---
id: wiki-2026-0508-lubrication
title: Lubrication
category: 10_Wiki/Topics
status: verified
canonical_id: self
aliases: [Tribology, Friction Reduction, Bearing Lubrication]
duplicate_of: none
source_trust_level: A
confidence_score: 0.88
verification_status: applied
tags: [tribology, mechanical-engineering, materials, maintenance]
raw_sources: []
last_reinforced: 2026-05-10
github_commit: pending
tech_stack:
  language: Python
  framework: scipy, sympy, openturns
---

# Lubrication

## 매 한 줄
> **"매 sliding/rolling surface 사이에 매 friction + wear 의 minimize 의 fluid/film 의 introduction"**. Reynolds (1886) 의 thin-film equation → modern: **EHL (elastohydrodynamic), MoS₂ / DLC solid lubricants, ionic liquids, ML-driven oil-condition monitoring**. 매 mechanical reliability 의 single biggest lever — 매 ~23% global energy 의 friction/wear 에 lost (Holmberg & Erdemir 2017).

## 매 핵심

### 매 Stribeck regimes
- **Boundary**: 매 surface contact, additives carry load (boundary additive ZDDP). μ ≈ 0.1.
- **Mixed**: 매 partial film + asperity contact.
- **Hydrodynamic**: 매 full fluid film, μ ≈ 0.001–0.01.
- **EHL (Elastohydrodynamic)**: 매 high-pressure point/line contact (gear teeth, ball bearing) — 매 oil viscosity rises with p, surfaces deform elastically.

### 매 lubricant types
- **Mineral oil** (refined crude).
- **Synthetic** (PAO, ester, PAG).
- **Grease** (oil + thickener, e.g., lithium-complex).
- **Solid** (graphite, MoS₂, PTFE, DLC coating).
- **Gas** (air bearing, hard-disk head).
- **Bio / water-based** (food, marine).

### 매 key parameters
- **Viscosity** η (Pa·s); kinematic ν = η/ρ (cSt).
- **Viscosity index (VI)**: 매 sensitivity to T (higher = better).
- **Film thickness** h (μm). 매 minimum h_min / σ (composite roughness) > 3 → 매 full EHL.
- **Pressure-viscosity coefficient** α — 매 EHL 의 critical.

### 매 modern (2026)
- **Online oil sensors**: dielectric, IR, viscometer-MEMS, ferrography.
- **ML CBM**: LSTM/transformer 의 RUL prediction from oil + vibration.
- **Sustainability**: re-refined oils, bio-based esters, dry / minimum-quantity lubrication (MQL) machining.

## 💻 패턴

### Reynolds equation (1-D, slider bearing)
```python
import numpy as np
from scipy.integrate import solve_bvp

eta = 0.05     # Pa·s
U   = 5.0      # m/s
L   = 0.05     # m
h0, h1 = 50e-6, 25e-6
def h(x): return h0 + (h1-h0)*x/L

def rhs(x, y):
    p, dpdx = y
    H = h(x)
    return np.vstack([dpdx, 6*eta*U/H**3 - 3*dpdx/H * (h1-h0)/L])

def bc(ya, yb): return np.array([ya[0], yb[0]])  # p=0 at both ends
xs = np.linspace(0, L, 50); y0 = np.zeros((2, xs.size))
sol = solve_bvp(rhs, bc, xs, y0)
W = np.trapz(sol.sol(xs)[0], xs)   # load capacity per unit width
```

### Hamrock–Dowson EHL film thickness (point contact)
```python
def hd_central_film(W, U, G, k=1.0):
    """central film thickness ratio H_c (Hamrock–Dowson, point contact)."""
    return 2.69 * U**0.67 * G**0.53 * W**-0.067 * (1 - 0.61*np.exp(-0.73*k))
# W = w/(E'*Rx²), U = η0*us/(E'*Rx), G = α*E'   (dimensionless groups)
```

