【作者】 张军华；

【导师】 胡瑾； 张海辉；

【作者基本信息】 西北农林科技大学 ， 农业电气化与自动化， 2022， 博士

【摘要】 日光温室是我国北方地区冬季农业生产的主要场所,冬季蓄热保温的特性使其在反季节生产中扮演着重要角色。卷帘机揭盖被是影响温室内小环境变化最主要的农事操作,如何合理决策揭盖被时间来增加光照时长,并保证室内适宜温度是提升冬季日光温室产量与品质的关键。现有日光温室揭盖被主要依赖于人工经验知识,部分自动控制系统虽已引入作物生理生境指标来进行阈值调控,但其忽略了揭盖被后外界环境因素变化对作物生理生长状态的动态影响,也忽略了对温室内环境变化的影响,本研究基于番茄低温条件下光合速率嵌套试验,分析番茄光响应、温度响应变化特征,获取不同外界条件下抑制番茄生长的低温限制点,同时构建以光补偿点的揭盖被光限制点动态预测模型;分析揭盖被后温室内环境变化规律,探寻揭被时不同内外环境使内部温度达到稳定的光平衡点,以及构建盖被后多因子关联的夜间最低温预测模型;在此基础上,研究多因子约束条件的日光温室卷帘机揭盖被决策方法,设计基于多传感器融合的日光温室卷帘机控制系统,并进行试验验证。主要研究内容及结论如下:（1）探明了低温条件下番茄光合响应特征及基于特征分析结果的调控限制点。本研究设计了7个低温温度梯度、11个光子通量密度梯度和5个CO₂浓度梯度的嵌套试验,通过LI-6400XT便携式光合仪完成了单叶光合速率的采集,分析了番茄低温下光响应与温度响应曲线特征,结合温度对光合速率提升的相对贡献率,确定了影响番茄光合速率显著变化的低温限制点为8℃;基于改进差分进化算法优化的支持向量机回归构建了低温条件下光补偿点预测模型,模型拟合决定系数为0.93,均方根误差为1.11μmol·m^-2·s^-1,为卷帘机揭盖被控制提供光限制条件依据;由于番茄单叶与整株光合存在差异,本研究搭建了高精度整株光合速率测量系统,构建了低温下整株光合速率试验,结果表明整株光合响应与单叶趋势基本一致,低温限制点位于8℃附近,光补偿点低于单叶（温度大于8℃）,证明单叶所分析的限制点能保证整株的正常生长。（2）明确了揭盖被操作对日光温室温度时空变化规律的影响。本研究搭建了256路温度场监测系统以及温室内外环境因子监测节点的部署,采集了日光温室不同天气条件下温度变化数据,经分析温室空间、土壤及后墙温度分布整体呈现西高东低,内高外低趋势,且低温主要靠近于外侧薄膜与东墙附近;采用Kriging插值构建了不同天气、不同时刻温度场插值模型,并采用差分进化算法对温度场最低温进行获取,结果表明,不同天气条件下最低温位置主要分布于靠近东墙与外膜[4.0 m,5.48 m]附近,因此将该点作为低温预测与卷帘机控制的传感器部署位置。（3）提出了基于多因子约束的卷帘机揭盖被决策方法。分析了日光温室揭被后室内温度的变化规律,由不同天气、不同温差条件下最低温监测点温度短时响应,获取了使温室内温度达到平稳的光辐射条件（光平衡点）,建立了不同温差条件下光平衡点预测模型,为卷帘机揭被决策提供温室温度响应的光限制条件;通过相关性分析获取了夜间最低温预测模型输入变量,融合气象预报积温、土壤温度、室内外初始温度参数,构建了精英策略遗传算法优化支持向量机回归的日光温室夜间最低温预测模型,拟合决定系数大于0.8,预测均方根误差为0.71℃,为盖被决策中温度阈值提供依据。融合作物低温限制点、光限制点、温室光平衡点、夜间最低温预测值与经验揭盖被决策时间等约束条件,提出了多因子约束的卷帘机揭盖被决策方法,为冬季卷帘机合理揭盖被提供了新的途径。（4）研发了多传感器融合的日光温室卷帘机控制系统。基于现有物联网平台,设计了适合于卷帘机等温室设备的网络通用通信协议,实现了环境监测节点、卷帘机物联网控制终端以及协调器之间的高效通信;开发了基于CC2530的日光温室小环境监测节点,实现了6类环境因子的实时高精度采集;系统融合红外对射技术用于卷帘机揭被上限位,实现了对卷帘机上行的可靠自动控制,有效解决过卷等不安全现象的发生。（5）开展了卷帘机控制系统的实地部署验证与调控性能分析。与卷帘机依靠传统经验控制相比,本文卷帘机揭盖被控制系统及方法有效提高了温室温度,日光温室夜间最低温低于番茄低温限制点的天数降低了43%。揭被操作中,温度基本不低于低温限制点,极端条件下温度最大下降0.3℃,最长21分钟后温度回升,证明了基于光平衡点决策的性能;试验期间,试验温室累计增加光照时间75.16 h,日光照时长平均延长1.25 h,相对于经验控制增加16.22%,总辐热积提升61.41 MJ·m^-2,提升25.89%,有效积温增加了22.28℃,对温室总体蓄热提升效能明显,试验温室前3次收获产量较对照温室高出30.74%,证明本文调控方法及系统有效提高了作物物质积累,为日光温室的提质增产提供了卷帘机控制方案。更多还原

【Abstract】 Solar greenhouse is widely used in winter agricultural production in northern China.The characteristics of heat storage and heat preservation in winter make it play an important role in out-of-season production.The most essential agricultural activity that impacts the changing of the tiny environment in the greenhouse is the curtain uncovering/covering of the shutter machine.The key to improving the production and quality of the solar greenhouse in winter is to decide on a reasonable cover time to extend the length of light and maintain a comfortable internal temperature.Existing solar greenhouse curtain uncovering/covering depends heavily on artificial experience and expertise.Some automatic control systems have included crop physiological habitat indicators for threshold regulation,however,they ignored the dynamic impact of external environmental factors on crop physiological growth after the curtain is opened/closed and on greenhouse environmental changes.Based on a nested test of tomato photosynthetic rate under low temperature settings,this study evaluated the features of tomato light response and temperature response,and investigated the low temperature limit points that impede tomato growth under various external situations.Simultaneously,a dynamic prediction model based on light compensation points was built for the light-restricted points of the curtain uncovering/covering.The full solar greenhouse’s temperature change law was taken into consideration while uncovering/covering the curtain in the winter.In addition,this study analyzed the light equilibrium point that causes the internal temperature of the greenhouse to attain a stable temperature under various internal and external settings,as well as the multi-factor