Pharmaceutical Cleanroom HVAC Ventilation Rate StudyAuthors Thomas R. Spearman Consultant Engineer IPPO Plant Engineering Donald R Moore Associate Senior Consultant Engineer Engineering Technical Center Rebecca J Elliott Research Scientist MS&T Statistics Eli Lilly and Company Lilly Corporate Center Indianapolis, IN 46285 Page 1 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Abstract The objective of this experiment was to challenge design practices that utilize high ventilation (air change) rates to maintain particulate levels within classified spaces. This experiment was designed to produce sufficient independent data on the correlation of ventilation rate to both particulate control (within classification limits) and recovery time (clean up from “in operation” particulate levels to one or two orders of magnitude cleaner when “at rest”). This study, which was performed in a pilot facility at Eli Lilly, as part of a pharmaceutical industry wide ‘Green’ Pharma HVAC Team initiative aimed at reducing HVAC energy costs while maintaining GMP compliance within the cleanroom. The team’s plan is to execute similar studies in other pharma cleanrooms to develop guidelines for the industry for more sustainable HVAC systems. The conclusions of this study are: • For this application, the room recovery rate test is a good measure of cleanroom performance. The theoretical equation provided good estimates of the actual performance. However, at an Adjustable Speed Drive (ASD) setting of 15 Hz (seven air changes per hour), the theoretical equation provided poor estimates of the actual performance. • A more consistent method of generating particles is needed for future testing. The generation rate needs to be independently determined or measured instead of using the equation based on supply air volume used in this report. Page 2 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Purpose The objective of this experiment was to challenge design practices that utilize high ventilation (air change) rates to maintain particulate levels within classified spaces. This experiment was designed to produce sufficient independent data on the correlation of ventilation rate to both particulate control (within classification limits) and recovery (clean up from “in operation” particulate levels to one or two orders of magnitude cleaner when “at rest”) time. The intent is to provide robust data for review by regulators and industry, thus encouraging investigation at a company level helping drive the industry to a science based assessment of the actual air change requirements, being more energy efficient and thus reducing operating costs (energy, filters, maintenance) in the process. Tests Executed This testing measured the room particulate concentrations with an aerosol (particle) challenge at varying ventilation rates. It also measured the recovery time after the aerosol challenge was stopped. The testing was performed with the room equipment not operating, so that this was a test of the HVAC system only. The testing was conducted per the pre-approved protocol. Testing was performed in the Cartridge Prep room at an Eli Lilly Pilot Plant, located from 22-Mar-2010 to 25-Mar-2010. Test equipment calibration and HEPA filter integrity tests were verified to be current. The HVAC system smoke detectors were disabled. The ventilation rate was varied using a remote keypad that changed the adjustable speed drive (ASD) setpoint on air handling unit (AHU) 5. For each ASD setting, the volumetric airflow rates for the room terminal HEPA filters and return grilles were measured and recorded. The room dimensions were measured and recorded. This information was used to calculate the room volume. The room volume includes the volume occupied by equipment. In other words, the equipment volume was not deducted from the room volume. Page 3 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study For each ASD setting, the following data was collected. • Room differential pressure relative to the adjacent mechanical room (room 16A) • Terminal HEPA volumetric airflow rate in cubic feet per minute (cfm) • Return air volumetric airflow rate in cubic feet per minute (cfm) • Particle counts for 0.5 micron particles and larger and 5 micron particles and larger were recorded in one minute intervals. This data was collected in two separate locations. Dilutors with a 10:1 ratio were used on the particle counters to allow for higher particle counts. The room layout with the equipment, motor control center (MCC), low level (LL) return, terminal HEPA, particle counter, and particle generator outlet (P) locations are shown in Figure 1. The particle counter designations (ID#) are also shown. The adjacent mechanical room was used as the reference for differential pressure measurements. Page 4 