ENV 1187 test method 2 Improvement of test equipment and procedure

Full text

(1)ENV 1187 test method 2 Improvement of test equipment and procedure Nordtest Project No. 1572-02. SP Fire Technology SP REPORT 2004:25 NT Techn Report 563. SP Swedish National Testing and Research Institute. Ingrid Wetterlund.

(2) Ingrid Wetterlund. ENV 1187 test method 2 Improvement of test equipment and procedure Nordtest Project No. 1572-02.

(3)

(4) 2. Abstract An improved design of the test equipment for test method 2 of ENV 1187 has given more precise results than in earlier comparisons. Deviations between the Nordic laboratories that have been seen before are no longer at hand. Tests performed have shown that the test equipment according to the improved design will give test results that still are on approximately the same Nordic mean level as before. Thus, earlier tested products will keep their classification. A revised definition of damaged length to be used for test method 2 will be needed to improve the precision of test result. It will be suggested to CEN for introduction into ENV 1187. New figures and more precise specifications for the equipment will also be suggested for the pre-standard. Key words: Small-scale fire test, roof coverings. SP Sveriges Provnings- och Forskningsinstitut SP Rapport 2004:25 ISBN 91-7848-999-7 ISSN 0284-5172 Borås 2004. SP Swedish National Testing and Research Institute SP Report 2004:25. NT Techn Report 563 Postal address: Box 857, SE-501 15 BORÅS, Sweden Telephone: +46 33 16 50 00 Telex: 36252 Testing S Telefax: +46 33 13 55 02 E-mail: info@sp.se.

(5) 3. Contents Abstract. 2. Contents. 3. Acknowledgement. 4. Sammanfattning. 5. 1. Introduction. 7. 2 2.1 2.2. The equipment for test method 2 of ENV 1187 Background New drawings and specifications. 9 9 9. 3. 3.3.1 3.3.2 3.3.3 3.3.4. Evaluation of the new apparatus for test method 2 of ENV 1187 New definition of damaged area Confirmation of consistence of test results at SP Confirmation of the turbulence of the airflow Confirmation of the improved precision of the test in the round robin Products tested in the round robin Statistical analysis of the round robin results Comparison with former round robins Air velocity measurements. 4. Conclusions. 20. 5. References. 21. Appendix A. New drawings for method 2 of ENV 1187. 22. Appendix B. Additional/revised specifications for method 2 of ENV 1187. 27. Appendix C. The interpretation of the definitions of damaged length. 30. Appendix D. Comparison of results from supervisory control. 32. Appendix E. The air velocity in the equipments used for ENV 1187 method 2. 34. Turbulence intensity measurements in the new test apparatus at SP used for ENV 1187 - method 2. 41. Results of the simple rebuilding of the old equipments. 43. 3.1 3.2 3.2.1 3.3. Appendix F Appendix G. 10 10 12 13 15 15 15 17 18.

(6) 4. Acknowledgement This work was sponsored by Nordtest (project 1572-02), which is gratefully acknowledged. It is also gratefully acknowledged that the following Nordic roof-covering producers have sponsored the work: Sveriges Tätskiktsfabrikanters Förening, Isola A/S, Icopal AS, Protan AS, Lemminkäinen Oyj, Katepal Oy, Icopal Oy, and Danske Tagpapfabrikanters Brancheforening. Sven-Ove Vendel performed the tests at SP and supported the development of the new equipment. Ulla Fridh arranged the purchase of the test material and supported the planning of the project. Sven-Ove and Ulla also gave valuable contribution to the evaluation of the test results. Elisabeth Wetterlund drew the new figures that will be suggested for the standard. Morgan Svensson at VST-Produkter and Bill Thoresson at SP Engineering workshop led the work on building the prototype and the new equipments for the participating laboratories. Martin Pauner at DIFT, Bjarne Kristoffersen at NBL and Jarmo Ruohomäki at VTT led the work at each of the participating laboratories..

(7) 5. Sammanfattning En förbättrad design av utrustningen till provningsmetod 2 i ENV 1187 har givit bättre precision i provningsresultaten än vad tidigare jämförelser har visat. Skillnader som har funnits tidigare mellan de nordiska brandlaboratorierna har därmed eliminerats. Provningar som genomförts visar att utrustningen enligt den förbättrade designen kommer att ge resultat som blir på ungefär samma nordiska medelnivå som tidigare. Det betyder att tidigare provade produkter kommer att behålla sin klassifikation. En reviderad definition för vad som menas med skadat material kommer att behövas för metod 2 för att precisionen ska förbättras. En sådan kommer att föreslås till CEN att föras in i ENV 1187. Nya figurer och en mer preciserad specifikation för utrustningen kommer också att föreslås till denna förstandard..

(8) 6.

(9) 7. 1. Introduction. The new European system for classification of building products implies that the fire performance of roof coverings shall be tested according to procedures given in the prestandard for testing ENV 1187 [1]. Three test methods are listed in ENV 1187. The work presented here concerns test method 2 of the pre-standard: a test method with burning brands and wind. It is based on the Nordtest method NT FIRE 006 [2], which has been used in the Nordic countries since the 1950’s. Comparative tests with NT FIRE 006 have been performed regularly among the Nordic fire laboratories listed in Table 1. The results received in these comparative tests have shown some deviations in the results. To reduce these deviations it was concluded in Nordtest project 1524-01 performed 2001 [3] that the specifications for the equipment needed further improvement. It was also found that there might be a need for additional recommendations on the test procedure itself. Table 1. The Nordic official fire laboratories. Laboratory Danish Institute of Fire and Security Technology, Denmark Norwegian Fire Research Laboratory, Norway VTT Building and Transport, Fire Technology, Finland SP Swedish National Testing and Research Institute, Fire Technology, Sweden Note: The order of the laboratories in the table above does not reflect the number the laboratories are given in the presentation of the test results.. This project initially started as a one-year-project. The aim was to investigate if some simple arrangements could be done to the existing equipments to improve the precision of the test results. It had already been identified in project 1524-01 that the difference in fire spread was due to different degree of turbulence of the air stream above the test specimen in each apparatus. The effect of additional numbers of perforated plates in the lower air channel was investigated. The results showed that the reproducibility was not improved, see Appendix G. The diversity in the construction of the equipments was too fundamental for such a simple correction to be successful. Plans were therefore drawn during the summer of 2002 on how to rebuild the equipments and later the same year Nordtest approved a prolongation and enlargement of the project. The work on rebuilding the equipments was performed during 2003. A prototype was first built and evaluated at SP. Copies of the prototype were then built for the other partners of the project. The comparative tests were performed around the turn of 2003/2004. The results of the evaluation of the prototype and of the comparative tests are given in chapter 3. The present version of ENV 1187 test method 2 contains only very simple figures of the equipment (see Figure 1) while there in NT FIRE 006 in addition also was a reference to more detailed drawings supplied by SP [4]. During the CEN-ballots there have been requests for better figures for test method 2 of ENV 1187. New figures have therefore been drawn within the project. They will be suggested for introduction into ENV 1187. The new figures, shown in Appendix A, reflect the construction of the current rebuilt Nordic equipments that were used in this project to achieve more precise results..

