张长兴, 王煜升, 刘玉峰, 孔祥强, 王清. 基于Hammerstein-Wiener模型的地埋管换热器出水温度预测[J]. 农业工程学报, 2018, 34(11): 181-186. DOI: 10.11975/j.issn.1002-6819.2018.11.023
    引用本文: 张长兴, 王煜升, 刘玉峰, 孔祥强, 王清. 基于Hammerstein-Wiener模型的地埋管换热器出水温度预测[J]. 农业工程学报, 2018, 34(11): 181-186. DOI: 10.11975/j.issn.1002-6819.2018.11.023
    Zhang Changxing, Wang Yusheng, Liu Yufeng, Kong Xiangqiang, Wang Qing. Prediction for outlet water temperature of borehole heat exchangers based on Hammerstein-Wiener model[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2018, 34(11): 181-186. DOI: 10.11975/j.issn.1002-6819.2018.11.023
    Citation: Zhang Changxing, Wang Yusheng, Liu Yufeng, Kong Xiangqiang, Wang Qing. Prediction for outlet water temperature of borehole heat exchangers based on Hammerstein-Wiener model[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2018, 34(11): 181-186. DOI: 10.11975/j.issn.1002-6819.2018.11.023

    基于Hammerstein-Wiener模型的地埋管换热器出水温度预测

    Prediction for outlet water temperature of borehole heat exchangers based on Hammerstein-Wiener model

    • 摘要: 针对土壤源热泵系统节能运行的控制需求,该文提出一种面向系统运行控制的地埋管换热器Hammerstein-Wiener(H-W)模型。基于H-W模型结构特性和非线性特征,结合地埋管换热器168 h的实时进出、水温度数据,利用Levenberg-Marquardt寻优算法对H-W模型进行辨识,在此基础上,以48 h的逐时进水温度作为模型输入,模型预测值与相应数据的拟合度为99.44%,验证了H-W模型的预测结果。在连续1 000次的H-W模型辨识与验证测试中,拟合度高于90%的占83%。地埋管换热器H-W模型的辨识速度快、预测结果精度较高,并在持续的在线预测中显示了较强的稳定性,为土壤源热泵系统在线优化控制的实施提供了保障。

       

      Abstract: Abstract: The ground-coupled heat pump system (GCHPs) has been recognized as being among the most energy efficient systems for space heating and cooling in residential and commercial buildings. Over the last several decades, many efforts had been made on appropriate design and optimal sizing of GCHPs. However, control and optimization of GCHPs were also important to improve their operating efficiency while providing satisfied indoor thermal comfort, and modeling of GCHPs was essential for appropriate analysis and improvement of its control system. The most challenging and important part of the model development for a particular application was the process of identifying the model order and the optimum parameters. The adjustment of parameters of a model was a necessary tool for the design, commissioning, operation, control, optimization and diagnostic processes in order to describe the behavior and dynamics of GCHPs. A properly identified model could provide good results, higher accuracy and minimum calculation time. As the most important part, borehole heat exchangers (BHEs) play a key role in improving the operating efficiency of GCHPs. For BHEs, time scales (larger amount of data), varying loads (thermal history) and multiple boreholes (thermal interactions) are the three major challenges to obtain the optimal design of BHEs based on the operating performance simulation of GCHPs. Most of BHEs models were based on either numerical methods or analytical approaches, and the incorporation models of the analytical and numerical solutions were also developed. Numerical methods had been used for development of response functions and research purposes, and numerical models for single BHE that utilize custom resistance networks inside the borehole had shown some promise. However, it is difficult for numerical methods or analytical approaches to be used to predict the operating performance so that on-line control strategy of GCHPs is applied. Considering the control requirements for energy-saving operation of GCHPs, a control-oriented modeling of borehole heat exchangers (BHEs) using the Hammerstein-Wiener (H-W) model was proposed in this paper. The H-W model could represent the dynamics of BHEs by a linear transfer function and capture the nonlinearities using nonlinear functions of inputs and outputs of the linear system, and combine one or two static nonlinear blocks in series with a linear block which is a discrete transfer function that represented the dynamic component of BHEs. Based on the structural characteristics and nonlinearity nature, the H-W model was identified by Levenberg-Marquardt optimization algorithm using the 168 h dynamic real-time inlet and outlet water temperature of BHEs. Furthermore, the other 48h dynamic inlet water temperatures were as the model input, the fitting goodness between the model predictive values and the corresponding data was 99.44%, so the prediction accuracy of H-W model was higher. In the 1000-time continuous tests of the H-W model identification and verification, more than 90% fitting goodness accounted for 83%. The identification of the H-W model of BHEs is fast, with high prediction accuracy and strong stability in continuous online prediction, which guaranteed the implementation of online optimization control of GCHPs.

       

    /

    返回文章
    返回