Extended field of application (EXAP) for reaction-to-fire Euro-classification of optical fibre cables

Transcription

1 SP Technical Research Institute of Sweden Extended field of application (EXAP) for reaction-to-fire Euro-ification of optical fibre cables Richard Johansson, Johan Post, Michael Försth SP Report 2015:32

2 Extended field of application (EXAP) for reaction-to-fire Euro-ification of optical fibre cables Richard Johansson, Johan Post, Michael Försth

3 3 Abstract Extended field of application (EXAP) for reaction-to-fire Euro-ification of optical fibre cables Reaction-to-fire tests for 66 optical fibre cables from different cable families were analysed with the purpose of proposing rules for extended field of application (EXAP). Four different candidates for EXAP rules were tested. These candidates were all based on the already existing EXAP rules for power cables, CLC/TS 50576, which was proposed in the CEMAC II project. It was found that the highest confidence was achieved by using EXAP rules where safety margin are added to the tested cables, which is identical to the procedure in CLC/TS 50576, and where the number descriptor n in the expression for the cable parameter χ is given by the number of so called units. The number of units replaces the number of conductors used in the expression for χ in the EXAP for power cables. A unit is defined as a combustible tubular item which may contain one or several fibres, excluding the fibres themselves or the outer jacket of the cable. The proposed EXAP rules, where safety margins are used and where χ is described in terms of number of units, gives a high confidence. Only 2% of the possible EXAP applications within the set of cables gives an erroneous non-conservative indication of Euro (B2ca, Cca, Dca, or Eca). Key words: EXAP, CPR, reaction to fire, optical fibre cables SP Sveriges Tekniska Forskningsinstitut SP Technical Research Institute of Sweden SP Report 2015:32 ISBN ISSN Borås 2015

4 4 Contents Abstract 3 Contents 4 Preface 5 1 Introduction 6 2 Background Reaction-to-fire ification of cables EXAP for power cables: The concepts of safety margin sm and cable parameter 10 3 Cable test data available for this study 15 4 Method 16 5 Results with n = number of optical fibres with n = number of units Summary of results and comparison with results from CEMAC II project (power cables) 23 6 Conclusions 24 7 References 25 8 Appendix: Proposal for EXAP-rules for optical fibre cables Definition of a product family for EXAP EXAP with safety margin Flaming droplets/particles 27 9 Appendix: Cable data Appendix: Detailed analysis χ calculated based on number of optical fibers, without using safety margin sm χ calculated based on number of optical fibers, using safety margin sm χ calculated based on number of units, without using safety margin sm χ calculated based on number of units, using safety margin sm 46

5 5 Preface This project was ordered by Europacable who also selected the cables and supplied all cable data and test data included in the analysis. All analyses were performed by SP.

6 6 1 Introduction The Construction Products Directive, CPD (89/106/EEC - CPD) [1], came into force in 1988 and has basically been aiming at creating a common market for construction products by producing documents on a European level that can be used for a common declaration of the properties of the products, verified through the CE-mark. A CE-mark is valid in more than 30 countries [2]. The Construction Products Regulation, CPR, (305/2011/EU - CPR) [3] replaced the CPD in July 1, The CPR intends to further clarify the concepts and the use of CE marking. Some procedures are simplified in order to reduce costs, in particular for small and medium sized enterprises (SMEs). The CPR is also increasing the credibility and reliability of the system by imposing stricter designation criteria to bodies involved in the assessment and the verification of construction products. A very important change is that CE-marking now becomes mandatory. This is expected to speed up the use of the CE-mark. Another implication may be that national voluntary marking systems become less important. In order to perform CE-marking, a so called harmonized product standard is needed in addition to the test and ification standards mentioned above. There are more than 400 harmonized EN standards which are cited in the Official Journal, the European official newspaper. The product standard describing construction of cable families is termed EN One plausible scenario is that this product standard becomes published in the Official Journal during fall This would mean that CE marking can start by fall 2015 and will be obligatory by fall 2016 (assuming the minimum default one year transition time). The number of cables to be tested described by the product standard is very large and testing of each cable would be excessively costly. Therefore, so called extended application of test results, EXAP, should be made available. An EXAP allows a family of products to be ified to a certain reaction to fire without testing all of the individual members of the family. The CEMAC II project [4, 5] included extensive testing of cables on the market. Using these data as the technical basis EXAP procedures could be developed for power cables. These rules where published in September 2014 as a formal document termed CLC/TS [6]. However, no EXAP procedures for optical cables or copper communication cables were developed as a result of the CEMAC II project. This report aims to analyse new test data for optical fibre cables and to propose EXAP procedures for such cables. The approach taken is to follow the procedures for power cables to the largest extent possible in order to achieve consistency within the regulatory framework. The most important requirement is however to deliver a technically sound proposal.

7 7 2 Background In this section a short background is given on the reaction-to-fire ification of cables within the CPR. Section 2.1 outlines the ification system. A more complete description can be found in reference [2]. Section 2.2 explain some specific features regarding the existing EXAP for power cables. These features, the safety margin sm and the cable parameter, will be important in the analysis and argumentation in this report. More information can be found in the CEMAC II report [4], from where the content of Section 2.2 has been abridged. 2.1 Reaction-to-fire ification of cables Interpretation of test results into Euroes is described in the new revision of the ification standard EN Details are found in a European Commission Decision [7]. This revision contain a new section, called EN [8], which describe ification of cables, in addition to the already existing ification of linings, floorings, and pipe insulation. The ification is based on heat release and flame spread, smoke production, burning droplets, and acidity. EN describes seven heat release and flame spread es of cables which are called A ca, B1 ca, B2 ca, C ca, D ca, E ca and F ca. The performances of the different heat release and flame spread es can approximately be described as: A ca. Level of highest performance corresponding to products that practically cannot burn, i.e. ceramic products. B1 ca. Products that are combustible but show no or very little burning when exposed to both the reference scenario experiments and the ification test procedure EN (30 kw flame source). B2 ca and C ca. Products that do not give a continuous flame spread when exposed to the kw ignition source in the horizontal reference scenario, that do not give a continuous flame spread, show a limited fire growth rate and limited heat release rate when tested according to EN (20,5 kw flame source). D ca. Products that show a fire performance approximately like that of wood when tested in the reference scenarios. When tested according to EN (20,5 kw flame source) the products show a continuous flame spread, a moderate fire growth rate, and a moderate heat release rate. E ca. Products where a small flame attack is not causing large flame spread. es B1 ca, B2 ca, C ca, D ca, and E ca are based on tests according to EN [9] but also require results from the small scale ignitability and flame spread test EN [10]. A ca requires testing according to EN ISO 1716 [11], however this is a rare. production is ified in the smoke es s1a, s1b, s1, s2, and s3. Testing according to standard EN [12] is required if compliance with the best smoke es, s1a and s1b, is sought. The other smoke es are based on results from the EN test. Burning droplets are ified into es d0, d1, and d2. These es are based on results from the EN test.

8 8 Acidity is ified in the acidity es a1, a2, and a3. Testing according to standard EN [13] is required for ification of acidity. The ification system from EN is summarized in Table 1 below. The EXAP document CLC/TS [6], which is only applicable to power cables, is only valid for es B2 ca, C ca, D ca, smoke es s1, s2, and s3, and droplet es d0, d1, and d2. This is also the scope for the EXAP proposal for optical fibre cables to be drafted in this report.

