SPRAYED "IN PLACE" polyurethane foam (PUF) roof systems have been

Transcription

1 Validation of the Polyurethane Foam Roof System D. KASHIWAGI US.A.F HG Air Force Ellgineering & Services Center Bldg. 22 Tyndall AFB, FL w. C. MOOR Arizona State University ASU School oj Engineering Dept. oj Industrial Engineering Tempe, AZ INTRODUCTION SPRAYED "IN PLACE" polyurethane foam (PUF) roof systems have been installed since the early 1960's. The lack of clear systems definition and requirements based on field performance of "successful" (20 year service life) PUF roofs, and an unbiased database ofpuf roof failures and successes have resulted in the continuing unpredictability of PUF roof performance. Coultrap's specification for PUF roof application [3], Alumbaugh, Keeton, Zarate, and Humm's work on experimental systems [2,9], and Jenkin's study of interrelationships between PUF roof system components [5], establishes a foundation for an initial definition of the PUF roof system. OBJECTIVE OF THE STUDY The objective of this study is to define the PUF roof system, performance This paper was presented at The Society of the Plastics Industry, Inc., (SPI) 30th Annual Technicall Marketing Conference, Toronto, Canada, October 15-17, The paper is being published herein from the conference proceedings after review by the Editorial Board,.,.:.\.:.' :torlary peer review process. JOURNAL OF THERMAL INSULATION Volume 1O-0ctober /86/ $ Technomic Publishing Co., Inc.

2 92 D. KASHIWAGI AND W. C. MOOR requirements and component requirements. The results will be used to define the PUF roof system performance criteria. Lastly, "rules of thumb" installation and priorities, procedures, and specification requirements will be covered. POLYURETHANE FOAM ROOF SYSTEM DESIGN, SCOPE AND ASSUMPTIONS The roof system is defmed as an economical, 20 year service life, low maintenance, sprayed "in-place" PUF roof system. It is generally accepted that a system with two inches of PUF has a greater probability of success than one with one inch of PUF [4]. There are two ways to design a "successful" system. The first is when there is a complete understanding of all factors affecting the performance of the system in a defined environment. The second approach is required due to a limited knowledge of the system operation in its environment. This methodology defmes the factors affecting the system as completely as possible, and supports it with experimental results, or in this case with data of existing "successful" PUF roof systems. SYSTEM PERFORMANCE MATERIALS INSTALLATION PROCEDURES D FIGURE 1. "Successful" PUF roof system. 20 YEAR PUF ROOF SYSTEM LESS THAN 20 YEAR PUF ROOF SYSTEM

3 Validatioll of the Polyurethane Foalll Roof System 93 There are three factors required to ensure a "successful" PUF roof system: 1. Materials to build the system components. 2. Installation procedure. 3. Performance of the system in a defined environment. All three components are present in a "successful" PUF roof system (Figure 1), but proven service life is the only factor that can independently identify the "successful" roof system. If a PUF roof system is consistently having a service life of more than ten years and a potential of 20 years barring unforeseen "acts of God;' it can be deduced that PUF roof system is a "successful" roof system that has a definable installation procedure and material defmition in a defmed environment. On the other hand, materials may have a service life of thirty years, but when installed as part of a system, the system may have a service life of only 5 years. Also, systems which are "successful" in a restricted environment, cannot be defined as "successful" roof systems in all environments. After defining the PUF system and component requirements, and validating it with inspection data and classical engineering theories, application "rule of thumb" principles and specification requirements will be covered. DEFINITION OF THE PUF ROOF SYSTEM The PUF roof system is made up of three components: PUF, a cover, and a vapor barrier. The most important component of the system is the PUF. The purpose of the PUF is to: 1. Form a closed cell waterproofmg barrier to protect the facility against water leakage. 2. Provide insulation. PUF is the best insulating material in the construction industry (R-value of 6.25 per inch). The seamless, continuous construction also increases the insulating quality of the roof. In a "successful" system, the PUF is required to maintain its waterproofing integrity for 20 years. Problems previously identified with the PUF are: 1. Condensation of moisture in the top layers of the PUF [5]. 2. Freezing of moisture laden PUF cells causing a loss of insulation, structural, and waterproofing capability [5]. 3. Improper cell formation [3,5]. 4. Disbonding of the PUF layers due to inadequate thicknesses, presence of moisture between passes, or "burnt" PUF surfaces. 5. Improper PUF structural capability.

