Modeling of Cross-contamination of Salmonella during Almond Processing

Introduction: Although low-moisture foods have been believed safe, recent outbreaks have brought safety concerns over how low-moisture food contamination occurs and how the contaminated foods travels through a system and are affected by the environment. To understand the nature of cross-contamination in low-moisture food products, first-principle models can be used to understand the interactions between dry particles and bacterial pathogens.

Purpose: The objective was to develop a first-principle based discrete element model of bacterial transfer and to assess the model performance.

Methods: Utilizing a Discrete Element Modeling software (LIGGGHTS: LAMMPS Improved for General Granular and Granular Heat Transfer Simulations), 200 almonds in a rotating drum being mixed with 5 inoculated almonds in a low-water activity environment at a rate of 8 rpms was simulated. The first model uses an analogy of heat transfer mechanics to bacterial transfer. Using a more fundamental approach, the second model depicts the electrostatic force between a simulated bacterial particle attached to a simulated almond particle. The Hamaker Constant of the Van der Waals equation for electrostatics was optimized to simulate realistic Salmonella adhesion properties. Both simulations were validated against experimental results of the same environment.

Results: Comparison between experimental results and simulation results using heat transfer as an analogy demonstrated an accurate representation of the system, with RMSE = 0.070 log(CFU/g). Comparison of results for electrostatic bacterial modeling and experimental results yielded RMSE = 0.108 log(CFU/g). Despite being less accurate of a fit, the electrostatic model provides fundamental insights into the mechanics of cross contamination and direct representation of the phenomena.

Significance: A first-principle based discrete element model can be used as a useful tool to elucidate the mechanism of bacterial transfer in low-moisture environment. Ultimately, the model will be used for scale-up validation and risk modeling to design safe process.