1、NA-04-7- 1 Using Certification Data of Cleanrooms to Determine Degradation and Retrofit Susan D. Morrison Member ASHRAE ABSTRACT What should be looked at in determining the retrofit of a cleanroom? This veT expense environment andpiece of infa- structure has a history or sewice life cost to its exis
2、tence. This history and case study of a cleanroom will show some of the overlookedparameters that should be used in this type of eval- uation, which are often overlooked by engineers. When we, as engineers, design a cleanroom and do the initial commissioning, it is probable that we will never see th
3、is cleanroom again. But if we are lucky enough to go back to a cleanroom that is having problems, we can try to find the “fly in the ointment,” so to speak, as I was able to do in the summer of 1999. One of the first questions we tend to ask ourselves is, “Where do 1 begin? It is important to rememb
4、er that many of us design, build, and certify cleanrooms. Yet we know little of the testing and validations as well as modifications that have taken place since our original involvement. Keeping this in mind, if we went right back to the original test documents and then requested all subsequent cert
5、ification tests conducted since that time, we might see when problems started to develop. Having done the initial, design, construction, and commissioning of this cleanroom complex, and having set the baseline, provided me insight as to what had happened, whether an overhaul was needed, and how to c
6、orrect it if needed. The first things needed were the original performance matrix and original test report for this facility. It was critical to know the performance baseline of the facility at commission- ing, including the original heat loads and particle generation levels in the at-rest mode. Thi
7、s was also a unique cleanroom, as it had been designed with a night setback. All of the tests had to be performed in the unoccupied mode to prove the environ- ment could sustain the metrology when we decreased the airflow to 50%. The cleanroom requirements utilized certifi- cation parameters, which
8、included more than just particle counts. These items should not be confused with control commissioning or HVAC commissioning and are parameters that are decided upon by the customer and designer for the interior of the cleanroom. These cleanroom requirements included lighting levels, which were to a
9、verage 100 or 70 foot candles, on average, depending on area; sound, which was not to exceed 80 decibels; airflow that was temperature of 22C; and humidity ofnot less than 30%. Pressure was decided to be a minimum of 10 to a maximum of 30 Pa between zones, and the air change rate was between 20 to 8
10、0 ach, based on the class of the room. Lets look at the data from some of the rooms from the initial commissioning throughout the development of operational problems. When we look at these data, it is useful to remember that these cleanrooms are unlike a standard cleanroom in that the heat load was
11、balanced against the makeup air for pressuriza- tion and the process exhaust; these factors determined the capacity of the coil. In these rooms, heat load was very tightly balanced and the units could be kept within their size limita- tion requirements. Heat load-both latent and sensible- generated
12、in the room was handled by the return retreating the air as well as process exhaust. Makeup air had to be treated to replace air consumed for process exhaust and pressurization; this volume of makeup air that flows through the coil only once needs to be conditioned in this one pass to the required t
13、emperature and humidity for the cleanroom. The volume of air needed to flow through HEPA filters in order to maintain the classification of the cleanrooms determined the recircu- lated air volume (see Figure 1). When we factor this all in to one room, we have a very tightly balanced metrology. That
14、results in a relationship in which very small changes to process loads will have a large effect on the overall performance of the cleanroom, not only in heat load, but also potentially in parti- cle generation and the retention of these particles in the airstream. Susan Morrison is facilities safety
15、 manager at Paratek Microwave, Columbia, Md. 02004 ASHRAE. 755 Table 1. Excerpt from Occupied Room Matrix I I Make Up AJr -D!schaqe- -Return Arc-. I t Figure I Air-handling schematic. Understanding this basic design concept and the matrix of the design (see Table i), we can then begin to look at wha
