基于光合速率时序模型的设施环境调控目标寻优方法研究

An appropriate and efficiently microclimate construction for plants to promote their photosynthesis and nutrient accumulation is the key problem of modern facility horticulture production. The difficulty is, under the influence of plant-environment interaction, to obtain regulation objective values accurately which integrated with regulation efficiency. In this project, take tomato growth environment regulation as an illustration, we propose a photosynthetic rate time-series model in the whole process based on Genetic Algorithms-Support Vector Machine (GA-SVM) algorithm and design several whole crop growth fine-grained nested factor tests to prove it. Based on the function of temperature and optimistic photosynthetic rate from previous step, we seek the photosynthetic rate sub-optimization map combined with regulation efficiency using curvature inflection point theory to obtain the appropriate temperature range for tomato growth and create desired photosynthetic rate value model. Further step, with the desired value model from previous step and the most optimized regulation financial cost as constraint conditions, we research on tomato growth environment regulation values optimal seeking method. And based on the tomato photosynthetic rate model and the previous optimal seeking method, we construct tomato growth converged data center and development multi-sensor information fashioned test platform to do method validation and improvement. This research will be of great importance in clarifying facility high-efficiency regulation principle, and will also lay the foundation for rapid development of modern facility horticulture.

营造与作物需求相适应的小气候，促进光合作用和物质积累是设施园艺高效生产的基础。针对作物生长过程中多环境因子相互作用对光合速率影响规律复杂，难以融合调控效率精准获取设施环境调控目标值的难题，拟面向设施番茄，研究光合速率时序模型与设施环境调控目标寻优方法。设计作物生长全程细粒度多因子嵌套试验，研究基于遗传支持向量机的全程光合速率时序模型建模方法；利用遗传算法建立温度与最优光合速率映射函数，基于曲率拐点理论研究生长适宜温度区间获取方法，探寻融合调控效率的光合速率次优目标值优化算法，建立温度关联的光合速率目标值模型；基于光合速率时序动态模型，以光合速率次优目标值为约束条件，研究调控成本最优的环境调控目标值寻优方法；以光合速率模型和目标值寻优方法为核心构建数据融合中心，设计多传感器信息融合试验平台，进行方法验证与完善。研究成果对揭示设施环境调控规律具有重要意义，将大力推动现代设施园艺的快速发展。

温室环境是影响温室作物产量与品质的关键因素之一，而多环境因子之间存在相互关联关系，且调控成本差异显著，如何在满足作物生长需求的同时降低调控成本，已成为温室环境高效调控亟待解决的科学问题和难点问题。.本项目在分析影响作物光合速率内外部影响因子的基础上，针对温度对光合速率的影响特性，研究基于径向基插值-鱼群算法的温度适宜区间获取方法，实现不同温度条件下的光与CO2二维寻优，获取光与CO2饱和条件下的温度与光合速率的关系;针对光与CO2对光合速率的影响特性，研究基于遗传-支持向量机算法的光合速率预测模型构建方法，进而探寻基于L离散曲率-粒子群算法的光合速率次优目标值模型获取方法，建立满足温度约束的光合速率次优目标值模型;融合光合速率次优目标值模型和实时信息感知，基于粒子群多目标寻优算法研究约束条件下的效益优先的光与CO2协同调控目标值在线寻优方法，获取效益优先的PPFD与CO2协同调控目标值。基于上述理论研究，设计并研发了温室环境协同调控系统并进行实际部署验证。调控系统原型部署于西北农林科技大学五泉试验基地，能耗数据统计显示，效益优先调控平均日消耗成本为4.48元，光合最优平均日消耗成本为6.96元。相对于自然对照区,效益优先协同调控区CO2日消耗1503.17L,日用电量2.97KW/h， 光合速率平均提升45.15%， 光合最优协同调控区CO2日消耗2290.61L，日用电量4.70KW/h，光合速率平均提升53.64%。本项目按计划完成所有研究内容，达到预期目标，融合作物光合速率次优目标值模型与调控效益，提出了温室环境多因子协同调控模型与方法，对温室环境高效调控和设施产业转型升级具有理论意义。