速度-高度-神经网络优化算法在水库用水计量中的应用

针对大口径非满圆型管超声波流量计量中存在精度差的问题,基于现有理论算法,提出了一种速度—高度—神经网络优化算法。首先通过多声路超声波流量计测量出圆形管道各层声路上的流速,然后通过液位计测得非满圆形管流态下的液位高度,再采用神经网络算法建立一个三层的BP神经网络,结合测得的流速、高度信息,计算出当前的流量,同时还通过高斯-雅可比数值算法确定了声路的数目及高度。仿真结果表明,该算法提高了大口径非满圆形管工况下用水测量精度。

A Proposal for Reservoir Volume Calculation in Rainwater Harvesting Systems

This paper introduces a proposal for reservoir volume calculation in rainwater harvesting systems. The proposed method can be used for reservoir volume design in rainwater harvesting systems and is based on three important variables. These variables are water demand, system efficiency and repayment time. Several simulations were carried out in different scenarios considering typical values of both catchment area （for low-income and medium-income households） and water demand, with fixed water and tank costs. Results showed that the integrated analysis of demand, efficiency and repayment time may assist designers to determine a more adequate reservoir volume.