BP神经网络在石油生产能耗分析预测中的应用

Application of BP neural networks to analysis and forecast of energy consumption in oil production

为系统分析和预测石油生产过程的能源消耗,通过分析石油生产系统的工艺流程,确立了影响能耗的主要因素及其评价能耗水平的综合性指标;采用人工神经网络BP算法,构建了能耗分析预测网络模型。以某采油厂能耗统计数据为样本,对模型进行培训和检验,在此基础上,分析了各因素对能耗的影响规律并对节能潜力进行预测。结果表明:机采系统效率对电耗的影响最大,机采系统效率每增长1个百分点吨油耗电将降低5.6 kwh,如果机采系统效率能达到预期的目标值,则年节电量近1.68亿kwh,节能潜力巨大。本研究思路和方法已在某采油厂应用,分析预测结果对该厂节能降耗工作起到了一定的指导作用。

To analyze and forecast energy consumption systematically in oil production process, the primary influence factors and comprehensive evaluation indexes and the artificial nerve network BP model about energy consumption were established in this article by analyzing technical process of oil production. Based on statistical data of energy consumption, the network model was trained and verified. By the network model, the influence rule of every factor on energy consumption was analyzed and energy conservation potential was forecasted. The results show that the efficiency of mechanical oil extraction system has the most significant influence on the electrical consumption. If the system efficiency grow 1%, the electrical consumption of per ton oil will reduce 5.6 kwh, and the total electricity saving will be 168,000,000 kwh in one year, if the pre-set target can be achieved. The energy conservation potential will be huge. The methods developed in this work have been applied in an oil production plant and the results of analysis and forecast played a guiding role to energy saving work of this plant.