PSO Algorithm–Based Inversion Model to Interpret a Production Profile from DTS for Horizontal Wells in Shale Gas Reservoirs

Significant difficulties still exist in production profile and fracture parameter diagnosis for horizontal wells in shale gas reservoirs (SGRs), especially when the complex fracture network is created after hydraulic fracturing. Although the distributed temperature sensing (DTS) technology provides enormous potential to solve this problem, due to the lack of robust inversion approaches, it is hard to quantitatively translate the DTS data to the production profile. First, a comprehensive inversion model combined with a forward temperature and particle swarm optimization (PSO) algorithm has been proposed. It has realized the quantitative interpretation of m×n dimensional unknown fracture parameters of horizontal wells in SGRs, which shows the significant novelty of this study. Subsequently, a synthetic case is presented to analyze the temperature characteristics and the application of this inversion approach. Obvious temperature drops denoting the cooling effect caused by the Joule–Thomson effect at each cluster location have been found. According to this characteristic, effective hydraulic fractures could be identified from the field DTS data. The sensitivity analysis of PSO parameters on production profile results has been performed to optimize the inversion model parameters. Finally, two field applications are provided to demonstrate the capability of the PSO inversion model. Satisfactory inversion results (the absolute inversion temperature errors <0.042°C, the mean absolute error of the gas production profile is 0.133×104 m3/d, and the R2 is 0.918) of Case 1 (Well FY26) validate the reliability of the developed inversion model. The interpreted production profiles of Case 2 are close to the production logging tool test results as well. Fracture parameters including the fracture half-length, permeability of the fracture network, and conductivity have been quantitatively diagnosed. The developed PSO algorithm–based inversion model proves to be a promising tool to perform the quantitative evaluation of the production profile of horizontal wells in SGRs. The findings of this study can help for better understanding of production profiling from distributed fiber optic sensing for horizontal wells in SGRs.