4 94 D. KASHIWAGI AND W. C. MOOR The second required component of the PUF roof system is the cover. Its requirements are: 1. Protect the PUF from ultraviolet rays. 2. Distribute point loadings out over a wide area to not exceed the structural requirements on the PUF. The last component of the PUF roof system is the vapor barrier. The purpose of the vapor barrier is to prevent moisture from entering the PUF if there is a possibility that the moisture will condense and saturate the PUF. INSPECTION OF PUF ROOF SYSTEMS The objective of the roof inspections and customer satisfaction survey questions, and other data collection is to: 1. Identify "successful" PUF roof systems. 2. Identify causes of failure of systems. 3. Determine whether the "successful" PUF roof system can be installed consistently. 4. Track the performance of the system as the system ages. 5. Record maintenance requirements. The system designed to quantify the inspection data for analysis included a roof rating and customer satisfaction questions. Each roof inspected (except for asphalt-coated PUF roofs which are listed under the AS column in Table 1) was given a roof condition rating. A rating system devised by Alumbaugh, Humm, and Keeton, slightly modified to allow for quantitive measurements, was utilized. The roof condition ratings in Table 1 were used solely to classify the roof's condition in this report. Failure of a roof system is currently a subjective determination in the roofmg industry. For example, a roof classified as "poor" would exhibit between 10 percent and 20 percent deterioration of its PUF and elastomeric coating. Deterioration was defined Table 7. PUF roof rating system. Rating E Excellent' VG Very Good G Good FR Fair P Poor F Fail % of PUF System Requiring Replacement O=X 0<X~1 1 < X ~ 5 5 < X ~ < X ~ ~ X * Excellent roofs have no deteriorotion. If a roof has any defects, it is classified as a VG roof.

5 Valiilatlclll of the Polyurethane Foalll Roof System 95 as areas containing coating of PUF requiring removal and replacement due to external influences (i.e., ultraviolet degradation, mechanical damage, etc.) to provide the PUF roof system with a 20 year service life. The percentage of deterioration of a roof was determined by measuring areas of deterioration on a roof system. The minimum area of deterioration for the study was fixed at one square foot. Roofs which failed due to structural movement or "acts of God" were not counted as failures. Roofs are not designed to withstand structural movement of buildings or "acts of God." Thus for the purposes of this study, these influences were factored out. Asphalt-coated roofs were not rated because of the difficulty in determining what percentage of the roof required removal. A rating system for the asphalt-coated PUF roof is an area of future research. The owner of each roof inspected was asked the following questions to assist in determining the performance of the PUF roof system and the applicator: 1. Would you use the PUF roof system in the future? (Question 1 in Table 3) 2. Are you satisfied with your PUF roof system? (Question 2 in Table 3) 3. Are you satisfied with the applicator who installed your PUF roof system? (Question 3 in Table 3) 4. What type of maintenance do you conduct on your PUF roof system? 5. Has your roof leaked, and was it covered under warranty? The database was constructed at Arizona State University's School of Engineering, Industrial Engineering Department in The database included records of 1250 PUF roof applications in six geographical areas: Phoenix, Arizona; Boulder, Colorado; Eau Clair, Wisconsin; Cranford, New Jersey; Louisville, Kentucky; and College Station, Texas [4]. The PUF roof applicators were selected on the basis of accurate records, longevity of the applicator's roofing experience, the applicator's consent to open up all of his records and participate in a valid nonbiased study, and varying environmental conditions in the different areas. All roof applications were selected with no previous knowledge of the roof performance. Information collected included: customer satisfaction, maintenance performed, percentage deterioration of the system and required replacement, and the coating and PUF roof system performance under different environmental conditions. The results included the following [4]: 1. Silicone coated PUF roof systems performed well in environments with hail, ice, snow and roof traffic. 2. Silicone and acrylic coated systems performed well in ponded roof areas. 3. Asphalt coated PUF roof systems performed well in the New Jersey east coast environment which includes high humidity, snow loads, and ice formation.

