From pore-scale physics to planetary-scale observation.
Our 2026–2030 plan for research direction, publication targets, infrastructure, and partnerships.
The team behind the plan.
17 active members and 10 alumni, led by Dr. Irwan Ary Dharmawan, Head of Laboratory, Department of Geophysics, Universitas Padjadjaran. He founded the lab, oversees its research in digital rock physics and remote sensing, supervises postgraduate students, and works with partners in Indonesia and abroad.
See all membersWhere the lab stands today.
What we have built
- Mature LBM workflow on RockExplorer (Xeon Gold, 128 GB)
- Hybrid LBM + ML pipeline (CNN, U-Net, RA-UNet)
- Open datasets: Digital Rocks Portal, NETL, DPM
- 10 landmark papers (2023–2026) in a reference book
- Two research streams: Rock Physics (18) and Remote Sensing (8)
- Active work with UI (ML/DL) and industry
What still needs work
- Synthetic-to-real generalization gap
- Little Indonesian volcaniclastic data
- 3-D model training is still too costly
- No standard national DRP benchmark yet
- Multi-phase flow and reactive transport not yet covered
- Real-time wellsite deployment is still a goal
A lab built around Indonesian reservoirs.
Our reference points for the 2026–2030 plan.
A digital rock physics and remote sensing lab for Indonesian reservoirs (volcaniclastic, carbonate, geothermal), with regular Q1 papers and active industry and university partnerships.
Five missions
- 01Keep developing the hybrid LBM + ML pipeline.
- 02Publish Q1 and top-tier papers regularly.
- 03Train graduate students in computational geophysics.
- 04Support Indonesian oil & gas, geothermal, and CCS.
- 05Keep international partnerships active.
Core values
“We bridge pore-scale physics and planetary-scale observation.”
Our mandate and strategic position.
Three strategic objectives frame our 2026–2030 work.
Combine physics-based simulation (LBM), machine learning, and remote sensing to characterize Indonesian reservoirs, from pore scale to surface observation.
Hybrid LBM + ML pipeline
Strengthen the LBM + ML workflow as the lab’s main research tool.
- PINN integration with Darcy and conservation laws
- Generative methods for image reconstruction
- Reduce the synthetic-to-real performance gap
- Multiscale connectivity (pore to reservoir)
- Open-source release of the pipeline
Indonesian datasets
Build a curated dataset of Indonesian rocks for ML training.
- Volcaniclastic reservoirs (geothermal)
- Pre-salt and tight carbonates
- Fractured sandstones (EOR targets)
- Coal-bed methane and shale gas
- National DRP benchmark suite
Industry deployment
Move from research code to tools that industry partners can use.
- Edge inference for wellsite use
- Software licensing and consulting
- Industry partnerships (Pertamina, Star Energy)
- CCS reservoir characterization
- Training and certification programs
Three priorities, five research pillars.
Each pillar links to ongoing PhD projects, expected papers, and industry pilots.
Academic excellence
Priority 01Q1 and top-tier papers, doctoral graduates, and reference textbooks.
Impact
Priority 02Industry consulting, CCS support, geothermal characterization, and NDC contribution.
Innovation
Priority 03Hybrid LBM + AI, PINN, generative AI, edge deployment, and software patents.
Five pillars, built in parallel
2026 – 2030Physics-informed AI
PINN, conservation-law constraints, and hybrid LBM-ML losses.
Generative imaging
GAN and diffusion models for super-resolution and data augmentation.
Multiscale modeling
Pore to core to reservoir, with MS-Net and GNN connectivity.
Multi-phase & reactive
Shan-Chen LBM, CCS, EOR, and two-phase flow simulation.
Edge AI & deployment
Real-time wellsite DRP, model distillation, and INT8 inference.
Open science, reproducible workflows, Indonesian datasets, mentorship for graduate students, and results before claims.
Four enablers
One theme per year.
Three phases, from consolidation to deployment.
Finish current outputs and address the synthetic-to-real gap.
Multiscale and generative methods; grow international work.
First edge-inference deployment; plan the next five-year cycle.