9 9 Table 1. es of reaction to fire performance for electric cables [8]. Test method(s) ification criteria Additional ification A ca EN ISO 1716 PCS 2,0 MJ/kg ( 1 ) B1 ca EN (30 kw flame source) and FS 1.75 m and THR 1200s 10 MJ and Peak HRR 20 kw and FIGRA 120 Ws -1 production ( 2,5 ) and Flaming droplets/particles ( 3 ) and Acidity ( 4, 7 ) B2 ca C ca EN H 425 mm EN (20,5 kw flame FS 1.5 m and source) THR 1200s 15 MJ and and Peak HRR 30 kw and FIGRA 150 Ws -1 EN H 425 mm EN (20,5 kw flame FS 2.0 m and source) THR 1200s 30 MJ and Peak HRR 60 kw; and FIGRA 300 Ws -1 production ( 2,5 ) and Flaming droplets/particles ( 3 ) and Acidity ( 4, 7 ) production (2,6) and Flaming droplets/particles (3) and Acidity (4, 7) D ca EN EN (20,5 kw flame source) and EN H 425 mm THR 1200s 70 MJ; and Peak HRR 400 kw; and FIGRA 1300 Ws -1 H 425 mm E ca EN H 425 mm production (2,6) and Flaming droplets/particles (3) and Acidity (4, 7) F ca No performance determined (1) For the product as a whole, excluding metallic materials, and for any external component (i.e. sheath) of the product. (2) s1 = TSP m2 and Peak SPR 0.25 m2/s s1a = s1 and transmittance in accordance with EN % s1b = s1 and transmittance in accordance with EN % < 80% s2 = TSP m2 and Peak SPR 1.5 m2/s s3 = not s1 or s2 (3) d0 = No flaming droplets/particles within 1200 s; d1 = No flaming droplets/ particles persisting longer than 10 s within 1200 s; d2 = not d0 or d1. (4) EN : a1 = conductivity < 2.5 µs/mm and ph > 4,3; a2 = conductivity < 10 µs/mm and ph > 4.3; a3 = not a1 or a2. No declaration = No Performance Determined. (5) The smoke declared for B1 ca cables must originate from the test according to EN (30 kw flame source) (6) The smoke declared for B2 ca, C ca, D ca cables must originate from the test according to EN (20,5 kw flame source) (7) Measuring the hazardous properties of gases developed in the event of fire, which compromise the

10 THR [MJ] EXAP for power cables: The concepts of safety margin sm and cable parameter In the CEMAC II project it was found that cables have a more complex fire behaviour than many other products that are more homogeneous products, such as mineral wool for example [14]. This is illustrated in the theoretical example in Figure 1 where the general trend is that THR decreases with increasing diameter, d, but where the fourth cable makes a sudden jump and breaks the monotonically 1 decreasing trend. It is clear that if the second and fifth cable would be tested and ification for all intermediate diameters would be based only on the worst tested result, i.e. the result for the second cable, ification would be too generous since the fourth cable, which belongs to D ca according to its THR value, would actually be ified as C ca according the EXAP outer diameter [mm] Figure 1 THR as a function of outer diameter. Theoretical example. For this reason a safety margin needs to be added to the worst result for the two tested cables. The magnitude of the safety margin will depend on how large the deviations from monotonicity are. This is described by max Equation 1 sm where max is the value used for ification according to respective ification parameter (peak HRR, THR, FIGRA, FS, peak SPR, and TSP), is the maximum, that is worst, test result of the tests that forms the basis of the EXAP, and 1 A monotonic function is a function that is always increasing or always decreasing. Constant plateaus are also allowed. In other words the slope does not change sign.

11 THR [MJ] 11 sm is the safety margin required for the particular ification parameter. Taking Figure 1 as an example the deviation from monotonicity occurs between the third and the fourth cable. THR for the third cable is 28 MJ and THR for the fourth cable is 31 MJ. The required safety margin in this example, sm, would therefore be 3 MJ. With such a safety margin the EXAP would never, for this particular cable type, allow a too generous ification of any non-tested cable included in the EXAP. Still, some cables exhibit a behaviour that would require such large safety margins that every application of the EXAP rules would result in Eca, see Figure 2 for example outer diameter [mm] Figure 2 THR as a function of outer diameter for cable group seven in CEMAC. The non-monotonic behaviour shows that the fire behaviour has little or non correlation with the outer diameter. This non monotonic behaviour remains also with other fundamental cable parameters as x-axis [4]. In order to obtain a smoother graph it is necessary to shift the outlier to one edge of the data set. It has been found that this can be successfully done by introducing the following parameter: c V 2 combust Equation 2 d with d [m] Outer diameter. V combust [m 2 ] Non-metallic 2 volume per meter ladder. c [ ] Number of conductors in one cable. 2 In the CEMAC II project V combust was defined as the non-metallic volume per meter ladder. For optical fiber cables, where typically no metals exist, V combust will instead be defined as the volume of combustible material per meter ladder, see Equation 9 and Section 8.2.

12 THR [MJ] 12 Using on the x-axis the graph transforms into Figure 3. The outlier is no longer an outlier since it is found on the right edge of the data set. Therefore it will never be an intermediate and non-tested cable in an EXAP. For any EXAP where this cable is included it will be one of the tested boundary cables upon which the EXAP is based. The high THR will therefore be reflected in max in Equation 1. Using Equation 1 and Equation 2 it was found that safety margins according to Table 2 gave a very high confidence in the EXAP rules for power cables [4]. These correspond to 10% of the limit for es B2, C and D, and to 20% of the limit for the smoke es s1 and s2, see also Table 1. Table 2 Safety margins v sm. B2ca Cca Dca s1 s2 Peak HRR [kw] THR [MJ] FIGRA [Ws -1 ] Flame spread [m] Peak SPR [m 2 s -1 ] TSP [m 2 ] Figure 3 THR as a function of for cable group seven. A phenomenological explanation to why can describe THR is given below. The quotient c/d 2 relates to the density of conductors in a cross section of the cable. When the flame hits a cable with a high conductor density the conductors can separate and air be entrained into the cable. This increases the oxygen supply, and thereby the intensity, of the combustion. Once the conductors have separated they can be viewed as separate cables with smaller diameter than the original cable. This speeds up the heating and therefore also the flame propagation along the cable. Multiplying the conductor density

13 13 c/d 2 by the amount of combustible volume of the ladder gives an estimate of how much material is combusted in total, which is an estimation of THR. Another contributing factor to increased flammability for cables with a high value of is that, for a given diameter, the ratio of insulation material to sheathing material increases with increased number of conductors, that is with increased. The insulation typically consists of a rather flammable material such as polyethylene while the protective sheathing consists of a more flame retardant material. Therefore, when the relative amount of insulation material increases the flammability of the cable also increases. Using as x-axis also gives a reasonably monotonic behaviour for most other ification parameters and cable types [4]. A plausible explanation of this is given below. From the FIPEC project, reference [15] p 150, it was concluded that for a majority of cables the most severe test is obtained by spacing the cables with a distance in the order of magnitude of their diameter. The mounting procedure suggested by the FIPEC project has been implemented in standard EN [9] and these procedures were used in the large scale tests performed within the CEMAC project. Since, typically, the cables are distributed over a width of 300 mm on the ladder and since the spacing between cables is d the following relation applies: Nd ( N 1) d 300 mm Equation 3 where Nd is the total width of the cables on the ladder and (N-1)d is the total width of the void spacing. Approximating N-1 by N gives: N 150mm Equation 4 d The combustible volume per meter cable, v combust, has been found to be approximately proportional to the cross section of the cable, that is to d 2 : 2 v combust ~ d Equation 5 The amount of combustible volume per meter ladder is therefore: 2 Vcombust Nvcombust ~ Nd Equation 6 and, from Equation 4: 150 V combust ~ Nd 2 2 d ~ d Equation 7 d

