How to Separate Walnut Shells and Kernels After Cracking

After cracking, walnut processors need to solve two different problems: remove shell efficiently and protect the kernels that carry the selling value. The strongest solution is usually a staged process that combines mechanical separation, infrared shell-kernel identification and AI optical sorting—not a single machine expected to correct every upstream problem.

Step 1: Stabilise the feed after cracking

Cracking creates a mixture of whole kernels, halves, pieces, shell, septum and dust. Large size differences and loose dust can affect the following separation steps. Screening, aspiration and controlled feeding should remove oversized material and free dust before the optical sorter sees the product.

Step 2: Use the right technology for shell and kernel

WESORT’s walnut solution recommends an infrared shell-kernel separation machine as one of the walnut processing options. Infrared material recognition is useful when shell and kernel need to be distinguished by material response, not only by visible colour. This stage can reduce the shell load entering final quality sorting.

For visible quality decisions after shell removal, WESORT’s AI walnut selection machine adds colour, shape, size, texture and area analysis. It can be programmed for black kernel, shrivelled kernel, discolored kernel, shell, septum and other visible targets using the customer’s own samples.

Step 3: Separate grades instead of creating one waste stream

A shell-and-kernel line often contains valuable secondary product. If all non-premium pieces are ejected together, a plant may lose usable kernels with the shell fraction. WESORT’s AI walnut machine is designed around multi-grade output: the published product information describes a nine-grade selection design and a return outlet. The actual outlet map should follow the processor’s sales grades.

What causes poor shell removal?

Common causes include excessive cracking, unstable feed, shell fragments attached to kernels, dust, overlapping material and a grade rule that has never been written down. A supplier should inspect the actual mixture and discuss the complete line—not simply quote a machine based on the word “walnut.”

Why WESORT is valuable after the cracking stage

WESORT’s published walnut equipment uses AI deep learning, intelligent CCD imaging, magnetic levitation valves and stable vibration feeding. The walnut selection machine also includes APP remote control and online monitoring features. These functions help the plant keep recipes, operating checks and technical communication connected after installation.

How to design the test