Solar Thermal Insulation: Molten Salt Tanks and Receivers

Concentrating solar power insulation is usually specified as a temperature-rating exercise: read the salt temperature, open a product catalogue, pick the material whose classification temperature sits comfortably above it, and move on. That approach survives right up to the first unplanned shutdown, and then it fails in ways the catalogue never described.

A concentrating solar power (CSP) insulation specification is credible only when it treats the hot boundary, the insulation system and the environment as one coupled assembly. Material temperature rating is an entry gate, never the selection criterion. The governing constraints are daily thermal cycling and transients; chloride and other leachable ions; molten-salt leak compatibility; and compressive, shear, attachment, access and wind loads. Every candidate must then be checked layer by layer for interface temperature, moisture state, settlement, jacket movement, inspectability, repair strategy and whole-life cost.

This article works through a nitrate-salt tower and trough plant system by system, because the four thermal boundaries inside one plant have almost nothing in common. The application-level overview of where CSP sits among generation technologies is on our CSP application page, and the wider plant context is on the power generation overview.

1. Why CSP insulation is not a fossil-plant insulation problem

A CSP plant contains several different thermal boundaries, and a material that is suitable for one is not automatically suitable for another. A defensible scope statement must distinguish the receiver, salt containment, external piping and equipment, and the CSP to power-block interface before discussing product grades or thicknesses.

This article covers tower external receivers, molten-salt tanks, salt piping, valves, steam generators and the CSP-facing portions of conventional steam systems. Parabolic-trough systems are included because their receiver vacuum architecture is routinely misunderstood. Linear Fresnel receivers may share aspects of linear focus and vacuum-tube construction, but their geometry, operating temperatures and heat-transfer fluids must be checked project by project. Direct-steam receivers, particle receivers and next-generation chloride salts sit outside the default nitrate-salt scope because their temperatures, containment materials and liner concepts differ materially.

The governing engineering rule is to define the boundary before selecting the product:

• External surface of a tower receiver: directly irradiated, non-uniform heat flux, wind loading, high tube-wall temperatures and access constraints.
• Trough receiver glass envelope: an optical and vacuum boundary that must transmit solar radiation while limiting infrared heat loss.
• External tank wall: a large curved hot surface with wind, seismic, attachment and maintenance constraints.
• Tank bottom: a load path carrying salt, tank and foundation loads while enforcing a temperature gradient toward soil or concrete.
• Pipe or valve: a removable or fixed system with supports, thermal movement, freeze protection, trace heating, precipitation ingress and leak-collection needs.
• Steam generator and steam line: a pressure-boundary interface governed by process temperature, personnel protection, corrosion under insulation, access and turbine-cycle transients.

ISO 12241:2022 provides calculation rules for heat transfer in industrial installations, but predominantly under steady-state conditions. It is useful for sizing and heat-loss checks, and it cannot by itself qualify an insulation system for CSP start-up, cloud passages, tank filling, drainage or other transients.

The anti-copy-paste rule follows directly from these differing boundaries. A fossil power-station insulation package may perform well on a steady steam line while failing on a 565 °C salt tank base, because its compressive creep, ion chemistry or leak cleanability is unsuitable. Conversely, dense refractory may be correct inside a tank-base load path but excessive, brittle and unnecessary on an accessible hot-salt valve. "Same temperature" does not mean "same boundary."

2. Three temperatures must be separated before anything is selected

Three commonly conflated temperatures have different physical meanings, and they must not be interchanged in insulation calculations. In the solar salt demonstration campaigns, nitrate salt left the tower receiver and entered hot storage at 565 °C after being heated from 290 °C. In one external-cylindrical-tubular receiver alloy study, about 600 °C represented a corrosion-related film-temperature limit for one austenitic grade, while other alloys had different limits. In a conceptual 110 MWe tower steam-generator study, live and reheat steam were both 540 °C while salt entered the generator at 565 °C.

The practical consequence is that a single insulation schedule cannot serve all three. A salt line at 565 °C, a superheater outlet at 540 °C and a receiver tube whose film temperature is limited at 600 °C sit in different thermal environments with different allowable interface temperatures, different support behaviour and different acceptable jacket temperatures.

