[{"data":1,"prerenderedAt":110},["ShallowReactive",2],{"hUz2AR114gwUoVd3C_v40r8RGEVg0lLY6HyYxuiMp8M":3,"portfolio-projects:/projects/cnc-tool-wear-prediction":6,"E6TgzWoYbxBv0uKBvJwEGJ4JLk1TkP58OgIpGtWx5e0":109},{"authenticated":4,"authMode":5},false,null,{"projects":7},[8,42,77],{"slug":9,"title":10,"summary":13,"problem":16,"solution":19,"detail":22,"tech":25,"outcomes":30,"period":34,"links":35,"coverImage":39,"galleryImages":40},"cnc-tool-wear-prediction",{"ko":11,"en":12},"CNC 공구 마모 예측 시스템","CNC Tool Wear Prediction System",{"ko":14,"en":15},"진동 센서 데이터 기반으로 공구 수명과 결함을 예측하는 딥러닝 시스템","A deep learning system that predicts tool life and defects from vibration sensor streams",{"ko":17,"en":18},"생산 라인에서 공구 마모를 사전에 감지하지 못해 불량률과 유지보수 비용이 증가했습니다.","Production lines lacked early wear detection, increasing defect rates and maintenance costs.",{"ko":20,"en":21},"시간/주파수 도메인 특징 추출과 엔벌로프 분석을 결합하고 TabNet, CNN, ResNet을 비교하여 일반화 성능을 개선했습니다.","Combined time-frequency feature extraction with envelope analysis and benchmarked TabNet, CNN, and ResNet to improve generalization.",{"ko":23,"en":24},"라인별 공구 마모 패턴을 정량화하고 데이터 분포 이동에 강한 피처 세트를 구성해 재학습 주기를 줄였습니다.","Quantified wear patterns by production line and built robust feature sets against data drift to reduce retraining cycles.",[26,27,28,29],"Python","TensorFlow","MATLAB","Signal Processing",[31,32,33],"Defect prediction lead-time improved","Overfitting reduced via feature selection","Model benchmark report delivered","2024",[36],{"label":37,"url":38},"Case Note","https://hello-jaemin.tistory.com","/images/project-tool-wear.svg",[41],{"image":39},{"slug":43,"title":44,"summary":47,"problem":50,"solution":53,"detail":56,"tech":59,"outcomes":66,"period":70,"links":71,"coverImage":74,"galleryImages":75},"real-time-industrial-data-collection",{"ko":45,"en":46},"실시간 산업 데이터 수집 시스템","Real-Time Industrial Data Collection System",{"ko":48,"en":49},"대규모 센서 데이터를 안정적으로 수집/저장/전달하는 백엔드 플랫폼","A backend platform for reliable ingestion, storage, and delivery of large-scale sensor data",{"ko":51,"en":52},"현장 센서 데이터가 분산되어 장애 대응과 장기 분석이 어려웠습니다.","Sensor data was fragmented across systems, slowing incident response and long-term analysis.",{"ko":54,"en":55},"NestJS API, InfluxDB 시계열 저장소, Redis 캐시, MySQL 장기 저장을 결합하고 Docker + NGINX로 운영 환경을 표준화했습니다.","Integrated NestJS APIs, InfluxDB time-series storage, Redis caching, and MySQL archival with Docker + NGINX for production standardization.",{"ko":57,"en":58},"수집-저장-조회 경로를 분리해 장애 영향 범위를 줄이고, 운영 모니터링 지표를 표준화했습니다.","Separated ingest, storage, and query paths to reduce failure blast radius and standardized operational observability metrics.",[60,61,62,63,64,65],"NestJS","InfluxDB","MySQL","Redis","Docker","NGINX",[67,68,69],"Improved data reliability in production","Lower API latency via cache layer","Containerized deployment workflow","2023-2024",[72],{"label":73,"url":38},"Architecture Summary","/images/project-data-collection.svg",[76],{"image":74},{"slug":78,"title":79,"summary":82,"problem":85,"solution":88,"detail":91,"tech":94,"outcomes":98,"period":102,"links":103,"coverImage":106,"galleryImages":107},"real-time-monitoring-dashboard",{"ko":80,"en":81},"실시간 모니터링 대시보드","Real-Time Monitoring Dashboard",{"ko":83,"en":84},"고밀도 시계열 데이터를 고성능으로 시각화하는 운영 대시보드","An operations dashboard for high-performance visualization of dense time-series data",{"ko":86,"en":87},"고주기 센서 데이터를 기존 차트로 렌더링할 때 성능 저하와 분석 지연이 발생했습니다.","Conventional charting caused performance bottlenecks and delayed analysis for high-frequency streams.",{"ko":89,"en":90},"SciChart 기반 멀티축 렌더링, 임계치 경보, 특징 기반 트렌드 분석을 도입해 실시간 추적 성능을 높였습니다.","Introduced SciChart multi-axis rendering, threshold alarms, and feature-driven trend analytics for faster real-time tracking.",{"ko":92,"en":93},"운영자 액션에 필요한 핵심 지표를 카드화하고, 이상치 상황에서 시각적 우선순위를 높여 대응 속도를 개선했습니다.","Prioritized operator action metrics via card-based layouts and improved anomaly response speed with visual urgency cues.",[95,96,97],"Vue","TypeScript","SciChart",[99,100,101],"Faster chart rendering under load","Alert visibility improved","Operator decision time reduced","2023",[104],{"label":105,"url":38},"Dashboard Notes","/images/project-dashboard.svg",[108],{"image":106},{"authenticated":4,"authMode":5},1772421941817]