储气库管柱冲蚀理论及预测研究进展

Research Progress on Erosion Theory and Prediction in Pipe String of Gas Storage

储气库的管柱冲蚀规律较复杂,且冲蚀预测难度较大,开展循环注采工况下的储气库管柱冲蚀理论、冲蚀预测方法研究具有必要性及迫切性。目前,针对储气库强采强注工况下的管柱冲蚀报道极少。综述了塑性材料及气主导体系管柱冲蚀机理、冲蚀预测方法,对保障储气库管柱注采安全具有一定借鉴意义。以塑性材料冲蚀机理为出发点,详细介绍了适用于管柱塑性材料的微切削、变形切削、挤压-薄片、局部变形、二次冲蚀及靶面热熔等六大典型冲蚀理论,并汇总其优缺点。指出适用于储气库管柱的冲蚀准则(材料的临界冲蚀流量、临界冲蚀流速、表面延展性、质量损失、体积损失、强度衰减)及冲蚀表征指标(冲蚀比、临界塑性应变、冲蚀损失率、材料剩余最小安全系数),针对这些指标介绍了目前国内外典型的旋转式、射流式、管流式、单颗粒冲蚀式实验系统,以及基于实验数据的冲蚀预测经验模型,如使用频率最高的E/CRC模型和Oka模型等。描述了基于CFD预测冲蚀的普遍流程及不足,分析了基于人工智能预测管柱冲蚀的可行性,这对保障储气库安全高效注采具有一定的理论指导及现实意义。

As natural gas develops to a specific scale in China,gas storage construction has become an inevitable tendency to realize the functions of natural gas peak regulation,emergency gas supply,and strategic reserve.Gas storage often owns a long operating life and adopts a multi-cycle fast and strong injection and extraction mode.Under the multi-cycle and high-intensity injection and extraction,the pipe string will inevitably suffer from the coupling effects of temperature,pressure and alternating cyclic load and will be exposed to an erosion risk.In addition,the gas storage is prone to fatigue damage and sand output under long-cycle alternating load,further aggravating the pipe string erosion.In conclusion,the erosion law of pipe string in gas storage is string and the erosion prediction is challenging,so it is necessary and urgent to study the erosion theory and the methods of erosion prediction for pipe string in gas storage under cyclic injection and extraction.At present,scholars from China and abroad have conducted much research on the erosion of plastic material and have achieved many results.However,these results mainly analyze plastic material erosion in gathering pipelines and specific working conditions,and few reports are available for the pipe string erosion in gas storage under the robust extraction and injection.Therefore,reviews on the erosion mechanism for plastic materials and pipe string in the gas-dominated system and erosion prediction are of specific significance to ensure the safety of pipe string in gas storage during gas injection and extraction.With erosion mechanism as the focus,six typical corrosion theories suitable for plastic materials of pipe string,such as micro-cutting,deformation-cutting,forging extrusion,local deformation,secondary corrosion and hot melt on the target surface were described,and the advantages,disadvantages,and application scope of each theory were clarified.The corrosion criteria(critical erosion flow volume and critical erosion flow rate,surface ductility,mass and volume loss,and strength decay)and the erosion indexes(flow rate ratio,critical plastic strain,a loss rate of mass and volume,and the remaining minimum safety factor)applicable to the pipe string in gas storage were summarized.The specific calculation methods of the indexes were also given.Typical experimental systems from China and abroad were introduced,such as rotating,jet,tubular flow,and single-particle erosion.Several empirical erosion models based on the experimental data,like the E/CRC model and Oka model with a higher frequency of utilization,were given.However,the related standards or specifications for erosion experiments have not been developed in China and abroad,so the erosion law of pipe string has not been systematically studied.The unified interpretation of experimental phenomena and erosion mechanisms under different working conditions and the establishment of universal models for erosion prediction have not been realized.Given the suitable selection of erosion models being a priori existence,this limitation becomes the key to improving the erosion prediction accuracy when using CFD.The general process and shortcomings of predicting erosion based on CFD are described,and the feasibility of predicting pipe erosion based on artificial intelligence is analyzed,which has certain theoretical guidance and practical significance for ensuring safe and efficient injection and production of gas storage.The erosion prediction methods based on artificial intelligence can improve the prediction accuracy of erosion rates.However,the method is in the initial exploration stage and is subject to many limitations in selecting machine learning algorithms and their application settings,requiring much experimental validation.