14 14 Inserting Equation 7 in Equation 2 gives the approximation: c c c Vcombust ~ d Equation d d d The approximate Equation 8 in essence explains why describes fire performance of different cable types in general. It is well known that combustion is more intense for cables with small diameter than for cables with large diameter. In other words combustion is more intense for large than for small. This is easily understood by making an analogy to matches and timber logs where the former is much easier to ignite. The exception is cable types which are completely combusted, such as Group 9 in the CEMAC project. In this case the relation is the opposite but still describes the fire performance in a fairly monotonic way, although with a different sign of the derivative. Furthermore the flame spread is, in general, facilitated if the number of conductors, c, is increased for a given diameter. The explanation of this is manifold but in essence more conductors mean a more porous cable in which the conductors more easy separate and where chimney effects is facilitated. A cable in which the conductors separate can be seen as several cables with smaller diameter, and therefore with more intense combustion according to the discussion above. As already mentioned above another reason for the increased flammability for cables with many conductors is that, for a given diameter, the ratio of flammable insulation material, typically polyethylene, to sheathing material increases with increased number of conductors.

15 15 3 Cable test data available for this study Cable selection was entirely performed by Europacable. Cable testing was performed by laboratories belonging to Europacable members or by research laboratories. The laboratories performing the tests were accredited according to ISO17025, except one out of five laboratories. All laboratories participated in the CENELEC round robin which was undertaken in the initial phase of the CEMAC II project. SP takes no responsibility regarding the representativity of the selected cables, nor for the quality of the performed tests. A detailed description of all families are presented in the appendix in Section 9. A summary of the cable families are given in Table 3. Table 3 Summary of cables included in the analysis of this report. Family Number of Diameter Units 1 Total account of fibers cables (min-max) [mm] The concept of unit is described in the beginning of Section 4.

16 16 4 Method Optical fibre cables cannot be characterized by the number of electrical conductors as defined in Equation 2 [4, 6]. Instead, the number of conductors must be replaced by another descriptor. From a functional point of view this descriptor could be the number of optical fibres. However, the fibres themselves are not combustible, and their coating contribute with negligible amounts of combustible materials, unless they are jacketed. Therefore the choice of number of fibres as a descriptor is not obvious since the purpose of including such a descriptor into the cable parameter was to describe the appearance of new highly combustible elements when the outer jacket was destroyed [4]. In this report two different number descriptors have been used for the cable parameter. The first is the number of optical fibres, as discussed above. The second is the number of combustible tubular units. This is similar to the number of tubes according to the optical cable nomenclature, but the number of units and tubes are not always identical. For example, a tight buffered fibre cable does not have any tube(s), whereas each buffered fibre is considered as a unit. A unit is defined as a combustible tubular item which may contain one or several fibres. The optical fibre with an outer diameter of approximately 250 µm is not considered a unit, nor is the outer jacket of the cable. A tubular item is not considered a unit if it contains other units. Examples of units can be; tight buffered fibre, microbundle, flextube, unitube and loose tube. The number of tubes attempt to describe the same feature as the number of tubular insulations around the conductors in a power cable. See appendix in Section 9 for definition of units for the different families within this study. The cable parameter χ is defined according to Equation 9, where the number descriptor n can be either the number of optical fibres or the number of units, as discussed above. n V 2 combust Equation 9 d with d [m] Outer diameter. V combust [m 2 ] Combustible 3 volume per meter ladder. n [ ] Number of optical fibres or units in one cable. This section describes the method used to quantify the confidence of a potential EXAP rule. This quantification allows a comparison of difference potential EXAP rules. Section 5 shows the results using the outlined quantification method. In this report four different potential EXAP rules for optical fibre cables are investigated and compared: calculating based on number of optical fibres, and not using the safety margin ν sm calculating based on number of optical fibres, and using the safety margin ν sm calculating based on number of units, and not using the safety margin ν sm calculating based on number of units, and using the safety margin ν sm 3 In the CEMAC II project V combust was defined as the non-metallic volume per meter ladder. For optical fiber cables, where typically no metals exist, V combust will instead be defined as the volume of combustible material per meter ladder.

17 17 Table 4 shows the analysis for cable family 1 where the tested EXAP rule is to use the number of optical fibres as the number descriptor n in Equation 9, and without using the safety margins given in Table 2. The five cables in the family are arranged in increasing χ-order (using n = number of optical fibres). The first column gives the name of the cable. The second contains the cable parameter χ. Column 3-8 contain the specific ification for the parameters flame spread, peak heat release rate, total heat release, peak smoke production rate, and total smoke production, respectively. The ifications are made based on the ification criteria in Table 1. Column 9 shows the total, which is the worst result from columns 3 (FS) to 6 (FIGRA). Column 10 shows the total smoke, which is the worst result from columns 7 (pspr) and 8 (TSP). Column 11 finally shows the reported droplet. As an example it can be seen in Table 4 that for THR there are three pairwise combinations that would give a non-conservative EXAP estimation for Product 4. In other words this EXAP approach would, for the THR parameter, estimate Product 4 as a cable with Dca for 3 combinations. These 3 combinations are considered as EXAP errors and are indicated on the second last row. For 5 cables, such as in this example, there are in total 6 pairwise combinations with at least one intermediate cable. The error rate is the percentage of errors in relation to the total possible number of combinations, that is 3/6 = 50% in this example. If the THR for Product 5 would have been Eca instead of Dca there would have been no EXAP errors for THR. The reason for this is that from the pairwise combinations it is the worst case that is used for the ification. If Product 5 would have been Eca the EXAP applications involving Product 5 would not have allowed Product 4 to erroneously be EXAP ificed as Dca. Table 4. Example of confidence analysis. Family 1 with n=number of optical fibres, and without using sm Combinations 6 Product 1 53 Dca Dca Dca Cca s1 s1 Dca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d0 Product Dca Dca Dca Dca s1 s1 Dca s1 d1 Product Dca Dca Eca Dca s1 s1 Eca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d1 Errors (No) Error rate % Table 5 shows a similar analysis as in Table 4, but with the safety margin included. The upper table is identical to Table 4. The lower table contain ifications based on the test results plus the safety margin as defined in Table 2. The tested pairwise combinations now contains ifications with an added safety margin, making the system more robust to misifications of intermediate cables. Notice that the investigation whether intermediate cables become non-conservatively ified or not should be based on the original ification, from the upper table. This is the reason why red arrows have been drawing from the upper table to the corresponding intermediate positions in the lower table. In this example the result was that adding safety margins reduced FIGRA errors from 2 to 0 since Product 5 obtains Dca instead of Cca using the safety margins. Therefore, for FIGRA, no pairwise combinations are possible that result in a nonconservative ification (e.g. Dca cable ified as Cca) of the intermediate cables. The number of total ifications did not change, for this particular case.