(10) 8. 430 1. 1. 2800. 2. 3. 3 7. 4. 4 1. 20. 00. –5. 6. 30 5. 2900. Key. Figure 1. 1 2 3 4. Fan Shield Hinged lid Removable wall channel sections. 5 Specimen 6 Wooden crib 7 Inspection cover. General view of the test equipment as given in ENV 1187. Apart from new figures additional technical information will also be needed for test method 2 of ENV 1187. The new text is gathered in Appendix B.. 1.

(11) 9. 2. The equipment for test method 2 of ENV 1187. 2.1. Background. The drawings referred to in NT FIRE 006 [4] are in some parts detailed enough and in other parts far too brief. Thus most of the instructions about the test specimen area and about the upper air channel are fully valid. For the lower air channel and for the blowing fan in the lower channel there has been too much freedom for design. Hence the brief instructions have resulted in different solutions to construct the test equipments used by the official fire laboratories and by roof covering producers. The original test apparatus, once designed at SP, was built in such a way that the air stream passing the test specimen became slightly turbulent. This turbulence is rather difficult to identify with the type of measuring instruments that normally are at hand by the fire testing laboratories. Complex airflow measurements could maybe solve the problem caused by the wide scope of constructing the apparatus according to the drawings in [4]. A more safe solution is instead suggested in this report: to prescribe more precisely how to build the equipment.. 2.2. New drawings and specifications. The new figures to be suggested for ENV 1187 are given in Appendix A. They are based on the drawings referred to in NT FIRE 006 [4]. The additions needed to the test specifications are presented in details in Appendix B. It should be noted that there is also a need for revising the definition of damaged material in EN 13501-5 [5]. Precise dimensions and details for the lower channel are given in the new figures. Thus it is prescribed that the entire lower channel from the fan to the opening at the test specimen must be at a fixed position relative the fan. It is also prescribed exactly how many perforated plates that must be mounted in the channel and how they must be designed. Finally, the fan shall be speed regulated for the adjustment of the wind speed. Today’s version of ENV 1187 prescribes that the wind shall be adjusted with a damper. The damper induces uncontrolled turbulence and a lot of boisterous noise. The reason for abandoning the possibility to use soft material and to allow the horizontal part of the lower air channel to be adjustable vertically vis-à-vis the fan is that it has been noticed that a non-fixed positioning of the channel and a varying shape of any flexible parts will induce very varying and uncontrolled turbulence of the airflow above the test specimen. Performance based instructions are given for those parts of the equipment where strict instructions are unnecessary. Thus the mounting of the specimen in the channel can be done in many different ways providing that the specimen finally ends up being mounted from under both the lower and the upper air channels. Options are also given to allow the loose wall channel section to be firmly and constantly mounted in the equipment instead of removing it between the tests. There will also be up to the user to decide whether a sound damper should be mounted on the air inlet. Finally, arrangements for extraction of the smoke through the upper air channel as well as extraction of smoke from the whole apparatus is made performance based..

(12) 10. 3. Evaluation of the new apparatus for test method 2 of ENV 1187. The evaluation of the new apparatus for test method 2 of ENV 1187 was performed in two steps: - The consistence in test results was first confirmed in the prototype apparatus at SP. - The improvement of the precision was confirmed by performing round robin tests among the four Nordic fire laboratories on equipments identical to the prototype.. 3.1. New definition of damaged area. During the evaluation of the first prototype at SP it was found that a new and precise definition of damaged area was necessary. Thus a new definition was given in the round robin protocol. The effect of this new definition on the evaluation of the prototype is further discussed below and in 3.2. The definition of damaged area given in ENV 1187 is: - “material that has been charred, melted or otherwise visually changed by heat. Discolouration and soot deposits are not to be regarded as damaged material”. New definition that was intermediately used for the round robin is: - “material that has been charred, melted or otherwise clearly damaged by heat. Areas that only has become shiny from the flames without being damaged are not to be regarded as damaged material, neither are discolouration and soot deposits”. The definition above was found to be very difficult to interpret even for the experienced operators at the Nordic fire laboratories. Since the definition of damaged area must be well established to avoid poor precision it was necessary to further develop the new definition of damaged area. The final suggestion is: - “material that has been charred, melted or otherwise clearly damaged by heat. For a product with protective surface this includes the area where the protective surface material has melted away. For a product without protective surface this includes the area where a damage of the surface from the combustion process has occurred. Areas that only have become shiny from the flames without being damaged by the combustion process or only contains tiny occasional holes from bubbles are not to be regarded as damaged material, neither are discolouration and soot deposits”. The three definitions are illustrated in Figure 2 and Figure 3. On bitumen roof coverings having a protective surface all three definitions give the same interpretation of damaged length. PVC-foil roof coverings show three different grades or zones of damage. The three definitions comply with each of the following zones: 1. The zone where the combustion process has occurred = The final suggestion for definition. The definition that was used for the final evaluation of the round robin. 2. The zone where the radiation from the flames have made the surface shiny = The ENV 1187 definition 3. The zone between 1 and 2 where the radiation has caused a deeper damage = The definition intermediately used for the round robin.

(13) 11. The dotted lines in Figure 3 show the zone of interpretation of the intermediate definition. The effect of this interpretation on the round robin results of the PVC-foil is given in Appendix C and the round robin results using the final definition are given in 3.3.2.. Figure 2. All three definitions of damaged material give only one evaluation of the test result when bitumen roof coverings are tested.. Figure 3. Three zones of different grades of damage on PVC-foil roof coverings that comply with the three definitions. The zones are further explained in the text..

(14) 12. 3.2. Confirmation of consistence of test results at SP. The test results at SP have been very close to the general mean of test results received at all the four Nordic fire laboratories in the round robins performed through out the years. It was therefore decided among these laboratories that the first prototype of the new apparatus should be built and evaluated at SP. The graphs in Figure 4 and Figure 5 show the results of the round robins performed between 1987 and 2001. The mean value received at SP is within ±35 mm from the great mean in all cases but one. The case when it was further away from the great mean is the bitumen used in the round robin performed 2001. This case is further discussed in 3.3.3.. 2 m/s, SP's result comp to General mean. Damaged length, mm. 600. s R upper/lower limits s r upper/lower limits. 500. General mean. 400. SP upper/lower limits. 300. SP mean. 200 100 0 0. 1 Foil-87. Figure 4. 2. 3. Foil-95. Bi-95. 4. 5. Foil-01. 6. Bi-01. Graphical presentation of the damaged length of foil and bitumen roof coverings tested at 2 m/s in round robins performed 1987 – 2001. SP’s mean value and standard deviation are compared to the general mean, sr, and sR.. 4 m/s, SP's result comp to General mean Damaged length, mm. 800. s R upper/lower limits. 700. s r upper/lower limits. 600. General mean. 500. SP upper/lower limits. 400. SP mean. 300 200 100 0 0. 1. 2. 3. 4. 5. 6. 7. Foil-87 Foil-95 Bi-95 Bi-99 Foil-01 Bi-01. Figure 5. Graphical presentation of the damaged length of foil and bitumen roof coverings tested at 4 m/s in round robins performed 1987 – 2001. SP’s mean value and standard deviation are compared to the general mean, sr, and sR..

