Research On HL Project Procurement Supplier Selection Based On BP Neural Network Model

With the increasing demand for petrochemical products and the vigorous development of petrochemical industry,a series of refining and chemical projects have been approved and put into operation in recent years.The second phase of HL refining and petrochemical project in South China of CNOOC(hereinafter referred to as HL project)is a scale of 10 million tons per year refining project and 1.2 million tons per year ethylene project.This refining and petrochemical project invested 46.6 billion yuan.The completion of the HL project will increase the capacity of the petrochemical industry in the Pearl River Delta.In the process of construction and operation,in order to safeguard the interests of the project and effectively control the investment cost,we should choose a better quality and more suitable supplier,and we must have a comprehensive understanding of all aspects of supplier performance.Therefore,there must be a reasonable selection method to locate suppliers.The unreasonable selection method can't effectively evaluate the true level of suppliers,which will seriously affect the enthusiasm and stability of suppliers' cooperation and damage the interests of enterprises.Aiming at the supplier selection problem,this paper will take HL project as an example to find out the existing problems in HL project supplier selection.By studying the relevant theories of supplier selection and artificial neural network knowledge at home and abroad,BP neural network model is selected to solve the problem of HL project supplier selection.In this paper,a three-layer BP neural network structure model is established,including an input layer,an implicit layer and an output layer.Considering the supplier characteristics of HL project,a supplier selection index system of HL project is established,which consists of 8 first-level indicators and 21 second-level indicators.46 suppliers in HL project vendor database were selected,and 21 index data of 36 suppliers were collected as training samples for input layer of the model,and the scores approved by 36 suppliers' project group were used as output layer values.The validity of the optimized model is verified by using the remaining 10 suppliers' data.The validation results show that the established supplier selection model has good generalization ability,high accuracy and practicability,and can be used in HL project supplier selection.This paper applies BP neural network to HL project supply selection,which can greatly shorten the selection time and make the supplier selection more efficient.The index weight calculation of this method is more scientific,and the evaluation results can accurately reflect the comprehensive ability of suppliers,which greatly improves the accuracy of HL project supply evaluation results.Using this result to guide purchasing work,we can achieve the goal of saving procurement cost and improving material quality.Therefore,the application experience of BP neural network model in CNOOC HL project can be an example of applying BP neural network method to supplier selection management in similar refining projects.