6 Table 2. Roof condition ratings of PUF roofs inspected. Roof Conditions Age (years) AS E % VG % G % FR % P % F % ? Total Table 3. Results of customer satisfaction of the PUF roofs surveyed. No. of Positive Responses Age of No. of Roofs Roofs Q1 % Q2 % Q3 % T otols

7 validation <?f the Polyurethll11c Foam Ro<?f System Urethane coated PUF roof systems at Texas A&M University proved to be robust and easily maintained. The "in-house" roofmg crew experienced a tremendous reduction in maintenance work due to the roof system reliability. The crews also performed all minor maintenance including installation of curbs, washdowns of coating, and recoating. 5. Ninety-four percent of the roofs inspected had less than 5 percent deterioration (Table 2) and 97 percent of the owners (Table 3) were satisfied with the roof system. In 1985, roofs were reinspected in three geographical areas: Phoenix, Arizona; Cranford, New Jersey; and Louisville, Kentucky. Asphalt, urethane rubber, silicone, and acrylic coated PUF roof systems were inspected. Thirteen roofs were inspected in Phoenix, Arizona and surrounding areas. Seven owners of the roofs inspected were interviewed. All the customers were satisfied with their roofs and would use it again (Table 4). Of the 13 roofs, two roofs had no deterioration, six roofs had less than 1 % deterioration, and one roof was removed (failed). Eleven roofs were silicone coated and two were acrylic coated PUF roof systems. All thirteen roofs were inspected two years earlier. Of the thirteen roofs, only two roof systems suffered increased deterioration in the past two years. A 16 year old acrylic coated PUF roof, which had been recoated two years earlier, had been removed. The roof was badly blistered, and also leaking near the penetrations. A 13 year old silicone coated roof system was disbonding from the BUR substrate it was installed over. It was installed with 3/4 inch layer of PUF, which was below the recommended one inch minimum [3]. Of the 13 roofs inspected, 7 were flat roofs. Only one roof contained moisture in the PUF, and it was located in a flat ponding area. All the roofs remaining have a high probability of having a 20 year service life. As the roofs get older, there has not been a decrease in customer satisfaction or roof rating for the roofs. Sixty percent of the older silicone coated systems had coating thicknesses below industry standards, but seem to be performing superbly. None of these roofs were heavy traffic roofs. On one of the 12 year old roofs, granules were installed on the walkways. The rest of the roof was not granulated, and the 12 year old solvent cured, silicone coating was becoming brittle, and cracking. This was not observed two years earlier. The granulated walkway sections, however, were in excellent condition. None of the owners performed maintenance on their PUF roof system. Thirty-four roofs were inspected in the Louisville, Kentucky area. Sixteen roofs were 9 years or older (Table 5). Ninety-seven percent of the owners interviewed would use the roof system again. Eighty-eight percent were satisfied with their roofs and 96 % were satisfied with the contractor. Three customers were dissatisfied with their roofs. Only one complaint was roofing

8 \D 00 Table 4. Roofs inspected at Phoenix, Arizona. Age of Roof # of Roofs Inspected (#/% positive responses) E Roof Rating VG G FR F Totals /100 1/100 1/ /100 4/100 3/ /100 1/100 1/ /7(100%) 6/6(100%) 5/5(100%)

9 Validation of the PolYllrethaile Foam Roof System 99 Tobie 5. Roofs inspected ot Louisville, Kentucky Roof Rating Age of # of Roofs Roof Inspected (#/% positive responses) E VG G F /66 3/ /100 6/100 6/ /100 1/50 2/ /100 3/100 3/ /100 4/100 4/ /100 1/100 1/ / / /100 3/100 3/ /100 1/100 1/ /100 1/100 1/ /100 2/100 2/ >1 2 2/100 2/100 2/ Totals 34 29/30(97%) 23/26(88%) 25/26(96%) related: there was a leak in a large industrial roof which had not been fixed. One customer felt that although his roof system performed adequately, it was not a good roofing material, and would fail. Another customer thought that she would have to recoat her entire roof, which, in her opinion, was far too expensive. Tobie 6. Roofs inspected ot Cranford, New Jersey Age of # of Roofs Roof Inspected (#/% positive responses) /100 1/100 1/ /100 2/100 2/ /50 1/50 1/ /100 1/100 1/ /100 2/100 2/ /100 3/100 3/ /100 3/100 3/ /100 2/100 2/ /100 1/100 1/ /100 3/100 3/ /100 4/100 4/ /100 5/100 5/100 3 ') v 3/100 3/100 3/ /100 6/100 6/100 >1 2 2/100 2/100 2/100 Totals 45 39/40(98%) 39/40(98%) 39/39( 100%)

