Factory-bonded district heating pipe assemblies conforming to EN 253:2019 for a 100 mm nominal steel carrier typically use a seamless or welded P235GH steel service pipe with an outside diameter of 114.3 mm and a high-density polyethylene casing outside diameter of 200 mm. When the steel service pipe is positioned concentrically, the nominal radial polyurethane foam thickness is 42.85 mm; after the permitted eccentricity between the service pipe and casing axes is applied, the local minimum insulation thickness at any one point is governed by the EN 253:2019 minimum foam thickness requirement and by the manufacturer’s declared dimensional tolerance. The factory-applied foam is a closed-cell rigid polyurethane produced by high-pressure metering of a formulated polyol blend and polymeric MDI, injected through the annular cavity while the pipe assembly rotates or is held in a horizontal foaming rig. Post-cure density in the core of a 100 mm assembly is usually controlled in the range 60–100 kg/m³ with an ISO 845 apparent density not less than 60 kg/m³; closed cell content by ISO 4590 is typically above 88%, and the aged thermal conductivity determined according to ISO 8497 at a mean temperature of 50°C is commonly declared in the range 0.025–0.027 W/(m·K). The pipe-to-foam and foam-to-casing adhesion is critical because EN 253:2019 treats the pipe, foam and casing as a bonded composite for axial shear transfer; factory shear strength values for the foam-steel interface determined by the EN 253 axial shear test are generally required to exceed 0.12 MPa at 23°C without adhesive failure. At every cut end of a factory pipe section, the annular insulation face is exposed and becomes the first pour foam boundary when the pipe is joined in the field. This boundary is not merely a thermal discontinuity; it is a mechanical joint where the site-applied two-component rigid polyurethane must bond to the cut polyurethane surface, the steel service pipe, and the inner high-density polyethylene casing wall simultaneously, under conditions that are far less controlled than the factory foaming cell.

At the cut end of a factory-insulated DN100/OD 114.3 mm pipe, the exposed annular polyurethane face usually exhibits a slightly higher surface porosity than the core because cutting severs cell walls and leaves open cells. The end cap or shrink film applied during transport is removed shortly before joint assembly; if the pipe end has been stored outdoors or in high humidity, the exposed foam absorbs water vapour until the local moisture content at the surface can exceed 1.0% by mass, which is enough to impair the bond with the site-poured foam. The site-poured joint cavity is formed by a polyethylene casing sleeve or heat-shrinkable joint cover, and the annular space between the steel pipe and the sleeve typically replicates the factory insulation thickness of 42.85 mm only if the joint kit is correctly dimensioned. In many field installations, the joint foam cavity is slightly thicker or thinner than the factory annulus because the outer sleeve is positioned over the factory casing ends; that step change in radial thickness creates a boundary where the exotherm, cure shrinkage and adhesion stress field differ from those in the factory pipe. Joint foam systems are generally supplied as two-component liquid systems with a mix ratio tolerance of ±2% by weight. High-pressure impingement mixing machines deliver the reacting mixture to the bottom of the cavity through fill ports at flow rates of 4–12 kg/min; the material then rises through the annular space. Cream time at 20°C is usually in the range 15–40 s, and the gel time is often 80–180 s. If the cavity is too cold, typically below 10°C, the cream time increases and the rising foam may not fuse adequately with the factory foam end face before gelation, leaving a visible cold boundary with reduced tensile strength. Moisture on the steel or polyethylene surfaces reacts with isocyanate to form polyurea and carbon dioxide, producing pinholes at the interface and lowering the peel strength of the bonded joint. Published data for this specific boundary configuration is limited, but field release testing on joint foam samples generally follows the same ISO 845 and ISO 844 methods as factory foam rather than a dedicated joint shear test.

When a 100 mm Carrier Pipe Meets a Site-Poured Joint Foam Cavity

The boundary between the factory-bonded pipe and a site-poured joint is not a simple butt joint; it is a three-material transition zone in which the polyurethane reaction front must fill a closed annulus while simultaneously bonding to aged polyurethane, steel and high-density polyethylene. The installation standard EN 13941-1 requires the joint cavity to be dry, clean and shielded from rain, snow and blowing dust before the pour is started. The polyethylene sleeve is normally centred with spacer clips that maintain the radial cavity width within 2–5 mm of the factory annulus, but the overlap where the sleeve covers the factory casing introduces a local expansion volume that is larger than the annular gap farther along the joint. When the liquid foam enters this expansion volume, the linear velocity of the rising front decreases, allowing air to escape from the foam surface but also increasing the risk of void formation if the venting is inadequate. The mixed liquid is injected at the lowest point of the cavity, and the fill rate is selected so that the rise time in the cavity is shorter than the mix system’s cream time; for a DN100 joint cavity with an annular cross-sectional area of approximately 21 000 mm², a fill rate of 6–10 kg/min typically produces a linear front velocity in the range 4–8 mm/s before density compensation. This front velocity is low enough to avoid turbulent air entrapment but high enough to complete the fill before the gel time is reached. At the upper end of the sleeve, vent holes allow displaced air and carbon dioxide to leave; the total vent area is often kept between 3% and 5% of the annular cross-sectional area on larger tank pours, but on small pipe joints the vent area is usually fixed by the kit design rather than by site calculation. Once the foam has filled the cavity and the vent has been sealed by rising material, the reaction continues under a confined volume with limited heat loss. The pressure generated during the rise is lower than that in a closed rigid mould, but it is still sufficient to deform a thin polyethylene sleeve if the fill rate is excessive; casing expansion of more than 1–2% in diameter during joint foaming is a practical indication that the fill rate should be reduced for subsequent pours. The cured joint is then checked for voids by tapping, by thermal imaging during the exotherm, or by destructive sectioning of first-article joints; acceptance is normally based on the absence of visible voids larger than 5 mm and on a core density that does not fall below 60 kg/m³.