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 1: Cartridge Prep Room Layout ID# 2 ID# 1 PG HEPA HEPA HEPA HEPA HEPA HEPA Tunnel LL Return LL Return MCC MCC Mech Room P Page 5 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study An ATI model TDA-6C aerosol generator was used as the particle generator. A hose was attached to the generator outlet. A partially opened diaphragm valve was placed on the end of the hose to reduce the particle generation rate. The valve position was determined and set during preliminary testing and was not readjusted during subsequent testing. The valve outlet was elevated approximately four feet above the floor. The Laskin nozzle pressure was set at 3 psig. The generator was started and stopped remotely, outside of the test room. Thirteen test runs were conducted. Run 13 was conducted as a replacement for Run 11 because the particle counter ran out of paper during Run 11. Table A lists the run number, adjustable speed drive (ASD) setting, and test date. Table A: Run ASD Settings and Test Dates Run # ASD setting Test Date Preliminary Various 22-Mar-2010 1 60 23-Mar-2010 2 15 23-Mar-2010 3 45 23-Mar-2010 4 30 23-Mar-2010 5 15 24-Mar-2010 6 30 24-Mar-2010 7 60 24-Mar-2010 8 45 24-Mar-2010 9 45 25-Mar-2010 10 60 25-Mar-2010 11 30 25-Mar-2010 12 15 25-Mar-2010 13 30 25-Mar-2010 Here are the protocol departures. They do not affect the test results. • Datasheets 3 and 4 were not used to setup the particle counters. The particle counters were setup by the Environmental Monitoring group. • The room does not have a routine cleaning procedure. This was marked as not applicable on Datasheet 5 Step 2. • A partially open diaphragm valve was added to the outlet of the particle generator to reduce the particle generation rate. • The room temperature and humidity were not recorded for runs 1 and 4. The particle counters also recorded the room temperature and humidity. This was marked as not applicable on Datasheet 6 Step 4. • The exact location of the particle counters was not measured and recorded per Datasheet 6 Step 6. The approximate locations are shown on Sketch 2 in the protocol and Figure 1 of this report. • The particle generator operation test duration was changed from 30 minutes to 15 minutes after preliminary testing. Page 6 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study • The ID#1 particle counter ran out of paper during run 11 at minute 41. Run 13 was conducted as a replacement for Run 11. Page 7 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Analysis The testing was divided into three time intervals: minute 1 to minute 10, minute 11 to minute 25, and minute 26 to minute 45. From minute 1 to minute 10, personnel were not present and the particle generator was not operating. This time period was used to determine At-Rest particle concentrations. From minute 11 to minute 25, the particle generator was in operation. This time period was used to determine average particle concentrations. From minutes 26 to 45, the particle generator was stopped. This time period was used to determine the two log reduction recovery time. The two log reduction recovery time is the time to achieve a 100 to1 reduction in 0.5 micron particle concentration. For each ASD setting, the following data was determined or calculated. • Maximum 0.5 micron particle concentration • Average 0.5 micron particle concentration • At rest 0.5 micron particle concentration • Supply air volumetric flow rate • Maximum particle generation rate • Average particle generation rate • Air change rate in changes per hour • Actual two log reduction recovery time in minutes • Theoretical two log reduction time in minutes • Difference between actual two log reduction recovery time and theoretical in minutes The maximum 0.5 micron particle concentration (Cmax) is the largest 0.5 micron particle concentration at any time during the run (between minutes 1-45). The average 0.5 micron particle concentration (Cavg) is the average of the 0.5 micron particle concentrations for the last five minutes when the particle generator was operating (minutes 21-25). The actual two log reduction recovery time (Ract) was determined by inspecting the particle concentration data in tabular form. When a two log reduction was not achieved in the case of low air change rates, the single log reduction recovery time was determined and multiplied by 2 to calculate the two log reduction recovery time. Page 8 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study The maximum particle generation rates (PGRmax) are calculated using the following equation 1 . P0Rmox = Cmox - µ Where: PGRmax =maximum particle generation rate in particles per minute Cmax =maximum particle concentration (minutes 1-45) in particles per cubic foot Q =supply air volumetric flow rate in cubic feet per minute (cfm) The average particle generation rates (PGRavg) are calculated using the following equation 1 . P0Ro:g = Co:g - µ Where: PGRavg =average particle generation rate in particles per minute Cavg =average particle concentration (minutes 21-25) in particles per cubic feet (ft 3 ) Q =supply air volumetric flow rate in cubic feet per minute The air change rates (N) are calculated using the following equation 1 . N = 6u - µ I Where: N =air change rate in changes per hour Q =supply air volumetric flow rate in cubic feet per minute V =room volume in cubic feet Page 9 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study The theoretical two log reduction recovery times (R th ) are calculated using the following equation 2 . N Rth ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ × − = 100 1 ln 60 Where: Rth =theoretical two log reduction recovery time in minutes N =air change rate in changes per hour Page 10 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Page 11 of 55 Results for Individual Tests Table B summarizes the recorded, measured, and calculated values for each run and particle counter location. For the thirteen test runs, Figures 2-14 plot the 0.5 micron particle concentrations, 5 micron particle concentrations for both particle counter locations. The time the particle generator (PG) is turned on and turned off are indicated. Pharmaceutical Cleanroom HVAC Ventilation Rate Study Table B: Run Data and Calculated Values Run # ASD ID# Cmax Cavg Cr Q V PGRmax PGRavg N Ract Rth Ract‐Rth ASD setting in Hertz max 0.5 particles per cubic foot average of 5 pts 0.5 particles per cubic foot at rest particles per cubic foot supply air volumetric flow rate in cubic feet per minute room volume in cubic feet particle generation rate in particles per minute particle generation rate in particles per minute air change rate in changes per hour actual 2 log recovery time in minutes theoretical 2 log recovery time in minutes difference between actual 2 log recovery time and theoretical in minutes 1 60 1 40,730 30,748 1 2881 3527 1.17E+08 8.86E+07 49.0 9 5.6 3.4 1 60 2 51,710 36,478 1 2881 3527 1.49E+08 1.05E+08 49.0 9 5.6 3.4 2 15 1 112,727 97,197 737 420 3527 4.73E+07 4.08E+07 7.1 36 38.7 ‐2.7 2 15 2 127,580 85,350 80 420 3527 5.36E+07 3.58E+07 7.1 30 38.7 ‐8.7 3 45 1 248,450 215,864 1 2032 3527 5.05E+08 4.39E+08 34.6 11 8.0 3.0 3 45 2 152,180 138,062 1 2032 3527 3.09E+08 2.81E+08 34.6 11 8.0 3.0 4 30 1 312,370 257,758 110 1199 3527 3.75E+08 3.09E+08 20.4 14 13.5 0.5 4 30 2 1,206,530 840,695 10 1199 3527 1.45E+09 1.01E+09 20.4 13 13.5 ‐0.5 5 15 1 1,387,630 1,329,324 540 405 3527 5.62E+08 5.38E+08 6.9 26 40.1 ‐14.1 5 15 2 1,767,430 1,498,644 530 405 3527 7.16E+08 6.07E+08 6.9 26 40.1 ‐14.1 6 30 1 620,660 56,886 1 1193 3527 7.40E+08 6.79E+07 20.3 13 13.6 ‐0.6 6 30 2 586,970 457,818 1 1193 3527 7.00E+08 5.46E+08 20.3 12 13.6 ‐1.6 7 60 1 84,590 41,926 1 2733 3527 2.31E+08 1.15E+08 46.5 9 5.9 3.1 7 60 2 112,480 51,040 1 2733 3527 3.07E+08 1.39E+08 46.5 9 5.9 3.1 8 45 1 42,909 11,526 1 1972 3527 8.46E+07 2.27E+07 33.5 11 8.2 2.8 8 45 2 63,970 53,256 1 1972 3527 1.26E+08 1.05E+08 33.5 11 8.2 2.8 9 45 1 574,370 47,610 1 2015 3527 1.16E+09 9.59E+07 34.3 9 8.1 0.9 9 45 2 747,350 355,496 30 2015 3527 1.51E+09 7.16E+08 34.3 9 8.1 0.9 10 60 1 88,960 53,258 1 2805 3527 2.50E+08 1.49E+08 47.7 8 5.8 2.2 10 60 2 104,820 66,492 1 2805 3527 2.94E+08 1.87E+08 47.7 8 5.8 2.2 11 30 1 245,091 39,418 70 1197 3527 2.93E+08 4.72E+07 20.4 18 13.6 4.4 11 30 2 864,130 702,218 60 1197 3527 1.03E+09 8.41E+08 20.4 14 13.6 0.4 12 15 1 332,131 24,404 3,200 442 3527 1.47E+08 1.08E+07 7.5 34 36.7 ‐2.7 12 15 2 549,640 476,074 490 442 3527 2.43E+08 2.10E+08 7.5 22 36.7 ‐14.7 13 30 1 388,657 20,538 140 1217 3527 4.73E+08 2.50E+07 20.7 14 13.3 0.7 13 30 2 746,370 601,772 1 1217 3527 9.08E+08 7.32E+08 20.7 13 13.3 ‐0.3 Note: Run 11 is excluded from analysis due to the particle counter running out of paper. Page 12 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 2: Run 1 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 1 60 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Page 13 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 3: Run 2 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 2 15 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG off PG on Page 14 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 4: Run 3 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 3 45 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Page 15 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 5: Run 4 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1.00E+07 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 4 30 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Page 16 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 6: Run 5 