These are specific design boundaries drawn from specific studies, not universal constants. Any project must take its own heat and mass balance, its own salt chemistry specification and its own alloy selection, and derive the interface temperatures from those.

3. A parabolic-trough receiver is a vacuum device, not a fibre-insulated pipe

The receiver's heat-loss control is created inside the component. Ceramic fibre applied to the outside acts at the wrong boundary and can neither reproduce the vacuum function nor address the principal degradation mechanisms.

A conventional trough heat-collection element comprises a metal absorber tube carrying heat-transfer fluid, a solar-selective coating on the absorber, a concentric glass envelope, an annular vacuum, glass-to-metal seals and bellows. The selective coating provides high solar absorptance and low thermal emittance. The glass transmits solar radiation while reducing infrared loss. The vacuum suppresses gas conduction and convection and limits oxidation of the selective surface. Bellows accommodate differential expansion between the metal absorber and the glass envelope.

The thermal-loss paths are consequently radiation between the selective absorber and the glass, residual-gas conduction and convection in the annulus, conduction through seals and supports, and external convection and radiation from the glass. A fibre blanket wrapped around the glass would add resistance only at the outer surface. It would not remove gas from the annulus, restore the coating, repair a seal or compensate for glass breakage, and it would not preserve the optical aperture.

The commonly quoted 0.013 Pa is an order-of-magnitude target, not a pass/fail number

Peer-reviewed receiver studies commonly state that annulus pressure should be kept at or below the Knudsen gas-conduction range, typically below 10⁻² Pa, for long-life performance. Because 0.013 Pa is 1.3 × 10⁻² Pa, it is reasonably described as being of the order of 0.01 Pa. It should not be converted into a blanket procurement requirement without identifying the receiver design, temperature, gas species, test method, getter arrangement, measurement location and acceptance protocol.

Residual-gas analysis and plant work show that the issue is not total pressure alone. Hydrogen generated by thermal degradation of the organic heat-transfer fluid can permeate through the absorber into the annulus. Getters absorb hydrogen and other gases until capacity or kinetic limits are reached. Material outgassing, seal leakage and glass failure can also raise pressure; air ingress through seals or broken glass may dominate in some plants.

The performance consequence is substantial. Published work reports that a hydrogen-filled annulus can increase heat loss by roughly a factor of four relative to a good vacuum under the conditions studied. That factor is evidence of mechanism and severity, not a universal multiplier for every receiver or climate.

A white or silvered appearance is a screening observation, not a diagnostic result

A change in glass or coating appearance can justify infrared screening, a visual inspection for broken glass, and a review of bellows and seals. It cannot establish annulus pressure or gas composition. Glass fouling, anti-reflection coating, soiling, condensation, normal reflectivity and viewing angle can all change appearance. An intrinsically hotter glass envelope is also non-specific: vacuum loss, selective-coating degradation, soiling and changed external conditions can all raise glass temperature.

A study of a 50 MW plant combined receiver temperatures with plasma-based annulus-gas evaluation. It found vacuum degradation in 9 percent of the surveyed receivers, coating oxidation in a subset, and no hydrogen in the evaluated failed sample. The result directly contradicts any assumption that every failed tube is hydrogen-filled, or that one visual signature identifies the cause.

A defensible condition assessment records glass condition, coating appearance, soiling and mechanical damage separately; normalises infrared thermography for irradiance, wind, viewing angle and operating temperature; trends receiver performance against like-for-like circuits; uses qualified non-destructive methods appropriate to the receiver type; and confirms representative failures through residual-gas or plasma methods where the outcome changes the campaign decision.

The receiver standard covering technical requirements and test methods for solar receivers, including optical and thermal performance and durability, states heat-loss tests that are component tests and do not reproduce every installed-field condition. Procurement should therefore reference the receiver standard while adding project-specific field-diagnostic requirements.

