What’s your cleanroom costing you?
To answer the question, we looked at the fact, on average, as much as 80% of the total energy used by a cleanroom is consumed by the heating, ventilation, and air conditioning (HVAC) system. Up to 50% of this energy is consumed by the fans that provide the high air change rates needed to keep these rooms clean.
Energy savings can be made by reducing the air change rates during periods of inactivity. This can be a fairly straightforward process for non-sterile environments; but what about sterile environments where you must consider non-viable and viable particle counts? From a quality assurance and regulator’s perspective, sterile environments are much more problematic.
Why should this be the case? What goes on in an inactive cleanroom? Does the air change rate actually affect microbial growth?
The team at PharmOut, using current and newly available technology, set out to show that during periods of inactivity, a sterile cleanroom can maintain conditions with a significantly reduced air change rate.
Disclaimer! This is not a definitive scientific study. This experiment was designed to explore the boundaries of cleanroom behaviour during periods of reduced air flow. The data presented here is not intended to be used to justify a turndown rationale, however, techniques used during this experiment can be used to gather your own data and qualify a cleanroom for turndown.
An experiment was performed in an empty Grade C cleanroom, serviced by an airlock. Both rooms had terminal HEPA filtration. The cleanroom had a low level return air and the air supplied to the airlock was discharged into an adjacent corridor.
A single air handling unit supplied the cleanroom and airlock, providing the cleanroom with up to 45 air changes per hour (AC/H). The system was balanced to provide 15 Pascals positive pressure to the airlock and 30 Pascals positive pressure to the cleanroom, relative to atmospheric pressure.
Air change rates and room pressures were set and maintained by adjusting a variable speed drive on one of the fans, a damper on the main return air duct, and the flow of air through the airlock door to the corridor. Temperature, room pressure, humidity, and airflow data was monitored and trended using a screen located in the corridor.
Throughout the experiment, the room was monitored for viable and non-viable particles using:
- a Lasair II particle counter, for continuous sampling of non-viable particles
- a MiniCapt air sampler, for occasional sampling of viable particles
- contact plates, for occasional sampling of viable particles
- a TSI BioTrak particle counter, for continuous sampling of both viable and non-viable particles.
The BioTrak particle counter was a new technology that allowed the team to monitor viable particles in real-time throughout the experiment. Additional details on the unit can be found here. The BioTrak was the primary device used for data collection and was located in front of the low level return to ensure the best chance of collecting any particles generated or present in the room.
Contact plates and initial viable particle counts were taken across the room after performing a full room clean. The data collected by the contact plates, MiniCapt air sampler and Lasair particle counter was largely in agreement. There was evidence (when compared with the MiniCapt) that the BioTrak overestimates the number of viable particles at low air change rates, allowing us to assume worst case when using real-time data.
Experimental ‘in operation’ activity was simulated during micro sampling activities and periods of full air flow (45 air changes per hour). While the simulation took place for approximately 30 minutes, it was conducted with the aim of exposing the cleanroom to what might be generally expected during an 8 hour ‘in operation’ period in a Grade C/ ISO 7 cleanroom. Simulation activities included door openings, changing of garments in the airlock and personnel movement throughout the room to different workstations.
Figure 1: Daily air change rates and sampling times over the trial period.
The data collected by the BioTrak particle counter showed a negligible increase in the number of both viable and non-viable particles during periods of inactivity with decreased air change rates. The total number of viable particles per air change rate are shown in the table below.
Table 1: Total viable counts (cfu/m3) at the tested air change rates.
An increase in the number of non-viable particles was noted during periods of low air change rates; however as seen in Figure 2 this increase has little effect, demonstrating that classification can be maintained with lower air change rates during periods of inactivity.
Figure 2: Graph of Air change rates and non-viable particle concentration during trial 3.
So how low can we go?
The level of room contamination dramatically increases when the HVAC is turned off (zero airflow) as control of temperature and humidity is lost and room pressures cannot be maintained. With the fan turned off the negative pressure in the return duct is lost, and air flows back into the room, potentially dragging external contamination into the room. An example of this is shown in this Youtube video.
Figure 3 shows the rapid increase in particle concentration as the air change rate is reduced then stopped. On average, room classification was lost in approximately 20 minutes, proving there are limits to the level of turndown taken to maintain acceptable conditions.
Figure 3: Graph of air change rates (low) and non-viable particle concentration over time during trial 3.
Note: This scenario is specific to the HVAC and filter configuration in this facility and may not be replicated in other facilities.
What points should you consider before testing and implementing periods of reduced air change rates?
Recommendations before reducing your air change rates during turndown periods:
- Initiate reduced air change rates manually instead of automatically to give users more flexibility in operation.
- Automate the process settings (fan speeds, damper positions, etc.) to provide smoother and more accurate maintenance of room pressures during the ramp down / ramp up times.
- Ensure your HVAC fans are capable of maintaining low air flow rates, as some fans designed for high flow may become inefficient and/or unstable at low air flows.
- Ensure rooms cannot be accessed during periods of low air flow to prevent the introduction of possible contamination.
- Risk assess areas of your rooms for extra testing to identify where there might be changes in airflow due to restrictions or the presence of equipment.
Most importantly, understand your system and gather data and report based evidence to justify the low air changes rates during downtimes to the satisfaction of your QA team and regulators.
Stay tuned for an upcoming whitepaper on this topic.
A special thanks to
Kenelec Scientific for providing the use of the BioTrak, Lasair II, and MiniCapt air samplers.
Wilkore Construction for providing the use of the cleanroom, energy, and low level return.
Park Avenue Laundry for providing the sterile garments and cleaning equipment.
PE009-8; PIC/S Guide to GMP Annex 1 – Manufacture of Sterile Medicinal Products
ISO 14644-1:1999; Cleanrooms and associated controlled environments – Part 1: Classification of air cleanliness
ISO 14644-3:2005; Cleanrooms and associated controlled environments – Part 3: Test methods
BS 8568:2013; Cleanroom energy. Code of practice for improving energy efficiency in cleanrooms and clean air devices.
Other posts about clean rooms – A basic clean room, clean room explained in simple terms, 15 things you should never see in a clean room, 12 deadly clean room sins, what is your clean room costing you, optimising your clean room, getting QA buy in, now you know it all, take the clean room quiz.
Excellent Experiment well done PharmOut Team ! Next step in evolution
BMS (HVAC Control) & EMS (Cleanroom monitoring) working on feedback
loops to automate and allow for shutdowns overnight and weekends ! Personally
been keeping an eye on this application. An irish study conducted a few years
back estimated the cost reductions I will post this white paper when I dig it up !
Ravindra Kumar Sharma