Grain-drying optimisation for agriculture using IoT sensing and machine learning

The Challenge

Grain must be dried to fractionally below the acceptable moisture threshold before sale: over-dry it and the estate wastes energy and sells lighter grain; leave it too moist and it fails the threshold entirely. With one drier processing thousands of tonnes per harvest and grain worth several pounds a tonne more when dried precisely, the difference adds up quickly — but the drying environment (up to ~100°C inside industrial driers, wind and rain in the field) is hostile to conventional sensing, and expert drying knowledge is scarce across simultaneous harvests.

The Solution

NquiringMinds developed a suite of agritech sensors gathering key harvest-data metrics across the full crop life cycle — sowing, growing, harvesting, drying, storage and delivery — combined with state-of-the-art machine learning and real-time responses to a dynamic system. Sensors operate in harsh environments with robust long- and short-range transmission and seasonal deployments lasting up to a year without battery change. Interactive cloud-based visualisations, analytics and control systems reduce energy consumption and environmental impact, and share scarce expert knowledge across multiple simultaneous harvests via remote intervention and machine-learned “expert knowledge”. Sensor hubs were built on the ST STM32 platform, and the platform is hardware neutral so it integrates onto existing plant and machinery using open-source and IoT standards.

Outcomes

The project delivered a working IoT network operating at up to 100°C with machine-learning-driven drying control, enabling grain to be dried confidently to just below the moisture threshold for maximum profitability. It raised the technology from TRL 2 to TRL 4 and was pivotal in unlocking follow-on work, including India rice drying and storage pilots with Brunel University that built directly on the Agrisense approach.

In Detail

Operating an IoT network in extreme conditions to dry grain. Data analytics and IoT devices are powerful tools for reducing spoilage and increasing yields in agriculture. Processes that rely on periodic human inspection and timely interventions fail frequently because of resource constraints, lack of expertise and the speed at which crops can be spoilt. Agrisense pairs harsh-environment sensing — surviving industrial driers at up to 100°C, wind and rain in the field, and year-long battery deployments — with machine-learning control that dries grain confidently to just below the moisture threshold for maximum profitability.

Agrisense is as much an industrial control story as an agritech one: NquiringMinds’ multi-award-winning platform has been applied to many industrial applications, including control systems for agricultural grain driers, with an emphasis on security, connectivity and analytics that makes it ideal for demanding industrial problems.