规模化养殖鸡舍环境温度虚拟采集方法研究

Virtual Collecting Method of Ambient Temperature in Large-scale Breeding Chicken House

摘要:

目前，在畜禽养殖高度集约化发展背景下，禽舍数智化管控尤为重要，海量信息收集支撑数据模型构建与物联网实体节点数量受限之间的矛盾是目前面临的一个问题。提出了一种结合当前参考点和历史数据的鸡舍温度虚拟采集方法。首先，采用计算流体动力学（CFD）模拟来分析和确定鸡舍内部温度分布和环境特点，根据CFD模拟结果初步划分采集区域。然后，结合灰色关联度与余弦相似度分析，有效识别出与关键未监测区域的温度高度相关的参考点。最后，采用XGBoost和WOA-BiLSTM等人工智能算法，预测未直接监测区域的温度。通过在河北省邢台市某蛋鸡养殖场测试，所部署的10个虚拟采集点数据与实际数据平均绝对误差小于025℃，保证数据可靠性的同时为禽舍数智化管控建模提供了足够的数据量，为智慧农业实践提供了技术基础。

Abstract:

At present, under the background of the highly intensive development of livestock and poultry breeding, the intelligent control of the number of poultry houses is particularly important. The contradiction between the construction of data model supported by massive information collection and the limited number of physical nodes of the Internet of Things is a problem currently facing. A virtual temperature acquisition method combining current reference point and historical data was proposed. Firstly, computational fluid dynamics (CFD) simulation was used to analyze and determine the temperature distribution and environmental characteristics inside the chicken house, and the collection area was preliminarily divided according to the CFD simulation results. Then combined with gray correlation degree and cosine similarity analysis, the reference points that were highly correlated with temperature of key unmonitored area were effectively identified. Finally, artificial intelligence algorithms such as XGBoost and WOA-BiLSTM were used to predict temperatures in areas not directly monitored. Through the test in a laying chicken farm in Xingtai City, Hebei Province, the average absolute error between the data of ten virtual collection points and the actual data was within 0.25℃, which ensured the reliability of the data and provided enough data for the intelligent control modeling of the number of poultry houses, and provided an important technical basis for the practice of smart agriculture.