4. Tower receivers: only the non-irradiated surfaces get insulated

The error of wrapping a trough receiver in fibre must not be inverted into the claim that all CSP receivers should remain bare. The correct treatment depends on which surface performs the optical function and which surface is an ordinary equipment boundary.

An external tubular tower receiver absorbs concentrated flux on its tubes. Its preferred surfaces must remain exposed to the heliostat field. Thermal protection is provided principally by material selection, tube geometry, flow distribution, coatings, circulation and control of peak flux. Insulation may be appropriate on rear shells, headers, downcomers, cavities, access platforms and adjacent non-irradiated equipment, but not on an irradiated absorber surface.

This distinction separates a tower receiver from a trough receiver. The tower tube is an exposed pressure and heat-transfer surface; the trough glass is part of a sealed annulus and optical system. A detail transferred from one to the other without examining the heat-flow path is likely to be wrong.

Ceramic fibre and high-alumina fibre systems are the usual candidates for rear shells and cavity backing inside their qualified temperature and atmosphere regime. The material family overview, including forms, densities and classification versus continuous-use limits, is on our ceramic fibre insulation hub.

5. Molten-salt piping: thickness comes from freeze survival, not from steady-state loss

A universal thickness table would conceal exactly the variables that control CSP performance. Thickness should emerge from coupled thermal, mechanical, operational and economic calculations.

The iteration begins with normal maximum, minimum and transient boundary temperatures, then assigns temperature-dependent thermal conductivity to each layer. Heat loss, surface temperature, freeze-protection hold time, adjacent-material temperatures, available space, structural load, jacket growth and access requirements are calculated next. The result is then tested against wind and rain, seismic restraints, maintenance removal, economic thickness and installed life.

For nitrate-salt lines the binding case is usually not the steady heat loss at full flow. It is the cooldown or stagnation case: how long the line can sit without circulation before salt reaches its freezing range, which for the common solar salt eutectic is around 220 °C. That single requirement can set the thickness, and it also sets whether trace heating is a backup or a primary function. The interaction between the two is covered in detail in our heat tracing and insulation system design guide.

Mineral wool and calcium silicate are the two workhorses for salt piping, and they are not interchangeable by temperature alone. Mineral wool is economical across the range but its compressive behaviour under support shoes, its water behaviour after a steam or rain event and its leachable-ion content all need product-specific data. Calcium silicate is rigid and dimensionally stable but brittle under impact and heavier per unit of thermal resistance. The range and limits of the mineral wool family are on our rock wool insulation hub, and the pipe form is on the rock wool pipe page.

Microporous and aerogel blankets have a legitimate role where available thickness or weight is the binding constraint: cramped skid packages, removable covers, and retrofits where the existing steel cannot carry more load. They are not a default. Cost per unit of thermal resistance is high, the attachment detail matters more than with a rigid section, and the hydrophobic behaviour of the composite has its own temperature ceiling. The material background is on our aerogel insulation hub and the product data on the aerogel blanket page.

6. Molten-salt tank shell and roof

Tank shells, roofs and pipes are ordinary but demanding external boundaries. External mineral wool, high-alumina fibre, calcium silicate, cellular glass, microporous or composite systems may all be valid candidates depending on temperature, mechanical exposure, moisture and access.

The design must accommodate support rings, nozzles, shell growth, platform penetrations, wind-driven rain and inspection requirements. Heat tracing must remain serviceable, and its effect on local interface temperatures must be included in the calculation.

Two details drive most of the failures on salt tank shells. The first is shell growth. A tank 14 m tall reaching 565 °C grows by roughly 100 mm over its height, and a rigid insulation system anchored to both the shell and a platform will either crack or tear its own anchors. The second is penetration density. A hot tank has nozzles for fill, drain, overflow, instrumentation, level and temperature measurement, and each one interrupts the insulation and the weather barrier at exactly the point where a leak would appear.

Where zone temperatures differ, the specification should differ too. The lower shell courses of a hot tank operate near salt temperature, the upper courses and the roof see a much lower temperature plus a vapour drive from below, and the roof in particular is exposed to rain and to standing water at the perimeter. A single material and thickness across the whole tank is usually either uneconomic or technically unsuitable.

