CAJ | 학술논문

棉花在诸多影响因素下,生长过程表现为复杂的非线性,使其水-盐的响应关系难以用传统的数学模型进行精确描述。本研究基于大田棉花膜下咸淡水滴灌试验成果,采用模糊优选BP神经网络模型,对籽棉产量与灌溉水量和灌溉水矿化度的响应关系进行了模拟。结果表明：该模型的模拟结果精度良好。模拟得到的连接权重矩阵可良好地表达籽棉产量与各生长阶段微咸水处理水平之间的响应关系,在微咸水灌溉技术中具有一定的指导意义。
면화재제다영향인소하,생장과정표현위복잡적비선성,사기수-염적향응관계난이용전통적수학모형진행정학묘술。본연구기우대전면화막하함담수적관시험성과,채용모호우선BP신경망락모형,대자면산량여관개수량화관개수광화도적향응관계진행료모의。결과표명：해모형적모의결과정도량호。모의득도적련접권중구진가량호지표체자면산량여각생장계단미함수처리수평지간적향응관계,재미함수관개기술중구유일정적지도의의。
The growth process of cotton expressed a complex non-linear characteristic influence by many factors,which made the water-salt response relationship be difficult to be accurately described with the traditional mathematical models.Based on the test results of drip irrigation with salt-fresh water under film of cotton,the relationships between seed cotton yield and irrigation water and salinity were simulated by fuzzy optimization BP neural network model.The result showed that the fuzzy optimization neural network model got higher simulation precision than Jensen model,and the weight matrix expressed the relationship between seed cotton yield and the saline water treatment levels in each growth stage well.It was of some guiding significance in saline water irrigation.