10 100 D. KASHIWAGI AND W. C. MOOR The only defects in the roofs inspected in the Louisville area were birdpecking (two roofs) and blistering (two roofs). The blistering occurred on roofs experiencing high moisture drives (apartment building, and classrooms). Six roofs had no slope. Only one roof (6%) had maintenance performed on it. Forty-five asphalt coated PUF roofs were inspected (Table 6) in Cranford, New Jersey and surrounding areas. Thirty-three percent of the roofs had service lives of over ten years. Ninety-eight percent of the customers were satisfied with the roof system and would use it again. One hundred percent of the customers were satisfied with the contractor. There was no decrease in satisfaction of the customers as the service life of the roofs increased. Only two roofs were inspected and maintained. Forty-four of the forty-eight roofs had no problems. Fifty percent of the roofs had no slope. Defects in the asphalt coated PUF roof system included delamination between passes, leaking near walls, penetrations, and edges, and moisture in the PUF. Eleven roofs had leaks, four roofs had moisture in the PUF, and six roofs had blistering and delamination. The older roofs (12-15 years) had a thick felt and asphalt cover. Three of these roofs contained moisture in the PUF. No moisture problems were identified on the newer roofs which had a light coating of asphalt and no felts. Saturation problems due to mechanical damage and open cells on the surface of the PUF were not evident on these roofs. Also the older asphalt coated roofs were not installed as expertly (smoothly and evenly) as the roofs inspected in Phoenix, Arizona; Louisville, Kentucky, or the roofs located at College Station, Texas; Eau Clair, Wisconsin; and Boulder, Colorado which were inspected in The newer roofs were of the same quality as seen elsewhere. The following conclusions can be made from the above data and previous data collected in PUF roofs can be installed consistently to provide 20 year roofs. If the PUF roof is a failure, one of the three essential factors (Figure 1) is missing. Either the 20 year field performance, the correct materials, or the correct installation procedure was not employed. 2. PUF roofs, if installed correctly, are a solution to flat roof problems. 3. "Successful" PUF roofs are low maintenance roofs. 4. Asphalt coatings and other non-elastomeric coverings can be used "successfully" with PUF roof systems. 5. Delamination of the PUF layers, saturated PUF, blisters in the PUF, and penetrations and edges are the most prevalent defects in the fmished product. 6. In all the roofs inspected in and , deterioration of coating has not been a cause for failure. Acrylic, urethane rubber, and asphalt

