Fig. 11 | Scientific Reports

Fig. 11

From: Optimizing a hybrid process of electrolysis ultrasound and persulfate for remediation of petroleum contaminated soils using AI models

Fig. 11

Boxplot, Violin Plot, and Kernel Density Estimation (KDE) of the simulated Removal (%) distribution obtained from Monte Carlo simulations (n = 10,000). The simulations were based on a quadratic polynomial regression equation involving six process variables: solution pH (X₁: 3–9), moisture content (X₂: 25–200%), applied voltage (X₃: 0.1–2.5), sodium persulfate concentration (X₄: 0.5–1.5), surfactant amount (X₅: 0–3), and ultrasonication time (X₆: 50–300). All variables were sampled from uniform distributions across their respective experimental ranges. The model output was clipped to the range [0, 100] to reflect the physical limits of percentage removal.

Back to article page