Date Gas Load Forecasting With OIF-Elman Network

摘要

Date gas load forecasting plays a significant role in making plans and dispatching of gas producting. In order to improve the forecasting accuracy, in accordance with the influence factors and characteristics of date gas load, a model has been established to forecasting date gas load with OIF (output-input feedback) Elman network. Compared with conventional Elman network, OIF-Elman network takes into account not only the hidden nodes feedback but also the output feedback so as to obtain more information from limited sampling spots. The emulation illustrates that OIF-Elman network is better than Elman network not only in training speed but also in accuracy when the sampling spots are less. OIF-Elman network improves the generalization. It also improves the forecasting accuracy with less sampling spots. Therefore it can be used to forecast the date gas load.