🌡️ Shell & Tube Heat Exchanger Fouling Dataset

Simulated dataset of thermal fouling resistance over time in a Shell & Tube heat exchanger, generated using the Epstein (1993) Arrhenius + shear-removal model. Designed for predictive maintenance, thermal degradation modeling, and ML benchmarking.

📐 Physical Model

The fouling resistance Rf [m²·K/W] evolves according to:

dRf/dt = A_dep · exp(−E_dep / RT) − A_rem · τ_w · Rf

Term	Description	Source
A_dep · exp(−E_dep/RT)	Arrhenius deposition rate	Polley et al. (2002)
A_rem · τ_w · Rf	Shear-driven removal	Kern & Seaton (1959)
τ_w = (f/8) · ρ · u²	Wall shear stress (Blasius)	Incropera & DeWitt (2002)
U = 1 / (1/U_clean + Rf)	Degraded overall HTC	Fouling factor definition

This is an asymptotic fouling model: deposition dominates at startup, removal grows with Rf until steady state.

---

🔧 Heat Exchanger Geometry

Parameter	Value	Unit
Type	Shell & Tube, single-pass counter-flow	—
Tube inner diameter	19	mm
Tube outer diameter	22	mm
Tube length	4.88	m
Number of tubes	100	—
Heat transfer area	~29.1	m²

---

💧 Fluid & Operating Conditions

Parameter	Value	Unit
Fluid	Water-based process fluid	—
Density (ρ)	980	kg/m³
Dynamic viscosity (μ)	3.5×10⁻⁴	Pa·s
Thermal conductivity	0.644	W/m·K
Mass flow rate	5.0	kg/s
Reynolds number	~16,000 (turbulent)	—
T_in range	50 – 125	°C
Simulation horizon	8,760	h (1 year)
Mass flow rate (nominal)	3.0 – 7.0	kg/s

---

📊 Dataset Columns

Column	Unit	Description
T_in_C	°C	Fluid inlet temperature
T_in_K	K	Fluid inlet temperature (Kelvin)
time_h	h	Simulation time (hourly)
Re	—	Reynolds number
u_m_s	m/s	Flow velocity
tau_w_Pa	Pa	Wall shear stress
Rf_m2K_W	m²·K/W	Fouling resistance
U_overall_W_m2K	W/m²·K	Degraded overall heat transfer coefficient
U_clean_W_m2K	W/m²·K	Clean (initial) heat transfer coefficient
Q_W	W	Heat duty
Q_clean_W	W	Reference clean heat duty
thermal_efficiency	—	Q / Q_clean ratio
dP_Pa	Pa	Pressure drop (Darcy-Weisbach)
fouling_factor_TEMA	—	TEMA fouling severity class (L/H)
scenario_id	—	Unique scenario identifier (e.g. T100_Q5.0)
m_dot_nominal_kg_s	kg/s	Nominal mass flow rate — scenario variable
R_wall_m2K_W	m²·K/W	Tube wall resistance — grows with corrosion (Arrhenius)
U_total_W_m2K	W/m²·K	Overall HTC including fouling + wall corrosion
Q_total_W	W	Heat duty accounting for all degradation sources
efficiency_total	—	Q_total / Q_clean — combined degradation efficiency
degradation_source	—	Dominant degradation: fouling, corrosion, or combined

---

📚 References

1. Epstein, N. (1993). The fouling of heat exchangers. 14th International Heat Transfer Conference, Washington D.C.
2. Polley, G.T., Wilson, D.I., Yeap, B.L., Pugh, S.J. (2002). Evaluation of laboratory fouling data for application to crude oil preheat trains. Chemical Engineering Research and Design, 80(7), 713–727.
3. Kern, D.Q., Seaton, R.E. (1959). A theoretical analysis of thermal surface fouling. British Chemical Engineering, 4(5), 258–262.
4. Incropera, F.P., DeWitt, D.P. (2002). Fundamentals of Heat and Mass Transfer, 5th ed. Wiley.
5. Sinnott, R.K. (2005). Chemical Engineering Design, 4th ed. Elsevier. (tube wall corrosion resistance model)

---

🚀 Reproduce It

git clone https://github.com/YOUR_USERNAME/shell-tube-fouling
cd shell-tube-fouling
pip install -r requirements.txt
python simulate_fouling.py
# → output/shell_tube_fouling_dataset.csv

---

🎯 Suggested ML Tasks

• Regression: Predict Rf or thermal_efficiency given T_in_C and time_h
• Classification: Predict TEMA fouling class (L or H)
• Time-series: Forecast fouling curve to schedule maintenance
• Anomaly detection: Detect abnormal fouling rate

---

License

Released under CC0 1.0 Universal (Public Domain). Use freely for research and commercial purposes.