An orifice plate does not suddenly become unreliable when it measures gas. The real problem is that the differential pressure signal must be converted into flow using a density term, and gas density changes whenever pressure, temperature, or molecular weight change.
If the calculation uses a fixed reference density while the process is actually operating at different conditions, the displayed flow can be misleading. In that sense, engineers sometimes say that the orifice plate is “lying”, although the primary element is simply responding to the physics of the process.
For liquids, density often changes only slightly over the normal operating range, so a fixed-density approximation may still be acceptable for many industrial calculations. For gases, the situation is very different. Even moderate changes in operating pressure or temperature can alter density enough to create a noticeable flow error.
The differential pressure transmitter is only measuring $\Delta P$. The flow computer, DCS, PLC, or transmitter characterization is what converts that signal into volumetric or mass flow. If that calculation assumes the wrong density, the final flow value is wrong even though the pressure measurement itself is correct.
This is especially important in these cases:
The starting point is the standard sequence used for an idealized differential pressure derivation.
Volumetric flow is velocity multiplied by area:
Mass flow is volumetric flow multiplied by density:
If the restriction geometry is grouped into one constant and the derivation is simplified for a given installation, volumetric flow can be expressed as:
and mass flow as:
| Symbol | Meaning | Typical units |
|---|---|---|
| $Q$ | Volumetric flow rate | m$^3$/s, m$^3$/h |
| $W$ | Mass flow rate | kg/s, kg/h |
| $v$ | Fluid velocity | m/s |
| $A$ | Cross-sectional area | m$^2$ |
| $\Delta P$ | Differential pressure across the primary element | Pa, kPa, bar |
| $\rho$ | Fluid density | kg/m$^3$ |
| $k$ | Lumped geometry and coefficient constant for the simplified form | depends on implementation |
These simplified forms are useful for understanding the compensation logic. They do not replace the full ISO 5167 treatment with discharge coefficient, expansibility factor, Reynolds number corrections, pipe diameter tolerance, installation effects, and other real-world terms.
The key message is simple:
So when density changes, both indicated values change even if the measured differential pressure stays the same.
The ideal gas law is:
Using:
and:
we can rewrite gas density as:
| Symbol | Meaning | Typical units |
|---|---|---|
| $P$ | Absolute pressure | Pa, kPa, bar abs |
| $V$ | Volume | m$^3$ |
| $n$ | Amount of substance | mol, kmol |
| $R$ | Universal gas constant | J/(mol K), J/(kmol K) |
| $T$ | Absolute temperature | K |
| $m$ | Mass | kg |
| $MW$ | Molecular weight | kg/kmol or g/mol |
| $\rho$ | Gas density | kg/m$^3$ |
This equation immediately shows why compensation is necessary. Density increases with pressure, decreases with temperature, and changes with gas composition through molecular weight.
If the gas composition changes, the error cannot be corrected by pressure and temperature alone. Molecular weight must also be updated.
Let the reference or design condition be identified by subscript $d$, and the actual operating condition by subscript $a$.
Then:
Taking the ratio:
| Symbol | Meaning | Typical units |
|---|---|---|
| $\rho_d$ | Density at design/reference conditions | kg/m$^3$ |
| $\rho_a$ | Density at actual operating conditions | kg/m$^3$ |
| $P_d$ | Design/reference absolute pressure | Pa, kPa, bar abs |
| $P_a$ | Actual absolute pressure | Pa, kPa, bar abs |
| $T_d$ | Design/reference absolute temperature | K |
| $T_a$ | Actual absolute temperature | K |
| $MW_d$ | Design/reference molecular weight | kg/kmol or g/mol |
| $MW_a$ | Actual molecular weight | kg/kmol or g/mol |
This is the most useful equation in the whole article because it links flow error directly to operating changes.
Several practical rules follow from it:
In other words, using a design density for a live gas process is only acceptable when the real process actually stays close to that reference state.
