Mud gas analysis: wetness, balance and character ratios

The mud log is often the first sign of what a reservoir holds, hours before any wireline log and days before a test. Read well, the light hydrocarbons in the mud tell oil from gas, productive from residual, and a reservoir from a coal. This article works through an 8½″ section with four shows, set in Western Desert conditions in Egypt, with DrillIQ's gas analysis engine: lag, the trap's extraction efficiency, Haworth's wetness, balance and character, and Pixler's ratios.

What the gas in the mud can tell you

As the bit cuts rock, the gas in its pores goes into the mud. At surface a trap stirs the returns and draws the gas to two detectors: a total gas (TG) detector and a chromatograph that splits out methane to pentane, C1 to C5. The amount of gas depends on porosity and saturation, but also on rate of penetration, flow rate, hole size and mud weight, so total gas alone is a poor guide to the fluid. The proportions of C1 to C5 are steadier, and that is where the ratios come in.

Two steps come before any ratio. The gas has to be put back at the depth it came from, and the chromatograph's readings have to be trusted: checked against total gas and corrected for what the trap fails to extract.

Lag: putting the gas back at its depth

Gas leaves the bit with the mud and reaches surface once the pumps have moved the whole annulus above the bit. The lag is set by a volume:

Pumps-off time at a connection adds to it, because no mud moves. DrillIQ builds the annulus from the casing, the open hole and the BHA from the bit up, and a carbide or other marker check scales the open-hole part for washout. In this example the hole is taken as in gauge.

The lag matters most in fast, thin sands. While the gas travels up, the bit keeps drilling. At 20 m/h and a 43.6 min lag the bit is 14.5 m deeper by the time the gas from the top of the sand reaches surface, which here is most of an 18 m sand.

From the trap to the ratios

A gas trap strips the light components from the mud more completely than the heavy ones. Uncorrected, it under-reads C3 to C5, and every ratio leans towards gas. The extraction efficiency of each component comes from a calibration run, the same mud through the trap and through a reference extractor, and each reading is divided by it. A data-quality check comes first: the gas quality ratio, GQR = TG / (C1 + 2C2 + 3C3 + 4C4 + 5C5), compares the two detectors and should sit near 1.

Haworth and co-authors (1985) defined three ratios from the corrected components:

Wh

wetness, %: below 0.5 very dry gas; 0.5–17.5 gas; 17.5–40 oil; above 40 residual or heavy oil

Bh

balance: above Wh leans gas, below Wh leans oil, above 100 excessively dry gas of low productivity

Ch

character: in the gas band, above 0.5 the gas carries liquids

Pixler's (1969) ratios C1/C2, C1/C3, C1/C4 and C1/C5 give an independent check. C1/C2 between 2 and 15 points to oil, 15 to 65 to gas, and outside that to non-productive rock. The four ratios should rise from C1/C2 to C1/C5. A ladder that falls points to a water-bearing, residual or tight zone.

DrillIQ types only shows, readings more than 1.2 times the rolling background and at least 50 ppm above it, and takes its confidence from agreement: Pixler's C1/C2 band and the side of Wh that Bh sits on must both agree with the Haworth class for high confidence. A falling Pixler ladder caps the confidence at low, and a missing extraction correction caps oil and wet-gas calls at moderate.

The example section

The example is a vertical well in a basin of the Western Desert, where Lower Cretaceous sandstones carry oil and deeper Jurassic sandstones, among coaly shales, carry gas and condensate. The 8½″ section runs from the 9⅝″ shoe at 3,250 m to 3,850 m with KCl-polymer mud.

A hand check on Sandstone A

1. Lag. Casing annulus: π/4 (8.681² − 5²) in² = 0.02552 m³/m, × 3,250 m = 82.9 m³. Open hole to 3,421 m, round the collars, jar and HWDP: 2.8 m³. Total 85.8 m³, at 520 gpm (1.968 m³/min): 43.6 min.
2. Extraction efficiency. In the calibration the trap read 684 ppm of C3 against the reference's 900: 0.76. At 3,430 m the trap read C1 to C5 as 3,960.3, 417.4, 264.4, 97.8 and 46.4 ppm; divided by 0.93, 0.86, 0.76, 0.64 and 0.52 they become 4,258.4, 485.3, 347.9, 152.8 and 89.2 ppm, 5,333.7 ppm in all.
3. Quality. On the trap's own readings, GQR = 6,448.3 / (3,960.3 + 2 × 417.4 + 3 × 264.4 + 4 × 97.8 + 5 × 46.4) = 6,448.3 / 6,211.5 = 1.04, inside 0.8–1.2.
4. Wetness. 100 × (485.3 + 347.9 + 152.8 + 89.2) / 5,333.7 = 20.2 %: the oil band. Uncorrected, the same reading gives 17.3 %, in the gas band.
5. Balance and character. Bh = (4,258.4 + 485.3) / (347.9 + 152.8 + 89.2) = 8.0, below Wh: oil. Ch = (152.8 + 89.2) / 347.9 = 0.70.
6. Pixler. C1/C2 = 4,258.4 / 485.3 = 8.8: the oil band.
7. Show strength. 6,448.3 ppm against a rolling background of 764.8 ppm is 8.4 times: a strong show (5–10).

The engine gives the same numbers, and types the reading as oil with high confidence: Pixler and the Bh side both agree.

