摘要
影响筒仓动态侧压力的影响因素十分复杂,如何全面考虑影响因素,高效、简单地预测筒仓动态侧压力是重要问题。针对此问题,尝试提出了基于支持向量机的预测模型。首先,将影响因素进行归一化处理,将归一化后的数据作为预测模型的输入向量,筒仓动态侧压力作为预测模型的输出向量;其次,以400组PFC模拟数据作为训练样本,运用交叉验证和网格搜索法寻优获得最优支持向量机参数,最终建立基于SVM的筒仓动态侧压力预测模型,并对105组PFC模拟数据进行筒仓动态侧压力预测。结果表明:SVM预测模型的均方误差MSE小于0.0005,相关系数R;大于0.98,模型具有较高的准确率和较好的泛化性能。将模型试验、数值模拟、公式计算与预测数据进行对比分析,结果拟合良好;利用该模型验证筒仓动态侧压力随着相关参数的变化趋势,结果与前人研究结果相一致。该预测模型与传统方法相结合对筒仓动态侧压力进行研究可行性较高,可为筒仓动态侧压力预测、影响因素研究提供一种新的方法。
The factors that influence dynamic side pressure of silo are very complex, so how to predict dynamic side pressure of silo efficiently and simply with comprehensive consideration of these factors is an important issue.To solve this problem, a prediction model based on support vector machine(SVM) is proposed.First, the influencing factors were normalized. The normalized data was used as the input vector of the prediction model, and the silo dynamic side pressure was used as the output vector of the prediction model.Secondly, 400 groups of PFC simulation data were taken as training samples, and cross-validation and grid search were used to optimize to obtain the optimal support vector machine parameters. Finally, a SVM-based silo dynamic side pressure prediction model was established, and 105 groups of PFC simulation data were used to predict silo dynamic side pressure.The results show that the MEAN square error(MSE) of SVM prediction model is less than 0.0005, and the correlation coefficient R;is greater than 0.98. Therefore, this model has higher accuracy and better generalization performance.By comparing and analyzing model test, numerical simulation, formula calculation and prediction data, the results fit well.The model is used to verify the variation trend of silo dynamic side pressure along with relevant parameters, and the results are consistent with previous research results.The results show that the prediction model combined with the traditional method is feasible to study the dynamic side pressure of silo, which can provide a new method for predicting the dynamic side pressure of silo and studying the influencing factors.
作者
徐志军
刘婷婷
李建平
原方
Xu Zhijun;Liu Tingting;Li Jianping;Yuan Fang(School of Civil Engineering and Architecture,Henan University of Technology,Zhengzhou 450001,China;National Engineering Laboratory for Grain Storage and Transportation,Zhengzhou 450001,China)
出处
《农机化研究》
北大核心
2022年第5期9-16,共8页
Journal of Agricultural Mechanization Research
基金
国家自然科学基金面上项目(51578216)。
关键词
筒仓
动态侧压力
支持向量机
网格搜索
预测模型
储粮
silo
dynamic side pressure
support vector machines
grid search
prediction model
grain storage
