Tunu Shallow Zone (TSZ) is one of producing zone in Tunu Field. Tunu Field is a giant gas field located in the present-day Mahakam Delta, East Kalimantan, Indonesia. The gas reservoirs are scattered along the Tunu Shallow Zone and correspond with fluvio-deltaic series and main lithologies are shale, sand and coal layes. The development of TSZ heavily relies on seismic to access and identify gas sand reservoirs as drilling targets. Anomaly seismic is correspond with the gas sand reservoirs, however with the conventional use of seismic that is difficult for differentiating the gas sands from the coal layers. We established Tunu reliable technology which is comprised four different analyses on stacks, CDP Gathers, AVA/AVO, and litho-seismic cube. We are hit high success rate in identifying gas but requires a lot of time to assess the prospect. But the challenge is to access more than 20, 000 shallow geobodies in time manner, faster and more efficient to fulfill our drilling sequences target and speed-up the development phase. Therefore, we are developing seismic driven supervised machine learning to fit learn geological Tunu characteristic to be gas reservoirs. Several machine learning algorithm has been tested and selected based on several criteria such as AVA/AVO, and amplitude of seismic. The algorithm used to learn behavior of seismic correspond with gas reservoir from data training then applied it to validation and blind dataset for evaluating final models. The final machine learning output is gas probability cube with precision of 70-80% precision from well drilled result in term of gas occurence. Furthermore, unsupervised machine learning has been used to extract potential prospecting targets as geobody targets. Initial test showed encouraging result to extract geobody targets in the shorter time compare with the conventional geomodeling. The final goals are optimizing our current workflow for screening shallow gas potentials, accelerate screening in the future well targets with more efficient, effective way and independent of subjectivity, allowing 2G (geologist and geophysicists) explore deeper and confident way when targeting next future shallow gas target. Usage of seismic driven machine learning for targeting shallow gas reservoir is one big step in the current oil and gas industry and in the same time opening more opportunity to maximize powerful machine learning in 4.0 industry era which is need accuracy, more precise, robust, faster and efficient.

Tunu is a giant gas field located in the present-day Mahakam Delta, East Kalimantan, Indonesia. Tunu gas produced from Tunu Main Zone (TMZ), between 2500-4500 m TVDSS and Tunu Shallow Zone (TSZ) located on depth 600 - 1500 m TVDSS. Gas reservoirs are scattered along the Tunu Field and corresponds with fluio-deltaic series. Main lithologies are shale, sand, and coal layers. Shallow gas trapping system is a combination of stratigraphic features, and geological structures. The TSZ development relies heavily on the use seismic to assess and identify gas sand reservoirs as drilling targets. The main challenge for conventional use of seismic is differentiating the gas sands from the coal layers. Gas sands are identified by an established seismic workflow that comprises of four different analysis on pre-stack and angle stacks, CDP gathers, amplitude versus angle(AVA), and inversion/litho-seismic cube. This workflow has a high success rate in identifying gas, but requires a lot of time to assess the prospect. The challenge is to assess more than 20,000 shallow objects in TSZ, it is important to have a faster and more efficient workflow to speed up the development phase. The aim of this study is to evaluate the robustness of machine learning to quantify seismic objects/geobodies to be gas reservoirs. We tested various machine learning methods to fit learn geological Tunu characteristic to the seismic data. The training result shows that a gas sand geobody can be predicted using combination of AVA gather, sub-stacks and seismic attributes with model precision of 80%. Two blind wells tests showed precision more than 95% while other final set tests are under evaluated. Detectability here is the ability of machine learning to predicted the actual gas reservoir as compared to the number of gas reservoirs found in that particular wells test. Outcome from this study is expected to accelerate gas assessment workflow in the near future using the machine learning probability cube, with more optimized and quantitative workflow by showing its predictive value in each anomaly.