(15) 13. To confirm that the mean of test results would be kept on the same level of fire performance, i.e. that products having passed the test in the old apparatus should still pass the test in the new apparatus and vice versa, a number of products were tested in the old apparatus at SP before it was rebuilt and also in the prototype. The products that were tested directly after rebuilding the apparatus were roof coverings of bitumen and plastic foil type. Since the apparatus also is used for testing of floor coverings according to NT FIRE 007 some floor coverings were tested as well. A sufficient number of test specimens were prepared so that tests on specimens from the same batches and prepared in an identical way could be performed in the old and the new apparatus. As is discussed in 3.1, the evaluation of the prototype showed that a new definition of damaged area was needed. The plastic foils are very sensitive to the slightly sharper flame profile that appears in the new apparatus. The sharper flames gave up to 120 mm longer damaged length in the prototype when evaluated according to the definition of damaged length given in ENV 1187. By not regarding “areas that only had become shiny from the flames without being damaged” the damaged length for the PVC-foil roof coverings in the new equipment was of approximately the same magnitude as in the old. The results complied very well with the results received in the old apparatus when the new definition of damaged area was used. Thus the average damaged length was kept very close to what it has been through out the years before the apparatus was rebuilt. The average of the difference in damaged length was within approximately ±10 mm and was randomly distributed around zero for those products taken from identical batches. In addition to the comparison performed in direct connection with the rebuilding of the apparatus, results from the supervisory control of roof and floor coverings have been studied. Also from these results it is concluded that the damaged length is very close to the level that has been recorded through out the years before the apparatus was rebuilt. Results from about 20 roof coverings and 25 floor coverings have been compared. The detailed results are listed in Appendix D. Although the results that are compared not at all are from the same batches the general mean of the difference in damaged length was less than 5 mm.. 3.2.1. Confirmation of the turbulence of the airflow. Before the actual testing on products was started in the prototype, several tests were performed on plain particleboard to confirm that the flow conditions and the turbulence characteristics in the prototype were as equal as possible as in the former apparatus. The experience has shown that more laminar airflow gives longer and sharper shape of the damage than a turbulent airflow does. Figure 6 and Figure 7 show the shape of the damaged area received on particleboard in the old and the new apparatus at SP. As can be seen the shape became rather similar in the prototype although a fully identical shape as in the old apparatus was not possible to achieve. The slight difference is due to the necessary reconstruction such as avoiding any soft material in the lower air channel discussed in 2.2. Figure 8 shows some examples of how more turbulent and more laminar airflow affects the shape of the damage..

(16) 14. Figure 6. The shape of the damaged area received on particleboard in the old SP apparatus at 2 m/s (left photo) and 4 m/s (right photo). Figure 7. The shape of the damaged area received on particleboard in the new SP apparatus at 2 m/s (left photo) and 4 m/s (right photo). Figure 8. Examples of how more turbulent airflow gives a truncated shape of the damage (left photo) while more laminar airflow gives a shaper shape of the damage (middle and right photo). The turbulence in the flow was also studied visually by using artificial smoke. The smoke was induced into the lower channel in the air inlet and was then video recorded. The final design of the new apparatus showed almost identical behaviour of the artificial smoke when passing above the specimen. To give a picture of the degree of turbulence in the test method the airflow was also evaluated in the new SP apparatus using a hot wire anemometer with a high resolution..

(17) 15. The relative turbulence intensity in percent (the standard deviation in relation to the mean velocity [6]) was found to be 1,5-3% at 2 m/s and 1-2% at 4 m/s. These values can be used as comparison between different test equipments and also to assure that no significant disturbances or changes occurs during time or when rebuilding the apparatus. Detailed results from the evaluation of the turbulence are presented in Appendix F.. 3.3. Confirmation of the improved precision of the test in the round robin. The precision of the new equipment was evaluated by means of a round robin performed between the four Nordic fire laboratories. The results received were more precise than results that have been received in former comparisons with the old equipment. The systematic deviations being seen between the laboratories are no longer at hand. (The comparison with former round robins is further presented in 3.3.3.). 3.3.1. Products tested in the round robin. Two types of roof coverings were tested: bitumen and PVC-foil. The roof coverings were cut in testing size before distributed to the participating laboratories. The test specimens from both roof coverings were cut in the perpendicular direction of the rolls from a set of rolls and grouped together in a randomised order before distribution. The roof coverings were nailed to standard non-combustible board, i.e. calcium silicate board by each of the participating laboratories before testing. Each laboratory used their standard silicate board.. 3.3.2. Statistical analysis of the round robin results. The round robin results were analysed statistically according to ISO 5725-2 [7] even though the round robin was not performed fully according to ISO 5725. (The standard suggests that more than five laboratories should participate to give a good estimation of the precision.) The repeatability and reproducibility standard deviations (sr and sR) as well as the repeatability and reproducibility coefficients of variation (COV(sr) and COV(sR)) were calculated. To evaluate the new definition against the definition in the standard two independent operators at each laboratory reported results. They measured the damaged length without making any marks were the damage was considered to end. Thus each of the operators was unaware of the result reported by the colleague. In some cases the operator had to interpret the definition on his/her own basis. This reflects how the case would be for a laboratory starting from scratch to use the test standard. After the test, the burnt specimens were sent to SP, where one specific operator measured the damaged length on all test specimens once again. This specific operator measured the specimens tested at SP as well. Thus three operators finally had measured the damaged length on each specimen. The result from each of these measurements was treated as one unique laboratory in the analysis, i.e. the four laboratories thus became 12 when the precision was calculated..

(18) 16. It very soon became clear that the intermediately used definition described in 3.1 would induce unnecessary errors. The definition was therefore further developed to the final suggestion and the PVC-foil specimens were again sent back to the participators for renewed measurements. This exercise confirmed that the final version of the definition to be suggested for the revision of ENV 1187 is as well established as ever the old definition was. More detailed results from the interpretation of the three definitions are given in Appendix C. The result of the final analysis of the round robin is given in Table 2. That the final definition of damaged material is as good established as the definition given in ENV 1187 is also visualized by the graphical presentation of the test results from the PVC-foil in Figure 9 and Figure 10. The graphs give the damaged length based on the final definition used in the round robin and on the old definition. As can be seen the agreement between the operators is very good for both definitions. The graphs also show that the new definition reduces the overall average value with 70 - 80 mm as was one of the aims of introducing a new definition.. Table 2. Precision values received in the new equipments, each specimen measured by three operators. General mean sr COV(sr) sR COV(sR). Fire spread results, averages, mm Bitumen on calcium silicate board 2 m/s 4 m/s 421 411 23 24 5% 6% 24 32 6% 8%. PVC-foil on calcium silicate board 2 m/s 4 m/s 233 300 23 20 10% 7% 28 20 12% 7%. 2 m/s, PVC-foil, new and old definition 600 New definition. Old definition. Damaged length, mm. 500 Lab 1. Lab 2. Lab 3. Lab 4. Lab 1. Lab 2. Lab 3. Lab 4. 400. Test 1 Test 2. 300. Test 3 200. General mean. 100. Figure 9. PVC-od, lab 4, SP. PVC-od, lab 4, op 2. PVC-od, lab 3, SP. PVC-od, lab 4, op 1. PVC-od, lab 3, op 2. PVC-od, lab 2, SP. PVC-od, lab 3, op 1. PVC-od, lab 2, op 2. PVC-od, lab 1, SP. PVC-od, lab 2, op 1. PVC-od, lab 1, op 2. PVC-nd, lab 4, SP. PVC-od, lab 1, op 1. PVC-nd, lab 4, op 2. PVC-nd, lab 3, SP. PVC-nd, lab 4, op 1. PVC-nd, lab 3, op 2. PVC-nd, lab 2, SP. PVC-nd, lab 3, op 1. PVC-nd, lab 2, op 2. PVC-nd, lab 1, SP. PVC-nd, lab 2, op 1. PVC-nd, lab 1, op 2. PVC-nd, lab 1, op 1. 0. Graphical presentation of the damaged length for the PVC-foil at 2 m/s when measured according to the final (new in the graph) and the old definition. The operators are given as op1, op2 and SP..