7. The tank base is a load-bearing thermal transformer, not a thick insulation pad

The principal function of a hot-tank base is to transform roughly 565 °C salt conditions into temperatures acceptable to structural materials and the ground while carrying very large loads. Thermal resistance alone does not satisfy that function. This is the one place in a CSP plant where the insulation is, functionally, a structural element.

A common functional sequence, from salt downward, is:

1. tank-bottom plate, sometimes with an internal liner or protection system;
2. levelling, bedding or sand layer where required by the design;
3. refractory hot-face or load-spreading layer;
4. insulating firebrick or another high-temperature load-bearing insulation;
5. cellular glass or another lower-temperature insulating and bearing layer;
6. reinforced-concrete slab or pedestal;
7. ventilation or natural-convection cooling and soil-temperature protection, where required.

This is a functional order, not a universal bill of materials. Some designs use external calcium-silicate insulation. Some hot tanks use internal refractory and a metal liner. Others use stainless hot-tank bottoms with externally insulated foundations. Historical national-laboratory studies describe external calcium-silicate insulation, foundation air passages, and internal refractory protected by thin corrugated corrosion-resistant alloy sheet in alternative concepts. The existence of several workable arrangements is itself evidence against a universal layer stack.

The widely quoted 90 °C and 480 °C values belong to one specific numerical study

A published numerical study examined a 16,500 m³ tank with a 41 m diameter and 14 m height. Its basis was a 565 °C steady-state salt and tank-bottom condition. To avoid loss of concrete strength through dehydration, it required the insulation-foundation bottom at the concrete interface to remain below 90 °C. It also required the cellular-glass top to remain below 480 °C.

The calculated steady-state temperatures were approximately 67 °C over most of the concrete top, rising locally to about 81 °C near a material transition, and a maximum of 453 °C at the cellular-glass top. The authors attributed the local concrete-temperature variation to the higher conductivity of the refractory brick.

These values support three conclusions: a high-temperature refractory layer can create local lateral heat-flow effects; interface temperatures, not nominal salt temperature alone, govern layer compatibility; and edge and transition detailing can control the maximum local temperature.

They do not support a universal 90 °C concrete threshold. Concrete temperature limits depend on the mix, moisture condition, design code, exposure duration, stress state and dehydration assessment. Nor do they support a universal 480 °C cellular-glass limit. The project must use the purchased grade's hot-surface rating and the manufacturer's agreed application limits.

A separate experimental tank study cited a 482 °C cellular-glass maximum working temperature and a 90 °C foundation-bottom prohibition, while calculating 384 °C and 46 °C for its own smaller-scale arrangement. The similarity in criteria shows a recurring design pattern; the differing results confirm that thickness, geometry, materials and test scale remain decisive.

Product grade, not material family, carries the number

Cellular glass illustrates why. ASTM C552 covers cellular-glass compositions and physical properties and sets an inherent standard scope up to 427 °C, while requiring actual application limits to be agreed between manufacturer and purchaser. A separate manufacturer's high-load product data sheet states 1,600 kPa lot-average compressive strength for one grade. Neither document justifies assigning that strength, or a 480 °C capability, to an unnamed "foam glass" in a bid.

The same reasoning applies to mineral wool. One manufacturer's low-leachable-ion product does not qualify all mineral wool. One fibre grade's classification temperature does not establish another grade's blowing or shrinkage resistance. Generic terminology identifies a material family; it is not a design allowables package.

Mineral wool is generally unsuitable for the tank-bottom load path, but it is not universally prohibited

The primary objection is mechanical rather than thermal. Loose or resilient batts are not a predictable, maintenance-free structural medium for concentrated vertical loads, differential settlement, construction traffic and long-duration compression. A tank base also requires a controlled load path, predictable settlement and resistance to crushing.

A composite design is not logically impossible if an independent structural system carries the load and mineral wool is used only as non-bearing thermal fill. Even then, moisture control, settlement, access, salt-leak consequences and replacement strategy must be demonstrated. "Independent structure" must mean a verified load path, not an unsupported assertion in a data sheet.