The pour foam boundary on large atmospheric district heating storage tanks is governed by a different geometric and thermal regime than the annular joint of a 100 mm buried pipe. A vertical tank sidewall is typically insulated by introducing liquid polyurethane into an annular cavity bounded on the inner side by the primary steel shell and on the outer side by cladding, lagging or a secondary containment plate. The cavity width is often kept in the range 50–150 mm, and the continuous pour lift height is commonly limited to 1.5–2.0 m per production shift to keep the hydrostatic pressure exerted by the rising liquid below the yield strength of the cell walls. When the lift height exceeds this range, lower sections of the foam can be compressed by the weight of the rising column before final gelation, producing a density gradient that can exceed 10–15% between top and bottom. This density gradient is measurable by ISO 845 core sampling and can be detected as a variation in compressive strength by ISO 844. The boundary between successive lifts is a distinct horizontal pour face where the previously cured foam has developed a closed skin; this skin must be pierced, scarified or mechanically roughened before the next lift unless the formulator has supplied a delayed-cure surface that remains tacky. If the interval between lifts is extended beyond the tack-free time, the exposed surface reacts with atmospheric moisture and carbon dioxide, forming a friable polyurea-rich layer that is poorly adhered to the next pour. On tank sidewalls, adhesion of the foam to the steel shell is normally checked by pull-off testing according to ISO 4624 on representative test panels prepared at the same time as the tank pour; values below 0.15 MPa are cause for rejection because the foam must resist thermal expansion shear between the steel shell and the external cladding during tank filling and draining cycles. The pour boundary at the tank base is even more sensitive because any water that accumulates in the annular cavity before foaming is trapped at the lowest point, and the reaction with isocyanate generates carbon dioxide gas that can create a weak cellular layer. Field procedures therefore require the annulus to be dried with oil-free compressed air at a dew point below −20°C and the steel surface to be warmed to at least 3 K above the ambient dew point before the first lift is poured.

What Limits Adhesion Across the Steel–Foam–HDPE Interface?

Adhesion at the pour foam boundary is not a single material property but a system response controlled by substrate cleanliness, surface energy, primer chemistry and local cure conditions. For the steel service pipe, the surface may be supplied with a thin factory-applied protective layer or may be grit-blasted to a surface preparation grade equivalent to Sa 2½ according to ISO 8501-1. The isocyanate component of the pour foam bonds mechanically to the steel surface roughness and chemically to any hydroxyl or oxide species present, but oil, grease, mill scale and water reduce the wetting area. For high-density polyethylene, the non-polar surface energy is generally below 35 mN/m; without surface treatment the peel adhesion of rigid polyurethane to high-density polyethylene is frequently below 0.05 MPa. Field practice for the polyethylene joint sleeve includes propane or butane flame treatment to raise surface energy above 40 mN/m, or use of a solvent-based primer based on chlorinated polyolefin. The factory-applied pipe casing has a higher surface energy on the inner surface because it is corona or flame treated before foaming in the production line; this treatment disappears with time and cannot be assumed to remain active at the cut end. The steel surface in the joint cavity should be cleaned and warmed to at least 3 K above the dew point before pouring, because condensation forms a water interlayer that weakens the interface. Pull-off adhesion according to ISO 4624 on steel coupons conditioned at 20°C and 50% relative humidity commonly yields values in the range 0.2–0.6 MPa for well-primed surfaces; cohesive failure in the foam is the preferred failure mode. For the high-density polyethylene side, values are usually lower, in the range 0.1–0.3 MPa, even after flame treatment, and the failure is often adhesive. The interfacial transition zone is not covered by a single pass/fail value in EN 253:2019; the bonded pipe standard requires a system axial shear strength rather than an interface-specific adhesion value, which means the quality of the pour foam boundary must be inferred from joint shear testing on representative mock-ups and from visual inspection of sectioned samples.

