The Problem

nquiringminds have been working with the Longwood Estate in Hampshire applying IoT sensor technology to the grain harvesting process. The task is a complex challenge as there are several variables to account for which affect the process. The agricultural environment is just one of the many applications of the nquiringminds IOT platform, and this case study is a good example of the flexibility of the technology and the robustness of our sensor technology in a challenging environment with legacy equipment demonstrating the true open source nature of our solution

IOT Sensors and recipe

  • nquiringminds custom made robust (high temperature tolerance) temperature sensors QBLOX
  • nquiringminds custom made robust (high temperature tolerance) Humidity sensors QBLOX
  • LoRa network for long range data transmissions
  • Nquire trusted data exchange (TDX)
  • Nquire ToolBox

 

Scenario

Grain drying is seasonal activity that takes place every year after harvest. Grain is sold by weight and the more moisture in the grain the greater the profit for the farmer. Because of this the buyers set a maximum moisture threshold above which the seller is heavily fined. So drying grain is essentially and optimisation problem to get the moisture content as close to the threshold as possible and achieve maximum yield on the product.

To make things more complicated the optimal moisture has a number of confounding variable including the grain type and of course the background temperature and humidity.

A grain drier is a harsh dusty environment with temperatures reaching over 100oC. The drier itself, is metal with poor radio transmission qualities and the control system can often be some distance from the drier itself.

To overcome these issues we constructed our own robust high temperature sensors coupled with thermodynamically resistant wire to an Atmel powered control board. Onto this control board we mounted some long range LoRa radio chips which transmitted status regularly to a computer in the central station.

All data is fed into our TDX data store, where both the historical and real time data can be analysed and visualised with the NQuire Toolbox.

We make extensive use of machine learning functions resident in the tool box, using the sensor data as inputs to optimise the drier temperature and speed at which the grain moves through the drier and maximise harvest yield.

  • With grain prices in the hundreds of pounds per tonne a farmer who harvests 10,000 tonnes in a year could (conservatively) increase their profits by £40,000 per year by retro fitting a sensor set and running the data through the NQuire toolbox.

 

  • Not only will farmers sell their grain for more, they will save on the fuel costs for drying the grain. The UK grain, seed and bean harvest is approximately 250 million tonnes a year. This would work out at approximately £60 million saved per year for each percent closer to the threshold the grain

 

  • Which is equivalent to a 1/3 billion kg of CO2 emissions per year for each percent closer to the threshold.

 

Outcome

We had to make sure that the grain was kept fractionally below the acceptable moisture level, to assure maximum profitability for the estate, and our sensors needed to be robust enough to cope with the harsh farming environment. By applying real time sensors and machine learning it is possible to confidently dry the grain to just below the threshold. And sell the grain for several pounds a tonne more than would otherwise be possible. With one grain dryer processing thousands of tonnes a harvest the additional revenue can quickly add up. The value we brought to the project was knowing how to operate an IOT network in extreme conditions of up to 100 degrees centigrade and then apply machine learning and real time responses to a dynamic system. As with everything we do the platform is hardware neutral and so can be easily integrated onto any existing equipment.

Download as PDF

Download