【作者】 吴彩丽；

【导师】 许迪； 白美健；

【作者基本信息】 中国水利水电科学研究院 ， 水利水电工程， 2015， 博士

【摘要】 目前,地面灌溉仍是我国应用最为广泛的田间灌水方法,但现有地面灌溉田间管理粗放,灌溉过程可控制性差,致使田间灌溉效率低下,地面灌溉性能指标值普遍偏低,此问题已成为制约我国地面灌溉技术发展的瓶颈。研究地面灌溉过程实时反馈控制技术,加强对灌溉过程的控制和管理,对于提高地面灌溉田间管理水平,改进地面灌水质量具有重要的理论和实际意义。基于大量典型田块的实测水流数据,通过分析不同畦田条件下利用部分数据替代全过程数据进行参数估算和模拟预测时的误差状况,进行畦灌过程实时反馈控制技术的适应性研究；基于贝叶斯统计方法,探讨根据先验灌溉信息准确预测控制后验灌溉过程所需的合理数据量；建立以水流推进时间和地表水深的实测值与模型模拟值的平均相对误差最小为目标函数的土壤入渗参数优化反求模型,并采用遗传算法进行求解入渗参数；基于大量典型田块的实测水流数据,借助灌溉模拟模型,系统分析考虑干湿田面糙率差异与否对模拟结果造成的影响；以平均灌水深度为控制目标,通过分析不同灌溉技术要素组合下,各灌溉性能指标值对灌溉控制目标的响应关系,确定不同技术要素组合下能使灌溉性能指标值最优的适宜控制目标；构建以灌溉信息实时获取为支撑条件,以灌溉过程模拟预测为核心基础,以灌溉方案决策和实施为应用平台的畦灌过程实时反馈控制决策技术体系。通过对以上内容的研究,主要创新点及结论在于：1、基于贝叶斯统计方法,从理论上推导得出了畦灌实时反馈控制中,由部分地表水流推进数据准确估算土壤入渗参数,并模拟预测灌溉全过程,所需要采集的合理数据量。结果表明,要使部分数据推求整个水流运动过程的模拟值与实测值的平均相对误差控制在40%、25%、15%、10%和7%左右,所需的合理数据量分别为40%、50%、60%、70%和80%。2、在优化反求土壤入渗参数中同时考虑水流推进时间和地表水深因素。采用综合目标函数,即以推进时间和地表水深的实测值与借助地面灌溉模拟模型得出的对应模拟值的平均相对误差最小为目标函数。该方法比此前单纯采用推进时间目标函数的方法更为准确,比采用推进时间和消退时间目标函数的方法更有时效性,能够满足实时反馈控制快速、准确的要求。3、基于地面灌溉实时反馈控制快速、准确的要求,根据田块的土质、地形、畦长、入畦单宽流量等灌水技术要素,将田间试验与数值模拟相结合,采用理论分析与实际经验相结合的方式,研究提出能够使灌溉性能参数最优的灌水深度控制目标建议值,并优化决策灌溉关口时间,实时生成反馈控制方案。结果表明,各种灌水技术要素组合下,砂壤和黏壤土质下畦长为50m时,灌水深度控制目标取Zlq=Zreq最佳；畦长为l00m时,零坡度下灌水深度控制目标取Zmin>0最佳,有坡度下取Zlq=Zerq最佳；畦长为150m时,除黏壤土有坡度田面平整精度好的情况采用Zlq=Zerq最佳外,其他建议采用Zmin>0。该研究成果提出的不同灌溉技术要素组合下的控制目标建议值可为畦灌过程实时反馈控制方案的制定提供科学依据。针对加强畦灌过程实时反馈控制的研究是当前改进地面灌溉技术的重点所在,有助于加强地面灌溉过程的可控性,提升地面灌溉田间灌溉管理水平,提高田间灌溉效率,该研究成果对于改善我国地面灌溉方式落后的现状、缓解农业水资源短缺的矛盾、促进灌溉农业的可持续发展,具有十分重要的现实意义和必要性。更多还原

【Abstract】 At present, surface irrigation is still the most widely used field irrigation method in China. The irrigation management in field is extensive,and the control of irrigation process is poor,so it results in low field irrigation efficiency and surface irrigation performance indexes. this problem has become a bottleneck to restrict the development of surface irrigation technology in China. It has important theoretical and practical significance which studying on real-time precision control technology of surface irrigation process to trengthen the control and management of the irrigation process.Through the analysis of error condition for using partial data to estimate whole process based on a large number of measured flow data typical field, this paper researches on the adaptation of basin irrigation process real-time precision control technology; To explore the reasonable amount of data for accurately predicting irrigation process according to prior irrigaiton information based on Bayesian statistical methods; To establish the reverse optimization soil infiltration parameter model which used genetic algorithm to solve; Based on the observed data from typical field experiment and applied surface irrigation simulation model, this paper analyzes the effect on irrigation simulation account of the differences of dry and wet surface roughness or not; Through pre determine the water depth control objective which enables irrigation performance indexes to be optimized based on different irrigation elements, it can directly choose suitable water depth control objective according to irrigation elements in real-time control process.To build the system of basin irrigation process real-time precision control technology which considers obtaining information in real-time as support conditions, simulating and predicting irrigation process as core, decision and implementation of irrigation schemes as application platform.Through the study of the above content, the main conclusions and innovation point this paper lies in:(1)The average relative error between simulation value and measured values can be controlled about 40%,25%,15%,25% and 7%, when adopt the data amount of irrigation information is respectively 40%,50%,60%,70% and 80%;(2)This paper has established optimized reverse model for infiltration parameters, chosen the minimum average relative error (AREadv and AREh) for the measured and simulated values of advance time and surface water depth as the object function of the optimized reverse method of infiltration parameters, and adopted Genetic algorithm to solve the model which could satisfy the requirement of real-time control that calculation accuracy and timeliness;(3)Suitable irrigation depth control objective value needs to meet the irrigation performance indexes comprehensive optimal. Overall, Ea can achieve the optimum when the irrigation water depth control objective Zmin>0, while CU and ES can achieve optimum when the irrigation water depth control objective Zmin=Zreq. Ea, CU, and Es cannot simultaneously achieve optimum under the same control objective. The real-time control objectives for different irrigation conditions are different, when(L=50m), Zlq=Zreq is the best; When(L=100m,S0=0), Zmin>0 is the best;When(L=100m, S0≠ 0), Zlq=Zreq is the best; When(L=150m, S0≠ 0), clay loam,Sd=2cm, Zlq=Zreq is the best; but for other conditions, Zmin>0 is the best.;Through carried out the compared validation trail between basins implemented real-time control technology and adopted traditional technology, the results showed that irrigation performance indexes of basin implemented real-time control technology were superior to the basin adopted traditional technology, and it has enhanced the irrigation process management and increased the water efficiency.更多还原