18 18 All analyses are presented in the appendix in Section 0. Table 5. Example of confidence analysis. Family 1 with n=number of optical fibres, and using sm Combinations 6 Product 1 53 Dca Dca Dca Cca s1 s1 Dca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d0 Product Dca Dca Dca Dca s1 s1 Dca s1 d1 Product Dca Dca Eca Dca s1 s1 Eca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d1 Errors (No) Error rate % Combinations 6 Product 1 53 Dca Dca Dca Dca s1 s1 Dca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d0 Product Dca Dca Dca Dca s1 s1 Dca s1 d1 Product Dca Dca Eca Dca s1 s1 Eca s1 d0 Product Dca Dca Dca Dca s1 s1 Dca s1 d1 Errors (No) Error rate %

19 FS phrr THR FIGRA pspr TSP Main 19 5 Results Table 6 to Table 9 shows the number of ification errors for all families and for all four tested EXAP procedures. Table 6 shows the result for an EXAP procedure where the number descriptor n is given by the number of optical fibres, and where no safety margin is used. Table 7 shows a similar analysis but where the safety margins of Table 2 are used. The families and parameters where the number of errors differs between the analyses without and with safety margins are highlighted in yellow. It can be seen that the safety margins reduced the number of errors at six positions. Table 8 shows the analysis where the number descriptor n is given by the number of units, and where no safety margin is used. The positions where the number of errors differs as compared to the case where n is the number of optical fibres are indicated in turquoise. This is the case at 13 positions and the difference can be both positive and negative. Table 9 finally shows the result when n is the number of units and when safety margins are used. Differences are highlighted with respect to the results in Table 6 to Table with n = number of optical fibres Table 6. Number of ification errors with n=number of optical fibres, and without using sm. Errors Fam. No. cables No. comb

20 FS phrr THR FIGRA pspr TSP Main 20 Table 7. Number of ification errors with n=number of optical fibres, and using sm. The numbers that differs as compared to Table 6 are marked in yellow. Errors Fam. No. cables No. comb

21 FS phrr THR FIGRA pspr TSP Main with n = number of units Table 8. Number of ification errors with n=number of tubes, and without using sm. The numbers that differs as compared to Table 6 (n=number of optical fibres) are marked in turquoise. Errors Fam. No. cables No. comb

22 FS phrr THR FIGRA pspr TSP Main 22 Table 9. Number of ification errors with n=number of tubes, and using sm. The numbers that differs as compared to Table 8 (without using sm ) are marked in yellow. The numbers that differs as compared to Table 7 (n=number of optical fibres, using sm ) are marked in turquoise. The numbers that differs as compared to both Table 7 and Table 8 are marked in green. Errors Fam. No. cables No. comb

23 FS phrr THR FIGRA pspr TSP Main FS phrr THR FIGRA pspr TSP Main Summary of results and comparison with results from CEMAC II project (power cables) The four rows in Table 10 show the condensed results for all families for Table 6 to Table 9, respectively. The EXAP procedure where n in Equation 9 is given by number of units in the cables, and where the safety margin is used, gives the best result with 2 main errors out of a total of 103 combinations. The same EXAP procedure gives as many as 21 smoke errors, and 16 droplet errors. The high number of smoke errors is entirely due to one cable in cable family 8. Table 10. Summary of results for the number of errors. The total number of combinations is 103. Error n sm OF no OF yes units no a 16 units yes a 16 a These 21 errors are all due to one single cable in family 8 with TSP=52.3 m 2, yielding an s2- due to the limit 50 m 2 between s1 and s2 es. All other cables in family 9 have s1. The cable in question had highest using n = number of optical fibres, whereas the cable had an intermediate using n = number of units. The total number of combinations in Table 10 is 103. Since this is close to 100 the number of errors in the table can also be approximated by the error rate in percent. This can be compared with the error rate in Table 11 for the EXAP for power cables as reported in the CEMAC II report. While the maximum error rate for an individual parameter for the EXAP for power cables is 2%, it is 16% for the EXAP with fewest main errors for optical fibre cables (n = number of units and using sm ). Table 11. Summary of results for the number of errors and error rate in the CEMAC II project. The total number of combinations was 166. Errors Error rate [%] a a a a a a a Not analysed.

24 24 6 Conclusions Four different approaches for EXAP procedures for optical fibre cables have been investigated. The test set consisted of 66 different cables split into 14 families. An analysis method for quantifying the confidence of an EXAP procedure was defined and tested on the four different EXAP approaches. Using an EXAP procedure where the cable parameter χ is described in terms of units in the cables, and using the same system with safety margin as for the EXAP for power cables, only two main ification errors occurred out of a total of 103 pairwise EXAP combinations. The number of smoke errors for this EXAP procedure was as high as 21, but all these errors where due to one single cable. The number of droplet errors for this EXAP procedure was 16. As compared to the other different EXAP procedures this procedure, with χ described in terms of units in the cables and using a safety margin sm, is considered to be the best alternative. The confidence of the proposed EXAP for optical fibre cables is considered as acceptable, with only 2% main error rate.

25 25 7 References 1. Council Directive 89/106/EEC of 21 December 1988 on the approximation of laws, regulations and administrative provisions of the Member States relating to construction products OJ No L 40 of 11 February 1989, E. Union, Editor Försth, M., et al. Status summary of cable reaction to fire reguations in Europs. in 62nd International Cable - Connectivity Symposium Charlotte, NC, USA. 3. Regulation (EU) No 305/2011 of the European Parliament and of the Council of 9 March 2011 laying down harmonized conditions for the marketing of construction products and repealing Council Directive 89/106/EEC, OJ L 88 of 4 April 2011, E. Union, Editor Journeaux, T., et al., CEMAC - CE-marking of cables, SP Report 2010_ , SP Technical Research Institute of Sweden: Borås. 5. Sundström, B., et al. Prediction of fire ification of cables, extended application of test data. in Interflam CENELEC, CLC/TS Electric cables - Extended application of test results COMMISSION DECISION of 27 October 2006 amending Decision 2000/147/EC implementing Council Directive 89/106/EEC as regards the ification of the reaction-to-fire performance of construction products (2006/751/EC), E. Union, Editor CENELEC, EN :2014 Fire ification of construction products and building elements Part 6: ification using data from reaction to fire tests on electric cables. 2014, CENELEC. 9. CENELEC, EN 50399, Common test methods for cables under fire conditions - Heat release and smoke production measurements on cables during flame spread test - Test apparatus, procedures, results. 2011, CENELEC. 10. EN Tests on electric and optical fibre cables under fire conditions - Part 1-2: Test for vertical flame propagation for a single insulated wire or cable - Procedure for 1 kw pre-mixed flame. 2004, CENELEC. 11. EN ISO 1716 Reaction to fire tests for products - Determination of the gross heat of combustion (calorific value) (ISO 1716:2010). 2010, CENELEC. 12. EN Measurement of smoke density of cables burning under defined conditions - Part 2: Test procedure and requirements. 2005, CENELEC. 13. EN Common test methods for cables under fire conditions - Tests on gases evolved during combustion of materials from cables - Part 2-3: Procedures - Determination of degree of acidity of gases for cables by determination of the weighted average of ph and conductivity. 1998, CENELEC. 14. Fire testing and ification protocol for mineral wool products, F.s.g.o.n.b.f.t. CPD, Editor Grayson, S., et al., FIPEC, Fire Performance of Electric Cables - new test methods and measurement techniques

