改进灰色模型在含蜡原油管道蜡沉积厚度预测中的应用研究  

Application of improved grey model in wax deposition thickness prediction of a waxy crude oil pipeline

作　　者：徐士祺[1,2] 殷彦龙 杜胤 秦干[4] 王磊 XU Shi-qi;YIN Yan-long;DU Yin;QIN Gan;WANG Lei(College of Petroleum Engineering,Xi'an Shiyou University,Xi'an 710065,China;Shaanxi Key Laboratory of Advanced Stimulation Technology for Oil&Gas Reservoirs,Xi'an Shiyou University,Xi'an 710065,China;The Second Oil Production Plant of Qinghai Oilfield,Haixi 817500,Qinghai,China;The Third Oil Production Plant of Changqing Oilfield,Yan'an 716000,Shaanxi,China)

机构地区：[1]西安石油大学石油工程学院,西安710065 [2]西安石油大学陕西省油气田特种增产技术重点实验室,西安710065 [3]青海油田采油二厂,青海海西817500 [4]长庆油田分公司第三采油厂,陕西延安716000

出　　处：《安全与环境学报》2022年第6期3041-3049,共9页Journal of Safety and Environment

摘　　要：为了对含蜡原油管道中的蜡沉积厚度进行准确预测,分别基于函数cos(x^(T))变换与函数cot(x^(T))变换建立了GM(1,1)蜡沉积厚度预测模型。通过理论证明比较了函数cos(x^(T))变换和函数cot(x^(T))变换对原始数据光滑度的提升能力。以现场管道结蜡数据和室内环道结蜡数据为例,对比了传统GM(1,1)模型、基于函数cos(x^(T))变换及基于函数cot(x^(T))变换建立的GM(1,1)模型之间的预测精度,并分析了T的取值变化对GM(1,1)模型预测精度的影响。结果表明:原始数据经函数cos(x^(T))变换相比函数cot(x^(T))变换具有更好的光滑度;基于函数cos(x^(T))变换的GM(1,1)模型精度最高,其次是基于函数cot(x^(T))变换建立的GM(1,1)模型,而传统GM(1,1)模型精度最差;在所取T范围内,基于函数cos(x^(T))变换与基于函数cot(x^(T))变换所建模型的精度均随T增大呈现出先增大后减小的趋势。In this paper,GM(1,1)wax deposition thickness prediction models are established based on function cos(x^(T))transformation and function cot(x^(T))transformation respectively.Through theoretical proof,the function cos(x^(T))transform and function cot(x^(T))transform are compared to improve the smoothness of the original data.Taking field pipeline wax deposition data and indoor pipeline wax deposition data as examples,the prediction accuracy between the traditional GM(1,1)model and GM(1,1)model based on function cos(x^(T))transformation and function cot(x^(T))transformation is compared,and the influence of the change of T value on the prediction accuracy of GM(1,1)model is analyzed.The results show that the original data transformed by function cos(x^(T))has better smoothness than that transformed by function cot(x^(T));GM(1,1)model based on function cos(x^(T))transformation has the highest precision,which can reduce the average relative prediction error to 0.0832%;while GM(1,1)model based on function cot(x^(T))transformation has the second precision,which can reduce the average relative prediction error to 1.2110%;while the traditional GM(1,1)model has the worst precision,with an average relative prediction error of only 6.4210%.For indoor wax deposition thickness prediction,GM(1,1)model based on function cos(x^(T))transformation has the highest accuracy,which can reduce the average relative prediction error to 1.1261%,followed by GM(1,1)model based on function cot(x^(T))transformation which can reduce the average relative prediction error to 1.2909%,while the traditional GM(1,1)model has the worst accuracy,with an average relative prediction error of only 16.88%.This method can provide a useful reference for the accurate prediction of wax deposition thickness of waxy crude oil pipelines,thus providing a guarantee for the safe operation of crude oil pipelines.

关 键 词：安全工程 蜡沉积厚度 GM(1 1)模型 函数变换 模型精度对比 

分 类 号：X93[环境科学与工程—安全科学]