Reference service life and replacement schedule: B4 cost in LCC
Service life is the parameter that moves an LCC result most after the discount rate. Where RSL comes from, the replacement formula N = ceil(T/SL) − 1, residual value, the B2/B3/B4 distinction and how long-lived products get unfairly penalised.
By clca Editorial TeamLast updated
Reference service life (RSL) is the period a product or component is expected to perform its function under defined conditions of use; in life cycle costing (LCC) it determines how many times the component is replaced within the analysis period and how much value it leaves at the end. The rule comes down to two formulas: number of replacements N = ceil(T / SL) − 1, and residual value = last investment × (remaining life / service life). Unless both mechanisms are set up correctly, a 25-year and a 40-year product can never be compared fairly. This article covers where service life comes from, how the replacement schedule is derived, how module B4 differs from maintenance and repair, and the mistakes that penalise long-lived products.
What is reference service life and where does it come from?
The ISO 15686 series defines service life through two concepts: reference service life (RSL) is the life declared by the manufacturer or a database for specific reference conditions; estimated service life (ESL) is that value adapted to the actual project conditions. The factor method of ISO 15686-8 makes that adaptation with seven multipliers: component quality, design level, workmanship, indoor environment, outdoor environment, conditions of use and maintenance level. A façade cladding on the coast has a shorter life than the same cladding inland, and the factor method quantifies the difference. EN 15804 requires RSL information in the EPD whenever use stage modules (B1-B7) are declared; a verified EPD is therefore the most reliable source of service life for LCC as well — same product, same life, two axes.
Number of replacements: N = ceil(T / SL) − 1
If the analysis period T is longer than the service life SL, the component is replaced within the period. The number of replacements is derived with the ceiling function: N = ceil(T / SL) − 1. The initial installation is not a replacement, hence the minus one; the ceiling function assumes the replacement happens the moment the life is used up. Each replacement is discounted at the end of the year it occurs in (k × SL). For a 50-year analysis period:
- SL = 20 years → N = ceil(2.5) − 1 = 2 replacements (in years 20 and 40)
- SL = 25 years → N = ceil(2.0) − 1 = 1 replacement (in year 25)
- SL = 30 years → N = ceil(1.67) − 1 = 1 replacement (in year 30)
- SL = 50 years → N = ceil(1.0) − 1 = 0 replacements
- SL = 60 years → N = ceil(0.83) − 1 = 0 replacements, 10 years of life remain at the end
Some methods use fractional replacements (T / SL − 1) instead of whole numbers; that is a shorthand which implicitly folds in the residual value. ISO 15686-5 permits both approaches but requires a single method across alternatives within one study. Whole-number replacements plus an explicit residual value are more transparent, because the report shows in which year how much money goes out.
Residual value: linear depreciation
If the component still has life left at the end of the period, the remaining life is converted to a value, discounted as an income in the final year and subtracted from NPC. The standard approach is linear depreciation: residual value = last investment cost × (remaining life / service life). In the examples above, for SL = 30 the component installed in year 30 has been used for 20 years by year 50, 10 remain and the residual value is one third of the investment; for SL = 60 there is no replacement, 10 years remain and the residual is one sixth; for SL = 20 the component installed in year 40 is halfway through its life in year 50 and the residual is half. A residual value discounted over 50 years shrinks — to about 0.23 of its size at 3% — but it is not zero, and it can change the ranking between long-lived alternatives.
Biases for and against the long-lived product
Three biases are common in service life calculations. The first is skipping the residual value: a 60-year product modelled over a 50-year period with no residual is treated as if one sixth of its life were given away for free, and it looks unfairly expensive. The second is a high discount rate: at something like 8%, a replacement in year 25 carries only about 15% of the weight of the investment, which favours the short-lived product. The third is an optimistic life declaration: a manufacturer's 50 years for laboratory conditions, used without adaptation to project conditions through the factor method, produces an incomplete replacement schedule. When the three act together the result measures modelling choices rather than product quality; that is why service life is always run through a ±20% sensitivity and the threshold life at which the ranking changes is reported.
