E-Elman神经网络在冰蓄冷空调系统建模中的应用

The application of modeling in ice storage air conditioning system based on the E-Elman neural network

摘要:

冰蓄冷技术能够为空调系统带来显著的节能效果, 已广泛地应用在现代建筑当中, 准确的监测数据预测值有利于系统合理地运行; 针对实际冰蓄冷空调工程中的能源管理控制系统（energy management and control system, EMCS）数据采样周期较长所导致在系统制冰与融冰阶段数据不足的问题, 提出了一种改进的机组运行状况预测模型; 模型算法以数据变化趋势为依据, 在传统Elman中引入评价层以约束网络输出值, 增加计算针对性, 从而提高模型输出的准确性; 仿真结果表明, 此种建模方法解决了系统融冰与制冰阶段的数据突变及网络输出值局部最优解等问题, 与传统Elman网络结构相比, 其输出值更为接近测量值, 有效地提高了模型输出的真实性; 通过关联函数, 设计的模型对冷水机组的能源消耗也可起到预测作用, 进一步说明了其实用性。

Abstract:

Ice storage has been widely used in modern buildings, because of the prominent energy saving effect for air conditioning systems. However, sampling periods of energy management and control system (EMCS) are long in some projects. This leads to data deficiencies at the time of making and melting ice, and influents the output of prediction model ulteriorly. This paper analyzes the engineering data of current percentage and chilled water temperature with the data variation trend in the actual project, and designs an improved prediction model based on the E-Elman neural network. The result shows that the improved model has enhanced the facticity of predicted outputs significantly, and solves the problems about data mutation and local optimum. Meanwhile, the improved model has superior performance.