認識 Vibromera.com —— 我們的新國際網站。 造訪 Vibromera.com →
故障排除

Why Balancing Doesn't Reduce Vibration: 9 Causes and How to Fix Each One

您執行了程序,安裝了校正配重,但振動沒有變化。甚至更糟。儀器沒有壞——問題在於某個動平衡校正從未設計用來修復的項目。以下是如何找出該項目的方法。

更新 閱讀時間 13 分鐘

核心問題:動平衡校正僅修復單一項目

動平衡校正修復旋轉部件中的質量不對稱。僅此而已。轉子的質量中心與其旋轉軸不重合,因此每轉一圈都會產生使機器震動的離心力。校正配重將質量中心移回軸心。振動隨之降低。

But vibration in rotating machinery has many sources, and unbalance is only one of them. This guide covers the nine that most often make a balancing job fail: resonance, mechanical looseness, misalignment, a bent shaft, bearing defects, a dirty rotor, an undersized trial weight, an angle error, and thermal distortion. Most of them produce vibration that 看起來 like unbalance: it is synchronous or near-synchronous, it's periodic, and it shakes the machine in the radial direction. The frustrating part is that adding correction weights to a machine suffering from looseness or resonance doesn't just fail — it can make things worse.

Balanset-1A is a balancer, but it's also a vibration analyzer with FFT spectrum analysis (F8 – Charts) and a vibration meter (F5 – Vibration Meter). These diagnostic tools are the key to identifying which of the nine causes you're actually dealing with — before you waste time on trial weights.

「偽不平衡」——5 種模擬不平衡的故障

故障 #1

共振

FFT: dominant 1×, phase wanders ±10–20°

Operating speed matches the structure's natural frequency. A small unbalance force gets amplified many times. Phase wanders even while the RPM readout looks steady — real drives never hold speed perfectly, and near a resonance a fraction of a percent of speed change is enough to move the phase. This is the diagnostic clue. The Balanset-1A cannot calculate a consistent correction angle because the angle keeps changing.

故障 #2

機械鬆動

FFT:2×、3×、4×、次諧波(0.5×、1.5×)

鬆動的螺栓、軟腳、裂紋的底板、磨損的軸承座。機器的響應變得非線性——當您添加試重時,系統的「移動」方式與數學預期不同。影響係數不正確,因此校正也不正確。

故障 #3

不對心

FFT:強烈的 2× + 軸向振動升高(>徑向的 50%)

驅動軸與被驅動軸之間的角向或偏移不對心。產生模擬不平衡的力,但具有強烈的 2× 分量。如果軸向振動超過徑向振動的約 50%,請在嘗試動平衡校正前懷疑不對心。

故障 #4

軸彎曲

FFT: strong 1× (+2×), axial 1× near coupling; balance holds at one speed only

Geometric eccentricity that doesn't behave like simple mass asymmetry. The rotor does respond to a trial weight, but the balance holds only at the speed where it was done — at any other speed vibration is worse and shaft stress increases. Check runout with a dial indicator — if it exceeds 0.02–0.05 mm, straighten or replace.

故障 #5

軸承缺陷

FFT: peaks at non-integer multiples of 1×

Damaged rolling elements, pitted races, or loose outer rings. Creates vibration at bearing defect frequencies, which are non-integer multiples of shaft speed — typically 3×–12× for BPFO/BPFI, so they can sit right among the low-order harmonics. What identifies them is that the peak does not land exactly on 2×, 3×, or 4×. Balancing has no effect; replace the bearing. The Balanset-1A measures vibration velocity in the 5–1000 Hz band (no envelope analysis), so it shows developed defects, not early-stage ones.

最昂貴的失誤

不斷在鬆動機台上增加試重的技師,正在造成最大的損害。每次迭代都會不可預測地改變響應。經過三、四次失敗嘗試後,先前運行留下的校正配重會以隨機位置焊接在機台上,使未來的動平衡校正更加困難。 規則:若首次試重未產生清晰、可重複的變化(振幅 ≥20–30% 或相位 ≥20–30°),請停止。在增加更多金屬前,先進行診斷。

共振:至少會困住每個人的陷阱

Near resonance, the phase angle between the unbalance force and the vibration response shifts rapidly with tiny speed changes. If the machine runs at 1,480 RPM and the structural 自然頻率 is at 1,500 RPM, a 1% speed drift can swing the phase by roughly 10–35°, depending on how lightly the structure is damped (the lighter the damping, the sharper the swing). The balancing software sees a different angle every run and computes a different correction every time.