11 Validatioll of the Polyurethane Foam Roof System 101 roofs have been recoated. There was not a case where a silicone coated PUF roof system has had to be recoated to increase the service life of the system. ECONOMIC ANALYSIS OF THE PUF ROOF SYSTEM The results of the survey and inspection of 246 PUF roof systems have identified the PUF roof system as a low maintenance, 20-year roof system. The next task was to determine the economic feasibility of this roof system. Due to the numerous factors involved in monitoring energy consumption, simulation was the only feasible solution. An economic analysis was performed using a computer model which performed a one-at-a-time sensitivity analysis on the following parameters to determine their effect on the equivalent uniform annual cost (EUAC) of the PUF roof system in comparison with a 4-ply coal tar pitch built-up roof (BUR) system. 1. General inflation rate. 2. Interest Rate. 3. Decreasing rate of insulative value of the PUF. 4. Service life of the PUF and BUR roof systems. 5. Thickness of the PUF. 6. Inflation rate of the cost of heating fuel and electricity. Each of these factors was incremented for two simulated cases using the environments of Phoenix, Arizona, and Eau Clair, Wisconsin. Both cases produced similar results. An analysis of the results showed that the most economic thickness of PUF depends upon the combination of factors, the most influential being the inflation rate of heating fuel and the environmental (:onditions. The analysis of the Phoenix, Arizona, case study shall be briefly discussed. The two roof systems used in the comparison were a silicone-coated, oneinch PUF system and a four-ply, coat tar BUR system. Both were installed over metal decks. Appendix 1 lists the components and R -values of the simulated BUR system. In the PUF system, 1 inch of PUF replaces the BUR and perlite insulation board. The heating power source was Number 2 fuel oil and the cooling power source was electricity. A general inflation rate of 5.5 percent and an inflation rate of the cost of fuel of 3.3 percent was used. A loss of insulative quality of the PUF of 2 percent a year for 20 years was implemented, in accordance with Alumbaugh's fmdings on the deterioration ofpuf's insulative qualities. A maintenance cost of$0.01 per square foot per year was used for the BUR system. The analysis provided the following results:

12 102 D. KASHIWAGI AND W. C. MOOR 1. A six-year $1.90 per square foot PUF roof system is an economical alternative to a 20-year $2.96 per square toot BUR system. 2. A lo-year, $1.90 per square foot PUF roof system is an economical alternative to a 20-year, $1.50 per square foot BUR system. It would be highly unlikely that a 20-year BUR could be purchased for $1.50 per square foot, but it is used here to show how economically feasible the PUF roof system can be if it has a lo-year service life. 3. The cost of the BUR systems increased at a slightly higher rate than the cost of the PUF roof system as the interest rate and the inflation rate were increased. 4. The most economical thickness ofpuf in this analysis was one inch. In the Eau Clair, Wisconsin, case study, the most economical thickness of PUF in this analysis was 1.25 inches. This increased dramatically when the fuel inflation rate was increased. 5. The impact of the rate ofloss of insulation value of the PUF on the evaluation of the PUF roof system is negligible when compared with the impact of varying service lives. The maximum documented rate is 2 % a year over ten years [9]. Therefore, service life and not maintenance of the insulating value is the most important factor in determining economic feasibility. POLYURETHANE FOAM ROOF SYSTEM ANALYSIS The purpose of this analysis is to determine the reason for the defects identified in the "successful" PUF roof systems, and propose solutions to eliminate them. The problems will be analyzed in the following order: delamination, blistering, saturated PUF, ice cracking and leaking at joints and penetrations. After the defects and their solutions are discussed, the design of the system and components (the coating, the PUF, and the vapor barriers) will be discussed in the next section. Whenever delamination within the PUF passes occurred, several common factors were identified: 1. The surface of the layer beneath was usually smooth, and slightly burned from UV rays. 2. The thickness of the upper pass of PUF was inconsistent. Both of these factors are PUF installation related. It is a common practice to put down the first pass ofpuf on one day and the next pass on the next day. This procedure serves several purposes: 1. The whole roof can be covered in one day to prevent rain from damaging the contents of a building if it is a retrofit job.

