摘要
建立污染影响率指标,包括水环境承载率和水环境应激率,根据公式分别获取数据值。采用回归算法,搭建线性调度神经网络,设置训练样本,计算综合影响率,获取不同指标的数据权值,并进行权值调整,将构建的调度神经网络作为模型内在联系的规划网络,即影响率分析模型的记忆部分,通过设置的外部对应任务门,输入影响率指标提取到的具体真实数据,获取输出门数据结果,实现模型影响率的分析测评。实验数据表明,水口水电站重建会对周边水环境产生不利影响,数据标准域绝对值降低了22%,获取的影响数据离散性降低了32%,可以证明该模型影响数据的检出限更高,更具代表性。
The impact of constructing current Shuikou hydropower station on the surrounding water environment is modeled and analyzed.The study established pollution impact rate indicators,including water environment carrying capacity and water environment stress rate,obtaining data values according to formulas,using regression algorithm,building linear dispatching neural network,setting training samples,calculating comprehensive impact rate,obtaining data weights of different indicators,and adjusting the weights.It uses the constructed dispatching neural network as the planning network of the model s internal connection.That is to say,setting up training samples,calculating comprehensive impact rate,and adjusting the weights of different indicators.In the memory part of the impact rate analysis model,the specific real data extracted from the impact rate indicators are input through the external corresponding task gates,and the output gate data results are obtained to realize the analysis and evaluation of the impact rate of the model.The experimental data show that the reconstruction of Shuikou Hydropower Station will have a negative impact on the surrounding water environment.The absolute value of the data standard domain is reduced by 22%,and the discreteness of the acquired impact data is reduced by 32%.This proves that the detection limit of the impact data is higher and more representative.
作者
沈蓓蓓
Shen Beibei(Department of Water Resources Engineering,Hubei Water Resources Technical College,Wuhan 430070,China)
出处
《环境科学与管理》
CAS
2019年第11期142-146,共5页
Environmental Science and Management
关键词
水口
水电站重建
周边水环境
数据权值
nozzle
reconstruction of hydropower station
surrounding water environment
data weight