### Stribeck curve simulator
```python
import numpy as np, matplotlib.pyplot as plt
def stribeck(eta_N_over_P):
    x = np.log10(eta_N_over_P)
    return 0.001 + 0.15*np.exp(-((x+5)**2)/1.5)  # toy μ(η·N/P)
xs = np.logspace(-7, -3, 200)
plt.semilogx(xs, [stribeck(v) for v in xs])
plt.xlabel("η·N/P"); plt.ylabel("μ")
```

### Walther viscosity-temperature
```python
def walther_nu(T_C, A, B):
    Z = 10**(A - B*np.log10(T_C + 273.15))
    return Z - 0.7   # cSt
# A, B 의 fit from two data points (e.g., 40 °C and 100 °C).
```

### Oil-condition ML (RUL with LSTM, sketch)
```python
import torch, torch.nn as nn
class RULNet(nn.Module):
    def __init__(self, in_dim=8, hidden=64):
        super().__init__()
        self.lstm = nn.LSTM(in_dim, hidden, 2, batch_first=True)
        self.head = nn.Linear(hidden, 1)
    def forward(self, x):
        h,_ = self.lstm(x); return self.head(h[:, -1, :])
# inputs: viscosity, TAN, water%, particle counts, Fe ppm, Cu ppm, T, p
```

### Bearing life (L10, ISO 281)
```python
def L10_million_revs(C_N, P_N, p_exp=3):  # 3 ball, 10/3 roller
    return (C_N / P_N) ** p_exp
def L10h_hours(L10_Mrev, n_rpm):
    return L10_Mrev * 1e6 / (60 * n_rpm)
```

## 매 결정 기준
| 상황 | Approach |
|---|---|
| Sliding journal bearing | **Hydrodynamic oil + Reynolds** |
| Rolling element / gear | **EHL oil with high VI + α** |
| High T (engine, turbine) | **Synthetic ester / PAO** |
| Vacuum / radiation / dry | **Solid lubricant (MoS₂, DLC)** |
| Food, medical | **H1 food-grade / bio-based** |
| Sealed-for-life | **Grease (NLGI 2)** |
| Precision machining | **MQL (minimum-quantity lubrication)** |
| Condition monitoring | **Online sensors + ML RUL** |

**기본값**: 매 industrial machinery → **mineral/PAO oil + ZDDP additive + ISO VG selected by viscosity-temp/load**, 매 critical → **online oil + vibration CBM**.

## 🔗 Graph
- 부모: [[Tribology]]

## 🤖 LLM 활용
**언제**: 매 lubricant 의 selection table 의 first cut, 매 viscosity-temp curve 의 fitting, 매 oil-analysis report 의 interpretation, 매 CBM model architecture 의 prototyping.
**언제 X**: 매 safety-critical (aerospace, nuclear) 의 final spec — 매 OEM/standards (ISO, DIN, MIL) 직접 검증.

## ❌ 안티패턴
- **Mixing greases (incompatible thickeners)**: 매 Li + Ca-sulfonate → 매 phase-separates → 매 bearing failure.
- **Over-greasing sealed bearing**: 매 churning heat → 매 seal blowout.
- **Wrong viscosity grade**: 매 too-low → boundary/wear; too-high → drag/heat.
- **Ignoring oil age / oxidation (TAN)**: 매 acid attacks bearing.
- **Topping up only**: 매 contaminants accumulate — 매 periodic full change + filter.
- **Same oil 매 high-T + low-T 모두**: 매 VI 의 inadequate → 매 multigrade or synthetic 사용.

## 🧪 검증 / 중복
- Verified (Reynolds 1886; Hamrock & Dowson 1977; Holmberg & Erdemir 2017; Stachowiak & Batchelor *Engineering Tribology* 4th ed.; ISO 281, ASTM D2270/D341).
- 신뢰도 A.

## 🕓 Changelog
| 날짜 | 변경 |
|---|---|
| 2026-05-08 | Phase 1 placeholder |
| 2026-05-10 | Manual cleanup — Stribeck/EHL theory + 6 patterns + decision matrix |