多久检查一次振动？何时应该进行动平衡检查？

设置合适的监测间隔

没有统一的时间表。"每月一次"并不总是合适的。"每季度一次"也并非总是错误的。正确的间隔取决于一件事： 从出现第一个可检测症状到功能失效，故障发展需要多长时间？ ISO 17359 calls this the "lead time to failure."

规则很简单：测量间隔应短于故障发生前一半的时间。例如，如果轴承从首次剥落到完全卡死通常需要两个月，那么至少每月测量一次。如果风扇叶轮在三周内积聚的灰尘足以改变振动模式，那么每十天检查一次。半周期规则至少能保证在故障发展过程中获得两个数据点——足以观察趋势并在故障发生前制定应对措施。

基于风险的区间选择

ISO 17359 requires a criticality and risk audit before choosing intervals, but it does not prescribe numbers. The intervals below are our field defaults, built on that framework — start there and adjust as your own trending data shows how fast faults actually develop.

这些只是起点。一旦检测到变化——例如振动水平逐渐升高、频谱中出现新的频率——就应立即提高测量频率。一台原本"每季度"测量一次的机器，一旦出现故障迹象，就应该改为"每周"测量一次。

连续型与周期型：两种方法，同一个目标

大多数工厂都会同时采用这两种方法。关键设备使用在线系统进行监控，其他设备则使用便携式仪器进行定期巡检。关键在于根据设备的重要性和故障发展速度选择合适的监控方法，而不是对整个工厂都采用同一种方法。

For periodic routes, the Balanset-1A has a built-in Route Inspection mode (F9 in the software): define your measurement points once, walk the round, and the software stores every reading and plots the overall RMS trend by date with ISO 10816 zone bands, plus separate trends for the 1× and 2× harmonics. Comparing each reading to the baseline and to previous rounds happens in the chart, not on paper.

振动趋势：应该追踪什么以及如何追踪

收集数据而不跟踪其随时间的变化是毫无意义的。振动趋势分析意味着将每次读数与基线值和之前的读数进行比较，以确定机器的性能是在改善、恶化还是保持不变。

确定基线

每台机器都需要一个参考点。在稳定且有据可查的条件下记录基线振动：匀速运转、正常负载、稳定温度。对于新机器，在调试完成后进行测量。大修后，在确定基线之前，应留出一段较短的磨合期（24-72 小时）——在磨合期间，随着轴承的磨合和部件的稳定，振动可能会发生变化。

记录振动数据的运行工况。如果没有转速、负载和温度等信息，振动读数几乎毫无意义——你无法将60%负载下的读数与100%负载下的读数进行比较。

需要追踪的内容：三层

第 1 层 — 整体 RMS 速度（毫米/秒）。 The simplest and fastest check. Compare to ISO 10816 zone boundaries (see table below; the ISO 10816 series has been superseded by the ISO 20816 series — the zone values are unchanged). A single number that tells you "good, acceptable, investigate, or act now." Use this for route efficiency — it takes 30 seconds per measurement point.

第 2 层 — 关键频率成分。 当整体水平上升时，你需要知道 为什么. Track the 1× RPM component (unbalance, looseness, buildup) and the 2× RPM component (misalignment, coupling). The Balanset-1A measures vibration velocity in the 5–1000 Hz band (its sensor response rolls off above roughly 550 Hz), so its FFT spectrum resolves the 1× component, its harmonics, and bearing tones that fall below 1000 Hz. Early-stage rolling-element bearing defects develop in a much higher frequency band and normally require envelope (demodulation) analysis, which the Balanset-1A does not provide — use it to catch unbalance, misalignment, and looseness, and a dedicated bearing analyzer for early bearing diagnostics.

