pvt-report-review

PVT Report Review

File-review skill for turning lab reports into usable engineering inputs. It is paired with pnge-well-engineering:petroleum-pvt, which provides the actual screening correlations and property calculations.

Important: Many PVT reports are internally consistent only within a single test type. Do not mix separator-test values, differential-liberation values, and CCE/CVD values without stating the intended use.

---

Preferred Inputs

• PDF or spreadsheet PVT report
• Test type labels: CCE, CVD, differential liberation, separator test, swelling
• Reservoir temperature and reference pressure
• Fluid composition if available
• Any black-oil table already exported by the lab

---

Workflow

1. Inventory what test types are actually present.
2. Extract the core fluid descriptors and reference conditions.
3. Check tables for monotonicity and obvious unit errors.
4. Separate values suitable for reservoir use from separator-use values.
5. Output a compact property package and list the unresolved issues.

---

Module 1 - Report Inventory

Use this checklist first:

• fluid type statement: black oil, volatile oil, gas condensate, lean gas
• reservoir temperature
• bubble point or dewpoint
• stock-tank oil gravity and gas gravity
• test tables present and missing
• black-oil summary table present or absent
• separator conditions and stage count

If the report lacks the reference conditions, stop and flag that before doing any extraction.

---

Module 2 - Table QC Helpers

def monotonic_direction(values):
    """
    Return 'increasing', 'decreasing', 'mixed', or 'flat' for a series.
    """
    diffs = [values[i] - values[i - 1] for i in range(1, len(values))]
    pos = any(d > 0 for d in diffs)
    neg = any(d < 0 for d in diffs)
    if pos and neg:
        return "mixed"
    if pos:
        return "increasing"
    if neg:
        return "decreasing"
    return "flat"

def linear_interpolate(x1, y1, x2, y2, x_target):
    if x2 == x1:
        return None
    return y1 + (y2 - y1) * (x_target - x1) / (x2 - x1)

Expected broad trends:

Table	Trend to sanity-check
Oil viscosity vs pressure below bubble point	Often decreases as pressure increases toward bubble point
Rs below bubble point	Increases with pressure
Relative volume in CCE above saturation	Changes smoothly, not erratically
Condensate dropout in CVD	Peaks then may partially recover with depletion

---

Module 3 - Black-Oil Extraction Targets

For most engineering workflows, extract:

• bubble point or dewpoint
• Rs, Bo, and oil viscosity versus pressure
• gas z, Bg, and gas viscosity versus pressure
• separator shrinkage and yields
• stock-tank and reservoir fluid descriptors

If the report already contains black-oil tables, prefer those over manual reconstruction from raw tests.

---

Module 4 - PVT Use Mapping

Need	Best report section
Reservoir simulation black-oil table	Lab black-oil summary or reconstructed DL/CVD results
Nodal analysis screening	Saturation pressure, Rs, Bo, viscosities, separator shrinkage
Facility flash / separator behavior	Separator test tables
Condensate banking risk	Dewpoint plus CVD dropout behavior

---

Module 5 - Common Report Problems

Problem	What to flag
Missing reservoir temperature	Entire report is hard to use
Units inconsistent across tables	Do not merge tables blindly
Bubble point conflicts between summary and test pages	Cite both and state discrepancy
Viscosity tables without pressure basis	Not safe to use in nodal or simulation
No composition and no pseudocomponents	EOS reuse will be limited

---

Output Format

When using this skill, structure the answer as:

1. Report inventory
2. Extracted key properties and conditions
3. QC findings and discrepancies
4. Recommended values for nodal / PTA / simulation use
5. Missing items to request from the lab

---

Integration Points

• Use pnge-well-engineering:petroleum-pvt for screening correlations and gap filling.
• Use pnge-well-engineering:nodal-analysis-multiphase for production system use.
• Use pnge-engineering-science:thermo-eos if the report includes compositional EOS data.