Abstract

In food processing of potato chips, one of the first unit operations is a peeler, which removes peel from the potatoes. The peeler consists of rotating cylindrical abrasive brushes that remove the material from the surface of the potato and an adjustable gate that controls the bed height and residence times of the potatoes in the peeler. When the gate position and brush rpm are controlled manually, process variations in potato type, size, and throughput can cause the system to remove too little peel, which results in a poor-quality product, or over-peel the potatoes, which removes potato flesh (or pulp) reducing productivity. Therefore, automating control of the process to maintain a consistent peel level would improve product quality and productivity. To develop automation of this peeling system, a digital twin of the peeler was created using discrete element method (DEM), and was used in combination with a machine vision system, that measured peel level. In the DEM simulation the potatoes were modeled as ellipsoids with a size distribution determined from the online camera. An equation for the material removal rate (MMR) from the potato was developed, with a dependence on the normal pressure applied to the potato when in contact with the brush and the relative velocity between the potato and the brush. The MRR and the distribution of forces and velocities acting on the potatoes over time were taken from the simulation and used to calculate the amount of peel left on the potato as well as the amount of pulp loss. The simulation was run at different operating conditions using data from the line including throughput, potato size, density, and peel level, and was used to create a surrogate model that could predict peel level and pulp loss in real time. This reduced model equation was used to create a virtual sensor for pulp loss to monitor and validate productivity savings for the system and used to create control algorithms. This study was funded by PepsiCo. The views expressed in this abstract are those of the authors and do not necessarily reflect the position or policy of PepsiCo, Inc.