correlation prediction model for the lowest temperature at night.On this foundation,a decision-making approach for curtain uncovering/covering for solar greenhouses with multi-factor restrictions was investigated,and a multi-sensor fusion-based control system for rolling shutters in solar greenhouses was constructed and tested.The key research findings and conclusions are as follows:（1）The characteristics of tomato photosynthetic response under low temperature conditions and the regulatory limit points based on the results of characteristic analysis were identified.In this study,7 low-temperature temperature gradients,11 photon flux density gradients and 5 CO₂ concentration gradients were nested,and the photosynthetic rate of single leaf was collected by LI-6400XT.The light response curve and temperature response curve characteristics of tomato at low temperature were analyzed.Considering the relative contribution rate of temperature to the increase of photosynthetic rate,the low temperature limit point that affected the significant change of photosynthetic rate was 8℃.A light compensation point prediction model for low temperature situations was built using the upgraded DE-SVR（Differential Evolution-Support Vector Regression）.The root mean square error（RMSE）was 1.11μmol·m^-2·s^-1,and the model fitting coefficient of determination（R²）was 0.93.This provided a foundation for controlling the rolling shutter machine’s light limiting constraints.There are differences in photosynthesis between a single leaf and a whole tomato plant.In this study,a high-precision whole-plant photosynthetic rate measurement system was built,and a whole-plant photosynthetic rate test at low temperature was constructed.The photosynthetic reaction of the entire tomato plant is essentially the same as the single leaf’s photosynthetic response.The total plant’s low temperature limit point is approximately 8°C,and the light compensation point is lower than the single leaf’s（more than 8°C）.It was proved that the restriction points analyzed by a single leaf can ensure the normal growth of the whole plant.（2）The impact of the curtain uncovering/closing curtain operation on the temporal and geographical change of the temperature of the solar greenhouse was investigated.A 256-channel temperature field monitoring system and the deployment of monitoring nodes for environmental elements within and outside the greenhouse were built in this study,and temperature change data was gathered under various weather circumstances.The temperature distribution of the greenhouse space,soil,and rear wall reveals a tendency to be high in the west and low in the east,high in the interior and low on the outside,and the low temperature is largely near to the outer film and the east wall.The temperature field interpolation model was built using Kriging interpolation in various weather conditions and times,and the differential evolution approach was utilized to find the lowest temperature in the temperature field.The results showed that the lowest temperature positions under different weather conditions and that they were primarily distributed near the east wall and the outer membrane near[4.0 m,5.48 m].This point serves as the sensor deployment location for low temperature prediction and roller shutter control.