13 Validatioll of the PolYlIrethane Foam Roof System It allows the removal of the gravel and, if required, the previous roof membrane in one day. 3. It also precludes match or cut lines in the PUF roof system. As previously mentioned, if this procedure is used, the probability of delamination between PUF layers is increased. During the time between PUF pass applications, dirt, grease, and moisture contaminate the surface of the ftrst pass of PUF. The problem is compounded by the burning of the surface of the PUF by UV rays, creating further dust contamination. Delamination is a delayed reaction to improper PUF application. It does not always result in immediate leaks, but is nevertheless a defect in the roof system which can contribute to a failure of the roof system. The solution to preventing delamination is to require same day, full buildup of the PUF thickness in 1/2 to 3f4 inch passes. Passes thinner than liz inch have the possibility of losing the heat released by the exothermic reaction between the polyol and the isocyanate, thus having insufficient heat to vaporize the blowing agent to form a uniform foam structure [5]. Blistering was identified in all coated systems. Blistering is caused by the loss of physical properties of the cell walls of the PUF. If moisture is present when the PUF is applied, an off-ratio foam results, leading invariably to blistering [3]. Water vapor condensation, and freeze-thaw cycles, also weaken the physical properties of the PUF [5]. The solution to mitigating blistering is similar to the solution to stop delamination of PUF layers. PUF must be brought to full depth on the same day, and no foaming should happen during wet or damp conditions. Along with blistering and delamination, saturation of the PUF was identified in the inspection as a major factor in PUF roof failure. The PUF is a closed cell structure which is watertight. For all practical purposes, PUF is waterproof. There are two ways moisture enters the PUF: 1. Open or damaged cells at the PUF surface. 2. Condensed moisture from vapor drive which usually condenses in the coldest regions of the PUF. There are several steps that can be taken to reduce the probability of saturation of the PUF: 1. Installation of a cover which will add a protective binding or cementious coating to the top layer of the PUF. 2. Stop the water vapor before it enters the PUF with a vapor barrier below the PUF. 3. Use a permeable cover, which protects the PUF by spreading point loadings, in conjunction with or without a vapor barrier on the bottom side of the PUF roof system. If a vapor barrier is not used, it is essential

14 104 D. KASHIWAGI AND W. C. MOOR to have a very permeable cover. This will be discussed further in the PUF systems design. Cracking in the PUF system has been attributed to the expansion and contraction of ice [8]. This cracking mechanism is caused by the continuous adhesive reaction to the top of the PUF roof system. A solution to this problem would be to disrupt the adhesive forces by using a ball bearing type concept. If gravel or granules were used, the ice would form around the aggregate, making the ice a discontinuous plate, free to expand and contract as required. The data from the inspection confirms that leaks occur mainly near penetrations and discontinuous sections of the roof. PUF is self flashing and if installed properly, should be sufficient to prevent leaking [3]. DESIGN OF THE PUF ROOF SYSTEMS As discussed in the inspection data and analysis, moisture in the PUF vapor drive is the biggest system design related problem. Using classical COATING OR COVER P (IMPERMEABLE COATING) lj... 0 a..: ~ w t MEABLE ING) PUF Temp l.9 4l. 5 5l '---.c... '---1..:::...-'- -...JI-l o P SAT PRESSURE (PSI) P(GRAVEL) P (PERMEABLE P (IMPERMEABLE COVER) COVER) FIGURE 2. Vapor pressure chart (PUF)-winter conditions.

15 Validation oj the Polyurethane Foam RooJ System ~Z:Z~~~z:LLZ:ZZ:Z"fZ2ZZZ2h 87.8 FL COVER 85.8 LL 0 0.: 83.8 ~ w I- PUF PRESSURE (PSI) Temp P s P(GRAVEL) P(PERMEABLE P (IMPER~EABLE COVER) COVER) So FIGURE 3. Vapor pressure chart (PUF)-summer conditions. steady state vapor drive equations, an analysis was done to determine the worst conditions for vapor condensation in a PUF system in the New Jersey area. The building system used under the PUF roof system is a four inch concrete structure, with a 16 inch clear space, and a false ceiling (Appendix II). The analysis was done for summer and winter conditions, to determine if a critical problem exists for vapor conditions, to determine if a critical problem exists for vapor drives going both ways through the PUF system. The vapor resistant factor is the critical factor in the condensation problem. Within the PUF roof system, the greatest vapor pressure drop occurs over the material with the lowest Perm rating or the highest resistance to vapor drive [7]. The first analysis was done for the summer design criteria for a PUF system with three different covers: 100% permeability cover such as a gravel layer, a permeable coating (> 1 Perm), and an impermeable coating «1 Perm). The results are shown in Figures 2 and 3. The same analysis was run for winter conditions. A critical condition exists when the vapor pressure exceeds the saturation pressure (when the Pv line is to the right of the Ps line). In those areas, theoretically, saturation occurs.