The first test is simple: open the Balanset-1A's Vibration Meter (F5 from the main screen), hold a constant speed and watch the phase — on screen, V1s= is the broadband RMS, V1o= the 1× component, and F1= is the 1× phase in degrees (not a function key). If the phase wanders more than 10–20° while the RPM readout is stable, you're near resonance. The Polar tab makes the drift visible: add each new reading as a point on the polar plot and watch the cluster smear. The fix is not more trial weights — it's either changing the operating speed (run at a different RPM) or modifying the structure's stiffness or mass to shift the natural frequency away from running speed.

The definitive test is a bump test or a coast-down, and the Balanset-1A has a built-in Bump Test mode: strike the housing with a soft hammer while the machine is stopped, and the software shows the structure's natural frequency along with the ring-down decay. Compare it with your running speed — if they are within a few percent, no amount of trial weights will help. On a coast-down, a resonance shows a clear amplitude peak with a ~90° phase shift at that speed.

鬆動:破壞數學模型的那一個

動平衡校正是線性代數。它假設不平衡力加倍,振動響應也會加倍。鬆動違反了這一假設。鬆動的軸承座可能在一個方向上剛性十足,而在另一個方向上卻鬆軟無力。軟腳會在特定振動振幅下將機台從一個安裝點抬起,改變循環中的有效剛性。

Before balancing any machine, check: all anchor bolts torqued, no soft foot (feeler gauge under each foot), no cracks in the baseplate, no play in bearing pedestals. If the Balanset-1A shows a "forest" of harmonics instead of a clean 1× peak — the F4-Harmonics (1x..Nx) tab in F8 – Charts displays the 1× to N× amplitudes directly — fix the structure first.

不對心:2× 特徵

聯軸器不對心主要產生 2× RPM(有時是 3×)的力。如果 Balanset-1A 的 FFT 顯示強烈的 2× 分量——尤其是伴隨高軸向振動時——問題在於對中,而非平衡。請先使用雷射對中軸系。然後檢查是否仍需要動平衡校正。通常不需要。

轉子狀態:髒污的葉輪與彎曲的軸

髒污轉子問題

灰塵、產品積聚、鈣化物沉積、腐蝕——這些任何一種附著在風機葉片、泵葉輪或離心機轉子上,都會造成質量分佈不均。機器會振動。誘惑在於「現狀」動平衡校正並恢復生產。

不要這樣做。Balanset-1A 會為髒污轉子產生校正方案。它不知道轉子髒了——它只是測量振動並計算。但這些沉積物在運行期間會剝落。在處理熱氣的風機中,一大塊水垢可能在週六凌晨 2 點掉落。現在轉子瞬間不平衡——更糟的是,因為您的校正配重原本是在補償剛剛掉落的髒污。現在這些配重成了不平衡源。

清潔後的陷阱

如果您對髒污轉子進行了動平衡校正,然後將其清潔——振動會立刻回來。您移除了已補償的質量,而校正配重仍然存在。解決方案:移除所有舊校正配重,徹底清潔轉子,然後從頭開始動平衡校正。將清潔視為第零步,而非事後補救。

彎曲軸:為何在單一速度下使用重配重無濟於事

彎曲軸會產生偏心——幾何中心與旋轉中心不重合。這在 1× RPM 時看起來像不平衡。關鍵差異在於:彎曲軸產生的振動對速度的依賴方式與簡單不平衡不同。有時您可以在特定速度下使用大校正配重降低振動,但在任何其他速度下振動都會更糟。而且軸應力會增加,縮短軸承和聯軸器的壽命。

驗證是機械性的:用手緩慢轉動軸時,使用百分表測量偏擺。如果總指示偏擺(TIR)超過機器公差——精密轉子通常為 0.02–0.05 mm,重型工業設備可達 0.1 mm——則必須校直或更換軸。動平衡校正無法修復幾何缺陷。