- Reference book published
- Systematic review (Q1)
- Talent pipeline expansion
- National benchmark draft
- 8 papers in pipeline (4 accepted, 4 under review)
- PINN prototype
- Domain adaptation methods
- Indonesian volcaniclastic dataset
- 4–5 Q1 papers
- First industry pilot
- MS-Net + GNN models
- Pore to reservoir bridge
- Generative super-resolution
- Top-tier targets
- 2 doctoral defenses
- Diffusion models for rocks
- Shan-Chen 2-phase LBM
- CCS reservoir suite
- International grant secured
- Software patent filed
- Wellsite edge inference trial
- Industry licensing live
- 15+ cumulative papers
- 5+ top-tier papers
- 2031+ plan published
What each year delivers.
Open a year to see its plan.
Research & publications
- Systematic review published (Results in Engineering)
- Reference book “Advancing Digital Rock Physics” launched
- Process 8 papers in the pipeline (4 accepted, 4 under review)
- Start the national DRP benchmark dataset
- First volcaniclastic micro-CT acquisition
Capacity & partnerships
- Recruit 2 doctoral candidates
- MoU with Pertamina and Star Energy
- First workshop with the UI ML team
- Apply for 1 international grant
- Open-source release: LBM utilities
Output targets
3–4 Q1 papers
1 reference book (done)
2 international conference talks
Infrastructure
RockExplorer GPU upgrade
Storage expansion (+50 TB)
CI/CD for the ML pipeline
Risk watch
Reviewer-2 capacity
Mid-year funding gap
Talent retention
Research highlights
- PINN-RA-UNet (Darcy + continuity loss)
- Domain adaptation: synthetic to real fractures
- Indonesian volcaniclastic dataset (geothermal)
- Tortuosity benchmark on Brazilian carbonates
- Uncertainty quantification (MC Dropout, ensembles)
Industry & international
- First industry pilot (Pertamina): Cepu reservoir
- Visiting researcher exchange (Imperial College / Kyushu University)
- Joint workshop on PINN and geosciences
- International grant secured (such as AUN/SEED-Net)
- Attend Interpore, AGU, and EAGE
Output targets
4–5 Q1 papers (1 top tier)
1 doctoral graduation
3 conference papers
Software output
DRPRS-PINN v0.1 (public beta)
Volcaniclastic dataset v1
Benchmark suite v1
Risk watch
PINN convergence challenges
Field acquisition delays
GPU bottleneck
Research highlights
- MS-Net for multiscale flow prediction
- GNN for pore-network connectivity
- PINN v2: full Navier-Stokes residuals
- Generative super-resolution (low-res to high-res CT)
- Pore to core to reservoir upscaling pipeline
Education & outreach
- 2 doctoral defenses
- MOOC on Computational Geophysics (Unpad)
- 3 international PhD co-supervisions
- Annual DRPRS Symposium launched
- Industry training certificate program
Output targets
5–6 Q1 papers (2 top tier)
2 doctoral graduations
1 software patent applied
Software output
DRPRS-MultiScale v1.0
DRPRS-PINN v1.0 (stable)
Public benchmark v2
Risk watch
Multiscale validation difficulty
GPU cost escalation
Talent competition (industry)
Research highlights
- Diffusion models for rock image synthesis
- Shan-Chen 2-phase LBM (oil-water, CO2-water)
- Reactive transport simulation (mineral dissolution)
- CCS-ready reservoir characterization suite
- Geothermal multi-phase modeling (Wayang Windu pilot)
Industry & patents
- Software patent: DRPRS-Hybrid (filed)
- Industry licensing: 1 commercial contract
- Second Pertamina pilot expanded
- BPMA Aceh carbonate project kickoff
- International grant renewal secured
Output targets
5–6 Q1 papers (2–3 top 10%)
1 doctoral graduation
1 software license signed
Software output
DRPRS-Generative v0.5
DRPRS-2Phase v1.0
CCS Toolkit v1.0
Risk watch
Diffusion training compute
Patent process delays
Industry timing misalignment
Research highlights
- Edge inference: model distillation, INT8 quantization
- Mobile-DRPRS: smartphone-based field acquisition
- Wellsite analytics dashboard prototype
- Federated learning across industry partners
- 2031–2035 work plan published
Cumulative outputs (2026–2030)
- 15+ new papers
- 5+ top-tier publications
- 3 software releases in industry use
- 1 national DRP benchmark suite
- 2 software patents filed
Output targets
4–5 papers, retrospective reviews
2 doctoral graduations
Second software patent
Software output
DRPRS-Edge v1.0
Mobile-DRPRS app (beta)
Federated learning toolkit
Legacy
Roadmap 2031–2035
Second reference book
Lab succession planning
Work that never stops.
