In this study, nanotube-type halloysites from weathered pegmatites were investigated to absorb Pb²⁺ in an aqueous solution. Also, a novel hybrid intelligent model based on the multiple layers perceptron (MLP) neural network and the Harris hawks optimization (HHO) algorithm (i.e., HHO-MLP neural network) was proposed for estimating the absorption of Pb²⁺ from an aqueous solution using this novel material. XRD, SEM–EDS, and TEM analysis revealed the existence of overlapping tubular halloysites in the studied sample, similar to the results of previous studies. Various conditions of contact time, solution pH, the adsorbent weight, and Pb²⁺ initial concentration were considered and evaluated using batch adsorption experiments with a total of 53 cases. Subsequently, an HHO-MLP neural network was developed and applied to predict Pb²⁺ absorption efficiency in water by the nanotube-type halloysite from weathered pegmatites. A traditional MLP neural network model (without optimized by the HHO algorithm) was also investigated to predict and compare with that of the proposed HHO-MLP neural network model. The experimental results indicated that the nanotube-type halloysite from weathered pegmatites is a potential material used in processing water and removing heavy metals, i.e., Pb²⁺, with a promising development. Furthermore, the obtained results of the proposed HHO-MLP neural network model showed that this model is a robust intelligent model for estimating the efficiency of the Pb²⁺ absorption in water using nanotube-type halloysite from weathered pegmatites (i.e., MSE = 1.647; RMSE = 1.283; R² = 0.931). It can be applied to increase the Pb²⁺ absorption efficiency to eliminate Pb²⁺ in an aqueous solution.

在这项研究中，研究了来自风化伟晶岩的纳米管型埃洛石在水溶液中吸收 Pb ²⁺的能力。此外，提出了一种基于多层感知器 (MLP) 神经网络和 Harris hawks 优化 (HHO) 算法（即 HHO-MLP 神经网络）的新型混合智能模型，用于估计水溶液中 Pb ²⁺的吸收。使用这种新颖的材料。XRD、SEM-EDS 和 TEM 分析表明研究样品中存在重叠的管状埃洛石，与之前的研究结果相似。接触时间、溶液 pH 值、吸附剂重量和 Pb ^{2+ 的}各种条件使用批次吸附实验考虑和评估初始浓度，共 53 个案例。随后，开发了 HHO-MLP 神经网络并将其应用于预测来自风化伟晶岩的纳米管型埃洛石在水中的Pb ²⁺吸收效率。还研究了传统的 MLP 神经网络模型（未经 HHO 算法优化）以预测并与所提出的 HHO-MLP 神经网络模型进行比较。实验结果表明，风化伟晶岩的纳米管型埃洛石是一种潜在的用于处理水和去除重金属，即 Pb ^{2+ 的材料。}，发展前景广阔。此外，所获得的所提出的HHO-MLP神经网络模型的结果表明，该模型是用于估计的Pb的效率健壮智能模型²⁺使用从风化伟晶岩（纳米管型埃洛石在水中的吸收即MSE = 1.647; RMSE = 1.283；R ²  = 0.931）。可用于提高Pb ²⁺吸收效率以消除水溶液中的Pb ²⁺。