Validating and Testing a Fluid Bed System--Part 2
Validation Group Manager, Vector Corp.
Back to Part 1
Snags and Snafus
This installation presented several challenges on validation and testing:
Pumping System
The solution was being batched using a <%=company%> flowmeter connected to a peristaltic pump. We were concerned about calibration for this pump/meter combination because the pump was supposed to be installed at the lowest point in the piping system, but in fact it had been installed at the highest point, and because of its pulsing flow. As it turns out, we were able to measure delivery to +/- 1 percent. Another problem: Floor scales were being calibrated around this new facility so several scaling systems had to be freshly calibrated to obtain the necessary flowrate and totalizer data. The peristaltic pump's hose blew on the last batch and had to be replaced, necessitating re-initiation of that test.
Personnel Availability
Training took place in segments, since not everyone was available at any given time during the two weeks Vector was on site.
Data Handling
The system uses a network LAN for communicating data to a central computer. Customer requests for re-configuration of this system and additional capabilities made our original test plans obsolete. Test plans had to be created and modified from the new configuration on the fly from specifications generated only days before. This resulted in some failed tests, misunderstandings, and general frustration as these tests needed to be done quickly due to their production requirements.
Supervisory System
The supervisory system providing the necessary batch reports had to be validated separately from the fluid bed system as well as integrated with it. Since the supervisor system was in a separate part of the plant, we had to use walkie-talkies to confirm the accuracy of the supervisory station.
Other Obstacles
The customer's QA department had not had a chance to review the documentation before conducting the tests. Two days into the validation, we had to break off and answer audit questions regarding the test plans. Also, deionized water for humidity control was not installed, and the humidifier system was extremely slow and had to be replaced by a "clean steam" unit. This unit was unavailable during routine validation and had to be separately re-tested.
The Good News
Factory testing the fluid bed using the prepared validation protocols reduced the number of problems we encountered. Calibration data on the equipment was available and current, so no re-calibrations were necessary. The room was presented in the same manner as it would be during regular processing. This helped answer GMP questions, location, and operator training issues. The excellent condition of installation wiring and documentation (tagging, labels, etc.) greatly assisted in identifying and testing the individual components and later troubleshooting.
Conclusions
All problem issues were addressed to the customer's satisfaction, including their QA department. Because we performed a complete validation of the system, all problems encountered were caught before running actual product. This approach to fluid bed validation, while time-consuming and costly, provides assurance that subsequent production runs on high-value material will proceed smoothly. In this instance the products to be run on the fluid bed cost, on a per batch basis, almost as much as the entire cost of validation. Therefore processing of two successful batches more than compensates for the time, effort, and cost of validating the fluid bed. By implementing validation procedures and structure protocols we can reduce warrantee costs by at least a third and reduce startup times for installation of new fluid beds from six weeks to two. Thorough validation of process equipment prevents damage to equipment and loss of valuable product, thereby saving manufacturers money.
For more information:
Don Rosendale, Validation Group Manager,
Vector Corp.,
675 44th St., Marion, IA 52302.
Tel: 319-377-8263, ext. 322.