论文部分内容阅读
采用现代控制理论中的多层递阶方法, 建立适合于动态预测的时变参数预测模型, 在此基础上, 动态预测天然气消费量; 模型中的天然气消费函数分别采用多元线性模型和多元几何模型。用两组算例两种方法(回归预测、多层递阶预测) 分别进行预测, 从计算结果分析可知, 多元几何模型的多层递阶预测方法预测精度最高。
The multi-level hierarchical method in modern control theory is used to establish a time-varying parameter prediction model suitable for dynamic prediction. On this basis, the natural gas consumption is dynamically predicted. The natural gas consumption function in the model uses multiple linear models and multiple geometric models . Two sets of examples are used to predict the two methods (regression prediction and multi-level prediction) respectively. From the analysis of the calculation results, the multi-level prediction method of multivariate geometric model has the highest prediction accuracy.