(19) 17. 4 m/s, PVC-foil, new and old definition 600 New definition. Old definition. Damaged length, mm. 500 Lab 1. Lab 2. Lab 3. Lab 4. Lab 1. Lab 2. Lab 3. Lab 4. Test 1. 400. Test 2 300. Test 3 General mean. 200 100. PVC-od, lab 4, SP. PVC-od, lab 4, op 2. PVC-od, lab 3, SP. PVC-od, lab 4, op 1. PVC-od, lab 3, op 2. PVC-od, lab 2, SP. PVC-od, lab 3, op 1. PVC-od, lab 2, op 2. PVC-od, lab 1, SP. PVC-od, lab 2, op 1. PVC-od, lab 1, op 2. PVC-nd, lab 4, SP. PVC-od, lab 1, op 1. PVC-nd, lab 4, op 2. PVC-nd, lab 3, SP. PVC-nd, lab 4, op 1. PVC-nd, lab 3, op 2. PVC-nd, lab 2, SP. PVC-nd, lab 3, op 1. PVC-nd, lab 2, op 2. PVC-nd, lab 1, SP. PVC-nd, lab 2, op 1. PVC-nd, lab 1, op 2. PVC-nd, lab 1, op 1. 0. Figure 10. Graphical presentation of the damaged length of the PVC-foil at 4 m/s when measured according to the final (new in the graph) and the old definition. The operators are given as op1, op2 and SP.. 3.3.3. Comparison with former round robins. The round robin performed in 2001 within the Nordtest project 1524-01 resulted in the precision values given in Table 3. The poor precision results for the bitumen roof covering tests at 4 m/s in that investigation were mainly due to the product, which had a tendency to sometimes keep on burning with small flames that slowly ascended up the specimen. This behaviour occurred on a random number of specimens thus giving the very high values of COV(sr) and COV(sR) for this product. More interesting for the precision of the method are the 4 m/s results for the PVC-foil. The COV(sR)-value (the reproducibility) for these results is much higher than the COV(sr)-value (the repeatability). This increase is a consequence of the systematic deviation between the laboratories at that air velocity. This deviation has been seen also in former round robins. Figure 11 shows the systematic deviation visually. As can be seen there has been a trend that lab 1 in most cases has received a longer damage and lab 4 shorter than the average when testing at 4 m/s. It has also been identified that the equipment at these laboratories was differently constructed compared to the others. Table 3. Precision values received in the round robin performed 2001. General mean sr COV(sr) sR COV(sR). Fire spread results, averages, mm Bitumen on mineral wool 2 m/s 4 m/s 466 574 33 123 7% 21% 37 157 8% 27%. PVC-foil on mineral wool 2 m/s 4 m/s 372 416 20 17 5% 4% 29 66 8% 16%.

(20) 18. 4 m/s, round robins 1995-2001 + new equipment. Test 1 Test 2. 800. 200 100. Lab 1 Lab 2 Lab 3 Lab 4. Lab 1 Lab 2 Lab 3 Lab 4. 300. Lab 1 Lab 2 Lab 3 Lab 4. 400. Lab 1 Lab 2 Lab 3 Lab 4. 500. Lab 1 Lab 2 Lab 3 Lab 4. Trend betw een labs. Lab 1 Lab 2 Lab 3 Lab 4. 600. General mean. Lab 1 Lab 2 Lab 3 Lab 4. Damaged length, mm. 700. Test 3. 1995 Bitumen. 1995 Foil. 1999 Bitumen. 2001 Bitumen. 2001 Foil. New eq Bitumen. New eq Foil. 0. Figure 11. Graphical presentation of the damaged length for products tested at 4 m/s in round robins performed 1995 - 2001 together with the results from the round robin in the new equipments. The blue lines show the systematic deviation between the laboratories that is further explained in the text.. 3.3.4. Air velocity measurements. In Nordtest project 1524-01 a comparison between the fire-spread results and the air velocity in the equipments was done to try to explain the differences in test results [3]. It was concluded that the air velocities were very diverse even if this did not fully explain the systematic deviation in damaged length between the laboratories. A renewed comparison of the air velocity in the equipments performed in this project showed that the air velocity precision was improved in the new equipments. In Table 4and Table 5 the maximum deviations of the air velocities at 2 and 4 m/s are compared. The present standard only requires that the air velocity shall be controlled at 2 m/s. Those requirements were fulfilled on the old equipments. When the reasons for the systematic deviations between the laboratories were investigated the wind profile was also measured along longitudinal lines on each side of the centre line. It was then found that the air velocity at 4 m/s deviated as much as 1,5 m/s from the nominal value in the old equipments. The air velocity in the new equipments does not fully fulfil the requirements of the present standard along the entire specimen. Thus, at S850 at 2 m/s the deviation is 0,2 m/s instead of the prescribed <0,1 m/s. This slight difference is due to the necessary reconstruction already discussed in 2.2 and 3.2.1. Table 4. Comparison of the min and max values of air velocities at 2 m/s from measurements at all laboratories in the new and old equipments. Dist, mm. New equipment Old equipment Min value Max value Min value -100 c +100 -100 c +100 -100 c +100 100 (S0) 1,9 1,9 1,9 2,0 2,0 2,1 2,0 2,0 1,7 600 (S500) 1,9 1,9 1,9 2,2 2,1 2,2 1,9 1,9 1,9 950 (S850) 1,7 2,0 1,7 2,3 2,2 2,3 2,0 2,0 2,0 Note: “-100” and “+100” refers to 100 mm away from the centre line. Max value -100 c 2,9 2,1 2,6 2,1 2,6 2,1. +100 2,6 2,5 2,6.

