ReliDC 可靠性方法
Technical Hierarchy
ISO 14224 9-level equipment taxonomy sourced from P&IDs and SLDs. Clean asset boundaries for chillers, HV transformers, PDUs, CRAHs and CDUs.
Asset Criticality Rating
6-step facilitated ranking. Safety, Environment, Reputation, Financial. 5x4 matrix. Classes A-D map to strategy, standby, spares and RCA candidates.
FMEA & Asset Strategy
RCM decision logic per SAE JA1011/1012. On-condition, scheduled, failure-finding, RTF or redesign — tuned to MTBF, cost and availability.
Embed in CMMS
Strategies, spares, RCA candidates and KPIs written back into your CMMS. Between audit cycles, the loop keeps improving.
為何是現在:香港的緊迫性。
傳統機房 5–15 kW/機櫃,AI 負載要求 40 kW+。
直冷晶片與後門熱交換器帶來新故障模式。
中電/港燈容量是硬限制。每一度閒置電力都是損失。
附表4第10/11類數據中心強制五年能源審核。
銀行須映射依賴並情景測試關鍵設施。
AI DCI 意味故障可跨 TKO ↔ 河套蔓延。
最新洞察
全部 →HKMA OR-2: the evidence pack your board needs before May 2026
A pragmatic checklist for banks and their outsourced data-center providers — mapped to severe-but-plausible scenarios, tolerance for disruption, and third-party dependency mapping.
Liquid Cooling Is Table Stakes. Reliability Isn't.
Everyone will have coolant loops. Few will have failure-mode libraries for them.
Commissioning liquid-cooling CDUs without breaking your reliability envelope
Field notes from three AI-density retrofits in Tseung Kwan O and Tsuen Wan — where the failure modes live, and the six telemetry points we now insist on before energisation.
為一間機房量身訂做的可靠性評估。
固定範圍、固定費用(視複雜度約 HKD 30–120 萬)。帶走資產基準、FMEA 風險清單與 OR-2 / BEEO 證據起步包。