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1.00E+07 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 5 15 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Page 17 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 7: Run 6 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 6 30 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Page 18 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 8: Run 7 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 7 60 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Page 19 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 9: Run 8 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 8 45 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Page 20 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 10: Run 9 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 9 45 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Page 21 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 11: Run 10 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 10 60 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Page 22 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 12: Run 11 Plot Note: The particle counter (ID#1) ran out of paper at minute 41. 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 11 30 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Page 23 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 13: Run 12 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 12 15 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Page 24 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Page 25 of 55 Figure 14: Run 13 Plot 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Minute Run 13 30 Hertz ID#1 0.5 micron per ft3 ID#2 0.5 micron per ft3 ID#1 5.0 micron per ft3 ID#2 5.0 micron per ft3 PG on PG off Pharmaceutical Cleanroom HVAC Ventilation Rate Study Results Figure 15 compares ASD setting to the air change rate. The R 2 fit of 0.999 is an indication of the repeatability of the ASD settings to supply air flow. See Appendix A for statistical analysis. The linear relationship is predicted by the Fan Laws 3 . Figure 15: Air Change Rate versus ASD Setting and Run Page 26 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Page 27 of 55 Table C examines the average particle generation rates. For the three runs at each ASD setting and for each particle counter location (ID#), the average and standard deviation are calculated. The high standard deviation relative to the mean is an indication of the high particle generation rate variability. Run 11 is excluded. Table C: ASD Setting Particle Count Average and Standard Deviation ASD setting ID# Run# PGRavg average standard deviation ID# PGRavg average standard deviation 60 1 1 88.6E+6 117.5E+6 30.5E+6 2 105.1E+6 143.7E+6 40.9E+6 60 1 7 114.6E+6 2 139.5E+6 60 1 10 149.4E+6 2 186.5E+6 45 1 3 438.6E+6 185.8E+6 222.0E+6 2 280.5E+6 367.3E+6 314.8E+6 45 1 8 22.7E+6 2 105.0E+6 45 1 9 95.9E+6 2 716.3E+6 30 1 4 309.1E+6 134.0E+6 153.1E+6 2 1.0E+9 762.2E+6 232.3E+6 30 1 6 67.9E+6 2 546.2E+6 30 1 13 25.0E+6 2 732.4E+6 15 1 2 40.8E+6 196.7E+6 296.3E+6 2 35.8E+6 284.4E+6 292.7E+6 15 1 5 538.4E+6 2 607.0E+6 15 1 12 10.8E+6 2 210.4E+6 Figure 16 compares the average particle concentration in particle per cubic feet to the air change rate in changes per hour. C1avg is the average particle concentration for particle counter ID#1 and C2avg is the average particle concentration for particle counter ID#2. Run 11 is excluded. Figure 17 plots maximum 0.5 micron particle concentration for both particle counter locations by ASD setting. Figure 18 plots average 0.5 micron particle concentration for both particle counter locations by ASD setting. Examining these figures, the test with the ASD at 15 Hertz is much different than the other tests. At an ASD setting of 30 Hertz, the two instruments are much different, but not at ASD settings of 45 Hertz and 60 Hertz. Figure 19 plots maximum 0.5 micron particle concentration for both particle counter locations by run. Figure 20 plots average 0.5 micron particle concentration for both particle counter locations by run. In both figures, there are three points that are high and “don’t fit”. The maximum and average particle concentrations increased over the testing period. For the average particle concentration, instrument 2 had an increase in concentration over the testing period. There is a clear difference in the instruments. Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 16: Average Particle Concentration versus Air Change Rate Page 28 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 17: Maximum Particle Concentration Variability Chart Page 29 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 18: Average Particle Concentration Variability Chart Page 30 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 19: Maximum Particle Concentration by Run Page 31 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Page 32 of 55 Figure 20: Average Particle Concentration by Run Pharmaceutical Cleanroom HVAC Ventilation Rate Study Table D shows the actual two log reduction room recovery rates for both particle counter locations. This information is sorted by ASD setting and run number. The average and standard deviation of the three runs at each ASD setting for each particle counter location is shown (see Appendix A for J MP output). The low standard deviation is an indication of high repeatability of this test. The higher standard deviation at ASD setting of 15 Hertz indicates that the test is not as repeatable at lower air change rates. Run 11 is excluded. Table D: Two Log Reduction Recovery Rate Averages and Standard Deviations Run # ASD N R1act R2act ASD setting in Hz air change rate in changes per hour actual 2 log recovery time in minutes average standard deviation actual 2 log recovery time in minutes average standard deviation 1 60 49.0 9 8.7 0.6 9 8.7 0.6 7 60 46.5 9 9 10 60 47.7 8 8 3 45 34.6 11 10.3 1.2 11 10.3 1.2 8 45 33.5 11 11 9 45 34.3 9 9 4 30 20.4 14 13.7 0.6 13 12.7 0.6 6 30 20.3 13 12 13 30 20.7 14 13 2 15 7.1 36 32.0 5.3 30 26.0 4.0 5 15 6.9 26 22 12 15 7.5 34 26 Figure 21 compares the actual and theoretical two log recovery times in minutes to the ASD setting in Hertz. R1act is the actual two log recovery time for particle counter location ID#1. R1th is the theoretical two log recovery time for particle counter location ID#1. R2act is the actual two log recovery time for particle counter location ID#2. R2th is the theoretical two log recovery time for particle counter location ID#2. Run 11 is excluded. Figure 22 compares the difference between actual and theoretical two log recovery times by ASD setting. Figure 23 compares the difference between actual and theoretical two log recovery time by run. The air change rate is significant. Only for the ASD settings of 30 Hertz are the results near zero, yet the instruments are different. For the ASD settings of 45 and 60 Hertz, the difference is consistent but actual is greater than theoretical. For the ASD setting of 15 Hertz, all the results are less than theoretical. Page 33 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 21: Two Log Recovery Time versus ASD Setting Page 34 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Figure 22: Difference between Actual and Theoretical 2 Log Recovery by ASD Setting Page 35 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Page 36 of 55 Figure 23: Difference Between Actual and Theoretical 2 Log Recovery by Run Pharmaceutical Cleanroom HVAC Ventilation Rate Study Discussion There was a delay in the particle generator starting during some runs. This is apparent in the following Figures and Runs. • Figure 3, Run 2 • Figure 4, Run 3 • Figure 5, Run 4 • Figure 6, Run 5 • Figure 9, Run 8 • Figure 10, Run 9 • Figure 12, Run 11 • Figure 13, Run 12 During some runs, the particle generation declined prior to turning off the particle generator. This is apparent in the following Figures and Runs. • Figure 5, Run 4 • Figure 6, Run 5 • Figure 7, Run 6 • Figure 10, Run 9 • Figure 12, Run 11 • Figure 14, Run 13 During some runs, the particle generation continued after the particle generator (PG) was turned off. This is apparent in the following Figures and Runs. • Figure 2, Run 1 • Figure 3, Run 2 • Figure 4, Run 3 • Figure 8, Run 7 • Figure 9, Run 8 • Figure 11, Run 10 The inconsistent particle generation may be attributable to: • Unstable aerosol production by the Laskin nozzles at 3 psig. • Aerosol back pressure building up upstream of the diaphragm valve and then bleeding off after the generator was turned off. • Variation in the room absolute pressure due to varying air supply flow rate which creates varying backpressure on the aerosol generator hose discharge. The unstable aerosol production may be due the flow surging because the nozzle did not reach sonic velocity, also known as choked flow. The 5 micron particles were significantly higher in concentration at lower airflow rates. This is expected since higher airflow rates have a higher transport velocity and carries larger particles more easily. Page 37 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study For plotting on semi-log scale, particle counts of zero were rounded up to one. For the 0.5 micron data, particle counts of less than 10 can be ignored since dilutors with a 10:1 ratio were installed on the particle counters. The particle counts at locations ID#1 and ID#2 were notably separated in the following Figures and Runs. • Figure 5, Run 4 (30Hz) • Figure 9, Run 8 (45 Hz) • Figure 12, Run 11 (30Hz) • Figure 13, Run 12 (15 Hz) • Figure 14, Run 13 (30 Hz) The particle concentration difference