(21) 19. Table 5. Comparison of the min and max values of air velocities at 4 m/s from measurements at all laboratories in the new and old equipments. Dist, mm. New equipment Old equipment Min value Max value Min value -100 c +100 -100 c +100 -100 c +100 100 (S0) 4,0 4,0 3,9 4,1 4,1 4,1 3,6 4,0 3,1 600 (S500) 3,6 3,7 3,6 4,5 4,3 4,4 3,1 3,4 3,1 950 (S850) 3,2 3,3 3,2 3,7 3,5 3,6 2,9 3,1 2,9 Note: “-100” and “+100” refers to 100 mm away from the centre line. Max value -100 c 5,5 4,3 5,0 4,2 4,7 4,1. +100 5,1 5,0 4,7. As a result of this investigation the revision and amendment of the standard will include new requirements for the air velocity (see Appendix B). The allowed deviation at the uppermost measuring point at 2 m/s will be increased. It will also be suggested that the position of this point is moved 50 mm to S800 since the bottom plate of the upper air channel hides the point S850. Finally, additional requirements for the air velocity will be given for longitudinal lines on each side of the centre line and also for 4 m/s were there have been no requirements at all. Graphical presentation and detailed description of how the air velocity measurements were performed is given in Appendix E..

(22) 20. 4. Conclusions. -. The design of the test equipment for test method 2 of ENV 1187 has been improved. The design is thoroughly described by new drawings and detailed prescription on how to build the apparatus.. -. The repeatability and reproducibility was improved by the new design, which was confirmed through tests performed among the Nordic fire laboratories.. -. A comparison of test results from about 60 products tested in the old and in the improved equipment at SP showed no deviation of test results between the old and the improved design. This confirms that the overall Nordic average of test results will remain on the same level.. -. A precise definition of damaged length has been developed. This does also improve the repeatability and reproducibility.. It is therefore recommended that the results of this project (drawings and more precise specifications for the equipment, detailed requirements of air velocity, and precise definition of damaged length) are introduced into ENV 1187, test method 2..

(23) 21. 5. References. 1. European Pre-standard – Test methods for external fire exposure to roofs. ENV 1187:2002 (E), CEN Brussels (2002).. 2. Roofings: Fire spread, Nordtest method NT FIRE 006, NORDTEST, Helsinki, 1985.. 3. I. Wetterlund, Examination of NT FIRE 006 -“Roofings: Fire spread”. Nordtest Project No. 1524-01, SP Technical Notes 2001:42, rev1, Borås 2001.. 4. Statens Provningsanstalt, NT FIRE 006/007, drawings no 342-002/1-9, Borås 1980.. 5. European Draft Standard – Fire classification of construction products and building elements – Part 5: Classification using data from external fire exposure to roof tests. prEN 13501-5:2003 (E). Brussels (2003).. 6. Goldstein, R. J. (editor), Fluid Mechanics Measurements (second edition), Taylor & Francis (1996).. 7. Accuracy (trueness and precision) of measurement methods and results  Part 2: Basic method for the determination of repeatability and reproducibility of a standard measurement method, ISO 5725-2:1994(E), ISO, Geneva, 1994..

(24) 22. Appendix A New drawings for method 2 of ENV 1187 The following drawings will be suggested for a revised version of test method 2 of ENV 1187: Figure 12 General assembly of test method 2 of ENV 1187 Figure 13 Detail drawing A - Lower air channel, side and top view Figure 14 Detail drawing B - Lower air channel, end view Figure 15 Detail drawing C - Perforated plate in lower air channel Figure 16 Detail drawing D - Lid section Figure 17 Detail drawing E - Narrow pass of upper air channel.

(25) 23. Figure 12. General assembly of test method 2 of ENV 1187.

(26) 24. Figure 13. Detail drawing A - Lower air channel, side and top view.

(27) 25. Figure 14. Detail drawing B - Lower air channel, end view (test specimen and side walls not shown). Figure 15. Detail drawing C - Perforated plate in lower air channel.

(28) 26. Figure 16. Detail drawing D - Lid section. Figure 17. Detail drawing E - Narrow pass of upper air channel.

(29) 27. Appendix B Additional/revised specifications for method 2 of ENV 1187 The below listed additional/revised specifications will be suggested for a revision of test method 2 of ENV 1187. It should be noted that the numbering of figures and tables need to be adapted when they are introduced into ENV 1187. The figure numbers in the text refer to the figures in Appendix A. The table numbers refer to tables in this Appendix. Definition 3.3 damaged material should be changed to: material that has been charred, melted or otherwise clearly damaged by heat. For a product with protective surface this includes the area where the protective surface material has melted away. For a product without protective surface this includes the area where a damage of the surface from the combustion process has occurred. Areas that only have become shiny from the flames without being damaged by the combustion process or only contains tiny occasional holes from bubbles are not to be regarded as damaged material, neither are discolouration and soot deposits. Clause 5.1.1 should be changed to: The equipment shall be built according to Figure 12 to Figure 17. The dimensions given are nominal unless tolerances are given. The air channels shall be made of 1 mm thick steel plates. The wall channel section placed between the lower and the upper air channels shall be made of steel plates and non-combustible fibre reinforced calcium silicate boards with a thickness of (11 ± 2) mm and a dry density of (680 ± 50) kg/m3. A lid, which is hinged to the lower end of the upper air channel, covers the top of this wall channel section (Figure 16). The wall channel section may be firmly mounted between the channels or removable as long as it is assured that the upper surface of the specimen can be mounted well in contact with the underside of the bottom of the lower and upper air channels. The lower air channel shall be built according to Figure 13 to Figure 15. No flexible material is allowed in the connections. The connection between the fan and the channel shall be equipped with a perforated steel plate in the upper half of the opening and a steel plate without perforation in the lower half of the opening. A pleated perforated steel plate shall be mounted in the lower air channel according to Figure 13 and Figure 15. The fan shall be of the centrifugal type and have a single air-inlet . The blades of the fan shall be backward curved. The wheel diameter shall be 280mm. The upper air channel shall be connected to the exhaust system in such a way that the air velocity in the narrow pass (Figure 17) can be kept according to clause 5.2.1.1. There shall be arrangements for applying and securing the test specimen in its test position under the air channels (no 2 in Figure 12). This can be made by means of a sliding platform and pneumatic lifts. There shall be a closable lid on the air inlet channel (no 3 in Figure 12). The channel may be equipped with a sound damper..

