ITRI brings AI real-time temperature calibration to advanced substrate lines
ITRI exhibited an "AI dynamic real-time temperature calibration system" at TPCA Show 2026 on October 20, combining multimodal infrared sensing with AI time-series correction to reduce temperature measurement errors on highly reflective composite materials. ITRI says the technology, made a standard feature of Group Up Industrial's smart thermal process equipment, has entered low-Earth-orbit satellite and HPC advanced substrate lines, cutting material scrap by about 20% and pushing outbound defect rates below 1%. The report notes conventional thermal imaging can err by more than 5°C on reflective materials, falling short of advanced-process specs of within 2°C error and 98% uniformity.
Thermal-cure uniformity in advanced substrates affects yields for AI server and LEO satellite boards, and ITRI claims its AI compensation addresses the measurement errors conventional thermal imaging hits on reflective materials.
隨著 AI 伺服器、高效能運算（HPC）及低軌衛星通訊需求持續成長，高階載板製程對溫度控制與良率的要求日益提高。工研院今（20）日於「TPCA Show 2026」展出「AI 動態即時溫測校準系統」，透過 AI 演算與紅外線感測技術，改善高反射板材的溫度量測誤差，並已與群翊工業合作，將技術列為智慧熱製程設備標準配備，導入國際低軌衛星及 HPC 先進載板產線，可降低約 20% 材料廢品。
根據台灣電路板協會（TPCA）與工研院產科國際所調查，受惠於 AI 伺服器、高速交換器及低軌衛星通訊等應用需求，預估 2026 年全球載板產值將突破 195 億美元，年增 32.3%。隨著高階載板朝細線路、高頻高速傳輸及複合材料發展，製程中的熱固化均勻度、能源使用效率及化學槽液潔淨度，已成為影響產品良率的重要因素。
#ITRI#Group Up Industrial#TPCA Show 2026#Taiwan Printed Circuit Association