程序錯誤:試重、角度與溫度

有時機器是健康的,問題出在程序上。這些錯誤會讓技師認為「儀器壞了」,而實際上是輸入資料錯誤。

試重太小

Balanset-1A 透過測量系統對已知試重的響應來學習系統。如果試重太小,振幅和相位的變化會掩埋在測量雜訊中。軟體從雜訊中計算影響係數,產生的校正結果本質上是隨機的。

Target: the trial weight should change amplitude by at least 20–30% or phase by at least 20–30°. If you add 10 g and the reading barely moves, try 20 g or 30 g. Start conservatively, but don't be afraid to go bigger if needed. The math needs a clear signal. You don't have to guess the starting mass: after Run 0 the Balanset-1A software offers a First Trial Weight Estimator that computes both the trial mass and the installation angle.

How big should a trial weight be?

A safe starting point is a weight whose centrifugal force stays within 5–10% of the rotor weight. As a formula:

m試重 = k · M · g / (R · ω²), k = 0.05–0.10

其中 M is the rotor mass (kg), g = 9.81 m/s², R is the mounting radius of the weight (m), and ω = 2π·RPM/60 (rad/s). Example: a 50 kg rotor at 3,000 rpm with the weight at R = 0.2 m gives m ≈ 0.07·50·9.81/(0.2·314²) ≈ 1.7 g — grams, not hundreds of grams. That looks surprisingly small for a 50 kg rotor, but check the force: at 3,000 rpm those 1.7 g already pull F = m·R·ω² ≈ 34 N — about 7% of the rotor's weight. Starting small is the point: stepping up is cheap, while an oversized weight is a real load on bearings and welds. The same force check works in reverse for any weight you are about to bolt on.

You don't have to compute this by hand: the Balanset software includes a First Trial Weight Estimator that recommends both the mass and the starting angle from your rotor data, and a 衝擊測試 to check that you are not working near a natural frequency.

A good trial weight changes the reading by at least 20–30% in amplitude or 20–30° in phase. If the response is smaller, increase the weight stepwise — never jump straight to a heavy weight "to see something happen."

角度測量錯誤

動平衡校正是向量數學。正確角度下的 10 g 配重會抵消不平衡。相同 10 g 在與正確角度相差 180° 的位置 會使 the unbalance. How bad an angle error is, exactly: the residual after a correction placed θ off target is 2·U·sin(θ/2) of the original unbalance U. 20° off leaves 35%, 45° off leaves 77% — still an improvement. At 60° you are back where you started, and beyond 60° the machine is worse than before you touched it.

Two common errors easily cost more than 60°: measuring the angle in the wrong direction — with the Balanset-1A the correction angle is always measured from the trial weight position in the direction of rotation, and for rotors that turn counter-clockwise as seen from the sensor there is a mirror option on the polar diagram — and moving the tachometer or reflective mark between runs, which shifts the zero reference.

兩者都是隱形殺手——軟體顯示自信的校正方案,您安裝後,振動卻跳升。如果安裝計算出的校正後振動增加,首先要檢查的是角度是否以正確方向測量。

熱變形:「今天早上還好好的」問題

在繞組溫度 20°C 時動平衡校正的馬達,在 80°C 時可能會嚴重振動。處理 200–400°C 工藝氣體的熱氣風機會產生熱彎曲——軸或葉輪隨溫度上升而輕微變形,改變質量分佈。冷態下達成的平衡在熱態下會消失。

解決方案:在最終微調動平衡校正運行前,讓機器運行至熱穩態(全運轉溫度、穩定條件)。對熱態運轉的機器進行「熱態」動平衡校正。如果機器冷態到熱態的振動變化顯著,請記錄兩種狀態——有些客戶接受較高的冷啟動振動,因為他們知道機器暖機後振動會下降。

先診斷。再動平衡校正。

Balanset-1A includes FFT spectrum analysis + vibrometer mode + 1/2-plane balancing. One device for diagnostics and correction. No separate analyzer needed for unbalance, looseness, misalignment, and resonance diagnostics.

決策表:頻譜告訴您什麼?

Open F8 – Charts in the Balanset-1A software and switch to the F5-Spectrum (Hz) tab. Look at the peaks. Match the pattern to the fault.

