论文部分内容阅读
研究了变风量空调系统神经网络预测优化控制方法,优化指标考虑了舒适性和耗能量,舒适性指标取PMV指标,耗能量包括风机和冷水泵能耗。系统的控制量为送风风速和冷水流量,被控参数为空调区域的温湿度,采用预测滚动优化控制算法训练多层前向神经网络,然后将其作为优化反馈控制器来求解变风量暖通空调系统的优化解,并在运行中实时预测空调区域的负荷。仿真结果表明,采用此方法,在模型环境、负荷参数变化的情况下,既可以达到节能的要求,又可以使空调区域的温湿度保持在舒适范围内。
The neural network predictive optimization control method for VAV air conditioning system is studied. The optimization index considers the comfort and energy consumption. The comfort index adopts the PMV index, and the energy consumption includes the energy consumption of the fan and the cold water pump. The control volume of the system is the air velocity and the chilled water flow rate. The controlled parameters are the temperature and humidity in the air-conditioned area. The predictive rolling optimization control algorithm is used to train the multi-layer forward neural network, which is then used as the optimal feedback controller to solve the problem of variable air volume Optimization of air-conditioning system solution, and real-time forecast in the operation of air-conditioned area load. Simulation results show that this method not only meets the energy-saving requirements but also keeps the temperature and humidity in the air conditioning area within the comfortable range under the circumstances that the model environment and the load parameters change.