Study On The Characteristics Of Short-term Natural Gas Load In A Certain Area And Its Application In Load Forecasting

Since the development of urban gas,companies have accumulated raw data that has begun to take shape.This precious data needs to be analyzed and processed to extract valuable information and ultimately serve the production and development of enterprises.The analysis and prediction of natural gas load is a way to explore the value of raw data.The purpose of load analysis is to understand the gas habits,characteristics and needs of natural gas users,and to guide the actual operation of the enterprise.Load forecasting is based on load analysis.Advanced load forecasting technology predicts the user’s gas demand,and load forecasting has an important guiding role for the company’s future dispatch or planning.Based on the measured data of natural gas users in a certain area within one year,this paper carried out research on load analysis and short-term load forecasting,and finally developed a simple load forecasting software.In the load analysis,the Pearson data detection technology was used to detect and analyze the correlation between the impact factor and the user load,and determine the four factors that significantly affect the user load: minimum temperature,average temperature,date type and climate.Based on the results of load analysis,this paper continues to carry out load forecasting research,focusing on the application of KPCA theory to the optimization of FOA-LSSVM model.It is found that KPCA can better optimize the prediction performance of the model,and the prediction accuracy of KPCA-FOA-LSSVM model is far.Compared with the other three intelligent load forecasting models,the model predicts that the maximum error can be accepted by the actual pipe network,and the model itself has good generalization ability.After the optimal load forecasting model is determined,a load forecasting software with the optimal model is successfully developed.After loading the original data,the software can automatically complete the process of data normalization,KPCA,LSSVM model parameter optimization,training and update prediction.The user interface is friendly,easy to operate,and has good promotion.