8. Steam generators, reheaters and steam lines: the most conventional zone, with one catch

Steam generators form the CSP to power-block interface, and they are the most familiar part of the plant for anyone coming from conventional generation. That familiarity is the risk. Hot-salt inlet regions can approach salt-side design temperatures, while feedwater, evaporator, superheater, reheater and casing zones operate at different temperatures. A single material and thickness across the whole vessel is therefore usually either uneconomic or technically unsuitable.

The catch is that a salt-heated steam generator shuts down and starts up with the solar field, not with the grid demand. Zone-by-zone salt and steam temperatures, warm and cold start-up rates, and salt freeze protection in the tubes all have to be handled by the same insulation system that provides personnel protection on the outside.

The same logic applies to main and reheat steam lines, where the design drivers are hanger loads, access and long-run jacket expansion rather than any CSP-specific chemistry. This is the zone where conventional practice transfers most successfully, and where a copy of a fossil-plant schedule is least likely to cause harm.

For rigid high-temperature pipe and equipment insulation, calcium silicate is the standard answer, and the material overview including type, shrinkage, strength and leachable-ion checks is on our calcium silicate insulation hub. The board form used for vessel and equipment cladding is on the calcium silicate insulation board page.

9. Chloride: why low conductivity is the wrong first question

Chloride control is separate from molten-nitrate contamination, and it is the more common cause of damage in service. Water-leachable chloride, fluoride, silicate and sodium ions migrate with moisture, concentrate at the metal surface as water evaporates, and attack austenitic stainless steel from the outside. Ammonium-content and pH requirements in the relevant product standards exist for the same reason.

These tests do not prove that a material is compatible with leaked solar salt. A product can pass stainless-steel stress-corrosion qualification and still be unsuitable after salt immersion, because of porosity, salt uptake, loss of strength or cleaning limitations.

The practical specification consequence is that for any insulation in contact with or draining toward austenitic stainless steel, the purchase order should ask for preproduction stress-corrosion qualification for the supplied product family, leachable-ion results on an agreed lot basis, and a written statement of the extraction method used. Asking for "low chloride" without a method and a basis is not a requirement, it is a wish.

10. Nitrate salt compatibility: three rules and a list of prohibitions

A leak path must be designed independently of the insulation, and the insulation must never be presented as secondary containment. Nitrate salts are strong oxidisers: they do not burn themselves, but they support the combustion of organic material, and they wet metal surfaces and freeze near 220 °C. Porous insulation may intercept, absorb or transport salt, but it cannot be assumed to retain it, and contamination can change both thermal performance and mechanical behaviour.

Experimental work on leaking salt in porous foundation material shows rapid migration before solidification, with temperature, porosity and leak geometry affecting penetration. The same work recorded an increase in the foundation's equilibrium temperature after leakage. That demonstrates that leaked salt changes both contamination state and heat transfer; it does not justify relying on the insulation as a barrier.

A robust leak strategy includes leak detection at likely sources, defined collection and drainage routes, thermal and process isolation, personnel and environmental precautions for oxidiser-containing nitrate residues, a pre-planned selective or full removal of contaminated insulation, inspection of the substrate and attachments, and documented waste acceptance with the treating or disposal facility.

Salt-contaminated insulation should not be returned to service merely because it can be dried

Heating may re-mobilise salt, redistribute contamination, increase thermal conductivity, corrode adjacent metal, and expose personnel to dust and nitrate or nitrite residues. Unless the owner's documented procedure and the receiving facility approve an alternative, remove the affected material and replace it with qualified material.

The following also require individual assessment: inserts, vapour retarders and jacket sections in the affected zone; fasteners, pins and support attachments; underlying carbon-steel or stainless-steel surfaces; drainability and hidden cavities; and the extent of penetration into porous materials.

11. Daily cycling is the governing load case

CSP cycling loads the insulation assembly as well as the pressure boundary, but the two fatigue mechanisms must not be conflated. Receiver-tube creep-fatigue calculations do not demonstrate that fibre, blankets, blocks or jacketing will survive. Insulation-system fatigue requires separate detailing, testing and inspection.

