increasing the burden of retrieving and interpreting relevant criteria during field inspections. As a result, inspectors must manually interpret scattered information across multiple documents, which can further reduce the efficiency and objectivity of the evaluation process. To address these challenges, this study proposes a decision support system for facility safety grade estimation based on a Retrieval-Augmented Generation (RAG) architecture combined with a Large Language Model (LLM). The proposed system converts unstructured maintenance documents into a vector-based knowledge index at the paragraph level and enables evidence-grounded response generation through semantic retrieval and contextual augmentation. By retrieving relevant document segments and generating responses grounded in referenced content, the system mitigates hallucination issues in conventional LLM-based systems. Moreover, it presents explicit document evidence with responses to enhance reliability and explainability in the decision-making process. Although this study focuses on functional validation rather than quantitative performance evaluation, the results demonstrate the practical applicability of the RAG-based approach as a decision support tool for infrastructure maintenance. The proposed system is expected to improve the consistency and reliability of safety assessments and to support the digital transformation of infrastructure management practices.

https://doi.org/10.11112/jksmi.2026.30.3.9

김성완(Sung-Wan Kim) ; 전법규(Bub-Gyu Jeon) ; 장성진(Sung-Jin Chang) ; 정영수(Young-Soo Jeong) ; 박동욱(Dong-Uk Park)

Steel pipe elbows in the piping systems of nuclear power plants are vulnerable to seismic loads, thus requiring critical consideration in seismic safety assessments. The limit state, defined as the occurrence of leakage due to fatigue cracking in steel pipe elbows, was quantitatively expressed using a damage index. In this study, this limit state was quantified using a damage model that accounts for both non-cumulative deformation and cumulative energy dissipation. Furthermore, the limit load was applied as the yield point criterion to reduce the variability in the damage index arising from different yield point definitions. For the limit state evaluation, constant-amplitude displacement loads were applied to the steel pipe elbows in the in-plane direction. The damage indices were calculated by applying both individual and representative yield points to the damage model, and the effect of the variability caused by the yield point definitions on the limit state evaluation was analyzed. The results confirmed that the damage index based on the representative yield point can quantitatively evaluate the limit state of steel pipe elbows.

https://doi.org/10.11112/jksmi.2026.30.3.19

오상혁(Sang-Hyuk Oh) ; 김현중(Hyun-Joong Kim) ; 도종남(Jong-Nam Do) ; 이지영(Ji-Young Rhee)

Underground expressway structures (culverts) in Korea account for approximately 44% of all road facilities, making them a major infrastructure asset. However, because they are excluded from the statutory management framework under the Facility Safety Act, they remain in a maintenance blind spot. In particular, 51.2% of these structures have been in service for more than 20 years, and there has been an increasing number of cases in which structural defects caused by long-term ground behavior and deterioration directly lead to damage in the overlying expressway pavement and related safety incidents. To address this issue, this study was conducted in two stages. First, a five-level condition assessment and inspection manual was developed for expressway underground structures, with particular emphasis on construction joints and leakage conditions in consideration of their unique characteristics. Second, full-scale concrete specimens were fabricated and subjected to a one-year exposure test in a 3% NaCl solution to quantitatively evaluate the resistance of different waterproofing methods to chloride attack. The experimental results showed that the single-layer asphalt coating system specified in the current KDS design standard achieved only about a 15% reduction in chloride ingress relative to the unwaterproofed control. In contrast, inorganic penetrating waterproofing and sheet-type composite waterproofing systems exhibited excellent chloride-blocking performance. These findings are expected to provide an engineering basis for establishing a proactive maintenance framework for underground structures and for revising the KDS 44 90 00 design standard based on durability performance.

https://doi.org/10.11112/jksmi.2026.30.3.28

윤창호(Chang-Ho Yun) ; 박종건(Jong-Gun Park) ; 허광희(Gwang-Hee Heo)

