neqsim-compressor-power-screening

name	neqsim-compressor-power-screening
calculation_basis	screening
version	0.1.0
description	Educational centrifugal-compressor power screening using the public polytropic-head equation (API 617 / API 619 style). USE WHEN: a task needs a public, screening-level estimate of polytropic head, discharge temperature, and gas power for a single compression stage before detailed compressor selection.
last_verified	2026-06-18
requires	{"python_packages":[],"java_packages":[],"env":[],"network":[]}

Compressor Power Screening

Use this skill for public, educational centrifugal-compressor power screening. It estimates polytropic head, discharge temperature, and gas power for a single compression stage using the open polytropic-head equation so an agent can scope a compression study and check against a rated driver power before detailed selection.

When to Use

• When a user asks roughly how much power a compression stage needs.
• When an agent needs a quick polytropic-head and discharge-temperature estimate.
• When examples must run without confidential compressor curves, vendor data, or company specs.

Inputs

• suction_pressure: stage suction pressure P1 in bar absolute.
• discharge_pressure: stage discharge pressure P2 in bar absolute.
• suction_temperature: stage suction temperature T1 in kelvin.
• mass_flow: gas mass flow in kg/s.
• molecular_weight: gas molecular weight M in g/mol.
• specific_heat_ratio: ratio of specific heats k, default 1.3.
• compressibility: average compressibility factor Z, default 1.0.
• polytropic_efficiency: polytropic efficiency, default 0.78.
• rated_power: optional rated driver power in kW for the comparison.

Outputs

• pressure_ratio: P2 / P1.
• polytropic_head_kj_kg: polytropic head.
• discharge_temperature_k: estimated discharge temperature.
• gas_power_kw: gas power on the shaft.
• power_margin_ratio: ratio of rated power to gas power, or null if no rated power supplied.
• compressor_warning: ok, watch, over-rated, or no-rating.
• assumptions: public assumptions used by the placeholder model.

Engineering Method

The Python class CompressorPowerModel uses the open polytropic-head equation only:

• the polytropic exponent uses (n - 1) / n = (k - 1) / (k * polytropic_efficiency).
• the polytropic head uses Hp = Z R T1 (n / (n - 1)) [(P2 / P1)^((n-1)/n) - 1], with R = 8.314 / M.
• the discharge temperature uses T2 = T1 (P2 / P1)^((n-1)/n).
• the gas power uses power = mass_flow * Hp / polytropic_efficiency.

This is educational and screening-only logic. It assumes a single stage, constant k and Z, and an ideal-gas head form. It does not include real-gas property variation, intercooling, mechanical losses, surge or stonewall limits, or vendor curves. It is not a replacement for validated compressor selection and a qualified rotating-equipment review.

Driver Power-Margin Sizing (operating margin to rated/max power)

When the binding limit is the driver / rated maximum power (not surge), a common operations question is how large a power margin to keep so the machine can run close to maximum without tripping. The screening relation is dimensionless:

• At near-constant pressure ratio, power ∝ mass flow, so throughput headroom [%] = power_margin / operating_power.
• Therefore the same absolute MW margin protects a smaller machine less — margin must be scaled to each machine's operating power, not applied as a flat MW. To give machine B the same % headroom as machine A: margin_B = margin_A × (P0_B / P0_A).
• Size the margin to the fastest credible power excursion before the control system reacts. The dominant excursions raise the pressure ratio: a suction-pressure drop or a discharge-pressure rise. A downstream pressure controller (e.g. an MPC holding delivery pressure fixed) removes a first-order power disturbance and so buys back margin: residual swing ≈ (∂power/∂P_discharge) × ΔP_controlled.

Obtain the local slopes ∂power/∂flow, ∂power/∂P_suction, ∂power/∂P_discharge, ∂power/∂T_suction by finite-differencing a validated Compressor.getPower() around the operating point (this screening class only gives the flow-proportional term). The driver rated power and performance curves must come from STID/vendor data before any margin is fixed. (Pattern first used for a Troll A export-compressor power-margin PEPR action; see the compressor-agent / enterprise-compressor-agent.)

Python Usage Pattern

from compressor_power_screening import CompressorPowerModel

model = CompressorPowerModel()
result = model.evaluate(
    suction_pressure=30.0,
    discharge_pressure=90.0,
    suction_temperature=313.15,
    mass_flow=25.0,
    molecular_weight=19.0,
    rated_power=8000.0,
)

print(result.compressor_warning)
print(result.gas_power_kw)
print(result.discharge_temperature_k)

Related NeqSim Functionality

For validated compressor calculations, redirect to existing NeqSim classes:

• neqsim.process.equipment.compressor.Compressor — rigorous polytropic compression with real-gas properties and curves.
• neqsim.process.equipment.compressor.CompressorChart — vendor performance maps and surge control.
• neqsim.process.mechanicaldesign.compressor.CompressorMechanicalDesign — compressor mechanical design and feasibility.

This skill is a public polytropic-head triage layer that decides when to invoke those validated compressor classes.

Validation Checklist

• Suction and discharge pressures are positive and P2 > P1.
• Temperature, mass flow, and molecular weight are positive.
• Tests cover a head and power calculation, a rating comparison, and invalid input.
• Results are described as educational screening indicators.
• Real selection is redirected to validated NeqSim compressor classes and qualified review.

Common Mistakes

Symptom	Cause	Fix
Discharge temperature too high	Single high-ratio stage assumed	Split into stages with intercooling
Power off significantly	Constant Z over a wide range	Use a real-gas method like the NeqSim Compressor
Head looks wrong	Molecular weight in kg/mol not g/mol	Keep M in g/mol

Limitations

• No proprietary compressor curves, vendor data, or company specs are included.
• No intercooling, mechanical losses, or surge and stonewall limits are modeled.
• No real-gas property variation across the stage is included.

References

• NeqSim repository: https://github.com/equinor/neqsim
• NeqSim Skills Guide: https://github.com/equinor/neqsim/blob/master/docs/integration/skills_guide.md