Point to Patch Convertor – MIDAS Support

Intro

A wheel is a point in the model and a contact patch on the bridge, and closing the gap between those two facts is done by hand on almost every assessment. The clause gives a contact area and two spread rates, so the effective patch has to be worked out layer by layer for each wheel; then the deck decides what happens next. On a plate deck the patch has to be clipped against the mesh and turned into a pressure on every element it touches, because a MIDAS pressure load covers a whole element face and cannot cover part of one. On a grillage it has to be split across the girders either side by the lever rule and then written as a partial-span UDL over the right length of the right elements, with the start and end stations expressed as fractions of each element. A single vehicle carries dozens of wheels, an assessment carries several vehicles at several positions, and every one of those numbers is arithmetic that has to come out to the load you started with. The Point to Patch Convertor does that conversion: it reads the concentrated loads already in the model, disperses each one through the surfacing and the slab to the plane the code names, draws the result before it writes anything, and applies it as native MIDAS pressure loads or partial-span beam loads with the total force conserved and confirmed by reading it back.

Developed with

MIDAS CIVIL NX 2026 (v2.2) — plugin version 1.0.0

Benefits of this plugin

Dispersal taken straight from the clause — CS 454 spreads 2V:1H through surfacing and fill under clause 6.7.1, 1V:1H through concrete under clause 6.7.2, and takes the load down to mid-depth of a reinforced concrete slab under clause 6.7.3(3). EN 1991-2 clause 4.3.6(2) spreads at 1H:1V through both layers to the centroid of the slab. The two codes are genuinely different on 100 mm of surfacing, and the plugin keeps them apart rather than averaging them into one convenient rule.
Fifteen contact areas already measured from the codes — the LM1 tandem wheel at 0.40 m square from clause 4.3.2(1)(a), LM2 at 0.60 m transverse by 0.35 m longitudinal from clause 4.3.3(4) and Figure 4.3, the fatigue wheels of Table 4.8, the LM3 axle lines of Annex A, the footbridge accidental vehicle, the 0.3 m square of CS 454 clause 6.11 and Annex B, and the HB wheel of Appendix C2, whose patch is not fixed at all but derived from the wheel load at an effective pressure of 1.1 N/mm².
Both of the places a wheel can already be — concentrated loads arrive as nodal loads and as point loads defined along a beam, because a wheel lands mid-member more often than it lands on a node. Both tables are read, and a concentrated moment is reported rather than silently dispersed as though it were a force.
The guide diagram is drawn to scale — a section through the wheel shows the surfacing band, the slab band, the contact patch, the dispersal cone kinking where the spread rate changes, and the target plane the load stops at. Both axes share one scale, so the angle you see is the angle being used; stretching the depth to fill the box would make every angle a lie.
Plate decks receive native pressure loads — the patch is clipped against every element it overlaps and each one is given the share of the load its overlap earns, smeared over its own area. That is not a shortcut: a MIDAS pressure load is keyed to an element and covers the whole face, so there is no way to load part of one, and clipping is the only faithful representation the table allows.
Grillages receive partial-span UDLs by the lever rule — the patch is sampled across its width and each strip is shared between the girders either side in inverse proportion to the lever arms, then written as a UDL over the patch footprint, split at every element boundary it crosses. Transverse members deliberately receive nothing, because the lever rule has already used the transverse stiffness and loading both ways counts the load twice.
Each share lands level with the wheel — the lever rule describes a strip of deck cut square across the traffic direction through the wheel, so every reaction it delivers is at the wheel's own position along the bridge. On a skew or fanned layout the nearest point on each girder is a different point entirely, metres away and a different distance for each one.
The model is treated as three-dimensional — joints are identified by their full coordinates, so a deck girder can never chain onto an arch rib or a bracing member passing over it, and member stations are true 3D arc lengths because MIDAS stores a load position as a fraction of an element's own length. A deck plane band keeps members far above or below the wheel out of the distribution, and names the ones it excluded.
Nothing is lost — where a dispersed patch runs past the end of a girder chain or past the edge of the mesh, that share is carried by the length or area that remains rather than dropped. The wheel is on the bridge and the bridge carries all of it. The overrun is still reported with the member, its length and the numbers, and the behaviour can be switched off to see the raw geometric truncation instead.
One output load case per source load case — loads read from different cases are dispersed and written separately, so HA and HB stay separable in the combinations that follow. Once two sets of loads share a case name in the model there is no way to tell them apart again, so merging them would quietly destroy something that cannot be recovered.
Replace or add — regenerating removes everything the plugin wrote before and rewrites it, so a re-run never accumulates duplicates or orphans; adding leaves all of it alone and creates new cases alongside, uniquifying any name already taken, for building up several patch cases and combining them by hand.
Coincident loads are dispersed separately, never merged — the j-end load of one element and the i-end load of the next land on the same node, and merging them looks harmless because the force is conserved. It is not: an HB patch is derived from the wheel load, so a merged load would change the patch size and every pressure with it. They are reported, so that a genuinely duplicated entry can still be spotted.
The mesh is judged against the patch — where an element is large relative to the smaller patch dimension the affected elements are named, and the report says plainly that local moments will be unreliable at that density, because the smeared pressure gives the right total force and the right global effects while averaging the local peak away.
The working model can be restricted — by structure group or by height band, governing both which elements may receive load and which source loads are read. An element with some nodes in the band and some outside straddles the boundary, so it is excluded and listed by number rather than quietly half-loaded.
Every write is confirmed by reading it back — the load cases, the pressure loads and the beam loads are read out of the model and counted per case after they are written, and other load cases sharing the same elements are read, preserved and written back untouched. Your source point loads are never modified or deleted.