Cold joint formation at interrupted pour boundaries remains one of the most frequent causes of field joint failure in 100 mm preinsulated pipe systems. When the site foam pour is interrupted, whether because of mixer malfunction, insufficient material, or cavity leakage, the partially risen foam stops expanding before the cavity is filled. The exposed surface then skins over and the internal exotherm decays. If the cavity is refilled after the initial foam has cooled below 40°C, the new pour can bond poorly to the smooth closed cells of the old surface; the resulting cold joint may show a discrete layer with higher density and a void line under X-ray or destructive sectioning. The cold joint is not necessarily detectable by external appearance, but it reduces the flexural and shear capacity of the joint and creates a pathway for water vapour migration. In this scenario, the only reliable remediation is removal of the affected foam back to full sound material and repouring the entire cavity under controlled temperature. Some joint kits include vent ports at the highest point of the cavity; if these vents are undersized or positioned incorrectly, the escaping air and blowing agent can become trapped, generating a void at the top of the annular cavity. For a DN100/OD 114.3 mm pipe with a casing sleeve of matching 200 mm outside diameter, the annular cavity cross-sectional area is approximately 21 000 mm², so a one-metre joint sleeve contains approximately 19–21 L of liquid reactant before density correction, depending on the actual casing wall thickness. A vent port of 10 mm diameter is often insufficient at high fill rates; a minimum total vent area of 3–5% of the annular cross-sectional area is applied on larger tank pours but is not always followed on small pipe joints. The intersection between the factory foam end face and the site-poured foam should be scarified with a wire wheel, vacuumed, and inspected for surface temperature before the pour; if the factory foam surface temperature is below 10°C or above 35°C, the curing profile of the liquid pour changes enough to create a measurable density and adhesion difference across the boundary.

Thermal Degradation and Exotherm Management in Thick Pour Boundaries

In the annular cavity of a DN100/OD 114.3 mm preinsulated joint or in a storage tank sidewall cavity, the polyurethane reaction exotherm is not uniformly distributed. The centre of the thickest cross-section develops the highest temperature because the surrounding foam and substrate act as thermal insulators. When the maximum cavity thickness exceeds approximately 50 mm, the core temperature can rise above 140°C; in confined tank cavities of 100 mm width, core temperatures in the range 150–180°C have been reported in manufacturer technical bulletins, although published data for this specific configuration is limited. Above about 180°C, oxidative degradation of the polyurethane network begins to accelerate; the visible result is a brown or amber discolouration, increased friability, and a decrease in ISO 844 compressive strength that can exceed 30% when compared with a non-scorched core from the same pour. The exotherm also affects the dimensional stability of the boundary. The freshly poured foam expands against the steel wall and high-density polyethylene casing; as it cools, the foam shrinks by roughly 0.5–1.5% linear because the thermal expansion coefficient of rigid polyurethane is in the range 4–8 × 10^-5 K^-1 and the temperature drop from the exotherm peak to ambient can be larger than 100 K. This shrinkage can create microgaps at the foam-steel or foam-high-density polyethylene interface, particularly if the substrate was not preheated or if the substrate temperature was below the dew point. In factory-bonded pipes, the axial shear test of EN 253:2019 is designed to detect gross detachment, but small microgaps below the detection limit can still increase water vapour transmission and reduce long-term thermal performance. For storage tanks, the pour boundary is also exposed to cyclic shell movement; the steel shell expands and contracts with operating temperature, while the external cladding remains closer to ambient. The differential strain across the foam layer at a tank operating between 20°C and 95°C is determined by the difference in the coefficients of thermal expansion of steel and the outer cladding material and by the radial thickness of the insulation. Rigid polyurethane with a compressive modulus below 10 MPa at 10% strain according to ISO 844 can accommodate some of this strain by elastic deformation, but repeated cycling beyond the proportional limit can cause fatigue cracking at the foam-cladding interface.

The intersection between the factory-applied annular foam end face and the site-poured joint foam on a DN100/OD 114.3 mm assembly is a high-risk discontinuity because the factory foam end face may carry cutting swarf, release agent, moisture or transport-related surface contamination. In practice, joint installers address this by cutting back the factory foam by 10–20 mm from the pipe end, exposing a fresh foamed surface before the outer sleeve is mounted. The exposed steel service pipe is then abrasive-blasted or wire-brushed to a clean profile with a surface roughness Rz in the range 50–100 µm. The inner surface of the high-density polyethylene joint sleeve is either flame-treated or primed, and the sleeve is centred to maintain a radial gap that matches the factory insulation thickness within 2–5 mm. Once the liquid foam is introduced from the lowest fill port, it rises through the annular space in a flow pattern that is affected by the spacing of the steel pipe and the sleeve. At the top of the joint, the vent port must release air and carbon dioxide without allowing foam to seal the port prematurely. A vent port of 12–16 mm diameter is common for small pipe joints; on larger tank cavities, vent spacing is based on the target fill time and the free-rise density of the foam. The pour is considered complete when foam of uniform colour and texture appears at all vent locations, but visual appearance alone does not confirm the absence of internal voids. Sectioned test joints for a 100 mm carrier pipe have shown that voids are most likely to occur at the abrupt change in cross-section where the site sleeve overlaps the factory high-density polyethylene casing, because the rising flow expands into the wider region and can create a low-velocity recirculation zone at the step. Packers and restriction rings are sometimes used to reduce this step and to guide the foam front, but they introduce an additional boundary layer that can become a crack initiation site if the inserted material is not fully wetted by the rising liquid. Published data for this specific configuration is limited, and most acceptance criteria are therefore based on destructive evaluation of first-article joints and on thermal imaging of the filled cavity after cure.