Suppose the uncompensated volumetric flow is calculated with design density:
The corrected volumetric flow must use actual density:
Dividing both expressions gives:
Substituting the density ratio:
| Symbol | Meaning | Typical units |
|---|---|---|
| $Q_u$ | Uncompensated volumetric flow (calculated with reference density) | m$^3$/s, m$^3$/h |
| $Q_c$ | Compensated volumetric flow | m$^3$/s, m$^3$/h |
| $k$ | Lumped geometry and coefficient constant for the simplified form | depends on implementation |
| $\Delta P$ | Differential pressure across the primary element | Pa, kPa, bar |
| $\rho_d$ | Design/reference density | kg/m$^3$ |
| $\rho_a$ | Actual density | kg/m$^3$ |
| Pd, Pa, Td, Ta, MWd, MWa | Same definitions as Section 6.1 | see Section 6.1 |
This equation tells us that an uncompensated gas volumetric flow reading must be multiplied by the square root of the density-ratio correction.
If actual density is lower than design density, the corrected volumetric flow is higher than the uncompensated value. That is why a fixed-density volumetric indication can under-report gas flow when the process becomes hotter or less dense.
Now consider the mass flow form.
If the system calculates mass flow using design density:
The corrected mass flow must be:
Dividing these expressions gives:
Substituting the gas-density ratio:
| Symbol | Meaning | Typical units |
|---|---|---|
| $W_u$ | Uncompensated mass flow (calculated with reference density) | kg/s, kg/h |
| $W_c$ | Compensated mass flow | kg/s, kg/h |
| $k$ | Lumped geometry and coefficient constant for the simplified form | depends on implementation |
| $\Delta P$ | Differential pressure across the primary element | Pa, kPa, bar |
| $\rho_d$ | Design/reference density | kg/m$^3$ |
| $\rho_a$ | Actual density | kg/m$^3$ |
| Pd, Pa, Td, Ta, MWd, MWa | Same definitions as Section 6.1 | see Section 6.1 |
This is the opposite trend from volumetric flow. If actual density falls below design density, the corrected mass flow is lower than the uncompensated value.
That distinction is essential in plant discussions. Operators often say “the flow is wrong”, but they do not always distinguish whether they mean actual volumetric flow, standardized volumetric flow, or mass flow. Those are not interchangeable quantities.
Consider a gas orifice installation configured with these design values:
Later, the process runs at:
The density ratio becomes:
Therefore:
If the uncompensated volumetric flow is $1000\,\text{m}^3/\text{h}$, the corrected volumetric flow is:
If the uncompensated mass flow is $1000\,\text{kg}/\text{h}$, the corrected mass flow is:
This example shows why the same uncompensated DP signal can mislead two different users in two different ways depending on which flow basis they are looking at.
The most frequent errors are not mathematical. They are conceptual.
Gas density relationships require absolute pressure. If gauge pressure is used directly, the compensation factor will be wrong.
The ideal gas law requires absolute temperature. Convert temperature to kelvin before applying any density-ratio formula.
Fuel gas, flare gas, biogas, and off-gas composition can drift significantly. If $MW$ changes, pressure and temperature compensation alone are not enough.
A compensated actual volumetric flow is not the same as a standard flow or normal flow. If commercial reporting is based on standard conditions, you still need the appropriate standardization step.
The formulas in this article explain the density effect clearly, but a production-grade orifice calculation should also consider expansibility, discharge coefficient, beta ratio, Reynolds number, straight-run effects, tapping arrangement, and applicable standards such as ISO 5167.
An orifice plate in gas service does not “lie” because the primary element is defective. The apparent lie appears when a correct differential pressure measurement is converted into flow using an incorrect density.
For a gas:
If the operating conditions move away from the design state, compensation in pressure, temperature, and, when necessary, molecular weight is not optional. It is part of obtaining a defensible flow value.