26 26 8 Appendix: Proposal for EXAP-rules for optical fibre cables The EXAP is applicable to ification into the es B2 ca, C ca and D ca, the smoke es s1, s2 and s3, and the droplet es d0, d1 and d2. The EXAP is not applicable to cables with a diameter < 5 mm. 8.1 Definition of a product family for EXAP For the purposes of applying these EXAP rules, a cable family should be defined as follows: A family of cables is a specific range of products of the same general construction and varying only in number of optical fibres and number of units. The specific family shall be produced by the same manufacturer using the same materials and the same design rules based on national or international standards, and/or company standards. The following properties are considered to have a negligible influence on the fire behaviour and therefore differences in these properties only do not necessarily imply that cables belong to different families: Fibre glass type Fibre type (single mode or monomode) Fibre colour Outer jacket colour Printing The full constructional and material details for the family shall be submitted to the certification body prior to the EXAP being applied. 8.2 EXAP with safety margin An EXAP is based on two tests. The parameter is used as independent cable parameter. is defined as: n 2 d V combust where d [m] Outer diameter. V combust [m 2 ] Combustible volume per meter ladder. n [ ] Number of units in one cable. A unit is defined as a combustible tubular item which may contain one or several fibres. The optical fibre with an outer diameter of approximately 250 µm is not considered a unit, nor is the outer jacket of the cable. A tubular item is not considered a unit if it contains other units. Examples of units can be; tight buffered fibre, microbundle, flextube, unitube and loose tube. The number of tubes attempt to describe the same feature as the number of tubular insulations around the conductors in a power cable.

27 27 All cables within the same family with a value of the cable parameter between the lowest and highest value of the cable parameters of the tested cables are included in the EXAP. ification is based on the maximum measured value plus a safety margin: max sm where max sm is the value used for ification according to respective ification parameter (peak HRR, THR, FIGRA, FS, peak SPR, and TSP), is the maximum, that is the worst, test results of the tests that forms the basis of the EXAP, and is the safety margin required for the particular ification parameter. The safety margins for the different es and ification parameters are given in Table 12. Table 12 Safety margins v sm. B2ca Cca Dca S1 S2 Peak HRR [kw] THR [MJ] FIGRA [Ws -1 ] Flame spread [m] Peak SPR [m 2 s -1 ] TSP [m 2 ] Flaming droplets/particles For flaming droplets/particles the cables within the cable parameter range for the EXAP are ified according to the worst result for the tested cables within this range.

28 28 9 Appendix: Cable data The information in this section was supplied by Europacable. Cable selection and description The optical fiber cable product family selected for the test program was made ensuring that they are representative for the European market. Only LSZH materials are used as sheathing compound. The emphasis was put on the fact that it was also necessary to have a wide range of burning behaviors to avoid a bias on the final analysis. For the selection of the different product families the following logics has been applied. - Grouping has been made on principal design methods covering the typical range of fibre counts per family. - Fibre types and fibre colours (i.e. various types of glass for single mode or multimode fibres) are considered to be irrelevant for the fire performance of the cable. - Outer jacket colours and various printing can be neglected for fire performance. - Families have been designed in a way that the principal material types (or even compound) remain unchanged when increasing the number the number of fibres leading to bigger diameters. The cable product are given below :

29 29 Product family 1-4 Distribution tight buffer cable Definition: This cable is designed with 900µm buffered structure for easy direct connectorization. It is a dry design for risers and horizontal deployment and may be installed in duct by pulling. Unit definition: an optical fiber of approximately 250 µm surrounded by a thermoplastic buffered material of approximately 900 µm in outer diameter. Material List: - Rigid central strength element made of glass fiber reinforced plastic - Optical fiber ø 250 µm sheathed up to ø 900 µm - Glass or Aramid yarns as reinforcement - Outer sheath, Low Halogen Free Flame Retardant material Design: Cable product family: Family Cables Diameter Units Total account of fibers (min-max) [mm]

30 30 Product family 5-7 Central Loose Tube cable, jelly filled Definition: Cylindar cable with one central loose tube which contains n optical fibers. The fiber count reaches from n = 2 up to n = 24 fibers. The loose tube is filled with jelly. Around the loose tube are placed glass roving tension elements and this cable core is protected by a low smoke zero halogen (LSOH) fire retardant sheath. Unit definition: optical fibers, each approximately 250 µm, surrounded by filling compound and a loose tube made of thermoplastic material. Material List: - OF: Optical Fiber with outer diameter of 250µm - Colored loose tube with jelly filling - glass yarns tension elements - swelling yarns - Colored Halogen free thermoplastic sheath of the cable Design: Cable product family: Family Cables Diameter Units Total account of fibers (min-max) [mm]

31 31 Central Loose Tube cable, jelly filled, with circuit and insulation integrity 90 min Definition: Cylindar cable with one central loose tube which contains n optical fibers. The fiber count reaches from n = 2 up to n = 24 fibers. The loose tube is filled with jelly. Around the loose tube are placed a non metallic tape and additional glass roving tension elements and this cable core is protected by a low smoke zero halogen (LSOH) fire retardant sheath. Unit definition: optical fibers, each approximately 250 µm, surrounded by filling compound and a central loose tube made of thermoplastic material. Material List: - OF: Optical Fiber with outer diameter of 250µm - Colored loose tube with jelly filling - flame barrier tape - glass yarns tension elements - two Polyester ripcords - Colored Halogen free thermoplastic sheath of the cable Design: Cable product family: Family Cables Diameter Units Total account of fibers (min-max) [mm] ,3 10,

32 32 Central Loose Tube cable, jelly filled Definition: Cylindar cable with one central loose tube which contains n optical fibers. The fiber count reaches from n = 2 up to n = 24 fibers. The loose tube is filled with jelly. Around the loose tube are placed glass roving tension elements and this cable core is protected by a low smoke zero halogen (LSOH) fire retardant sheath. Unit definition: optical fibers, each approximately 250 µm, surrounded by filling compound and a central loose tube made of thermoplastic material. Material List: - OF: Optical Fiber with outer diameter of 250µm - Colored loose tube with jelly filling - glass yarns tension elements - swelling yarns - Colored Halogen free thermoplastic sheath of the cable Design: Cable product family: Family Cables Diameter Units Total account of fibers (min-max) [mm] 7 3 7,7 8,

33 33 Product family 8-9 Indoor distribution or mini break-out cable with buffered fibres (2 24) and fire resistant outer jacket Definition: Circular cable with strength members of glass or aramid yarns under the jacket covering the buffered 2-24 fibres. This distribution or mini-break-out cable can be used for many indoor applications. The cable features buffered fibres. Typical cable applications include: LAN and WAN backbones, central office interconnections, backbones in data centres among others. The cable features glass yarns for ease of installation and is suited for installation in ducts and on trays. The cable features a resistant FireRes sheathing. Unit definition: an optical fiber of approximately 250 µm surrounded by a thermoplastic buffered material of approximately 900 µm in outer diameter. Material List: - OF: Optical Fiber with outer diameter of 250µm - Colored Halogen Free thermoplastic buffer (typically 900µm) - glass yarn or aramid yarn as strength members - UV stabilized and Colored Halogen free thermoplastic sheath of the cable Design: - Buffered fibre - Strength member (glass/aramid yarn) - LSZH outer jacket Cable product family: Family Cables Diameter Units Total account of fibers (min-max) [mm]