For nitrate-salt towers, the daily boundary often moves between approximately 290 °C cold-salt conditions and about 565 °C hot-salt conditions. Demonstration plants showed this operating range and could undergo rapid temperature changes during cloud passages. Experimental corrosion studies commonly approximate the operating cycle as 12 hours of daytime heating and 12 hours of night-time cooling, with separate thermal-shock tests imposed for rapid weather events. These are representative cycles, not a universal plant schedule.

During heat-up, the pipe, vessel or tank grows relative to the insulation and jacket. During cool-down, the movement reverses. Repeated cycles produce a fairly predictable set of failure modes.

Rain, wash-down water or humid air enters through any of the openings above. Subsequent heating drives off moisture, concentrates dissolved ions and creates a wet-dry corrosion cycle at the metal surface. This is an insulation-system failure mechanism, not the same calculation as creep-fatigue of a receiver tube.

Published structural assessments describe through-thickness, circumferential and axial thermal gradients caused by non-uniform flux and axial fluid-temperature variation. Daily cycles and cloud events load the tubes, while prolonged high-temperature holds introduce creep. One annual simulation found start-ups dominant for fatigue while long high-temperature holds dominated total damage. Those findings support transient operating cases and component-specific damage models. They do not justify treating insulation as structurally benign. A multilayer system with dissimilar coefficients of expansion, frictional slip and external wind loading can accumulate local damage even when the pressure boundary remains within its allowable fatigue usage.

The detailing consequence is that movement should be made repeatable rather than restrained. Useful measures include specifying cyclic movement at each support and termination; selecting segmented or compliant details where differential growth is high; avoiding hard bearing of rigid insulation against moving nozzles; providing expansion captures and slip planes in jacketing; sealing penetrations without bonding moving parts together; and draining all low points. Vertical towers also require wind-vibration, dead-load and rain-shedding checks. The weather-barrier side of this is covered in our cladding and jacketing guide.

The acceptance target is not "no movement." It is controlled movement without progressive opening, crushing, pumping or loss of weather protection.

12. The standards landscape: a real gap, stated precisely

The standards position on CSP insulation is more nuanced than the version usually repeated, and getting it right matters more than getting it dramatic. The official published-parts list for the solar thermal electric plants series is broad: it covers terminology, solar-resource datasets, plant performance testing, heat-transfer fluids, thermal-energy-storage characterisation, parabolic-trough plant design and collector testing, trough receivers, reflector measurements, tower-plant design, heliostat-field control and linear Fresnel collectors.

The important qualification is this. A dedicated project for thermal insulation of solar thermal electric plants has been identifiable in the technical committee's project list. As of the published-parts list reviewed for this article, no dedicated insulation part had been verified as published. So the accurate statement is narrow: as of the list reviewed, a dedicated CSP insulation project existed, but no dedicated insulation part had been verified as published.

Saying "there is no CSP insulation standard" overstates the case and is easy to disprove. Saying "the series does not contain a published insulation part, although a project exists" is defensible, and it has a practical consequence: until such a document is contractually applicable, the specification has to be assembled from components.

No generic standard should be represented as containing CSP-specific requirements that it does not state. The comparison between the European and American mineral-wool pipe standards, including what each declares and where they diverge, is set out in our EN 14303 versus ASTM C547 comparison.

Where a future insulation document changes the agreed design basis, the contract should establish whether the project follows the edition current at bid, the edition current at a defined design-freeze date, or a later edition only by change order.

13. Selection matrix: the constraint that governs, by location

No material family is universally suitable for CSP insulation. Candidate screening should first eliminate products whose chemistry, mechanical behaviour or geometry conflicts with the boundary; only the remaining products should enter thermal and economic optimisation.

The single most useful discipline in this table is the third column. Most expensive CSP insulation mistakes are not material mistakes; they are boundary mistakes, where a detail that was correct in one location was transferred to another with the same nominal temperature and a completely different mechanical or chemical duty.

