Enhancing the vaporization surface area of the solar distillers through the implementation of a novel configurations based solar distiller represents a cost-efficient strategy for maximizing the distilled water output of conventional solar stills. The study introduces a pioneering wicked prismatic-shaped solar distiller equipped with wick materials and feed spaying nozzles, aimed at augmenting vaporization rates inside the still trough and, consequently, increasing the yield of freshwater. Two solar distillers with double slope covers are constructed and tested include a modified solar still with a prismatic basin, two incline covers, and spraying nozzles (MSS) and a reference double slope solar still (RSS). Furthermore, we have constructed a hybrid artificial intelligence framework, integrating a long short-term memory (LSTM) neural network fine-tuned through the utilization of great wall construction algorithm (GWCA). This model has been designed for the purpose of forecasting both the saltwater temperature and the associated freshwater product within the two examined solar distillers, whereas, the time, solar flux, wind velocity, and ambient temperature are considered as inputs. GWCA is effectively employed to optimize the LSTM model by determining the optimal parameter values to enhance its performance. The experimental results revealed that the daily freshwater production for the MSS reached 7.94 kg/m²/day, while the RSS achieved 5.31 kg/m²/day. This represents a substantial 49.53% improvement when compared to the RSS. Additionally, the daily energy efficiency of the MSS and RSS was assessed at 57.40% and 39.80%, respectively, whereas the daily exergy efficiency was 3.80% and 2.20%, respectively, signifying a notable 44.23% and 72.74% increase the energetic and exergetic efficiencies over RSS. Furthermore, the prediction findings demonstrated that during the testing phase, the coefficient of determination for saltwater temperature prediction of the MSS was calculated at 0.996 for LSTM-GWCA and 0.963 for LSTM. In the case of freshwater product prediction, these values were 0.983 for LSTM-GWCA and 0.922 for LSTM, respectively.

中文翻译：

---

带棱柱吸收池的创新太阳能蒸馏系统：实验分析和 LSTM 机器学习建模与长城施工算法相结合

通过实施基于新型配置的太阳能蒸馏器来增强太阳能蒸馏器的汽化表面积，代表了一种最大化传统太阳能蒸馏器的蒸馏水输出的具有成本效益的策略。该研究引入了一种配备了灯芯材料和饲料喷射喷嘴的开创性的灯芯棱柱形太阳能蒸馏器，旨在提高蒸馏槽内的蒸发率，从而提高淡水产量。建造并测试了两个具有双斜面盖的太阳能蒸馏器，包括一个带有棱柱形盆、两个斜面盖和喷嘴 (MSS) 的改进型太阳能蒸馏器以及一个参考双斜面太阳能蒸馏器 (RSS)。此外，我们构建了一个混合人工智能框架，集成了通过利用长城构建算法（GWCA）进行微调的长短期记忆（LSTM）神经网络。该模型的设计目的是预测两个被检查的太阳能蒸馏器内的盐水温度和相关的淡水产品，而时间、太阳通量、风速和环境温度被视为输入。 GWCA 被有效地用来优化 LSTM 模型，通过确定最佳参数值来提高其性能。实验结果显示，MSS的日淡水产量达到7.94公斤/平方米/天，而RSS的日淡水产量达到5.31公斤/平方米/天。与 RSS 相比，这意味着大幅提升了 49.53%。此外，MSS和RSS的每日能源效率分别为57.40%和39.80%，而每日火用效率分别为3.80%和2.20%，这意味着能量和火用效率显着提高了44.23%和72.74%通过 RSS。此外，预测结果表明，在测试阶段，LSTM-GWCA 的 MSS 盐水温度预测的决定系数计算为 0.996，LSTM 的为 0.963。在淡水产品预测的情况下，LSTM-GWCA 的这些值分别为 0.983，LSTM 的这些值分别为 0.922。