In recent years, radiant floor heating systems have been favored by more and more consumers because of their better comfort, high stability, and energy efficiency. This is especially true in southern China, where heating demand is increasing, and its application is becoming more common. This study explored the method of establishing a state-space model of a variable-flow radiant heating system without selecting and applying extensive measurement data and the application of the model predictive control method to its control optimization. The average errors between the state-space model and the experimental data in terms of the zone air temperatures and radiant floor surface temperatures were only −0.21–0.07 °C. The state-space model saves 76%~95% of the computation time compared to the Trnsys model, particularly suitable for large volumes and longtime building simulations. For the intermittent operation of the radiant floor heating with an air-source heat pump, the MPC controller reduced the response time by about 90 min compared with that with the PID controller, a reduction of approximately 56%. Besides, the MPC controller can effectively reduce energy consumption by 14.9% and improve the COP of the air-source heat pump by 24.5% compared with the PID controller.

近年来，辐射地板采暖系统因其更好的舒适性，高稳定性和能源效率而受到越来越多消费者的青睐。在华南地区尤其如此，那里的供热需求不断增加，其应用也越来越普遍。本研究探索了无需选择和应用大量测量数据即可建立变流辐射供暖系统状态空间模型的方法，以及将模型预测控制方法应用于其控制优化的方法。就区域空气温度和辐射地板表面温度而言，状态空间模型与实验数据之间的平均误差仅为-0.21-0.07°C。与Trnsys模型相比，状态空间模型节省了76％〜95％的计算时间，特别适用于大批量和长期的建筑模拟。对于使用空气源热泵的地板辐射采暖间歇性操作，与使用PID控制器相比，MPC控制器将响应时间减少了约90分钟，减少了约56％。此外，与PID控制器相比，MPC控制器可有效降低能耗14.9％，并将空气源热泵的COP提高24.5％。