Forecasting energy consumption of long-distance oil products pipeline based on improved fruit fly optimization algorithm and support vector regression

Abstract

Predicting the energy consumption of oil pipelines is an important part of pipeline companies’ energy-saving and consumption-reduction plans and the realization of refined management. In order to predict the energy consumption of the long-distance product oil pipeline faster and better, this manuscript innovatively uses the normal distribution function to improve the search mode of the fruit fly optimization algorithm (FOA). It establishes the normal distribution fruit fly optimization algorithm (NFOA). It enhances search accuracy in the central area and effectively expands the search scope. Experimental results show that the accuracy and stability of the algorithm are improved by 100% and 900%. Then, NFOA combined with support vector regression (NFOA-SVR) is used to predict the three long-distance product pipeline data sets in China. The results show that the optimization speed and prediction accuracy of NFOA-SVR in LCY-Others set and LW-total set are significantly better than the other two algorithms. In the LCY-Pump set, NFOA-SVR has the same accuracy as the other two algorithms. Finally, experiments on random data sets show that the accuracy and stability of NFOA-SVR gradually decrease with the increase of the standard deviation of the data set.