The Forecasting Of Natural Gas Flow Load

With the development of the national green and smart cities, the intelligence of natural gas pipeline network is imperative. Yinchuan, as one of the national green city pilot area, its demand of gas amount will increase day by day and the requirements of the natural gas system will be stricter. As the load forecasting for the key technology of intelligent natural gas system, the accuracy of the forecasting will directly affect the safety and scientific operation of the whole network.Based on the background of Yinchuan Tianjia Energy Technology co., LTD, this paper focuses on the study of the short-term load forecasting. Through the study and research of domestic and foreign natural gas load forecasting methods, BP neural network is selected as the main frame structure of prediction. The study shows that the weights and thresholds of the network is the key of BP model, and single BP neural network in the network learning process is prone to overfitting and local optimization. To avoid affecting prediction accuracy when defects as overfitting occur in the single BP network model forecasting, in this paper, the genetic algorithm and BP neural network algorithm are combined and finally formed the GA-BP combination forecasting model, and by using laboratory data to verify the adaptability and effectiveness of the GA-BP model. As the actual data of gas in company is not complete and may have other problems, this paper employs the characteristic curve method to repair the data. At the same time, the paper adopts the correlation coefficient between the data to select the input dimension of the GA-BP model to ensure the richness and accuracy of the model. At last, the optimal combination forecasting GA-BP model and single BP model are applied to the actual data prediction simulation of Yinchuan Tianjia Energy Technology Co., Ltd.. The prediction results show that GA-BP model is more superior than single BP model.It lays a foundation for the establishment of the intelligent network of the company.