How much does psi value junction modelling cost for building fabric heat loss calculations?

Psi value junction modelling — the numerical calculation of linear thermal transmittance (ψ-values) at building junctions such as wall/floor intersections, lintels, jambs, sills and roof/wall junctions — typically costs between £150 and £600 per junction detail when commissioned from a specialist UK consultancy, or between £0 and roughly £2,000 per year if you model junctions in-house using software you already own. The full range depends on who does the work, which tool is used, how many variants are needed, and whether the output must be accredited for SAP, SBEM or Passivhaus certification.

What a psi value actually is and why it carries a cost

Also worth reading: What are the most effective thermal bridge mitigation strategies for urban building envelopes in extreme heat climates? · What is the true urban heat mapping technology cost comparison for municipal planners? · How do automated building permit review systems accelerate municipal housing approvals?

A psi value quantifies the extra heat lost through the geometric and material discontinuity where two building elements meet, measured in W/mK. It is not something that can be read off a datasheet for your specific construction; it has to be derived from a 2D (or occasionally 3D) steady-state thermal simulation of the junction geometry, solved to convergence with appropriate boundary conditions and internal surface resistances. Because each junction is a separate finite-element or finite-difference model, each one represents a discrete unit of engineering labour plus software time.

The cost exists because the process is genuinely skilled work. A modeller must draw the junction accurately, assign correct conductivity values for every material layer, set external and internal temperatures (commonly 20°C inside and 0°C outside for UK conventions), apply surface resistances of 0.13 m²K/W internally and 0.04 m²K/W externally, mesh the model finely enough around thermal bridges, and then extract the flux integral correctly. Errors at any stage produce a psi value that looks plausible but is wrong by 10–30%, which propagates directly into SAP calculations, Part L compliance figures and ultimately the EPC rating.

It is worth being blunt about the economics: a single junction model might take an experienced analyst 1–3 hours including checking, while a junior or an automated pipeline might do it faster but with higher error risk. At typical consultancy rates of £60–£120 per hour in the UK as of 2026, the arithmetic alone puts a floor under pricing before any commercial margin is added.

Typical price ranges in 2025–2026

Market rates have been fairly stable over the last few years, with modest inflation. Based on published price lists from UK thermal modelling consultancies and BRE-linked assessors, the following ranges are representative:

Volume discounts are common and meaningful. Ordering ten or more junctions at once frequently brings the per-junction rate down by 15–25%, so a full set of 15–25 junctions for a typical dwelling might cost £2,500–£4,500 rather than the sum of individual prices. Some consultancies offer fixed-price packages: for example, a complete dwelling junction set for SAP purposes commonly lands between £1,800 and £3,500 depending on construction complexity.

Passivhaus certification pushes costs upward because PHI requires specific junction treatments, tighter tolerances, and often 3D analysis for corner details. Expect £400–£800 per complex junction and a total certification-support modelling budget of £4,000–£8,000 on a demanding project.

How the modelling process works step by step

Understanding the workflow explains where the money goes and helps you interrogate quotes. First, the modeller collects drawings or CAD sections showing every material layer at the junction, including insulation continuity, cavity closers, lintels and fixings. Missing or ambiguous information is the single biggest cause of delays and rework fees, so supplying clean DWG or PDF sections with dimensions and product names up front genuinely reduces cost.

Second, the geometry is built in a 2D conduction solver — TRISCO, THERM, PSI-Value tools within packages like Build Desk or Physibel, or open-source alternatives. Material conductivities are assigned from manufacturer data or standard references such as BRE Digest or ISO 10456 tabulated values. Third, boundary conditions are applied: the standard UK convention uses an internal temperature of 20°C and external of 0°C, giving a temperature difference of 20K, with unheated spaces handled via adjustment factors.

Fourth comes meshing and solving. Convergence checks matter: a coarse mesh can over- or under-predict fluxes near sharp corners. Fifth, the psi value is extracted by subtracting the 1D element heat losses from the total 2D flux through the modelled length. Finally, the result is documented in a report showing geometry, materials, boundary conditions, mesh sensitivity and the resulting ψ-value, formatted so a SAP or SBEM assessor can drop it straight into the calculation. That report is what you are largely paying for — a number without documentation is close to worthless for compliance purposes.

Accredited Construction Details versus bespoke modelling

The cheapest option is to avoid modelling altogether by using pre-calculated values. England's Approved Document L (2021 edition onwards) and the Scottish Technical Handbook reference sets of Accredited Construction Details (ACDs) with published psi values for standard junctions. Using ACDs exactly as drawn — same materials, same dimensions, same build-ups — lets you claim those values at zero modelling cost.

The catch is the word "exactly". Deviate from the published detail by changing insulation thickness, swapping a lintel type, altering the cavity width, or substituting a different insulation product, and strictly speaking the published value no longer applies. Assessors vary in how rigorously they enforce this, but under the 2021+ Part L regime, with its tighter fabric targets (notional spec requiring psi values around 0.05–0.08 W/mK for many junctions), the tolerance for hand-waving has narrowed considerably. If your build-up differs materially from the ACD, bespoke modelling is the defensible route, and the £150–£400 spent per junction is cheap insurance against a failed compliance check or an expensive post-construction air test and re-assessment cycle.

There is also a middle path: some manufacturers publish tested psi values for their own lintels, cavity closers and balcony connectors. These are free to use when the specified product appears in your junction, and they are often better than ACD defaults because they were measured rather than assumed.

Software options compared

If you model more than roughly 10–15 junctions per year, buying software starts to beat outsourcing financially. Here is how the main routes compare:

THERM's free licence makes it the obvious entry point for practices experimenting with in-house capability, though you should validate early results against known benchmarks (the BRE IP 1/06 validation exercises are the classic reference set). Paid European packages offer better reporting formats aligned to UK assessment conventions and formal validation certificates, which matters when an assessor scrutinises your evidence.

One honest caveat: in-house modelling done badly is worse than no modelling at all, because it produces confident-looking numbers that fail scrutiny late in a project. Budget for training — a two-day course typically costs £500–£900 per person — and for a period of shadow-checking against an experienced consultant before you rely on your own outputs.

Common mistakes that inflate cost or invalidate results

The most frequent error is commissioning models too late. Junctions are decided on site and in detailing packages months before anyone thinks about psi values; retrofitting a thermal bridge solution after the structural design is frozen usually means accepting a worse value or paying for redesign. Involve the modeller at RIBA Stage 3–4, when changing a lintel or extending insulation return still costs nothing.

Second, incomplete drawing information triggers rework charges. A quote based on "send us sections" becomes more expensive when the consultant spends hours chasing cavity closer products and insulation lambda values. Provide product datasheets with your first submission and most consultancies will hold their quoted price.

Third, people confuse psi values with U-values or with chi factors (point thermal bridges, W/K). A steel balcony connection needs a chi value from a 3D model, not a psi value, and quoting the wrong quantity in SAP produces errors an assessor will reject. Fourth, ignoring repeatability: if you use the same junction across a housing development, model it once properly and reuse it — but re-model it whenever the specification changes, even slightly, because a 20mm insulation reduction can shift a psi value by 0.02–0.05 W/mK, which is material at Part L current standards.

Finally, beware of suspiciously cheap services quoting £40–£80 per junction with same-day turnaround. Some of these apply template geometry with minimal checking. Ask to see a sample report including mesh screenshots and boundary condition assumptions before committing; a legitimate provider will supply one without hesitation.

When bespoke modelling pays for itself

Do the arithmetic against the alternative outcomes. A dwelling failing its Part L target because assumed default psi values were pessimistic may need additional PV, a heat pump upsizing, or fabric upgrades costing £1,000–£5,000 to compensate. Conversely, accurate low psi values from good detailing reduce the required offset measures, sometimes saving more than the entire modelling fee. On a 50-unit scheme, spending £3,000–£6,000 on a full junction set that demonstrates psi values 0.03–0.05 W/mK better than defaults can eliminate tens of thousands of pounds of compensating measures.

Act early in the design programme: commission modelling alongside the Stage 3 technical design, allow 1–2 weeks for a full dwelling set, and build one revision round into your budget. If you are a small practice doing fewer than ten projects a year, outsource; if you run high-volume housebuilding work, invest in software and training, where break-even typically arrives within 12–18 months.

For context on adjacent performance topics, resources such as the CPD material on roof window efficiency published via building.co.uk illustrate how manufacturers increasingly publish junction-tested data for whole products, a trend worth exploiting before paying for custom models.