第 3 层——变化率。 增长率与绝对值同样重要。一台运行稳定12个月、速度为4.5毫米/秒的机器，与一台三周前速度为2.0毫米/秒、现在速度为4.5毫米/秒的机器截然不同。快速增长意味着故障快速发展——应缩短维护周期并立即采取措施。缓慢的线性增长则有利于在下一个合适的时机进行计划性维护。

何时重新平衡：4 个基于条件的触发因素

平衡并非日程安排任务。不要在没有数据支持的情况下，将平衡工作"每 6 个月"或"每年"安排一次。只有当数据表明需要平衡时才进行平衡——并且只有在确认不平衡是主要原因时才进行平衡。

为什么平衡后振动很快又会恢复

如果在平衡作业后几天或几周内振动再次出现，请不要再次进行平衡作业——而应进行调查。反复出现的振动意味着平衡作业只是在缓解症状，而不是解决根本原因。

转子脏了。 沉积物会移动或脱落，破坏平衡。如果平衡一个脏污的叶轮，校正配重块可以补偿污垢。但当污垢移动后，这些配重块就会成为新的不平衡源。解决方法：在平衡之前，先将叶轮清洁至裸金属状态。

热变形。 转子在高温下会发生弯曲或不均匀膨胀，导致质量分布发生变化。在20°C绕组温度下进行冷态平衡的电机，在80°C时可能会出现剧烈振动。解决方法：在工作温度下进行平衡。

宽松版型。 转子在轴上发生位移、轮毂打滑或键在启动和停止过程中松动。每次启动都会使位置略微改变，因此平衡也会发生变化。解决方法：在进行平衡之前，先修复机械配合问题。

谐振。 接近结构固有频率的运行速度会放大微小的残余不平衡。由于微小的质量变化（热膨胀、沉积物移动）会被放大，机器似乎需要不断地"重新平衡"。解决方案：改变速度或修改结构以改变固有频率——请参阅我们的 隔振指南.

Field Report: 14 Months Without a Rebalance

A food processing plant in Central Europe had four identical 30 kW centrifugal fans on a drying line, each running at 2,920 RPM. The maintenance team was rebalancing all four every 3 months — a technician came in for a full day, balanced each fan, and left. Four campaigns a year, four fans each time: 16 scheduled balance jobs per year.

We set up a monthly monitoring route using the Balanset-1A in vibrometer mode. The first three months of data showed: Fan 1 and Fan 3 were stable at 1.8–2.2 mm/s overall (Zone B — acceptable for long-term operation, no action needed). Fan 2 was climbing slowly — 2.4 → 3.1 → 3.8 mm/s — with a rising 1× component indicating unbalance from product buildup on the impeller blades. Fan 4 had a strong 2× component suggesting coupling misalignment, not unbalance at all.

结果：我们对风扇 2（清洁后）进行了平衡调整，并校准了风扇 4 的联轴器。风扇 1 和 3 未做任何调整。十四个月后，风扇 1 和 3 仍然无需进行平衡调整——它们的转速分别为 2.0 毫米/秒和 2.3 毫米/秒。

The savings came from stopping unnecessary work. Two fans didn't need balancing at all. One needed alignment, not balancing. Only one actually had an unbalance problem. Monthly monitoring with a portable instrument cost 30 minutes per round — the data told the team exactly which machine needed what, and when.

ISO 10816 严重性参考

ISO 10816-3 applies to industrial machines with a rated power above 15 kW and nominal operating speeds between 120 and 15,000 RPM (the often-quoted "up to 50 MW" is the cap of its Group 1 category, not the scope of the standard). The table below shows the zone boundaries for Group 2 machines (15–300 kW) on rigid foundations. Use these as reference thresholds for your trending program. Your plant may set tighter limits based on experience.

Note: the ISO 10816 series has been superseded by the ISO 20816 series; the numerical zone boundaries below are unchanged in ISO 20816-3.

These limits apply to broadband RMS velocity measured on the bearing housings or machine structure in the 10–1000 Hz band (2–1000 Hz for machines running at or below 600 RPM). Do not compare a single narrow-band component — for example the 1× peak — with these zone boundaries; they are overall values.

这些数值适用于刚性基础上的第二组（15–300 kW）机组。对于第一组（>300 kW）机组和柔性基础，阈值有所不同——请参阅完整标准。要点：A/B 区 = 正常监测。C 区 = 调查并制定计划。D 区 = 立即采取行动。

Note that the Route Inspection trend chart on the Balanset-1A draws its zone bands from the ISO 10816-1 machine classes (I–IV), not from the ISO 10816-3 group scheme in the table above — so the zone shown on the screen and the zone from this table can differ for the same reading. Pick one scheme for your plant standard and stay with it.

常见问题