(30) 28. A note should be added to clause 5.1.1: Further information about the fan is available at SP Fire Technology, Borås, Sweden. Clause 5.2.1.1 should be changed to: The basic calibration shall be carried out following the installation of the apparatus and whenever changes occur which could affect the performance of the apparatus. a) With the dummy specimen (see 5.2.1.1 b) properly inserted in the specimen holder (see 5.7.1) and with the apparatus running and properly adjusted in accordance with 5.2.1.2 the air velocity on the dummy specimen as measured with the vane wheel anemometer (see 5.2.1.2a) in the direction of the flow shall be within the limits given in Table 6 and Table 7. The air velocity at the central line of the narrow pass (60 ± 2) mm in the upper air channel (see 5.2.1.2a) shall be (6,0 ± 0,5) m/s. Furthermore the air velocities shall be verified with the hot-wire anemometer mounted through the left and right anemometer probe insertion holes, respectively. These are located in the bottom of the channel. Table 6. Allowed deviation from nominal air velocity at 2 m/s. Lengthwise position (see 5.2.1.1b) S0 S500 S800 Table 7. Allowed deviation, m/s At centre line At lines 100 mm from centre line ±0,1 ±0,1 ±0,1 ±0,2 ±0,2 ±0,3. Allowed deviation from nominal air velocity at 4 m/s. Lengthwise position (see 5.2.1.1b) S0 S500 S800. Allowed deviation, m/s At centre line At lines 100 mm from centre line ±0,1 ±0,1 ±0,3 ±0,5 ±0,5 ±0,8. b) A dummy specimen shall be made of an ordinary particleboard with a thickness of (19 ± 2) mm and density of (680 ± 50) kg/m3 at normal conditioning atmosphere (see 5.5.2) and cut to the size of the specimen (400 mm x 1000 mm). It shall be plane and marked at the following points along the centre line and along longitudinal lines on each side of and 100 mm from the centre line. S0: 100 mm from the bottom edge S500: 600 mm from the bottom edge S800: 900 mm from the bottom edge. Clause 5.6 needs an additional paragraph: The test apparatus may be mounted in a fume cupboard. The extraction of fumes shall be provided in such a way that the conditions given in 5.2 can be kept..

(31) 29. Clause 5.7.1 should be changed to: The specimen is mounted in test position so that the upper surface of the specimen is in contact with the underside of the bottom of the lower and upper air channels. Care shall be taken so that the joint between the lower edge of the specimen and the bottom of the lower air channel is airtight. In practice, the bottom plate of the lower air channel should overlap the lower end of the specimen by 2 mm to 3 mm. In some cases it may be necessary to pack air gaps with mineral wool. The position where the burning wood crib is to be placed shall be marked on the specimen. Clause 5.7.3 should be changed to: The wall channel section (if it is removable) is placed on the specimen and the lid is closed. The air velocity is controlled (see 5.2.1.2)..

(32) 30. Appendix C The interpretation of the definitions of damaged length When the results of the round robin for the PVC-foil were analysed it very soon became clear that the intermediately used definition described in 3.1 was very difficult to understand even among the experienced operators at the Nordic fire laboratories. The deviating interpretation of the damaged length between the operators is visualized in Figure 18 and Figure 19. In the graphs the damaged length based on the final definition to be suggested for the revision of ENV 1187 and on the intermediate definition used in the round robin are compared.. 2 m/s, PVC-foil, final and intermediate definition 600 Final definition. Intermediate definition. Damaged length, mm. 500 Lab 1. 400. Lab 2. Lab 3. Lab 4. Lab 2. Lab 1. Lab 3. Lab 4. Test 1 300. Test 2 Test 3. 200 100. Figure 18. PVC-def2, lab 4, SP. PVC-def2, lab 4, op 2. PVC-def2, lab 3, SP. PVC-def2, lab 4, op 1. PVC-def2, lab 3, op 2. PVC-def2, lab 2, SP. PVC-def2, lab 3, op 1. PVC-def2, lab 2, op 2. PVC-def2, lab 1, SP. PVC-def2, lab 2, op 1. PVC-def2, lab 1, op 2. PVC-fin, lab 4, SP. PVC-def2, lab 1, op 1. PVC-fin, lab 4, op 2. PVC-fin, lab 3, SP. PVC-fin, lab 4, op 1. PVC-fin, lab 3, op 2. PVC-fin, lab 2, SP. PVC-fin, lab 3, op 1. PVC-fin, lab 2, op 2. PVC-fin, lab 1, SP. PVC-fin, lab 2, op 1. PVC-fin, lab 1, op 2. PVC-fin, lab 1, op 1. 0. Graphical presentation of the damaged length for the PVC-foil at 2 m/s when measured according to final and intermediate definition. The operators are given as op1, op2 and SP..

(33) 31. 4 m/s, PVC-foil, final and intermediate definition 600 Final definition. Intermediate definition. Damaged length, mm. 500 Lab 1. 400. Lab 2. Lab 3. Lab 4. Lab 2. Lab 1. Lab 3. Lab 4. Test 1 300. Test 2 Test 3. 200 100. Figure 19. PVC-def2, lab 4, SP. PVC-def2, lab 4, op 2. PVC-def2, lab 3, SP. PVC-def2, lab 4, op 1. PVC-def2, lab 3, op 2. PVC-def2, lab 2, SP. PVC-def2, lab 3, op 1. PVC-def2, lab 2, op 2. PVC-def2, lab 1, SP. PVC-def2, lab 2, op 1. PVC-def2, lab 1, op 2. PVC-fin, lab 4, SP. PVC-def2, lab 1, op 1. PVC-fin, lab 4, op 2. PVC-fin, lab 3, SP. PVC-fin, lab 4, op 1. PVC-fin, lab 3, op 2. PVC-fin, lab 2, SP. PVC-fin, lab 3, op 1. PVC-fin, lab 2, op 2. PVC-fin, lab 1, SP. PVC-fin, lab 2, op 1. PVC-fin, lab 1, op 2. PVC-fin, lab 1, op 1. 0. Graphical presentation of the damaged length of the PVC-foil at 4 m/s when measured according to new and old definition. The operators are given as op1, op2 and SP.. The effect of the different interpretation of the definition on the precision values was also calculated. The precision results of the measurements performed on the PVC-foil according to all three definitions are given in Table 8. Table 8. Precision values from measurements on the PVC-foil specimens, each specimen measured by three operators. General mean sr COV(sr) sR COV(sR). Fire spread results, averages, mm PVC-foil on calcium PVC-foil on calcium silicate board, old silicate board, definition intermediate definition 2 m/s 4 m/s 2 m/s 4 m/s 315 370 265 323 23 23 25 25 7% 6% 10% 8% 30 23 51 36 9% 6% 19% 11%. PVC-foil on calcium silicate board, final definition 2 m/s 4 m/s 233 300 23 20 10% 7% 28 20 12% 7%.

(34) 32. Appendix D Comparison of results from supervisory control Results from the supervisory control of roof and floor coverings performed before and after the apparatus at SP was rebuilt are compared in Table 9. The comparison includes all products that have been controlled from the time the new apparatus at SP was installed until September 2004. To avoid that any results can be identified to a certain product the information about the product type has been left out as well as information about air velocity. The results from the tests in the old apparatus were collected during 1999 - 2003. The standard deviation is given for those products where more than single results were included in the comparison. The high standard deviation for some of the products indicates that those products are not very stable in their results. Thus, fail results were occasionally noted for some of the products; in some cases the product had burnt as much as 900 mm at one testing occasion. Since the results were collected during up to four years the production might also have been altered during the period. In any case, none of the results for any of the products listed were from the same batch. For a comparison of this type it is therefore not possible to focus on the individual test result for each of the products. Due to the reasons discussed above products may display very different results from time to time. Thus the only values that should be compared are the general mean values of the average values. The general mean value of the standard deviation values has also been calculated to indicate whether the spread in results has increased or decreased. As can be seen the difference of the general mean value of the damaged length is within less than 5 mm for the two cases. The spread in results has decreased. Table 9. Results from the supervisory control of roof and floor coverings. Result Damaged length in old no apparatus, mm Average or Standard single result* deviation 1 508 2 380 3 459 4 611 95** 5 522 6 272 7 466 8 411 9 361 10 415 62 11 341 12 381 1 13 483 18 14 359 15 357 *Explanation on next page. Damaged length in new Comment apparatus, mm Average or Standard single result* deviation 544 335 424 509 **Fail results are included 481 225 525 457 497 480 423 339 510 21 509 20 314 -.

