一种新的天然气管网负荷预测方法

A NEW METHOD FOR PREDICTING LOAD OF NATURAL GAS PIPELINE NETWORK

对管网负荷的准确预测是保证管网经济安全运行的前提,在总结前人大量的管网负荷预测方法的基础上,提出了基于新鲜度函数的模糊变权重组合预测管网负荷的方法。该方法是单模型预测技术的自然推广,通过将新鲜度函数引入组合预测,可有效地刻画时间序列的变化趋势,从而实现了对模糊自适应变权重组合预测方法的有效改进。改进后的组合预测方法不仅综合利用了各种预测方法的有用信息,而且该最佳组合预测的权系数也是时变的,是一种进一步提高预测精度的方法,具有预测精度高、计算简便的特点。

Precise predictions on pipeline networks are the precondition of economical and secure operation. The numerous existing methods for prediction of pipeline network load were reviewed, and a multi-model combined forecast method based on fresh degree function was developed. This method was derived from the single-model prediction technology. The fresh degree function is brought into the multi-model combined forecast method, which can effectively describe the variation trend of the data list and improve the self-adaptive recurrent algorithm and fuzzy variable weight combined forecast method. The improved combined forecast method integrates helpful information of all other prediction methods. Moreover, the weight coefficient of the optimum combined forecast is the function of time, making it a high-precision, easy-calculation method.