in the two instruments is also indicated in Figures 16 and 17. There is more separation between instruments at an ASD setting of 30 Hertz than other settings. Significantly different airflow patterns between higher and lower air flow rates may account for this separation. Computational Fluid Dynamic (CFD) computer models may be a useful tool to determine if this is the case. Figure 20 also indicates that the separation between instruments became greater over time, specifically runs 11-13. Figure 22 shows a difference in the instruments at an ASD setting of 30 Hertz. Table D and Figures 21 and 22 show that the room recovery rates are consistent between runs at the same ASD settings of 30, 45, and 60 Hertz. This indicates that this test is a good measure of the room recovery rate. This is also indicated by the low standard deviations in Table D. The room recovery rate was inconsistent at the ASD setting of 15 Hertz. Sources for error include: • Shortridge AirData Multimeter calibration • Supply air flow rate measurement technique • Particle counter calibration • Dilutor calibration • HEPA filter integrity (Supply air particle counts) • Room pressurization variability between cartridge prep room and adjacent rooms. • Room pressurization variability, the effect of the room absolute pressure on the aerosol generator • Repeatability of the particle generation rate • Not deducting equipment volume from the room volume calculations Page 38 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Conclusions For this application, the room recovery rate test is a good measure of cleanroom performance. The theoretical equation provided good estimates of the actual performance. However, at an ASD setting of 15 Hz (seven air changes per hour), the theoretical equation provided poor estimates of the actual performance. A more consistent method of generating particles is needed for future testing. The particle generation needs to begin immediately when the particle generator is started. The particle generation rate needs to be consistent from minute to minute when the particle generator is operating. The particle generation needs to end when the particle generator is stopped. The generation rate needs to be independently determined or measured instead of using the equation based on supply air volume used in this report. References 1. ISPE Good Practice Guide Heating, Ventilation, and Air Conditioning, First Edition, International Society of Pharmaceutical Engineers (ISPE), 2009, p. 253 2. DiGiovanni, M and Spearman T, “Recovery Testing a Pharmaceutical Cleanroom Case Study”, Performance Review, Spring 2006 edition, Controlled Environment Testing Association (CETA), p. 7-14. 3. 2008 ASHRAE Handbook – HVAC Systems and Equipment, page 20.4. Page 39 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Appendix A Statistical Analysis J MP 7.01 was used for all analyses. The ASD setting and the air change rate is the same for both ID#1 and 2. This regression analysis uses ID#1 and includes Run #11. Page 40 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study This table was generated using Tables Summary. ASD setting in Hertz ID# N Rows Mean(actual 2 log recovery time in minutes) Std Dev(actual 2 log recovery time in minutes) 15 1 3 32 5.29150262 15 2 3 26 4 30 1 3 13.6666667 0.57735027 30 2 3 12.6666667 0.57735027 45 1 3 10.3333333 1.15470054 45 2 3 10.3333333 1.15470054 60 1 3 8.66666667 0.57735027 60 2 3 8.66666667 0.57735027 Page 41 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Appendix B Photographs Page 42 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B1: Particle Counter with Dilutor Page 43 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B2: Particle Counter with Dilutor Page 44 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B3: Low Level Vent Page 45 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B4: Low Level Vent Page 46 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B5: Aerosol Generator Diaphragm Valve and Hose Page 47 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B6: Aerosol Generator Page 48 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B7: Vail Washer Protected with Plastic Page 49 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B8: Aerosol Generator Diaphragm Valve Page 50 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B9: Aerosol Generator Diaphragm Valve Page 51 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B10: Cartridge Washer Page 52 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B11: Sterilization Tunnel Page 53 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Photo B12: Sterilization Tunnel Page 54 of 55 Pharmaceutical Cleanroom HVAC Ventilation Rate Study Page 55 of 55 Photo B13 (left to right): Norm Goldschmidt, Don Moore, Tom Spearman, Bill Johnson, Rick Tracy