This study investigated the effect of enhanced electrical and mechanical properties of three-phase composite conductive concrete with multi-walled carbon nanotubes (MWCNT), carbon fiber (CF), and steel fiber (SF). Accordingly, a specimen was prepared by varying the mixing contents of MWCNT, CF, and SF as experimental variables, and electrical and mechanical properties were evaluated through electrical resistance measurement, flow table test, and strength test. In addition, the microstructure of the three-phase composite conductive concrete was analyzed through scanning electron microscope (SEM) and energy-dispersive X-ray spectroscopy (EDS). The experimental results confirmed that the fluidity of the PC mixture was 215 mm, and when only MWCNT was mixed, there was a slight increase, whereas the fluidity decreased as the mixing contents of CF and SF increased. Meanwhile, in the case of three-phase composite conductive concrete mixed with MWCNT, CF, and SF, it was found that the conductive network was effectively formed, resulting in a significant reduction in electrical resistance and improved electrical properties. In addition, the mechanical properties were found to contribute to the improvement of flexural strength by enhancing crack control and load transfer capabilities through the incorporation of CF and SF. Furthermore, SEM-EDS analysis confirmed the formation of multi-scale conductive pathways in the cement matrix, which are considered to play a significant role in the simultaneous improvement of electrical and mechanical properties. Therefore, the three-phase composite conductive concrete developed in this study is expected to have high potential for use as a multifunctional construction material capable of self-heating and smart infrastructure applications.

https://doi.org/10.11112/jksmi.2026.30.3.39

김일순(Il Sun Kim) ; 최소영(So Yeong Choi) ; 양은익(Eun Ik Yang)

This study evaluates the effects of replacing natural fine aggregate with coal gangue on the workability, mechanical properties, and durability of concrete, with supplementary mortar tests. As the replacement ratio increased, slump and air content decreased, whereas mortar flow increased, indicating different flow behaviors depending on the presence of coarse aggregate. Compressive strength increased with replacement ratio, reaching up to an 11.2% improvement at 91 days. This enhancement is attributed to the combined effects of fine particle filling, pore refinement, and a slight reduction in the effective water?cement ratio. In contrast, the elastic modulus showed no significant change. The chloride diffusion coefficient and electrical resistivity improved with curing age, while the influence of the replacement ratio was limited. Overall, concrete incorporating coal gangue exhibited improved compressive strength while maintaining comparable durability. However, reduced workability requires mix design optimization, including the use of chemical admixtures, for practical application.

https://doi.org/10.11112/jksmi.2026.30.3.47

권유진(U-Jin Kwon) ; 백종하(Jong-Ha Paik) ; 백은림(Eun-Rim Baek) ; 이상호(Sang-Ho Lee)

Low-rise reinforced concrete (RC) piloti buildings are vulnerable to seismic damage because of the vertical irregularity of the open first story. When the core wall is placed eccentrically, torsional irregularity may also occur, causing displacement concentration and damage amplification in weak-side columns. However, current seismic design provisions generally evaluate torsional effects based on the elastic center of rigidity, which cannot fully reflect the inelastic shift of the center of rigidity caused by member yielding and damage. This study investigates the inelastic shift of the center of rigidity and the resulting increase in eccentricity in low-rise RC piloti buildings through one-directional nonlinear static pushover analyses. The main variables are the plan aspect ratio (AR), wall placement (W), wall-to-total rigidity ratio (R), and wall-to-total strength ratio (S). The results show that a higher wall strength ratio increases deformation concentration in weak-side columns, while a larger plan aspect ratio causes earlier column yielding under the same floor rotation. The center of rigidity moves continuously during the inelastic response, and its direction depends on the yielding sequence of the wall and columns. In some cases, the inelastic eccentricity increased up to 2.5 times the initial elastic eccentricity. These results show that the current elastic-rigidity-based evaluation cannot properly represent the actual torsional demand of torsionally irregular piloti structures. Therefore, this study discusses the need for an inelastic eccentricity ratio that reflects the effects of wall location, rigidity ratio, and strength ratio in practical torsional design.

https://doi.org/10.11112/jksmi.2026.30.3.58

김민준(Min-Jun Kim) ; 유석형(Suk-Hyeong Yoo)

This study proposes a limited-measurement-based algorithm for extracting capacity curves of buildings using acceleration data recorded at a limited number of floors. The capacity curve extraction method proposed by Kusunoki et al. (2018) can directly derive capacity curves from measured acceleration responses without requiring a detailed structural model or prior knowledge of structural properties. However, since the method requires acceleration data from all floors, its application is constrained in the Korean instrumentation environment where accelerometers are installed only at selected floors. To address this limitation, the proposed algorithm estimates the relative responses of unmeasured floors using only acceleration records at the base, an intermediate floor, and the roof. The relative displacements of unmeasured floors are estimated using a continuous displacement shape based on two-segment cubic polynomials satisfying the measured floor displacements, and the relative accelerations are obtained by linear interpolation between measured floors. Nonlinear time-history analyses of a 10-story reinforced concrete building were conducted for verification. The acceleration records from all floors were input into the existing capacity curve extraction algorithm to obtain reference results, while only the records from three measured floors were input into the proposed algorithm. The maximum representative displacement was reproduced at approximately 97.5?98.0% of the full-story measurement result, and the maximum representative acceleration was slightly overestimated at 105.8?115.5%. The overall envelope shape of the capacity curve and the representative hysteretic response were reproduced with reasonable accuracy, confirming the practical applicability of the proposed algorithm under limited measurement conditions.

