In this paper, we propose a rigorous model for representing the operation of an air cooled chiller of a large office building and some simple correlations for predicting global heat exchange coefficient and pressure drop of the evaporator. The proposed models and a simplified physical one that assumes constant evaporator's heat exchange coefficient and pressure drop are used to simulate and optimize the operation of the chiller. The rigorous model shows that large variations of the order of 10% can be observed in the prediction of global heat exchange coefficient of the evaporator. We also present correlations for predicting specific enthalpies and entropies of superheated R-410, capable of representing Chemour's tabulated values within a maximum error of 2,5% for the entire expected operating region of chillers for air conditioning systems. All models are programmed as non-linear programming problems and solved by a sequential quadratic programming algorithm. Simulation and optimization results show that all physical models are easily and rapidly solved and that the more rigorous model should be preferred to optimize the operation of the chiller and to adequately predict limiting operating points and off-design situations. CPU time for all simulations was below 0,3 s using an Intel Centrino Duo type processor. Optimizing the operational point by means of the rigorous model might enable significantly reducing energy consumption depending on the disturbances affecting the operation of the chiller (e.g. of the order of 7%). We further show that the simple correlations are capable of predicting the operation of the chiller within acceptable deviations (predicting errors in compressor's power are up to 3%) and might be used to preliminarily test advanced control applications. It is also shown that evaporator's pressure drop should not be neglected as is a common practice, since prediction errors in superheating might be of the order of 3 °C. Finally, the rigorous model should be used to analyze limiting operating points and the effect of changing manipulated variables on important process constraints like superheating.

在本文中，我们提出了一个严格的模型来表示大型办公楼的风冷式冷水机组的运行，以及一些用于预测全球热交换系数和蒸发器压降的简单相关性。所提出的模型和假设蒸发器的热交换系数和压降恒定的简化物理模型用于模拟和优化冷水机组的运行。严格的模型表明，在预测蒸发器的全局热交换系数时，可以观察到 10% 量级的较大变化。我们还提供了预测过热 R-410 的比焓和熵的相关性，能够在最大误差 2 内表示 Chemour 的列表值，空调系统冷水机组的整个预期运行区域为 5%。所有模型都被编程为非线性规划问题，并通过顺序二次规划算法求解。仿真和优化结果表明，所有物理模型都可以轻松快速地求解，应该首选更严格的模型来优化冷水机组的运行，并充分预测极限运行点和非设计情况。使用 Intel Centrino Duo 型处理器时，所有模拟的 CPU 时间都低于 0.3 秒。根据影响冷水机组运行的干扰（例如，7% 左右），通过严格模型优化运行点可能会显着降低能耗。我们进一步表明，简单的相关性能够在可接受的偏差范围内预测冷水机组的运行（预测压缩机功率的误差高达 3%），并可用于初步测试高级控制应用。它还表明，蒸发器的压降不应被忽略，因为这是一种常见的做法，因为过热的预测误差可能在 3°C 的量级。最后，应该使用严格的模型来分析限制操作点以及改变操纵变量对重要过程约束（如过热）的影响。s 压降不应被忽略，因为这是一种常见的做法，因为过热的预测误差可能在 3°C 的量级。最后，应该使用严格的模型来分析限制操作点以及改变操纵变量对重要过程约束（如过热）的影响。s 压降不应被忽略，因为这是一种常见的做法，因为过热的预测误差可能在 3°C 的量级。最后，应该使用严格的模型来分析限制操作点以及改变操纵变量对重要过程约束（如过热）的影响。