Talent pipeline
- 2 new doctoral candidates per year
- Yearly undergraduate summer school
- Postdoc fellowships (from year 3)
- Industry–academia exchange
- Mentorship: 1 senior per 3 juniors
Open science
- All code on GitHub, permissive license
- Datasets on Zenodo with DOIs
- CI/CD for reproducible pipelines
- Yearly code-review hackathons
- Public lab notebooks
International engagement
- 3+ active international collaborations by 2028
- Yearly international DRPRS Symposium
- Visiting researcher exchanges
- PhDs co-supervised with EU, US, and Japan
- Interpore, AGU, EAGE, and SEG presence
Indonesia-first datasets
Goal: the largest curated, AI-ready set of Indonesian rocks, covering volcaniclastic (geothermal), pre-salt and tight carbonates, fractured sandstones (EOR), coal-bed methane, and shale gas.
Industry partnership program
Shared benchmarks, joint pilots, and consulting with Pertamina, Star Energy, BPMA Aceh, and geothermal operators, to turn research into reservoir-scale impact.
How we will measure progress.
Each chart shows today (2026), the 2028 plan, and the 2030 target.
Research output
-
Cumulative published paperssince 2021
×1.7 by 2030 -
Q1 / top-tier papers
×2 by 2030 -
Doctoral graduatescumulative
×4 by 2030
Partnerships and data
-
Active international partners
×5 by 2030 -
Industry pilots / contracts
New by 2028 -
Indonesian dataset volumeTB
×10 by 2030
Software and funding
-
Public software releases
New by 2028 -
Software patents filed
New by 2028 -
Yearly research budgetRp bn, indicative
×4 by 2030
By 2030: a steady 5–6 Q1 papers a year, an active doctoral pipeline, two software releases used in industry, and the first national DRP benchmark.
Clear rhythm, named risks.
Decision & review rhythm
- WeeklyLab stand-up and PhD check-ins
- MonthlyPublication pipeline review
- QuarterlyRoadmap KPI review with the Faculty
- Twice a yearIndustry partner roundtable
- YearlyExternal advisory board and strategy refresh
Oversight bodies
- FMIPA Research Committee
- Department academic council
- External advisory board (industry + academia)
- Internal ethics & data-stewardship committee
- Joint LPDP / national grant oversight
What it takes to deliver.
Compute
~Rp 1–2 bn / year
- 2 GPU nodes (8× A100 / H100) by 2027
- Storage up to ~100 TB by 2028
- Cloud burst budget for diffusion models
- Mobile and edge dev kits (year 4)
- Yearly lifecycle cost: ~Rp 1–2 bn
People
8 PhDs over 5 years
- Lab head + 2 senior co-investigators
- 2–3 postdocs (from year 3)
- 4–6 MSc students per year
- 2 visiting researcher slots per year
Funding mix
5 sources
- National: LPDP, BRIN, Riset Inovasi
- International: AUN/SEED-Net, JSPS, Royal Society
- Industry pilots and consulting
- Software licensing royalties (year 4+)
- University internal seed grants
Critical success factors
Talent
Keep top PhD candidates with mentorship and a software IP share.
Discipline
Hold quarterly KPI reviews and avoid scope creep.
Industry alignment
Co-design pilots early and deliver value often.
Open science
Reproducibility is required, and all major outputs are public.
Partners who share the plan.
Active research in Europe, the Middle East, and Southeast Asia on digital rock physics, AI for reservoir characterization, remote-sensing geophysics, and geophysical computation, with shared PhD supervision. UTP (Malaysia) is our newest partner in Southeast Asia.
Europe
- KIT: computational geosciences, HPC
- UGent: pore-scale imaging
- Univ. Konstanz: statistical physics, ML
Middle East
- UAEU: carbonate reservoir characterization
- Joint workshops and data sharing
- Doctoral co-supervision
Southeast Asia New
- UTP: reservoir and petroleum geoscience
- Joint Indonesia–Malaysia DRP benchmarks
- Shared geothermal and volcaniclastic agenda
Questions, feedback, or a proposal?