Abstract: The monthly market-price forecasting framework for the Oilfield Class-A materials, taking a large proportion in total purchasing cost, is considered based on the ARIMA model in time series method. The framework includes the sample set module and the ARIMA module. The sample set module provides a sample input for forecasting and updating in real time. The ARIMA module includes how to fit, test, forecast, evaluate and dynamically revise the model. According to the framework, in China Daqing Oilfield, market-prices in three places from January to December in 2011 of the Small Deformed Steel Bar(20-HRB335)in Class-A materials are predicted, including Tianjin, Shijiazhuang and Shenyang. The accuracies of the prediction are no more than 2.13%,1.64% and 1.82%, respectively, which is given a high evaluation by users. It provides the basis for Daqing Oilfiled Materials Corporation in making optimal material purchasing decision. Finally, suggestions of improving this framework are presented.

Key words: applied mathematics, market price forecasting, time series method, ARIMA model, class-A oilfield materials of Daqing

摘要： 本文研究了油田A类物资的市场价格预测。采用时间序列方法中的ARIMA模型,结合油田物资历史价格数据,分析并提出了一套针对油田A类物资的市场价格预测模式。该模式包括两个模块:样本集模块和ARIMA模块。样本集模块的主要功能是样本集的输入和实时更新;ARIMA模块包括了价格预测模型的拟合、检验、预测、评价、动态反馈和调整等主要环节。在该模式的指导下,以大庆油田A类采购物资中的小螺纹钢(20-HRB335)(天津、石家庄和沈阳3个产地)为例,对2011年各月的市场价格进行了模拟预测,预测的平均相对误差分别控制在2.13%,1.64%和1.82%,该结果得到了用户的认可。该预测模式的运用对大庆油田物资集团制定合理的物资采购方案提供了依据。结论部分对该预测模式的意义及存在问题进行了分析,并给出改进建议。

关键词: 应用数学, 市场价格预测, 时间序列方法, ARIMA模型, 大庆油田A类物资