5次现场排查：终结 -45°C 液击"悬案"

A low-temperature refrigeration plant serving a −45°C process duty experienced repeated large-compressor seizures. The owner and OEM initially treated each event as a discrete machine defect. Site access was intermittent; instrumentation logs were incomplete; and spare machines were swapped without a system-level failure tree.

Reported symptoms included sudden current spikes before trip, metallic noise at start/stop, and progressive bearing distress on more than one compressor body. Oil level sight-glasses often looked “normal,” which delayed suspicion of liquid return or lubrication collapse.

Early hypotheses focused on manufacturing tolerance, bearing metallurgy, and oil grade. Those hypotheses were plausible for a single machine—but they failed the multi-unit, multi-visit pattern.

Across five visits we reconstructed suction superheat trends, evaporator circuit balance, and start-up liquid hold-up. Quantitative plant tags were incomplete; where numeric logs were missing, we relied on repeatable qualitative checks (frost patterns, circuit ΔT feel, and trip chronology) rather than inventing missing sensor values.

The failure chain was system-side: evaporator design at −45°C duty allowed intermittent liquid carry-over → suction line liquid slug → sealing-line / bearing film collapse → seizure. The compressor was the victim, not the root cause.

Corrective work focused on evaporator circuiting / liquid management and suction protection—not on serial bearing swaps alone. After the system-side fix, remaining compressors were preserved rather than condemned as defective stock.

Verification criteria were operational: repeat start/stop without liquid-noise signature, stable suction superheat on the repaired circuits, and no recurrence across a multi-week observation window agreed with the owner. Exact COP or power numbers are not published here because they were not part of the sealed failure narrative.

• Treat multi-unit seizures as a system failure tree before blaming OEM metallurgy.
• Map evaporator circuit balance at the design minimum temperature.
• Require suction superheat / liquid-return checks in commissioning and after major overhauls.
• Do not accept “oil level looks OK” as proof of an oil film under liquid-slug risk.

This note is a field-engineering postmortem for learning. It is not a PE-stamped design review, not a warranty determination, and not a substitute for OEM service manuals or local pressure-vessel / refrigeration codes.

Plant identity, exact capacity, OEM model strings, and customer names are withheld. Temperatures and qualitative failure modes are retained where they are essential to the engineering lesson and were established on site.

某服务于 −45°C 工艺负荷的低温制冷装置反复出现大型压缩机抱死。业主与主机厂最初按单机缺陷处理。现场准入间歇、仪表记录不完整，备机更换也未建立系统级故障树。

报告现象包括跳闸前电流突升、启停金属异响，以及多台机身轴承损伤逐步加重。油位视镜常显示“正常”，延误了对回液或润滑失效的怀疑。

早期假设集中在制造公差、轴承冶金与油品牌号。对单机尚可解释，但无法解释多机、多次现场的复发模式。

五次现场中，我们重建了吸气过热度趋势、蒸发器回路平衡与启动积液行为。工厂测点并不完整；缺少数值日志时，采用可重复的定性检查（结霜形态、回路温差手感、跳闸时序），而不是编造缺失测点数值。

根因链在系统侧：−45°C 工况下蒸发器设计导致间歇带液 → 吸气管液击 → 密封线/轴承油膜崩溃 → 抱死。压缩机是受害者，不是根因。

纠正措施聚焦蒸发器回路/积液管理与吸气保护，而非单纯连环换轴承。系统侧修复后，其余压缩机得以保全，不再被当作缺陷库存报废。

验证标准偏运行侧：修复回路启停无液击特征、吸气过热度稳定，并在与业主约定的数周观察窗内无复发。本文不公布精确 COP/功率——它们不属于此次密封故障叙事的必要证据。

• 多机抱死先建系统故障树，再谈主机冶金。
• 在设计最低温度下核对蒸发器回路平衡。
• 调试与大修后强制检查吸气过热度/回液风险。
• 液击风险下，“油位看起来正常”不等于油膜存在。

本文是现场工程复盘，供学习参考；不是盖章设计评审、不是质保判定，也不能替代主机服务手册或当地压力容器/制冷规范。

工厂身份、精确容量、主机型号串与客户名称已脱敏。温度与定性失效模式在构成工程教训且经现场确认的范围内予以保留。