基于随机不确定性与非概率不确定性量化分析的中央空调系统性能预测与优化配置方法研究

Accurate performance prediction provides the basis for performance assessment, design and control optimization of centralized air-conditioning systems. However, uncertainties exist generally in the system including stochastic uncertainties in the indoor and outdoor environment variables, and non-probabilistic uncertainties related to the reliability of the system such as the failures and aging of equipment. These uncertainties make the performance prediction and design optimization more challenging. How to address the uncertainties in the centralized air conditioning systems appropriately is very important for the energy efficiency of buildings. To improve the life-cycle performance of the centralized air conditioning systems, this project attempts to develop a performance prediction and design optimization method by quantifying the stochastic and non-probabilistic uncertainties simultaneously. The Monte Carlo method is used to quantify the impacts of the stochastic uncertainties existing in both the demand and supply sides of the centralized air-conditioning system. The Markov method is adopted to quantify and analyze the non-probabilistic uncertainties related to the reliability of centralized air-conditioning systems. By integrating the quantitative results, the performance of the centralized air-conditioning system can be predicted reliably. The impact of each uncertainty is analyzed by conducting global sensitivity analysis based on random forest and influential factors are identified. By solving the non-linear equations with multiple objectives representing the system, the robust optimal design of the centralized air conditioning system can be obtained. Then the method is tested by being implemented in a practical project, and verified by comparing the performance of the obtained system with that from the conventional method. Then a software would be developed to integrate the proposed method with the building energy simulation software. This project will provide a new strategy and theory for the optimization of centralized air-conditioning systems, which is influential to achieve energy saving in buildings.

准确的性能预测是中央空调系统性能分析、设计与运行优化的基础。然而，在气象参数、室内热扰等随机不确定性和设备老化、故障等非概率不确定性的共同作用下，中央空调系统性能预测与优化配置具有一定挑战性。科学处理中央空调系统的不确定性是保证室内热舒适、提高系统能效的关键。因此，本项目以优化系统全生命周期性能为目标，提出一种基于多源不确定性量化分析的中央空调系统性能预测与优化配置方法。首先，采用蒙特卡洛法和马尔科夫法分别量化中央空调系统的随机不确定性和非概率不确定性；进而耦合两类不确定性量化结果预测系统性能，并探索各个不确定性的影响机理，辨识重要因素；然后，构建多目标非线性优化方程，求得系统配置最优解，并将该方法应用于实际项目，与传统方法结果比较，验证该方法；最后，开发可与建筑能耗模拟平台集成的不确定性分析软件，以便工程应用。本项目将为中央空调系统的优化提供新思路和理论基础，对建筑节能具有重要意义。

准确的性能预测是中央空调系统性能分析、设计与运行优化的基础。然而，在模型和参数多源不确定性的共同作用下，中央空调系统性能预测与优化配置具有一定挑战性。科学处理中央空调系统的不确定性是保证室内热舒适、提高系统能效的关键。本项目以水源热泵为研究对象，分析在水源热泵系统性能预测和设计中的不确定性来源，确定不确定性量化分析方法，考察多源不确定性对系统性能预测和设计的影响，并通过随机森林识别具有重要影响的不确定性参数。结果表明，不确定性对系统的性能预测和设计具有重要影响，忽略不确定性的影响可导致模拟评估结果偏离真实情况，可导致方案选择和设计决策错误。其中对办公建筑冷热负荷影响较为重要的因素为室外空气干球温度、太阳辐射、室外空气相对湿度、人员与新风量；影响系统能耗预测的主要不确定因素为机组COP、室外空气干球温度、太阳辐射、室外空气相对湿度、热泵机组模型、人员及新风量。当对系统进行性能评估或者设计优化时，对以上重要性参数应重点考虑，减小其不确定性，从而减小负荷或系统性能预测的不确定性。本项目将为中央空调系统的优化提供新思路和理论基础，对建筑节能具有重要意义。

内容获取失败，请点击重试