34 34 Product family 10 - Compact Tube premise distribution riser cable for permanent accessibility. Definition: Cylindar cable with rigid strength members into the sheath made up of n Compact Tube (unit from 2 to 12 fibres). This cable product family has been designed to allow permanent accessibility to fibers through the sheath for splicing with customer cables. The cable, protected by a low smoke zero halogen (LSOH) fire retardant sheath, comprises micro modules (called Compact Tube ). Each micro module contains 2 to 12 fibers protected by an easily strippable thermoplastic skin (easy strip technology). Unit definition: optical fibers, each approximately 250 µm, surrounded by a finger peelable sheath made of thermoplastic material, called compact tube or micromodule. This micromodule can be jelly filled, dry watertight. Material List: - OF: Optical Fiber with outer diameter of 250µm - Colored Halogen Free thermoplastic sheath of OF unit - 2 FRP strength members - UV stabilized and Colored Halogen free thermoplastic sheath of the cable Design: Cable product family: Family Cables Diameter Units Total account of fibers (min-max) [mm]

35 35 Product family Loose Tube Indoor Cable up to 72 fibers Definition: The loose tube cable construction, by isolating the fibers from installations and environmental rigors, provides stable and highly reliable transmission parameters. The SZ stranded construction further reduces installation and environmental influences on the transmission parameters and allows mid-span access. The cables can be installed in conduits and shafts inside buildings. Unit definition: optical fibers, each approximately 250 µm, surrounded by filling compound and a loose tube made of thermoplastic material. Material List: A. FRNC/LSZH jacket B. Buffer tube, gel-filled with 12 fibers C. Fiber, 250 µm D. Dielectric central member E. Ripcord Design: Cable product family: Family Cables Diameter Units Total account of fibers (min-max) [mm]

36 36 Definition: This cable family is designed with highly flexible subunits for EDGE application. The fiber range is defined from 12f up to 144f. Aramid yarns inside the cable guaranty the required tensile load. A flame retardant, low smoke jacket will provide the appropriate fire performance. Unit definition: optical fibers, each approximately 250µm, some aramid yarns, surrounded by a thin, flexible tube made of flame retardant thermoplastic material. Material List: A. FRNC/LSZH jacket B. Subunits with 12 fibers C. Fiber, 250 µm D. Aramid yarns as strength elements Design: Cable product family: Family Cables Diameter Units Total account of fibers (min-max) [mm]

37 37 10 Appendix: Detailed analysis 10.1 χ calculated based on number of optical fibers, without using safety margin sm Family 1 Combinations 6 Product 1 53 Dca Dca Dca Cca s1 s1 Dca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d0 Product Dca Dca Dca Dca s1 s1 Dca s1 d1 Product Dca Dca Eca Dca s1 s1 Eca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d1 Errors (No) Error rate % Family 2 Combinations 1 Product 1 53 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Errors (No) Error rate % Family 3 Combinations 0 X FS phrr THR FIGRA pspr TSP (m) (kw) (MJ) (W/s) (m2/s) (m2) Product Dca Cca Cca B2ca s1 s1 Dca s1 d1 Product Cca Cca Cca B2ca s1 s1 Cca s1 d1 Errors (No) NA NA NA NA NA NA NA NA NA Error rate % NA NA NA NA NA NA NA NA NA Family 4 Combinations 6 Product 1 68 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product Cca Cca Cca B2ca s1 s1 Cca s1 d0 Errors (No) Error rate % Family 5 Combinations 1 Product 1 52 Dca Dca Dca Dca s2 s2 Dca s2 d2 Product Dca Dca Dca Dca s2 s2 Dca s2 d2 Product Dca Dca Dca Eca s2 s2 Eca s2 d1 Errors (No) Error rate %

38 38 Family 6 Combinations 1 Product 1 13 Dca Dca Dca Dca s1 s2 Dca s2 d2 Product 2 76 Dca Dca Dca Dca s1 s1 Dca s1 d0 Product Dca Dca Dca Cca s1 s2 Dca s2 d2 Errors (No) Error rate % Family 7 Combinations 1 Product Dca Dca Dca Dca s1 s2 Dca s2 d1 Product Dca Dca Dca Dca s1 s2 Dca s2 d1 Product Dca Dca Dca Dca s1 s2 Dca s2 d2 Errors (No) Error rate % Family 8 Combinations 45 Product 1 18 Cca Cca B2ca B2ca s1 s1 Cca s1 d0 Product 2 18 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product 4 82 B2ca B2ca B2ca Cca s1 s1 Cca s1 d0 Product 3 83 B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product Cca Cca Cca B2ca s1 s1 Cca s1 d0 Product Cca Cca Cca B2ca s1 s1 Cca s1 d0 Product B2ca Cca Cca Cca s1 s1 Cca s1 d0 Product B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product Cca Cca Cca Cca s1 s1 Cca s1 d0 Product Cca Cca Cca B2ca s1 s2 Cca s2 d0 Errors (No) Error rate % Family 9 Combinations 15 Product 1 69 Dca B2ca Cca Cca s1 s1 Dca s1 d0 Product 2 69 Dca B2ca Cca Cca s1 s1 Dca s1 d0 Product B2ca Cca B2ca B2ca s1 s1 Cca s1 d0 Product Cca Cca B2ca Cca s1 s1 Cca s1 d1 Product Dca Dca Dca Cca s1 s1 Dca s1 d2 Product Cca Cca Cca Cca s1 s1 Cca s1 d0 Product Dca Cca Cca Cca s1 s1 Dca s1 d0 Errors (No) Error rate %

39 39 Family 10 Combinations 21 Product 3 87 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d2 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product Cca B2ca Cca B2ca s1 s1 Cca s1 d2 Product B2ca B2ca Cca B2ca s1 s1 Cca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d2 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d2 Product B2ca B2ca Cca B2ca s1 s1 Cca s1 d1 Errors (No) Error rate % Family 11 Combinations 1 Product Dca Dca Eca Cca s2 s2 Eca s2 d2 Product Dca Dca Eca Cca s2 s2 Eca s2 d2 Product Dca Dca Eca Dca s2 s2 Eca s2 d2 Errors (No) Error rate % Family 12 Combinations 3 Product Cca B2ca B2ca B2ca s1 s1 Cca s1 d1 Product Dca Cca B2ca B2ca s1 s1 Dca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca Cca Cca B2ca s1 s1 Cca s1 d2 Errors (No) Error rate % Family 13 Combinations 0 Product Dca Cca B2ca B2ca s1 s1 Dca s1 d0 Product Cca Cca B2ca B2ca s1 s1 Cca s1 d0 Errors (No) NA NA NA NA NA NA NA NA NA Error rate % NA NA NA NA NA NA NA NA NA Family 14 Combinations 1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Errors (No) Error rate %

