Abstract A Seawater greenhouse is a desalination plant that, using solar energy and seawater, humidifies the interior of the greenhouse and produces water from the humid air. The produced water can be used both for irrigation and human consumption. Many factors affect the performance of a seawater greenhouse. In this study, using artificial neural networks, the effects of greenhouse width and length, first evaporator height, and roof transparency on the water production and energy consumption of a seawater greenhouse were examined with the help of Support Vector Regression (SVR) method. A suitable structure was obtained for this method, and %AARE, RMSE and R2 statistic measures were used for evaluating the performance of the network. This method shows the favorable correspondence with experimental data. Using the prepared optimized network, the effect of each parameter on water production and energy consumption was examined for a wide range of variations in the parameter values. Finally, a 125 m wide, 200 m long greenhouse with a 4 m high evaporator and permeability of 0.6 was found to be the optimum configuration, offering a daily water production of 161.6 m3 for 1.558 kWh of energy consumed per cubic meter of water produced.

中文翻译：

---

基于支持向量回归的加湿除湿海水温室性能参数研究

摘要 海水温室是一种海水淡化厂，它利用太阳能和海水对温室内部进行加湿，并从潮湿的空气中产生水。产出的水可用于灌溉和人类消费。许多因素会影响海水温室的性能。在这项研究中，使用人工神经网络，借助支持向量回归 (SVR) 方法研究了温室宽度和长度、第一蒸发器高度和屋顶透明度对海水温室产水量和能源消耗的影响。该方法获得了合适的结构，并使用 %AARE、RMSE 和 R2 统计量来评估网络的性能。该方法与实验数据显示出良好的对应关系。使用准备好的优化网络，检查了每个参数对水产量和能源消耗的影响，以了解参数值的各种变化。最后，发现具有 4 m 高蒸发器和 0.6 渗透率的 125 m 宽、200 m 长温室是最佳配置，每立方米产生的水可提供 161.6 m3 的日产水量和 1.558 kWh 的能源。