Powder Formulation Efficiency Improvements

Team: Spencer Deveraux, Carston Gerlach, Quaid Goodey, Matt Jensen, Jax Stookey and John Watson

Problem Description

Cytiva is seeking an improvement to their powder weighing process to reduce ergonomic strain on their operators. The current process has operators manually scooping powder, resulting in high ergonomic risk. The powder weighing system automates the weight measurement process by using a hopper and dispenser integrated with a Mettler-Toledo scale. The scope of the project includes powder measurements from 2 kg to 20 kg.

Design Overview

Powder Dispenser System Diagram

Custom-made powder dispenser

  • Hopper
  • CORA Valve
  • Control system
    • Raspberry Pi
    • Motor Driver
    • Power Supply
    • Mettler-Toledo Scale
  • Cart
CORA valve CAD rendering
Control System

Hot Swappable Design:

  • Removable liner allows hopper to be reused between powders
  • Each powder has own CORA valve per formulation
Digital Display Interface
System requirements, constraints, goals and measured performance
Requirement/Constraint/Goal Threshold Target Measured Performance
Powder Measurement Accuracy 2 kg ± 10 g
20 kg ± 100 g		 2 kg ± 10 g
20 kg ± 100 g		 2 kg ± 52.60 g
20 kg ± 59.83 g
Cleanability CIP: 3 hours CIP: 2 hours 15 minutes
COP: 3 hours COP: 2 hours 2 hours
Labor Reduction 2 kg/min 3 kg/min 2 kg: 6.11 kg/min
10 kg: 10.78 kg/min
20 kg: 10.97 kg/min
Space Constraint < 6'x5'x8' < 6'x5'x8' 3.5'x2.5'x7'
Safety 4.75° 5.93° (25% SF) 6°

Analysis and Testing

Safety Test: The system was tested on a 6° incline in the fully loaded (worst case) condition.

Cleaning Time Estimation: A mock cleaning test accounting for all steps of the cleaning process was performed in 15 minutes.

Setup Time: A mock system setup and teardown test was performed in 20 minutes.

Powder Measurement Accuracy: Fifteen tests of powder measurement accuracy tests were performed at three weights.

Powder measurement accuracy test results
Target Weight (g) Dispensed Weight (g) Dispensed Weight Error
2000 1947 2.70%
10000 9967 0.33%
20000 19940 0.30%

Powder Dispense Rate: Fifteen tests of powder dispense rate were performed at three weights. Testing entailed dividing total weight dispensed by total fill time.

Powder dispense rate test results
Target Weight (g) Fill Rate (kg/min)
2000 6.11
10000 10.78
20000 10.97

Lessons Learned

Design iteration helps to dial in on what is most important.

System prototypes can identify which paths to pursue

3D printing is useful for fast prototyping, but not strong enough for structural components

When dealing with abrasive powders, choosing an overpowered motor is best

Photo of green gear component
Motor Driver Control Box

Design Satisfaction

Powder Measurement Accuracy: Lower dispense volumes resulted in higher error due to the discrete volume of each CORA valve wedge.

Cleaning Time: The cleaning time for the system is estimated to be under Cytiva's standard procedure.

Labor Reduction: The system exceeds customer requirements for powder dispense rate, with faster rates at higher fill volumes.

Space Constraint: The system is within maximum bounds.

Cart Tipping Test: The system passed the test on a 6° incline, proving stability during transportation.

Future Work

Round Hopper Component
  • Final design to use round hopper
  • Final design to use real CORA valve, not 3D printed
  • Transfer code system from Raspberry Pi to CORA proprietary system