B4 replacement versus B2 maintenance and B3 repair
EN 15804 and EN 16627 split the use stage into the same boxes. B2 maintenance is planned, recurring work to preserve the component's life: painting, cleaning, filter changes. B3 repair is unplanned, partial intervention: replacing a broken pane, renewing a leaking seal. B4 replacement is the removal of the component as a whole and installation of a new one, and its cost is not just the price of the new product: removal labour, waste management of the old component, transport and installation of the new one are all booked to B4 — just as environmental B4 includes the A1-A3, A4 and A5 burdens of the new product. B5 refurbishment is planned intervention at building level covering several components. These boundaries are not arbitrary: booking the same cost to two modules double-counts, booking it to none under-counts, and the row-by-row comparison of EPD and LCC breaks.
Where does service life data come from?
- A verified EPD — if B modules are declared, RSL and reference conditions are written in the document; the most reliable source
- Manufacturer technical declaration and warranty period — a warranty is not a life, it is read as a lower bound
- National service life tables and component databases — typical values; adapted to project conditions with the factor method
- Accelerated ageing tests and field observation under ISO 15686-2 — costly but project-specific
- Maintenance history — real replacement intervals for an existing building, the best calibration data
Frequently asked questions
- What is the difference between reference service life and estimated service life?
- Reference service life (RSL) is the life declared by the manufacturer or a database for specific reference conditions. Estimated service life (ESL) is RSL adapted to the actual project conditions through the ISO 15686-8 factor method — multipliers for material quality, design, workmanship, indoor and outdoor environment, use and maintenance. In LCC the replacement schedule is built with ESL.
- How is the number of replacements calculated?
- With the formula N = ceil(T / SL) − 1, where T is the analysis period and SL the service life. For a 20-year component over a 50-year period, ceil(2.5) − 1 = 2 replacements are scheduled (in years 20 and 40). The initial installation does not count as a replacement. Each replacement is discounted at the end of the year it occurs in and booked to module B4.
- Why does residual value matter?
- Because without it long-lived products are systematically penalised. Life unused at the end of the analysis period is converted to value by linear depreciation, discounted as an income in the final year and subtracted from NPC. Discounted over 50 years it shrinks but does not vanish; between two long-lived alternatives of similar cost it can decide the ranking.
- What is included in the B4 replacement cost?
- Not just the price of the new product: removal of the old component, waste management, transport of the new product and installation labour are booked to B4 too. This mirrors environmental B4, which covers the A1-A3, A4 and A5 burdens of the new product. Planned maintenance goes to B2, unplanned partial repair to B3 and building-level refurbishment to B5.
- Can the warranty period be used as the service life?
- No; a warranty is a commercial commitment and usually falls well short of the technical life. The warranty period is at most a lower bound. For service life, the RSL in a verified EPD, the manufacturer's technical declaration or national service life tables are used, and the value is adapted to project conditions with the factor method.
Tags
- reference service life
- replacement schedule
- life cycle costing
- residual value
- B4 replacement
- ISO 15686-5
- EN 16627
Related reading
All articles in Data & Methods- Data & Methods
How to calculate LCC: life cycle costing step by step
Life cycle costing (LCC) in seven steps: analysis period and functional unit, assigning items to EN 15804 modules, the real-versus-nominal decision, discounting and NPC, EAC, a worked example and sensitivity.
11 min read - Data & Methods
Discount rate in LCC: real or nominal, and which rate?
The discount rate is the single assumption that moves an LCC result most. The real-versus-nominal distinction, the Fisher relation, typical ranges for public and private projects, the advantage of real terms under high inflation and a sensitivity example.
8 min read - Data & Methods
Energy recovery or material recycling? Comparing the credits
The same tonne of plastic displaces electricity if burned and granulate if recycled — but the two credits are not the same size and cannot both be granted to one stream. The waste hierarchy, EN 15804 Table 13 and the formulae.
11 min read - Data & Methods
Transport in LCA (A4): tkm, load factor and empty return
A4 is the transport of the product from factory gate to site, calculated as mass × distance × a per-tkm factor. Vehicle class, utilisation, empty return and the trap of volume-limited products; the EN 15804 scenario table and a worked example.
11 min read