论文部分内容阅读
对 Box-Jenkins 型多变量时序传递函数模型参数估计,以及基于此模型对序列末来值预报,属于非线性估计预报问题。通过线性化处理,本文得到一种估计预算新方法。文中还指出了噪声统计特性有限记忆自适应估计器和双置信区不良数据检测法,使算法具有较强的自适应跟踪和检测功能。仿真结果验证了算法的有效性。在对无水库水电站曰发电能力和河道来水量的预报中,其精度达到了实际生产的要求。
The Box-Jenkins type multivariate time series transfer function model parameter estimation, as well as based on the model of the sequence of the last value forecast, belong to the nonlinear estimation and prediction problems. Through the linearization process, we get a new method to estimate the budget. The paper also points out the finite memory adaptive estimators of noise statistics and the double-confidence zone bad data detection method, which makes the algorithm have strong self-adaptive tracking and detection functions. Simulation results verify the effectiveness of the algorithm. In the non-reservoir hydropower station generating capacity and river water forecast, the accuracy reached the actual production requirements.