頻譜模式相位行為最可能的故障措施
清晰的 1× 峰值,無其他諧波穩定不平衡繼續進行動平衡校正
Strong 1×, phase wanders ±10–20° while RPM reads steady不穩定共振改變轉速或修改結構
多項諧波:2×、3×、4×、次諧波不規則機械鬆動鎖緊、修正軟腳、檢查機座
強烈的 2× 訊號伴隨軸向振動升高穩定不對心使用雷射對準軸系
Strong 1× (+2×), axial 1× at the coupling; balance holds only at one speed穩定軸彎曲檢查偏擺,校直或更換
Peaks at non-integer multiples of 1× (bearing defect frequencies)不適用軸承缺陷更換軸承
Clean 1×, balance drifts back within weeks穩定Dirty rotorClean to bare metal, remove old weights, rebalance
Trial weight changes amplitude by under 20–30% and phase by under 20–30°Barely changes試重太小Increase the trial mass — use the First Trial Weight Estimator
暖機後位移的 1× 峰值隨溫度變化而位移熱變形在運轉溫度下進行動平衡校正
1× 訊號,但校正後情況惡化穩定角度誤差確認旋轉方向與參考相位
5 分鐘診斷法則

Before starting any balancing job, spend 5 minutes in the spectrum (F8 – Charts → F5-Spectrum (Hz)). If the spectrum shows a clean 1× peak with stable phase — proceed. If it shows anything else — diagnose first. This single habit eliminates the majority of failed balancing attempts. Five minutes of spectrum analysis saves an hour of futile trial weight runs.

現場報告:那台反覆出問題的風機

一家穀物加工廠來電反映一台大型引風機(45 kW,轉速 1,470 RPM)的問題。他們在六個月內已校正過三次。每次校正後:振動降至約 2 mm/s,但在 3–4 週內又回升至 8 mm/s 以上。前一位技師在每次校正後都焊接了校正配重——三次到訪共焊接了三組配重,目前仍留在葉輪上。

First thing I did was open the Balanset-1A spectrum (F8 – Charts). The FFT showed a clean 1× peak at 24.5 Hz (shaft speed) — so it looked like unbalance. Phase was stable. No looseness. No misalignment signature. That part checked out.

接著我檢查葉輪。發現厚重的穀物粉塵覆蓋,厚度 3–5 mm,分佈不均。前一位技師每次都是針對粉塵進行校正。粉塵累積、位移、部分剝落——導致振動回升。三次到訪的校正配重現在互相干擾。

我們移除了所有先前的校正配重(共三組,11 個配重)。將葉輪清潔至裸露金屬。從頭開始校正。單次雙平面校正:前平面 22 g,後平面 15 g。

現場資料 — 反覆振動

45 kW 引風機,1,470 RPM,穀物加工 — 六個月內校正 3 次

根本原因:針對隨時間位移的粉塵沉積物進行校正。移除先前三組校正配重。葉輪清潔至裸露金屬。重新進行雙平面動平衡校正。

8.4
校正前 mm/s(髒污狀態)
0.9
校正後 mm/s(清潔狀態)
89%
降低幅度
6 個月以上
穩定(仍維持中)

The plant installed a monthly cleaning schedule for the impeller. Vibration was 0.9 mm/s right after the job and 1.1 mm/s six months later — no rebalance needed. The three previous visits — removal of old weights, welding, measurement — cost more in total than a single correct diagnosis would have.

For context: a 45 kW machine falls into ISO 10816-1 Class II (medium machines, 15–75 kW), where the vibration severity zone boundaries are 1.12 / 2.80 / 7.10 mm/s RMS (A/B, B/C, C/D). The original 8.4 mm/s was deep in zone D — vibration severe enough to cause damage — while both 0.9 and 1.1 mm/s sit in zone A, the as-new range. See our 振動嚴重度圖表 for other machine classes.