(35) 33. Table 9, continued Result no. Damaged length in old apparatus, mm Average or Standard single result* deviation. Damaged length in new apparatus, mm Average or Standard single result* deviation. Comment. 16 605 418** 359 **Fail results are included 17 335 372 18 433 395 19 307 368 20 507 435 21 425 522 22 443 462 23 422 400 24 545 492 25 452 412 17 26 541 58** 475 2 **Fail results are included 27 425 440 28 536 46** 473 **Fail results are included 29 368 333 30 341 14 300 31 316 64 333 32 329 13 350 33 360 16 344 34 310 23 342 35 345 34 340 11 36 369 28 361 37 382 367 38 316 310 39 395 367 40 434 59 391 41 512 93** 397 **Fail results are included 42 216 25 300 20 43 285 15 377 69 44 188 5 232 35 45 192 26 295 52 46 53 9 68 4 47 469 42 448 75 48 275 117 322 7 49 491 70** 469 1 **Fail results are included 50 518 93** 478 65 **Fail results are included 51 222 47 303 52 384 91 544 53 418 98 406 117 54 372 67 370 General mean 392 50 395 34 * “Average” refers to the average of the mean values of damaged length reported at each control occasion. “Single value” refers to the mean value from a single control occasion..

(36) 34. Appendix E The air velocity in the equipments used for ENV 1187 - method 2 The air velocity in the test equipments was measured according to an expanded procedure based on the prescriptions given in clause 5.2.1.1 in ENV 1187. Thus the air velocity was measured along three longitudinal lines at the positions S0, S500, and S850. The standard only prescribes measurements along the centre line. The point at 100 mm from the lower end, S0, is the control point for the daily verification of the air velocity. The three longitudinal lines were (looking from the lower fan): the centre line and lines on each side of the centre line. The line to the left of the centre line, is referred to as the “–100 mm line” and the line to the right “+100 mm line” in the reporting. In addition to the measuring points on the test specimen, the velocity was also measured at points in the lower air channel. The measuring points there were located 80 mm from the front edge of the channel. These points were also located at the centre line and at the “-100 mm” and “+100 mm” lines. The air velocity is reported to be at “-80 mm”. The test specimen during air velocity measurements was a non-burning particleboard. The results are presented as graphs for each air velocity and for each longitudinal line. The results from the new equipments are presented together with the results from the old equipments for each of the velocities and longitudinal lines. The air velocity measurements at 2 m/s are presented in Figure 20 - Figure 22, and the 4 m/s measurements are presented in Figure 23 - Figure 25. It should be noted that the graphs from the air velocity measurements performed in the old equipments are taken directly from [3]. The order of the laboratories is therefore not the same as for the measurements in the new equipments. Thus laboratory 1 etc. does not necessarily mean the same laboratory for both graphs. The comparison here only aims to show that the overall precision of the air velocity has been improved..

(37) 35. Air velocity, m/s. 2 m/s, along -100 mm line 3,5 3,0 2,5 2,0 1,5 1,0 0,5 0,0. Lab 1 Lab 2 Lab 3 Lab 4. -200. 0. 200. 400. 600. 800. 1000. 1200. Distance along test specimen, mm. Air velocity, m/s. 2 m/s, along -100 mm line 3,5 3 2,5 2 1,5 1 0,5 0. Lab 1 Lab 2 Lab 3 Lab 4. -200. 0. 200. 400. 600. 800. 1000. Distance along test specimen, mm. Figure 20. Air velocity at 2 m/s along the longitudinal line "–100 mm" (to the left of the centre line). Upper graph = new equipment. Lower graph = old equipment..

(38) 36. Air velocity, m/s. 2 m/s, along centre line 3,5 3,0 2,5 2,0 1,5 1,0 0,5 0,0 -200. Lab 1 Lab 2 Lab 3 Lab 4 Contr point 0. 200. 400. 600. 800. 1000. 1200. Distance along test specimen, mm. Air velocity, m/s. 2 m/s, along centre line 3,5 3 2,5 2 1,5 1 0,5 0 -200. Lab 1 Lab 2 Lab 3 Lab 4 Contr point 0. 200. 400. 600. 800. 1000. Distance along test specimen, mm. Figure 21. Air velocity at 2 m/s along the longitudinal centre line. The daily verification of the air velocity is performed at the point denoted as “Contr point”. Upper graph = new equipment. Lower graph = old equipment..

(39) 37. Air velocity, m/s. 2 m/s, along +100 mm line 3,5 3,0 2,5 2,0 1,5 1,0 0,5 0,0 -200. Lab 1 Lab 2 Lab 3 Lab 4. 0. 200. 400. 600. 800. 1000. 1200. Distance along test specimen, mm. Air velocity, m/s. 2 m/s, along +100 mm line 3,5 3 2,5 2 1,5 1 0,5 0 -200. Lab 1 Lab 2 Lab 3 Lab 4. 0. 200. 400. 600. 800. 1000. Distance along test specimen, mm. Figure 22. Air velocity at 2 m/s along the longitudinal line "+100 mm" (to the right of the centre line). Upper graph = new equipment. Lower graph = old equipment..

(40) 38. Air velocity, m/s. 4 m/s, along -100 mm line 7,0 6,5 6,0 5,5 5,0 4,5 4,0 3,5 3,0 2,5. Lab 1 Lab 2 Lab 3 Lab 4. -200. 0. 200. 400. 600. 800. 1000. 1200. Distance along test specimen, mm. Air velocity, m/s. 4 m/s, along -100 mm line 7 6,5 6 5,5 5 4,5 4 3,5 3 2,5. Lab 1 Lab 2 Lab 3 Lab 4. -200. 0. 200. 400. 600. 800. 1000. Distance along test specimen, mm. Figure 23. Air velocity at 4 m/s along the longitudinal line "–100 mm" (to the left of the centre line). Upper graph = new equipment. Lower graph = old equipment..

(41) 39. Air velocity, m/s. 4 m/s, along centre line 7,0 6,5 6,0 5,5 5,0 4,5 4,0 3,5 3,0 2,5 -200. Lab 1 Lab 2 Lab 3 Lab 4 Contr point. 0. 200. 400. 600. 800. 1000. 1200. Distance along test specimen, mm. Air velocity, m/s. 4 m/s, along centre line 7 6,5 6 5,5 5 4,5 4 3,5 3 2,5 -200. Lab 1 Lab 2 Lab 3 Lab 4 Contr point. 0. 200. 400. 600. 800. 1000. Distance along test specimen, mm. Figure 24. Air velocity at 4 m/s along the longitudinal centre line. The daily verification of the air velocity is performed at the point denoted as “Contr point”. Upper graph = new equipment. Lower graph = old equipment..

