Yongjian Jin

Executed equipment calibration for transformers, ensuring compliance with safety protocols and industry standards.

• Participated in the maintenance of 110 kV substations, mastering the skills of substation patrols.
• Developed standardized calibration procedures for SF6 density relays, reducing measurement errors by 25%.

Supported drone-based 3D mapping for grid construction, aiding pre-planning and safety control efforts.

• Conducted drone inspections of transmission lines, utilizing HD cameras and thermal imaging for defect detection.

Executed secondary safety measures for 110kV protection panels, marking 30+ cable loops for error-free restoration.

• Operated WD-7900S tester to conduct TA volt-ampere characteristic tests and validate compliance.
• Led acceptance of substation panel cabinets, identifying missing labels and coordinating rectification with manufacturers.
• Resolved telemetry data discrepancies in 35kV switchgear, ensuring 100% alignment with field conditions.

Implemented dual labeling for 220kV grounding switch wiring, eliminating risks of cross-connections.

• Participated in debugging wide-area protection systems at Hunan Electric Power Research Institute, verifying interstation strategies.
• Reduced the risk of misoperation with reduced protection devices to zero through standardized safety procedures.

• Developed simulation system for aircraft fuel measurement utilizing C# and MySQL, enhancing measurement accuracy in diverse flight attitudes.
• Engineered core algorithm in C# for accurate fuel volume calculation through integration of 3D tank models with dynamic simulations.
• Designed and managed MySQL database featuring four-dimensional function modeling for precise fuel-level mapping.

• Developed a big data model for plant classification using YOLOv5, requiring dataset annotation and model training.
• Preprocessed over 2,000 plant images while establishing annotation standards for accuracy.
• Collaborated with an international team to design project frameworks and facilitate weekly discussions on model architecture.
• Presented final model online, achieving 92% recognition accuracy in biodiversity research applications.