Overview

The machine seen on the right is NxStage’s System One at home dialysis unit. A dialysis machine functions by pumping patient blood on one side of a membrane and a dialysate fluid on the other. This process removes impurities from the blood, acting as an artificial kidney. This device utilizes peristaltic pumps; this pumps move fluid by compressing a flexible tube with a series of rollers.

Problem

During my co-op, the company was interested in studying why they were seeing variability in the fluid volume output of their pumps. When first turned on, the pumps start at sub-par performance and work their way up to a steady state peak performance. I was tasked with finding what variables affected the time it takes for the pumps to reach that steady state performance.

Action

After conducting a root cause analysis, 3 variables were deemed worth investigating: pump tube temperature, pump roller speed, and fluid temperature. A high, low, and nominal condition was then set for each variable. An exploratory DOE style investigation was then conducted, testing every permutation of variable conditions.

The data that was collected was then analyzed in Minitab to search for significant differences across the permutations. This was a pilot study with the goal of informing more in depth investigations in the future, so the sample size was limited.

Result

This limited sample did however show correlations for which variables had positive and negative impacts on performance. This study also had an interesting side effect in that it suggested that the start up procedure for the pump could be optimized. The normal startup for the pumps involve running the pumps at maximum speed to reach that steady state as fast as possible, but this investigation suggested that under certain conditions it is actually more beneficial to run at the minimum speed during start up. More data would be required to make any concrete assessments, but the investigation was successful in informing future experiments.