{"id":"W3157575065","doi":"10.1016/j.petrol.2021.108735","title":"Optimization of steam injection in SAGD using reinforcement learning","year":2021,"lang":"en","type":"article","venue":"Journal of Petroleum Science and Engineering","topic":"Reservoir Engineering and Simulation Methods","field":"Engineering","cited_by":35,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Alberta","funders":"Canada First Research Excellence Fund; University of Alberta","keywords":"Reinforcement learning; Steam injection; Time horizon; Bellman equation; Production (economics); Reservoir simulation; Function (biology); Mathematical optimization; Computer science; Petroleum engineering; Engineering; Mathematics; Artificial intelligence","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008958043,0.00007535329,0.0001683247,0.0004741202,0.00002905934,0.00003333647,0.000072698,0.00003422849,0.000004935569],"category_scores_gemma":[0.0002550301,0.00007743268,0.0000290653,0.0006605239,0.00001966705,0.0004414798,0.00002031008,0.0001992032,7.389036e-8],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001513497,"about_ca_system_score_gemma":0.00007239813,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003392337,"about_ca_topic_score_gemma":3.578365e-7,"domain_scores_codex":[0.9991143,0.0000101313,0.0003527744,0.00006750522,0.0003042902,0.0001509449],"domain_scores_gemma":[0.9995997,0.00003499564,0.00006471016,0.00006213385,0.0001805796,0.00005789733],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000002364677,0.000003202492,0.0008768807,0.00005163117,0.000005596749,0.00000478599,0.0001225707,0.9182462,0.0803493,0.00001793119,6.749663e-7,0.0003188934],"study_design_scores_gemma":[0.0002693649,0.00003899346,0.001185996,0.0001182931,0.000005843095,0.000059469,0.0001620793,0.9856473,0.01236953,0.000002547935,0.00007056558,0.00007005221],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6152221,0.0003089216,0.3841443,0.000003053716,0.0002014085,0.000009696669,5.790973e-8,0.00001177208,0.00009875195],"genre_scores_gemma":[0.9631031,0.0001399978,0.03669196,0.000001749693,0.00004494102,3.207427e-7,1.884815e-7,0.00000944275,0.000008354255],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.347881,"threshold_uncertainty_score":0.3157613,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01382787458441278,"score_gpt":0.2541123855040688,"score_spread":0.240284510919656,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}