HX · Predictive Intelligence

// HEAT EXCHANGER ML DASHBOARD · THERMAL EFFICIENCY MONITORING

MODEL ACTIVE
--:--:--
Best Model R²
0.9985
XGBoost · Test Set
Best RMSE
0.0106
XGBoost · Test Set
Dataset Size
10K
10 features · 20 cols
CV R² (XGB)
0.9967
5-Fold ± 0.0004
Maint. Alerts
39.6%
198 / 500 readings

Model Performance Comparison

TEST SET · N=1,500
Model R² Train R² Test RMSE MAE R² Visualization
XGBoost BEST 0.9996 0.9985 0.01060 0.00763
ANN (MLP) 0.9983 0.9976 0.01332 0.00978
Random Forest 0.9987 0.9957 0.01791 0.01288
Linear Regression 0.9159 0.9104 0.08206 0.06211
RMSE Comparison

5-Fold Cross Validation

R² SCORE
XGBoost
0.9967 ± 0.0004
Random Forest
0.9928 ± 0.0009
ANN (MLP)
0.9895 ± 0.0014
Linear Reg.
0.9153 ± 0.0054
CV R² Distribution

Feature Importance — Random Forest

MEAN DECREASE IMPURITY

Predictive Maintenance Monitoring

XGBoost · 500 READINGS
✅
Normal Operation (Steps 0–299)
Efficiency stable above baseline median
⚠️
Fouling Degradation Zone (Steps 300–500)
18% gradual LMTD reduction simulated
🚨
198 Maintenance Alerts Triggered (39.6%)
Threshold: 12% below baseline median

Operational Optimization Surface

m_cold × T_hot → Efficiency
0.500
Optimal m_cold (kg/s)
350.0 K
Optimal T_hot_in
0.7625
Peak Efficiency

Live Inference Simulator

XGBoost PREDICTOR
Hot Inlet Temperature
K
Cold Inlet Mass Flow
kg/s
Hot Outlet Temperature
K
Cold Outlet Temperature
K
Heat Load
kW
LMTD
K
PREDICTION RESULT
—
THERMAL EFFICIENCY
AWAITING INPUT
Model: XGBoost
Threshold: 0.5335
Dataset: 10,000 samples
Features: 10