The Research And Application Of Prediction Of Oil Tank Storage Of Gas Station Based On BP Neural Network

Gas station store fuel in the underground oil tanks, but too much storage will leads to the fuel inventory backlog, being adverse to manage and wasting storage costs. Too little inventory may results in insufficient oil supply and can’t meet customer demand, damaging image and ultimately declining profitability. Therefore, it is significant to determine the fuel inventory. Taken the features of the stock of oil tanks of gas station into consideration and based on historical data, BP neural network is used to study the prediction of the stock of oil tanks of gas station.First, several experienced methods of forecasting are studied. After the brief comparative analysis of the pros and cons of different methods, BP neural network is selected as a basic approach for forecasting the stock of oil tank of gas station. Then, the structure of BP neural network and its algorithm are studied, as well as implementation of the algorithm by some tools, etc. Finally, correlative algorithm based on BP neural network is put forward, which forecasts the stock of oil tanks of gas station.This study aims to build the model of the stock tank of gas stations and to use BP neural network to predict the stock of oil tanks, being provided a theoretical basis for the gas station oil management.