動平衡校正前檢查清單

在任何機器上放置試重前,請確認此清單上的每一項。若任何檢查未通過,請先修正。對未通過其中一項檢查的機器進行動平衡校正只是浪費時間。

  1. 1
    轉子是否清潔?
    裸露金屬。無粉塵、無沉積物、無產品堆積。若無法清潔,請記錄風險並告知客戶動平衡校正可能無法維持。
  2. 2
    軸是否平直?
    使用指針式量規檢查。全跳動值(TIR)需在機器公差範圍內(精密機器 0.02–0.05 mm,重型工業機器 0.1 mm)。若超差,請校直或更換。
  3. 3
    無鬆動?
    所有螺栓已鎖緊至規定扭矩。每個機腳下使用塞尺檢查——無軟腳。底板無裂縫。軸承座堅固。頻譜:無諧波「叢林」現象。
  4. 4
    對心是否可接受?
    Axial vibration less than 50% of radial. The Balanset-1A's two channels normally mount radially — turn one sensor to the axial direction for a separate run to check this. No strong 2× in spectrum. If suspect, laser-align first.
  5. 5
    遠離共振?
    Phase stable — wandering less than ±10–20° while the RPM readout looks steady. If it wanders more, run the built-in Bump Test; change speed or modify structure before balancing.
  6. 6
    是否處於運轉溫度?
    對於高溫運轉機台:請在熱穩態下進行動平衡校正,而非冷態。若冷/熱態差異顯著,請同時記錄兩者。
  7. 7
    轉速計與參考點是否固定?
    反光標記已就位。轉速計已固定。角度方向已確認(順或逆轉)。首次運轉後,切勿移動任何參考點。
  8. 8
    Acceptance criterion agreed?
    Know the ISO 10816 zone limit for the machine before you start, and the residual-unbalance grade G (ISO 1940-1, now ISO 21940-11) you are aiming at. The Balanset-1A software has both built in: an ISO 1940 tolerance calculator on the plane panel and ISO 10816-1 machine classes in the route/trend module.

常見問題

Three common causes: (1) The correction weight was placed at the wrong angle — an error beyond 60° leaves the machine worse than before, and a full 180° error doubles the unbalance. (2) The machine is running near a resonance. There the phase response is extremely steep, so the angle measured in one run is not repeatable in the next, and the influence coefficient the software derives is unreliable — the correction lands at the wrong angle and the resonance amplifies the result. The correction mass itself does not shift the natural frequency: a few tens of grams on a rotor of hundreds of kilograms changes it by less than 0.01%. Confirm with the built-in Bump Test. (3) Mechanical looseness (loose bolts, soft foot, cracked baseplate) makes the system nonlinear, and the trial-weight calibration produces an incorrect correction. Run the FFT spectrum: if you see strong 2×, 3×, or sub-harmonic peaks, the problem is not unbalance.
The instrument will produce a correction, but you shouldn't use it. Deposits flake off later, instantly destroying the balance. Worse: your correction weights become the new unbalance source. Clean to bare metal first, then balance.
大多數情況下不行。彎曲的軸會產生幾何偏心,而非單純的質量不對稱。您可能在單一轉速下降低振動,但在其他轉速下會惡化,且軸應力增加。請使用千分表檢查偏擺——若超出公差(精密轉子為 0.02–0.05 mm),請在動平衡校正前校直或更換。
熱變形。大型馬達與高溫氣體風機在溫度上升時會產生不均勻膨脹。冷態下平衡的轉子,在熱態時質量分佈會不同。解決方案:在最終微調平衡前,運轉至熱穩態。
FFT spectrum. Clean 1× with stable phase = unbalance. Many harmonics = looseness. Strong 2× + high axial = misalignment. 1× that responds to a trial weight but holds the balance at only one speed = bent shaft. Phase that wanders ±10–20° while the RPM readout looks steady = resonance. Spend 5 minutes in the spectrum (F8 – Charts → F5-Spectrum (Hz) tab) before starting the balancing routine.
Increase the trial weight. If it changes amplitude by less than 20–30% and phase by less than 20–30°, the influence coefficients are unreliable. Size the trial weight from the rotor, not from a table: m = k·M·g/(R·ω²), with k = 0.05–0.10. Or skip the arithmetic — after Run 0 the Balanset-1A software offers a First Trial Weight Estimator that gives both the mass in grams and the installation angle. The weight should produce a visible change without making vibration dangerously high.

停止猜測。開始診斷。

Balanset-1A:FFT 頻譜 + 振動計 + 雙平面動平衡校正,整合於單一套件。診斷真實故障,修復並驗證。透過 DHL 全球配送。2 年保固。無訂閱費。


0 Comments

發佈留言

Avatar placeholder
WhatsApp
Balanset-1A · €1975諮詢工程師