16 106 D. KASHIWAGI AND W. C. MOOR The results confirm that if a vapor drive exists, the critical vapor flow is from inside out, and not outside in. Therefore, an impermeable characteristic for a coating is not an important requirement for preventing the condensation of water vapor in PUF. If a vapor drive exists, it becomes a "worst condition" state. The best cover for a PUF roof system with vapor drive is a "totally" permeable cover. The data collection, analysis, and previous work on PUF roof systems confirm the initial hypothesis of the study. It dictates the following requirements on the PUF cover. 1. Reduce the mechanical loading requirement of the PUF by spreading the point loadings. 2. Protect the PUF from the UV rays. The requirements of the PUF are: 1. Waterproofmg membrane. This is the primary requirement. 2. Insulation. The requirement of the vapor barrier is to stop water vapor from entering the PUF roof system from the interior of the structure. PUF ROOF SYSTEM DESIGNS There are two different designs for PUF roof systems. The first is for systems that will allow a vapor drive to exist through the roof system. This design includes a minimum of one inch ofpuf installed in 112 inch lifts, with a very permeable cover. The type of cover would depend on the traffic loading on the roof, and the structural design and slope of the roof. Roof conditions allowing, the priority of coverings are: 1. Gravel or slag. 2. Gravel w/1s mil coating of asphalt. 3. One coat of 10 mils (.015 inches) of silicone or acrylic with granules embedded. 4. Single coat of 15 mil asphalt mil of acrylic coating. For systems which do not allow the existence of a vapor drive, the design includes a permeable or impermeable cover, PUF, and a vapor barrier. Great care must be taken in the installation of the vapor barrier if an impermeable cover is being used. The following covers are recommended: 1. Gravel or slag. 2. Gravel w/1s mil coating of asphalt.

17 Validation oj the PolYllrethaile Foam RooJ System One coat of 20 mils (0.15 inches) of silicone or acrylic with granules embedded mil asphalt mil of acrylic coating mil of urethane rubber coating. In both systems, installing more than one inch of PUF would probably lengthen the service life of the roof system because of added structural stability. In all cases, each layer of PUF has to be a minimum of 1f2 inch in depth. SPECIFICATIONS OF THE "SUCCESSFUV' PUF ROOF SYSTEM Specifications ensuring the application of a "successful" PUF roof system must include quantification of the produced work and proof of skill to install the work. The responsibility of the specification lies with the design engineer or architect, the installation of the roof with the contractor, and the warranty of performance of the roof after it is inspected and accepted as a proper application with the manufacturer. In today's PUF roofing industry, it is the manufacturer of the coating that issues the performance warranty. If a PUF system with a gravel cover is installed, the PUF manufacturer should issue the performance warranty. The specifications should include: 1. A certification document from the manufacturer of the coating or PUF manufacturer that the applicator is an approved and trained applicator. 2. A certification from the contractor listing ten jobs of equivalent size of the project that the applicator has installed. 3. A certification from the manufacturer that they will warrant the PUF roof if it is installed by the approved applicator. 4. A list of ten roofs in the immediate vicinity which are at least five years old with the specified system. 5. A minimum of 4 cores should be taken at random to ensure the specifications have been met. Four of the five samples should be within 20% of the specified values for pass thickness, PUF density, adhesive strength between passes, and with the coverng and roof deck, and flexural strength. CONCLUSION The "successful" 20 year PUF roof system is a proven, procurable product. The PUF is the waterproofing membrane, and a covering is required to

