基于遗传算法的空调风系统突发污染快速溯源

发布时间：2018-03-16 11:48

  本文选题：风系统　切入点：溯源　出处：《同济大学学报(自然科学版)》2017年08期 　论文类型：期刊论文

【摘要】：为了在空调风系统内突发污染发生的短时间内辨识到污染源,进而采取相应的应急措施保障室内人员安全,建立了不同污染场景下的风系统传感器模拟浓度数据库,基于遗传算法,以传感器模拟计算浓度与实测浓度差的平方为适应度函数,研究了风系统采用不同的传感器数目、不同的传感器监测位置与时间、传感器性能参数变化及节点间距对反演结果总平均相对误差(ξ)的影响.算例结果表明:针对某一特定风系统,经位置优化后的传感器数目越多,ξ越小,同时考虑ξ的降低幅度及传感器成本,该风系统内宜设置3个传感器;在三传感器系统条件下,溯源计算收敛子代数少且收敛时间短(5s),可实现快速溯源;传感器监测位置与首次监测到污染物的时间对溯源误差的作用是耦合的,传感器监测误差对源位置x0反演没有影响,而传回数据时间间隔越大,源位置x0反演相对误差越大;最后风系统内节点总数存在一个适当范围兼顾全局寻优及降低计算负荷的作用.
[Abstract]:In order to identify the pollution sources in a short period of time when the sudden pollution occurred in the air conditioning system, and then take corresponding emergency measures to ensure the safety of indoor personnel, a simulated concentration database of air system sensors under different pollution scenarios was established. Based on genetic algorithm, the square of the difference between the simulated and measured concentration is taken as the fitness function, and the number of sensors, the monitoring position and the time of the sensor in the wind system are studied. The change of sensor performance parameters and the effect of node spacing on the total average relative error (尉) of the inversion results. The numerical results show that: for a particular wind system, the more sensors are optimized, the smaller 尉 is. At the same time, considering the reducing amplitude of 尉 and the cost of sensor, three sensors should be set up in the wind system, and under the condition of three-sensor system, the convergence subalgebra of traceability calculation is few and the convergence time is short, so the fast traceability can be realized. The effect of sensor monitoring position and the time of first monitoring pollutant on the tracing error is coupled. The sensor monitoring error has no effect on the source position x 0 inversion, but the larger the time interval is, the greater the source position x 0 inversion relative error is. Finally, the total number of nodes in the wind system has an appropriate range to take into account the global optimization and reduce the calculation load.
【作者单位】： 同济大学机械与能源工程学院;
【基金】：国家自然科学基金(51278370)
【分类号】：TU83;X51
，

本文编号：1619783