（3）A multi-factor constraint-based decision-making method for rolling shutter machines’curtain uncovering/covering was proposed.This study further analyzed the variation law of the indoor temperature after the curtain was uncovered.The light radiation parameters（such as light balancing point）that caused the temperature in the greenhouse to achieve a stable level were established by observing the short-term response of the temperature of the lowest temperature monitoring point under diverse weather and temperature differential circumstances.The light balancing point prediction model under varied temperature difference settings gives the light limit conditions of the greenhouse temperature response for the rolling shutter machine decision.After the curtain was covered,the temperature changes in the greenhouse was not only related to the initial indoor and outdoor temperatures,but also closely related to the back wall of the greenhouse and the soil heat storage.The GA-SVR solar greenhouse nighttime minimum temperature prediction model improved by the elite strategy was constructed by integrating the accumulated temperature of weather forecast,soil temperature,and initial indoor and outdoor temperature.The R² of the model was more than 0.8,and the RMSE was 0.71°C,providing a foundation for the temperature threshold in covering curtain decision-making.Combining the constraints of crop low temperature limit point,light limit point,greenhouse light balance point,night minimum temperature prediction value,and empirical decision time,a multi-factor constrained decision method for curtain uncovering/covering was proposed.It offers a novel approach to the uncovering and covering of the curtains of rolling shutter machines in the winter.（4）A multi-sensor fusion solar greenhouse roller shutter control system was developed.Based on the existing Internet of Things（Io T）platform,a general network communication protocol suitable for greenhouse equipment such as rolling shutter machines was designed,which realizes efficient communication between environmental monitoring nodes,rolling shutter machine Io T control terminals and coordinators.The solar greenhouse micro-environment monitoring node was designed and developed based on the CC2530,which can realize real-time high-precision collecting of 6 types of environmental factors.The use of infrared radiation device for the method of lifting the upper limit of the roller shutter machine can effectively solve the problem of over-rolling occurrence,ensuring the safe operation of the roller shutter machine.（5）The field deployment verification and regulation performance analysis of the control system of the rolling shutter machine were performed.Compared with the traditional experience control of the roller shutter machine,the control system can effectively improve the greenhouse temperature and reduce the number of days when the nighttime temperature of the solar greenhouse in winter was lower than the tomato low temperature limit point by43%.During the uncovering curtain,the temperature was basically higher than the low temperature limit point.Under extreme conditions,the temperature drops by a maximum of0.3°C,and the temperature rises after a maximum of 21 minutes,which proves the performance of decision-making based on the light balance point.The average length of sunlight in the experimental greenhouse was extended by 1.25 h,which was an increase of16.22%compared with the empirical control.The cumulative increase in the light time during the entire test（60 days）was 75.16 h,and the radiant heat product increased by 61.41MJ·m^-2,an increase of 25.89%.The effective accumulated temperature increased by22.28℃,which significantly improved the overall heat storage efficiency of the greenhouse.Based on the actual growth indicators of crops,the plants in the experimental greenhouse were more robust and the average growth rate was higher.The yield of the first three harvests in the experimental greenhouse was 30.74%higher than that in the control greenhouse.It was proved that the control method and system in this paper can effectively improve the accumulation of crop material,and provide a control scheme for the roller shutter machine for the improvement of the quality and yield of the solar greenhouse.更多还原