18 108 D. KASHIWAGI AND W. C. MOOR protect the PUF from UV rays and reduce point loadings by transferring the loads to a greater area. The recognition and design to handle vapor drive is the most critical design factor. If a vapor drive exists through the PUF roof system, a permeable cover will be installed. In cases where an impermeable vapor barrier is installed below the PUF, any of the discussed covers are sufficient. The installation of the PUF is the most critical phase in the installation procedure. It should be well specified in quantifiable and testable terms. ACKNOWLEDGEMENTS We wish to thank Keith Coultrap, Irv Stumler, Roger Morrison, and Ai Goertz and all their associates in Sprayfoam Southwest, Inc., Urethane of Kentuckiana, Inc., and Industrial Roofmg, Inc. for their expertise, cooperation, and encouragement. Their great desire to understand the performance of the polyurethane foam roof system has made this study possible. REFERENCES 1. Adler, N. A. "Economic Optimization of Roof Insulation Thermal Resistance," 1985 Second National Symposium 011 Ro(~fillg Tc(hnology, National Symposium 011 Roof ing Tcchllology, National Ro(~fillg Colltractors.4ssociatioll, Chicago, fl, pp (1985). 2. Alumbaugh, R. L., J. R. Keeton and E. F. Humm. "Experimental Polyurethane Foam Roof Systems - II;' TN No. N-1656, Naval Civil Engineering Laboratory, Port Hueneme, CA Ganuary, 1983). 3. Coultrap, K. H. "Principles of Urethane Foam Roof Application," PO No. 79-MR-461, Naval Facilities Engineering Command, Port Hueneme, CA Gune 1980). 4. Kashiwagi, D. "The Economic Feasibility of the Polyurethane Foam Roof System;' Arizona State University, Tempe, AZ (December, 1983). 5. Jenkins, A. L. "A Study of Some Structural Interrelationships Between Polyurethane Foam and Protective Coatings:' 1985 Second Illtemational Symposilllll (~f RO(?F illg Techllology, Nati01zai Roo.fillg Cmtm((ors Associatiol1, Chicago, IL, pp (1985). 6. National Roofing Contractors Association. "Good Roofs Save Energy;' NRCA, Oak Park, IL (September, 1977). 7. Ramsey, C. G. and H. R. Sleeper. 4rchitc(tllral Graphic Stil11dards. USA:John Wiley & Sons, Inc., pp (1981). 8. Riedel, R. G. "A Theory to Explain Roof Splitting By Ice;' 1985 Sewl1d National Symposium on Roo.fing Tcchnology, National Symposillm on Ro(?fing Tcchnology, National Ro(?fing Contractors,4ssociatioll, Chicago, IL, pp (1985). 9. Zarate, D. A. and R. L. Alumbaugh. "Thermal Conductivity of Weathered Polyurethane Foam Roofing;' TN No. N-1643, Naval Civil Engineering Laboratory, Port Hueneme, CA (September, 1982).

19 Validation of the PolYllrethane Foam Roof System 109 APPENDIX I. SIMULATED BUR SYSTEM IN ECONOMIC ANALYSIS R-Value Roof Component Cooling Heating Outside Air Film BUR One Inch Perlite Insulation Vapor Retarder Steel Deck Air Space /8 Inch Ceiling Tile Inside Air Film TOTAL Assumptions - Roof System APPENDIX II. VAPOR DRIVE ANALYSIS Roof Systems R-Values Air Space.17 Cover " Concrete Deck.33 16" Clear Space.94 False Ceiling 1.81 Air Space.61 Perm Rating 2.0,0.5, TOTAL Temperature and Relative Humidity Inside Outside Summer 75 F/40% RH 88 F/lOO% RH Winter 75 F/40% RH 11 F/I00% RH* *l()ivx) RH is chosen because it creates :1 nlinimum pressure differcntial. BIOGRAPHIES Dean Kashiwagi Dean Kashiwagi was born in Honolulu, Hawaii, October 27, He graduated in 1978 with a Bachelor of Science Degree in Civil Engineering

20 110 D. KASHIWAGI AND W. C. MOOR from the University of Hawaii. That same year he was commissioned as a Second Lieutenant in the USAF and assigned to Holloman AFB, New Mexico. In 1983, Dean graduated from Arizona State University with a Masters of Science Degree in Industrial Engineering. He was then assigned to the Air Force Engineering and Services Center at Tyndall AFB, Florida. While at Arizona State University, Dean began studying polyurethane foam roof systems. He has since developed a database, written a performance specification, and continued to analyze the problems associated with its installation and maintenance. Dean is married to Judy Kashiwagi and they have six children. William C. Moor William C. Moor was born in St. Louis, MO in 1941, growing up in southwestern Illinois. He earned the B.S.I.E. (1963) and M.S.I.E. (1965) degrees from Washington University (St. Louis) and the Ph.D. (1968) from Northwestern. He joined Arizona State University in 1968 where he is an Associate Professor of Engineering. He is married to the former Marilyn Nichols and the father of two children.