16、t has occurred to the cleanroom when we graph these parameters. One of the first parameters we will begin with is the airflow while the room is occupied (see Figure 2). The first thing that has to be established under IES-RP-CC-006-84-T Sections 3 and 4, which deal with air flow uniformity, is to me
17、et *20% uniformiy of the design airflow. This percentage of uniformity between individual filters in a room must be achieved as agreed upon by the buyer and seller. As shown in Figure 2, at the time of initial certification, we fell well within these parameters. The difference between total airflow
18、that was tested in 1995 and design was *20% or less of design. We should look at the deviation not only as shown in Figure 2 but also filter- to-filter. If we take a look at the data in Figure 2, the first thing we notice is the strange failure of the airflow total of Rooms 9 and 10 in 1998. If we w
19、ere to look at the uniformity from HEPA filter to HEPA filter, we would find uniformity between these filters at less than *8%. But why reference and use an out-of-date IEST standard? It was out of date at the time this project took place, but it was what the customer wanted and demanded! When we lo
20、ok at a project, it always helps to look at what was required in the contract at the time it was constructed, not what it is today. Sometimes, in an evaluation of an existing space, we neglect to find out how the space was intended to work and apply standards based on todays knowl- Air Flow 6000 500
21、0 gDesign ml995 01997 Y 1998 z 4000 2 3000 D g 2000 Il999 1000 O 14 131211 10 9 2 8 7 5 4 6 Room # Figure 2 Occupied airflow. edge without regard for the original design intent. It is impor- tant to remember that there are many recommended standards and practices out there, and some customers want w
22、hat they want even if you tell them it is out of date-a good example of this was IES-RP-CC-006-84-T, which was used on this job site. They wanted all air in cubic meters an hour and particles in concentration per cubic foot of air. This can get very confus- ing, but this type of mix exists out there
23、, and it is necessary for us to maintain a library of these items for reference when reviewing certification data for analysis. Lets look at the particle counts (see Figure 3) for the actual locations from the same time periods to see if there is a correlation. We find that the particle counts were
24、higher in those rooms by as much as 250% in 1998 compared to the 1997 and 1999 tests. These particle counts were below the limit estab- lished by Federal Standard 209E of 10,000 particles at 0.5 um per cubic foot of air. It was curious to see what could have happened between 1997 and 1999. When we l
25、ook at the graph in Figure 3, the room would still pass certification, but it did not do as well as the year prior to or after it. What could have happened? Maybe we should look at some other parameters to see what might have caused this deviation. If we then looked at sound, we would find that the
26、sound levels in those rooms were, on average, 20% higher in the 1998 report than in the 1997 report. What correlation does using sound have? Remember, if something gets noisy, there must be 756 ASHRAE Transactions: Symposia Room 9 810 Partiile Counts imn ,i - 1 . - - L 1200 ? - 1000 9 800 r? % q 200
27、 s O 600 400 O 9 Room # 10 Figure 3 Particle counts in suspect rooms. Temperature 4 6 Room # Figure 4 Temperature. a reason for it! In checking the maintenance records for this site, we found that the motor bearings were replaced in late 1998-before the 1999 certification took place. In checking wha
28、t kind of motor and fan this site was using, we found that the GE motors used have a specified lubrication life for the bearings. In that application they would have a bearing lubri- cation life of 1.5 years (GE Motor Manual). Information from this manual is as follows from Table 19, “Severe Motors,
29、” for fans that run 24 hours a day, at 10-40 HP, and have re-lubri- cation intervals of 1.5 years. These bearings were sealed in this model of GE motor, and they had been in continuous operation since the summer of 1995. In early 1998, the noisy motor was tripping on thermal overload and was slowed
30、down until repairs could be accomplished. Unfortunately, this took over nine months to accomplish. So again we come back to yet another item to review, which is the service record for this facility. Have the prefilters been changed on a timely fashion? Have ail changes and major repairs been logged?