https://doi.org/10.11112/jksmi.2026.30.3.69

서우석(Woo-Seok Seo) ; 이현호(Hyun-Ho Lee) ; 허무원(Moo-Won Heo) ; 김현생(Hyun-Saing Kim)

This study conducted cyclic loading tests on three reinforced concrete frame specimens to evaluate the seismic strengthening performance of reinforced concrete frames: an unreinforced specimen (BF), a strengthened specimen with a knee-brace damper installed at the lower part of the frame (RV2), and a newly proposed strengthened specimen with a steel rod damper (SRD-4). The moment?drift ratio behavior, yielding characteristics, stiffness degradation, cumulative energy dissipation capacity, longitudinal rebar strain, and damper strut strain of the specimens were comparatively analyzed. The results showed that BF exhibited the poorest behavior in terms of strength and deformation capacity, while RV2 showed superior performance in terms of initial stiffness and peak strength. In contrast, although SRD-4 exhibited a slightly lower peak moment than RV2, it maintained more gradual post-peak strength degradation and stable behavior, demonstrating superior ductile behavior up to larger displacement levels. In addition, SRD-4 showed the largest cumulative energy dissipation capacity, and the comparisons of rebar strain and damper strain indicated that deformation was more evenly distributed in SRD-4. Therefore, the proposed SRD-4 strengthening system using a steel rod damper can be considered an effective seismic strengthening method for improving the ductility and energy dissipation capacity of reinforced concrete frames.

https://doi.org/10.11112/jksmi.2026.30.3.79

김우인(Woo-In Kim) ; 오태근(Tae-Keun Oh)

This study analyzes the effectiveness and field applicability of preventive methods for reducing caught-in and collision accidents during excavator operations on construction sites. A questionnaire survey was conducted among field practitioners with experience in excavator-related work, including site supervisors, safety managers, signal persons, general workers, and equipment operators. Five preventive methods?LED lights, signal persons, warning alarms, work-zone demarcation, and rear cameras?were evaluated in terms of hazardous-zone recognition, access prevention, and ease of field application. The results showed that respondents had a very high awareness of caught-in and collision risks around excavators and strongly recognized the need for access control and visual hazard identification. Among the evaluated methods, work-zone demarcation showed the highest overall effectiveness, particularly in preventing worker access to hazardous zones, while LED lights were highly effective for visualizing hazardous zones under nighttime, low-visibility, and high-noise conditions. In contrast, rear-view cameras were considered more suitable as operator-assistance devices than as direct access-prevention methods. These findings suggest that excavator-related caught-in and collision accidents should be prevented through site-specific combinations of spatial separation, visual warnings, auditory warnings, and signal communication.

https://doi.org/10.11112/jksmi.2026.30.3.87

김경원(Kyungwon Kim) ; 심성한(Sung-Han Sim) ; 이영주(Young-Joo Lee) ; 이준화(Junhwa Lee) ; 이승준(Seungjun Lee)

Accurately measuring the long-term displacement trend of bridges during and after construction is essential for structural safety assessment. Conventional total-station-based displacement measurement requires the instrument to remain fixed for long periods, and self-motion compensation using a single fixed-point reflector has inherent limitations. This study proposes a method that places four reflectors at the outer corners and one at the center of the bridge underside, defines a base plane from the four outer points, and computes the reference-plane-based camber of the deck as the signed perpendicular distance from the center reflector to the plane. The derived camber is mathematically invariant to rotation and translation of the total station coordinate frame; thus, repeated installation and removal of the instrument does not break the temporal continuity of the camber time series, making the method suitable for intermittent monitoring over long durations. A 16-day field validation on a vehicle/pedestrian bridge showed that the proposed method repeatedly observed daily camber variations of 2-4 mm above an outer-plane residual RMS noise floor of about 0.77 mm. The weather comparison used ambient temperature and shortwave radiation from a public weather archive. The zero-lag correlation with ambient temperature was weak (Pearson r=+0.23), while a 5 h lag comparison increased the correlation to r=+0.61, indicating a delayed thermal response tendency rather than direct accuracy validation.