Total gas, background and event gas

The shales give 750–1,300 ppm, and the rolling background follows them. The four sandstones and the coal stand out above it. The dashed trace is the same gas plotted at the bit depth when it reached surface: every show sits 4–17 m too deep, most where the rate of penetration is highest. Sandstone C's first gas, from 3,706 m, came up when the bit was at 3,723 m.

Two other things stand out. Connection gas, the gas that seeps into the hole while the pumps are off, shows at every stand at 1.7–2.0 times the gas around it, and the trip gas after the bit trip at 3,600 m reaches 6,123 ppm. Both are kept out of the background and are never a show; their size is a check on the mud weight's margin, and a rising trend is a warning. And the top of the coaly shale reads as a weak dry-gas show for 23 m, until the 30 m background window fills with the new, gassier rock.

Reading wetness and balance

In the shales Wh is about 9.5 % and Bh about 24, above it: gas, as shale gas usually is. DrillIQ leaves it untyped, since background gas is not a reservoir show.

In Sandstone A Wh rises to 20.0 % and Bh falls to 8.1, crossing below it: oil. Without the extraction correction (the dashed curve) Wh is 17.1 %, just inside the gas band, and the engine types the sand as light oil or oil-associated gas at moderate confidence. The correction moves it into the oil band at high confidence.

In Sandstone B Wh is 43.8 % and Bh 1.8: heavy or residual oil. The show is moderate, 2.2 times the background, and a reservoir engineer might be tempted. Pixler's ladder falls, so the confidence is low, as the next section shows.

In the coal the gas is almost pure methane: Wh 2.9 % and Bh 126, beyond Haworth's 100. That is excessively dry gas of low productivity. A coal can give a show of 5,000 ppm and nothing to produce.

In Sandstone C Wh is 10.4 % and Bh 18.2, above it: wet gas.

Reading character and the Pixler ladder

Character separates a dry gas from one carrying liquids. Sandstone C reads 0.44, below 0.5: wet gas without much liquid, so the class stays wet gas. Sandstone A reads 0.69 and Sandstone B 1.58, both heavy-end-rich, as oils are. DrillIQ only uses character when C4 and C5 each reach 20 ppm, about ten times a chromatograph's detection limit, because near-noise C4 and C5 can turn a dry gas into a false light oil.

Sandstone B's ladder falls at the second step: C1/C3 is 4.0 against C1/C2 of 7.1. Pixler read that as rock that will not produce what its gas suggests, often a water-bearing sand holding residual oil. That is the call here: a moderate show with heavy, residual composition and a falling ladder is a reason to wait for the wireline logs before planning a test.

Practical guidance

Lag-correct before you correlate. A show logged 15 m deep can sit opposite a shale on the wireline log. Check the lag with a carbide or another marker at least once a section, and after any change of hole size, BHA or flow.

Calibrate the trap. Without an extraction-efficiency correction, light oils drift into the gas band. One calibration run per mud system is enough to correct every reading.

Read the ratios only on shows. Background gas is a mix of recycled gas, cavings and slow seepage; its ratios describe the mud as much as the rock.

Use two methods. Where Haworth and Pixler agree, the call is firm. Where they don't, or the ladder falls, say so in the report and wait for other data.

Watch the event gas. Connection and trip gas that grow from stand to stand say the hole is taking gas while static: a mud-weight question for the well-control team before it is a geology one.

A checklist for the rig team

• Enter the hole geometry, BHA and flow, and check the lag with a carbide each section.
• Run a trap calibration for each mud system and apply the extraction efficiency.
• Check GQR on every show: outside 0.8–1.2 read the ratios with care, outside 0.5–2 don't type the fluid.
• Type only shows against the rolling background, and note where the background is still settling.
• Read Wh and Bh together, then Ch in the gas band.
• Check the Pixler ladder before calling a show productive.
• Log connection and trip gas separately and watch their trend.
• Treat coal gas as a show of low productivity, whatever its size.

How DrillIQ handles it

DrillIQ's Gas Analysis module takes the chromatograph and total-gas channels from WITSML or a file, in depth and in time, and lag-corrects them from the hole geometry, BHA, pump strokes and lag checks, with pumps-off time accounted for. It applies each mud interval's base-oil baseline and extraction efficiency, checks GQR and data quality, and computes composition, Haworth's ratios, Pixler's ratios and ladder, the rolling or hand-picked background and show strength. Only shows are typed, each with every method's vote, its confidence and the warnings behind it, and connection, trip and recycled gas are kept apart.

Show intervals go to a review workflow and into the gas-log report, with prognosis against actual and comparison with offset wells, and H₂S and very strong shows raise alerts. Everything above was worked out with the same engine.

Book a demo → See all capabilities

References

1. Haworth, J.H., Sellens, M. and Whittaker, A. Interpretation of Hydrocarbon Shows Using Light (C1–C5) Hydrocarbon Gases from Mud-Log Data. AAPG Bulletin 69(8), 1985.
2. Pixler, B.O. Formation Evaluation by Analysis of Hydrocarbon Ratios. Journal of Petroleum Technology 21(6), 1969 (SPE 2254).
3. Whittaker, A. Mud Logging Handbook. Prentice Hall, 1991.

About this example. The well in this article is a representative scenario set in typical Western Desert, Egypt conditions and worked through with DrillIQ's engines. It is not a client's well, and its figures are not field results.