Optimization of impermeability parameter design for fireproof layer in ceramsite concrete pebble oil storage pits using improved PSO (Particle Swarm Optimization)–SVM (Support Vector Machine) hybrid model

Maintaining the integrity of oil-storage pits lined with ceramsite-aggregate concrete pebble substrates is crucial for preventing environmental contamination, especially during fire events when a sudden increase in permeability can result in hydrocarbon leakage. Traditional trial-and-error mix designs struggle to navigate the high-dimensional parameter space needed to maintain ultra-low permeability after exposure to temperatures up to 50°C. Data-driven optimization offers a systematic alternative. This study introduces an enhanced hybrid framework combining Particle Swarm Optimization (PSO) with a Support Vector Machine (SVM) regression model to predict and minimize the permeability coefficient k of fireproof mortar mixes. An adaptive PSO algorithm incorporates dynamic inertia weighting and Gaussian mutation to efficiently tune the SVM’s RBF-kernel hyperparameters (C,γ). A dataset of 240 mortar specimens varying water-cement ratio and superplasticizer dosage was evaluated under both ambient and post-fire conditions. High-definition microstructural analyses (fluorescence microscopy, SEM, MIP, XRD, TGA) elucidated the mechanistic links between mix composition, pore-network evolution, and thermal damage. On a hold-out test set, the optimized PSO-SVM achieved an RMSE of <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mn>0.96</mml:mn> <mml:mi mathvariant="normal">x</mml:mi> <mml:msup> <mml:mn>10</mml:mn> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>10</mml:mn> </mml:mrow> </mml:msup> </mml:mrow> </mml:math> m/s and <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:msup> <mml:mi>R</mml:mi> <mml:mn>2</mml:mn> </mml:msup> </mml:mrow> </mml:math> = 0.89, outperforming default parameter SVM and Response Surface Methodology baselines by 13% and 34% error reduction, respectively. Sensitivity and ANOVA analyses identified water-cement ratio as the most influential factor. The optimized formulation reduces predicted post-fire permeability by 35% compared to a conventional mix. Practical mix-design guidelines: w/c of 0.30-0.35, superplasti-cizer dosage of 0.25%–0.35%, and ceramsite content of 55%–65% are proposed to achieve <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mi mathvariant="normal">k</mml:mi> <mml:mo>&lt;</mml:mo> <mml:mn>1.0</mml:mn> <mml:mo> </mml:mo> <mml:mi mathvariant="normal">x</mml:mi> <mml:mo> </mml:mo> <mml:msup> <mml:mn>10</mml:mn> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>10</mml:mn> </mml:mrow> </mml:msup> </mml:mrow> </mml:math> m/s while retaining compressive strength above 32 MPa, streamlining the development of robust, fire-resilient linings for hy-drocarbon containment.