40 χ calculated based on number of optical fibers, using safety margin sm Family 1 Combinations 6 Product 1 53 Dca Dca Dca Dca s1 s1 Dca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d0 Product Dca Dca Dca Dca s1 s1 Dca s1 d1 Product Dca Dca Eca Dca s1 s1 Eca s1 d0 Product Dca Dca Dca Dca s1 s1 Dca s1 d1 Errors (No) Error rate % Family 2 Combinations 1 Product 1 53 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Errors (No) Error rate % Family 3 Combinations 0 X FS phrr THR FIGRA pspr TSP (m) (kw) (MJ) (W/s) (m2/s) (m2) Product Dca Cca Cca B2ca s1 s1 Dca s1 d1 Product Dca Cca Cca B2ca s1 s1 Dca s1 d1 Errors (No) NA NA NA NA NA NA NA NA NA Error rate % NA NA NA NA NA NA NA NA NA Family 4 Combinations 6 Product 1 68 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca Cca B2ca B2ca s1 s1 Cca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product Cca Cca Cca B2ca s1 s1 Cca s1 d0 Errors (No) Error rate % Family 5 Combinations 1 Product 1 52 Dca Dca Dca Dca s2 s2 Dca s2 d2 Product Dca Dca Dca Dca s2 s2 Dca s2 d2 Product Dca Eca Dca Eca s2 s2 Eca s2 d1 Errors (No) Error rate %

41 41 Family 6 Combinations 1 Product 1 13 Dca Dca Dca Dca s1 s2 Dca s2 d0 Product 2 76 Dca Dca Dca Dca s1 s1 Dca s1 d0 Product Dca Dca Dca Cca s1 s2 Dca s2 d2 Errors (No) Error rate % Family 7 Combinations 1 Product Dca Dca Dca Dca s1 s2 Dca s2 d1 Product Dca Dca Dca Dca s1 s2 Dca s2 d1 Product Dca Dca Dca Dca s1 s2 Dca s2 d2 Errors (No) Error rate % Family 8 Combinations 45 Product 1 18 Cca Cca B2ca B2ca s1 s1 Cca s1 d0 Product 2 18 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product 4 82 B2ca B2ca B2ca Cca s1 s1 Cca s1 d0 Product 3 83 B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product Cca Cca Cca B2ca s1 s1 Cca s1 d0 Product Cca Cca Cca B2ca s1 s1 Cca s1 d0 Product B2ca Cca Cca Cca s1 s1 Cca s1 d0 Product B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product Cca Cca Cca Cca s1 s1 Cca s1 d0 Product Cca Cca Cca B2ca s1 s2 Cca s2 d0 Errors (No) Error rate % Family 9 Combinations 15 Product 1 69 Dca B2ca Cca Dca s1 s1 Dca s1 d0 Product 2 69 Dca B2ca Cca Dca s1 s1 Dca s1 d0 Product B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product Dca Cca Cca Cca s1 s1 Dca s1 d1 Product Dca Dca Dca Cca s1 s1 Dca s1 d2 Product Dca Cca Cca Cca s1 s1 Dca s1 d0 Product Dca Cca Cca Cca s1 s1 Dca s1 d0 Errors (No) Error rate %

42 42 Family 10 Combinations 21 Product 3 87 B2ca B2ca B2ca Cca s1 s1 Cca s1 d2 Product B2ca B2ca B2ca B2ca s1 s1 B2Ca s1 d1 Product Cca B2ca Cca B2ca s1 s1 Cca s1 d2 Product B2ca B2ca Cca B2ca s1 s1 Cca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d2 Product Cca B2ca B2ca B2ca s1 s1 Cca s1 d1 Product Cca B2ca B2ca B2ca s1 s1 Cca s1 d2 Product Cca B2ca Cca B2ca s1 s1 Cca s1 d1 Errors (No) Error rate % Family 11 Combinations 1 Product Dca Dca Eca Dca s2 s2 Eca s2 d2 Product Dca Dca Eca Cca s2 s2 Eca s2 d2 Product Dca Dca Eca Dca s2 s2 Eca s2 d2 Errors (No) Error rate % Family 12 Combinations 3 Product Cca B2ca B2ca B2ca s1 s1 Cca s1 d1 Product Dca Cca Cca Cca s1 s1 Dca s1 d1 Product B2ca B2ca Cca B2ca s1 s1 Cca s1 d1 Product Cca Cca Cca B2ca s1 s1 Cca s1 d2 Errors (No) Error rate % Family 13 Combinations 0 Product Dca Cca B2ca Cca s1 s1 Dca s1 d0 Product Cca Cca Cca B2ca s1 s1 Cca s1 d0 Errors (No) NA NA NA NA NA NA NA NA NA Error rate % NA NA NA NA NA NA NA NA NA Family 14 Combinations 1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Errors (No) Error rate %

43 χ calculated based on number of units, without using safety margin sm Family 1 Combinations 6 Product Dca Dca Eca Dca s1 s1 Eca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d1 Product Dca Dca Dca Dca s1 s1 Dca s1 d1 Errors (No) Error rate % Family 2 Combinations 1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Errors (No) Error rate % Family 3 Combinations 0 X FS phrr THR FIGRA pspr TSP (m) (kw) (MJ) (W/s) (m2/s) (m2) Product Dca Cca Cca B2ca s1 s1 Dca s1 d1 Product Cca Cca Cca B2ca s1 s1 Cca s1 d1 Errors (No) NA NA NA NA NA NA NA NA NA Error rate % NA NA NA NA NA NA NA NA NA Family 4 Combinations 6 Product Cca Cca Cca B2ca s1 s1 Cca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Errors (No) Error rate % Family 5 Combinations 1 Product 1 13 Dca Dca Dca Dca s2 s2 Dca s2 d2 Product Dca Dca Dca Dca s2 s2 Dca s2 d2 Product Dca Dca Dca Eca s2 s2 Eca s2 d1 Errors (No) Error rate %

44 44 Family 6 Combinations 1 Product Dca Dca Dca Dca s1 s2 Dca s2 d0 Product Dca Dca Dca Dca s1 s1 Dca s1 d0 Product Dca Dca Dca Cca s1 s2 Dca s2 d2 Errors (No) Error rate % Family 7 Combinations 1 Family 8 Product 2 14 Dca Dca Dca Dca s1 s2 Dca s2 d1 Product 1 15 Dca Dca Dca Dca s1 s2 Dca s2 d1 Product 3 16 Dca Dca Dca Dca s1 s2 Dca s2 d2 Errors (No) Error rate % Combinations 45 Product 1 71 Cca Cca B2ca B2ca s1 s1 Cca s1 d0 Product 2 71 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product Cca Cca Cca Cca s1 s1 Cca s1 d0 Product Cca Cca Cca B2ca s1 s2 Cca s2 d0 Product 9 76 B2ca Cca Cca Cca s1 s1 Cca s1 d0 Product 7 82 Cca Cca Cca B2ca s1 s1 Cca s1 d0 Product 8 82 Cca Cca Cca B2ca s1 s1 Cca s1 d0 Product B2ca B2ca B2ca Cca s1 s1 Cca s1 d0 Product B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Errors (No) Error rate % Family 9 Combinations 15 Product 3 71 B2ca Cca B2ca B2ca s1 s1 Cca s1 d0 Product 6 71 Cca Cca Cca Cca s1 s1 Cca s1 d0 Product 7 75 Dca Cca Cca Cca s1 s1 Dca s1 d0 Product 4 75 Cca Cca B2ca Cca s1 s1 Cca s1 d1 Product 5 76 Dca Dca Dca Cca s1 s1 Dca s1 d2 Product 1 82 Dca B2ca Cca Cca s1 s1 Dca s1 d0 Product 2 82 Dca B2ca Cca Cca s1 s1 Dca s1 d0 Errors (No) Error rate %