https://doi.org/10.11112/jksmi.2026.30.3.99

이용재(Yong-Jae Lee) ; 김욱종(Ook-Jong Kim)

This study aimed to develop steel-wire integrated deck plates that do not require end vertical reinforcement by experimentally investigating deck plates with modified structural behavior directions of lattice truss steel wires at the ends. A total of 18 specimens were tested to compare and analyze structural behavior with and without end vertical reinforcement, focusing on deflection under construction loads, residual deflection after unloading, and ultimate load capacity. The results confirmed that specimens without end vertical reinforcement exhibited stable structural behavior similar to those with reinforcement.

https://doi.org/10.11112/jksmi.2026.30.3.109

김효정(Hyo-Jung Kim) ; 알레무 아벨 쉬페라우(Alemu Abel Shiferaw) ; 현창진(Chang-Jin Hyun) ; 김형기(Hyeong-Ki Kim) ; 김윤용(Yun-Yong Kim)

This study evaluates the self-healing performance of 3D-Printable SHCC incorporating cactus stem powder (CS), a natural biopolymer material. Additionally, it verifies whether 3D-Printable SHCC incorporating CS (3DCP) exhibits strain-hardening behavior similar to that of general SHCC, as seen in conventional 3D-Printable SHCC. Tensile test results showed that the 3DCP specimen exhibited a lower maximum tensile strain than the conventional SHCC specimen. This is attributed to the relatively higher cracking strength of the 3DCP compared to conventional SHCC. Permeability tests revealed that while the 3DCP specimens showed higher initial permeability than the SHCC specimens, the permeability of the CS-incorporated 3DCP specimen decreased more rapidly as the healing age increased. When CS was incorporated, the interior of the cracks was observed to be more thoroughly filled with a healing material presumed to be calcium carbonate.

https://doi.org/10.11112/jksmi.2026.30.3.116

김봉오(Bongo Kim) ; 오태근(Tae-Keun Oh)

This study analyzes the relative importance of safety and health evaluation items for specialty contractors using the Analytic Hierarchy Process (AHP). Although specialty contractors perform a substantial portion of actual construction work, their safety management conditions differ from those of general contractors in terms of organizational size, safety personnel, documentation capacity, and dependence on prime contractors. Based on a previously developed framework, this study established an AHP hierarchy with five evaluation domains and nineteen core indicators. A pairwise comparison survey was conducted with 50 construction safety experts, and inconsistent responses were excluded using the consistency ratio criterion. The results showed that the safety and health management system had the highest priority, followed by legal compliance, hazard identification and improvement, implementation capacity, and accident response system. Key indicators included safety organization and responsibility system, management participation, emergency response, recurrence prevention, risk assessment, and legal compliance. These findings provide quantitative evidence for improving safety and health assessment systems tailored to specialty contractors.

https://doi.org/10.11112/jksmi.2026.30.3.125

이재훈(Jae-Hoon Lee) ; 허광희(Gwang-Hee Heo) ; 전승곤(Seung-Gon Jeon) ; 김충길(Chung-gil Kim)

This study developed a state-space equation-based Digital Twin (DT) model for a model cable-stayed bridge and verified its damage localization performance using the Improved Mahalanobis Distance (IMD) damage assessment method. To this end, finite element (FE) analysis and modal testing were conducted to obtain the dynamic characteristics of the target structure, and an intact-state DT model consistent with the experimental results was constructed through modal updating. Subsequently, the variations in the mass and stiffness matrices (, ) between the non-updated intact and damaged FE models were calculated and incorporated into the modal-updated intact system matrices to construct DT models for the single cable damage case (C4) and the multiple cable damage case (C3 and C5). The El-Centro earthquake record was applied to the constructed DT models as the input excitation, and the acceleration response data generated under each damage condition were analyzed using the IMD method. The damage localization performance was then evaluated based on the Upper Control Limit (UCL) exceedance ratio and the RMS values. The results showed that, in the single damage case, sensor S4 adjacent to the damaged cable C4 exhibited the highest IMD RMS value of 1.108. In the multiple damage case, sensors S3 and S6, adjacent to the damaged cables C3 and C5, exhibited RMS values of 1.185 and 1.179, respectively, confirming a double-peak distribution corresponding to the two damage locations. In addition, the DT-based IMD assessment results showed spatial distribution patterns similar to those of the measurement-based results. These findings confirm that the state-space equation-based DT model can effectively reproduce the relative response variation patterns caused by cable damage and can be used to support damage scenario assessment and damage localization of bridge structures under limited measurement conditions.