基于随机降雨的水稻优化灌溉制度  

Optimizing Irrigation Schedule for Rice Based on Stochastically Generated Precipitation

作　　者：雷彩秀 熊卫红 姜忠 张继芳 

机构地区：[1]湖北省水利水电规划勘测设计院,武汉430064

出　　处：《灌溉排水学报》2017年第5期66-72,共7页Journal of Irrigation and Drainage

摘　　要：为了确定合理的水稻灌溉制度,在水稻灌溉试验资料的基础上,建立了以旬灌溉水量为决策变量的水稻灌溉制度优化模型,采用了蒙特卡罗方法和漳河灌区30 a实际降雨资料模拟出500 a的旬降雨,运用遗传算法解决以模拟降雨作为输入的优化模型,求得了每个模拟年份各旬在5种设定的灌溉定额条件下的灌溉水量,并通过对结果的统计分析,得到了不同灌溉定额条件下作物生长情况和各旬灌水量的概率分布。结果表明,漳河灌区30 a长序列旬降水服从伽玛分布。灌溉定额为120 mm时,中稻不发生枯萎;灌溉定额为180 mm时,基本保证中稻生长良好。以30 mm为灌水定额步长,灌溉定额为180 mm时,6月上旬和8月上旬各灌30 mm,6月中旬和7月下旬各灌60 mm。6月中旬、7月下旬和8月上旬缺水对产量的影响较大,其需水量应优先被满足。A new method based on stochastically generated precipitation series was presented in this paper to de- termine rational irrigation schedule for rice with the water required by the crop in each ten days as the decision variables; the objective was to maximize the relative yield. A long series of precipitations were first generate dusing the Monte Carlo method based onmeasured rainfall data. The genetic algorithm was then used to solve the model in which the generated precipitations were taken as input parameter. The optimal irrigation schedule for rice was obtained by statistical analysis. The model was applied to the Zhang River Irrigation District （ZID）. The results showed that 30-year precipitation series in ZID could be described by the gamma distribution. Irrigating 120 mm of water to the middle-season rice was sufficient to avoid wilting. It was also found that irrigating 180 mm of water could keep the rice grow normally. If the overall irrigation was 180 mm and the amount of each irri- gation is 30 mm, the rice would need two irrigations in early June and early August respectively, each with an amount of 30 mm, and further two irrigations in mid-June and late July respectively each with 60 mm of water. It is relatively dry in ZID from mid-June to early August, and an optimal irrigation is hence the only way to meet the demands of rice for water.

关 键 词：灌溉制度 蒙特卡罗法 水稻 随机降雨 遗传算法 

分 类 号：S274[农业科学—农业水土工程] S511[农业科学—农业工程]