(42) 40. Air velocity, m/s. 4 m/s, along +100 mm line 7,0 6,5 6,0 5,5 5,0 4,5 4,0 3,5 3,0 2,5 -200. Lab 1 Lab 2 Lab 3 Lab 4. 0. 200. 400. 600. 800. 1000. 1200. Distance along test specimen, mm. Air velocity, m/s. 4 m/s, along +100 mm line 7 6,5 6 5,5 5 4,5 4 3,5 3 2,5 -200. Lab 1 Lab 2 Lab 3 Lab 4. 0. 200. 400. 600. 800. 1000. Distance along test specimen, mm. Figure 25. Air velocity at 4 m/s along the longitudinal line "+100 mm" (to the right of the centre line). Upper graph = new equipment. Lower graph = old equipment..

(43) 41. Appendix F Turbulence intensity measurements in the new test apparatus at SP used for ENV 1187 - method 2 To give a picture of the degree of turbulence in the test method the airflow was also evaluated in the new SP apparatus using a hot wire anemometer with a high resolution. A hot wire anemometer type Swema Air 300 with a resolution of ≤ 0,03 m/s in the velocity range 0,05-1,0 m/s and ≤ 3 % in the velocity range 1,0-30 m/s was used to capture the velocity field in the test channel. 100 records with an interval of 0,5 s were taken at each measuring point. The velocity was measured in the area near the crib position along the centre line at 100, 200, and 300 mm from the bottom end of the specimen and at three different heights above the specimen. Values were also taken at the same heights at positions 100 mm to the left and to the right of the crib position (at 100 mm from the bottom end). The heights above the specimen was chosen to represent three levels of the vane-wheel anemometer that is prescribed to be used for measuring the air velocity over the specimen. The levels were at the centre, and at the lower and the upper wings of the 100 mm diameter of the vanewheel. The lower measuring point had to be at 15 mm above the specimen since the sensing point of the Swema Air 300 anemometer is positioned 15 mm from the end of the probe. Thus the upper measuring point was similarly chosen to be at 15 mm from the upper edge of the wings or 35 mm above the centre of the vane-wheel. The relative turbulence intensity in percent (the standard deviation in relation to the mean velocity [6]) is presented separately for each level above the specimen and for each air velocity. The air velocity measurements at 2 m/s are presented in Table 10 - Table 12 and the 4 m/s measurements are presented in Table 13 - Table 15.. Table 10. Turbulence intensity at 15 mm above specimen at 2 m/s. Distance from lower end of specimen, mm 100 200 300 Table 11. Relative turbulence intensity, % At the centre line At the +100 line 2,5 2,9 2,0 1,8 -. Turbulence intensity at 50 mm above specimen at 2 m/s. Distance from lower end of specimen, mm 100 200 300 Table 12. At the -100 line 2,2 -. At the -100 line 1,7 -. Relative turbulence intensity, % At the centre line At the +100 line 2,3 2,6 1,9 2,1 -. Turbulence intensity at 85 mm above specimen at 2 m/s. Distance from lower end of specimen, mm 100 200 300. Relative turbulence intensity, % At the -100 line At the centre line At the +100 line 1,9 1,6 1,9 1,5 1,7 -.

(44) 42. Table 13. Turbulence intensity at 15 mm above specimen at 4 m/s. Distance from lower end of specimen, mm 100 200 300 Table 14. Relative turbulence intensity, % At the centre line At the +100 line 1,7 2,2 1,5 1,3 -. Turbulence intensity at 50 mm above specimen at 4 m/s. Distance from lower end of specimen, mm 100 200 300 Table 15. At the -100 line 1,6 -. At the -100 line 1,5 -. Relative turbulence intensity, % At the centre line At the +100 line 2,2 2,2 1,3 1,4 -. Turbulence intensity at 85 mm above specimen at 4 m/s. Distance from lower end of specimen, mm 100 200 300. Relative turbulence intensity, % At the -100 line At the centre line At the +100 line 1,4 1,4 1,4 1,1 1,2 -.

(45) 43. Appendix G Results of the simple rebuilding of the old equipments During the first year of the project an investigation regarding a simple rebuilding of the equipments to try to improve the conformity between the laboratories was performed. Two additional perforated plates were inserted in the lower channel. The fire spread on particleboard at 4 m/s was used to study if the extra plates would give any improvement. As discussed in 3.3.3 it had been noted in former comparisons that laboratory 1 often received longer damage and laboratory 4 shorter than the average when testing foil products at this air velocity. As can be seen from the results denoted “normal equipment” in Figure 26 this fact is also obvious for particleboard. (The results are taken from Nordtest project 1524-01 reported in [3].) Figure 26 also shows the results after the simple rebuilding. The results are presented for each of the laboratories as the average of the change in damaged length when extra plates were inserted. As can be seen a simple rebuilding would not have solved the problem. To provide the desired improvement of the precision the flame spread at laboratory 1 needed to be much shorter and the flame spread of laboratory 4 needed to be much longer. The figure shows that the damaged length received at laboratory 1 was reduced all right, but this reduction is in fact an illusion. The capacity of their lower fan was far too low to even come close to producing 4 m/s when the extra plates were inserted; thus the air velocity was only 2,4 m/s when their “4 m/s tests” were performed. In fact none of the laboratories managed to set the air velocity to 4 m/s with the extra plates. For laboratory 4 which needed a longer flame spread to be in line with the Nordic average, the damaged length was reduced even more when the extra plates were inserted. Thus the conclusion was to improve the apparatus design..

(46) 44. WoodPart 300. 250. Damaged length, mm. 200. 150. Normal equipment, 2 m/s. Normal equipment, 4 m/s. 100 Simple rebuilding, two extra plates Results at "4" m/s. 50 Overall average standard deviation. 0. -50. Figure 26. Lab 4, 3.3 m/s. Lab 3, 4 m/s. Lab 2, 3.7 m/s. Lab 1, 2.4 m/s. Lab 4, 4, 2001. Lab 3, 4, 2001. Lab 2, 4, 2001. Lab 1, 4, 2001. Lab 4, 2, 2001. Lab 3, 2, 2001. Lab 2, 2, 2001. Lab 1, 2, 2001. -100. Damaged length for particleboard. The results received before adding extra plates are given as absolute damaged length. The results after adding extra plates are given as change in damaged length..

(47)

(48) SP Swedish National Testing and Research Institute develops and transfers technology for improving competitiveness and quality in industry, and for safety, conservation of resources and good environment in society as a whole. With Swedens widest and most sophisticated range of equipment and expertise for technical investigation, measurement, testing and certfication, we perform research and development in close liaison with universities, institutes of technology and international partners. SP is a EU-notified body and accredited test laboratory. Our headquarters are in Borås, in the west part of Sweden.. SP Fire Technology SP REPORT 2004:25 ISBN 91-7848-999-7 ISSN 0284-5172 NT Techn Report 563. SP Swedish National Testing and Research Institute Box 857 SE-501 15 BORÅS, SWEDEN Telephone: + 46 33 16 50 00, Telefax: +46 33 13 55 02 E-mail: info@sp.se, Internet: www.sp.se.

(49)

No results found