Stop Spacing Road Gullies at 30 m: Inlet Spacing Is Calculated, Not Copied

Almost every drainage drawing carries the note "gully spacing 30 m". Ask where it came from and the answer is "experience" or "we have always done it that way". It is neither a code value nor a general rule of thumb - it is the answer to one specific set of conditions: roughly a 12 m contributing width per side, a flat single-grate inlet, and a 1.5 times margin. Change any one of those and the number is void.

The consequences go both ways. On small catchments it wastes money - where the contributing width per side is only 6-8 m, the calculation allows 46-93 m, so forcing 30 m means 40-67% more inlets than necessary. On large catchments it is not enough - an undivided single-carriageway road, a plaza or a car park with a 16-20 m contributing width works out at 18-23 m, so 30 m leaves the kerb line overflowing and the pavement ponding, which fails inspection and means cutting the road open again.

This article works the whole verification chain through, so you can see whether your project should be at 25 m or 50 m.

1. The code gives a range and leaves the calculation to you

Start with the actual clauses, so nobody argues from memory:

Clause 5.7.1 is the point. Spacing is an output of flow and capacity; the 25-50 m band is simply where a sensible answer lands. The commentary to the code explicitly criticises laying inlets out uniformly by road length without calculation, on the grounds that it wastes investment and still drains badly - which is almost a description of the 30 m rule of thumb.

2. How much water one grate can actually take

Everything rests on one number: the flow a single inlet can swallow in design. The authoritative figures come from the national standard drawing set for road gullies (S2, cast iron frame and grate), measured on a 1:1 hydraulic model at 3-3.5 per mille longitudinal gradient, 1.5% crossfall and 40 mm grate head:

These are nominal figures and must not be used raw. The test condition includes 40 mm of head at the grate. On site, if the crossfall is short or the grate level has not been set below the pavement, that head does not develop and the nominal figure is not achieved. Hence the blockage reduction below, before any margin is applied.

3. The counter-intuitive step: the vertical inlet is lower rated but performs better

Leaf litter, grit and debris are the normal condition. The commentary to GB 50014-2021, 5.7.2 gives an explicit reduction basis:

• Flat inlets: design assuming 50% blocked
• Vertical inlets: design assuming 10% blocked

(The earlier GB 50014-2006, 4.6.3 instead said inlets liable to blockage should have their capacity multiplied by 0.5-0.7, and 0.7 is common in practice. Both bases coexist; the figures below follow the current 2021 commentary.)

The ranking reverses. The vertical inlet is nominally 15 L/s against the flat inlet's 20 L/s - 25% lower. After blockage it is 13.5 L/s against 10.0 L/s - 35% higher. Selecting on nominal capacity gets it backwards. The price of the vertical inlet is that the channel must hold some depth and the kerb face must stand high enough, which is covered in section 5.

4. The spacing limit, and what falls out of it

Write out how much water one inlet serves. With inlets on one side only, each serves a contributing area of "contributing width B x spacing L":

• psi, runoff coefficient: 0.85-0.95 for concrete or asphalt pavement (0.9 used here); weight by area where planting is included, using 0.15 for soft landscape
• q, design rainfall intensity: L/(s.ha), from the local intensity formula q = 167 x A₁(1 + C x lgP) / (t + b)ⁿ, with return period P and time of concentration t. Ordinary roads take a 1-3 year return period; flood-prone and underpass locations take more
• 1.5: the lower margin required by GB 50014-2021, 5.7.2; important and flood-prone locations take 3.0
• B: contributing width per side, in metres. For symmetric two-sided layout, B is half the carriageway; for an undivided road, a plaza or a car park drained from one side only, B is the full width

Contributing width table (psi = 0.9, q = 200 L/(s.ha), margin 1.5)

That table is the whole answer, and it reads in three lines:

1. Where B is 8 m or less, the governing condition is the 50 m code cap, not the flow. Spacing at 30 m here gives 50-67% more inlets than 45-50 m. Those extra inlets will never work near capacity - pure waste
2. At B = 12 m the calculation gives 30.9 m. That is where the folk 30 m comes from: a 12 m contributing width, a flat single grate and a 1.5 times margin. It also explains why it "always seems to work" - the half width of an urban access road is very often around 12 m
3. At B of 16 m or more the calculation gives 18-23 m, so 30 m is insufficient. Plazas, car parks and undivided roads are a real ponding risk at 30 m

Check: on a 500 m access road, is 30 m really the limit

Take a 12 m contributing width, psi = 0.9 and q = 200 L/(s.ha), and test a single flat grate at different spacings:

Worth showing a designer. At 30 m with B = 12 m, only 3% of the 1.5 times margin is left - nine tenths of the code margin has been eaten. Move any single variable up a notch - return period from 1 year to 3, a revised rainfall intensity formula, runoff from adjacent planting - and it stops being adequate. Two ways to restore margin: tighten to 25 m, or swap the flat single grate for a combined single grate (q_eff = 15.0, adequate even at 45 m). The second is usually cheaper.

The cost of the extra inlets

On the same 500 m single-sided run with B = 8 m:

At an all-in rate of about CNY 450 per gully (brick shaft, frame and ductile iron grate, C250), typical of our municipal and estate projects, those 5 extra inlets cost roughly CNY 2 250 on 500 m one side, or CNY 4 500 both sides. Against a 500 m estate road drainage package of CNY 150 000-300 000 that is 1.5-3% - not fatal, but it shows the selection step was never calculated, and the same drawing copied across eight estates is close to CNY 40 000.

Basis statement: the CNY 450 per gully is an experience rate used by ASPER in municipal and estate projects. It varies widely with regional labour, shaft depth, masonry and grate class - and far more with proprietary plastic or precast concrete gullies. It is not a code or schedule rate. Formal pricing must be re-built from the local schedule of rates and current material prices.

5. Slope and grate level: skip these and the calculation means nothing

The water reaches the inlet because the longitudinal and cross falls push it there. If those are wrong, the capacity calculated above is discounted in service.

Grate level is the one most often missed. The drawing says "inlet to standard detail" and the site sets it flush with the pavement, so the 40 mm head in the test condition never develops and real capacity drops below half the nominal figure. Call this out separately at the pre-construction briefing and measure every one at handover.

6. Six locations that always need more inlets, however good the spacing

1. Road junctions and upstream of pedestrian crossings - the flow direction changes and pedestrians cannot be left in standing water
2. Vertical curve low points and gradient transitions - flow piles up where the fall flattens
3. Low points at main/slip road entries and the foot of underpass approaches - the most critical ponding locations; raise the return period and take the margin at 3.0
4. Bus stops - the platform interrupts the drainage path and footfall is high
5. Metro entrances - spacing should not exceed 30 m
6. Road bends and transitions into sunken planting - intercept before the flow path turns

Five common mistakes

1. Treating 30 m as a code value - the code gives 25-50 m and requires calculation. 30 m is one particular answer for B around 12 m with a flat single grate
2. Laying out uniformly by road length instead of calculating from contributing area - explicitly criticised in the code commentary for wasting investment while still draining poorly
3. Using nominal capacity with no blockage reduction - flat inlets at 50% blocked (10 L/s effective) and vertical inlets at 10% (13.5 L/s effective). Skipping this doubles the assumed capacity
4. Getting flat and vertical inlets the wrong way round - the vertical inlet is nominally lower at 15 L/s but effectively higher at 13.5 L/s against the flat inlet's 10.0 L/s. Selecting on nominal figures reliably picks the wrong one
5. Calculating the road but not the run-on - runoff from adjacent planting, neighbouring plots and plazas crosses the kerb into the carriageway. This is the single most common reason inlets are overwhelmed in practice. Include it in the contributing area, and check kerb face height and any interception measures at the same time

In one line: inlet spacing is what one inlet can take divided by what one metre of road produces. The code only fixes the 25-50 m band in which the answer lands. Work it through and you will find small catchments should sit at 45-50 m, large ones at 18-23 m - and 30 m falls in the middle, wrong in both directions.