31、 Has anyone changed the setting on the airflow or motor RPM and so on? What do we check these against? Use the TAB data from the original installation or something looked at even less that should become part of the initial certification package-the “birth certificate”-of the equipment. When units ar
32、e initially started, someone has set it up. They shouldnot have checked its operation before the TAB and initial commissioning was started, but generated a start-up certificate. When we compare this to the TAB report, we find the data for fan amperage, fan speed, discharge air tempera- ture, or make
33、up air temperature, which can be used to track or compare to diagnose a failure or, in some cases, predict a fail- ure. Temperature can always become a problem in a tightly balanced system. Referring back to Figure1, ifwe were to look at the fact that we treat the makeup air and enough of the return
34、 air to overcome the process load-both latent and sensible- you can see by the basic design that it does not take much to create a temperature failure or a humidity failure with changes that are based on the statement, “We had enough space to add more equipment!” Change order control used during the
35、 facil- ity life cycle can prevent this, as someone should review what is being added to discover the load change and impact to the room. Lets look at a room in the same area that was fine on particles but started to exceed design temperature (see Figure 4). When we looked at these two rooms, we not
36、iced that they were out of the temperature tolerance of 22”C*2”C. We had found the “we had room so we put it in” syndrome. Now the ASHRAE Transactions: Symposia 757 question was what to do! They needed the equipment. We then looked back at the original design documents again. After inspection, we fo
37、und we could increase the return air and decrease the recirculated air to return the room to the proper metrology. When we look further we should look at lighting levels at the same time, since we know fluorescent light degrades over time. We decided that since we were opening the ceiling and rebala
38、ncing anyway, it was a good time to re-lamp and re- ballast the entire cleanroom. This was done to save any further downtime on the facility since, over the five-year period, the average light levels had dropped between 7- and 15-foot candles in each room. Again we used certification data and mainte
39、nance records to make an informed decision to save downtime to the facility and save costs that would have been required for a second cleaning and certification if this became a necessary repair during the intervening period. The same data as listed in the above, such as particles and airflow, could
40、 be graphed over the unoccupied period and show the same data trend. More importantly, the unoccupied setback allowed the facility to save money in the overall utility cost. This savings more than justified the second full set of certification tests that needed to take place. But again came the stat
41、ement, “It cannot be setback and still be a cleanroom.” This input came from several of the vendors consulted by the customer. Why? Can it not still be certified as a cleanroom if all of the testing parameters can be met (all except perhaps air change rate)? The answer comes back again to look at th
42、e commissioning document to determine the operation of the setback and the unoccupied matrix for the room. Though many of our colleagues will write a sequence of operation, many more do not understand the actual hardware used by control companies or how to diagnose problems when they occur. Many of
43、the problems with the airflow of Rooins 9 and 1 O were blamed on this. Yet this customer had an actual validation package for the cleanroom, and it was easy to look at the historical trend data and find the operation parameters for the room. Historical trending-if it is available-is a great diag- no
44、stic tool. We could keep reviewing this and graphing data, but I think by now the point is well taken. This cleanroom did not need a total overhaul but some good old-fashioned engineer- ing. But the above discussion shows that from our involvement in initial design and construction of the cleanroom,
45、 our work does not end at the initial commissioning. We need to help the client develop a strong document control plan, a good service and review plan, and a life-cycle change order plan to maintain the metrology of this environment. But most of all we need to learn how this system is working in fac
46、t, not only in theory. The final conclusion is that by working with the certifi- cation test data and maintenance data and not just depending on theoretical norms, we can provide the information to our customers to make an informed decision. Many people had told customers that the rooms would need t
47、o be overhauled and have all the HEPA filters replaced to remove the discrepancies and problems. What a cost to the customer! We need to become more familiar with the test data we have and how it relates to the initial design. We need to become more familiar with the work done by the mechanics and t
48、echnicians in the field on the system. This room did not need to be overhauled, but it did need some minor rework for the year 2000 to restore it to back to its metrology. We also need to become familiar with codes and/or standards used in the TAB and commission- ing of cleanrooms, as well as inform
49、ation that may be there that we dont use in such evaluations. We need to follow other groups best practices and see what we can learn from them. Use of that information will let us make informed decisions to help the customer reduce their service life costs. Reducing downtime will automatically increase their productivity. Through this life-cycle management approach to cleanroom problem solving and evaluation, you not only ensure that the customer saves money, but you will have helped save valuable resources! I leave you with something I was told by an engineer who trained me t
