(201b) Method for Characterization of Loss-in-Weight Feeder Equipment
In pharmaceutical manufacturing, powder cohesion causes large variability in the flow rate of ingredients fed by screw/auger dispensing feeders. This can pass problems of composition and flowrate variability to subsequent unit operations in continuous manufacturing. For instance, if powder flows from multiple feeders into a continuous mixer have significant flowrate variability, than this can lead to undesirable and inconsistent blends being discharge by the mixer regardless of mixer performance.
Loss-in-Weight feeders have improved the ability to control feedrate and minimize flow variability due to density changes associated with the emptying of the feeding hopper (Hopkins 2006}), which is helpful once a feeding system is setup. Unfortunately, the selection and setup process for a feeding system is highly based on experience and knowledge that is not readily available. For loss in weight feeders, most of the knowledge regarding to either (i) the effect of powder properties on flow rate intermittence, or (ii) the effect of feeder design and operation on powder properties resides with the equipment manufacturers. There has been some published work on improving feeder performance by using devices at the discharge (Kehlenbeck and Sommer 2003), but actual specification and sizing is lacking. Feeder tooling (screw, discharge screen, etc.) selection currently is heavily based on trial and error, and there has been little work focused on optimizing the feeding of materials.
In this work a feeder characterization testing plan has been designed which uses a Schenck Accurate AccPro II as a ?catch? scale in order to monitor powder fed from the loss-in-weight feeder. Through the use of different powders pharmaceutical powders of various cohesiveness and flowability as well as the various configurations of the feed tooling (screw, screens, etc.) the feeder can be characterized for optimal performance. Flowability of each powder is quantified through the use of the flow index measured from a Gravimetric Displacement Rheameter (GDR) and dilation in a simple drum tumbler.(Alexander et al, 2006; Chaudhuri et al, 2006; Faqih et al, 2006)
A method for filtering and analysis has been developed to optimally remove noise from the experimental signal while leaving actual signal variability intact. An Analysis of Variance (ANOVA) will be used to investigate the different sources of variability caused by the process variables (powder cohesion, screw speed, screw type, and screen) and their impact on feeder performance. Feeder performance is defined as the feeder's ability to maintain a consistent powder feedrate at a desired setpoint.
For this work, a KTron KT35 twin screw Loss-in-Weight feeder for pharmaceutical applications has been evaluated using specific characterization techniques. A parametric study of the 3 screen conditions (no screen, coarse square screen, and fine square screen), 4 paired sets of screws (coarse concave, fine concave, coarse auger, and fine auger) and the 3 powders (Foremost 316 Fast Flo Lactose, FMC Biopolymer Avicel PH-102, and Asahi-Kasei Ceolus KG-802) at 3 feedrate setpoints has been evaluated. It was found that all investigated feeding process variables had a significant impact on feeder performance. The experimental data was further investigated to determine the significance of each process variable.
A large range of consistent feed rates can be achieved out of a single feeder simply by changing the screws and screens. Adding the screens enabled the feeder to feed more consistent flow, but it also lowered the maximum controllable feed rate. Reducing the size of the screen holes further reduced the maximum flowrate, but at the same time improved performance and consistency. For all of the combinations of screw and screen in the screw speed range of 20-80% there was the same observation of FastFlo Lactose (the most free flowing powder) having the best performance and Ceolus (the poorest flowing powder) having the poorest performance with Avicel 102 falling in the middle. It is also shown that by using the flow index and dilation numbers for each of these powders it is possible to correlate flowability and feeder performance.
Alexander AW, Chaudhuri B, Faqih A, Muzzio FJ, Davies C, Tomassone MS. 2006. Avalanching flow of cohesive powders. Powder Technol 164(1):13-21.
Chaudhuri B, Mehrotra A, Muzzio FJ, Tomassone MS. 2006. Cohesive effects in powder mixing in a tumbling blender. POWDER TECHNOLOGY 165(2):105-114.
Faqih A, Chaudhuri B, Muzzio FJ, Tomassone MS, Alexander A, Hammond S. 2006. Flow-induced dilation of cohesive granular materials. AICHE JOURNAL 52(12):4124-4132.
Faqih A, Chaudhuri B, Alexander AW, Davies C, Muzzio FJ, Silvina Tomassone M. 2006. An experimental/computational approach for examining unconfined cohesive powder flow. Int J Pharm 324(2):116-27.
Hopkins M. 2006. LOSS in weight feeder systems. MEASUREMENT& CONTROL 39(8):237-240}.
Kehlenbeck V and Sommer K. 2003. Possibilities to improve the short-term dosing constancy of volumetric feeders. Powder Technol 138(1):51-6.