基于KF-LSE的建筑室内温度场实时准确重建方法

Real-time and accurate reconstruction method for indoor temperature field of building based on KF-LSE

实时、准确的建筑室内温度分布信息反馈是暖通空调（HVAC）系统控制与室内热环境优化研究的难点。针对这一问题,提出了基于卡尔曼滤波-线性随机估计（KF-LSE）的室内温度场实时准确重建方法。该方法利用本征正交分解（POD）方法将室内温度场映射为与POD模式系数相关的低阶线性系统,采用LSE方法获取POD模式系数动态模型,基于KF算法构建POD模式系数估计器,最后通过估计器与在线采集的室内温度值进行POD模式系数最优估计,快速重建出温度场。对比传统的POD-LSE方法进行建筑室内仿真实验,结果表明本方法具有更好的噪声抑制能力和温度场重建精度。

Real-time and accurate feedback of indoor temperature distribution information of buildings is the research difficult point of heating,ventilation and air conditioning（ HVAC） system control and indoor thermal environment optimization. Aiming at this problem,a real time and accurate reconstruction method of indoor temperature field based on Kalman filter-linear stochastic estimation（ KF-LSE）is proposed in this paper. Firstly,this method uses proper orthogonal decomposition（ POD） method to map the indoor temperature field to a low order linear system related to POD mode coefficients. Then,the LSE method is adopted to obtain the dynamic model of the POD mode coefficients,and the POD mode coefficient estimator is constructed based on KF algorithm. Finally,the POD mode coefficients are optimally estimated using the estimator and the indoor temperature collected online,and the fast indoor temperature field reconstruction is realized. The indoor simulation experiment was carried out. Experiment result shows that compared with traditional proper orthogonal decomposition-linear stochastic estimation（ POD-LSE） method, the proposed algorithm has better noise suppression ability and temperature field reconstruction precision.