● AI · Petrophysics ANFIS ANN JCM 2017 · Malang

MICP prediction
for sealing capacity
analysis

An ANFIS model trained on 35 mercury injection capillary pressure datasets from 10 wells — predicting MICP curves from open-hole logs, blind-tested across two fields, and confirming a major gas reservoir with strong sealing capacity.

35
MICP datasets · 10 wells
.997
ANFIS · all params
0.910
Blind test R² · 3-param model
C-8
Reservoir confirmed

Why MICP matters beyond saturation

Mercury injection capillary pressure data serves three distinct purposes in reservoir evaluation, each involving different parts of the MICP curve:

  • Initial fluid saturation: Predicts the hydrocarbon-water contact and saturation distribution in the reservoir
  • Seal capacity (displacement pressure): The minimum capillary entry pressure of the cap-rock determines maximum hydrocarbon column height the seal can hold
  • Relative permeability input: Additional data for simulation model parameterization

MICP₂.₅ — the MICP value at 2.5% mercury saturation — was used as the target output, representing the threshold entry pressure most relevant for seal capacity assessment.

The prediction challenge

Direct MICP measurement requires core samples with laboratory mercury injection testing — equipment and analysis not available for every well.

The six inputs are all constituents of regular open-hole logs: Depth, DT (sonic), GR (gamma ray), RHOB (density), RT (resistivity), and NPHI (neutron).

  • 35 data points from 10 wells
  • Kutai-basin asset, East Kalimantan
  • Output: MICP at 2.5% mercury saturation (psia)
  • Resistivity converted to log scale; all inputs normalized
  • Outputs converted to log scale to prevent negative predictions

From data preparation to real-case confirmation

01

Data preparation & processing

Data selection, calibration, normalization, and grouping. Six log inputs selected. Resistivity converted to logarithmic scale before normalization. All outputs log-transformed.

02

Pattern / function identification & generating initial networks

ANN (trainlm, 2 hidden layers: 30 & 48 neurons) and ANFIS (cluster radius 0.7) trained on full 6-parameter dataset. Establishes baseline R² for each method.

03

Determining three most influential parameters

Networks run without each input parameter; RMSE calculated for each omission. Parameters ranked from highest RMSE drop (most influential) to lowest.

04

Generating prediction networks (generalization)

Final networks trained using early stopping (ES) with validation set. Both all-parameter and 3-parameter models built — enabling blind-field testing with minimal inputs.

05

Selecting best networks & trial on real case data

Best networks selected by R². Prediction network applied to C-8 reservoir — a target with no existing MICP data — to assess sealing capacity.

GR, DT, and RHOB drive MICP prediction

Sensitivity analysis identified three parameters where omission caused the highest RMSE increase — confirming them as the most influential inputs:

GR
Most influential
DT
Highly influential
RHOB
Highly influential
RT
Moderate
Depth
Lower influence
NPHI
Lower influence

Using only these three parameters (GR, DT, RHOB) reduces the data requirement for new wells while maintaining good prediction accuracy — important for practical deployment.

ANFIS outperforms ANN at every stage

Pattern identification
ANN (trainlm, 6 params)
0.937
Pattern identification
ANFIS (6 params) ← best
1.000
Prediction · all params
ANN (6 params)
0.988
Prediction · all params
ANFIS (6 params) ← best
0.997
3-param model · all data
ANFIS (GR + DT + RHOB)
0.960
Blind test · neighbouring field
ANFIS · 3 params only ← key result
0.910

The blind test result (R² = 0.910) used only GR, DT, and RHOB — the three most influential parameters — applied to a neighbouring field the model had never seen during training. This confirms the model generalizes beyond the training dataset.

A gas reservoir confirmed by predicted cap-rock integrity

After validating the prediction network on the blind-test field, the ANFIS 3-parameter model was applied to the C-8 reservoir — a deeper interval with no existing MICP measurements. GR logs were used to estimate the top and bottom of the C-8 sand.

The MICP predictions across the C-8 interval indicated strong sealing capacity in the overlying cap-rock — confirming C-8 as a significant gas-bearing reservoir with seal integrity sufficient to retain a major hydrocarbon column.

/ Conference paper · JCM 2017
Hadad, A.F., Prasetyo, A.B., Riadi, R.S. & Permana, R.C. (2017).
Prediction of Mercury Injection Capillary Pressure as source for Sealing Capacity Analysis Study.
Joint Convention Malang 2017 (HAGI–IAGI–IAFMI–IATMI), Ijen Suites Hotel, Malang, 25–28 September 2017.
Affiliations: Vico Indonesia (1, 3, 4) · SKK Migas (2)