一类储粮与环境联动系统的建模及分布式控制方法研究

Food security is Strategic cornerstone of national security, and the food safety storage is a recognized problem. In the granary environmental factors such as temperature, humidity and gas is the important premise to assure the safety of grain storage. Because the dynamic relationships such as the environment of warehouse between inside and outside, warehouse environment and the quality of grain storage that is not fully considered by the existing grain storage technology, combined with that the internal environment of grain heap is gradient distribution and the transmission of control variables in the grain heap is lag and uneven, makes the existing storage methods lead to a large grain storage loss with the effect of poor. So, a class of large warehouse of wheat storage will be the background in this project, we will focus on: the analysis and modeling of the linkage system of grain storage and environment. the anomaly detection and prediction method within the warehouse environment based on food quality constraints,. the research of optimal operation and control strategy for the indicators of energy conservation and efficiency. Focus on solving the key scientific problems exist in the process of building the above methods: food quality changing with the environment between inside and outside of the warehouse the establishment of their time-delay nonlinear relationship, the structure of method of the food quality prediction, the establishment of a complex multivariable coupling relationship between grain storage environments. To get some innovative research achievements, for lay a solid theoretical foundation and the key technology in the popularization and application of this method in actual reserves.

粮食安全是国家安全的战略基石，而粮食安全储藏却是一个公认难题。粮仓中的温度、湿度与气体等环境因素是保证粮食安全储藏的重要前提。因现有储粮技术未充分考虑仓内与仓外环境之间、仓内环境与储粮品质之间的动态关系，加之粮堆内部环境呈梯度分布及所施加的控制变量在粮堆中的传播滞后及不均匀等因素，使得现有储粮方法因效果不佳而造成较大的储粮损失。为此，本NSFC-河南联合基金将以一类高大平房仓的小麦存储为背景，开展以下研究内容：储粮与环境联动系统的分析与建模；基于粮食品质约束的仓内环境异常检测及预测方法；以节能增效为指标的优化调度和控制策略研究。重点解决在建立上述方法过程中存在的一些关键科学问题：粮食品质随仓内外环境时滞变化的非线性关系的建立、粮食品质预测方法的构造、储粮环境复杂多变量耦合关系的建立。取得一些创新性研究成果，为该方法在实际中的推广应用打下坚实的理论基础和关键技术储备。

粮食安全是国家安全的战略基石，而粮食安全储藏却是一个公认难题。针对不同储藏环境下储粮品质的复杂变化规律，项目负责人和项目组成员对国内外现有方法进行充分论证的基础上，开展了一系列储粮品质检测与控制方面的研究，具体内容如下：1）针对不同储粮环境下储粮品质的变化机理展开研究：相继提出基于改进PLS的非线性建模方法，用以处理粮食储藏过程中的动态非线性特性；针对储粮环境多传感器融合感知系统中出现偏差故障信息时的容错融合估计问题，提出一种最优的实时的序贯式Kalman容错融合估计方法。2）针对不同储粮环境下储粮品质的异常检测与分类展开研究：针对储粮品质变化特点，提出一种多级多粒度的异常检测方法；为实现无破坏性、高准确度、适应性强的快速小麦品种和品质识别归类，研究基于完全梯度聚类和深度卷积网络的小麦品种归类方法；为有效判断储粮品质当前所处的状态，提出一种基于时段划分的间歇过程实时监测方法。3）针对储粮环境与品质的定位及控制问题展开研究：为实现劣变源的定位，提出一种基于序贯无迹卡尔曼滤波的事件源定位方法；针对由多个复杂交互的分布式子系统所构成的储粮系统，提出一种基于约简分类的分布式模型预测控制方法。相关成果已发表学术论文11篇，其中6篇已被SCI检索，申请发明专利7项，获河南省科技进步奖2项，承办全国性学术会议3次。该项目研究内容及成果对国家粮食安全具有重要的理论指导意义和实际应用价值，并且该部分相关成果已经进行实际推广应用。所建设的初步智能化的粮情在线检测与控制系统的储粮试点仓库，已作为河南省粮食局“仓储智能化升级改造”项目，该项目是全国首批的智能粮库试点示范工程之一。