45 45 Family 10 Combinations 21 X FS phrr THR FIGRA pspr TSP (m) (kw) (MJ) (W/s) (m2/s) (m2) Product 5 10 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product 3 44 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d2 Product 4 48 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d2 Product 7 67 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d2 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca B2ca Cca B2ca s1 s1 Cca s1 d1 Product Cca B2ca Cca B2ca s1 s1 Cca s1 d2 Product B2ca B2ca Cca B2ca s1 s1 Cca s1 d1 Errors (No) Error rate % Family 11 Combinations 1 Product Dca Dca Eca Dca s2 s2 Eca s2 d2 Product Dca Dca Eca Cca s2 s2 Eca s2 d2 Product Dca Dca Eca Cca s2 s2 Eca s2 d2 Errors (No) Error rate % Family 12 Combinations 3 Product 2 51 Cca B2ca B2ca B2ca s1 s1 Cca s1 d1 Product 1 62 Dca Cca B2ca B2ca s1 s1 Dca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca Cca Cca B2ca s1 s1 Cca s1 d2 Errors (No) Error rate % Family 13 Combinations 0 Product 1 48 Dca Cca B2ca B2ca s1 s1 Dca s1 d0 Product Cca Cca B2ca B2ca s1 s1 Cca s1 d0 Errors (No) NA NA NA NA NA NA NA NA NA Error rate % NA NA NA NA NA NA NA NA NA Family 14 Combinations 1 Product 2 52 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product 1 62 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Errors (No) Error rate %

46 χ calculated based on number of units, using safety margin sm Family 1 Combinations 6 Product Dca Dca Eca Dca s1 s1 Eca s1 d0 Product Dca Dca Dca Cca s1 s1 Dca s1 d0 Product Dca Dca Dca Dca s1 s1 Dca s1 d0 Product Dca Dca Dca Dca s1 s1 Dca s1 d1 Product Dca Dca Dca Dca s1 s1 Dca s1 d1 Errors Error rate Family 2 Combinations 1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Errors (No) Error rate % Family 3 Combinations 0 X FS phrr THR FIGRA pspr TSP (m) (kw) (MJ) (W/s) (m2/s) (m2) Product Dca Cca Cca B2ca s1 s1 Dca s1 d1 Product Dca Cca Cca B2ca s1 s1 Dca s1 d1 Errors (No) NA NA NA NA NA NA NA NA NA Error rate % NA NA NA NA NA NA NA NA NA Family 4 Combinations 6 Product Cca Cca Cca B2ca s1 s1 Cca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product B2ca Cca B2ca B2ca s1 s1 Cca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Errors (No) Error rate % Family 5 Combinations 1 Product 1 13 Dca Dca Dca Dca s2 s2 Dca s2 d2 Product Dca Dca Dca Dca s2 s2 Dca s2 d2 Product Dca Eca Dca Eca s2 s2 Eca s2 d1 Errors (No) Error rate %

47 47 Family 6 Combinations 1 Product Dca Dca Dca Dca s1 s2 Dca s2 d0 Product Dca Dca Dca Dca s1 s1 Dca s1 d0 Product Dca Dca Dca Cca s1 s2 Dca s2 d2 Errors (No) Error rate % Family 7 Combinations 1 Product 2 14 Dca Dca Dca Dca s1 s2 Dca s2 d1 Product 1 15 Dca Dca Dca Dca s1 s2 Dca s2 d1 Product 3 16 Dca Dca Dca Dca s1 s2 Dca s2 d2 Errors (No) Error rate % Family 8 Combinations 45 Product 1 71 Cca Cca B2ca B2ca s1 s1 Cca s1 d0 Product 2 71 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Product Cca Cca Cca Cca s1 s1 Cca s1 d0 Product Cca Cca Cca B2ca s1 s2 Cca s2 d0 Product 9 76 B2ca Cca Cca Cca s1 s1 Cca s1 d0 Product 7 82 Cca Cca Cca B2ca s1 s1 Cca s1 d0 Product 8 82 Cca Cca Cca B2ca s1 s1 Cca s1 d0 Product B2ca B2ca B2ca Cca s1 s1 Cca s1 d0 Product B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Errors (No) Error rate % Family 9 Combinations 15 Product 3 71 B2ca Cca B2ca Cca s1 s1 Cca s1 d0 Product 6 71 Dca Cca Cca Cca s1 s1 Dca s1 d0 Product 7 75 Dca Cca Cca Cca s1 s1 Dca s1 d0 Product 4 75 Dca Cca Cca Cca s1 s1 Dca s1 d1 Product 5 76 Dca Dca Dca Cca s1 s1 Dca s1 d2 Product 1 82 Dca B2ca Cca Dca s1 s1 Dca s1 d0 Product 2 82 Dca B2ca Cca Dca s1 s1 Dca s1 d0 Errors (No) Error rate %

48 48 Family 10 Combinations 21 X FS phrr THR FIGRA pspr TSP (m) (kw) (MJ) (W/s) (m2/s) (m2) Product 5 10 B2ca B2ca B2ca B2ca s1 s1 B2Ca s1 d1 Product 3 44 B2ca B2ca B2ca Cca s1 s1 Cca s1 d2 Product 4 48 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d2 Product 7 67 Cca B2ca B2ca B2ca s1 s1 Cca s1 d2 Product Cca B2ca B2ca B2ca s1 s1 Cca s1 d1 Product Cca B2ca Cca B2ca s1 s1 Cca s1 d1 Product Cca B2ca Cca B2ca s1 s1 Cca s1 d2 Product B2ca B2ca Cca B2ca s1 s1 Cca s1 d1 Errors (No) Error rate % Family 11 Combinations 1 Product Dca Dca Eca Dca s2 s2 Eca s2 d2 Product Dca Dca Eca Dca s2 s2 Eca s2 d2 Product Dca Dca Eca Cca s2 s2 Eca s2 d2 Errors (No) Error rate % Family 12 Combinations 3 Product 2 51 Cca B2ca B2ca B2ca s1 s1 Cca s1 d1 Product 1 62 Dca Cca Cca Cca s1 s1 Dca s1 d1 Product B2ca B2ca Cca B2ca s1 s1 Cca s1 d1 Product Cca Cca Cca B2ca s1 s1 Cca s1 d2 Errors (No) Error rate % Family 13 Combinations 0 Product 1 48 Dca Cca B2ca Cca s1 s1 Dca s1 d0 Product Cca Cca Cca B2ca s1 s1 Cca s1 d0 Errors (No) NA NA NA NA NA NA NA NA NA Error rate % NA NA NA NA NA NA NA NA NA Family 14 Combinations 1 Product 2 52 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product 1 62 B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d1 Product B2ca B2ca B2ca B2ca s1 s1 B2ca s1 d0 Errors (No) Error rate %

49 SP Sveriges Tekniska Forskningsinstitut SP Technical Research Institute of Sweden Our work is concentrated on innovation and the development of value-adding technology. Using Sweden's most extensive and advanced resources for technical evaluation, measurement technology, research and development, we make an important contribution to the competitiveness and sustainable development of industry. Research is carried out in close conjunction with universities and institutes of technology, to the benefit of a customer base of about organisations, ranging from start-up companies developing new technologies or new ideas to international groups. SP Technical Research Institute of Sweden Box 857, SE BORÅS, SWEDEN Telephone: , Telefax: info@sp.se, Internet: SP Report 2015:32 ISBN ISSN More information about publications published by SP: