【作者】 李爱传；

【导师】 衣淑娟；

【作者基本信息】 黑龙江八一农垦大学 ， 农业机械化工程， 2017， 博士

【摘要】 随着水资源的日趋紧张,世界各个国家都在积极寻求有效的节水途径和措施。水稻是世界上栽培面积很广的主要细粮作物,用水量巨大。因此提出水稻用水的高效利用理论体系与方法刻不容缓。中国水稻全生育期用水量占农业灌溉总用水量的65%,是中国农业用水的最大用水户,对水稻应用节水灌溉措施和技术能减少农业用水的浪费,迫切且必要。我国北方种植水稻的区域尤其是三江平原地区在地理位置上属于温带湿润、半湿润大陆季风气候地区,产能很大但是水资源不太充足,地下水位下降厉害,发展后劲不足已经凸显。三江平原是我国重要的农业生产基地,面积大且平坦规则,为此,本文通过对三江平原地形、土壤、气候特点和对三江平原水稻控灌条件下用水规律进行分析和2013-2016年在黑龙江省洪河农场进行的试验,使用数据拟合技术建立洪河农场水稻日用水量与气象因子关系模型,定量分析环境平均温度、平均湿度、日均风速、日照时数、日辐射量对水稻日用水量的影响,并利用提出的寒地水稻用水模型和GALM混合算法研究寒地水稻节水灌溉自动控制系统,研究为寒地水稻控制灌溉的信息化管理提供了样本。取得的相关成果如下。1)围绕寒地水稻用水模型及算法的建立及理论开展研究,为提高水稻用水效率及水稻节水灌溉运行管理提供科学依据。2)提出了将日均温度、日均湿度、日均风速、日照时数、太阳辐射量作为自变量,把水稻日用水量作为因变量,建立了5输入1输出的多元回归方程。提出了一种分生育阶段进行高阶多元多项式建模方法,根据生育期中各生育阶段的数据拟合出适合当前生育阶段的回归经验公式。3)建立了基于遗传算法和LM优化算法的混合学习算法模型,GALM算法有效地将二者的优点相融合。实验结果表明,GALM网络经358步达到预定误差,GALM混合算法提高了网络的收敛速度,比其他算法误差更小。4)设计了以高智能芯片为核心的节水灌溉控制器,对水稻生育期的周围环境温湿度、日辐射量、日照时数、风速、日降雨量、逐日蒸发量、土壤热通量、不同埋深的泥温等因素进行数据采集,并按照GALM混合算法模型进行灌水。节水灌溉控制器采用低功耗设计和分布式控制模式,可单机独立工作,通过试验对比,提高了终端控制的可靠性。5)进行适合于北方的寒地水稻节水灌溉控制平台的设计。平台硬件设备操作方便、能耗低。平台软件具有实时数据监测、历史数据查询分析、灌溉信息预测等功能。经过测试,通信距离在2800m内均可实现正常通信,数据传输正确率100%,系统抗干扰性优异。通过现场整体性能实际测试,系统按下开始按钮10s后就开始接收各种终端传感器采集来的数据,并在5s内完成数据的向外发送。上位机接收到数据后可在20s之内完成数据的融合及转发。并可通过手机或者其他无线掌上设备发出指令实时控制,控制效率高。6)进行了多个农场田间试验,对比试验结果表明:采用用水模型及算法的控制灌溉与人工充分灌溉相比,平均节水37.68%、平均增产6.47%、平均每亩增收节支148.97元、3年累计增收43.51万元,应用前景广阔。更多还原

【Abstract】 With the water resource increasingly tense,the world’s big responsible countries are actively seeking effective ways and measures for water-saving.Rice is the second biggest cultivated area and output in the world after the main refined grain crops of wheat,the water consumption is huge.Therefore,it is urgent to put forward the theory and method of efficient use of rice water.The China’s rice consumption in the whole growth period accounts for 65% of the total water consumption of agricultural irrigation,is the largest water users in China’s agricultural water use,the water-saving irrigation measure and technique to reduce agricultural waste,which is urgent and necessary.The northern region of rice cultivation,especially in Sanjiang plain region geography position belongs to temperate humid,semi humid continental monsoon climate region,with a large capacity but less abundant water resources,also a drop in groundwater level and a lack of development potential.In the China’s northern rice cultivation area,especially in the Sanjiang Plain region is geographically arid and semi-arid areas,with a large capacity but less abundant water resources,it has a drop in groundwater level and a lack of development potential.The Sanjiang plain is the important agricultural production base in China which is large,flat and neat,therefore,the article focus on the Sanjiang plain terrain,soil and climate characteristics,Based on the analysis of the water use law under the conditions of rice irrigation in Sanjiang Plain and the experiment conducted in Honghe Farm in Heilongjiang Province from 2013 to 2016,established the Honghe farm rice requirement and meteorological factors relationship model,the quantitative analysis of environmental temperature,relative humidity,wind speed and sunshine time influence on rice water requirement through the data-fitting technique,and in-kind production,based on the proposed model of paddy field water use and the GALM hybrid algorithm,the paper studies the automatic control system of water saving irrigation in cold region.,the related results are as follows.1)Study on the theoretical system of rice water use model and algorithm in cold region,and provide scientific basis for improving the efficiency of agricultural water use and the large area of rice field irrigation management.2)The daily temperature,daily average humidity,daily wind speed,sunshine hours and solar radiation are taken as independent variables,and the daily water consumption of rice is taken as the dependent variable.The multiple regression equation of 5 input and 1 output is established.In this paper,a new method of high order multivariate polynomial modeling is proposed,which is based on the data of each growth stage.3)A hybrid learning model based on genetic algorithm and LM optimization algorithm is established,and the GALM algorithm effectively combines the advantages of the two methods.The experimental results show that the GALM hybrid algorithm improves the convergence speed of the network,and it is better than other algorithms in the rate of false positives.The GALM network achieves the predetermined error through 358 steps,the maximum absolute error is about 0.1720,and the average absolute error is about 0.0434.4)Design of water-saving irrigation controller with high intelligent chip as the core,the fixed frequency on rice growth period around the environmental temperature and humidity,solar radiation,sunshine duration,wind speed,rainfall,daily evaporation,soil heat flux,different depth of soil temperature factors such as data acquisition,and in accordance with the genetic LM algorithm and optimization model of irrigation based on hybrid learning algorithm.The water saving irrigation controller is designed with low power consumption and distributed control mode,which can work independently.5)The design of water saving irrigation monitoring platform suitable for paddy field in northern China was carried out.The platform hardware equipment has the advantages of small footprint,easy operation and low energy consumption.The platform software has the functions of real-time data monitoring,historical data query and analysis,irrigation information prediction and so on.It can also be used in the field of irrigation controller.After testing,the communication distance can be realized in 2800 m,the data transmission accuracy is 100%,and the system has good anti-jamming performance.Through the actual test of the overall performance of the site,to the device on the power,press the start button 10 s began to receive a variety of terminal sensors to collect the data,and in the 5S to complete the outgoing data.After the host computer receives the data,the data can be fused and forwarded in 20 s.The control data can also be displayed on the mobile phone or other wireless terminals,and can be controlled in real time by the mobile phone or other wireless handheld devices.The utility model has the advantages of fast response speed and high control efficiency.6)Tests were carried out in a number of farm field.By comparing the control irrigation and artificial full irrigation,the contrast test results show that the average consumption of water was 37.68 percent,the average grain output increased by 6.47 percent,the increase in income was 148.97 yuan per acre,the cumulative value of the three years increased by 435 thousand and 100 yuan,and the application prospects were broad.更多还原