[vc_empty_space][vc_empty_space]
A physic-based scaling curve of gas production under compaction: A leading way to better machine-learning prediction
Sajjad F.a, Chandra S.d, Naja S.b, Suganda W.c
a PT Pertamina Hulu Energi Offshore Northwest Java
b PT Pertamina Hulu Energi Offshore Northwest Java
c PT Pertamina Hulu Energi Offshore Northwest Java
d Institut Teknologi Bandung, Indonesia
[vc_row][vc_column][vc_row_inner][vc_column_inner][vc_separator css=”.vc_custom_1624529070653{padding-top: 30px !important;padding-bottom: 30px !important;}”][/vc_column_inner][/vc_row_inner][vc_row_inner layout=”boxed”][vc_column_inner width=”3/4″ css=”.vc_custom_1624695412187{border-right-width: 1px !important;border-right-color: #dddddd !important;border-right-style: solid !important;border-radius: 1px !important;}”][vc_empty_space][megatron_heading title=”Abstract” size=”size-sm” text_align=”text-left”][vc_column_text]Copyright 2019, Society of Petroleum EngineersWe present a simple analytical solution to diagnose gas production under compaction. This solution scales production profile of different wells and collapses them into a single general curve. The curve will later serve as the “learning” function for physic-based machine-learning prediction. A rapid growing flood of big data in the oil and gas industry reveals a substantial opportunity to the better understanding of hydrocarbon reservoir. With machine learning, one can turn a numerous amount of data to predict future production and determine field economics. However, the quality of the prediction from machine learning is dependent on the learning function selected that most of the time does not concatenate any physical aspects of the problem. In this paper, we offer a better machine learning with a physics-based function to estimate future gas production under severe compaction. We construct a physic-based master curve by solving the coupled Darcy-Biot equation for vertical gas well under reservoir compaction. We assume that the flow is radial and the porosity is transiently changing by the reduction in pore pressure due to gas production. Finally, we reduce the complexity of the coupled non-linear equation to two scaling optimization parameters: a mass scaling factor to scale the recovery factor and time scaling factor to scale the diffusion time. We verify our model with a field case from KLX field, Indonesia. This gas field produces an enormous amount of gas with subsidence as the side effect. The subsidence was identified by knowing the change in platforms level. By collapsing the production profile of all existing wells into a single master curve, we capture the universal scaling parameters that represent the behavior of gas flow under reservoir compaction. Furthermore, we can substitute the resulted master curve as the learning function for to the machine-learning model to predict and diagnose other fields in the future that undergo the same phenomena. We find that reservoir compaction leads to a higher recovery factor of gas for a long term. However, the high subsidence rate is not a favorable condition for the offshore field as the production facilities on the platform will submerge under sea level in a matter of years. Thus, the field owners must consider some subsidence mitigations such as injection and maintaining critical production rate. Our novelty is to produce a general scaling to describe gas production under compaction, which is later useful for the development of our machine-learning process to simplify the prediction process, not involving extensive and expensive numerical simulation.[/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”Author keywords” size=”size-sm” text_align=”text-left”][vc_column_text]Favorable conditions,Hydrocarbon reservoir,Machine learning models,Oil and Gas Industry,Optimization parameter,Production profiles,Reservoir compaction,Time scaling factors[/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”Indexed keywords” size=”size-sm” text_align=”text-left”][vc_column_text][/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”Funding details” size=”size-sm” text_align=”text-left”][vc_column_text][/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”DOI” size=”size-sm” text_align=”text-left”][vc_column_text]https://doi.org/10.2118/196254-ms[/vc_column_text][/vc_column_inner][vc_column_inner width=”1/4″][vc_column_text]Widget Plumx[/vc_column_text][/vc_column_inner][/vc_row_inner][/vc_column][/vc_row][vc_row][vc_column][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][/vc_column][/vc_row]