Abstract
Acidizing is widely used to remove near-wellbore damage and enhance the inflow performance of the reservoir to boost the well oil production rate. However, when the injected acid exposes to reservoir oil, either asphaltic sludge or emulsion forms as acid-induced damages. Therefore, laboratory compatibility tests are required before every acidizing job to determine both the acid sensitivity of oil samples and the optimal dosage of chemical inhibitors that should be used to prevent sludge and emulsion formation. The lack of knowledge to predict the risk of asphaltic sludge and emulsion damages for different oil and acid systems necessitates expensive and time-consuming compatibility tests prior to each acidizing process to select the best additives and optimize their dosages. A modified API RP-42 as the well-known standard procedure of the acid-sludge test is introduced here to overcome nuisance effects, estimate damage type and mimic the down-hole condition. Then, a robust database of 240 data (120 mass and 120 vol) is created for six different oil samples and three operational variables: 1) HCl mass fraction, 2) acid volume concentration, 3) iron contamination. Besides, statistical analysis is used to control the quality of the testing method and the prepared databank. Furthermore, an artificial intelligence (AI) model is trained by the database prepared to estimate sludge formation potential of all crude samples by utilizing their Saturate to Aromatic ratio (S/Ar) and Asphaltene to Resin ratio (A/R). According to the results, sludge formation potential strongly depends on the crude samples. Also, both A/R and S/Ar are good representations of crude samples. The statistical analysis shows that individual oil models are more than 90% accurate. Likewise, the error of the AI model developed is 10.5% that indicates good predictability. A logical criterion for damage type diagnosis (pore-clogging or tight emulsion block) is proposed. The models' performance is sound for HCl acid solution without additives, and further work is needed to generalize it by adding other effective parameters. • In detail procedure for acid-sludge test is proposed. • Robust database of acid-sludge formation for six different oil samples is prepared. • An AI based model is developed for mass and original volume of sludge. • The quality of experimental procedure, data, and models controlled by statistical analysis. • Logical criterion for assessment of damage type is presented based on the models.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have