14. Procurement and inspection: making the specification testable

The purchase order should define evidence, boundaries and responsibilities, not merely a product name. A generic description such as "1,260 °C ceramic fibre" or "foam glass" leaves the principal design questions unanswered.

Four purchase-order clauses do most of the work, and they are worth quoting in substance rather than paraphrase:

• Design boundary. The supplier states product and system limits for temperature, compression, moisture, chemical exposure and attachment, and a catalogue classification temperature may not be substituted for a demonstrated continuous-use limit at the specified hot-face and interface temperatures.
• Stainless-steel compatibility. For insulation in contact with or draining toward austenitic stainless steel, the supplier provides preproduction stress-corrosion qualification for the supplied product family and leachable-ion results on an agreed lot basis, complying with the project's acceptance requirements.
• Mechanical duty. The supplier provides compressive or bearing deformation data at the project's stated temperature, load duration and moisture condition, and products whose strength is demonstrated only at ambient temperature are not used as tank-bottom load-bearing layers.
• Salt release. Insulation is not represented as secondary containment, and the supplier identifies salt-penetration, cleanability, removable-section and disposal limitations.

Inspection has to address the installed system, not only the delivered material. A material can conform to its data sheet while the installed system still fails. Certificate review can confirm declared product identity, nominal density, dimensions, test-report traceability and conformity with the project specification. Random checks can identify gross transport damage, incorrect product, wrong facing and clearly nonconforming thickness.

A third party normally cannot create project acceptance by a single site observation. It cannot verify an unstated load history, reproduce years of cycling, establish salt compatibility from a visual check, or measure the steady-state conductivity of an installed multilayer system without an appropriately designed test. Where those issues are critical, the purchase order must call for witnessed factory tests, sample retention and, where necessary, a field mock-up.

During operation, the useful checks are differentials after cloud passages and rapid load changes rather than steady operating values alone; local jacket hot spots, vapour plumes, salt staining, abnormal drip points and new vibrations; removable covers clearing valve stems, actuators and drain points; and platform, ladder and support penetrations that remain both sealed and drained. At overhaul, the priority is opening representative high-risk locations rather than judging by exterior appearance, then sampling suspect insulation for moisture and leachable ions and investigating salt staining as potential leak evidence.

The document-verification side, including how to check a certification body's scope rather than just its existence and what each certificate actually proves, is covered in our industrial insulation supplier selection guide.

15. Conclusion: move temperature rating out of the first column

A robust design closes the loop from operation back to procurement. Thermal performance, insulation condition and leak response must all be tied back to the original design basis.

A practical design-closure package contains a tagged boundary drawing identifying hot-face, cold-face and intermediate interface temperatures; material schedules with declared conductivity curves and density ranges; heat-loss and surface-temperature cases for maximum, minimum and transient operation; support, anchor and jacket movement details; leak detection, drainage and removal details; coating and corrosion-under-insulation or stress-corrosion assessment for carbon and stainless steels; procurement traceability and acceptance test records; and inspection and repair manuals.

The final engineering test is falsifiability. The specification should state which temperature, moisture, movement, heat-loss or contamination observation would require investigation. If no such criterion exists, the design is not being actively managed.

The three governing conclusions are:

• do not transplant a steady fossil-station insulation schedule without demonstrating cyclic movement and moisture recovery;
• do not select on temperature rating alone;
• do not specify one material as the answer to every tank, pipe, valve and receiver boundary.

Reliable CSP insulation is achieved by assigning each boundary its correct function, qualifying the complete layer system, and retaining the ability to inspect, remove and replace it over the life of the plant. On a solar plant, the insulation is not a consumable wrapped around the design. It is part of the design, and it is the part that decides whether the other parts reach their intended life. Where the whole-life cost of that decision sits, and how the material families compare once installed performance and replacement are included, is set out in our insulation material cost comparison.

If a project is at the point of specifying CSP insulation, the CSP application page is the right starting point for the system-level questions, and our engineering team can work through the boundary-by-boundary detail against a specific salt chemistry, tank design and operating profile.