Greenhouse energy modeling is a prevalent tool for optimizing greenhouse energy consumption. However, for a model to serve its intended use, it is imperative to have a high level of confidence in the precision of its predictions. In this paper, a validated greenhouse energy model for a typical small-scale greenhouse in a cold climate is developed. The model is created using TRNSYS, a building performance simulation tool, with detailed energy modeling components and a user-defined crop model. The model is calibrated to fix uncertain parameters. A sensitivity analysis is used first to identify sensible uncertain parameters, followed by a multi-stage automated calibration. The automated calibration method uses a multi-objective genetic algorithm to adjust the uncertain parameters, calibrating the model for the measured indoor air temperature and relative humidity. The model performed well during the free-floating and ventilated stages (56 days) with a combined root mean square error (RMSE) of 1.6 °C for indoor air temperature and 8.3 % for the air relative humidity. The validation process involved assessing the applicability of the calibrated model using two additional datasets. For all the cases, comparing the simulation results with indoor environment measurements resulted in an RMSE of less than 2 °C for air temperature and less than 10 % for air relative humidity; these values compare favorably to the literature. The model achieved a 3.7 % mean relative error (MRE) in estimating monthly energy consumption for a minimally heated greenhouse. Given these results, the model is deemed sufficiently accurate and applicable for future investigations.

中文翻译：

---

使用 TRNSYS 对小型温室进行能源建模、校准和验证

温室能源建模是优化温室能源消耗的常用工具。然而，为了使模型满足其预期用途，必须对其预测的精度具有高度的信心。本文针对寒冷气候下的典型小规模温室开发了一个经过验证的温室能源模型。该模型是使用 TRNSYS（一种建筑性能模拟工具）创建的，具有详细的能源建模组件和用户定义的作物模型。该模型经过校准以修复不确定的参数。首先使用灵敏度分析来识别敏感的不确定参数，然后进行多阶段自动校准。自动校准方法采用多目标遗传算法调整不确定参数，针对测量的室内空气温度和相对湿度来校准模型。该模型在自由漂浮和通风阶段（56 天）表现良好，室内空气温度的组合均方根误差 (RMSE) 为 1.6 °C，空气相对湿度的组合均方根误差为 8.3%。验证过程涉及使用两个附加数据集评估校准模型的适用性。对于所有情况，将模拟结果与室内环境测量结果进行比较，得出空气温度的 RMSE 小于 2 °C，空气相对湿度的 RMSE 小于 10%；这些值与文献相比毫不逊色。该模型在估算最低供暖温室的每月能源消耗时实现了 3.7% 的平均相对误差 (MRE)。鉴于这些结果，该模型被认为足够准确并且适用于未来的调查。