How to use this plugin?

Open the bridge model in CIVIL NX and start the plugin — the connection details are supplied automatically, and the panel opens on the Model step with nothing read yet.
Read the model — press Read model. Nodes, elements, plate thicknesses, structure groups, load cases and the model's own units are read in one pass and summarised, and every number shown from that point on is in the model's units.
Restrict the working model if it holds more than the deck — tick the structure groups that make up the deck, or set a height band. On a model that also carries piers, an arch or bracing this is worth doing: it sharpens the plate-or-grillage detection and keeps the dispersal on the deck where it belongs.
Check the model type — the plugin reports whether it reads the selection as a plate deck or a grillage, with its reasoning and a confidence, and can be overridden. A slab with a downstand girder in it and a true grillage are genuinely ambiguous, so the detection recommends rather than decides.
Choose the source loads — filter by load case, by node or element number, or by a magnitude threshold, and see every load that will be converted with its position, its magnitude and any warning attached to it before anything is calculated.
Choose the vehicle and the spread — on the Dispersal step pick the code and the wheel, then give the surfacing or fill thickness and the slab thickness. The guide diagram redraws as you type, and the effective patch, the dispersal depth and the patch area are stated underneath it.

Check the preview — the plan view shows every dispersed patch on the deck, the elements that receive load shaded by how much they receive, and the source point loads. The source total, the applied total and the force check sit above it, computed from the same calculation the write will use, so the two cannot disagree.

Read the checks — whole-job notes come first and are never truncated: how the girders chained, which members were excluded and why, the coarse-mesh warning with its element numbers, and any patch that overruns a member end, with the member's length, the patch length and how far it runs past.
Write to the model — the panel lists what it will create before you press the button: each source load case, the case it will be written as, how many loads and how much force. Choose whether to replace the plugin's previous work or add alongside it, then write. Progress is shown phase by phase, and the result names the cases created and the item and element counts confirmed by read-back.
Export the report — the Report step gives the whole conversion as markdown or CSV: every source load, the dispersal parameters and target plane used, the effective patch, and every element that received load with its pressure or its intensity, its start and end stations and its force. Add the new cases to your combinations, and exclude the source cases so the traffic is not counted twice.

Conclusion

The Point to Patch Convertor takes the mechanical part of putting wheel loads on a deck and does it the way the code describes, on the model you already have. The dispersal follows the clause rather than a rule of thumb, the result is drawn before it is written, the total force that comes out is the total force that went in, and every case is written back separately so the combinations that follow still mean what they should. What is left for the engineer is the part that needs judgement: which vehicle, which position, which plane the load should be taken to, and whether the mesh is fine enough for the effects being checked — and the plugin states its answer to each of those in the report rather than leaving it to be inferred.